&EPA
United States Industrial Environmental Research EPA-600/8-83-009
Environmental Protection Laboratory April 1983
Agency Research Triangle Park NC 27711
Research and Development
Control Technology
Appendices for
Pollution Control
Technical Manuals
! >- f
-------
EPA-600/8-83-008
April 1983
CONTROL TECHNOLOGY APPENDICES
FOR
POLLUTION CONTROL TECHNICAL MANUALS
Program Manager: Gregory G. Ondich
Office of Environmental Engineering and Technology (RD-681)
U.S. Environmental Protection Agency
401 M Street, SW
Washington, DC 20460
Project Officer: William J. Rhodes
Industrial Environmental Research Laboratory-RTF
Research Triangle Park, NC 27711
-------
NOTICE
This document has been reviewed in accordance with
U.S. Environmental Protection Agency policy and
approved for publication. Mention of trade names
or commercial products does not constitute endorse-
ment or recommendation for use.
11
-------
PREFACE
This collection of appendices supplements the Pollution Control Technical
Manuals (PCTMs) by providing more detailed information on individual pollution
control technologies. Additionally, these appendices present the references
from which PCTM information on control technology performance, applications,
and cost was derived. Acid gas removal processes, depending upon their speci-
fic designs, can serve both process and environmental concerns; thus, they are
also included with the gaseous appendices.
These appendices are arranged in three groups. Information on gaseous,
aqueous, and solid waste control technologies can be found in Appendices A, B,
and C, respectively. Each control technology write-up is subdivided into sec-
tions on process description, applicability, performance, secondary waste
streams, reliability, economics, and references.
Users of the PCTMs who have an interest in a specific facility design and
site are encouraged to use these appendices as a means of understanding and
adjusting for differences between the PCTM illustrative applications examples
and their particular cases.
i i i
-------
ABSTRACT
The Environmental Protection Agency (EPA), Office of Research and
Development, has undertaken an extensive study to determine synthetic fuel
plant waste stream characteristics and to evaluate applicable pollution con-
trol systems. The purpose of the Pollution Control Technical Manuals (PCTMs)
is to convey this information in a manner that is readily useful to designers,
permit writers, and the public.
The Control Technology Appendices (CTA) is a supplementary volume that
describes the technical aspects of over 50 pollution control processes.
The Control Technology Appendices address the major multimedia (gaseous,
aqueous, and solid waste) control processes that are commercially available
and could be applied to synfuel plants to alleviate discharge problems. The
CTA Volume provides detailed information on the control techniques that are
specifically addressed in the PCTMs.
The control processes in this volume are grouped by media (gaseous,
aqueous, and solid waste). For each process(es) the CTA provides the follow-
ing information: process description, process applicability, process perfor-
mance, secondary waste streams from the control process(es), process reliabi-
lity, and process economics.
-------
CONTENTS
Preface ii
Abstract iii
Acknowledgement iy
Acronyms and Abbreviations vii
Conversion Factors x
Appendix
Number
AIR POLLUTION CONTROL APPENDICES
Amine Acid Gas Removal Processes Al
Rectisol Acid Gas Removal Process A2
Selexol Acid Gas Removal Process A3
Benfield Acid Gas Removal Process A4
Catacarb Acid Gas Removal Process A5
Clans Process A6
Stretford Process A7
SCOT Process A8
Beavon Process A9
Wellman—Lord Process A10
Cyclones All
Fabric Filtration Process A12
Electrostatic Precipitation Process A13
Venturi Scrubbing Process A14
Flares A15
Thermal Incineration Processes A16
Catalytic Incineration Processes A17
Fugitive Organics Control A18
Fugitive Dust Control Techniques A19
Flue Gas Desulfurization Processes A20
NO Control Technologies A21
WATER POLLUTION CONTROL APPENDICES
Gravity Separation Bl
Coagulation/Flocculation B2
Air Flotation B3
Filtration B4
Solvent Extraction B5
Wet Air Oxidation B6
Steam Stripping B7
Steam Stripping Ammonia Recovery: Phosam-W Process B8
Steam Stripping Ammonia Recovery: Chevron WWT Process B9
Activated Sludge Process BIO
Trickling Filters Bll
Rotating Biological Contactors B12
-------
CONTENTS (Continued)
Appendix
Number
Lagoons B13
Anaerobic Digestion B14
Activated Carbon Adsorption B15
Chemical Oxidation B16
Thermal Oxidation B17
Cooling Tower Oxidation B18
Chemical Precipitation B19
Ion Exchange B20
Ion Exchange Removal of Ammonia Using Clinoptilolite B21
Membrane Separation B22
Biological Processes for Removal of Reduced Nitrogen Species. . B23
Cyanide Removal by Polysulfide Addition B24
Chemical Oxidation of Reduced Nitrogen and Sulfur Species . . . B25
Forced Evaporation: Vapor Compression Evaporators B26
Deepwell Injection B27
Surface Impoundment B28
SOLID WASTE MANAGEMENT APPENDICES
Landfill Cl
Surface Impoundment C2
Land Treatment C3
Solid Waste Incineration C4
Chemical Fixation and Encapsulation C5
COST METHODOLOGY D
v i
-------
ACRONYMS AND ABBREVIATIONS
AAS air activated sludge
ACEAS activated carbon enhanced activated sludge
ADA anthraquinone disulfonic acid
ADIP Shell-patented diisopropyl amine-based acid gas
removal process
AGR acid gas removal
API American Petroleum Institute
BBF bias burner firing
BOD biochemical oxygen demand
BOOS burner out of service
BSRP Beavon Sulfur Removal Process
BV bed volume
CE Chemical Engineering
CM combustion modification
COD chemical oxygen demand
CPI corrugated plate interceptor
CRA compression-refrigeration-absorption
DAF dissolved air flotation
DEA diethanolamine
DIC direct installed cost
DIPA diisopropanolamine
DMPEG dimethyl ether of polyethylene glycol
DOE Department of Energy
EGR exhaust gas recirculation
EP extraction procedure
EPA Environmental Protection Agency
EPRI Electric Power Research Institute
ESP electrostatic precipitator
FBC fluidized bed combustion
FGD flue gas desulfurization
FGR flue gas recirculation
FGT flue gas treatment
Vll
-------
ACRONYMS AND ABBREVIATIONS (Continued)
GHSV gas hourly space velocity
GET Gesellschaft fur Kohle-Technologic
GT gas turbine
HETP height equivalent to a theoretical plate
HPC hot potassium carbonate
IAF induced air flotation
IBL installed battery limits
1C internal combustion
IERL Industrial Environmental Research Laboratory
kmol kg-mole
K-T Koppers-Totzek
LEA low excess air
L/G 1iqnid-to-gas ratio
LNB low NO burners
x
LNCS low NO combustor system
LPG liquified petroleum gas
MEA monoethanolamine
MDEA methyldiethanolamine
MLVSS mixed liquor volatile suspended solids
NO nitrogen oxides
NPDES National Pollutant Discharge Elimination System
NSPS New Source Performance Standards
OFA overfire air
ORD Office of Research and Development
OSC off-stoichiometric combustion
PAH polycyclic aromatic hydrocarbons
PCTM Pollution Control Technical Manual
PM premix burner
PNA polynuclear aroma tics
POM polycyclic organic matter
PSD Prevention of Significant Deterioration
PVC polyvinyl chloride
Vlll
-------
ACRONYMS AND ABBREVIATIONS (Continued)
RAP reduced air preheat
RBC rotating biological contactors
RCRA Resource Conservation and Recovery Act
RL reduced load
SASOL South African Coal, Oil and Gas Corporation. Ltd.
SCA specific collection area
SCOT Shell Claus Off-Gas Treatment
SCR selective catalytic reduction
SNG substitute natural gas
SNPA Societe Nationale des Petroles d'Aquitaine
SO sulfur oxides
SRC Solvent Refined Coal
TON Thermal De-NO
x
TDS total dissolved solids
TEA triethanolamine
TGT tail gas treatment
THC total hydrocarbons
TIC total installed cost
TOC total organic carbon
TSP total suspended particulates
TSS total suspended solids
VOC volatile organic compounds
VSS volatile suspended solids
W-L Wellman-Lord
IX
-------
CONVERSION FACTORS
1.0 kg [kilogram]
1.0 Mg [megagram (metric ton)]
1.0 kg/min [kilogram per minute]
1.0 m3 [cubic meter]
1.0 Nm3/hr [normal cubic meter
(at 273 K) per hour]
1.0 GJ [gigajoule]
1.0 MW [megawatt]
1.0 MJ/s [megajoule per second]
1.0 kWh [kilowatt hour]
1.0 MJ/Nm1 [megajoule per
normal cubic meter
(at 273 K)]
1.0 g/Nm3 [gram per normal cubic
meter (at 273 K)]
1.0 kPa [kilopascal]
1.0 kg mole
2.205 Ib [pound (mass)]
1.102 ton [short ton (2000 Ib)]
132.3 Ib/hr [pound per hour]
264.2 gal [gallon]
37.32 scfh [standard cubic feet
(at 60°F) per hour]
0.9479 x 10* Btu [British thermal
unit]
3.413 x 10* Btu/hr [British thermal
unit per hour]
3.413 x 10* Btu/hr [British thermal
unit per hour]
3413 Btu [British thermal unit]
25.40 Btu/scf [Btu per standard
cubic foot (at 60°F)]
0.413 gr/scf [grains per standard
cubic foot (at 60°F)]
0.00987 atmosphere
22.4 Nm3 (32°F)
Prefixes
T = tera = 1012
G = giga = 10*
M = mega = 10«
k = kilo = 103
-------
ACKNOWLEDGEMENT
Technical and background information for these Control Technology
Appendices for the Pollution Control Technical Manuals was prepared for the
EPA by the Environmental Division, TRW, Inc., Redondo Beach, California and
the Radian Corporation, Austin, Texas. TRW's effort was under Contract num-
bers 68-02-3647 and 68-02-3174, Task 90, and Radian's effort was under Con-
tract number 68-02-3137. The Project Managers were R. Orsini and C.
Shih for TRW and W. Thomas for Radian.
-------
APPENDIX Al
AMINE ACID GAS REMOVAL PROCESSES
The alkanolamines are the most widely used solvents for removal of acid
gases from gas streams. The two amines that have seen extensive commercial
use in gas purification are monoethanolamine (MEA) and diethanolamine (DEA).
Triethanolamine (TEA) was the first amine to become commercially available,
although its use has been largely displaced by other amines (1). Diisopro-
panolamine (DIPA) and methyldiethanolamine (MDEA) have both been used in the
ADIP process because of their selectivity for H^S in the presence of C02. The
removal of acid gases by absorption with MEA, DEA, TEA, DIPA, and MDEA are
discussed in the following sections.
1. Process Description
The basic flow scheme for all alkanolamine acid gas absorption processes
is shown in Figure Al-1. The gas to be purified is passed upward through a
tray or packed tower, countercurrent to the downward flow of the regenerated
lean amine solution. The rich amine solution from the bottom of the tower is
heated by exchange with the lean amine solution from the bottom of the regen-
erator. In absorption systems treating sour hydrocarbon gases at high pres-
sure, the rich amine solution is often flashed at an intermediate pressure to
remove dissolved and entrained hydrocarbon gases. The flash stage can be lo-
cated either before or after the heat exchange step. The flashed and heated
rich amine solution flows to the regenerator, where the absorbed acid gases
are removed by heating and steam stripping. The acid gases and steam which
come overhead from the regenerator are cooled with cooling water and then flow
to the reflux vessel where vapor—liquid separation takes place. Steam conden—
sate saturated with acid gases is returned to the regenerator as reflux. The
acid gases saturated with water vapor are sent to the sulfur plant. The lean
amine solution from the regenerator, after partial cooling in the heat ex-
changer, is further cooled with cooling water and then fed to the top of the
Al-L
-------
ABSORBER
AMINE
REGENERATOR
REFLUX VESSEL
TREATED GAS
FEED GAS
LEAN RICH
AMINE AMINE
RICH
AMINE
0
REGENERATED
LEAN AMINE
ACID GAS AND STEAM
STEAM
CONDENSATE
fte
—
ACID GAS
Figure Al-1. Basic flow scheme for amine process
-------
Appendix Al
Amine Processes
absorption tower to complete the cycle. Because of the presence of solid par-
ticles and degradation products, a portion of the recirculating amine solution
(typically 10 to 20 percent) is often sent through a leaf-type precoat filter
followed by an activated carbon filter for the removal of impurities.
The primary differences in amine processes are in the different types of
amines used and in solution concentrations. Because of differences in molec-
ular weights, higher amine concentrations (on weight basis) for secondary or
tertiary amines are generally required than for primary amines. Typical con-
centrations of MEA range from 12 to 25 percent, and a design concentration of
15 percent has been recommended (1). DEA solutions used for treatment of
refinery gases range in concentrations from 10 to 25 percent, while concentra-
tions of 25 to 35 percent are typically used for natural gas purification, as
in the SNPA-DEA process (1). The ADIP process employs relatively concentrated
aqueous solutions of DIPA or MDEA, typically 30 to 40 percent and higher
(1,2). TEA solutions, because of their lower acid gas carrying capacity (per
unit weight of TEA), are also used in higher concentration levels than MEA and
DEA solutions.
In the absorption tower, a primary amine such as MEA (represented by RNH2
where R is the alcohol radical; R is CH2CH2OH for MEA) or a secondary amine
such as DEA or DIPA (represented by R,NH; R is CH2CH2OH for DEA and CH2CHCH}OH
for DIPA) reacts with H2S to yield amine hydrosulfide (1,2):
H2S + RNH2 = HS~ RNH3+ (1)
H2S + R2NH = HS~ R2NH2+ (2)
C02 can either react directly with the primary or secondary amine to form the
amine carbamate or with water or hydroxyl ions to form carbonic acid or bicar-
bonate ion:
Al-3
-------
Appendix Al
Amine Processes
C02 + 2RNHa = RNHCOO RNH,"1" (3)
C02 + 2R2NH = R2NCOO~R2NH2+ (4)
C02 + H20 = HjCO, (5)
C02 + OBT = HCOa~ (6)
These acids then react with the amine to yield the amine bicarbonate
(HCOj'RNH,"1" or HCOj'RjNH^) and the amine carbonate (CO^RNH 3 + )2 or
» _f.
C03 (R2NH2 )2). Reactions 1 and 2 are fast, reactions 3 and 4 are
moderate, and reactions 5 and 6 are known to be slow.
Tertiary amines such as MDEA or TEA (represented by RjR'N; R = CH2CH2OH
and R' = CH3 for MDEA; R = R' = CH2CH2OH for TEA) cannot react directly with
carbon dioxide. The only possible reactions for tertiary amines are:
H2S + RaR'N = HS~R2R'NH+ (7)
and reactions 5 and 6. followed by acid-base neutralization reactions such
as:
H2COj + R2R'N = HCOj~R2R'NH+ (8)
Because reaction 8 is limited by the extremely slow reactions 5 and 6 and
reaction 7 is fast, tertiary amines exhibit significantly higher selec-
tivity for H2S. The ADIP process, for example, can be designed by employing
MDEA and gas—liquid contact time in the absorption tower that is sufficiently
long for the removal of H2S but only long enough for partial co-absorption of
C02.
Al-4
-------
Appendix Al
Ami lie Processes
The absorption reactions shown above are all favored at ambient tempera-
tures. The temperature of the lean amine solution entering the absorption
tower is typically 310 K. At elevated temperatures in the regenerator, the
reactions are reversed with the decomposition of sulfide, carbonate, and car-
bonate salts and the release of acid gases.
The alkanolamines have been used commercially to treat sour gases for
many years, and the number of installations has been estimated to be in excess
of 1,000 (3). With the exception of a few proprietary processes such as the
SNPA-DEA and ADIP processes, designs of most alkanolamine processes are avail-
able from a large number of engineering and construction companies. The SNPA-
DEA process was developed by Societe Nationale des Petroles d' Aquitaine
(SNPA) of France and licensed by the Ralph M. Parsons Company. It is cur-
rently in use to sweeten 140 million m3/day of raw gas (3). The ADIP process
was developed by Shell and licensed by Shell Development Company of Houston
for U.S. projects and Shell Internationale Research Mij. B.V. of the Hague for
projects outside the U.S. More than 150 ADIP plants are in operation or under
construction. This figure includes a number of large COS extraction plants,
but excludes the selective absorption plants in SCOT units.
2. Process Applicability
The MEA process is particularly suitable for applications where low re-
sidual H2S and COZ concentrations are required simultaneously at low absorber
acid gas partial pressures. MEA is the most reactive amine. However, the use
of MEA is generally inappropriate if the feed gas contains significant amounts
of COS and CS2 because of the formation of irreversible reaction products when
MEA combines with these sulfur compounds and the resultant chemical losses.
Al-5
-------
Appendix Al
Amine Processes
DBA, on the other hand, has been used for the treatment of refinery gases
which contain appreciable amounts of COS and CSa. This is because secondary
amines are much less reactive with COS and CS2 than primary amines. The ad-
vantages of DEA over MEA are resistance to degradation from COS and CS4, lower
vaporization loss, and lower heat requirements for amine regeneration.
As discussed previously, the use of TEA has been displaced largely by
either MEA or DEA. The principal advantage of TEA and other tertiary amines
such as MDEA is their selectivity for hydrogen sulfide. The high molecular
weight of TEA (compared with MEA and DEA) results in low acid gas carrying
capacity at moderate amine concentrations (on TEA weight basis), and generally
leads to higher capital and operating costs.
Both DIPA and MDEA have been used in the ADIP process. The ADIP process
is designed to remove HaS and substantial amounts of COS without detrimental
effects to the solution and with only limited co-absorption of C02. It is
particularly suited for the enrichment of H^S in Claus feed streams, as most
Claus units are not equipped to treat acid gases containing less than 15 per-
cent HaS. MDEA based solvent is used in the ADIP process when the feed gas
contains high levels of C0a and higher selectivity for HaS is desired.
Another important use of the ADIP process is for Clans tail gas treatment. In
this application the ADIP unit is an integral part of the SCOT process, as
discussed in Appendix A8.
3. Process Performance
The performance of amine processes in gas purification depends on four
primary factors: number of trays in the absorption tower, recirculation rate
of the amine solution, temperature of the lean amine solution entering the
absorption column, and the residual amount of H^S in the lean amine solution.
In a great number of cases, the treated gas stream leaving the top of the
absorption tower is found to be in equilibrium with the inlet lean amine
Al-6
-------
Appendix Al
Amine Processes
solution with respect to H2S content. Thus, the critical parameter deter-
mining process performance is often the residual H2S in the lean solution,
which is largely dependent on the amount of stripping steam employed in the
amine regenerator. Also, given a particular lean solution loading, low H2S
concentrations are favored by high operating pressures. To achieve a low HaS
specification at low operating pressures, it would be necessary to reduce resi-
dual H2S in the lean amine solution to very low levels. Another consideration
is that primary amines are more reactive than secondary and tertiary amines
and primary amine solutions will have a lower acid gas vapor pressure at
equivalent acid gas loadings. Thus, plants designed to meet 4 ppmv H2S speci-
fication (the conventional pipeline gas specification) with primary amine
solvents such as MEA only need to regenerate the amine solution to a typical
H2S loading of 0.05 to 0.1 mol/mol lean solution (4). For secondary amines
such as DEA, it is necessary to reduce the residual H2S in the regenerated
amine solution down to the order of 0.01 to 0.02 mol/mol lean solution to
achieve the same H2S specification.
The typical MEA design involves absorption in a 24 to 30 tray tower with
a MEA circulation rate of 0.03 to 0.04 m3 of 15 to 18 wt % MEA solution per m'
of acid gas removed (equivalent to 0.40-0.56 mol acid gas removed/mol MEA)
(1). Regeneration is carried out in a 20 tray tower, with typical steam
requirements of 120 to 140 kg/mj of MEA solution. At these design and oper-
ating conditions, H2S levels of less than 4 ppmv are obtained in the treated
gas at operating pressures above 0.8 MPa (1,4,5).
The performance of MEA absorption systems can be estimated using experi-
mental data on the equilibrium of H2S and C02 over MEA solutions. However,
data are generally very limited in those regions of low acid gas partial pres-
sure (less than 7 Pa). Kent and Eisenberg have devised a model to correlate
vapor-liquid equilibria data for the absorption of H2S and C02 in MEA (5).
This model considers the amine's reaction with ions in solution, the ionic
Al-7
-------
Appendix Al
Amine Processes
dissociation reactions for HaS, C0a, and HaO, and relates the equilibrium par-
tial pressures of CO, and HaS to the free concentrations of CO, and HaS in
solution by the Henry's law constants. The model shows excellent agreement
with published vapor pressure data and can be used to predict absorber per-
formance as most commercial MEA absorbers are capable of achieving a very
close approach to equilibrium at the top of the tower.
The typical DBA process design involves absorption in a 25 tray tower.
With high DBA circulation rates (0.06 m3 of 20 wt % DEA solution per m3 of
acid gas removed, equivalent to approximately 0.35 mol acid gas removed/mol
DEA) and high steam rates for regeneration (180 kg/m3 of DEA solution), sulfur
levels of less than 10 ppmv are obtained in the treated fuel gas. This is
because of the low residual levels of sulfur compounds in the lean DEA solu-
tion. With lower and more typical DEA circulation rates (0.04 m3 of 20 wt %
DEA solution per m3 of acid gas removed, equivalent to approximately 0.53 mol
acid gas removed/mol DEA) and steam rates for regeneration (120 kg/m3 of DEA
solution), however, sulfur levels of approximately 100 ppmv are obtained in
the treated fuel gas. Performance data for selected DEA absorption systems
are summarized in Table Al-1. Data for Plant A were obtained in the early
development phases of the SNPA-DEA process. The SNPA-DEA process is specifi-
cally designed to treat natural gas streams at operating pressures of 3.55 MPa
or higher to meet the conventional pipeline gas specification of 4 ppmv HZS or
less. It will be noted that higher sulfur levels from 34 to 192 ppmv are
reported for Plants B, C, D, and E. This is because in refinery operations,
the treated gas in often used for fuel within the refinery and a high degree
of purity is not specified.
As with MEA, the performance of DEA absorption systems can also be esti-
mated by using experimental data on the equilibrium of H2S and C0a over DEA
solutions, or by using the vapor-liquid equilibria correlation developed by
Kent and Eisenberg (5).
Al-8
-------
TABLE Al-1. PERFORMANCE DATA FOR DEA ABSORPTION SYSTEMS
>
M
I
Design and Operation Variables
Feed Gas
Plant A
Natural Gaa
Plant B
Plant C
Plant D
Syntheals Gas Synthesis Gaa Refinery Gas
Plant E
Refinery Gas
Plant F
Pilot Plant
Synthesis Gaa
Gaa Feed Rate, baol/hr
Feed Gaa Analysis:
BaS, percent
C0a, percent
COS , ppmv
CSa, ppmv
Outlet Gas Analysis :
HaS, ppmv
C0a, percent
COS , ppmv
CSa, ppnv
DEA:
Solution Concentration, vt %
Reel rculs tlon rate -
at'/m* acid gas removed
kmol acid gas/kBol DEA
Absorber :
Number of trays
Pressure, MPa
Regenerator :
Number of trays
Pressure, HPa
Steam rate. kg/m> DEA solution
Reboller temperature, K
gata Source: Plants A to E (Reference
5 cm rings
16 m of packing
1770
15.0
10.0
300
600
4.5
0.002
<1
No Data
20
0.033
0.63
30
7.0
20
0.27
120
406
1). Plant F (Reference
104
0
15
No
No
192
2
No
No
35
0
0
a
2
No
No
No
No
6)
.208
.0
Data
Data
.5
Data
Data
.027
.45
.5
Data
Data
Data
Data
86
0.
19.
No
No
34
4.
No
No
41
0.
0.
a
2.
No
No
No
No
120
4
Data
Data
2
Data
Data
031
34
4
Data
Data
Data
Data
No
0
No
No
No
96
No
No
No
No
No
No
b
1,
No
No
No
No
Data
.42
Data
Data
Data
Data
Data
Data
Data
Data
Data
.8
Data
Data
Data
Data
No
2.
No
No
No
80
No
No
No
No
No
No
c
1.
No
No
No
No
Data
.4
Data
Data
Data
Data
Data
Data
Data
Data
Data
,1
Data
Data
Data
Data
No
8
3
0
0
10
0
0
0
20
0
0
16
0
No
No
No
No
Data
.7
.0
.42
.049
.43
.45
Data
Data
Data
Data
-------
Appendix Al
Amine Processes
Typical design and operating conditions for TEA absorption systems are
not readily available from the published literature because these processes
are only in limited use. The performance of TEA for H2S removal is expected
to be similar to MEA and DEA absorption systems; and H2S levels of less than 5
to 10 ppmv can be obtained in the treated gas, provided the residual H2S level
in the lean TEA solution is rendered sufficiently low by stripping steam (1).
The TEA process can be designed for limited co-absorption of C02 if an acid
gas stream of higher H2S/C02 ratio than the feed gas is desired.
The ADIP process involves absorption in a tower with 15 to 25 trays. The
temperature of the ADIP solution entering the absorber is typically 310 K, but
may be as high as 330 K. The ADIP process can reduce H2S content in natural
gas to less than 4 ppmv and in fuel gas to less than 100 ppmv (3). Perfor-
mance data for the ADIP process are summarized in Table Al-2. The performance
of the ADIP process can generally be described by using the steady-state film
theory with chemical absorption (8).
The data presented for the ADIP process are based on using DIPA as the
absorption solution. The ADIP process can also use HDEA as the absorption
solution (3), but commercial application of MDEA has been quite limited to
date (9). According to Shell (10), MDEA is used in approximately 25 ADIP
absorption systems in SCOT units (SCOT units are designed to treat the tail
gas from Claus sulfur recovery plants). MDEA is used in place of DIPA when
the C02 loading in the Claus tail gas is high and higher selectivity for H2S
over C02 is required. MDEA has shown good selecivity for H2S over C02, espe-
cially at low operating pressures (1). However, the selectivity decreases and
the H2S level in the treated gas increases markedly after a critical acid gas
loading level is surpassed (6). Both the selectivity and the H2S level in the
treated gas are also sensitive to the temperature of the lean MDEA solution
entering the absorber. These are some of the reasons the MDEA solution has
not been extensively used. At acid gas loadings of less than 0.79 imol/kmol
Al-10
-------
TABLE Al-2. PERFORMANCE DATA FOR THE ADIP PROCESS USING DIPA SOLVENT
Design and Operation Variables
Feed Gas
Gas Feed Sate, knol/hr
Feed Gas Analysis:
H ,S , percent
CO, , percent
Outlet Gas Analysis,
H.S, ppmv
CO,, percent
Absorber :
Nuaber of trays
Pressure, HPa
Temperature, E
Begeners tor :
Stream rate, kg/kg acid gas removed
Plant A
Cracked Gaa
Catalytic
Cracker
500
10.4
2.5
00
0.2
20
1.97
308
1.83
Plant B
Plant C
Residual Gaa Gaa from
Froa Hydrogen
Bydrodeaulf ur i- Purification
za tlon
5000
15.6
--
100
—
15
0.48
314
2.29
5190
3.0
91.2
500
98.1
No Data
0.10
283
No Data
Plant D
Plant E
Gaa from Fluid Natural Gas
Catalytic
Cracking
1000
2.4
1.9
160
1.0
No Data
1.34
310
2.73
185
7.1
21.3
130
17.8
No Data
0.28
316
3.53
Plant F
Synthesis Gas
Fron Oil
Gasification
838
0.5
5.5
2
No Data
25
2.5
40
1.3
Data Source: Plant A - E deference 7), Plant F (Reference 1)
-------
Appendix Al
Amine Processes
MDEA and lean MDEA temperatures lower than 320 K, H^S levels of less than 10
ppmv was obtained when a refinery gas containing 8.5 percent H^S and 1.4 per-
cent C02 was fed to a 20 tray absorption tower (4).
4. Secondary Waste Generation
Reactions of amines with gas contaminants lead to chemical degradation of
the amine solutions and the subsequent generation of secondary wastes. The
most common degradation loss of MEA and DBA is its irreversible reaction with
C02 present in the gas to form carbamate or carbonate (1). Monoethanolamine
carbonate (or monoethanolamine carbamate) is first converted to oxazolidone-2
and subsequently to N—(2—hydroxyethyl) — ethylenediamine and C02. Diethanola—
mine carbamate or carbonate is converted to 3-(2-hydroxyl)oxazolidone-2 and
then to N,N'-di(2-hydroxyethyl)piperazine and other nitrogen-containing degra-
dation products (1). MEA reacts readily with about 15 to 20 percent of the
COS in the feed gas to form irreversible products, including oxazolidone, imi-
dazolidone, and diethanol urea (1). Both primary and secondary amines also
react with carbon disulfide to form dithiocarbamates. In the MEA or DEA pro-
cess, degradation products are often removed by filtration of the recircula—
tory amine solution. The secondary waste generated will be a filter residue
containing iron sulfide (a corrosion product), liquid hydrocarbons, and MEA or
DEA degradation products. In the ADIP process, degradation products are some-
times removed by blowdown of the recirculating ADIP solution. In addition to
degradation products, this secondary waste stream will also contain trapped
impurities, hydrocarbons, and corrosion products.
5. Process Reliability
One reason why alkanolamine processes are used extensively today is that
they are relatively free from operating problems. Nevertheless, operating
difficulties sometimes arise due to problems caused by corrosion, foaming, and
Al-12
-------
Appendix Al
Amine Processes
the plugging of equipment. In MEA or DEA processes, the most severe corrosion
is generally found in the tube side of the amine heat exchanger and the steam
stripper (1). The corrosion observed in DEA plants, however, is considerably
milder than that found in plants using MEA. Corrosion in MEA or DEA plants
can be controlled by the continuous or periodic removal of corrosion—promoting
agents [It has been proposed that such compounds as hydroxyethylethylenedi-
amine, which is a reaction product of MEA and C02, act as chelating agents for
iron in the hot sections of the process and promotes corrosion (1)1 and sus-
pended solids from the recircnlating solution by filtration and addition of
corrosion inhibitors. Plants using aqueous solutions of DIPA are reported to
have essentially no corrosion problems.
Foaming is usually caused by contamination of the amine solution by amine
degradation products, finely divided suspended solids such as iron sulfide,
and condensed light hydrocarbons (1). The condensation of light hydrocarbons
can be controlled by keeping the temperature of the lean amine solution above
that of the feed gas and by flashing of the rich amine solution. Suspended
solids and amine degradation products can be removed by filtration or solution
blowdown. Foaming can also be controlled by the addition of foam inhibitors
such as silicone compounds.
In general, after the initial shakedown period, alkanolamine absorption
systems are not expected to cause unscheduled plant shutdowns.
6. Process Economics
Data on the capital investment costs and operating requirements of alka—
nolamine processes are limited. Most of the available data relate to the
operation of DEA units. For estimation purposes, it can be assumed that the
capital investment cost and operating requirement for MEA units are similar to
those for DEA units. The cost of steam for regeneration of the MEA solution
Al-13
-------
Appendix Al
Amine Processes
is the major contributor to the overall operating cost of MEA units, and the
regeneration steam requirement for MEA and DEA units are approximately the
same on a kg/ma amine solution basis.
The capital investment cost for a DEA unit with a DEA circulation rate of
397 m*/hr and designed to treat sour gases at 3.6 MPa in a SRC-II commercial
plant was estimated to be $15 million based on 1980 dollars (11). In a SRC-I
demonstration plant, the capital investment cost for a DEA system with a total
DEA circulation rate of 192 m'/hr and designed to treat both low pressure
(0.63 MPa) and high pressure (12.4 MPa) sour gases in two separate absorption
towers was estimated to be $8.1 million based on 1980 dollars (12). The Gas
Processing Handbook reported in 1970 a capital investment cost of $0.48
million (adjusted to 1980 dollars) for a 22.7 m3/hr DEA unit operating at 0.45
MPa (13). Gary and Handwerk presented cost curves for amine treating units at
low operating pressures (14). These cost curves show that at capacities above
22.7 m3/hr, the costs increase with the 0.84 power of the capacity ratio. The
cost data from these sources are plotted in Figure Al-2. The higher cost
curve, based on the two data points provided in the SRC-I and SRC-II designs.
is appropriate for DEA units operating at pressures exceeding 3.6 MPa. For
this higher cost curve, the capital investment costs increase with the 0.85
power of the capacity ratio. The lower cost curve, based on a single data
point provided by the Gas Processing Handbook, is plotted assuming that costs
also increase with the 0.85 power of the capacity ratio. This lower cost
curve is appropriate for DEA units operating at pressures of 0.45 MPa, and
will provide conservative cost estimates when compared with the cost curve
presented by Gary and Handwerk.
The operating requirements for DEA absorption systems are presented in
Table Al-3. The operating requirement for any DEA unit can be calculated by
first determining the DEA circulation rate. The cost of steam for regenera-
tion of the DEA solution is usually the major contributor to the overall oper-
ating cost.
Al-14
-------
SI-TV
Amine Solution Circulation Rate, m "7hr
o
o
o
o
o
o
o
c
o
ODO
05
c
O o
_[u 0)
a
w
CL
a*
CL
VD
00
O
N)
O
50 po
n o
i i
n
o o
CO O
-------
Appendix Al
Amine Processes
TABLE Al-3. OPERATING REQUIREMENTS FOR DEA ABSORPTION SYSTEMS (11-14)
DEA Scrubbing with High
Circulation Rates
Cost Item (0.06 m3/m3 acid gas)
1 MPa Steam, kg/m3 solution
Electricity, kWh/m3 solution
Cooling Water, m3/m3 solution
DEA Consumption, mg/mol gas scrubbed
Operating Labor
180
2.6
6.6
1.2
1 man/shift
DEA Scrubbing
with moderate
Circulation Rates
(0.04 m3/m3 acid gas)
120
2.6
4.4
0.8
1 man/shift
Capital investment costs for ADIP units are presented in Figure Al-3.
These costs were estimated based on the assumption that:
ADIP cost = SCOT cost - (Beavon Cost - Stretford Cost)
This is a reasonable assumption because: 1) ADIP is used in the back section
of the SCOT process; 2) the front sections of the Beavon and SCOT processes
are similar to COS hydrolysis sections; and 3) Stretford is used in the back
section of the Beavon process. The operating requirements for ADIP are pre-
sented in Table Al-4.
TABLE Al-4. OPERATING REQUIREMENTS FOR ADIP PROCESS (3,15)
Steam 32-96 kg/kmol of sulfur removed
Electricity 71 kWh/Ntaol feed gas
DIPA JO.OS/Mmol feed gas
Operating Labor 1 man/shift
Al-16
-------
1000
500
200
100
50
i
to
TD
0)
20
10
I I I I I I
i
I I I I
0.1
0.2
0.5 1.0 2.0 5.0
Capital Investment, $10' mid-1980
10
20 30
Figure Al-3. Capital investment cost for AD IP process
Al-17
-------
Appendix Al
Amine Processes
7. References
1. Kohl, A. and F. Riesenfeld. Gas Purification Second Edition. Gulf
Publishing Company, Houston, Texas, 1974.
2. Blanc, C., J. Elgue, and F. Lallemand. MDEA Process Selects H3S,
Hydrocarbon Processing, 60(8):111-116, August 1981.
3. Gas Processing Handbook. Hydrocarbon Processing, 58(4):99-170, April
1979.
4. Butwell, K. F., D. J. Kubek and P. W. Sigmund. Alkanolamine Treating.
Hydrocarbon Processing, 61(3):108-116, March 1982.
5. Kent, R. and B. Eisenberg. Better Data for Amine Treating. Hydrocarbon
Processing, 55(2):87-90, February 1976.
6. Vidaurri, G. C. and L. C. Kahre. Recover H2S Selectively from Sour Gas
Streams. Hydrocarbon Processing, 56(11):333-337, November 1977.
7. Duncan, J. M., Shell Development Company, Houston, Texas to M. Ghassemi,
TRW Environmental Division, Redondo Beach, California. The Shell ADIP
Process. December 8, 1977.
8. Ouwerkerk, C. Design for Selective H2S Absorption, Hydrocarbon
Processing, 57(4):89-94, April 1978.
9. Edwards, M. S. H2S Removal Processes for Low-Btu Coal Gas.
ORNL-TM-6077 , Report prepared by Oak Ridge National Laboratory, Oak
Ridge, Tennessee, January 1979.
10. Kuijpers, N. G. M. The Shell Claus Off-Gas Treating Process. Paper pre-
sented at the Gas Sweetening and Sulphur Recovery Seminar, Amsterdam, the
Netherlands, November 9-13, 1981.
11. U.S. Department of Energy. SRC-II Demonstration Project Conceptual
Commercial Plant Design. July 31, 1979.
12. U.S. Department of Energy. SRC-I Coal Refinery Phase 0. 6000 TPD
Demonstration Plant, Tasks 1 and 2 - Preliminary Demonstration Plant
Design, Cost Estimates, and Long Lead Procurement Planning - Summary.
July 31, 1979.
Al-18
-------
Appendix Al
Amine Processes
13. Gas Processing Handbook. Hydrocarbon Processing 49(9):257. September
1970.
14. Gary, J. H. and G. G. Handwerk. Petroleum Refining Technology and
Economics. Marcel Dekker, Inc., New York, 1975, p 185-190.
15. Handbook of Gasifiers and Gas Treatment Systems. FE-1772-11, Report pre-
pared by Dravo Corp. for the U.S. Energy Research and Development
Administration, February 1976.
Al-19
-------
APPENDIX A2
RECTISOL ACID GAS REMOVAL PROCESS
1. Process Description
Rectisol is an acid gas removal process which removes carbon dioxide,
hydrogen sulfide, carbonyl sulfide, organic sulfur compounds, hydrogen cya-
nide, ammonia, benzene, and gum-forming hydrocarbons from synthesis gases by
means of physical absorption in an organic solvent (especially cold methanol)
at temperatures below 273 K. Operation is based upon the fact that these com-
pounds, particularly the reduced sulfur species and carbon dioxide, are very
soluble at high pressure in cold methanol and are readily recoverable by flash
desorption. This is demonstrated in Figure A2-1, which presents carbon
dioxide solubility as a function of partial pressure (1). Consider, for exam-
ple, the absorption of carbon dioxide at a partial pressure of 1.0 MPa.
Figure A2-1 shows that at least 90 percent of the dissolved carbon dioxide may
be desorbed by isothermal flashing at methanol temperatures of 258 K or lower.
Solubility data for compounds at a partial pressure of 0.1 MPa over
methanol are presented in Figure A2-2 (2). It should be noted that gas
solubilities generally increase with increasing partial pressure but that
solubility coefficients (the ratio of solubility to partial pressure) do not
increase appreciably with pressure until partial pressures exceed 0.1 to 0.2
MPa. Solubility coefficients of hydrogen sulfide and carbon dioxide are seen
to increase substantially with decreasing temperature while those of major
product gases such as hydrogen, carbon monoxide, and methane are relatively
temperature independent. For this reason, Rectisol absorption columns operate
at low temperatures, typically in the range of 253 to 213 K (1,3,4). Low
temperature operation also reduces solvent losses by reducing the partial
pressure of methanol in the product streams.
A2-1
-------
TO
i-i
(D
ho
I
ho
ho
I
3 M
(B i-n
rt i-h
!T fD
PI O
O
Ml
C02 Partial Pressure, MPa
I-i
rt
TO
cn
en
e
o
3
cn
O
C
o"
p-
H-
rt
o
M
&
H-
rt
O
Hi
O
O
<
o
<:
o
s
cr
o
H-
O
-------
Solubility Coefficient (A) at one Atmosphere Partial Pressure,
kmol of Dissolved Gas/(Mg of Solvent x MPa Partial Pressure of Gas)
I
UJ
OQ
re
in
3
re
rr
61
O
0)
p
o
H
(B
-------
Appendix A2
Rectisol Process
Because the solubilities of reduced sulfur species (e.g., hydrogen sul-
fide and carbonyl sulfide) in methanol are substantially greater than that of
carbon dioxide at the same partial pressure, the Rectisol process is capable
of selective recovery of reduced sulfur species versus carbon dioxide; to some
degree, this holds for all physical absorption solvents capable of absorbing
reduced sulfur species and carbon dioxide.
The Rectisol process was jointly developed by Linde Aktiengesellschaft
(Munich, Germany) and Lurgi Mineraloltechnik (Frankfurt, Germany), and is
currently licensed by both companies. It is also available through their
U.S. subsidiaries, Lotepro Corp. (New York, NY) and Lurgi Corp. (River Edge,
NJ) , respectively. The Gelleschaft fur Kohle Technologic (GKT, Essen,
Germany) also has a limited Rectisol license applying to Koppers-Totzek (K-T)
gasification facilities.
Selective Rectisol Process Configurations
A variety of selective Rectisol units are currently being used in
applications such as ammonia and methanol synthesis, medium-Btu gas synthesis,
natural gas purification, and refinery hydrogen production. Although selec-
tive Rectisol designs are site- and process-specific, common key features
include low temperature operation, sequential hydrogen sulfide-carbon dioxide
absorption, discrete methanol regeneration columns for hydrogen sulfide and
carbon dioxide recovery, and separation of methanol and water by distillation.
However, there are significant differences among the designs in use which
relate to both the feed gas composition and the product specifications.
Examples of selective Rectisol process configurations used in coal
gasification applications are presented in Figures A2-3 and A2-4. The process
presented in Figure A2-3 is used by AECI Limited at Modderfontein, Republic of
South Africa, and desulfurizes an essentially hydrocarbon-free quenched K-T
A2-4
-------
ro
I
> HjS-RlCH GAS
REGENERATION
COLUMN
ICOjl
COjRICHOAS
Figure A2-3. Process flow diagram of the Modderfontein selective Rectisol section (5,6)
-------
PRODUCT GAS
CRUDE
PRODUCT
GAS FROM
GAS
PRODUCTION
SECTION
PURE PRODUCT
GAS TO AMMONIA
PLANT
GAS
CONDENSATE
TO MEDIUM
OIL SEPARATOR
IN TAR OIL
SEPARATION
SECTION
BY PRODUCT
NAPHTHA TO
BY PRODUCT
STORAGE
SECTION
CYANIC WATER
TO TAR OIL
SEPARATION
SECTION
CLEAN PRODUCT
GAS TO GAS
DISTRIBUTION
RICH WASTE
GAS TO BURNER
C02 RICH WASTE
MAKEUP
METHANOL
Figure A2-4. Process flow diagram of the Kosovo selective Rectisol section (7,8)
-------
Appendix A2
Rectisol Process
gas prior to carbon monoxide shift conversion and subsequent carbon dioxide
removal (5,6). Methanol is added to the feed gas prior to cooling and
hydrogen sulfide absorption to prevent icing. Moisture in the feed gas is
removed from the hydrogen sulfide absorber in solution with methanol, which is
recovered by distillation. Hydrogen sulfide and carbonyl sulfide are absorbed
from the feed gas using sulfur-free methanol from the carbon dioxide regenera-
tion column. Rich methanol from the hydrogen sulfide absorber is partially
flashed to liberate absorbed hydrogen and carbon monoxide which is compressed
and combined with the cold feed gas. Additional flashing and stripping in the
concentration column, with reabsorption of reduced sulfur species in sulfur-
free methanol, produces a sulfur-rich methanol stream for hot regeneration and
a carbon dioxide offgas. Hydrogen sulfide is recovered by stripping with
methanol vapor in the regeneration column.
Carbon dioxide is removed from shifted process gas by absorption in
regenerated methanol. Methanol is added to the shift gas prior to cooling and
carbon dioxide absorption to prevent icing, and moisture in the shift gas is
removed from the carbon dioxide absorber in solution with methanol. Rich
methanol from the carbon dioxide absorber i$ partially flashed to recover
absorbed hydrogen which is compressed and combined with the cold feed gas to
the hydrogen sulfide absorber. Carbon dioxide is recovered by flashing and
stripping with nitrogen in the carbon dioxide regeneration column.
It should be noted that desulfurization prior to shift conversion enables
the use of conventional shift catalysts (e.g., iron-chromium and copper-zinc)
and can enhance process selectivity by absorbing hydrogen sulfide in the
presence of a minimum of carbon dioxide (approximately 10 to 12 percent by
volume for K-T coal gasification, 18 to 20 percent for Texaco coal gasifica-
tion, and 5 to 6 percent for gas produced by partial oxidation of oil). How-
ever, in conjunction with partial oxidation of liquid hydrocarbons for hydro-
gen or ammonia production, shift conversion employing sulfur tolerant cobalt-
molybdate shift catalysts precedes acid gas removal. Selective Rectisol
A2-7
-------
Appendix A2
Rectisol Process
configurations for such systems are similar to that presented in Figure A2-3
except that no gas processing occurs between hydrogen sulfide absorption and
carbon dioxide absorption. Shift conversion prior to acid gas removal results
in an increased concentration of carbon dioxide in the hydrogen sulfide
absorber feed gas (up to about 42 percent volume). Owing to the less favor-
able carbon dioxide to hydrogen sulfide ratio after shift conversion, a
greater degree of methanol enrichment is required to achieve the same
selectivity attainable with an unshifted feed gas.
The process presented in Figure A2-4 is used at the Kosovo Gasification
Plant near Pristina, Yugoslavia for the production of medium-Btu fuel gas and
hydrogen for ammonia synthesis (7,8). Feed gas to the Rectisol unit is
generated by gasification of lignite in oxygen-blown Lurgi-type gasifiers.
Cooled crude gas from gasification is further cooled by sequential washing
with cold water and methanol in the two stage cooler. Condensed gas liquor
from the water wash section is flashed to liberate dissolved sour gases, and
the organic phase is recovered from wash water in the naphtha separator. Con-
densed gas liquor from the cold methanol wash section is flashed, and methanol
and condensed moisture are recovered from the naphtha phase by extraction with
water. Dissolved organics in the aqueous phase are recovered by distillation.
Naphtha from the naphtha separator and the naphtha/methanol/water extractor is
sent to byproduct storage via the naphtha surge tank. Cyanic water from
naphtha separation and methanol/water distillation is sent to tar/oil separa-
tion.
Gas from the two stage cooler is scrubbed with carbon dioxide—rich
methanol in the hydrogen sulfide absorber for bulk removal of reduced sulfur
species. Carbon dioxide is removed from the first absorber top gas in two
carbon dioxide absorbers. Bulk carbon dioxide removal is achieved in the
first absorber by washing with carbon dioxide-lean methanol and regenerated
methanol. Overhead gas from the first carbon dioxide absorber is fed directly
A2-8
-------
Appendix A2
Rectisol Process
into the fuel gas distribution system. When a higher purity gas is required
for feed to the cryogenic hydrogen separation unit, additional carbon dioxide
removal is achieved in the second carbon dioxide absorber using regenerated
methanol.
Hydrogen sulfide-rich methanol is regenerated by multistage flashing in
the hydrogen sulfide flash tower, and steam stripping in the methanol regener-
ation column. Hydrogen sulfide-rich waste gas from methanol regeneration is
combined with flash gas from the naphtha separator and the methanol prewash
flash tank prior to disposal. Carbon dioxide-rich methanol is regenerated by
multistage flashing and nitrogen stripping in the carbon dioxide flash tower.
Based upon publicly available data, it is not known how the Kosovo
Rectisol design compares with other selective Rectisol units currently pro-
cessing Lurgi crude gas. Several selective Rectisol designs have been pre-
pared for proposed Lurgi gasification facilities in the United States (e.g.,
facilities for Wesco, El Paso Natural Gas Co., Hampshire Energy Co., and
Nakota Co) (9). However, data with respect to process configuration are gen-
erally proprietary.
Configurations of the two units presented in Figures A2-3 and A2-4 differ
in several respects. Principal differences result from 1) the fact that Lurgi
crude gas contains significant levels of condensible hydrocarbons (approxi-
mately 0.01 kg Cj+ aliphatics, benzene, toluene, and other aromatics per
kg MAP coal) which must be removed prior to acid gas removal (10), 2) the need
for two-stage acid gas removal if sulfur intolerant catalysts are used for
shift conversion, and 3) the fact that at Kosovo all hydrogen sulfide contain-
ing offgases are simply burned so that high hydrogen sulfide concentrations
are not necessary, as would be the case for Claus processing. These designs
differ substantially with regard to selectivity.
A2-9
-------
Appendix A2
Rectisol Process
Nonselective Rectisol Process Configurations
Nonselective Rectisol processes differ from selective processes in that
all acid gas constituents are absorbed simultaneously and no carbon dioxide
regenerator or reabsorber is used to produce a high purity carbon dioxide vent
gas. An example of a commercial nonselective Rectisol unit is presented in
Figure A2-5 which is a simplified schematic of the South African Oil, Coal,
and Gas Corporation's SASOL I acid gas removal system (1). Feed gas to acid
gas removal is crude or partially shifted Lurgi gas from Fischer-Tropsch syn-
thesis. The feed gas is split into three streams which are cooled in each of
two stages by refrigeration, heat exchange with cold high pressure flash gas
(including carbon dioxide-rich flash gases above 100 tPa) and heat exchange
with cold product gas. Condensed moisture and hydrocarbons are recovered from
the combined feed gas following the first cooling stage, and methanol is added
to prevent icing in the second gas cooling stage. Following the second gas
cooling stage, the condensed gas liquor is recovered from the coal gas and
sent to the naphtha separator for byproduct and methanol recovery.
Cooled gas is washed with cold methanol in three consecutive stages. In
the first absorption or prewash stage, the cooled gas is washed with flashed
methanol from the expansion tower to remove the final traces of condensible
organics along with some hydrogen sulfide, carbon dioxide, and organic sulfur
compounds. Rich methanol from the first stage absorber is combined with gas
liquor from the gas cooling second stage and sent to the naphtha separator.
Separator feed is flashed and extracted with water to yield an aqueous
methanol phase and a byproduct naphtha phase containing organic sulfur
compounds. Methanol is recovered from the aqueous phase by distillation.
Bulk acid gas removal is achieved in the second or main wash stage or
absorber by washing with flashed methanol from the expansion tower. Rich
methanol from the second stage absorber is regenerated along with the
A2-10
-------
FEED GAS
HIGH PRESSURE
FLASH GAS
PRODUCT GAS •*
ATMOSPHERIC
HOT
REGENERATOR
CYANIC WATER
LOW PRESSURE
FLASH GAS
Figure A2-5. Process flow diagram of the SASOL I non-selective Rectisol section (1)
-------
Appendix A2
Rectisol Process
methanol/water still overhead in an expansion tower. Regeneration is by
pressure reduction in six stages to a final pressure of about 30 kPa. High
pressure flash gas consisting primarily of carbon dioxide, carbon monoxide,
and hydrogen is used to cool the Rectisol feed gas and then used as onsite
fuel gas. Low pressure flash gas is compressed and flared.
The third or finewash stage absorber effects final gas purification by
washing the second-stage absorber effluent gas with completely stripped
methanol from the hot regenerator. Rich methanol from the third-stage
absorber is partly regenerated by flashing to atmospheric pressure and then
competely stripped of acid gas in a distillation column. Atmospheric flash
gas from the hot regenerator is released for incineration. Cold product gas
is used to precool the Rectisol feed gas and then sent to liquid synthesis.
Based upon publicly available data, it is not known how the SASOL I
Rectisol design compares with other commercial nonselective Rectisol
processes, although a similar design has been used in the SASOL II facility
which was commissioned in 1980 (11). Nonselective Rectisol designs have been
prepared for several proposed Lurgi gasification facilities in the United
States including those proposed by Great Plains Gasification Associates
(currently under construction), Wycoalgas Inc., Tenneco Coal Gasification, and
El Paso Natural Gas Co. (9,12). A schematic of the Great Plains nonselective
Rectisol section is presented in Figure A2-6 (13). This schematic indicates a
similar configuration to that of the SASOL I facility but includes details
such as the prewash flash vessel and the azeotrope distillation column which
are not included in Figure A2-5.
2. Process Applicability
The Rectisol process is used in three typical applications: 1) removal of
carbon dioxide, hydrogen sulfide, carbonyl sulfide, organic sulfur compounds,
A2-12
-------
ho
I
U)
Figure A2-6. Process flow diagram of the Great Plains non-selective Rectisol section (13)
-------
Appendix A2
Rectisol Process
hydrogen cyanide, ammonia, benzene, and gum-forming hydrocarbons from crude
gas produced by coal gasification for syngas and SNG production; 2) removal of
hydrogen sulfide, carbonyl sulfide, and carbon dioxide from gas produced by
partial oxidation for syngas or hydrogen production; and 3) used in conjunc-
tion with low temperature liquefaction and fractionation plants for removal of
acidic components present at moderate levels. Process limitations in these
applications primarily relate to requirements for high pressure, low tempera-
ture operation, and methanol contamination by minor constituents present in
the feed gas.
As with any other physical absorption process, the minimum circulation
rate of solvent required for complete removal of a gaseous constituent is
inversely proportional to the partial pressure of the constituent in the feed
gas and to the solubility coefficient for the constituent in the solvent
used. Process economics depend mainly upon the solvent circulation rate
because the circulation rate influences the size of all equipment and,
therefore, the capital costs. Solvent circulation rate also affects the
operating costs since pumping costs are proportional to circulation rate and
regeneration costs are nearly proportional to the circulation rate (14).
Therefore, the economics of physical absorption processes improves with
increasing acid gas partial pressures. Physical solvent type acid gas removal
processes are typically selected when acid gas partial pressures are greater
than about 1.0 to 1.4 MPa (1,15). Feed acid gas partial pressures at existing
Rectisol units in coal gasification and partial oxidation applications are in
the range of 0.4 to 2.6 MPa (3,5,6).
As indicated in Figure A2-2, the solubilities of most gases of interest
increase with decreasing methanol temperature. Thus, for reasons mentioned
above, Rectisol economics improve with decreasing methanol temperature.
Rectisol absorption columns operate at low temperatures, typically in the
range of 253 to 213 K (1,3,4). An additional benefit of low temperature
operation is the attendant reduction of methanol losses. Vapor pressure data
A2-14
-------
Appendix A2
Rectisol Process
for methanol are presented in Figure A2-7 (1). These data indicate that meth-
anol losses can be decreased by a factor of about three to four for each 20 K
temperature reduction down to 253 K and by about one order of magnitude for
each 20 K temperature reduction below 253 K.
Minor constituents such as ammonia, hydrogen cyanide, and nitrogen oxides
which may be present in the Rectisol feed gas can complicate operation or
result in fouling. Ammonia and hydrogen cyanide, which are very soluble in
methanol, make the regeneration process more complicated and result in addi-
tional steam requirements (2). Further, the presence of ammonia and hydrogen
cyanide in the hydrogen sulfide fraction is not desirable due to the potential
for adverse reactions during subsequent sulfur recovery. These contaminants
may be removed from the feed gas by employing a prewash of either cold water
or methanol. This prewash also provides feed gas drying (particularly the
methanol prewash) and, in low temperature gasification applications, removes
condensible hydrocarbons.
One coal gasification facility has reported Rectisol fouling which is
attributed to the presence of oxygen and nitrogen oxides in the Rectisol feed
gas (16,21). Oxygen in the Rectisol feed gas results in oxidation of a por-
tion of the hydrogen sulfide to elemental sulfur. The presence of nitric
oxide with oxygen accelerates the rate of sulfide oxidation. Deposits of
sulfur in columns resulted in reduced solvent circulation rates, and fouling
of heat exchangers resulted in insufficient cooling capability to achieve the
required degree of gas purification.
It has been determined that this fouling can be reduced by allowing low
levels of hydrogen cyanide and ammonia to enter the Rectisol unit to solubil-
ize sulfur by formation of ammonium thiocyanate which is ultimately removed
with the methanol/water distillation bottoms. When insufficient hydrogen
A2-15
-------
p
b
CO
00
1 I
p
o
I I
Methanol Vapor Pressure, nun of Mercury
I IT
I I I
OQ
NJ
CO
00
00
NJ
I
ft)
-a
o
1-1
fD
0)
tn
H
fD
X)
ft
l-t
P
rr
C
i-i
CD
ro
CJl
Co
ro
CD
00
CO
00
00
00
00
-------
Appendix A2
Rectisol Process
cyanide is present in the feed gas, sodium cyanide solution is injected into
the methanol. A more fundamental solution which has been implemented is the
hydrogenation of nitrogen oxides and oxygen over a cobalt molybdate catalyst
upstream of the Rectisol unit. Formation of elemental sulfur and the associ-
ated fouling of the Rectisol unit have not occurred since installation of the
catalytic hydrogenation unit (21).
3. Process Performance
Depending upon the product requirements and other site specific con-
straints, the Rectisol process can be designed to yield a product gas contain-
ing less than 0.1 ppmv total sulfur and less than 10 ppmv carbon dioxide.
The carbon dioxide content achievable in the purified gas is independent of
the type of Rectisol process employed (i.e., selective or nonselective
Rectisol). However, in the case of a nonselective Rectisol process, the util-
ities (steam, cooling water, and refrigerant) increase to obtain a product gas
with ppmv levels of carbon dioxide. Publicly available data indicate that in
gasification applications involving an essentially hydrocarbon-free feed gas,
selective Rectisol processes can produce a sulfur—rich offgas containing 25—
75 percent hydrogen sulfide and a carbon dioxide-rich offgas containing less
than 10 ppmv total sulfur. The presence of moderate quantities of hydrocar-
bons in the feed gas (9 to 16 percent) has no influence on the selectivity of
hydrogen sulfide recovery; hydrogen sulfide concentrations of 25 to 35 percent
in the hydrogen sulfide-rich offgas can be achieved along with a carbon
dioxide-rich offgas containing 10 ppmv total sulfur. However, C3 and C4
hydrocarbons present in the feed gas, will tend to concentrate in the hydrogen
sulfide-rich offgas.
Performance data for selective Rectisol units treating essentially hydro-
carbon free feed gases are summarized in Table A2-1. Plants 1 and 2 produce
hydrogen and ammonia synthesis gas, respectively, by partial oxidation of
A2-27
-------
TABLE A2-1. SELECTIVE RECTISOL PERFORMANCE DATA FOR HIGH TEMPERATURE GASIFICATION APPLICATIONS'
N>
I
M
CO
G.s
Component
",
N,+Ar
CO
CH.
CO,
B,S
COS
Flow Kate,
kmol/hr
Pressure, MPa
'Plant 1 is a
Feed Gas, Mole %
Plant 1
62.35-
63.74
0.12-
0.52
3.24-
4.13
0.13-
0.17
31.62-
33.23
0.26-
0.49
riant 2 Plant 3
61.59
0.41
2.60
0.33
34.55
0.52
10-63ppu
3562-
3992
303-313
3.2-3.3
6112
7.3
27.5-
29.3
1.52
36.62
0.10
11.8-
13.3
0.59-
0.75
0.10
4691-
4801
311
3.0-3.1
Purified Gas, Mole % Stripping Gal, Mole *
Plant 1
93.58-
94.08
0.17-
0.82
4.86
0.19
0.24
(lOppiD
-------
Appendix A2
Rectisol Process
oil. These plants utilize sulfur-tolerant shift conversion catalysts which
enable shift conversion prior to acid gas removal. Therefore, the feed gases
to Plants 1 and 2 contain 31 to 35 percent carbon dioxide, 62 to 64 percent
hydrogen, and less than about 5 percent carbon monoxide (2,3,6,17). Plant 3
is a coal gasification facility producing ammonia synthesis gas. This plant
employs a two—stage Rectisol system which removes sulfur species prior to
shift conversion and removes carbon dioxide subsequent to shift conversion
(refer to Figure A2-3 for example process flow diagram). Feed gas to the
Plant 3 sulfur absorber therefore contains only 12—13 percent carbon dioxide,
27-29 percent hydrogen, and about 57 percent carbon monoxide (2,6). Feed gas
to the Plant 3 carbon dioxide absorber, which is not included in Table A2-1,
contains 42-43 percent carbon dioxide, 53-54 percent hydrogen, and about 3
percent carbon monoxide.
These selective Rectisol units are seen to perform similarly in most
respects over a wide range of operating pressures although there is a substan-
tial range in the concentration of hydrogen sulfide, 25-72 percent, in the
sulfur—rich waste gas from Plant 3. Lotepro Corporation has indicated that
the higher hydrogen sulfide concentration is attainable at the expense of
higher refrigeration and stripping gas requirements (6). The amount of strip-
ping gas is a function of the hydrogen sulfide concentration desired in the
hydrogen sulfide-rich offgas, the type of Rectisol process, the feed gas pres-
sure, and the carbon dioxide, hydrogen sulfide, and carbonyl sulfide concen-
trations in the feed gas. Under given conditions, an increase in stripping
gas of about 60 percent is necessary to increase the hydrogen sulfide concen-
tration from 25 to 70 percent.
Performance data for the Kosovo selective Rectisol unit (taken at partial
load and not fully representative of normal performance), which treats crude
gas from Lurgi-type gasification, are summarized in Table A2-2 (refer to
Figure A2—4 for the process flow diagram). Data were obtained during three
A2-19
-------
TABLE A2-2. PERFORMANCE DATA FOR SELECTIVE RECTISOL UNIT AT KOSOVO MEDIUM-BTU COAL
GASIFICATION PLANT (7,18)a
i
ho
o
Gas Conponent
B,
i
0,
a
N.
1
co
CO,
1
CH
VM ^
C,H,
C»H4
C,
1
C4
C,
C
1
Benzene
Toluene
Xylene and
Ethy Ibenz ene
Phenol «
HaS
COS
CH.SB
C.H.SH
1 I
HCN
pH
Totil Solids, Bg/L
Total Nonvolatile
Solids, .|/1.
Total Suspended
Solids, Bg/L
Total Dissolved
Solids, mg/L
COD (as BgO,/L)
Pernangana te
(•< BgO,/U
Total Sulfur, mg/L
Flo* Rate, knol/hr
Flow Rate, B'/hr
Temperature, K
Crude Product Gas
(Stream 7.3), Hole *
Value Range
38.1
0.36
0.64
15
32
11. 5
0.47
0.04
0.19
0.074
0.044
0.064
750ppBv
230ppBV
lOOppBV
-IppBV
0.60
97ppav
590ppBv
3.3ppBV
320ppmv
703
295
36-46
0.09-2.6
0.04-1.6
9.6-17
21-40
8.9-14.5
~lppBV-0.76
-IppBV-O.ll
0.07-0.40
0.02-0.24
0.01-0.06
0.02-0.20
660-840ppnv
200-260ppnv
16-110ppBV
—
0.44-0.78
63-120ppav
460-700ppBv
-------
Appendix A2
Rectisol Process
sampling campaigns from September 1977 to November 1978. Tabulated data
represent the best overall data obtained during these tests, and the ranges of
the available data. As indicated previously, the presence of moderate levels
of hydrocarbons in the feed gas has no influence on process selectivity. Dif-
ferences in process selectivities indicated in Tables A2-1 and A2-2 primarily
reflect differences in process requirements. At the Kosovo facility, the sul-
fur-containing gases are burned and, therefore, high sulfur concentrations in
these offgases are not necessary. Thus, unlike the facilities cited in Table
A2—1, the Kosovo facility does not utilize an enrichment stage. Also,
Kosovo's hydrogen and methane rich flash gases from the carbon dioxide and
hydrogen sulfide loaded methanol streams are added to the hydrogen sulfide
fraction rather than being recycled to the feed gas.
Available performance data for the SASOL I nonselective Rectisol unit,
which also treats crude gas from Lurgi gasification, are presented in Table
A2-3 (refer to Figure A2-5 for the process flow diagram). As initially
designed, the high pressure flash gas is used as an onsite fuel gas, the low
pressure flash gas is flared, and the atmospheric flash gas is vented to the
atmosphere through the power stack (19). More recently, a Stretford unit was
designed to treat the atmospheric flash gas which contains about 90 percent of
the sulfur species absorbed (17,19). Proposed designs for U.S. facilities
indicate that at least a portion of the high pressure flash gas is recycled to
the gasification plant for recovery of carbon monoxide, hydrogen, and methane;
some fraction of the high pressure flash gas may be combined with the other
waste gases for sulfur recovery (12,20). Therefore, the performance indicated
in Table A2—3 may require some adjustment.
A2-21
-------
TABLE A2-3. NON-SELECTIVE BECTISOL PERFORMANCE DATA FOR SASOL I (LURGI GASIFICATION) (17)a
Offaases, Mole %
Gas Component
Ha
CO
CH4
COa
Na+Ar
HaS
COS
CSa
RSH
Thiophene
Total Sulfur
ct
Flow Rate, Nm3/hr
Temperature, E
Pressure, MPa
Rectisol Feed
Gas, Mole %
40.05
20.20
8.84
28.78
1.59
0.30
lOppmv
NA
20ppmv
NA
NA
0.54
381,000
303
2.6
Product
Gas, Mole %
57.30
28.40
11.38
0.93
1.77
ND
NA
NA
NA
NA
0.04ppmv
263,000
288
2.4
High- Pressure
Flash Gas
21.4
18.2
11.4
46.7
1.5
0.32
NA
NA
NA
NA
NA
0.7
4,600
273
1.3
Low- Pressure
Flash Gas
2.6
4.8
7.2
83.4
0.8
0.49
NA
NA
NA
NA
NA
1.1
15,000
273
0.48
Atmospheric
Flash Gas
0.14
0.0
0.9
97.2
0.03
0.88
30ppmv
2ppmv
280ppmv
2ppmv
NA
0.7
98,000
268
0.11
aRefer to Figure A2-5 for process flow diagram.
-------
Appendix A2
Rectisol Process
4. Secondary Waste Generation
Secondary waste streams produced by the Rectisol acid gas removal process
are: 1) hydrogen sulfide-rich offgases, 2) carbon dioxide-rich offgases
(selective Rectisol processes only), and 3) methanol/water distillation
bottoms. Available characterization data for the offgas streams have been
summarized in Section 3 for each of the three basic Rectisol process config-
urations. The sulfur-rich offgas is typically sent to the sulfur recovery
unit, either Claus or Stretford, or flared. When the Rectisol process is used
in conjunction with low temperature coal gasification systems (e.g., Lurgi
gasifiers) the Rectisol feed gas contains significant concentrations of Ca
hydrocarbons relative to the concentration of hydrogen sulfide. The naphtha
fraction is recovered from the feed gas by washing prior to acid gas removal.
Lighter hydrocarbons largely pass through the prewash and are, to some extent,
absorbed with the acid gases. These light hydrocarbons, particularly the
CjS and C4s, tend to concentrate in the hydrogen sulfide-rich offgas and may
also be present in the carbon dioxide-rich offgas. Therefore, unless special
precautions are taken, high levels of these C3 and C4 hydrocarbons in the
Rectisol feed gas may result in off-color sulfur if Claus sulfur recovery is
employed or excessive tail gas hydrocarbon emissions if Stretford sulfur
recovery is employed. An approach proposed in conjunction with Wesco and
Hampshire Energy Co. selective Rectisol units involves the use of an amine
unit (AOIP) to separate hydrocarbons from the Claus feed gas (9,20).
The carbon dioxide-rich offgas from selective Rectisol units is either
sold as byproduct or vented to the atmosphere at existing facilities. As
discussed above, light hydrocarbons present in the Rectisol feed gas are
coabsorbed to some extent with the acid gases and may be present in the carbon
dioxide-rich offgas. Further, steps taken within the Rectisol process to
minimize hydrocarbon levels in the hydrogen sulfide fraction will likely
result in increased hydrocarbons levels in the carbon dioxide offgases.
A2-23
-------
Appendix A2
Rectisol Process
Similarly, carbon monoxide is coabsorbed and will be present in the carbon
dioxide-rich offgas due to its low solubility in methanol. Of course the
extent of carbon monoxide coabsorption, and therefore its potential concentra-
tion in the carbon dioxide-rich offgas, depends upon its partial pressure.
Thus, for similar acid gas removal systems, processes requiring only a partial
shift conversion (e.g., SNG, methanol, or acetic acid syntheses) would be
prone to higher concentrations of carbon monoxide in the carbon dioxide-rich
offgas. Therefore, proposed designs in Lurgi-based coal gasification applica-
tions indicate either incineration of the carbon dioxide—rich offgas for con-
trol of carbon monoxide and hydrocarbon emissions, or sale of the offgas as
byproduct; direct discharge to the atmosphere is not being proposed. Also, at
least one non-Lurgi coal gasification plant currently under construction, the
Tennessee Eastman Kingsport, Tennessee Texaco gasification project, proposes
catalytic incineration of the carbon monoxide enriched portion of the carbon
dioxide offgas for control of carbon monoxide emissions (21).
Publicly available characterization data for the methanol/water distilla-
tion bottoms are extremely limited. This is apparently due to the fact that
the size of the still bottoms stream is generally quite small relative to
other wastewater streams requiring similar wastewater treatment (e.g., gas
liquor and synthesis condensates). Thus, from an operational standpoint, the
still bottoms are likely to be of minor significance other than for checking
still operation and methanol losses. One set of data, provided by SASOL per-
sonnel (19), are presented in Table A2-4. At the SASOL facility, this waste
stream is sent directly to biological treatment where it comprises less than
2 percent of the feed to this system.
A2-24
-------
Appendix A2
Rectisol Process
TABLE A2-4. CHARACTERIZATION DATA FOR METHANOL/WATER
DISTILLATION BOTTOMS AT SASOL (19)
Parameter/Component Value
PH 9.7
Phenol, mg/L 18
Cyanides (as CM), mg/L 10.4
(includes thiocyanate)
Ammonia (as N), mg/L 42
Sulfides (as S) Trace
COD, mg/L 1,686
5. Process Reliability
The original Lnrgi nonselective Rectisol unit built at SASOL in 1955 has
operated with an on-stream factor of about 97 percent (17). Normal mainte-
nance includes partial shutdowns about once per year for cleaning of critical
equipment, and complete shutdown every two years during the normal plant
downtime. Major upsets in the Rectisol unit requiring process adjustments
rarely occur (19).
As discussed in Section 2, plugging problems in a two-stage selective
Rectisol unit at a coal gasification facility have been reported (16). This
problem has been attributed to deposition of elemental sulfur resulting from
the presence of oxygen and nitrogen oxides in the Rectisol feed gas. Fouling
was at least partially controlled by allowing low levels of hydrogen cyanide
and ammonia to enter the Rectisol unit to solubilize sulfur by formation of
ammonium thiocyanate. A more fundamental solution is the hydrogenation of
nitrogen oxides and oxygen over a cobalt molybdate catelyst ahead of the
Rectisol unit. Detailed operating data from this facility are not available.
A2-25
-------
Appendix A2
Rectisol Process
6. Process Economics
Available capital costs and utility requirements for the Rectisol process
are summarized in Tables A2-5 and A2-6, respectively. Tabulated capital costs
are primarily conceptual design cost estimates while tabulated utility
requirements are published data for existing units. It should be noted that
the cost of a Rectisol unit is influenced by a variety of considerations
including the feed gas flow rate, pressure, acid gas content, and heavy
hydrocarbon content, and the desired levels of selectivity and product
purity. Due to the number of variables and associated interdependencies of
these variables which influence cost, costs of Rectisol systems tend to be
highly case specific.
TABLE A2-5. CAPITAL COSTS FOR RECTISOL ACID GAS REMOVAL UNITS
Selective
Nonselective
Dry
Feed Gas,
kmol/hr
6,100
96,384
52,786
57,574
Total
Pressure,
MPa
7.8
2.9
2.8
2.8
CO,,
vol %
35
28.9
31.4
34.2
HjS,
vol %
0.25
0.25
0.135
0.13
Capital
Cost, ilO«
(adj. to
1980 basis)
13.5*
150.6
91.8°
81.9°
Reference
24
25
23
22
aThe feed gas to this unit does not contain heavy hydrocarbons. Cost includes
refrigeration unit, erection, and plant startup. This is the same unit which
is identified as Plant 5 in Table A2-6.
Data are based upon a conceptual design cost estimate. Details of the cost
estimate are not available. The feed gas to this unit does not contain heavy
hydrocarbons.
CData are based upon a conceptual design cost estimate. The feed gas to this
unit contains heavy hydrocarbons. Reported cost includes naphtha and methanol
recovery and erection. It is not specified whether the costs for a
refrigeration unit and unit startup are included.
A2-26
-------
TABLE A2-6. UTILITY REQUIRMENTS FOR RECTISOL ACID GAS REMOVAL UNITS
to
I
to
Selective Rectisol1
Flow Rate, knol/hr
Pressure, MP.
Electric Power, kfh/kmol
Low Pressure Stoim, MJ/kmol
Cooling Water, MJ/knol
Stripping Nitrogen, bnol/knol
Makeup Methanol, kg/kmol
Refrigeration, KI/)uaol
(at 227 to 235 I)
Plant 1
3692-3992
3.2-3.3
0.14-0.15
5.14-5.54
1.20-1.42
0.067-0.072
0.0085-0.0092
2.09-2.29
Plant 2
6112
7.3
0.31
3.44
6.43
0.031
0.0057
Included
above In power
and cooling
water
Plant 3
7112
3.0
0.18
4.16
1.92
0.048
0.012
1.90
Plant 4
6350
3.3
0.57
5.09
9.52
0.067
0.0079
Included
above in power
and cooling
wa ter
Plant 5
6100
7.8
0.168
2.77
1.91
0.043
0.007
1.24
Non-Selective
Rectlsol6
Plant 6
16993
2.6
No data
3.27
0.682
No data
0.013
No data
Plant 1 is a refinery producing hydrogen by partial oxidation of oil; shift conversion occurs prior to acid gas removal (2,6).
Refer to Table A2-1 for performance data.
Plant 2 produces ammonia synthesis gas from crude hydrogen generated by partial oxidation; shift conversion occurs prior to
acid gas removal (3). Utility requirements reflect the use of compression refrigeration. Refer to Table A2-1 for performance
data.
Plant 3 produces ammonia synthesis gas by coal gasification; shift conversion follows hydrogen solfide removal but precedes
carbon dioxide removal (2,6). After sulfur removal the gaa is increased from 3 MPa to 5 MPa by compression; the additional
power required for compression is not included in the tabulated electric power requirement. Tabulated data are based on gas
flow rate after shift conversion. Refer to Figure A2-3 for process flow diagram, and to Table A2-1 for performance data.
Plant 4 uses a Rectisol unit for purification of hydrogen from partial oxidation of heavy crude oil; shift conversion occurs
prior to acid gas removal (3).
Plant 5 produces ammonia synthesis gas by partial oxidation of oil; shift conversion occurs prior to acid gas removal.
Approximately 62% of the Incoming carbon dioxide la provided as a carbon dioxide fraction containing less than 1.5 ppmv sulfur
for urea production (24).
b
Plant 6 is the SASOL coal gasification facility (17). Refer to Figure A2-5 for process flow diagram, and Table A2-3 for
performance data.
-------
Appendix A2
Rectisol Process
7. References
1. Kohl, A. and F. Reisenfeld. Gas Purification. Gulf Publishing Co.,
Houston, Texas, 1974.
2. Ranke, G. Acid Gas Separation by Rectisol in SNG Processes. Linde AG,
Munich, Germany. Copy of presentation obtained through Letepro
Corporation, New York, N.Y.
3. Scholz, W.H. Rectisol: A Low-Temperature Scrubbing Process for Gas
Purification, Advances in Cryogenic Engineering, Vol. 15, 1969.
4. Maddox, R.S. Gas and Liquid Sweetening, Campbell Petroleum Series, 1974.
5. Zee, C.A., J. Clausen, and K.W. Crawford. Environmental Assessment:
Source Test and Evaluation Report, Koppers-Totzek Process. EPA-600/7-81
009. January 1981.
6. Lotepro Corporation brochure.
7. Lee, K.W., W.S. Seames, R.V. Collings, K.J. Bombaugh, andG.C. Page.
Environmental Assessment: Source Test and Evaluation Report - Lurgi
(Kosovo) Medium-Btu Gasification, Final Report EPA-600/7-81-142.
August 1981.
8. D.S. Environmental Protection Agency. Environmental and Engineering
Evaluation of the Kosovo Coal Gasification Plant, Yugoslavia. Symposium
Proceedings: Environmental Aspects of Fuel Conversion Technology, III,
September 1977, Hollywood, Florida. EPA-600/7-78-063. April 1978.
9. D.S. Environmental Protection Agency. Comparison of Environmental Design
Aspects of Some Lurgi-Based Synfuels Plants. Symposium of Environmental
Aspects of Fuel Conversion Technology, Denver, Colorado, October 26-30,
1981. EPA-600/9-82-017. September 1982.
10. Trials of American Coals in a Lurgi Gasifier at Westfield, Scotland.
Woodall-Duckman, Ltd., Sussex, England. ERDA R&D Report No. 105, 1974.
11. Cameron Synthetic Fuels Report. Rocky Mountain Division, The Pace
Company Consultants & Engineers, Inc. Volume 18-Number 4, December 1981.
12. D.S. Environmental Protection Agency. Evaluation of Background Data
Relative to New Source Performance Standards for Lurgi Gasification.
EPA-600/7-77-057, June 1977.
A2-28
-------
Appendix A2
Rectisol Process
13. U.S. Department of Energy. Final Environmental Impact Statement: Great
Plains Gasification Project, Mercer County, North Dakota. Vol. I.
August 1980.
14. Hochgesand, G. Rectisol and Purisol. Industrial and Engineering
Chemistry, Vol. 62, No. 7. July 1970.
15. Fleming, O.K. Acid Gas Removal Systems in Coal Gasification. Ammonia
from Coal Symposium. Tennessee Valley Authority. May 8-10, 1979.
16. Engelbrecht, A.D. and L.J. Partridge. Operating Experience on a 1000-
ton/day Ammonia Plant at Modderfontein. Ammonia from Coal Symposium.
Tennessee Valley Authority. May 8-10, 1979.
17. U.S. Environmental Protection Agency, Emission Standards and Engineering
Division. Control of Emissions from Lurgi Coal Gasification Plants. EPA-
450/2-78—12 (OAQPS No. 1.2-093). March 1978.
18. Bombaugh, K.J. and W.E. Corbett. Kosovo Gasification Test Program
Results - Part II, Data Analysis and Interpretation. Symposium on
Environmental Aspects of Fuel Conversion Technology, IV. Hollywood,
Florida. April 17-20, 1979.
19. Data provided to EPA's Industrial Environmental Research Laboratory,
Research Triangle Park, N.C. by South African Coal, Oil and Gas Ltd.
(SASOL). November 1974.
20. U.S. Department of the Interior - Bureau of Reclamation. Final
Environmental Impact Statement. Western Gasification Company (WESCO)
Coal Gasification Project and Expansion of Navajo Mine by Utah
International Inc., New Mexico. Vol. I, II. January 14, 1976.
21. Review comments provided to TRW by Linde AG, April, 1982.
22. Oak Ridge National Laboratory, Oak Ridge, Tennessee. Liquefaction
Technology Assessment - Phase I: Indirect Liquefaction of Coal to
Methanol and Gasoline Using Available Technology. ORNL-5664. February
1981.
23. U.S. Department of Energy. Research Guidance Studies to Assess Gasoline
from Coal by Methanol—to—Gasoline and Sasol-Type Fischer—Tropsch
Technologies. FE-2447-13. August 1978.
24. Information provided to TRW by Lotepro Corporation, January 1983.
25. U.S. Department of Energy. Conceptual Design of a Coal to Methanol
Commercial Plant. Volume IVA. FE-2416-35.
A2-29
-------
APPENDIX A3
SELEXOL ACID GAS REMOVAL PROCESS
The Selexol process was developed and is licensed by the Norton Company.
The process purifies gases through physical absorption of H2S, C0a, COS, CS2,
and mercaptans. This process can be applied to treat gas streams such as sour
natural gas; raw gas from the gasification of coal, oil, and light hydro-
carbons; hydrogen-rich shift gas and excess synthesis gas from the gasifica-
tion of coal liquefaction bottoms slurry; synthesis gases from steam reforming
or partial oxidation of hydrocarbon liquids; and refinery gases.
The Selexol solvent used in removing acid gas is the dimethyl ether of
polyethylene glycol (DMPEG). The Selexol process can be designed for non-
selective or selective operation. In the nonselective application, hydrogen
sulfide and carbon dioxide are removed simultaneously and only one acid gas
stream is produced from the regeneration of the DMPEG solvent. In the selec-
tive application, hydrogen sulfide and carbon dioxide are removed sequentially
and two acid gas streams are produced. One contains essentially all the
hydrogen sulfide while the other contains the bulk of the C02 and a small
amount of sulfur compounds. The advantage of selective acid gas removal is
the generation of an H2S-rich acid gas stream which can be handled in Claus
sulfur recovery plants.
The solubility of several gases in the Selexol solvent and the physical
properties of the solvent are presented in Figure A3-1 and Table A3-1,
respectively. The solubility of acid gas in the Selexol solvent is directly
proportional to the partial pressures of acid gases in the feed gas. There-
fore, the higher the partial pressures of the acid gases, the greater the
amount of acid gas that can be absorbed by the solvent (3). The Selexol
A3-1
-------
250
200 _
c
CU
150
E
V)
e
„ 100
•rl
3
O
en
50 .
0 0.5 1.0 1.5 2.0 2.5 3.0
Gas Partial Pressure, MPa
3.5
4.0
Figure A3-1. Solubility of selected gases in Selexol solvent (1)
A3-2
-------
Appendix A3
Selezol Process
TABLE A3-1. PHYSICAL PROPERTIES OF SELEXOL SOLVENT (2)
Property Value
Freeze point, K 251 to 244
Flash point, K 424
Vapor pressure at 298 K, mm.Hg <0.01
Specific heat at 298 K, cal/g-°C 0.49
Density at 298 K, gm/cc (Ib/gal) 1.03 (8.60)
Viscosity at 298 K, cp 5.8
Odor very mild
Toxicity nil
process is most economical when the partial pressure of acid gases in the feed
stream is above 1.4 MPa. When acid gas partial pressures fall below 1.4 MPa,
other processes may be preferred (3). The Selexol process is particularly
suited for application to some synfuel plants, especially since most of the
coal gasification and liquefaction processes operate at high pressures and
high acid gas partial pressures. The operating pressure and acid gas content
for selected sour gas streams from some coal gasification and liquefaction
processes are listed in Table A3-2.
The requirements for the acid gas removal units in coal gasification/
liquefaction plants vary according to the specification of the product gas pro-
duced. For example, the SNG plants require H2S removal to below 1 ppmv
because of the sulfur-sensitive methanation reaction catalyst (1). Also,
removal of C02 is often required to increase the heating value of the product
gas. SNG product gas specifications call for a 1 vol % or below limit for C02
(1). The Selexol process is commercially proven in natural gas and refinery
applications to handle all the above situations.
A3-3
-------
Appendix A3
Selexol Process
TABLE A3-2. CHARACTERISTICS OF SOME POTENTIAL SELEXOL FEED GASES
FROM SYNFDELS FACILITIES
Process
Solvent Refined
Coal- I
Solvent Refined
Coal-II
H-Coal
Process
Stream
H2-rich Shift Gas
Hj-rich Shift Gas
Excess Synthesis Gas
H2-rich Shift Gas
Excess Synthesis Gas
Pressure,
MPa
5.1
8.0
8.2
4.1
4.6
Acid Gas
Vol% HaS
0.82
0.46
0.38
0.42
0.59
Content
Vol% C02
39.6
27.0
0.65
38.3
7.1
Exxon Donor Solvent
(Market Sensitivity
Flexibility Design)
Lurgi (SASOL I)
Koppers-Totzek
(SNG Design)
H2-rich Shift Gas
Hj-rich Shift Gas
Synthesis Gas
H2-rich Shift Gas'
6.2
2.6
3.0
3.0
0.30
0.30
1.0
0.0007
25.1
28.8
8.5
32.7
Assumes selective removal of HaS prior to shift conversion.
Data Source: Conceptual commercial plant designs for SRC-I, SRC-II, H-Coal,
and Exxon Donor Solvent processes (4,5,6,7), Lurgi (8), and
process model for K-T process developed by TRW.
1. Process Description
Nonselective Selexol Process
One application of the nonselective Selexol process is in the bulk remov-
al of C0a from a gas with a high C02 to H2S ratio. Figure A3-2 shows a schem-
atic flow diagram for the nonselective removal of CO.,. The gas to be purified
is passed upward through a tray or packed tower, countercurrent to the down-
ward flow of the regenerated Selexol solution. The solvent enters the
A3-4
-------
SELEXOL
I PUMP
u>
i
TREATED
GAS
INLET
GAS
TO INLET
GAS COOLER
TO INLET
GAS COOLER
TURBO
EXPANDER
ABSORBER
(TRAY OR
PACKED
TOWER)
INTERMEDIATE
PRESSURE
FLASH DRUM
TO INLET
GAS
RECYCLE GAS COOLER
COMPRESSOR
RECYCLE GAS
( ) PUMP
\_ SELEXOL
PUMP
LOW PRESSURE
\ FLASH DRUM
) (ATMOSPHERIC
PRESSURE)•
HIGH-PRESSURE
FLASH DRUM
HYDRAULIC
RECOVERY
TURBINE
Figure A3-2. Flow scheme for nonselective Selexol C02 removal (1)
-------
Appendix A3
Selexol Process
absorber at slightly below ambient temperature. The treated gas leaves the
absorber at the top and may be passed through a heat exchanger to cool the
feed gas stream.
The rich Selexol solvent from the absorber is flashed in several stages
to remove most of the acid gas from the solvent. Gases from the first high
pressure flash are usually sufficiently rich in hydrocarbons to justify recom—
pressing and recycling back to the absorber. Hydraulic turbines are often
utilized to recover energy from the high pressure streams. C02 vapors are
also expanded through a turboexpander to recover additional power. Vapors
from the intermediate and low pressure flash tanks are vented to the atmos-
phere.
The nonselective Selexol process can also be applied to treat gases with
a relatively high H2S to C02 ratio. Enrichment of the H2S content of the acid
gas stream from solvent regeneration for Claus plant feed is not needed in
this case. Figure A3-3 shows a schematic flow diagram for the nonselective
removal of H2S. The gas to be purified is passed upward through a tray or
packed absorber where it first comes into contact with semi—lean Selexol sol-
vent feed in the lower section of the absorption tower. Most of the H2S and
C02 are removed at this point. In the upper part of the absorber, stripped
lean solvent contacts the feed gas and removes the remaining H2S and C02 to
the desired specifications. The rich Selexol solvent is flashed to remove most
of the acid gas. Flash gases from the high pressure flash tank are recycled
to recover hydrocarbon gases. Part of the solvent is then returned to the
absorber. A semi-lean solvent and the remaining solvent are fed to a
stripper/regenerator. Air stripping is preferred when the regenerated gases
contain only CO and are suitable for venting to the atmosphere without
A3-6
-------
H-
0-3
i-i
tt>
I
OJ
o
en
o
nr
§
m
Ml
o
3
o
en
ro
CO
(D
n>
x
o
§
O
-------
Appendix A3
Selexol Process
farther processing. However, if the regenerated gas is to be fed to a sulfur
recovery unit, steam stripping will be necessary to produce a concentrated H^S
stream. The stripper overhead and flashed acid gas from the absorber comprise
the feed to the sulfur recovery unit.
Selective Selexol Process
The selective Selexol process produces two acid gas streams. One con-
tains essentially all the HaS while the other is basically C02 and air. The
solubility of HaS in the Selexol solvent is approximately nine times greater
than the solubility COZ which allows preferential absorption of HaS. The flow
scheme for a selective Selexol process is shown in Figure A3-4. The absorp-
tion of HaS and C0a takes place in two separate columns. The first column
removes essentially all the HaS along with some COS and other sulfur com-
pounds. The overhead gas from the first absorber is sent to the C02 absorber
where the bulk of the CO, along with additional COS and other sulfur compounds
are removed. The Selexol solvent (rich in HaS) from the bottom of the first
absorber is fed to 'the HaS stripper unit, where the absorbed H2S is removed by
heating and steam stripping. The HaS stripper overhead gas is fed to the sul-
fur recovery unit. In the C02 stripper, air or inert-gas is used to strip
C0a from the solvent. Overhead gas from the C0a stripping column is normally
vented to the atmosphere. The rich Selexol solvent from the H^S absorber is
often flashed at an intermediate pressure to recover and recycle hydrocarbons
absorbed by the solvent (Personal communication with J.B. Grant, May 27,
1981).
2. Process Applicability
In synfuel plants, the Selexol process would most likely be used in cases
where the partial pressure of acid gases in the feed gas is higher than 1.4
MPa. The ability to selectively remove H2S from gases containing C02 gives
A3-8
-------
ACID GAS TO
SULFUR RECOVERY
FEED GAS
SOUR WATER
TREATED GAS
C02 RICH GAS
^COMPRESSED
* AIR
SOLVENT
FLASH DRUM
Figure A3-4. Flow scheme for selective Selexol process (9)
-------
Appendix A3
Selezol Process
the Selezol process an advantage over chemical absorption processes such as
the nonselective DEA process. The Selezol process is preferred in many
commercial applications due to the ability to effectively remove organic
sulfur, naphtha-range hydrocarbons, and water vapor from the gas, without any
degradation or loss of the solvent. The Selezol process is also currently
favored in many applications due to its overall simple operation.
The Selezol process has also demonstrated applicability in the treatment
of natural gas, where performance has proven to be dependable, flexible, and
very economical in a number of natural gas plant facilities (10).
3. Process Performance
The performance of the Selezol process depends on the imposed operating
conditions. The process works best when the acid gas partial pressure is
above 1.4 MPa and temperature is below 310 K. The temperature of the Selezol
solution entering the absorber is usually kept below 293 K. Product gases
from a Selezol absorber typically contain less than 1 ppm total sulfur (11).
The HaS-rich stream leaving a selective Selexol unit can generally be pro-
cessed by a Glaus unit. The sulfur content in the 002 stream may be low
enough to allow venting to the atmosphere without further treatment (12).
A typical Selezol process design involves absorption in a 24-tray tower.
The amount of HaS absorbed in the absorption tower is a function of the liquid
to gas molar ratio (L/G) in the tower, the partial pressure of H2S, contact
temperature, number of absorption stages, and leanness (amount of resid-
ual H2S) of the Selezol solvent. Removal of H2S may be enhanced by increasing
the solvent circulation rate and holding other variables constant. Generally,
the solvent circulation rate is selected such that the value of the absorption
factor A for H2S lies in the range of 1.1 to 2.0. The absorption factor is
A3-10
-------
Appendix A3
Selezol Process
defined as the ratio of the slope of the operating line L/G to that of the
equilibrium curve m; A = L/mG. Complete removal of HaS is possible only when
the absorption factor for HaS is greater than unity. A design solvent loading
of greater than 75 m3 gas per m* solvent is often used in the treatment of
natural gas (10).
In the selective Selexol process, the degree of H2S selectivity designed
into the system governs the degree of COS removal in the HaS absorption tower.
COS is intermediate in solubility to HaS and C0a. Based on solubility data it
is estimated that the absorption factor for COS is only about 27 percent of
the absorption factor for HaS for any specific L/G ratio. It can be shown
that about 50 to 70 percent of the COS present in the feed gas will be emitted
with the COj-rich stream. Thus, venting of the C0a-rich stream is only prac-
tical when the COS content of the feed gas is sufficiently low.
In order to substantially reduce COS emissions from the vent stack, the
Selexol circulation rate in the HaS absorption tower may be increased. How-
ever, this will result in additional C0a absorption and an acid gas stream
unsuited for feed to a Claus plant. Another possibility is to further treat
the C0a vent gas for COS removal. However, the additional costs incurred may
make the selective Selexol process economically unattractive. Process designs
have been developed which will absorb considerable amounts of COS while still
producing a suitable feed gas for a Claus plant (15-25 percent HaS gas) (1).
Knowing the inlet gas concentrations of C0a and COS, the COS content of
the C0a vent gas can be approximately calculated if the absorption factor for
HjS is given or estimated based on separation requirements. An equation to
estimate the COS content of the C0a-rich vent gas was developed by TRW using
the Kremser-Souders-Brown equation (13) and the relative solubilities of COS,
C0a, and HaS in the Selexol solvent:
A3-11
-------
Appendix A3
Selezol Process
Mole fraction of COS in C02 vent gas =
- 0.268A yrn
-------
0.9 i-
0.8
co
n
0.7
0.6
1.0 1.25 1.5
Absorption Factor "A" for
1.75
2.0
Figure A3-5. "R" factor as a function of absorption factor A for
A3-13
-------
TABLE A3-3. PERFORMANCE DATA FOR NONSELECTIVE SELEXOL SYSTEMS
Deil|n and Operation
Varliblil
Feed Gn
Flint Capacity, 10* .'/day
Feed Gil Aialyili:
H,S, percent by vol.
COj , percent by vol .
COS, pp.v
CS,. W»v
Outlet Git Anilyiii:
HtS. ppaiv
C0}, percent by vol.
COS, pp«v
CS,, ppmv
Abiorber :
Inlet Prenure, HP a
Teaperature K
Steia ConauaptloD
Rcboller
Pu.pi
Plant A Plant B Plant C Flint D Plant E Flint F Plait Q
Syntheili Syntneiii Souf Fuel Nitnril Nitnril Natural Natural
Gil Gil Gil On On Gil GII
7.4 9.6 4.5 5.7 7.8 3.7 7.1
0.23 0.03 0.5» 0.0013 0.0016 130 130
6.76 5.09 7.1 27.1 43 11.0 3.5
70 4 370 0 0 0 0
No Ditl No Ditl No Ditl 0000
<4 <4 44444
0.69 4.1 5.8 .03 3.5 2.5 3.0
No Ditl No Ditl 200 0000
No Ditl No Ditl No Ditl 0 0 0 0
7.7 5.« 4.6 6.2 7.0 6.9 6.9
310 310 310 310 No Diti No Data No Ditl
5720 k|/hr 31,900 kf/hr No Diti 0.25 GJ/hr No Diti No Diti No Data
(100 MPi and (1.4 HFi and
110 GJ/hr) 63 GJ/hr)
Motor 15,000 k|/hr No Diti No Diti No Data No Ditl No Data
Drive (4.1 Ufa and
640 I Inlet;
1.4 MPa ind
530 I eihiuit)
Motor 92,500 k|/hr No Diti No Dati No Diti No Ditl No Diti
Drive (4.1 KPa and
640 I inleti
1.4 MPa and
530 I oihauit)
Diti Source. Plant! A and B (1), Flint C (Private coKBualci tion, J.B. Grant, Utf, 1981), Flint D (14), Plinti £ to G
-------
TABLE A3-4. NONSELECTIVE SELEXOL PROCESS MODEL*
Constituent
H2
°.
N2 + Ar
CO
coz
CH
4
C2H4, C2H,
HjS
COS
HCN
NH3
H2°
Temperature, K
Pressure, MPa
Feed Gas
%b.
100
100
100
100
100
100
100
100
100
100
100
c
305
2.8
H S
Acid1 Gas Treated Gas
_b _b
W 7W
0.01 99.99
100
0.06 99.94
0.07 99.93
Design Dependent
0.55 99.45
16.69 33.31
100 1.0 ppmv
100
100
100
c c
320 305
0.18 2.7
o
Developed from data in Reference 12 and relative solubilities of gases in
Selexol solvent.
Percentages of feed gas.
Saturated with water at temperature and pressure of stream.
A3-15
-------
I
M
(^
TABLE A3-5. SELECTIVE 8,8 AND CO, REMOVAL
ABSORPTION MODE, T10 STAGES
- MATERIAL BALANCE SELEXOL
IGT Steam-Oxygen
Hygas (vol%)
H,
CO
CO,
CH4
C,H,
N,
H,S
COS
H,0
c,+
kg-»ole/hr
«g/hr
Temperature, E
Pressure, MPa
CO, vent gas,
Mg/hr
Feed Gas
38.40
11.23
33.17
15.52
1.21
0.24
0.005
.08
0.15
20,800
450
310
7.6
Treated Gas
57.82
16.87
0.80
23.03
1.47
0.007
13,800
143
290
7.5
287
BC8 Bi-Gas
(vol*>
Feed Gas
39.40
12.31
34.50
13.33
0.17
0.20
0.008
0.08
23.600
512
310
7.8
Treated Gas
60.00
18.70
1.09
19.95
0.26
0.003
15,500
158
290
7.6
335
Syn thane
(vol%)
Feed Gas
29.15
7.40
46.50
14.56
1.79
0.20
0.29
0.003
0.10
0.01
23,700
564
310
6.4
Treated Gat
54.87
13.86
1.52
26.73
2.65
0.37
0.007
12,600
136
290
6.0
360
Lurgi*
(vol%)
Feed Gas
42.02
11.58
34.44
10.63
0.51
0.30
0.26
0.26
25,000
534
310
2.6
Treated Gas
62.80
17.31
2.79
15.88
0.77
0.45
0.005
16,800
172
290
2.5
351
Feed-gas composition and condition estimated for Lurgi process (1),
-------
TABLE A3-6. NONSELECTIVE SELEXOL PROCESS MODEL*
Constituent
Ha
o,
Na + Ar
CO
CO
1
CH4
CiH«- C»H«
H*S
COS
HCN
NHj
HjO
NO
X
CS
so,
Temperature, K
Pressure. MPa
Feed Gas
*"
100
100
100
100
100
100
100
100
100
100
100
c
305
7.6
HjS-Rich CO^-Rich
Acid Gas Acid Gas Treated Gas
b b b
0.023 0.33 99.65
100
0.57 99.43
0.10 0.79 99.11
14.28 Design Dependent
0.41 1.07 98.52
1.57 15.12 83.31
99.99 0.01
30-50 50-70 1.0 ppmv
100
75 25
c c c
100
100
100
320 320 305
0.17 0.1 7.3
Based on data in Reference 1 and relative solnbilities of gases in Selezol
.solvent.
Percentages of feed gas.
Saturated with water at temperature and pressure of stream.
A3-17
-------
Appendix A3
Selexol Process
Data published by Dravo Corporation for a conceptual SNG plant using the
selective Selexol process to remove H2S and C0a from a 3.5 MPa gas stream show
an H2S content of less than 0.1 ppm in the treated gas (11). C0a was reduced
from 35 percent in the feed gas to 11 percent in the treated gas. Solvent
loss was estimated at 8 kg/10* m3 of feed gas (1).
4. Secondary Waste Generation
Two main gas streams are generated from nonselective Selexol units. The
first is the purified gas stream and the second is the concentrated acid gas
stream. The acid gas stream requires further treatment in a sulfur recovery
unit before discharge to the atmosphere.
In the application of the selective Selexol process, three gas streams
are generated: 1) a treated gas stream, 2) an H^S-rich gas stream from the
H2S stripper, and 3) a C0a~rich gas stream from the CO^ stripper. The H2S-
rich gas stream usually contains sufficient H2S (15 to 25 percent) to be
treated in a Claus unit. The CO^—rich gas stream generally contains very
little or no suflur and may often be vented directly to the atmosphere. How-
ever, if the C02-rich gas stream contains sulfur compounds (usually COS) over
the specified emission levels, further treatment may be necessary.
No solvent reclaimer is required since there is no degradation of the
solvent. Therefore, no spent solvent will be generated from the Selexol unit
(14). However, a sour waste stream may be generated from regeneration of the
Selexol solvent, depending on the water content of the feed gas and the
specific design of the Selexol process.
A3-18
-------
Appendix A3
Selezol Process
5. Process Reliability
Commercial experience shows that the Selexol process is highly reliable
and simple to operate. Operating difficulties are minimized to a great extent
because of the noncorrosive and nonfoaming properties of the Selexol sol-
vent. However, corrosion can still occur due to equipment exposure to the
wet acid gases. Corrosion in Selexol plants can be controlled by limiting the
solvent circulation rate to ranges which have proven to minimize corrosion or
erosion problems in the past. Corrosion test probes may be installed to
detect corrosion problems for correction during scheduled maintenance periods.
Selexol solvent maintenance is important to the reliability and oper-
ability of a Selexol plant. The build-up of materials such as solid particul-
ates, heavy hydrocarbons, compressor oil, etc., may cause operating problems
(10). Solids and free liquid may be filtered from the inlet gas. A side-
stream filter may be utilized to remove solids which may have developed in the
process.
6. Process Economics
In general, plants based on physical methods of treating and purifying
gas streams cost less to build and operate than plants based on chemical
absorption methods (2). Table A3-7 presents the economics of the Selexol
process relative to those of both a MEA unit (20 percent MEA solution) and a
hot potassium carbonate unit in the treatment of natural gas (1).
In a study of commercial design concepts for coal gasification plants
(16), nonselective Selexol acid gas removal combined with the Stretford pro-
cess for sulfur recovery and the FMC, Inc. process for flue gas desulfuri-
zation (dual alkali) showed the lowest capital and operating cost for western
coals. For eastern coals the same study showed selective Selexol (combined
with a Claus unit and incineration of Claus plant tail gas) to be the most
economical process.
A3-19
-------
TABLE A3-7. RELATIVE ECONOMICS OF NATURAL GAS TREATING
SOLVENTS* (1)
Relative Economics. % of MEA Cost
20% Aqueous Hot Potassium
Parameter MEA Carbonate Selexol
Capital Cost 100 80 70
Direct Annual Operating
Costs
Electricity 100 95 20
Steam 100 70 10
Process Water (cooling) 100 95 25
Total Direct 100 75 40
Indirect Annual Operating 100 80 75
Costs
Total Annual Operating 100 75 50
Costs
aPlant operating conditions: capacity, 2.8 million standard mj/day; feed gas
pressure, 6.9 MPa; inlet C0a. 30 mole %; and inlet HaS, 320 ppmv.
Includes cost items such as losses, operating labor, and maintenance, along
with steam, electricity, and water costs.
A3-20
-------
Appendix A3
Selexol Process
Capital costs for nonselective Selexol removal of H2S and C02 have been
published for two coal-derived SNG plants and are shown in Table A3-8. Costs
have been escalated to first quarter of 1980. Capital cost information for
selective Selexol removal of H2S and C02 for four SNG plant designs are
presented in Table A3-9.
The operating costs for any Selexol unit can be estimated by determining
the Selexol circulation rate. The costs of regeneration and pumping of the
Selexol solution are usually the major contributors to the overall operating
cost. Energy costs of compressing and recycling hydrocarbon-rich flash gases
back to the absorber are also major contributors. These energy requirements
are mainly affected by the operating pressure of the absorber.
In many cases, the solvent regeneration method is greatly influenced by
the presence of H2S in the stripper inlet gas. If H2S is present, air strip-
ping cannot be used because the oxygen would react with H2S to produce elemen-
tal sulfur in intolerable amounts. Therefore in the presence of H2S, steam
stripping of the solvent is usually indicated (17).
Allied Chemical Corporation provided utilities requirements for selective
removal of H2S from a low-Btu gas and these are presented in Table A3-10. The
unit operated at 2.2 MPa and treated a gas containg 0.61 vol% and 8.9 vol%
C02. The Dravo Corporation presented typical utility requirments for the
selective removal of H2S and C02 from a 3.5 MPa gas stream containing 0.5 per-
cent H2S and 35 percent C02. The treated gas in this case contained less than
0.1 ppmv H2S and 11 percent C02. The estimated operating requirements for
28,000 m3 of gas treated by this system are: 900 kWhr electricity, 1360 kg
steam, 130 m3 cooling water, and 2.34 kg solvent loss (1). Operating require-
ments for additional selective Selexol designs are presented in Reference 1.
A3-21
-------
TABLE A3-8. NONSELECTIVE SELEXOL - CAPITAL COSTS (mid-1980) FOR TWO
COMMERCIAL PLANT DESIGNS* (1)
Purchased Equipment
Columns
Reactors
Pressure vessels
Tanks
Exchangers
Pumps
Compressors and blowers
Refrigeration package
Cartridge filters
Miscellaneous
Total Purchased Equipment
Total Installed Cost
IGT Steam - Iron Hygas,
*
3,016,800
627,900
649,000
9,800
2,621,900
771,400
1,970,800
4,441,400
8,500
734.100
14,851,600
52.8 million
Conoco CO 2
Acceptor,
*
840,400
633,500
76,000
11,300
2,240,400
330.800
254.800
1.385,900
2,800
277,300
6,053,200
25.3 million
aFeed gas—IGT Steam-Iron Hygas: 13,700 kg-moles/hr; 7.7 MPa
Conoco CO Acceptor: 17,800 kg-moles/hr; 5.8 MPa
A3-22
-------
TABLE A3-9. SELECTIVE B^S AND CO^ REMOVAL - CAPITAL COST BREAKDOWN
(mid-1980) FOR FOUR COMMERCIAL PLANT DESIGNS (1)
Purchased Equipment
Columns
Reactors
Pressure vessels
Tanks
Exchangers, Shell
and Tube
Exchangers, Air Coolers
Pumps
Compressors and blowers
Refrigeration package
Cartridge filters
Jet ejectors
Miscellaneous
Total Purchased Equipment
Total Installed Cost
IGT Steam-
Iron Hygas,
i 10*
6.30
0.70
1.10
0.02
2.35
1.14
1.86
4.23
4.18
0.04
0.02
1.11
23.1
76.0
BCR
Bi-Gas,
i io«
7.40
0.70
1.50
0.02
3.16
0.90
1.95
4.18
4.18
0.01
0.02
2s'.3
80.2
Syn thane
$ 10«
8.60
0.60
2.30
0.02
3.04
1.60
2.48
4.41
4.32
0.04
0.02
1.36
28.8
89.4
, Lnrgi,
i io«
4.20
0.40
2.40
0.02
2.91
1.21
2.10
4.24
4.18
0.39
0.02
1.06
23.1
82.4
The costs are based on material balances and operating parameters shown in
Table A3-5.
A3-23
-------
TABLE A3-10.
OPERATING REQUIREMENTS FOR SELECTIVE SELEXOL REMOVAL OF
FROM A LOW-BTU GAS. (1)
Utilities
Regenerator heat load
Power
Cooling load
Chemicals
Selexol Solvent
Labor
Operators
18 GJ/hr
450 kW
12 GJ/hr
23 kg/day
1 man/shift
Capacity:
H2S in inlet gas:
C02 inlet gas:
Pressure:
117,000 kg/hr
0.61 vol %
8.87 vol %
2.1 MPa
A3-24
-------
Appendix A3
Selexol Process
7. References
1. Edward, M.S. H^S Removal Process for Low-Btu Coal Gasification, Oak
Ridge National Laboratory, ORNL/TM-6077. January, 1979.
2. Lemmon, A.W. Jr. B.C.B. Hsieth, and A.E. teller. Jr. Report on
Evaluation of Low-Temperature Cleanup System for Low Btu Gas. Battelle
Columbus Laboratories for the Tennessee Valley Authority Power Research
Staff, August 29, 1975.
3. Fluor Engineers and Constructor, Inc. Economics of Current and Advanced
Gasification Processes for Fuel Gas Production. EPRI AF-244. Research
Project 239, Final Report. July 1976.
4. U.S. Environmental Protection Agency. Preliminary Baseline Design
Package for the SRC-1 Demonstration Plant. December 19, 1980.
5. U.S. Department of Energy. SRC-II Demonstration Project Phase Zero.
Task Number 3. Deliverable Number 8. Vol. 2 of 4. Conceptual
Commercial Plant Description. July 31, 1979.
6. Grant, J. Material for EPA PCGD Preparation Supplied by Ashland
Synthetic Fuels, Inc. September 30, 1981.
7. DeGeroge, C.W. and M.R. Wise. EDS Information Response to EPA Requests
of September 9 and September 25, 1980, for Assistance in Preparation of
Direct Liquefication Pollution Control Guidance Document. November 18,
1980.
8. U.S. Environmental Protection Agency, Emission Standards and Engineering
Division. Control of Emissions from Lurgi Coal Gasification Plants. EPA-
450/2-78-012 (OAQPS No. 1.2-093). March 1978.
9. U.S. Depatment of Energy and Gas Research Institute. Sulfur Recovery in
a Coal Gasification Plant. FE-2240-50. August 1978.
10. Raney, Donald R. Remove Carbon Dioxide with Selexol. Hydrocarbon
Processing. pp. 73-75. April, 1976.
11. Dravo Corporation. Handbook of Gasifiers and Gas Treating Systems.
FE-1772-11. For the Energy Research and Development Administration,
Washington DC. February, 1976.
A3-25
-------
Appendix A3
Selexol Process
12. Christensen, K.G. and W.J. Stupin. Comparison of Acid Gas Removal
Processes. C.F. Braun and Co., Alhambra, California. FE-2240-49, April
1978.
13. Sherwood, T.K. and R.L. Pigford. Absorption and Extraction. McGraw-Hill
Book Publishing Co. 1952.
14. Judd, Donald K. Selexol Unit Saves Energy. Hydrocarbon Processing.
pp. 122-124. April, 1978.
15. Kohl, Arthur, and Fred Kiesenfeld. Gas Purification. Gulf Publishing
Company, Houston, 1972
16. Gas Research Institute, American Gas Association, Department of Energy
and International Gas Union. Proceedings of Tenth Synthetic Pipeline Gas
Symposium. October-November 1978.
17. Eickmeyer, A.G. and H.A. Gongriwala. The Role of Acid Gas Removal in
Synfuels Production. Presented at the Symposium of Gas Purification,
AICHE 1981 Spring National Meeting and llth Petrochemical Exposition,
Houston, Texas.
8. Personal Communication
Grant, J.B., Ashland Synthetic Fuels, Inc., to B. Bakshi, TRW
Environmental Division. May 27, 1981.
A3-2 6
-------
APPENDIX A4
BENFIELD ACID GAS REMOVAL PROCESS
1. Process Description
The standard Benfield process uses a hot potassium carbonate solution
containing additives, corrosion inhibitors, and activators to chemically
absorb C0a and H2S. Developed by Benson, Field, and coworkers at the U.S.
Bureau of Mines in the 1950s (1), the process is currently licensed by the
Union Carbide Corporation (2).
Simplified process configurations for the four major alternative Benfield
designs of the hot potassium carbonate (HPC) system are shown in Figures A4-1,
A4-2, A4-3, and A4-4. Starting with the simplest system in Figure A4-1, the
processes are modified such that product gas purities are improved (increased
removal of acid gases) through systems two, three, and four (3,4,5).
In the simplest design (Figure A4-1), lean/regenerated potassium carbo-
nate solution is fed to the top of the absorption column and the process gas
is fed near the bottom of the absorber. Intimate contact of the counter-
currently flowing streams is maintained by use of packed beds or trays.
Purified gas leaves the top of the absorber while the acid-gas rich solution
is fed to the top of the regenerator. Steam fed near the bottom of the regen-
erator strips the solution of absorbed gas which, with the steam, leaves the
regenerator at the top. The stripped solution is then recycled to the top of
the absorber.
In the split flow absorber (Figure A4-2), part of the regenerated solu-
tion is directly pumped to the absorber near its midpoint while the remainder
is cooled to between 340 K and 370 K and fed at the top of the absorber. The
cooler temperature reduces the acid gas equilibrium over the solution and thus
increases acid gas removal.
A4-1
-------
PURIFIED
GAS
ABSORBER
H2S
Figure A4-1. BenCield process with single stage absorber (2, 3)
A4-2
-------
PURIFIED
GAS
ABSORBER
H2S
Figure A4-2. Benfield process with split flow absorber (2, 3)
A4-3
-------
PURIFIED
GAS
ABSORBER
FEED
GAS '
H2S
Figure A4-3. Benfield process—two stage system (2, 3)
A4-4
-------
PURIFIED
GAS
ABSORBER
FEED
GAS"
Figure A4-4. Benfield HiPure process (5)
A4-5
-------
Appendix A4
Benfield Process
Figure A4-3 shows the two stage system in which the regenerator is also
in two sections. Most of the absorbent is drawn off at the midpoint of the
regenerator and fed to the midpoint of the absorber. The remainder undergoes
a more thorough stripping then is cooled prior to feeding to the top of the
absorber. This increased stripping permits larqer quantities of acid gas to
be absorbed prior to reaching acid gas equilibrium pressures over the absor-
bent.
For hiqh product purity, two independent absorbent streams are used. In
fact, the second/final absorption section in this Benfield HiPure process
(Figure A4-4) may use an absorbent other than a carbonate solution to meet
specific product requirements. In the independent countercurrent systems, the
hot carbonate solution flows through the first/bottom section of the absorber
and top section of the regenerator and is similar to the single stage process.
The trim solution is used for final purification at operating conditions best
suited for the desired application. This trim section is smaller than the hot
carbonate section and uses a lower flow rate and an excess of stripping steam.
This steam is used subsequently to strip the carbonate solution (2,4,6).
More recent modifications to these processes (Benfield LoHeat) have been
concerned with methods for conserving energy and have not changed the basic
process schematics or the acid gas removal efficiencies (7).
Historically, K2C03 solution was used to absorb C02 at temperatures below
315 K by the 1920s, and alkali carbonate solutions were used to absorb H2S
by 1938 (3). In the Benfield process, absorption temperatures were raised to
380 K which is near the atmospheric boiling point of the solution. Acid gases
are removed by contacting an aqueous solution of 25 to 35 percent K2C03-
containing corrosion inhibitors and activators to catalyze the rates of
absorption and desorption.
A4-6
-------
Appendix A4
Benfield Process
Acid components react reversibly with the KjCOj as follows:
K2C03 + CO, + H20 = 2KHCO, (1)
K2CO, + H2S = KHCO, + KHS (2)
K2CO} + RSH = KHCO, + RSK (3)
Apparently the absorbing solution catalyzes the hydrolysis of COS and CS2:
COS + H20 = C02 + H2S (4)
CSa + 2H20 = CO, + 2H2S (5)
such that the above reactions occur at rates that are orders of magnitude
faster than in water. C02 and H2S are subsequently absorbed (8).
Additionally, ammonia (NH,) is absorbed in an analogous manner to that
of the mercaptans (RSH). S02 and HCN are assumed to react as follows:
K2C03 + S02 = K2S03 + C02 (6)
K2C03 + HCN = KCN + KHCO, (7)
and have the same order of magnitude removal efficiencies as for H^S. Heavy
hydrocarbons and particulates should not react with the carbonate solution but
may cause foaming (4,5).
Using C02 as a reference, relative rates and capacities of absorption for
a few selected substances are listed in Table A4-1. With the proper selection
of operating conditions and additives, the rate of absorption of H S can be
increased to over 50 times that of C02 instead of the four times as shown in
Table A4-1 (4,5). Thus, the system can be utilized to selectively remove H2S
from a large excess of C02.
A4-7
-------
Appendix A4
Benfield Process
TABLE A4-1. RELATIVE ABSORPTION CAPACITY AND RATE INTO
HPC SOLUTION AT 380 K (4)
Component
co^
H*S
COS
csa
CHjSH
NH3
Relative
Capacity
1
1.41
Hydrolyzes
Hydrolyzes
0.03
0.10
Relative
Rate5
1
3.6
0.36
0.10
1.2
3.5
Q
Capacities measured to equilibrium partial pressure of 14 kPa
,in volume of gas to volume of absorbent.
Rates measured at solution loading to the equilibrium partial
pressure of 14 kPa.
2. Process Applicability
It is estimated that some 600 HPC systems of all types are in operation
worldwide (7). The staged system is currently the most popular Benfield
configuration. Similar systems are also fabricated and sold by Vetrocoke,
Catacarb. and Carsol. However, the most successful has been the Benfield
system, with over 230 applications. This process is usually used for the
removal of C0t from synthesis gas streams produced by the catalytic reforming
of natural gas (9).
Principal applications of the Benfield system have been in the production
of (3,7,10):
-------
Appendix A4
Benfield Process
ammonia and/or hydrogen O150 plants)
town gas O180 plants)
natural gas (18 plants)
SNG from naphtha (14 plants)
specialty gases (28 plants)
The plant numbers for the last two categories are 1974 data and may be
significantly changed. Also, for a number of years the HPC process has been
applied to coal gasification at the Lurgi plant in Westfield, Scotland. As of
1981, there were 15 Benfield HiPure units in operation and 30 Benfield LoHeat
processes installed (9). This LoHeat process involves plant modifications to
increase energy efficiency with minimal effect on capital cost or efficiency.
In addition to the wide variety of commercial applications in which
the HPC process is used, the application of HPC to fuel conversion processes
has been studied extensively. These studies include (8):
• Production of SNG from coal or heavy oil.
• Liquefaction of coal,
• Coal or heavy oil gasification for medium or low Btu gas,
• In situ coal gasification,
• Oil shale processing,
• Geothermal power generation, and
O Fuel cell processes.
Although generally applicable, the Benfield process may not be economical
for feed gases with C02 partial pressures below 34 to 100 kPa depending on
the mode of operation. Also, H2S—containing gas streams require the presence
of some COZ for treatment. Without C02, C0a from the carbonate would be lost
A4-9
-------
Appendix A4
Benfield Process
from the stream during regeneration, allowing buildup of non-regenerable
potassium snlfide. Around one part of CO, to six parts H2S will essentially
prevent this deactivation (3).
The Benfield process can be operated within a wide range of conditions as
indicated by its use in the purification of natural gas. Operating pressures
range from 1.4 MPa to 7.0 MPa processing feed gas containing up to 44 percent
CO, and 16 percent H2S. In some cases gases containing more H2S than CO are
also being processed (10).
The applicability of the Benfield process to coal gasification has been
shown at the Lurgi plant at Westfield, Scotland (3). and a unit is installed
in a Synthane pilot plant in Pittsburgh, PA (11).
3. Process Performance
As a general guide, it should be possible to obtain the following C02
concentrations when treating a 2.9 MPa gas by the specified Benfield process
configurations (4):
Final CQ2
Single stage >1%
Split flow absorber 0.1 to 1%
Two stage system 500 to 1000 ppm
Benfield HiPure process 10 to 500 ppm
H2S removal/purity is a function of the feed composition, operating
conditions, and the degree of H2S selectivity designed into the system (3).
However, H2S will always be reduced to a greater extent than C02 in an HPC
absorber. Data from several commercial applications of the Benfield systems
are presented in Table A4-2. For coal gasification, the unit installed in
A4-10
-------
Appendix A4
Benfield Process
1960-61 at Westfield, Scotland uses a 35 percent solution of K,C03 which is
regenerated in a conventional steam stripper (13). This unit uses another
stage for final removal of H,S.
TABLE A4-2. PARTIAL LIST OF COMMERCIAL BENFIELD PROCESS GAS
PURIFICATION OPERATIONS OR INSTALLATIONS
Capacity,
Application 10* mj/day
(Ref.)
Liquified
Natural Gas (6) 14.2
Partial
Oxidation (8)
Partial
Oxidation (8)
Coal
Gasification (8)
0.18
1.5
0.68
Natural gas (12) 1.7
Pressure,
MPa CO.
Feed Gas
v
Purified Gas
CO
H,S
(ppmv)
Natural
Natural
Natural
gas
gas
gas
(8)
(8)
(6)
4.
9.
0.
2
3
9
7
7
4
.7
.0
.1
7
16
43
16% 1
50 ppm 1
11% <0.1
500
1
2
5.3
1.0
3.1
4.9
4.7% 50 ppm <1
26
5.2
1%
0.01
2.3 28
6.3 17.6
0.93% 0.95 45
0.6% 0.8 70
3.8% 0.5 <15
Additional data generated at the Westfield plant show that the conven-
tional Benfield process removes trace organic sulfur compounds as follows (3)
Component
Mercaptans
Thiophene
Carbonyl sulfide
Carbon disulfide
Percent Removed
Over 90
85
75
75
A4-11
-------
Appendix A4
Benfield Process
Carbonyl sulfide removal has varied from 75 to over 99 percent in exist-
ing commercial units. Although removal to less than 1 ppm has been demon-
strated in a HiPure unit, assurance of such COS concentrations may not coin-
cide at times with the required C0a and H2S removal (4,5).
Carbon disulfide hydrolyzes to C02 and H2S (equation 5) in two steps,
first forming COS and H2S. CS2 is thus absorbed more slowly than COS. Under
the same condition for 99 percent COS removal, CS2 reduction is 71 percent.
The HiPure process can remove about 85 percent of the feed CS2 (4,5).
Methyl mercaptan is reduced 68 percent to 34 ppm at the same conditions
at which 99 percent of the COS is removed. In a simulated HiPure process for
the same feed concentration, 92 percent of the mercaptan was removed (4,5).
Tests have shown that thiophene does not react chemically with activated
K2COj. Its removal is apparently due to physical solubility. NHS reacts
with the KjCOj and is rapidly absorbed although its solution capacity is low.
Feed concentration and operating parameters of the HPC system determine the
NH3 removal efficiency. Removal efficiencies for SOS and HCN are of the same
order of magnitude as for H2S (4,5).
Table A4-3 shows the application of a split flow absorber for the removal
of C02 and H2S in an ammonia plant. Gas purification data for the conventional
Benfield process and two energy conservation modifications were similar.
Table A4—4 shows data from three representative applications for selective H2S
absorption. Pilot plant studies have shown that 90 to 96 percent of the H2S
in the feed can be recovered while limiting C02 recovery to 15 to 30 percent
(3). Since the difference in absorption rates is normally used to effect this
selective removal, less contact time between feed gas and absorbent is used.
Thus, increased levels of residual H2S are left in the product, and subsequent
trace removal of H2S may be required prior to discharge of the C02 rich acid
gas stream.
A4-12
-------
TABLE A4-3. APPLICATION OF A SPLIT STREAM HPC ABSORBER FOR
COi REMOVAL IN AN AMMONIA PLANT* (7)
Flow Rate (Dry), kmol/hr
Steam, kmol/hr
Composition (Dry), vol %
C°2
H2
Other
Feed
6065
2370
18.4
60.8
19.9
0.9
Purified
Gas
4950
45
0.1
74.4
24.4
1.1
Acid
Gas
1115
188
99.3
0.6
0.1
0.01
Q
Operations at a plant producing 907 Mg per day NH from
natural gas.
TABLE A4-4. SELECTIVE HiS ABSORPTION WITH THE HPC PROCESS (5)
1
Pressure, MPa 2.9
HPC Application
2
2.2
3_
0.9
C02, vol %
Input 11.0 13.90 7.47
Output 8.71 10.39 5.28
Sul fur , ppm
Input
Output
5,800
100
7,600
540
8,000
170
A4-L3
-------
Appendix A4
Benfield Process
Process performance can be altered significantly in the Benfield systems
by varying 1) the additives and their concentrations, 2) concentration of
potassium carbonate, 3) operating configuration, and 4) the cleanup solvent in
the HiPure process. The Westfield data obtained by using a simple Benfield
process are not indicative of the H2S removal possible with more current
systems. In coal gasification, H2S can be removed before or after shift with
either a selective or nonselective system. Additionally, significantly
different synthesis gas compositions due to either the gasification process
or the coal being used may require plant specific design modifications.
4. Secondary Waste Generation
Only the purified gas and acid gas streams are expected discharges from
the Benfield processes. Only those acid gas streams meeting emission require-
ments would be discharged to the atmosphere. Depending on the feed stream,
the location of the acid gas removal module in the overall coal conversion
process scheme, and the Benfield configuration(s) used, there may be one
or more acid gas streams with HaS concentrations ranging from below 1 ppm to
that sufficient to be treated in a Claus unit (e.g., >20 percent H2S) without
subsequent concentration. Most of the HaS streams may not be "rich" enough
for Claus units and, thus, will require either enrichment or use of other bulk
sulfur recovery units such as Stretford process.
Chemical losses may occur due to the loss of absorption solution through
leaks, upsets, and surplus water discharges. These losses can be avoided by
proper maintenance, design, and monitoring of chemicals (corrosion inhibi-
tors), and operations. Trace losses may occur by entrainment with the product
or acid gas discharges. At a sour gas plant in Wyoming (12), the net steam
requirement for regeneration was stated to be less than 60 kg/m3. Surplus
water condensed from the COj vent is being used to generate low pressure steam
without any treatment (14). This would indicate negligible loss of chemicals
through the discharge.
A4-14
-------
Appendix A4
Benfield Process
The literature does not include the need or required frequency, if any,
for change of absorption solutions. However, about 30 percent per year
replacement of the Benfield solution was anticipated in one study (11).
5. Process Reliability
Contact with engineering firm personnel who are familiar with and have
evaluated acid gas removal systems indicated that there should be no
reliability problems associated with the hot potassium carbonate system. Any
scheduled plant shutdown for general maintenance would be more than sufficient
for Benfield systems, as long as they are maintained and well run. BPC sys-
tems have been used extensively in ammonia plants without causing any known
shutdowns.
Other than normal instrumentation problems, potential problem areas
include pump failures (the only moving part in the system) and foaming.
Foaming is most likely to occur in a natural gas plant if liquid hydrocarbons
get into the system. A good liquid/gas separator upstream prevents this
occurrence (Personal communication, D.M. McCrea, March 1982). At the Table
Rock, Wyoming natural gas facility, filters including one made of charcoal
have prevented foaming (Personal communication with S. P. White, March 1982).
The Table Rock facility has had a few pump seal problems. There are two
50 percent duty pumps; a request for a third as a spare could not be jus-
tified on the basis of down time for seal replacement. The onstream factor
has been over 90 percent, and after conversion to a reliable seal flush system
to prevent seals from drying out, very few problems have occurred. Recently,
some efficiency drop has been noted due to wear of the pump impeller. This
may result in the first major work on the system after over four years of
operation (Personal communication with S.P. White, March 1982).
A4-15
-------
Appendix A4
Benfield Process
6. Process Economics
Comparison of equipment purchase costs for a conventional Benfield split
stream absorber process with one of its energy conservation modifications is
given in Table A4-5. These first quarter 1981 costs show no difference in
major equipment costs, but the recovery of heat from the hot regenerated
solution and/or process feed can reduce thermal regeneration energy require-
ments by 32 percent (and eliminated the need for outside steam) for the sub-
ject plant (7). There is negligible difference in the operations of the
absorber and regenerator, thus, process performance is not changed (7).
TABLE A4-5. EQUIPMENT PURCHASE COST FOR CO, REMOVED
IN A 910 Mg/day NH, PLANTa (7)
Equipment (Costs in $1000)
Absorber, Packing, Internals
Regenerator, Packing, Internals
Reboilers and Exchangers
Gas Heated K2C03 Reboiler
Steam Heated K,C03 Reboiler
Gas Heated Condensate Reboiler
Regenerator Cooler— Condenser
Lean Solution Cooler
Pump and Turbine
Flash Vessel
Total, Major Equipment
Split Stream
Absorber
$414
400
294
196
—
426
47
273
—
2050
Benfield
LoHeat
$438
416
235
—
221
342
41
26
91
2051
First quarter 1981 dollars. LoHeat process shown requires 32 percent less
regeneration energy than conventional
process shown.
Additional energy
savings are possible with increased cost
McCrea and Field (5) present procedures for estimating capital and
operating costs for the Benfield processes. Estimates of capital costs are
A4-16
-------
Appendix A4
Benfield Process
shown in Figure A4-5 as a function of C0a removal. These estimates were made
on an installed basis but do not include costs of items such as extra boiler
capacity, extra electrical generating capacity, or extra cooling water
capacity needed to support the Benfield processes. Also, the given ranges are
rough approximations, since there are many variables to be considered in
computing capital costs in addition to process configuration, feed gas
pressure, and composition (5). For comparison, the equipment cost data in
Table A4-5 is represented as an X in Figure A4-5.
Steam requirements for regenerating the HPC solution are related to the
degree to which the carbonate in the absorber is converted to KHC03 (equation
1). Steam requirements, which thus depend on the C02 concentration in the
feed gas, determine the required regenerator heat input. Figure A4-6 shows
the typical regeneration energy requirement as a function of C02 partial
pressure. Applications in which fuel costs are low or where a simple configu-
ration is desired are represented by the upper boundary. The lower heat
requirement can be met by optimizing the system for energy conservation.
Values extracted from Reference 7 on the LoHeat process are shown as X's in
Figure A4-6 for the conventional system and the alternative LoHeat processes.
In addition to heating, energy is required to pump K2C03 solution from
the regenerator to the absorber. This energy requirement is affected by the
total pressure of the absorber and the partial pressure of the C02 in the
feed. Estimates for pumping energy can be made by using Figure A4-7 as a
guide. Net values assume that a turbine operating on high pressure rich
solution supplies a portion of the required energy (5).
These guidelines and estimates provided for C0a absorption are reasonable
for estimating costs for combined C02 and HaS absorption as long as the
C02/H2S ratio is greater than 8:1. For calculations assume that the total
acid gas is C02. H2S will always be reduced to a greater extent than C02.
A4-17
-------
Installed Costs,
$/kg-mole of COa removed per day
i
M
OO
H-
OQ
C.
Ln
to m
ro en
D rt
i-h H-
H. 3
ro CB
I—1 rr
P- ID
a
re
T3 H
n 3
w
O rt
r1- rt>
"~d o-
c
en
o-
^N O
(U
>-f
cn
-p-
o
a-
o
CD
o
o
o
n
o
i-l
ro
tn
05
C
K3
Ul
O
O
p
a.
H-
n
Ca
rt
(t
CO
ex
CD
ro
HI
ro
i-(
ro
3
n
ro
-------
c
o
n3
S-i
00 .S
QJ K
* „
^8
d u
Q)
-o
o 2
rn 6
S-i
o
QJ
60
^
60
0)
e
160
120
80
40
O Indicates data from reference 7
250
500
750
1000
C02 Partial Pressure, kPa
Figure A4-6.
Typical regeneration energy required by conventional
Benfield or HiPure processes (5)
A4-19
-------
•n
0)
n
'
00
Total Pressure, MPa
0.2 -
250
500
750
1000
C02 Partial Pressure, kPa
Figure A4-7. Typical pumping energy required by Benfield
HPC or HiPure processes (5)
A4-20
-------
Appendix A4
Benfield Process
For selective absorption of HaS, case by case evaluation is necessary to
estimate performance, cost, and utility consumption (5). However, representa-
tive applications have been studied and are shown in Tables A4-4 and A4-6.
Observations have shown that capital costs of units for purifying coal gases
are relatively independent of operating pressure and degree of H2S removal.
Energy consumption, however, increases with increased product gas purity.
TABLE A4-6. SELECTIVE H,S ABSORPTION WITH THE HPC PROCESS (5)
HPC Application
1 2
Pressure, MPa
2.9
2.2
0.9
Percent removal
Residual ppm
Capital Cost,
»/m3/day of feed gas
Energy Consumed,
MJ/m3 of feed gas
98.3
100
0.22
0.50
92.9
540
0.21
0.29
97.8
170
0.20
0.41
See Table A4-4 for additional performance information.
1974 dollars.
7. References
1. Benson, H.E., and J. H. Field. Method of Separating CO, and HZS from Gas
Mixtures. U.S. Patent 2,866,405, May 1959.
2. The Benfield Process for Energy Savings in Acid Gas Removal. Union
Carbide EPfflP Brochure No. L-5100, 1981.
3. Parrish, R.W. and J.H. Field. The Benfield Process in Coal
Gasification. Presented ,at the 24th Annual Gas Conditioning Conference,
The Univ. of Oklahoma, Norman, Oklahoma, March 1974.
A4-21
-------
Appendix A4
Benfield Process
4. Parrish, R.W. and H.B. Neilson. Synthesis Gas Purification Including
Removal of Trace Contaminants by the Benfield Process. Presented at the
167th National Meeting of the ACS, Div. of Industrial and Engineering
Chem., Los Angeles, California, April 1974.
5. McCrea, D.H. and J.H. Field. The Purification of Coal Derived Gases:
Applicability and Economics of Benfield Processes. Presented at the
AIChE, 78th National Meeting, Salt Lake City, Utah, August 1974.
6. Benson, H.E. and R.W. Parrish. Hipure Process Removes COa/H.jS. Hydro-
carbon Processing, April 1974.
7. Baker, R.L. and D.H. McCrea. The Benfield LoHeat Process: An Improved
HPC Absorption Process. Presented at the AIChE, 1981 Spring National
Meeting, Houston, Texas, April 1981.
8. U.S. Environmental Protection Agency. The Benfield Activated Hot
Potassium Carbonate Process: Commercial Experience Applicable to Fuel
Conversion Technology. In Symposium Proceedings: Environmental Aspects
of Fuel Conversion Technology II (Dec. 1975, Hollywood, Florida), EPA-
600/2-76-149, June 1976.
9. Brown, F.C. Criteria for Selecting C02 Removal Processes. Presented at
the UNIDO Technical Conf. on Ammonia Fertilizer Technology for Promotion
of Economic Cooperation Among Developing Countries, Beijing, China, March
1982. (Author from Humphreys & Glasgow Ltd., London)
10. Clayman, M.A. and J.R. Clark. Low Energy Natural Gas Purification Using
Benfield Processes. Presented at the 30th Annual Gas Conditioning
Conference, The Univ. of Oklahoma, Norman, Oklahoma, March 1980.
11. Christensen, K.G. and W.J. Stupin. Merits of Acid Gas Removal Processes.
Hydrocarbon Processing, Feb. 1978.
12. White, S.P. and D.J. Morgan. Sour-Gas Plant Boosts CIG's Supply. Oil
and Gas Journal, June 1979.
13. Trials of American Coals in a Lurgi Gasifier at Westfield, Scotland.
Prepared by Woodal1-Duckham Ltd., Crawley, Sussex, England for ERDA and
the American Gas Association, Rep. No FE-105.
14. Parrish, R.W. and R.O. Seidel. Environmental Considerations in Acid Gas
Removal Systems. AIChE, Second Pacific Chemical Engineering Congress,
Denver, Colorado, August 1977.
A4-22
-------
Appendix A4
Benfield Process
8. Personal Communications
D.H. McCrea, Benfield Process Union Carbide Corporation, Tarrytown, New
York to TRW Environmental Division. March 1982.
S.P. White, Colorado Interstate Gas Co., Colorado Springs, Colorado to
TRW Environmental Division. March 1982.
A4-23
-------
-------
APPENDIX A5
CATACARB ACID GAS REMOVAL PROCESS
The Catacarb process is a modification of the hot potassium carbonate
process. Invented and developed by A. G. Eickmeyer, the process uses an
aqueous solution containing primarily alkali metal salts, potassium carbonate,
potassium borate, small amounts of catalytic activator, and a corrosion
inhibitor. The process is currently licensed by Eickmeyer and Associates of
Prairie Village, Kansas (1,2).
1. Process Description
The flow scheme for a Catacarb plant is identical to that used for the
hot potassium carbonate process. A simplified process diagram is shown in
Figure A5-1. Solution is fed to the top of a packed or trayed absorber column
and the process gas is fed near the bottom of the absorber. Purified gas
leaves the top of the absorber. The hot rich Catacarb solution leaving the
bottom of the absorber is depressurized, usually through a hydraulic turbine,
and flashed into the top of the regenerator. Here the descending solution
meets with rising steam which strips out C02 and H2S from the solution. The
regenerated lean solution is recycled back to the absorber. The steam is usu-
ally condensed in an overhead condenser. The acid gas is either vented (if
only C02 absorption is involved) or sent to a sulfur recovery plant (if coab-
sorption of C02 and H^S are involved).
A split flow design is used when a higher degree of purification is
needed. In the split flow design (shown in Figure A5-2) a portion of the lean
solution is cooled and fed to the top of the absorber. The cooler temperature
reduces the back pressure of CO., and HaS and thus achieves a higher degree of
purification.
In the case where a very high degree of purification of the gas is
required or the use of steam for regeneration heat is to be minimized, a two-
stage design may be used. A typical flow diagram for a two-stage Catacarb
A5-1
-------
PURIFIED
GAS
ABSORBER
RAW GAS
CO2 + H2S
STEAM
Figure A5-1. Catacarb process with single stage absorber (3)
A5-2
-------
PURIFIED
GAS
ABSORBER
FEED
GAS"
CO2 + H2S
STEAM
Figure A5-2. Catacarb process with split flow absorber (4)
A5-3
-------
Appendix A5
Catacarb Process
absorption system is presented in Figure A5-3. In this design, a major por-
tion of the solution is withdrawn as semi-lean solution from the middle of the
regenerator. Only a minor portion of the solution is fully regenerated to
become lean solution. The semi-lean solvent enters at the midpoint of the
absorber, below which the bulk of the C02 and H2S is removed. The lean solu-
tion, usually after some cooling, enters the top of the absorber where the
final purification is achieved (1,5, Personal communication with E. A.
Wiberg, April 22, 1982).
The Catacarb catalyst greatly increases the rate of absorption and
desorption of acid gases without changing the equilibrium concentration of C02
and H2S in the potassium carbonate solution. In the Catacarb process, the
absorption of acid gas takes place near the atmospheric boiling point of the
solution or at an elevated temperature of at least 330 K. A hot aqueous solu-
tion of 15 to 50 percent by weight K2C03 containing minor amounts of an amine
group (alkanolamines and ethylene polyamines) is used to absorb and desorb
acid gas (2, Personal communication with E. A. Wiberg, April 22, 1982).
The potassium carbonate solution is not highly ionized and thus contains
few hydroxyl ions which can react directly with C02. Therefore, it is
believed that C02 must first combine with water to form carbonic acid, which
in turn reacts with a carbonate ion to form two bicarbonate ions:
C02 + H20 = H2C03 (1)
H2C03 + K2C03 = 2KHC03 (2)
H2S and RSH components react reversibly with the K2C03 as follows:
K2C03 + H2S = KHC03 + KHS (3)
K2C03 + RSH = KHC03 + RSK (4)
A5-4
-------
PURIFIED
GAS
ABSORBER
FEED_
GAS*
SEMI-LEAN
SOLUTION
H2S + CO2
STEAM
LEAN
SOLUTION
Figure A5-3. Catacarb process with two stage absorber and
regenerator (4)
A5-5
-------
Appendix A5
Catacarb Process
The Catacarb solution also catalyses the hydrolysis of COS and CS2 :
COS + H20 = CO, + H4S (5)
CSa + 2HaO = COj + 2H2S (6)
Increased absorber temperature (near boiling point of the solution) and con-
tact time are necessary for optimum COS and C02 absorption (Personal communi-
cation with E. A. Wiberg, April 22, 1982). S02 and HCN are believed to react
as follows:
KaC03 + S02 = K2S03 + C02 (7)
K2C03 + HCN = ECN + KHO>3 (8)
The removal efficiencies for S02 and HCN are in the same order of magnitude as
for H S. Past Catacarb designs show very little removal of aromatics from the
feed gas (6).
The effect of various amounts of Catacarb catalyst on the activity of two
commercial plant solutions is shown in Figure A5-4 (4). Catalyst concentra-
tions in the range of 8 to 10 percent will make the solution several times
more active than pure potassium carbonate. Figure A5—5 shows the effect of
contamination on the activity of carbonate solution (4).
It is estimated that over 100 Catacarb plants belonging to 51 com-
panies are in operation in 27 countries. The plants remove over 40 million
mj/day of acid gas from about 280 million m»/day of gas. The Catacarb process
has been selected by Exxon to purify and remove 2000 Mg per day of C02 from a
1,600 Mg/D ammonia plant at Redwater, Alberta, Canada. This plant is one of
the first projects to use the new Catacarb technology to produce low energy
ammonia (8).
A5-6
-------
400 r-
o
o
0)
S-l
3
CL
M-i
O
4-1
•H
•H
4J
U
c
o
•H
O
CO
300
« 200
100
100 r-
o
o
0)
M
3
4-1
•H
•H
4J
O
c
o
•H
10
15
Catacarb Content, 7, by Volume
Ratio of Contaminated/Active Potassium
Figure A5-4.
Activity of catalyzed carbonate
solutions (4)
Figure A5-5. Effect of contamination on activity
of carbonate solutions (4)
-------
Appendix A5
Catacarb Process
2. Process Applicability
Catacarb technology is commercially proven in the purification of low,
medium, and high pressure gas streams by removing CO^, HjS, and COS. Typical
applications include (3,7,8):
• Ammonia and urea-gas purification,
• Hydrogen purification,
• Natural gas sweetening and Btu improvement,
• Substitute natural gas purification,
• Ethylene oxide recycle gas purification,
• Iron ore reduction gases purification,
• Power plant stack gas for enhanced oil recovery,
• Hethanol synthesis gas purification,
• Coal gasification streams for Btu improvements,
• Heavy oil upgrading (e.g., Dynacracking), and
• Molecular sieve regeneration gas purification.
Although generally applicable in all the above areas, the most common applica-
tion for the Catacarb process is in hydrogen, ammonia, and urea-gas purifica-
tion. Catacarb has not been applied to any synthetic fuel conversion plants
(9).
Hot carbonate processes such as Catacarb are economically attractive when
the partial pressure of acid gas is below 1.4 MPa. With acid gas partial
pressures above 1.4 MPa, physical solvent processes become more attractive.
Generally, the hot potassium carbonate (HPC) processes are not very selective
with respect to H S-CO removal. However, the Catacarb process can be effec-
tively applied in conjunction with a selective process to reduce overall unit
cost (Personal communication with E. A. Wiberg, April 22, 1982).
A5-J
-------
Appendix A5
Catacarb Process
The Catacarb processes can be operated over a wide range of conditions.
Operating pressures range from 7 to 70 MPa. Gas temperatures vary between
290 K and 510 K depending on the mode of operation. Typical feed gas compo-
sitions are 15 to 30 percent C02 for ammonia and hydrogen purification and in
SNG plants and 5 to 50 percent acid gas for natural gas sweetening and other
applications (3).
In a synfuel plant, the Catacarb process can be applied for nonselective
removal of H2S and C02. However, a second process (such as the MDEA) may be
required to concentrate the H2S for sulfur recovery by a Claus plant. Cata-
carb may also be applied for selective removal of H2S and C02. With proper
design, an H2S^-rich stream suitable for processing in a Claus plant may be
produced. However, the C0a vent gas stream usually contains relatively high
concentrations of reduced sulfur compounds and, therefore, may require further
treatment before venting. The Catacarb process is more likely to be applied
in the bulk removal of C02 after the shift reaction, in designs where H2S is
removed prior to the shift reaction ("sweet" shift).
3. Process Performance
The vent gas from the Catacarb absorber typically contains less than 4
ppmv of H2S and usually between 0.05 and 2 percent C02. The H2S removal is a
function of feed gas composition, process cycle, and the degree of selectivity
designed into the system. Since in most applications C02 and H2S are removed
together, the acid gas from the regenerator contains these two components in
nearly the same proportion as in the feed gas (Personal communication with E.
A. Wiberg, April 22, 1982).
A5-9
-------
Appendix A5
Catacarb Process
In a study performed to evaluate and compare unit processes which may be
used in coal gasification plants, the Catacaxb process was considered for two
applications: 1) to treat high pressure, low acid gas content gas and 2) to
treat low pressure gas but with an intermediate acid gas content. A block
flow diagram of the first of these Catacarb applications is shown in Figure
A5—6. The high pressure and low acid gas stream contains about 5 to 18 per-
cent by volume of acid gas. C0a content of the feed gas is 8 percent by
volume. The absorber operating pressure is 5.9 MPa. The Catacarb unit pro-
duces an acid gas stream with about 0.7 mole percent H2S. A second process
(such as an MDEA unit) will be required to concentrate HaS for sulfur recovery
by the Claus process (6). Material balances of gas streams around the
Catacarb process for this application are presented in Table A5—1.
A block flow diagram of the second noted Catacarb application is shown in
Figure A5-7. In the low pressure and intermediate acid gas content case,
selective removal of H4S to produce an HaS-rich acid gas stream was required.
The operating pressure of the acid gas is 1 MPa. The C02 content of the feed
gas is 15 percent by volume. The Catacarb design produced an H^S stream rich
enough for processing in a Claus unit. However, the C02 vent stream contains
appreciable amounts of H2S and may require further treatment before venting.
Material balances of gas streams around the Catacarb process for this second
application are presented in Table A5—2.
As a general guide, a 2.9 MPa gas should be purified to the following C0a
content when treated by the specified Catacarb process configuration (Personal
communication with E. A. Wiberg, April 29, 1982)!
A5-10
-------
FEED GAS
FROM RAW
GAS QUENCH
NONSELECTIVE
CATACARB
ACID GAS
REMOVAL UNIT
H2S + CO2 TO
SULFUR RECOVERY
TREATED GAS
Figure A5-6. Nonselective Catacarb design (6)
FEED GAS FROM
SHIFT CONVERSION
SELECTIVE
CATACARB
ACID GAS
REMOVAL UNIT
H2S-RICH OFFGASTQ
SULFUR RECOVERY
CO2-RICH OFFGASTO
SULFUR RECOVERY
TREATED GAS
Figure A5-7. Low pressure, selective Catacarb design (6)
-------
TABLE A5-1. MATERIAL BALANCE: LOW ACID GAS CONTENT,
NONSELECTIVE CATACARB DESIGN (6)
Component
H , kg-mole/hr
N , kg-mole/hr
CO, kg-mole/hr
C02, kg-mole/hr
H S, kg-mole/hr
CH4, kg-mole/hr
H^O, kg-mole/hr
Total, kg-mole/hr
Total, kg/hr
Temperature, K
Pressure, MPa
1
Feed Gas
11,237
60
3,445
1,428
11
2,443
22
18,646
223,500
310
5.9
Stream Number and
2
Acid Gas
25
0
8
1,411
11
5
165
1,625
65,800
330
0.17
Name
3
Treated
11,212
60
3,437
17
Gas
1 ppmv
2,438
37
17,201
161,000
330
5.8
a.
Stream numbers correspond to those shown in Figure A5-6,
TABLE A5-2. MATERIAL BALANCE: LOW PRESSURE,
SELECTIVE CATACARB DESIGN (6)
Stream Number and Name
Component
HZ, kg-mole/hr
N , kg-mole/hr
CO, kg-mole/hr
CO , kg-mole/hr
H S, kg-mole/hr
COS, kg-mole/hr
CS2, kg-mole/hr
CH , kg-mole/hr
C6H6+, kg-mole/hr
Phenol, kg-mole/hr
HZ0, kg-mole/hr
Total, kg-mole/hr
Total, kg/hr
Temperature, K
Pressure, MPa
1
Feed Gas
29,600
240
9,170
7,540
180
5
1.4
2,630
120
0.5
305
49,792
718,500
310
1.1
2
Acid Gas
2
—
1
425
162
-
-
-
-
0.5
70
661
25,600
-
0.17
3
CO Gas
2
25
-
7
6,694
24
-
-
1
-
-
592
7.343
306,300
317
0.1
4
Treated
Gas
29.573
240
9,162
421
0.2
0.5
1.4
2.629
120
-
4.668
46.815
477,200
377
1.0
Stream numbers correspond to those shown in Figure A5-7.
A5-12
-------
Appendix AS
Catacarb Process
Final CO^
Single stage 0.3 to 0.4 vol%
Split flow absorber 0.1 to 0.2 vol%
Two-stage system 500 to 1000 ppm
A low heat version of the Catacarb process has been recently selected by
California Synfuels Research Corporation for acid gas removal in their first
commercial-scale plant using the Dynacracking route to process heavy oils.
The following acid gas is treated in the Catacarb unit:
coa
H,S
COS
The rate of hydrolysis of COS is highly dependent on temperature of the
solution. Because the hydrolysis rate is primarily dependent on temperature
and because the overall rate of hydrolysis is slow relative to C02 removal,
maximum COS removal can only be achieved by higher operating temperatures
(near boiling point of solution) and longer residence times. If the system is
primarily designed to remove H2S and C02, removal of COS by hydrolysis is esti-
mated to be around 90 percent (Personal communication with E. A. Wiberg, April
29, 1982).
Removal of carbon disulfide (CS2) requires an additional stage of hydro-
lysis (equation 6). Because of this, CS3 removal is always lower than COS
removal. If the system is mainly designed for H2S and C02 removal, removal of
CS2 by hydrolysis is estimated to be around 50 to 60 percent (Personal commu-
nication with E. A. Wiberg, April 29, 1982).
Sour Gas
24.7 vol%
1.6 vol%
0.1 vol%
Sweet Gas
0.2 vol%
<20 ppm
<140 ppm
Removed
230 Mg/day
12 Mg/day
A5-13
-------
Appendix A5
Catacarb Process
4. Secondary Waste Generation
The only secondary waste stream generated by the nonselective Catacarb
process is the acid gas stream from the regeneration of the Catacarb solution.
The only secondary waste stream generated by the selective Catacarb process is
the CO -rich vent gas. Depending on the concentration of reduced sulfur
compounds in these streams, they may be vented directly to the atmosphere or
may require further treatment (e.g., in a sulfur recovery plant).
5. Process Reliability
The Catacarb solution has proven to be nonfouling and noncorrosive,
resulting in low maintenance and an onstream factor of over 97 percent (8).
Laboratory tests indicated the Catacarb catalyst to be noncorrosive to low
carbon steel. The calculated penetration rate was only 2.5 to 5.0 \aa per year
(7). Foaming in the Catacarb unit is most likely to occur if liquid hydrocar-
bons or phenols get into the system. However, foaming can be controlled by
the addition of foam inhibitors such as silicone compounds.
6. Process Economics
The economics and energy requirements of Catacarb depend on C02 and H^S
content of feed gas, degree of removal required, system pressure, tower
intervals, and size of the towers. There are also interrelated capital
investment and operating cost factors which depend on the specific design.
The capital investment costs for the Catacarb C02 removal process which
uses steam regeneration are presented in Table A5-3. The capital investment
costs appear to be rather flat for the pressure range selected (10). Annual
utility costs are plotted in Figure A5-8, for the feed gas defined in Table
A5-3. These costs are shown to decrease rapidly with pressure.
A5-14
-------
TABLE A5-3. CAPITAL INVESTMENT COST FOR CATACARB O>2 REMOVAL (10)
Basis: 13.8 million NmVday of inlet gas at 1.5 MPa, 2.8 MPa, and 4.2 MPa
Composition
H,S
CO,
. CO,
1.5 MPa
$21,320,000
Inlet Gas
22%(vol)
Balance
Capital Investment Costs
2.8 MPa
$21,260,000
Purified Gas
1 ppm
0.5%(vol)
4.2 MPa
$22,410,000
1981 basis
o
cn
o
0)
0)
-H
6.4p-
6.2
6.0
5.8
5.6
5.4
5.2
5.0
Steam = $4.40/Mg
Power = 2.5c/kWh
Cooling Water = 1.3c/m3
Pressure, MPa
Figure A5-8. Utilities cost for CO?, removal with Catacarb (10)
A5-15
-------
Appendix A5
Catacarb Process
Eickmeyer and Associates have optimized COj/H^S removal designs taking
into consideration the most economical use of heat, power, and capital invest-
ment (Personal communication with E. A. Wibert, April 22, 1982). The process
design combines conventional know-how with low energy technology to produce
significant savings. With new Catacarb designs, savings of up to 25 to 60
percent of the heat requirements of the conventional designs have been
achieved. The economics of the Catacarb process as applied to a hydrogen
plant and also to an ammonia plant for C02 redaction are shown in Tables A5-4
and A5-5 (Personal communication with E. A. Wiberg, April 22, 1982).
A5-16
-------
TABLE A5-4. ECONOMICS OF THE CATACARB PROCESS
APPLIED TO A HYDROGEN PLANT
Basis: 2.5 million m3/day gas capacity
C02 in feed gas: 20.5 vol%
CO^, in treated gas: 0.15 vol%
Pressure: 1.9 MPa
Conven-
tional
NEW CATACARB
1981 Capital Investment. $10*
Waste heat from gas, GJ/hr
Steam, Mg/hr
Power, kWh
Cooling water, 103, m3/hr
0.44 MPa steam generated for export,
(Boiler cost not incl.)
Heat efficiency, MJ/kg-mole
of CO removed
2
5.
106
15
750
2.
0
110
8
Case 1
5.8
106
0
750
1.3
0
Case 2
6.2
78
0
1640
1.2
17 Mg/hr
86
63
Source: Personal communication with E.A. Wiberg, April 22, 1982
TABLE A5-5. ECONOMICS OF THE CATACARB PROCESS APPLIED TO AN AMMONIA PLANT
Basis:
1,000 Mg/day ammonia plant
CO. in feed gas: 20.5 vol%
0.15 vol%
CO in treated gas:
Pressure: 3.1 MPa
1981 Capital Investment, ilO*
Waste heat from gas, GJ/hr
Steam, Mg/hr
Power, kWh
Cooling water, 103 , m3/hr
Heat Efficiency, MJ/kg-mole of
CO removed
2
Source: Personal communication with
Conven-
tional
6.7
57
20
1024
1.1
75
E.A. Wiberg, At
NEW
CATACARB
7.2
57
1.5
2060
0.98
45
>ril 22, 1982
A5-17
-------
Appendix A5
Catacarb Process
7. References
1. Kohl, Arthur and Fred Riesenfeld. Gas Purification. Gulf Publishing
Company, Houston, Texas, 1974.
2. Eickmeyer. U.S. Patent 4,271,132, June 2, 1981.
3. Catacarb C0a and HjS Removal. Hydrocarbon Processing, April 1975.
4. Maddox, R. N. Gas and Liquid Sweetening. 1974.
5. The Catalytic Process for Acid Gas Removal. Hydrocarbon Processing,
September 1974.
6. Christensen, K. G. and W. J. Stupin. Comparison of Acid Gas Removal
Processes. FE-2240-49. U.S. Department of Energy Final Report, April
1978.
7. Eickmeyer, A. G. Catalytic Removal of CO.,. Chemical Engineering
Progress 58:4, April 1962.
8. Eickmeyer & Associates. Catacarb Catalog. Overland Park, Kansas.
9. Christensen, K. G. and W. J. Stnpin. Merits of Acid-Gas Removal
Processes. Hydrocarbon Processing, February 1978.
10. Eichmeyer, A. G. and H. A. Gangriwala. The Role of Acid Gas Removal in
Synfuels Production. Presented at the Symposium on Gas Purification at
the AIChE 1981 Spring National Meeting and llth Petrochemical Exposition
at Houston, Texas.
8. Personal Communications
Wiberg, E.A., Eickmeyer & Associates, Overland Park, Kansas to TRW
Environmental Division. April 22, 1982.
Wiberg, E.A., Eickmeyer & Associates, Overland Park, Kansas to TRW
Environmental Division. April 29, 1982.
A5-LS
-------
APPENDIX A6
CLAUS PROCESS
1. Process Description
The Clans sulfur removal process is a vapor-phase, dry, high temperature
process in which hydrogen sulfide-rich gases from acid gas removal processes
are catalytically reacted with sulfur dioxide (produced by air oxidation of
hydrogen sulfide) to form elemental sulfur. Sulfur recovery units of the
Claus type have been used extensively for more than 30 years in oil
refineries, natural gas treating plants, and coking operations when large
quantities of hydrogen sulfide are produced (1). Modern Claus units are
capable of greater than 95 percent sulfur recovery and produce a high-quality,
salable sulfur, as well as steam. The Claus unit offgas generally contains
several thousand ppmv sulfur and may require further processing to reduce
emissions of sulfur species (e.g., SCOT, Beavon, or Wellman-Lord processes).
A typical Claus unit (Figure A6-1) consists of a thermal reaction furnace
and one or more catalytic stages, depending on the desired level of sulfur in
the treated tail gas. The overall Clans reaction is given by:
3H2S + 1.502 = 3H20 + 3S (1)
The overall Claus reaction actually takes place in two steps. Sulfur dioxide
is formed by oxidation of a portion of the hydrogen sulfide in the thermal
stage of a Claus unit:
H2S + 1.502 = S02 + H20 (2)
A6-1
-------
T
LEGEND:
1 ACID GAS FEED
2 COMBUSTION AIR
3 TAIL GAS
4 CONDENSATE
5 BOILER FEED WATER
6 LOW PRESSURE STEAM
7 HIGH PRESSURE STEAM
8 LIQUID SULFUR
9 SPENT CATALYST
(LOCATION NOT KNOWN)
i
REACTION
FURNACE 6
« i A
J
, V
I CONDENSER NO 1
T
I
i
f CONVERTER.
REHEAT. 1
I
T
>
)
^
C CONVERTER: n ")
REHEAT: n
i
J J CONDENSER'NO
f
n
i
3
M
| | CONDENSER NO n + 1 J
t _ .
AIR BLOWER Q
I
SULFUR PIT
0
I
Figure A6-1. Straight-through Glaus process (2)
-------
Appendix A6
Glaus Process
The prevailing combustion conditions (1370 to 1800 K and pressure slightly
above atmospheric) also result in the formation of sulfur vapor from hydrogen
sulfide and sulfur dioxide with a 50 to 70 percent yield (3,4,5,6):
2H2S + S0a = 3S + 2H,0 (3)
The process gas is cooled in a steam-producing waste heat boiler and the
sulfur is condensed and removed. Additional formation of sulfur from the
remaining hydrogen sulfide and sulfur dioxide is promoted by either bauxite or
alumina catalyst in one or more catalytic stages. Each catalytic reactor is
preceded by a process gas reheat unit and followed by a sulfur condenser.
The Claus process is controlled by regulating the flow of combustion air
to the reaction furnace so that only about one-third of the hydrogen sulfide
is oxidized to sulfur dioxide. This provides the optimum quantity to react
with the hydrogen sulfide to produce elemental sulfur. However, high carbon
dioxide contents (such as those encountered in coal gasification processes)
often dilute the acid gas to the point where stable combustion of one-third of
the hydrogen sulfide is very difficult. Therefore, Claus plant designs are
generally based on one of three combustion configurations, depending on the
carbon dioxide concentration in the feed gas: "straight-through," "split-
stream," and the "sulfur burning" mode (2,7,8).
For carbon dioxide levels below about 30 percent by volume (corresponding
to hydrogen sulfide levels above 40 percent by volume), the "straight-through"
process is generally used (i.e., the entire feed gas stream is sent to the
Claus furnace (Figure A6-1)). At carbon dioxide concentrations above about 30
percent by volume (corresponding to hydrogen sulfide concentrations below
about 40 percent by volume), free flame combustion with stoichiometrie air
Ab-3
-------
Appendix A6
Glaus Process
becomes unstable and the "split-stream" process must be used. In this design
(Figure A6-2), approximately one-third of the acid gas is completely oxidized
to form sulfur dioxide and the remaining two-thirds is bypassed around the
combustion chamber. When the acid gas contains less than 5 to 10 percent by
volume hydrogen sulfide (corresponding to very high carbon dioxide levels)
combustion temperatures are too low to sustain the thermal sulfur reactions
and side reactions become troublesome, especially those involving hydrocar-
bons. Hydrocarbons in the feed gas influence "split-stream" applicability,
since they may limit the bypassing of gas directly to the catalytic conver-
ters. At very low hydrogen sulfide levels, the "sulfur burning" mode may be
used, in which liquid sulfur is recycled to the combustion chamber to supply
heat and sulfur dioxide for reaction with hydrogen sulfide downstream in the
process.
There are several other ways to get around the problem of unstable com-
bustion. The combustion temperature can be "bolstered" by preheating the acid
gas, the air, or both. Multizone combustion chambers could be used. Other
combustible materials could be added to the combustion chamber, but this in-
creases the plant size. Pure oxygen can be used instead of air, although
oxygen enrichment is better than using pure oxygen, because nitrogen will help
moderate the combustion temperature. The acid gas streams produced in most
coal conversion processes as a result of acid gas treatment will contain car-
bon dioxide in excess of 30 percent by volume; therefore, the "split-stream"
Glaus process will probably be the most applicable.
Although the Glaus process is generally described by the simple reactions
given by equations 2 and 3, the actual combustion reactions of hydrogen
snlfide and air are very complex as shown in Table A6-1. The presence of
hydrocarbons, carbon dioxide, ammonia, and other impurities in the feed also
multiply the number of possible reactions that can occur as shown in Table
A6-2. Secondary reactions taking place are mainly the formation reactions of
A6-4
-------
1 ACID GAS FEED
2 COMBUSTION AIR
3 TAIL GAS
4 TWO-THIRDS OF ACID GAS
STREAM VOLUME
5 ONE-THIRD OF ACID GAS
STREAM VOLUME
6 WATER
7 HIGH PRESSURE STEAM
8 LOW PRESSURE STEAM
9 BOILER FEED WATER
10 LIQUID SULFUR
11 SPENT CATALYST
(LOCATION NOT KNOWN)
Figure A6-2. Split-stream Glaus process (2)
-------
TABLE A6-1. COMBUSTION REACTIONS (1,2,5,6,8,9,10)
H2S + 02 = H0a + HS
H2S = HS + H
H + H2S = HS + H,
HS + 02 = SO + OH
OH + H2S = H20 + HS
OH + H02 = H20 + 02
SO + 02 = S02 + 0
0 + H2S = OH + SH
SO + 0 = S02 = S02 + hv
2SO = S202
2HS = H +
HS + HO = H20 + S
S + 0 = SO
2S = S2
S02 + 0 + M = SOs + M
SO + 02 + M = S03 + M
S0
= S02 + H20
TABLE A6-2. POSSIBLE SIDE REACTIONS IN THE FURNACE AND CATALYTIC REACTORS
(1,2,5,6,8,9,10,11)
CO
H20 =
C0
2C02 = 2CO
Hydrocarbons = C02, CO, H20, H2
C02 + S02 = COS
CH4
S02 = COS
C02 + S02 = COS
CO + S = COS
H20
H20
1.50
CH
2S
CS
2H2S
2CO + S2 = CO^ + CS2
COS + H20 = C02 + H2S
CS + H 0 = COS + H S
CS2 + 2H20 = C02
2H2S
2NH + 1.50, = N + 3H 0
3 222
3SO, + Al.O, = Al (SO)
i 23 2 4 A
A6-6
-------
Appendix A6
Glaus Process
carbonyl sulfide and carbon disulfide in the combustion zone and the partial
hydrolysis of these compounds to hydrogen sulfide in the catalytic reactors
(6). Carbonyl sulfide and carbon disulfide are slow to react over the Claus
catalyst and, therefore, represent sulfur that is not totally recovered as ele-
mental sulfur. The potential presence of ammonia in the feed gas gives rise
to secondary reactions which may disturb the operation of the units.
2. Process Applicability
More than 15 million tons of sulfur are recovered annually from treatment
of acid gases by the Claus process (6). There are at least 170 Claus plants
in the United States used in a wide variety of industries including petroleum,
natural gas, and coke production (4). Applicability of the Clans process to
coal conversion process gas purification systems has not been entirely esta-
blished. A Claus plant is featured in the design of the Hygas pilot plant at
Chicago, Illinois for processing the acid gas stream from a diglycol amine
unit (12). A number of firms are experienced designers of Claus plants in the
United States including the following: Black, Sivalls & Bryson, Inc.; Ford,
Bacon & Davis, Inc.; and The Ralph M. Parsons Company.
Claus plants can be designed to operate at various temperatures and pres-
sures and with a wide variation of feed stream compositions. Operating tem-
peratures will vary as a function of feed stream conditions and plant design.
The combustion temperatures are generally 1370 to 1800 K. The catalytic con-
verter temperatures will be progressively lower. Pressures are usually low
(e.g., below 0.17 MPa) with only enough extra pressure to overcome the pres-
sure drop through the plant.
A6-7
-------
Appendix A6
Clans Process
Hydrogen sulfide concentration is the most important parameter in Glaus
plant design and operation. The presence of various contaminants in the acid
gas feed (e.g., carbon dioxide, ammonia, water, hydrocarbons, heavy organics,
and hydrogen cyanide) can decrease the hydrogen sulfide concentration, lower
the sulfur recovery ability of the Glaus process, and increase the size of the
plant due to larger volumetric flows (2-11, 13, 14). There are limited data
available on the maximum allowable concentration of these various contami-
nants. However, decreased sulfur removal efficiencies appear to be due to de-
creased hydrogen sulfide concentrations rather than the presence of impurities
in the acid gas feed.
Carbon dioxide is a diluent that often results in combustion control
problems. Several combustion scheme alternatives related to the carbon diox-
ide content (i.e., "straight-through" process, "split—stream" process, and
"sulfur burning" mode) are discussed in Section 1. Due to the high carbon
dioxide contents encountered in coal gasification processes (greater than 30
percent volume), the "split-stream" Glaus process will probably be the most
applicable. High carbon dioxide concentration levels in the feed gas would
not create major operating problems unless the gas also contains greater than
500 ppmv ammonia (since these conditions can lead to ammonium sulfide deposi-
tion in the catalyst beds) (2). High carbon dioxide levels, however, increase
the total sulfur content of the Glaus tail gas by the formation of carbonyl
sulfide and carbon disulfide.
A large majority of the ammonia in the acid gas feed burns to nitrogen
and water:
2NH3 + 1.50Z = N2 + 3H20 (4)
A6-8
-------
Appendix A6
Claus Process
The nitrogen produced tends to depress the Claus reaction (equation 2) by dil-
ution as an inert. Water formed is an inert and also tends to depress the
Claus reaction, since it is a reaction product of both equations 2 and 3. Ad-
ditional combustion air requirements also cause inert-dilution problems due to
the nitrogen associated with the air. Burning appreciable amounts of ammonia
generates high temperatures in the furnace, increasing waste heat boiler duty
and total condenser duty. Under certain conditions significant amounts of
nitrogen oxides (NO ) are formed (6), which can cause serious corrosion pro—
x
blems. Nitrogen dioxide (the predominant NO species) catalyzes oxidation of
sulfur dioxide to sulfur trioxide even in small amounts. Sulfur trioxide can
cause corrosion due to sulfuric acid (HaS04) formation in cool spots and
bauxite catalyst (A120}) deactivation due to sulfation by the reaction:
3S03 + A1203 = Alt(S04)f (5)
Formation of NO can also present a serious pollution problem from NO emis-
sions to the atmosphere via the tail gas. The reducing atmosphere in the
combustion furnace does not provide enough oxygen to completely burn the am-
monia present in the acid gas. Ammonia reacts with hydrogen sulfide and sul-
fur dioxide to form salts (e.g., ammonium hydrosulfide, ammonium polysulfides,
ammonium carbonates or carbamates, and/or ammonium sulfate) that precipitate
throughout the Claus unit, causing plugging problems which frequently result
in premature plant shutdowns (11). Catalyst plugging problems can occur when
the ammonia concentration exceeds 500 ppmv in combination with a carbon
dioxide concentration greater than 30 vol % (2). If the carbon dioxide level
is lower in the feed gas, a higher level of ammonia (up to 18 vol %) can be
handled with design modifications (15).
A6-9
-------
Appendix A6
Claus Process
Hydrocarbons in the acid gas feed can be decomposed to hydrogen and car-
bon monoxide and react with sulfur compounds to form carbonyl sulfide and car-
bon disulfide, which are undesirable byproducts and interfere with sulfur
conversion. Combustion of hydrocarbons consumes oxygen and produces water
which suppresses formation of sulfur since water is also a product of the
hydrogen sulfide—sulfur dioxide reaction (equation 3). Excess combustion air
requirements cause inert-dilution problems similar to those resulting from
burning ammonia. Additional heat released from combustion of hydrocarbons
also increases waste heat boiler duty and total condenser duty. If high
ammonia and high hydrocarbon contents occur simultaneously, temperatures well
in excess of 1900 K can be experienced (11). Carryover of high molecular
weight hydrocarbons can also cause deactivation of the catalyst by coke forma-
tion in the catalyst beds (11, 13). Coke formation can also discolor the
recovered sulfur. Problems of high hydrocarbon contents can be minimized by
using adequate surge tanks, skimming facilities, flash tanks, filters, acid
gas condensers, and/or knockout drums so that the acid gas reaches the Claus
plant at a low temperature (e.g., below 325 K) to minimize hydrocarbon
content.
When several acid gas streams, of differing ammonia and hydrocarbon con-
tents, are being combined for the Claus process, a dual chambered reactor may
be used (7,11,14). The streams high in ammonia and hydrocarbon contents and
an excess amount of air with respect to the stoichiometry of the ammonia,
hydrogen sulfide, and hydrocarbon combustion enter the primary chamber where
ammonia and hydrocarbons are preferentially burned in an oxidizing atmosphere.
The remainder of the acid gas is then mixed with the effluent gas from the
primary chamber for combustion under reducing conditions. Heavy organics
A6-10
-------
Appendix A6
Glaus Process
(e.g., aromatics, cresols, phenols, xylenols) are converted to carbon in the
furnace. Even small amounts (e.g., 500-1000 ppmw) of heavy organics can cause
enough carbon formation to deactivate the catalyst and form dark sulfur (11).
The presence of hydrogen cyanide in the acid gas can lead to excessive
equipment corrosion and catalyst deactivation via formation of thiocyanates
(1,3). There may be other cyanides present as well that could cause problems.
Substantial quantities of hydrogen cyanide may survive the combustion zone and
react in the first catalytic reactor to produce high molecular weight com-
pounds that plug the Claus unit even more severely than the ammonium salts.
Hydrogen cyanide in the acid gas feed must be removed if a Claus plant is to
operate successfully. Elimination of hydrogen cyanide from the feed gas by
catalytic oxidation and hydrolysis prior to hydrogen sulfide combustion has
been achieved in several coke oven operations (3):
HCN + 2H2S + 0.502 = CS2 + NH, + H20 (6)
HCN + H20 = NH3 + CO (7)
3. Process Performance
Modern Claus units are capable of greater than 95 percent sulfur
recovery. However, the sulfur removal efficiency is dependent on several
variables including the following: number of catalytic conversion stages,
inlet feed stream composition, operating temperatures and catalyst mainte-
nance, maintenance of the proper stoichiometric ratio of hydrogen sulfide to
sulfur dioxide, and operating capacity factor (4,5,6).
The Claus reaction is reversible and is limited by chemical equilibrium
as shown in Figure A6-3. Sulfur recovery is enhanced by removing heat since
the desired reactions in the Claus process are exothermic. Operation at as
A6-11
-------
100
c
o
•H
ID
C
o
o
70 -
50
400
600
800
1000
1200
1400
1600
Temperature, K
Figure A6-3. Theoretical equilibrium conversion of hydrogen sulfide to
sulfur vapor (16)
A6-12
-------
Appendix A6
Clans Process
low a temperature as possible in the catalytic reactors without condensing
sulfur vapor on the catalyst is desired, but adequate reaction rates must be
maintained. The percentage of sulfur recovery increases with the number of
catalytic stages and the acid gas feed hydrogen sulfide concentration as indi-
cated in Table A6—3. The addition of a fourth catalytic stage can give recov-
eries above 99 percent. However, the Claus reaction rate is significantly
reduced at high sulfur conversion because of the lower temperatures imposed by
thermodynamics and, especially, the lower hydrogen sulfide and sulfur dioxide
concentrations. Claus plants in the U.S. generally have two or three cataly-
tic stages (4).
TABLE A6-3. SULFUR RECOVERY VARIATION RELATIVE TO ACID GAS FEED
COMPOSITION AND NUMBER OF CATALYTIC STAGES (4,5)
Mole Percent H2S
In Acid Gas Feed
15
50
90
90
90
Number of
Catalytic Stages
2
2
1
2
3
Percent Sulfur
Recovery
90
93
85
94
97
Sulfur recovery is also dependent upon catalyst performance. Very active
catalysts which resist aging are required to maintain fast reaction rates,
especially at lower temperatures. Sulfation (equation 5) is the most signifi-
cant cause of activity loss, but inorganic or organic deposits also lead to
catalyst deactivation (6). Sulfation is limited by equilibrium conversion and
increases with low hydrogen sulfide concentrations, high sulfur trioxide and
oxygen concentrations, and low temperatures. Thus, a catalyst in the second
A6-13
-------
Appendix A6
Clans Process
and third catalytic conversion stages is more susceptible to deactivation by
sulfation, even though these later stages require a more active catalyst to
maintain reasonable reaction rates at the low temperatures. Catalyst life
generally varies from 2 to 5 years depending on plant operation and contami-
nants in the feedstock (5).
Deviation above or below the 2:1 stoichiometric ratio of hydrogen snlfide
to sulfur dioxide results in a loss of Claus efficiency (5,8,17). At less
than stoichiometric amounts of air, the sulfur yield falls due to reduced
flame temperatures. At greater than stoichiometric amounts of air, it
declines due to oxidation of elemental sulfur. Operation of a Claus plant
below capacity may also impair sulfur recovery efficiency. Although the
impact of these process variables represents, in most cases, a loss of only 1
to 3 percent efficiency, at the relatively high efficiencies typical of Claus
operations (95 percent), a 1 to 3 percent efficiency loss represents a 20 to
60 percent increase in uncontrolled sulfur emissions (10).
Accurate data on the exact composition of the gas leaving the furnace is
limited due to the complexity of the reactions taking place inside the flame
and experimental difficulties encountered, particularly those of analytically
identifying the various chemical species involved (6,18,19). For example,
elementary reaction mechanisms have been developed which account for the
phenomena of increased formation of carbon sulfides (e.g., carbonyl sulfide
and carbon disulfide) observed in industrial units for increased carbon
dioxide and hydrocarbon concentrations in the acid gas feed (8,9,11,19).
However, carbon disulfide concentrations predicted by equilibrium calculations
fall below values reported in plant test data.
A6-14
-------
Appendix A6
Claus Process
In the models, carbon disulfide is formed in the flame by reaction be-
tween methane and sulfur vapor:
CH4 + S2 = CS2 + 2H2 (8)
CH4 + 2S2 = CS2 + 2H2S (9)
These reactions occur instantaneously at high temperatures (i.e., above
1300 K) but slowly at low temperatures (i.e., below 1150 K). The carbon
disulfide is then partially and slowly decomposed by water and sulfur
dioxide:
CS2 + H20 = COS + H2S (10)
2CS2 + S02 = 2COS + 1.5S2 (11)
Carbonyl sulfide is formed from carbon monoxide and elemental sulfur:
CO + 0.5S2 = COS (12)
There are several possible mechanisms for carbon monoxide formation. At high
temperatures, carbon dioxide dissociates into oxygen and carbon monoxide.
Hydrogen sulfide also dissociates thermally at high temperatures. The free
hydrogen reduces carbon dioxide and leads to the formation of carbon monoxide:
C02 + H2 = CO + H20 (13)
The formation of carbonyl sulfide by equation 12 is slow at any temperature.
For a given acid gas, the formation of carbonyl sulfide increases constantly
with temperature.
A6-15
-------
Appendix A6
Claus Process
Carbon sulfides formed in the reaction furnace undergo a partial hydro-
lysis in the catalytic converters because of the high water vapor content of
the process gas:
CSa + HaO = COS + H2S (14)
COS + H20 = CO + H2S (15)
The hydrolysis rates increase rapidly with temperature. This hydrolysis is
most noticeable in the first converter and becomes negligible in the following
converters. Hydrolysis conversions currently obtained are approximately 90
percent for carbonyl sulfide and 70 to 75 percent for carbon disulfide (6).
Despite these relatively high rates, sulfur losses as carbon sulfides are
about the same as losses through hydrogen sulfide and sulfur dioxide. Hydro-
lysis conversion of carbon sulfides in the catalytic converters could be
improved by raising the operating temperature of the first converter and using
a more active and specific catalyst. A number of alternative catalysts are
under development.
Actual operating data and predictions based on proprietary models are
shown in Tables A6-4, A6-5, and A6-6. The treated gas from the Claus process
generally contains several thousand ppmv (dry basis) of sulfur composed of
hydrogen sulfide, sulfur dioxide, carbonyl sulfide, carbon disulfide, and sul-
fur vapor. Depending on the exact nature of the acid gas feed and the operat-
ing conditions, combined carbonyl sulfide and carbon disulfide levels as high
as 5000 ppmv may exist in the tail gas (5). However, the "split-stream" pro-
cess efficiently suppresses the formation of carbon sulfides since the primary
reaction products are sulfur dioxide and water with virtually no elemental
sulfur being formed.
A6-16
-------
TABLE A6-4. REPORTED GLAUS PLANT OPERATING DATA
I
M
^J
Dilly Sulfur Production
(10' kg)
Number of Citllytic Stigei
Acid Gn Feed Source
Acid Git Feed Quint 1 ty
(Nm'/hr)
Acid Gn Feed Conpoiltion
(mol *)
H,S
x,
CO,
CH<
H,0
CS,
NH,
BCN
SO,
H,
Till Gi< Quantity !Nm'/hr)
Till Git Competition (mol %)
H,S
SO,
s
a>.
N>
H,0
CS,
COS
H,
CO
CH,
Efficiency, %
"sttuffer Chemlcil Co. (4)
Shell Oil Co. (4)
PUnt
A'
360
2
S tr li gh t-
Through
CS, Flint
12,350
96
0
0
3
0
33,380
0
0
0,
1.
64,
33,
QC
JO ,
.3
.2
.2
.2
.2
.9
.4
.3
.1
.2
,1
PUnt B * Pl.nt Cb
230 120
2 3
Strilght- Striight-
Through Through
CS, Flint Refinery
6,640 2. 570
95.5 90
0.4
0.6 8
3.4 2
0.5
21,220 7.260
1.0 1.0
0.5 0.5
0.3
1.3 3.5
64.3 65.0
32.6
94 .0 95 . 3
PUnt
D«
410
(4 uniti)
2
Straight-
Through
Refinery
7.200
83
0
11
0
4
0
40,710
0
0
0
4
60
33
90 ,
.5
.4
.5
.5
.0
.1
.8
.4
.3
.4
.2
.9
,0
PI
• nt E°
N.R.°
3
Stnlgfit-
Through
Coke Oven
N.
75
3
18
2
0
0
0
0
N.
0
0
11
87
0
0
0
98
R.
.11
.48
.20
.07
.29
.45
.05
.33
R.
.22
.13
.57
.40
.11
.15
.42
. 5
PUnt Fb PUnt
185 90
3 2
0*
Stnight- Striight-
Through Through
Sulfinol Refinery
Proceti
8,490 4,500
68 50
1
30 46
2 2
15,630 10,050
0.40 0
0.24 0
0
17.5 21
79.0 51
25
.3
.0
.1
.5
.7
.3
.2
.8
.5
.5
PUnt Hd
N.R.
2
Split-Streim
Gn PUnt
N.
8
0
90
0
N.
0
0
79
19
R.
.79
.22
.67
.32
R.
.20
.12
.18
.67
Trice
2.5
98.3 94
.7
0
0
0
0
95
.03
.41
.10
.28
.6
^Homc Oil Co. (Personal communication with C.E. LoUclle, December 13, 1979)
Not reported
-------
TABLE A6-5. PREDICTED CLAUS PLANT FURNACE PRODUCT GAS COMPOSITION
Acid Gat Feed Composition (aol %)
H,S
CO,
CH4
H,0
NBi
Furnace Product Oat Concentration (mol %)
CO
COS
CO,
CS,
> NHi
T NO
1— '
CO NO,
SN
BCN
Sulfur Conversion Efficiency
(first condenser), *
Gate 1
89.72
4.98
0.80
4.50
0.47
0.02
1.29
6.7i!0"5
3.5xlO"S
4.3x!0"6
8.4X10"13
1.6i!0"5
2.2x!0"8
71.52
Case 2
49.84
45.21
0.44
4.50
1.23
0.20
18.89
4.8X10"14
l.lxio"5
3.7xlo"8
-------
TABLE A6-6. PREDICTED GLAUS PLANT TAIL GAS COMPOSITION
Daily Snlfur Production (10' kg)
Acid Gas Feed Quantity (nole/day)
Acid Gas Feed Conposition (»ol %)
B»S
COi
CH4
CiH,
NBi
BiO
Tail Gas Quantity (mole/day)
Tall Gas Composition (BO! %)
> Ni
T COa
E BiO
BiS
SO.
S< + Si
Entrained liquid S°
Sulfur Conversion Efficiency, %
Case 1
130
11,635
90
3
0
0
0
5
32,047
60
1
34
0
0
0
0
95
.1
.6
.8
.4
.1
.6
.4
.3
.8
.4
.02
.06
.4
Case 2
127
16,631
63.
2.
0.
0.
18.
14.
49,246
60.
0.
35.
0.
0.
0.
0.
93.
0
5
6
3
7
9
4
9
9
8
4
02
06
2
Case 3
127
14,923
70
2
0
0
20
5
47,523
62
0
33
0
0
0
0
93
.2
.8
.7
.3
.9
.1
.6
.9
.7
.8
.4
.02
.06
.7
Case 4
127
20,686
50
43
0
0
0
5
41,110
47
21
27
1
0
0
0
93
.7
.5
.5
.2
.1
.2
.9
.9
.0
.5
.02
.06
.0
Case 5
126
14,028
74
3
14
1
0
5
54,835
62
1
29
0
0
0
0
92.
.7
.0
.3
.9
.1
.2
.3
.8
.8
.4
.02
.06
5
Case 6
127
12,973
80
3
7
3
0
5
45,777
61
1
31
0
0
0
0
93
.8
.2
.7
.2
.1
.8
.3
.0
.8
.4
.02
.06
.4
Ci»e 7
128
15.855
66.1
2.6
0.6
0.3
0
30.4
36,312
53.5
1.2
41.7
1.0
0.5
0.02
0.06
94.0
Calculations based on proprietary model of a Straight-Through Clans plant with 2 catalytic conversion stages (11)
-------
Appendix A6
Glaus Process
4. Secondary Waste Generation
Several secondary waste streams are generated by the Clans process and
include the following: knockout drum condensate, spent catalyst, and catalyst
regeneration offgases. The condensate from the knockout drum will contain
many of the species present in the acid gas feed; however, no data are avail-
able on this stream. Regeneration of the catalyst is performed intermittently
at some facilities when the efficiency of the process drops below a specified
level. The offgas contains sulfur dioxide and can be a potentially signifi-
cant emission depending on the method of regeneration employed. Catalyst life
generally varies from two to five years depending on plant operation and
contaminants in the feedstock (5).
5. Process Reliability
Available information indicates no special maintenance problems or
unusually hazardous conditions created by the Claus process. Principal
problems result from lack of feed or from upsets in feed rate and composition
caused by operating problems in upstream units (Personal communication with
C.L. Black, June 20, 1978). Plant shutdowns generally result due to plugging
problems. Carryover of hydrocarbons present in the acid gas feed may lead to
coke formation in the catalyst beds. High concentrations of carbon dioxide
and ammonia in the acid feed gas (e.g., greater than 300,000 ppmv carbon
dioxide and 500 ppmv ammonia) can result in precipitation of ammonium salts
throughout the Claus unit. Hydrogen cyanide in the acid gas can produce high
molecular weight compounds that plug the Claus unit even more severely than
the ammonium salts. Hydrogen cyanide can also lead to excessive equipment
corrosion.
Information concerning on-stream time for any of the Claus systems was
not present in the available literature.
A6-20
-------
Appendix A6
Claus Process
6. Process Economics
The cost of a Clans plant varies as a function of the acid gas hydrogen
snlfide content and the daily sulfur production capacity. The presence of
inert diluents (e.g., carbon dioxide, nitrogen, and air) and combustible
contaminants (e.g., ammonia, hydrocarbons, carbon monoxide, hydrogen cyanide,
carbonyl sulfide, and carbon disulfide) increase the initial investment cost
of the plant. The combustible impurities also affect the operating costs.
The Claus capital investment costs, as a function of both hydrogen
sulfide concentration in the Claus feed gas and the kg moles of sulfur removed
in the Claus process, are presented in Figures A6-4 and A6-5. The first
graph. Figure A6-4, is based on data for Claus feed streams containing 15 and
50 vol % hydrogen sulfide and includes installed costs only (4,20,21, personal
communication with C.E. Loiselle, December 13, 1979). The second graph,
Figure A6-5, is an extension of the first graph and presents total capital
costs, which include installed costs as well as contingency costs, engineering
design costs, and contractors' fees.
The utility and labor requirements estimated for the Claus process are
presented in Table A6-7. These values are based on information in published
literature (10, 15, 21-24).
A6-21
-------
10,000
7
1000
O
oo
T3
0)
100
10
A Personal Comnunicatiqn with C.E. Loiselle. December 13. 1979
D Reference 4
O Reference 4
• Reference 21
O Reference 20
0.1
I i I I I I
I I I I 1 I I
10
I I I I I I I
7 100
Installed Costs, $106 (1st Quarter 1980)
Figure A6-4. Glaus plant installed costs
A6-22
-------
10,000
7
1,000
O
60
T3
0)
O
Ki
3
U-l
100
40% sulfur in feed
I
15% sulfur in feed
I I I I I
(1st Quarter 1980; includes
contingency, engineering,
design, and contractor fee)
I I I I I I
Figure
2 4 7 10 2 4 7 100
Capital Investment, $10
A6-5. Estimated Glaus plant capital investment cost
(developed from data base, Figure A6-4)
A6-23
-------
TABLE A6-7. OPERATING REQUIREMENTS FOR THE GLAUS PROCESS (10,15,21-24)
Utilities
Steam (generated), 10* kg/kg mole sulfur removed 0.10 at 0.7 MPa
0.01 at 4 MPa
Boiler feedwater, mj/kg mole sulfur removed 0.125
Electric power, kW/kg mole sulfur removed 3.0
Process Materials
Chemicals, i (1st quarter 1980)/kg mole sulfur 0.012
Manpower
Labor, manhours/yr 1577
A6-24
-------
Appendix A6
Clans Process
7. References
1. Beavon, O.K. and R.P. Vaell. Prevention of Air Pollution by Sulfur
Plants. Presented at the Eighth Annual Technical Meeting, Southern Cali-
fornia Section, American Institute of Chemical Engineers, April 20, 1971.
2. Chute, A.E. Tailor Sulfur Plants to Unusual Conditions. Hydrocarbon
Processing, April 1977, p 119.
3. Homburg, O.A. and A.H. Singleton. Performance and Problems of Claus
Plant Operation on Coke Oven Acid Gases. Journal of the Air Pollu-
tion Control Association, Vol. 25, No. 4, April 1975, p 375.
4. Beers, W.D. Characterization of Claus Plant Emissions. Process Re-
search, Inc., PB 220 376, April 1973.
5. D.S. Environmental Protection Agency. Standards Support and Environ-
mental Impact Statement, Volume 1: Proposed Standards of Performance
for Petroleum Refinery Sulfur Recovery Plants. EPA-450/2-76-016a,
September 1976.
6. Grancher, P. Advances in Claus Technology, Part 1: Studies in Reac-
tion Mechanics. Hydrocarbon Processing, June 1978, p 155.
7. Grancher, P. Advances in Claus Technology, Part 2: Improvements in
Industrial Units and Operating Methods. Hydrocarbon Processing, Sep-
tember 1978, p 257.
8. Meisen, A. and H.A. Bennett. Consider All Claus Reactions. Hydro-
carbon Processing, November 1974, p 171.
9. Maadah, A.G. and R.N. Maddoz. Predict Claus Properties. Hydrocarbon
Processing, August 1978. p 143.
10. Dravco Corporation. Handbook of Gasifiers and Gas Treatment Systems.
ERDA, FE-1772-11, February 1976.
11. Goar, G.B. Impure Feeds Cause Clans Plant Problems. Hydrocarbon
Processing, July 1974, p 129.
12. Gas Processing Handbook. Hydrocarbon Processing, Vol. 54, No. 4, April
1979, p. 104.
13. Pearson, M.J. Determine Claus Conversion from Catalyst Properties.
Hydrocarbon Processing, April 1978, p 99.
A6-25
-------
Appendix 6
Glaus Process
14. Wiley, S. Off-gas Aids Clans Operations. Hydrocarbon Processing,
April 1980, p 127.
15. Effects of Sulfur Emission Controls on the Cost of Gasification
Combined Cycle Power Systems. Fluor Engineering and Construction,
Inc., EPRI AF-916, October 1978.
16. Kohl, Arthur and Fred Riesenfeld. Gas Purification. Gulf Publishing
Company, Houston, Texas, 1974.
17. Taggart, G.W. Optimize Claus Control. Hydrocarbon Processing, April
1980, p 133.
18. Fisher and Peterson. Control of Hydrocarbons and CO Emissions in the
Tail Gases from Coal Gasification Facilities. August 1978.
19. Fischer, H. Burner/Fire Box Design Improves Sulfur Recovery.
Hydrocarbon Processing, October 1974, p 125.
20. Conceptual Design of a Coal to Methanol Commercial Plant, Volume TVA.
Badger Plants Incorporated, Cambridge, Massachusetts, FE-2416-24 (Vol.
4A). February 1978.
21. U.S. Environmental Protection Agency. Control of Emissions from Lurgi
Coal Gasification Plants. EPA-450/2-78-012, March 1978.
22. Gas Processing Handbook. Hydrocarbon Processing, Vol. 58, No. 4, April
1979, p. 132.
23. U.S. Environmental Protection Agency. McGlamery, C.G., et al. Detailed
Cost Estimates for Advanced Effluent Desulfurization Processes. EPA-
600/2-75-006, 1974.
24. U.S. Environmental Protection Agency. Capital and Operating Costs of
Selected Air Pollution Control Systems. EPA-450/5-80-002, December,
1978.
8. Personal Communications
Black, C.L., The Ralph Parsons Co. to TRW Environmental Division. June
20, 1978.
Loiselle, C.E., Home Oil Company Limited, Calgary, Alberta, Canada to TRW
Environmental Division. December 13, 1979.
A6-26
-------
APPENDIX A7
STRETFORD PROCESS
1. Process Description
The Stretford process primarily removes hydrogen sulfide from gas streams
by chemical absorption. Liquid phase oxidation of hydrogen sulfide to ele-
mental sulfur occurs in an alkaline solution of sodium vanadate and anthra-
quinone disulfonic acid (ADA) salts. Hydrogen sulfide contents of 1 ppmv can
be achieved in the treated gas at operating pressures ranging from atmospheric
to 7.0 MPa. Sulfur of 99 percent purity can be produced molten or as a cake.
The North Western Gas Board of the British Gas Corporation originally
developed the Stretford process. A number of improvements and refinements
have been made in the basic Stretford process by various firms. Some of the
engineering and construction firms which license the Stretford process in-
clude: Woodall-Duckham (USA) Limited of Pittsburgh, Pennsylvania; Peabody
Engineered Systems of Stamford, Connecticut; Wilputte Corporation of Murray
Hill, New Jersey; Black, Sivalls, and Bryson, Inc. of Houston, Texas; The
Pritchard Corporation of Kansas City, Missouri; and The Ralph M. Parsons
Company of Pasadena, California.
The Stretford process (1,2,3,4,5,6), shown in Figure A7-1, is basically a
one-step process to convert hydrogen sulfide to elemental sulfur according to
the following overall reaction:
2HaS +0a = 2S + 2H20 (1)
The raw acid gas stream is contacted conntercnrrently in the absorber with an
aqueous solution of sodium carbonate/sodium bicarbonate containing the
disodium salt of the 2,7 isomer of anthraquinone disulfonic acid (ADA) and
A7-1
-------
©->•
CIRCULATING SOLUTION
ABSORBER
LEGEND:
1. Raw Gas
2. Air
3. Chemical Make-up
4. Water Make-up
5. Steam
6. Product Gas
7. Sulfur Cake
8. Molten Sulfur
9. Solution Purge
10. Oxidizer Vent Gas
COOLING
TOWER
SURGE TANK
ALTERNATIVE
SULFUR RECOVERY
METHODS
FILTRATION
FILTRATION AND
AUTOCLAVE
CENTRIFUGATION
CENTRIFUGATION
AND HEATING
AQUEOUS
EFFLUENT
•KD
Figure A7-1. The Stretford sulfur removal process
-------
Appendix A7
Stretford Process
pentavalent vanadium. The hydrogen sulfide is absorbed by the alkaline solu-
tion and reacts with the sodium carbonate:
HaS + NajCO, = NaHS + NaHCOj (2)
The alkaline solution also absorbs hydrogen cyanide and sulfur dioxide from
the gas but does not absorb carbon dioxide. Pentavalent vanadium reacts
almost immediately with the sodium bisulfide (NaHS) to produce elemental sul-
fur:
2NaV03 + NaHS + NaflCO, = S + Na^O, + Na^CO, + H20 (3)
Citric acid is used as a completing agent to keep tetravalent vanadium dis-
solved in the alkaline solution. The liquor from the absorber is held in a
reaction tank to allow sufficient time (about 10 minutes) for these reactions
to be completed. The purified gas exits at the absorber top.
Solution from the reaction tank flows to an oxidizer where air is sparged
into the solution at ambient temperature and atmospheric pressure to reac-
tivate the Stretford solution by oxidizing tetravalent vanadium to pentavalent
vanadium:
ArjA
2Na1V,0J + Oa = 4NaV03 (4)
This reaction is very slow and needs a catalyst in the form of ADA to act as
an oxygen carrier. The air also causes the sulfur to float as a heavy froth
on the liquid surface.
A7-3
-------
Appendix A7
Stretford Process
The sulfur froth overflows from the oxidizer to a froth tank while the
sulfur-free reactivated Stretford solution flows through a cooling tower
before falling into a surge tank and being recirculated to the absorber. The
cooling tower is designed to evaporate sufficient water to maintain the system
in water balance and to control the solution temperature which is normally
kept at 305 to 314 K. A solution heater (not shown in Figure A7-1) is usually
provided in case additional heat is required to promote evaporation of water
to maintain the water balance.
The sulfur froth can be processed several ways to recover the sulfur dep-
ending on the desired end product, total sulfur production, and utilities
cost. For small sulfur capacities, simple filtration of the sulfur froth may
be economic. For larger sulfur production rates, one or more stages of centri-
fuging (or filtering) followed by heating may be used. In this mode of opera-
tion, the sulfur cake is reslurred with steam condensate to reduce the
Stretford reagent loss. The slurry is pumped through a sulfur melter and then
into a steam-jacketed separator, where molten sulfur is recovered and the
remaining dilute Stretford solution is returned to the system. Any wash water
and reslurry water provide additional water load to the cooling tower duty.
In an alternative mode of operation (i.e., direct autoclaving), the sulfur
froth is pumped through a sulfur melter into an autoclave, where the phases
are separated and the hot Stretford solution is returned to the surge tank.
This mode of operation has the advantages of simplicity of operation, lower
capital investment, and reduction of water evaporation load on the cooling
tower.
Various side reactions proceed simultaneously with the absorption and
regeneration reactions (2,6,7). Some of the sodium bisulfide is converted to
sodium thiosulfate (NaJS10J) and to a less extent, sodium sulfate (NaJS04).
Hydrogen cyanide in the feed gas reacts irreversibly with the Stretford solu-
tion to form sodium thiocyanate (NaCNS) which is stable in solution. These
A7-4
-------
Appendix A7
Stretford Process
salts are inactive in the desulfnrization reactions. The accumulation of
these salts reduces the solubility of vanadium and ADA in the solution, de-
creases the rate of Stretford solution reactivation, and could lead to crys-
tallization of the salts. To maintain these salts at an acceptable level of
20 to 30 wt %, a portion of the Stretford solution must be purged (8). Pre-
washing could also be used for high hydrogen cyanide concentrations (e.g.,
above 50 ppm) since hydrogen cyanide is stable in solution (6).
Over 100 Stretford units are currently in operation world-wide, with
capacities ranging from 2.7 x 10» to 5.4 x 1012 Nm*/D with sulfur removal
rates from 0.45 to 82 Mg/D (1,9). Applications include purification of Claus
unit tail gases, amine regenerator off-gases, coke oven gases, Rectisol off-
gases, coal gasification product streams, low Btu gas streams, oil gasifica-
tion SNG streams, reformed petroleum products, and vent gases from geothermal
steam turbines. A commercial scale Stretford unit is being installed at the
SASOL Lurgi coal gasification facility in South Africa (10). Stretford sulfur
recovery processes are also included in the designs of at least four Lurgi
coal gasification facilities in the U.S. (i.e., projects sponsored by Tenneco
Coal Gasification, Nokota Company, WyCoal Gas, Inc., and Great Plains Gasifi-
cation Associates) (11).
2. Process Applicability
The reactions upon which the Stretford process is based are essentially
insensitive to pressure (1,7,8). Absorber operating pressures vary from near
atmospheric to 7 MPa. The solution oxidizer and sulfur recovery systems oper-
ate at atmospheric pressure. When the Stretford process is operated on
streams containing high levels of hydrocarbons, a flash drum may be incor-
porated between the absorber and the primary oxidizer such that hydrocarbons
A7-5
-------
Appendix A7
Stretford Process
are recovered as a low pressure fuel gas stream rather than released to the
atmosphere (Personal communication with A.J. Grant, December 5, 1977). Oper-
ating temperatures throughout the unit are in the range of ambient to 322 K
(1,8).
Inlet hydrogen sulfide concentrations as low as 300 ppmv and as high as
95 percent can be processed in a Stretford unit (6). However, the Stretford
process is most applicable to streams containing less than 15 percent hydrogen
sulfide (7). The Stretford process can tolerate significant amounts of carbon
dioxide. The Stretford solution reaches an equilibrium with respect to the
carbon dioxide in the gas and only relatively small amounts of carbon dioxide
are removed by the process. However, extremely high carbon dioxide concentra-
tions in the feed gas can cause pH reduction and reduced efficiency (7). Pre-
liminary information at the SASOL Lurgi coal gasification facility in South
Africa indicate that biological growth due to high carbon dioxide levels in
the feed is primarily responsible for plugging problems in the Stretford
absorption towers. Hydrogen cyanide concentrations as high as 2000 ppmv can
be handled although accumulation of sodium thiocyanate produced from hydrogen
cyanide increases the purge requirement (6).
3. Process Performance
Hydrogen sulfide concentrations can be reduced to less than 1 ppmv in the
purified gas stream (2,6,7). Hydrogen sulfide in the feed gas stream is con-
verted to sodium bisulfide and subsequently to elemental sulfur, as discussed
earlier. Some of the sodium bisulfide is also converted to sodium thiosulfate
and, to a lesser extent, sodium sulfate. Approximately 1 to 2 moles of sodium
thiosulfate are formed for every 100 moles of hydrogen sulfide in the feed gas
(7). All of the sulfur dioxide in the feed gas is converted to sodium sul—
fate. Ninety percent of the methyl mercaptans present in the feed gas are con-
verted to sodium bisulfide and subsequently to elemental sulfur (Personal
A7-6
-------
Appendix A7
Stretford Process
communication with C.A. Vancini, February 2 and 6, 1980). Other sulfur com-
pounds (e.g., carbonyl sulfide and carbon disulfide) are not removed by the
Stretford process (6,7). Virtually all of the hydrogen cyanide in the feed
gas is converted to sodium thiocyanate. Approximately one mole of sodium
thiocyanate is formed for every mole of hydrogen cyanide in the feed gas (7).
Carbon dioxide is not removed from the feed gas (6).
Although there are many installations where the Stretford process is
employed, actual operating data are limited. Stream characterizations of most
of the effluent streams, including trace and minor constituents, are lacking.
Data from a design for a Stretford unit that would treat lean acid gas from
the Rectisol unit at El Paso Natural Gas Burnham coal gasification facility
are shown in Table A7-1 (12). Typical compositions of the Stretford abosrber
offgas from several Beavon sulfur removal plants which utilize the Stretford
process are presented in Table A7-2 (2,4,13). The recovered sulfur normally
consists of 99.5 percent sulfur with small amounts of contaminants such as
vanadium salts and sodium thiocyanate. Unreacted mercaptans are split between
the purified gas stream and the oxidizer vent stream (Personal communication
with C.A. Vancini, February 2 and 6, 1980). Carbonyl sulfide, carbon disul-
fide, and the lower hydrocarbons (e.g., containing 5 carbon atoms or less)
essentially appear in the purified gas stream. Higher hydrocarbons as well as
ammonia will be present in the oxidizer vent stream, if present in the
absorber feed gas.
4. Secondary Waste Generation
Three secondary waste streams are generated in the Stretford process:
Stretford solution purge, oxidizer vent gas, and evaporation and drift from
the cooling tower.
A7-7
-------
TABLE A7-1. DESIGN DATA FOR A STRETFORD DNIT TREATING RECTISOL OFFGAS AT EL
PASO NATURAL GAS BURNHAM COAL GASIFICATION FACILITY (12)
Component
CO %
H S, ppmv
COS , ppmv
CS , ppmv
HCN, ppmv
CO, %
CH4, %
C H , %
C H f %
»:•**
Na S 0 , %
Na§Q$, 3%
NaVO , %
ADA,3*
NaHCO + Na CO , %
3 23
Raw Gas
96.0
7400
77
2
30
0.17
0.53
0.22
0.30
0.43
1.6
Absorber
Offgas
99.0
8
75
2
0
0.16
0.52
0.22
0.29
0.42
4.32
Solvent
Purge
80.0
10.8
4.4
0.7
1.1
3.0
TABLE A7-2. TYPICAL COMPOSITIONS OF STRETFORD ABSORBER OFFGAS
IN BEAVON SULFDR REMOVAL PLANTS
p
Component Facility A
H S , ppmv
S02 , ppmv
COS , ppmv
CS2, ppmv
(JH So , ppmv
CO , ppmv
CH4, ppmv
N2 + AR, %
CO , %
E * %
i
5
<20
50
20
<5
NR
NR
NR
NR
NR
Facility Bb
<1
<1
30
9
NR
500
200
89
7
3
Facility C°
<0.1-7
<5
7-23
1
NR
565
221
NR
5.7
5.6
Facility D°
<1
<0.4
5
0.5
NR
250
NR
NR
NR
NR
Facility E°
<1
<1
15
0.1
NR
670
NR
NR
NR
NR
.Reference 4
Reference 13
NR - Not reported
A7-*
-------
Appendix A7
Stretford Process
As discussed previously, a portion of the Stretford solution must be
purged to prevent excessive accumulation of sodium thiosulfate, sodium sul-
fate, and sodium thiocyanate. Further treatment of this liquor would be
expected before discharge, because it contains materials toxic to aquatic life
and has a high COD value due to its high sodium thiosulfate content. The
purge liquor also contains valuable chemicals: ADA, vanadium salts, and
sodium carbonate.
Several methods have been proposed for the treatment of the spent
Stretford solution. The Ralph M. Parsons Company offers a process for treat-
ment of this stream with hot sulfuric acid, which converts most of the thio-
sulfate to sulfate (2). As sulfate builds up and at relatively high thio-
sulfate levels, the sulfate is crystallized from another slipstream. With
this process, essentially only sodium sulfate crystals are discarded.
In another process, the spent Stretford solution is acidified with sul-
furic acid to decompose the thiosulfate to sulfur and sulfur dioxide, and
limed to remove added sulfate and restore the pH of the solution (8). The
thiosulfate decomposition reaction takes place in steps according to the
following equations:
• Protonation of thiosulfate ions
SaO§" + H+ = HS^Oj" (5)
HS^,' + H+ = H2Sa03 (6)
• Internal redox reaction
* Hiso,
A7-9
-------
Appendix A7
Stretford Process
• Evolution of SOa
HaS03 = HaO + S0a (8)
Sulfur dioxide is stripped from the acidified solution by using steam and re-
quires further treatment (e.g., recovered sulfur dioxide could be recycled to
the Claus plant). The sulfur floats to the top and is recovered by skimming
or filtration. The clear Stretford solution is then reacted with lime to
raise the pH and precipitate sulfate/sulfite ions as CaS04/CaSOj. The sulfate
removed in this step includes both sulfate from byproduct formation in the
Stretford process as well as sulfate from sulfuric acid addition in the acid-
ification step. The slurry is settled in a thickener, and the solution is
recycled to the Stretford absorber. The calcium salts are disposed of after
washing and filtration.
In the reductive incineration method, the spent Stretford solution is
burned with air and a feed gas (usually a small stream taken from the main
feed to the Stretford plant). This produces a gas stream containing hydrogen
sulfide and water vapor and a solid residue containing soda ash and reduced
vanadium salts (6,14). The salts are returned to the Stretford process as
makeup chemicals and the hydrogen sulfide gas and water vapor are recycle to
the absorber. Thus, the reductive incineration process recovers expensive
chemicals while effectively attaining a "zero" discharge of solution purge.
Other methods proposed for treatment of spent Stretford solution include:
evaporative or spray drying, biological degradation, and oxidative combustion
(6).
The Stretford oxidizer vent will contain mainly nitrogen, oxygen, and car-
bon dioxide and will be saturated with water. However, half of the unreacted
mercaptans and all of the ammonia will also be present (Personal communication
A 7-10
-------
Appendix A7
Stretford Process
with C.A. Vancini, February 2 and 6, 1980). The oxidizer vent may also
include some of the heavier hydrocarbons (7,10).
Cooling tower evaporation and drift are only present in designs that
employ a cooling tower to evaporate sufficient water in order to maintain
water balance for the system. The evaporation losses can be estimated from
the water balance. Drift losses should be about 0.005 percent of the
Stretford solution recirculation rate. In contrast to evaporation losses, the
drift droplets contain the same chemicals as the recirculated Stretford solu-
tion and purge stream (e.g., ADA, vanadium, sodium carbonate, sodium thio-
sulfate, and sodium sulfate).
5. Process Reliability
The oxidation vessels and slurry tanks in Stretford units are normally
coated with plastic to avoid corrosion by deposited sulfur. Corrosion can
occur in other parts of the Stretford unit if sulfur is allowed to accumulate.
However, corrosion is not normally considered a problem with Stretford units.
Operating experience indicates that the most common problem is plugging
in the absorber and sulfur froth lines (15,16). Preliminary information at
the SASOL Lurgi coal gasification facility in South Africa indicates that bio-
logical growth due to high carbon dioxide levels in the feed is primarily
responsible for plugging problems in the Stretford absorption towers. Conven-
tional absorbers consist of packed towers. Plugging normally occurs in the
bottom 2 to 3 meters of packing after 6 to 12 months operation (2). The use
of a venturi scrubber followed by a short conventional absorber will result in
appreciable reductions in investment and maintenance costs since plugging is
limited to the venturi scrubber which can be more easily cleaned than a con-
ventional absorber (2,17). Stretford designs with two venturi scrubbers in
A 7-11
-------
Appendix A7
Stretford Process
parallel (one spare) followed by a short absorber can conceivably eliminate
plant shutdown for absorption equipment cleanout. Plugging of sulfur froth
lines is greatly reduced by using long-radius elbows and flushing the lines
before shutdown. Some provision must be made for sour gas disposal during
absorber cleanouts and any mechanical failures.
6. Process Economics
Investment and operating costs are affected by operating pressure, hydro-
gen sulfide content of feed gas, and disposition of sulfur products. The cap-
ital investment costs for a Stretford unit are presented as a function of sul-
fur removal capacity in Figure A7-2. The capital investment costs were ob-
tained from studies recently performed by Catalytic Inc. (7), U.S. Environ-
mental Protection Agency (18), and Oak Ridge National Laboratory (19), and
from Black & Veatch (20). From data provided by J. F. Pritchard Co. to
Black Si Veach (20) an equation was developed relating capital cost to sulfur
recovered and flow rate as follows:
(9)
where: A = sulfur removed, kmol/hr,
B = feed gas flow rate, m3/min per kmol sulfur removed
per hour, and
C = total capital cost less interest during construction, tlO*.
Operating requirements for the Stretford process as presented in Table
A7-3 were extracted from information provided by Ralph M. Parsons Co. to
Black fi Veatch with respect to the Beavon sulfur removal process. Other cost
data in published literature (6,7,10,13,18,19,21-25) were also reviewed.
A7-12
-------
1000
7
M
0)
r-t
o
e
00
-o
0)
>
o
I
Pi
100
10
7
8
o
D
O
O
i i i i i i
D
a
A Reference 7
O Reference 18
Q Reference 19
O Reference 20
I I I I I I
4 7 10 2 4
Capital Investment, $106 (1980 Dollars)
7 100
Figure A7-2. Stretford capital investment cost (lass interest
during construction)
A7-13
-------
Appendix A7
Stretford Process
TABLE A7-3. OPERATING REQUIREMENTS FOR THE STRETFORD PROCESS*
For 36.1 kmol/hr (27.3 Mg/day) unit
cooling water = 5.68 ms/min
electricity = 610 kW
For 41 kmol/hr (31.45 Mg/day) unit
chemicals = J350,000/year
steam = 7.28 Mg/hr (16050 Ib/hr)
Labor estimated at one half man per shift
Data extracted from information provided to Black & Veatch by the Ralph
Parsons Company for the Beavon Sulfur Removal Process, Nov. 1980.
7. References
1. Gas Processing Handbook. Hydrocarbon Processing, 54(4):104, April 1975.
2. Kouzel, B., R. H. Fuller, E. J. Jirus, and B. B. Woertz. Treat Low
Sulfur Gases with Beavon Sulfur Removal Processes and the Improved Stret-
ford Process. Paper presented at the 27th Annual Gas Conditioning Confer-
ence, University of Oklahoma, March 7-8, 1977.
3. Gas Processing Handbook* Hydrocarbon Processing, 58(4):132, April 1979.
4. GPA H2S Removal Panel. Processes Clean up Tail Gas. The Oil and Gas
Journal, pp. 160-166, August 28, 1978.
5. Beavon, D. K. and R. P. Vaell. Prevention of Air Pollution by Sulfur
Plants. Paper presented at the Eighth Annual Technical Meeting, Southern
California Section, American Institute of Chemical Engineers, April 20,
1971.
6. Dravo Corporation, Handbook of Gasifiers and Gas Treatment Systems.
ERDA FE-1772-11, February 1976.
7. U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina. The Stretford Process. December 15, 1976.
A7-14
-------
Appendix A7
Stretford Process
8. Yan, T. Y. and W. F. Espenscheid. Removal of Thiosulfate/sulfate
from Spent Stretford Solution. Environmental Science & Technology,
14(6): 732-735, June 1980.
9. Gray, J. A. Gasification of Carbonaceous Feedstocks. Chemical Engin-
eering Progress, p. 73, March 1980.
10. Atkins, T. W. Problems Associated with Controlling Sulfur Emissions
from High-Btu Coal Gasification Plants. C. F. Braun and Company report
prepared for ERDA under Contract No. E(49-18)-2240, December 1976.
11. Beychok, M. R. and W. J. Rhodes. Comparison of Environmental Design
Aspects of Some Lurgi-Based Synfuels Plants. Presented at the Sixth
Symposium on Environmental Aspects of Fuel Conversion Technology, Denver,
Colorado, October 26-30, 1981.
12. Cameron, D. S. Treating Problems Associated with Coal Gasification.
El Paso Natural Gas Company, El Paso, Texas. Proceedings of Gas Condi-
tioning Conference, ERDA Document No. COMF-740365, March 1974.
13. U. S. Environmental Protection Agency. Standards Support and En-
vironmental Impact Statement. Volume I: Proposed Standards of Perform-
ance for Petroleum Refinery Sulfur Recovery Plants. EPA-450/2-76-016a,
September 1976.
14. Vasan and Sirini. The Holmes-Stretford Process for Desulfurization
of Tail Gases from Acid-Gas Systems. Peabody Process Systems, Inc.
Paper presented at the Ammonia-from-coal Symposium, Muscle Shoals, Ala-
bama, May 9, 1979.
15. Rucker, J. E. of American Petroleum Institute to R. M. Statnick of
U.S. Environmental Protection Agency. Reliability Data on Refinery Tail
Gas Cleanup Systems. July 30, 1981.
16. Laengrich, A. R. and W. L. Cameron. Tail Gas Cleanup Addition May
Solve Sulfur-Plant Compliance Problem. The Oil and Gas Journal, pp. 159-
162, May 27, 1978.
17. Haynes, W. P. Synthane Process Update, mid-1977. Fourth Interna-
tional Conference on Coal Gasification, Liquefaction and Conversion,
August 2, 1977.
18. U. S. Environmental Protection Agency. U. S. EPA Control of Emis-
sions from Lurgi Coal Gasification Plants. EPA-450/2-78-012, March 1978.
A7-15
-------
Appendix A7
Stretford Process
19. Edward, M. S. H^S Removal Process for Low-Btu Coal Gasification.
Oak Ridge National Laboratory, ORNL/TM 6077, January 1979.
20. J. F. Pritchard Co., Kansas City. Data obtained through J. Geick,
Black & Veatch, Consulting Engineers, Kansas City, Missouri, 1980.
21, Effects of Sulfur Emission Controls on the Cost of Gasification Combined
Cycle Power Systems. Fluor Engineering and Construction, Inc. EPRI-AF-
916, October 1978.
22. Oby, Vasan, and Sirini. Holmes-Stretford Process Offers Economic
H»S Removal. Peabody Systems, Inc. The Oil and Gas Journal, pp. 78-80,
January 2, 1975.
23. Economics of Current and Advanced Gasification Processes for Fuel
Gas Production, EPRI-AF-244, July 1976.
24. U. S. Environmental Protection Agency. Detailed Cost Estimates for
Advanced Effluent Desulfurization Processes. EPA-600/2-75-006, 1974.
25. D. S. Environmental Protection Agency. Capital and Operating Costs
of Selected Air Pollution Control Systems, EPA-450/5-80-002, December
1978.
8. Personal Communications
Grant, A.J., Woodall-Duckham to TRW Environmental Division. December 5,
1977.
Vancini, C.A., Peabody Process Systems, Inc., Stamford, Connecticut
to TRW Environmental Division. February 2 and 6, 1980.
A7-16
-------
APPENDIX A8
SCOT PROCESS
1. Process Description
The Shell Clans Offgas Treating (SCOT) process was developed by the
Royal Dutch Shell Laboratories in the Netherlands and is licensed in the U.S.
by the Shell Development Company in Houston. The process is designed to
reduce emissions of sulfur species from Claus plant tail gas.
A typical flow diagram for the SCOT process is presented in Figure A8-1.
The process essentially consists of two sections: a reduction section and an
alkanolamine absorption section. In the reduction section, all sulfur com-
pounds and any free sulfur in the Claus offgas are converted into hydrogen
sulfide over a cobalt-molybdate catalyst at a temperature of about 570 K in
the presence of a reducing gas. The reducing gas can be hydrogen or a mixture
of hydrogen/carbon monoxide mixtures or hydrogen/hydrocarbon mixtures supplied
from an outside source, or can be generated by substoichiometric combustion of
a fuel gas in the direct heater preceding the reduction reactor.
In the presence of hydrogen and steam, the following reactions take place
in the reduction reactor:
S0» + 3H2 = His + 2Ha° (1)
S, + 8H2 = 8H2S (2)
COS + H20 = C02 + H2S (3)
CS2 + 2H20 = C0± + 2H2S (4)
When carbon monoxide is also present as a reducing agent, the following
additional reactions may occur:
S02 + 3CO = COS + 2C02 (5)
8S + SCO = 8COS (6)
A8-1
-------
TOCLAUSUNIT
INCINERATOR
REDUCING GAS —^
oo
I
to
CtAl«UNIT ^
FUEL GAS — *•
AIR ->
LINE HEATER
f"^
REAC
X
v^.
^1
TOR
X, L f STEAM
r8,
JL
_UI_
COOLING
TOWER
T"
S-^
(
i
' — «s ^ A'H OR
A ^ WATER
^-^
. | CONOENSATE |
....*»! T0 1
LEAN AMINE
tn
SOUR WATER
STRIPPER
Figure A8-1. Shell Glaus off-gas treating (SCOT) process schematic flow diagram
-------
Appendix A8
SCOT Process
CO + HaO = CO, + Ha (7)
CO + H,S = COS + H2 (8)
The reactor effluent is subsequently cooled in a heat exchanger and a
quench tower. The excess condensate from the quench, which contains a small
amount of hydrogen sulfide and possibly NH3, is sent to a sour water stripper.
Any H2S and NH, present is released and recycled to the Claus feed.
The cooled gas, which contains up to 3 mole percent hydrogen sulfide and
up to 20 mole percent carbon dioxide, is countercnrrently scrubbed with an
alkanolamine solution in an absorption column. The treated gas, containing
traces of hydrogen sulfide and carbonyl sulfide, is incinerated. The rich
amine solution is regenerated in a conventional stripper, and the desorbed
hydrogen sulfide is recycled to the front of the Claus plant for recovery of
additional sulfur.
As a result of design requirements in different applications, four major
differences among SCOT designs have been developed. These relate to 1) Claus
tail gas pressure, 2) reducing gas specification, 3) amine solvent selection,
and 4) the level of integration of the solvent system with upstream processes.
The following brief descriptions of these design differences are excerpts from
Reference 1.
With or without gas compression—
Depending on the pressure of the tail gas at the outlet of the Claus
unit, process gas compression may be required to overcome the pressure drop
over the SCOT unit. A typical value for this pressure drop is 25 kPa.
A8-3
-------
Appendix AS
SCOT Process
To minimize the compression energy the gas compressor, if required, is
intalled at a point of low volumetric gas flow (i.e., in between quench tower
and absorber). On occasions it is economically more attractive to install a
two-stage quench tower with gas compression in between the two stages. In
this case the compression energy (i.e., process gas temperature increase, is
dissipated in the second quench stage and, therefore, does not present an addi-
tional heat load for the absorber solvent circulation).
From a process point of view there is no objection to a subatmospheric
pressure at any point in the unit. However, to exclude the possibility of air
ingress — which can lead to corrosion — above—atmospheric process gas pres-
sures should be preferred.
External/Internal reducing gas—
SCOT units located in oil refineries are usually supplied with external
reducing gas, since the availability of hydrogen or a hydrogen-rich gas does
not present a problem. In other locations (e.g., in natural gas treating
plants where hydrogen is not available) reducing gas is generated by firing
the SCOT line burner at a snbstoichiometric air/fuel gas ratio, thus convert-
ing part of the fuel gas to hydrogen and carbon monoxide. The two functions
of this burner, viz. supplying heat and generating reducing gas, are quite
compatible and present no control problems ia commercial units. With the pro-
prietary substoichiometric burner firing at air fuel gas ratios of down to 70
percent of stoichiometry is feasible. It is interesting to note that a number
of SCOT units are operated without the addition of reducing gas. The tail gas
of a Claus unit operating at a high recovery and a high main combustion cham-
ber temperature usually contains sufficient hydrogen and carbon monoxide for
the reduction reactions in the SCOT unit.
A8-4
-------
Appendix AS
SCOT Process
Different solvents —
The selection of the type of solvent used in SCOT absorbers is largely
determined by the composition of the Claus tail gas, more specifically its C02
content. For amine treating processes using MEA (monoe thanol amine) or DEA
(die thanol amine) or the ADIP solvents based on DIPA (diisopropanol amine) and
MDEA (me thyl die thanol amine ), the rate of H2S absorption is high and practi-
cally the same, since the absorption is gas-film-controlled. Therefore, to
achieve a given purity of the treated gas the same number of absorption trays
is required for these four solvents.
However, the rate of absorption of C02 - present in the Claus tail gas in
varying quantities - is quite different for the various amine solutions. The
secondary amines DEA and DIPA (and also the primary amine MEA) react with C02
to form predominantly car hamate, whereby one C02 molecule binds two amine
molecules :
C02 + 2R2NH2 = R2NCOO~ + R2N+H2 (9)
Since the rate of carbamate formation for DIPA is lowest, this solvent is
more selective.
Tertiary amines — being fully substituted — are even more selective and
react with C02 to form bicarbonates in accordance with the relatively slow
reaction:
C02 + H20 + RaNR' = HO> + R2NRn (10)
A8-5
-------
Appendix AS
SCOT Process
Contrary to the case of carbamate formation, in this reaction one molecule
of C02 binds only one amine molecule. This fact, in combination with
fully utilizing the difference in absorption rate between H,S and C02, allows
for the design of a most selective ADIP absorption system. Reported C02 co-
absorption in DIPA systems are up to 20-30 percent (2,3). CO^ coabsorption in
MDEA systems is less than 10 percent (4).
Calculations have shown that the lower maximum H2S loading of a lower-
alkalinity solvent does not necessarily result in a solvent circulation
greater than with a stronger amine solution. The advantage of lower steam
requirement for the regeneration of low-alklanity solvent (per unit volume)
can therefore be realized.
Different levels of integration—
Three different levels of integration of the solvent systems have been
applied, viz. add-on, common regeneration, and cascaded lineup. The term add-
on refers to a self-contained SCOT unit with a fully independent solvent
system, including solvent regeneration which as such can be "added-on" to one
or more Claus units.
In the two other systems the SCOT solvent section is integrated with the
solvent section of the primary gas treating unit. Obviously in an integrated
system all treating units must use the same type of solvent.
In the common regeneration system the loaded SCOT solvent is regenerated
together with the fat solvent stream(s) from the primary gas treaters in one
common regenerator - usually considered part of the primary gas treaters.
A8-6
-------
Appendix A8
SCOT Process
Owing to the low H,S partial pressure in a SCOT absorber, the solvent leaving
this column is only partly loaded with HaS. The loading capacity left can be
utilized for the removal of HaS from a gas with a high H2S partial pressure.
In the cascaded line-up this further loading takes place by introducing the
SCOT "fat" solvent into a primary gas absorber. From this column the solvent
is routed to the (again common) solvent regenerator.
Most SCOT units built so far are of the add-on type. However, in recent
SCOT designs there is a distinct trend towards the application of integration,
in particular the cascaded line-up.
The first two commercial SCOT units began operation in the fall of 1973.
By late 1980, 32 plants in the U.S. and 25 plants outside the U.S. were in
operation (5). Approximately 8 plants in the U.S. and 7 plants outside the
U.S. are in various stages of planning, design, and construction. These
plants are located mostly in petroleum refineries and range in size from 3 to
2100 Mg/day of sulfur recovered in the Claus unit. The largest SCOT unit in
the U.S., at 850 Mg/day of sulfur, is located in the Shell Oil Plant at
Eustace, Texas.
2. Process Applicability
The SCOT process is applicable for treatment of tail gases from Claus
plants operating under a variety of conditions. As discussed previously, fac-
tors such as the Claus tail gas pressure and composition, available reducing
gas, and desired level of integration with upstream processes influence the
overall SCOT design.
A8-7
-------
Appendix AS
SCOT Process
3. Process Performance
The vent gas from the SCOT absorber typically contains 200 to 500 ppmv
hydrogen sulfide (2,6). In Table A8-1, typical compositions of Glaus and
SCOT unit gas streams are presented. For this example case, the SCOT offgaj
contains 300 ppmv H2S, 10 ppmv COS, and 1 ppmv CSa. The low concentration of
COS could be the result of the low C03/EtO ratio (less than 0.1) in the Claus
offgas, leading to favorable conditions for the hydrolysis of COS in the SCOT
reactor. The calculated equilibrium COS concentration for this case, however,
is even lower at 1 ppmv. This is in conflict with the observation that the
concentration of COS approaches thermodynamic equilibrium (7).
Three other sets of sulfur emission data have been collected (8) and are
presented in Table A8-2. However, none of the data presented provides infor-
mation on the concentrations of COS in the SCOT offgas. Also, all three
plants started up prior to 1978 and were probably not designed to meet the
current New Source Performance Standards (NSPS) for petroleum refinery Claus
sulfur recovery plants.
The total sulfur emissions from the SCOT unit are the sum of H^S, COS,
and CSa in the absorber effluent. The amount of HaS present in the absorber
effluent is a function of the following factors:
• Absorber height (No. of trays),
• Rich amine solution acid gas loading,
• Steam rate (which determines the amount of residual acid gas
present in the lean amine solution),
• Temperature at the top of the absorber, and
• Operating pressure of the absorber.
A8-8
-------
TABLE A8-1. TYPICAL COMPOSITIONS OF GAS STREAMS IN CLAUS AND SCOT UNITS (7)
(Basis: 94% sulfur recovery in Claus unit)
Incinerated
Composition, % vol.
H S
sS,
S vapor and mist
c6s
cs,
CO
coa
HC (MW:30)
H
H*0
N
°2
Total
Temperature, K
Pressure, kPa
Gas quantity, mole
TABLE
Plant Location
Claus
Intake
89.9
___
4.6
0.5
—
5.0
—
—
100.0
310
147
relative 1
Claus
Offgas
0.85
0.42
0.05
0.05 10
0.04 1
0.22
2.37
1.60
33.10
61.30
—
100.00
410
127
3.0
SCOT
Offgas
0.03 <10
ppm vol
ppm vol
3.05
0.96
7.00
88.96
—
100.0
310
98
2.2
SCOT
Offgas
ppm vol
0.02
—
4.42
—
9.84
83.94
1.78
100.0
920
98
3.5
A8-2. SULFUR CONCENTRATION IN SCOT OFFGAS (8)
Sulfur
Concentration
Before Incineration
in SCOT Offuas
After Incineration
Shell/Deer Park, Texas
Gulf/Port Arthur, Texas
Texaco/Port Arthur, Texas
517 ppmv
200-500 ppmv H2S
average 300 ppmv
160 - 350 ppmv,
average 200 ppmv on the
basis of 50% excess air
267 ppmv
A8-9
-------
Appendix A8
SCOT Process
Removal of H2S is enhanced by increasing the absorber height, the steam
rate, and the absorber operating pressure and by decreasing the rich amine
solution acid gas loading and the absorber temperature. Currently, most SCOT
units are designed to yield a typical H2S concentration of 200 ppmv in the
absorber effluent (Personal communication with J. Schilk, December 3, 1981).
COS and CS2 are largely converted to H2S by hydrolysis in the SCOT
reduction reactor. The level of CS2 in the reactor effluent is typically 1
ppmv. The amount of COS present, however, will be determined by:
(COS) = i_ 1 (H*S)
-------
Appendix A8
SCOT Process
Based on available information, the vent gas from the absorber in the
current design of the SCOT process can be expected to contain less than 250
ppmv of total reduced sulfur compounds. This vent gas is normally inciner-
ated before discharge to the atmosphere.
4. Secondary Waste Generation
Two secondary waste streams are generated by the SCOT unit. The first is
the sour water stream from the quench tower. This sour water stream results
fron condensation of water vapor present in the reduction reactor effluent.
The flow rate of this sour water stream can be estimated from knowledge of the
amount of water vapor present in the Claus tail gas, the operating pressure
and temperature of the quench tower (typically 310 K), and by assuming that
the gas stream leaving the quench tower is saturated with water vapor. The
sour water stream is sent to the sour water stripper for treatment. Any H S
present and possibly NH? are released and recycled to the Claus feed.
The second waste stream generated is spent catalyst from the reduction
reactor. The initial catalyst requirement is approximately 28 wt % of the
initial Claus catalyst requirement (9), which translates to 850 kg catalyst
per kmol/hr H2S converted from S0a, COS, and CS2. For a 150 Mg/D Claus unit,
the initial catalyst requirement for the SCOT unit is calculated to be 12.6
Mg. With a catalyst service life of 5 years (8), this amount of spent cata-
lyst will require disposal every 5 years. The spent catalyst will be cobalt
molybdate on alumina with traces of sulfates, sulfides, elemental sulfur, and
carbon deposits.
Normal degradation losses for the DIPA solution used are negligible (5).
Therefore, no spent amine solution will be generated by the SCOT unit.
A8-11
-------
Appendix AS
SCOT Process
5. Process Reliability
The SCOT process is generally considered to be highly reliable and easy
to operate. Three sets of process reliability data have been collected (8).
These data indicate that:
At the Shell/Deer Park, Texas facility, the SCOT unit has 99.5
percent on-stream amime (excluding three weeks scheduled main-
tenance shutdown every three or four years).
At the Gulf/Port Arthur, Texas facility, typical on-stream time
for the SCOT unit ranges from 97.9 percent in 1977 and 1978 to
99.9 percent in 1980. Due to a heater tube failure in 1979 and
a process blower failure in early 1981, the on-stream time
during these two periods is considerably lower.
At the Texaco/Port Arthur, Texas facility, on-stream time for
the SCOT unit is 99.16. Operational problems causing down-
time include: 1) leaking amine exchanger, 2) high incinerator,
temperature, 3) blower failure, and 4) malfunction and repair
of regulators, control valves, and emergency shut-down instru-
ments.
6. Process Economics
The capital investment cost for a SCOT unit designed to treat the tail
gas from a 150 Mg/D Claus unit operating at 94 percent sulfur recovery was
quoted by Shell Development Company to be $4 million based on mid-1979 dollars
(5). Earlier 1973 data reported by Shell indicated the following capital
investment costs for SCOT units: $1.8 million for a 100 Mg/D unit, $3.2 mil-
lion for a 200 Mg/D unit, and $7.2 million for a 1000 Mg/D unit (7). All
these costs have been converted to mid-1980 dollars and a U.S. location cost
basis, and all SCOT unit sizes refer to the capacity of the preceding Claus
unit. It is also known that the capital investment for the add-on SCOT unit
corresponds to about 100 percent of the capital investment of the preceding
A8-12
-------
Appendix AS
SCOT Process
Claus unit (7). For the integrated SCOT unit (sharing amine regeneration and
sour water stripping facilities), the capital investment cost of the SCOT unit
equals about 75 percent of the Claus unit (7). In Figure A8-2, the capital
investment costs for the SCOT unit are presented as a function of the capacity
of the preceding Claus unit (7). Two cost curves for the SCOT unit are
plotted. The higher cost curve, based on a single mid-1979 data point pro-
vided by Shell is plotted assuming that costs increase with the 0.6 power of
the capacity ratio. The lower cost curve is based on the three data points
reported by Shell in 1973. As shown in Figure A8-2, there is considerable
difference between the two cost curves. For comparison purposes, a cost curve
for the Claus unit based on two 1980 data points is also plotted in Figure A8-
2. These plots show that the higher cost curve should be used in estimating
the capital cost for the SCOT unit because: 1) it is based on recent cost
data, and 2) it closely approximates the capital cost of Claus units based on
recent cost data.
The operating requirements for the SCOT process are presented in Table
A8—3. These operating requirements are based on recent data for a SCOT unit
associated with a 150 Mg/D Claus unit. Operating requirements for SCOT units
of other sizes can be estimated from these data. The hydrogen requirements
can be estimated by assuming twice the stoichiometric requirement for reacting
with the S02 present in the Claus tail gas.
A8-13
-------
n-gv
Capital Investment, 10f) Dollars (mid-1980)
T!
H-
TO
C
CO
I
to
n
•O
H-
H-
3
ro
en
rr
g
0)
n
o
en
n
o
H
C
3
H-
-------
TABLE A8-3. OPERATING REQUIREMENTS FOR THE SCOT PROCESS (5)
(Basis: 150 Mg/D Clans unit with 94% sulfur recovery)
Utilities
Steam 0.4 MPa, net import
Cooling water, 17K rise
Electric power*
Fuel gas , heat absorbed by
process (in-line burner)
Process Materials
SCOT catalyst, replacement cost
SCOT solvent, replacement cost
Hydrogen
5.4 Mg/hr
6.4 m'/min
50 kW
5.3 GJ/hr
ilO.OOO/year
i 1,000/year
dependent on Claus
tail gas composition
Manpower
Operators
1/6 man/shift
*Claus tail gas at 410 K, 0.13 MPa
Replacement cost converted to annual basis
"Reflects estimated mechanical losses
A8-15
-------
Appendix A8
SCOT Process
7. References
1. Knijpers, N.G.M.J. The Shell Clans Offgas Treating Process. Presented
at the Gas Sweetening and Sulfur Recovery Seminar, Amersterdam, The
Netherlands, November 9-13, 1981.
2. Goar, G. Claus Tail Gas Cleanup - Cost, Air Regulations Affect Process
Choice. The Oil and Gas Journal, August 18, 1975, pp. 109-112.
3. Goar, G. and T.O. Arrington. More Guides Offered for Handling Sour Gas.
The Oil and Gas Journal, pp 54-57, July 3, 1978.
4. Herfkens, A.M. One Company's Experience with TGT Hydrocarbon Processing,
November 1982, pp. 199-103.
5. Information provided to TRW by D.F. Vance, R.E. Bayles, and J.M. Duncan,
Shell Development Company. The Shell Claus Off-Gas Treating (SCOT)
Process. January 1980.
6. GPA H2S Removal Panel. Process Clean Dp Tail Gas. The Oil and Gas
Journal, August 28, 1978, pp 160-166.
7. Naber, J.E., J.A. Wesselingh, and W. Groenendaal. New Shell Process
Treats Claus Offgas. Chemical Engineering Progress, 69(12): 29-34.
December 1973.
8. Rucker, J.E. of American Petroleum Institute to R.M. Statnick of U.S.
Environmental Protection Agency. Reliability Data on Refinery Tail Gas
Cleanup Systems. July 30, 1981.
9. McNamee, G.P. and G.A. White. The Effect of Purchased Power and Steam
Turbine Drive on the Solvent Refined Coal Process. Report prepared by
the Ralph M. Parsons Company for the Electric Rower Research Instiute.
EPRI AF-741 (Supplemental Report). April 1978.
10. Information provided to TRW by G. Koutclas, J.F. Pritchard & Co.
8. Personal Communications
Schilk, J., Shell Development Company, Houston, Texas to C. Shih, TRW
Environmental Division. December 3, 1981.
A8-16
-------
APPENDIX A9
BEAVON PROCESS
1. Process Description
The Beavon Sulfur Removal Process (BSRP) for the clean-up of Claus plant
tail gas was developed jointly by The Ralph M. Parsons Company and Union Oil
Company of California and licensed by the latter company. A flowsheet for
the BSRP is shown in Figure A9-1.
The BSRP consists of two main steps: 1) catalytic hydrogenation and
hydrolysis of all sulfur species to H2S, and 2) conversion of all H S to
elemental sulfur by the Stretford process. In the first section, the Claus
plant tail gas is mixed with a hot reducing gas containing sufficient hydro-
gen and carbon monoxide to convert all sulfur compounds to H2S. The quan-
tity of reducing gas is regulated to provide a temperature of 620 to 640 K for
the mixture of Claus tail gas and reducing gas. This gas mixture then flows
to the catalytic reactor containing a cobalt-molybdenum catalyst in the sul-
fided state. In the catalytic reactor, the following reactions take place
(1):
S02 + 3H2 = H2S + 2H20 (1)
S, + 2H2 - 2H2S (2)
S. + 8H2 = 8H2S (3)
COS + H20 = C02 + HaS (4)
CS2 + 2HaO = CO, + 2HaS (5)
CH3SH + H2 = H,S + CH4 (6)
CO + H20 = C02 H, (7)
Conversion of COS and CS, to H2S is predominantly by hydrolysis, whereas
conversion of other sulfur species to H2S is by hydrogenation. Also, at
catalytic reactor conditions in which there is generally considerably more
water vapor than COJf the carbon monoxide in the reducing gas reacts almost
A9-1
-------
4
3
1
1
{REDUCING
Tl
4
LEGEND:
044 MPa
5 STEAM
t
('
9
1EACTOH J
IEFFLUE
1 COOLE
BOILER
FEED
WATER
1
^r~|
R I
6
I
r-o
1 FEED GAS
2 OFFGAS
3 AIR
4 FUEL GAS
5. HOT GAS CONTAINING H2S
6 COOL GAS CONTAINING H2S
7 SULFUR
8. SOUR WATER
9 SPENT CATALYST
Figure A9-1. The Beavon sulfur removal process
-------
Appendix A9
Beavon Process
completely with water to form CO, and hydrogen through the water gas shift
reaction (1). Temperature in the catalytic reactor is in the 640 to 670 K
range. There is always a temperature rise through the catalyst bed, mainly
due to the exothermic hydrogenation reaction of S0a.
The reactor effluent is cooled in a waste heat boiler to temperatures
substantially below 670 K, and then enters a gas cooler where it is further
cooled by direct contact with a circulating sodium carbonate-bicarbonate solu-
tion. This circulating solution is kept at a pH above 7 to absorb any S0a
which may leave the catalyst bed in cases of overloading. In most current
designs, the gas cooler gaseous effluent, near its dewpoint, flows directly to
a Stretford absorber. In some of the earlier designs, the gas cooler effluent
is cooled to about 310 K in a contact condenser before entering the Stretford
absorber. This additional cooling step leads to considerable condensation of
water.
The Stretford process uses an aqueous solution of sodium carbonate-sodium
bicarbonate containing the disodium salt of the 2:7 isomer of anthraquinone
disulfonic acid (ADA) and pentavalent vanadium. In the absorber, the H2S is
absorbed in the solution:
HaS + NaaCO, = NaHS + NaHCO, (8)
Some of the hydrogen sulfide undergoes side reactions and is converted to
sodium thiosulfate and sodium sulfate. The accumulation of these salts
reduces the solubility of vanadium and ADA in the solution, and decreases the
rate of Stretford solution reactivation. To maintain these salts at an
acceptable level of 20-30 wt %, a portion of the Stretford solution must be
purged (2). After absorption of H2S and excess moisture, the Stretford
solution is sent to the reaction tank. In the reaction tank, pentavalent
A9-3
-------
Appendix A9
Beavon Process
vanadium reacts almost immediately with the sodium bisulfide (NaHS) to produce
elemental sulfur:
NaHS + NaHCOj + 2NaV03 = S + NajVjO, + NaaC03 + H,0 (9)
The reaction tank effluent flows by gravity to a three-stage oxidizer. Air is
sparged into the solution to reactivate the Stretford solution in the presence
of ADA, which promotes the oxidation of tetravalent vanadium to pentavalent
vanadium:
ADA
Naivi°s + °-502 = 2NaV03 (10)
The air also floats the sulfur to the liquid surface as a heavy froth. This
froth overflows the oxidizer to a froth tank while the sulfur-free reactivated
Stretford solution flows through a cooling tower before falling into a balance
tank. The cooling tower is designed to evaporate sufficient water to maintain
the system in water balance, and to remove sufficient heat from the solution
to prevent excessive thiosulfate production. Solution temperature is normally
kept at 305 to 315 K.
Two modes of operation are available for froth processing. In one mode
of operation, a sulfur cake is produced by filtering, after which the cake is
reslurried with steam condensate. This slurry is pumped through a sulfur
melter and then into a steam-jacketed separator, where molten sulfur is
recovered; the dilute Stretford solution obtained is returned to the
system. The wash water for filter cake, filter belt and roller, and the
reslurry water all provide additional water load to the cooling tower duty.
A9-4
-------
Appendix A9
Beavon Process
In the direct antoclaving mode of operation the sulfur froth is pumped
through a sulfur melter into the autoclave, where the phases are separated and
the hot Stretford solution is returned to the balance tank. This mode of
operation has the advantages of simplicity of operation, lower capital invest-
ment, and reduction of water evaporation load on the cooling tower.
The BSBP has been in commercial operation since early 1973. By May 1980,
27 BSRP units in the U.S. and 3 BSRP units outside the U.S. were in operation
(3). Another 17 units in the U.S. and 5 units outside the U.S. are in desiqn
and construction stages. These BSRP units range in size from 0.7 to 54 Hg/day
of sulfur (corresponding capacity of the preceding Claus units range in size
from 10 to 1286 Mg/day of sulfur recovered). The largest operating BSRP unit
in the U.S., which recovers 32.5 Mg/day of sulfur, is located in the Exxon re-
finery at Baytown, Texas.
2. Process Applicability
Unlike the SCOT process, the BSRP can handle Claus tail gases containing
high concentrations of C0a. This is because the HaS removal efficiency of the
BSRP is relatively independent of C0a content, and because the BSRP does not
recycle H2S (along with large quantities of C02) back to the Claus plant for
conversion to elemental sulfur. The BSRP is also suitable for treating the
tail gas from smaller as well as larger Claus plants, as evidenced by the size
range of BSRP units currently in operation. These BSRP units are located
either in petroleum refineries or natural gas processing plants.
3. Process Performance
The treated gas stream from the BSRP unit normally contains less than 100
ppmv total sulfur compounds and less than 10 ppmv HaS (1,4). In Table A9-1,
the typical compositions of the BSRP offgas from the Stretford absorber are
A9-5
-------
TABLE A9-1. TYPICAL COMPOSITIONS OF GAS STREAMS IN BSRP UNITS
Beavon Demonstration Plant
Component
HaS, ppmv
S02 , ppmv
COS , ppmv
CSj, ppmv
CHjSH, ppmv
CO , ppmv
CH4 , ppmv
N2+A, %
coa, %
H,. %
Data Source :
Data Source :
Da ta Source :
Claus Plant
Tail Gas
30,000
40,000
1,400
2,500
<5
NR
NR
NR
NR
NR
Reference 5.
Reference 1.
Reference 6.
Beavon Reactor
Effluent
40,000
<20
50
20
<5
NR
NR
NR
NR
NR
Concentrations
temperature of
Concentrations
located in Los
Stretford Absorber
Offgas
5
<20
50
20
<5
NR
NR
NR
NR
NR
reported are on dry basis.
Stretford Absorber Offaas
Facility" Facility0
A
<1
<1
30
9
NR
500
200
89
7
3
Offgas is
B
<0.1-7
<5
7-23
1
NR
565
221
NR
5.7
5.6
Facility0
C
<1
<0.4
5
0.5
NR
250
NR
NR
NR
NR
essentially saturated with
Facil
D
<1
<1
15
0.
NR
670
NR
NR
NR
NR
vater at
ity°
1
the
the Stretford solution entering the absorber
reported for sulfur species
Angeles county.
are on dry
basis. Facil
ities B, C,
and D are
all
NR - Not Reported
-------
Appendix A9
Beavon Process
presented. Three other sets of sulfur emission data are presented in Table A9-
2. However, none of the API data provides information on the concentrations
of individual sulfur species in the BSRP offgas.
As indicated by the data presented in Tables A9-1 and A9-2, the BSRP
offgas often contains significantly less than 100 ppmv of total sulfur
compounds. The level of H2S in the offgas ranges from 0.1-7 ppmv, whereas
the levels of COS and CS2 range from 5-50 ppmv and from 0.1-20 ppmv,
respectively. Therefore, COS is the major contributor to the total sulfur
emissions.
COS is converted to H2S by hydrolysis in the Beavon reactor. The amount
of COS present is determined by:
1 1 (H2S)(C02) (11)
COS= "I A (HO)
P *
where: (COS), H2S), (C02) and H20) denote the concentrations or
partial pressures for these four compounds,
K is the equilibrium constant for the COS hydrolysis
reaction, and
A is the value for approach to equilibrium (A = 1.0 at
equilibrium).
At the operating temperature of the Beavon reactor (~670 K), K = 335. Thus,
P
under conditions encountered in the Beavon reactor, with H2S concentrations
of approximately 10000 to 15000 ppmv and C02/H20 <1.0, the equilibrium COS
concentration would be less than 45 ppmv. In practice, A is dependent on the
space velocity and can be significantly less than 1.0. Beavon reactors are
probably designed with A values in the 0.4 to 0.8 range (8). At high (C02)/
(H20) ratios such as 1.0, the Beavon reactor may be designed with lower space
velocities to yield an A that approaches 1.0, so that the COS levels in the
A9-7
-------
Appendix A9
Beavon Process
reactor effluent would still be less than 50 ppmv. Neither COS nor CS2 is re-
moved in the Stretford absorber.
TABLE A9-2 SULFUR CONCENTRATION IN BSRP OFFGAS
FROM STRETFORD ABSORBER (7)
Plant Location Sulfur Concentration in Off Gas
Union/Wilmington, California 7-33 ppmv, average 13 ppmv
Exxon/Baytown, Texas Unit 1 2-5 ppmv
Exxon/Baytown, Texas Unit 2 2-9 ppmv
4. Secondary Waste Generation
Five secondary waste streams may be generated from the BSRP unit. These
are: 1) sour water stream from the condensation of water vapor present in the
Beavon reactor effluent; 2) evaporation and drift from the coolinq tower used
to maintain water balance; 3) Stretford solution purge stream due to thiosul-
fate and sulfate formation; 4) Stretford oxidizer vent; and 5) spent catalyst
from the Beavon reactor.
The sour water stream is only present in some of the earlier plants
where the desuperheater effluent is passed through a contact condenser before
entering the Stretford absorber. The flow rate of this sour water stream can
be estimated from knowledge of the amount of water vapor present in the Claus
tail gas, the operating pressure and temperature of the contact condenser
(typically at atmospheric pressure and 310 K), and by assuming that the gas
stream leaving the contact condenser is saturated with water vaoor. The sour
water stream contains small amounts of dissolved H2S and should be sent to a
sour water stripper for treatment.
A9-8
-------
Appendix A9
Beavon Process
The cooling tower evaporation and drift is only present in designs that
employ cooling towers to maintain the water balance for the system. In these
designs the evaporation rate can be estimated from the water balance. Drift
losses from the cooling tower should average 0.005 percent of the recircnlat-
ing rate. In contrast to evaporation losses, the drift droplets contain the
same chemicals as the recirculating solution. Thus, the drift droplets are
expected to contain ADA, vanadates, sodium carbonate, sodium thiosulfate, and
sodium sulfate. The concentrations of these chemicals in the cooling tower
drift are approximately the same as their concentrations in the Stretford
solution purge stream.
The Stretford oxidizer vent from the BSRP unit is not expected to cause
any problem. This is because of the absence of ammonia and mercaptans in the
inlet gas stream to the Stretford absorber. The Stretford oxidizer vent will
therefore contain only nitrogen, oxygen, and carbon dioxide and will be satu-
rated with water.
As discussed previously, a portion of the Stretford solution must be
purged because of thiosulfate accumulation. The composition and flow rate of
this purge stream are presented in Table A9-3. Several methods have been
proposed for the treatment of the spent Stretford solution. The Ralph M.
Parsons Company offers a process for treatment of this stream with hot sul-
faric acid, which converts most of the thiosulfate to sulfate (1). As sulfate
builds up and at relatively high thiosulfate levels, the sulfate is crystal-
lized from another slipstream. With this process, essentially only sodium
sulfate crystals are discarded.
In another process, the spent Stretford solution is acidified with sul-
furic acid to decompose the thiosulfate to sulfur and sulfur dioxide and
limed to remove added sulfate and restore the pH of the solution (2). The
thiosulfate decomposition reaction takes place in steps according to the
following equations:
A9-9
-------
Appendix A9
Beavon Process
• Protonation of thiosulfate ions
SaO,= + H+ = HSaO,~ (12)
„„ n - „+ „ „ n /,,*
uO f\ i TJ __ ¥¥ O fl I i ^ 1
Jlij \j T n — n j v/. V A«J /
• Internal redox reaction
H S 0 = H SO + S (14)
213 21
• Evolution of S0a
H SO = H 0 + SO (15)
23 2 2
The acidified solution is stripped of SOj using steam, with the SOZ recycled
to the Claus plant. The sulfur floats to the top and is recovered by skim-
ming or filtration. The clear Stretford solution is then reacted with lime to
raise the pH and precipitate sulfate/sulfite ions as CaS04/CaS03. In this
step, the sulfate removed includes both sulfate from byproduct formation in
the Stretford process as well as sulfate from sulfuric acid addition in the
acidification step. The slurry is settled in a thickener, and the solution is
recycled to the Stretford section. The calcium salts are disposed of after
washing and filtration.
TABLE A9-3. CHARACTERISTICS AND RATES OF STRETFORD SOLUTION PURGE
Concentration of a
Purge Stream, Rate of Purge,
Component wt % g/kg sulfur
ADA
Vanadium
Na CO,
2 3
Na*SO
HO
Total
0.19
0.38
1.45
8.95
4.02
85.01
100.00
0.58
1.14
4.30
26.64
11.97
253.09
297.72
Sulfur processed in the BSRP unit
Data Source: Personal communication with M.H. Griebe, November 19,1980.
A9-10
-------
Appendix A9
Beavon Process
The initial catalyst requirement for the Beavon reactor is approximately
28 wt % of the initial Glaus catalyst requirement (9). This is equivalent to
850 kg catalyst per kg-mole/hr of HaS converted from S02, COS, and CS2. For a
Claus plant processing 150 Mg/day of sulfur, the initial catalyst requirements
for the BSRP unit is 12.6 Mg. With a catalyst service life of two to three
years (1), this amount of spent catalyst will require disposal every two to
three years. The spent catalyst will be cobalt-molybdate in alumina with
traces of sulfates, sulfides, elemental sulfur, and carbon deposits.
5. Process Reliability
Operating experience has indicated a number of problems with the BSRP.
The most common problem reported appears to be absorber plugging and plugging
in the sulfur froth lines (1). Absorber plugging by sulfur normally occurs
in the bottom 2 to 3 meters of packing after 6 to 12 months of operation (1).
Current evidence suggests that freshly formed sulfur causes plugging.
Three sets of process reliability data have been collected (7). These
data indicate that:
• At the Union/Wilmington, California facility, on-stream time
including 19 day scheduled maintenance every 8 month cycle, is
92 percent. On-stream time is almost 100 percent if scheduled
maintenance down-time is excluded. Minor upsets were sometimes
caused by abrupt change in Claus plant feed or loss of
Stretford solution circulation pump, which required the combus-
tor on the BSRP unit to be relit. These upsets have been
infrequent and never exceed more than one day of down—time
during the eight month cycle.
• At the Exxon/Baytown, Texas facility, on-stream time was 89.4
percent in 1979 and 89.6 percent in 1980 for BSRP unit No. 1.
In 1979, the unit was down for 21 days due to planned mainte-
nance and down for 19 days due to operational problems. In
1980, the unit was down for 3 days due to planned maintenance
and down for 35 days due to operational problems. The opera-
tional problems causing shutdown included: 1) solids in auto-
clave, pit and pumps, 2) overloaded Stretford solution, 3) loss
A9-11
-------
Appendix A9
Beavon Process
of combustor, 4) plugged product piping - autoclave to pit,
5) high pressure drop across the reactor, 6) loss of Stretford
solution overflow, and 7) sticking feed diverter valve.
For BSRP unit No. 2 in the Exxon/Baytown, Texas facility, on-
stream time has averaged 98.5 percent since startup in 1977.
Operational problems causing shutdown included: 1) hole in
oxidation tank, 2) corrosion in Stretford circulation valve,
and 3) Stretford solution overload.
Both BSRP units in the Exxon/Baytown facility started up in 1977. Capacity of
the preceding Claus plant is 300 Mg/day for unit No. 1 and 800 Mg/day for unit
No. 2. The reasons for the differences in on-stream time between the two
BSRP units are not known.
Modifications have been made in the BSRP design to alleviate two of the
problems that cause shutdown (1). First, the cooling tower incorporated in
later BSRP designs eliminates fouling problems in shell and tube heat
exchangers. Second, the combination of a venturi scrubber followed by a short
conventionally packed absorber in the design provides adequate H2S removal
while limiting the plugging to the venturi. The venturi can then be conven-
iently cleaned if a plugging problem arises. Stretford designs with two
venturi scrubbers in parallel (one spare) followed by a short absorber can
conceivably eliminate plant shutdown for cleanout of absorption equipment.
6. Process Economics
The capital investment cost for a BSRP unit processing 31.45 Mg/day of
sulfur in a coal liquefaction plant was quoted by The Ralph M. Parsons Company
to be $12.1 million based on mid-1980 dollars. (Personal communication with
M. H. Griebe, November 19, 1980). EPAs 1975 data for petroleum refineries
indicated the following capital investment costs for BSRP units (adjusted to
mid-1980 dollars): $1.04 million for a 0.25 Mg/day unit, $1.28 million for a
0.51 Mg/day unit, and $4.05 million for a 5.08 Mg/day unit (6). The Gas
Processing Handbook reported a capital investment cost of $2.74 million for a
A9-12
-------
Appendix A9
Beavon Process
5.08 Mg/day BSRP unit (4). The capital investment cost for a BSRP unit is
approximately equal to the capital investment cost of the preceding Claus unit
(10). In Figure A9-2, the capital investment costs for the BSRP unit are
presented as a function of the capacity of the BSRP unit. At capacities above
2 Mg/day of sulfur, the costs increase with the 0.6 power of the capacity
ratio.
The operating requirements for the BSRP process are presented in Table
A9-4. These requirements are based on recent estimates provided by Parsons
for a 31.45 Mg/day BSRP unit in a coal liquefaction plant. Operating
requirements for BSRP units of other sizes can be estimated from these data.
The hydrogen requirement can be met by the substoichiometric combustion of
fuel gas, in most cases. If a hydrogen rich stream is available as reducing
gas, the hydrogen requirement can be estimated by assuming twice the stoichio—
metric requirements for reacting with the S0a present in the Claus tail gas.
A9-13
-------
I
M
-P~
c
0)
W
cn
OJ
C
p.
n)
U
30
2
10
7
0.1
O Reference 9
^ Reference 6
O Reference 4
I I I I
D
I I 1 II _L
0.6 slope
10
Capacity of BSRP Unit, Mg/Day of Sulfur
100
Figure A9-2. Capital investment cost for BSRP units (mid-1980 dollars)
-------
Appendix A9
Beavon Process
TABLE A9-4. OPERATING REQUIREMENTS FOR THE BSRP
(Basis: 31.45 Mg/day BSRP unit)
Utilities
Steam 0.44 MPa, net export
Steam 1.1 MPa, used
Process Water
Cooling Water
Electric Power
Fuel Gas
4680 kg/hr
610 kg/hr
0.05 m /min
23 mVmin
3,350 kw
34 GJ/hr
Process Materials
ADA - initial charge
annual makeup
Vanadium - initial charge
annual makeup
Na CO - initial charge
annual makeup
Hydrogen
4880 kg
9120 kg
9740 kg
9120 kg
36,500 kg
223,700 kg
Dependent on Claus
tail gas composition
Data Source: Personal communication with M.H. Griebe, November 19,1980.
7. References
1. Kouzel, B., R.H. Fuller, E.J. Jirus, and B.B. Woertz. Treat Low Sulfur
Gases with Beavon Sulfur Removal Processes and the Improved Stretford
Process. Presented at the 27th Annual Gas Conditioning Conference, the
University of Oklahoma, Norman, Oklahoma, March 7—8, 1977.
2. Yan, T.Y. and W.F. Espenscheid. Removal of Thiosulfate/Sulfate from
Spent Stretford Solution. Environmental Science and Technology, 14(6):
732-735, June 1980.
3. Technology Sales Department, Union Science and Technology Division, Union
Oil Company of California. Beavon Sulfur Removal Process (BSRP)
Commercial Installations. May 16, 1980.
A9-15
-------
Appendix A9
Beavon Process
4. Gas Processing Handbook. Hydrocarbon Processing, 58(4): 132, April
1979.
5. GPA H2S Removal Panel. Processes Clean Dp Tail Gas. The Oil and Gas
Journal, August 28, 1978, pp 160-166.
6. U.S. Environmental Protection Agency. Standards Support and
Environmental Impact Statement. Volume I: Proposed Standards of
Performance for Petroleum Refinery Sulfur Recovery Plants.
EPA-450/2-76-016a, September 1976.
7. Rucker, I.E. of American Petroleum Institute to R.M. Statnick of U.S.
Environmental Protection Agency. Reliability Data on Refinery Tail Gas
Cleanup Systems. July 30, 1981.
8. Beavon O.K. and R.P. Vaell. Prevention of Air Pollution by Sulfur
Plants. Paper presented at the Eighth Annual Technical Meeting, Southern
California Section, American Institute of Chemical Engineers, April 20,
1971.
9. McNamee, G.P. and G.A. White. The Effect of Purchased Power and Steam
Turbine Drives on the Solvent Refined Coal Process. Report prepared by
the Ralph M. Parsons Company for the Electric Power Research Institute,
EPRI AF-741 (Supplemental Report), April 1978.
10. Laengrich, A.R. and W.L. Cameron. Tail Gas Cleanup Addition May Solve
Sulfur-Plant Compliance Problem. The Oil and Gas Journal, May 27, 1978,
pp 159-162.
8. Personal Communications
Griebe, M.H., The Ralph M. Parsons Company to TRW Environmental Division.
November 19, 1980.
A9-16
-------
APPENDIX A10
WELLMAN-LORD PROCESS
1. Process Description
The Wellman-Lord S0a recovery process was developed by Davy McKee
Engineers and Constructors (formerly Davy Powergas, Inc.) in the 1960s to
produce a concentrated sulfur dioxide gas from lean sulfur dioxide offgas
streams. Davy McKee is also the licensor for the process.
A flowsheet for the Wellman-Lord S02 recovery process, as applied to the
treatment of Claus tail gas, is presented in Figure AlO-1. The process con-
sists of two principal areas: incineration-absorption and solution regenera-
tion (1). In the incineration-absorption area, tail gas from the Claus sulfur
recovery unit is first incinerated to convert essentially all sulfur compounds
to sulfur dioxide. The hot combustion gas from the incinerator is cooled to
590 K in a waste heat boiler. The gas is then quenched to its adiabaic satu-
ration temperature in a packed column by a circulating acidic water solution.
Small amounts of S02 are removed in this quench step. The saturated gas
leaves the quench column at 345 K and is further cooled to 320 K in a shell
and tube exchanger. Water condensed from the gas stream is returned to the
quench column. The gas leaving the cooler is sent through a cyclonic type
entrainment separator prior to entering the absorber.
The absorber consists of two packed stages for S02 removal, one inter-
stage collector tray and a mesh type mist eliminator. The absorbing solution
for each stage is individually recirculated. In the absorber, the principal
reaction is between sulfur dioxide in the incinerated Claus tail gas and the
sodium sulfite in the lean absorber solution, forming sodium bisulfate:
SO, + NaaSO, + H,0 = 2NaHSO, (1)
The treated gas leaves the absorber at 326 K, is reheated to 340 K, and then
discharged to the atmosphere through a stack. Some oxidation of the sodium
A10-1
-------
WASTE HEAT QUENCH & COOLING SO2
INCINERATOR BOILER SECTION ABSORBER
EVAPORATOR
DISSOLVING
TANK
CLEANED
PRODUCT SO2
RECYCLED TO
CLAUS PLANT
Figure A10-1. The Wellman-Lord S02 recovery process
-------
Appendix A10
Wellman-Lord Process
sulfite to sodium sulfate takes place in the absorber due to the oxygen pre-
sent in the incinerated Claus tail gas:
Na2S03 + 0.50a= NaaS04 (2)
Additional sodium sulfate may be formed in the absorber if sulfur trioxide is
present in the incinerated Claus tail gas:
2NaaS03 + S03 + HaO = Na2S04 + 2NaHSO, (3)
The sodium sulfate formed by these reactions is removed from the absorber
solution by purging a small amount of absorber feed solution prior to soda ash
addition.
In the regeneration area, the rich sodium bisulfite solution is first sent to
a double-effect evaporator. In the evaporator, sulfur dioxide is released and
sodium sulfite crystals precipitate according to the following reaction:
2NaHSO, = NaaS03 + S02 + H^O (4)
A disproportionation reaction also takes place in the evaporator at high
temperatures, forming sodium sulfate and sodium thiosulfate:
2Na SO, + 2NaHSO, = 2NaaS() + Na,S 0 + HO (5)
13 9 24 ZZ3 2
The sodium thiosulfate formed is removed from the absorbing solution in the
absorber feed solution purge.
The slurry product from the double-effect evaporator flows to the dis-
solving tank, where it is mixed with cooled stripped condensate from the
A10-3
-------
Appendix A10
Wellman-Lord Process
condenser system to redissolve the Na>SOJ crystals. This condensate is
derived from steam stripping of the condensates from the second effect
evaporator and the heater for the second effect evaporator. The S02 vapor
removed from the sour condensate is compressed and recirculated to the Claus
sulfur recovery unit.
A small portion of the regenerated solution from the dissolving tank is
purged to remove the sodium sulfate/thiosulfate byproducts formed. The
remainder of the regenerated solution is pumped to the absorber feed tank
where soda ash is added to replace the sodium lost in the byproduct stream
purge:
Na2C03 + 2NaHSOs = 2Na2SO, + H20 + C02 (6)
The Wellman-Lord S02 recovery process has been in commercial operation
since 1970. By the end of 1981, 11 Wellman-Lord units in the U.S. and 18
units outside the U.S. (all in Japan) were in operation (1). Of the 29 oper-
ating units, eight have been designed to treat Claus plant tail gas. In the
U.S., Wellman-Lord units for Claus plant tail gas treatment are located at
Standard Oil Company of California refineries in El Segundo, California (2
units) and Richmond, California (2 units), and at the Chevron U.S.A. (a sub-
sidiary of Standard Oil Company of California) refinery in Perth Amboy, New
Jersey (1 unit). These Wellman-Lord units range in size from 150 to 400
Mg/day of sulfur recovered in the preceding Claus unit (2,3).
2. Process Applicability
The Wellman-Lord S02 recovery process has been used to treat offgas
streams from coal- and oil-fired boilers, sulfuric acid plants, and Claus sul-
fur plants. Its best application is treating gases which contain 1500 to
A.10-4
-------
Appendix A10
Wellman-Lord Process
30,000 ppmv SO, (4). For Clans tail gas treatment, the Wellman-Lord process
is more expensive than other cleanup processes and is generally attractive
only under the following conditions (5):
• Treatment of tail gas from a single-train large Claus plant
(exceeding 200 Mg/day sulfur) or multiple train Claus units
(100-150 Mg/day sulfur) at a single location. This is because
several Wellman-Lord absorbers are often used with only one
central regeneration system.
• For Claus tail gas containing appreciable amounts of CO, (50-
70 vol % or greater), the Wellman-Lord process may be a good
choice because it does not create a CO, recycle problem.
The Wellman-Lord S0a recovery process may be suitable for Claus tail gas
treatment in coal gasification/liquefaction plants because of the size of
these facilities and the high CO, content of Claus tail gas streams.
3. Process Performance
The Wellman-Lord S02 recovery process is typically designed to achieve
effluent levels of less than 150 ppmv to 250 ppmv SO,. Effluent levels of
less than 100 ppmv SO, in the stack gas have been consistently achieved in
commercial installations (6). In Table A10-1, the compositions of the stack
effluent from the Wellman-Lord unit in the Standard Oil of California El
Segundo refinery are presented. These data indicate total sulfur emissions
(mostly as SO,) in the 12-59 ppmv range. This particular Wellman-Lord unit
was designed for 250 ppm S0a (2). However, the data were collected during
periods when the Wellman-Lord unit was operating at half or less of its design
capacity due to low refinery throughputs. Emissions of S0a when the Wellman-
Lord unit is operating at design capacity may be considerably higher.
A10-5
-------
Appendix A10
Wellman-Lord Process
TABLE AlO-1. COMPOSITIONS OF STACK EFFLUENT FROM WELLMAN LORD S0a
RECOVERY UNIT IN THE STANDARD OIL COMPANY OF
CALIFORNIA EL SEGUNDO REFINERY (2)
Component
C02, Vol. %
02, Vol. %
CO, ppmv
S02 , ppmv
COS , ppmv
CSa, ppmv
H2S, ppmv
NO , ppmv
Tofal sulfur
compounds, ppmv
THC, ppmv
Water vapor. Vol. %
Flow rate, dscm/min
Test
A!
5.3-7.2
0.8-3.0
39-100
5.9-38
0.9-3.2
1 .1-3 .4
<0.1
9.0-21.0
11 .7-43 .8
4.6-7.5
10.6-13.0
135.4-209.7
Test
A,
6.6
1.3
3000
10-15
0.3
1.5
NR
21 .7-25
11.8-16.8
3.0
NR
NR
Test
A,
17.8-21.6
NR
670-3600
31-47
<1-15
1-13
<0.1-1.7
9.1-14.5
41-59
17-41
10.0-14.0
127.4-229.4
NR - not reported
Except for water vapor, all compositions as Vol. % dry or ppmv dry.
THC is total hydrocarbons as methane.
Davy Powergas guaranteed stack effluent levels of less than 250 ppmv
total sulfur as SO, on the first U.S. Claus plant application of the Wellman-
Lord process (2). More recently, Davy McKee has designed a Wellman-Lord S0a
absorption system to treat the Claus tail gas in a coal liquefaction plant, so
as to reduce the S0± concentration in the treated stack gas to a maximum of
150 ppmv (1). In the design of the Breckinridge project (based on the H-coal
process) the Claus plant tail gases are incinerated in the steam generation
plant boiler fire boxes and then treated in the Wellman-Lord unit (7). These
steam plant boilers are fired with coal middlings as well as the Claus plant
tail gases. The Davy McKee design for this specific Wellman-Lord unit allows
for a maximum of 200 ppmv S0a in the treated flue gas.
A10-6
-------
Appendix A10
WeiIman—Lord Process
Based on available data, the effluent from the Wellman-Lord process is
normally designed to contain less than 150-200 ppmv SOZ, depending on specific
requirements. Actual S02 emission levels can be expected to be significantly
lower than design SO^ emission levels and often less than 100 ppmv. Because
the Claus tail gas is incinerated prior to absorption in the Wellman-Lord
unit, however, the S0a emission levels are based on the incinerated offgas.
In the Davy McEee design of the Wellman-Lord unit to treat the Claus tail gas
in a coal liquefaction plant, the volume of the stack gas discharged is
approximately 1.48 times the volume of the Claus tail gas treated (1). Thus,
a 100 ppmv S02 emission level in the Wellman-Lord stack gas is equivalent to
sulfur emission levels of approximately 150 ppmv in the offgas from the Beavon
process or the SCOT process (prior to incineration).
4. Secondary Waste Generation
Two secondary waste streams are generated from the Wellman-Lord S02
recovery unit. The first is the acidic wastewater purge from the quench
column. This acidic wastewater results from condensation of water vapor
present in the effluent from the tail gas incinerator. The flow rate of this
acidic wastewater stream can be estimated from knowledge of the composition
and flow rate of the Claus tail gas, the composition and flow rate of the fuel
gas used for incineration, and by assuming that the gas stream leaving the
heat exchanger after the quench column is saturated with water vapor at 320 K.
This acidic wastewater purge has a pH value between 1 and 2 (1).
The second waste stream generated is the thiosulfate/sulfate byproduct
purge. For a Wellman-Lord unit treating the tail gas from a 1046 Mg/day Claus
A10-7
-------
Appendix A10
Wellman-Lord Process
sulfur plant, the generation rate of this purge stream is 463 kg/hr. The
composition of the purge stream is given by Davy McKee as follows: NaaS03,
41.5 wt %;, Na^O,, 18.1 wt %; Na.,S04, 10.0 wt %; Na2Sa03, 1.1 wt %; H20,
29.2 wt % (1).
5. Process Reliability
One set of process reliability data has been collected by API for EPA from
the Standard Oil Company of California Richmond refinery (8). The data indi-
cate that:
• The Claus and Wellman-Lord units are normally shut down about
10 percent of the year for routine maintenance. The shutdown
period can range from 8 to 20 percent of the year.
• Two major operating problems are: 1) sulfur particles plug-
ging the Karbate cooler, and 2) sodium sulfite plugging of
the absorber column. These problems tend to limit the run
length of the Wellman-Lord unit. There is also a problem
with corrosion of the quench column by sulfuric acid.
Davy McKee indicated that they do not receive regular reports on the reliabi-
lity of Wellman-Lord units in service but are informed that on-line avail-
ability of well over 95 percent is expected and achieved on a yearly basis
(1). This appears to be in some conflict with the Richmond refinery's experi-
ence. On the other hand, the Standard Oil Company of California's first
Wellman-Lord unit has been in service since 1972 and the company has added
four more units since that time. This is an indication that the Standard Oil
Company of California is generally satisfied with operating efficiency and
reliability of the Wellman-Lord process.
6. Process Economics
The capital investment cost for a Wellman-Lord S02 recovery system
designed to treat the tail gas from a 1050 Mg/day of sulfur Claus plant
A10-8
-------
Appendix A10
Wellman-Lord Process
operating at 95 percent sulfur recovery was quoted by Davy McKee to be $35
million based on 1980 dollars (1). The system consists of two incinerator-
absorber trains and a single regeneration unit. Each of the two incinerator-
absorber trains represents approximately 37 percent of the capital investment,
while the single regeneration unit represents the remaining 26 percent of the
capital investment. On this basis, the capital investment for each
incinerator-absorber train is $12.95 million and $9.1 million for the single
regeneration unit.
The Oil and Gas Journal extrapolated 1972 cost data and reported the
following capital investment costs for Wellman-Lord S0a recovery systems:
$3.07 million for a 200 Hg/D of sulfur unit, and $6.8 million for a 1520 Mg/D
of sulfur unit (5). All these costs have been converted to mid-1980 dollars,
and all Wellman-Lord unit sizes refer to the capacity of the preceding Claus
plant. Earlier 1975 data reported by EPA indicated a capital investment cost
of $2.56 million for a 100 Mg/D of sulfur Wellman-Lord unit (adjusted to mid-
1980 dollars) (2). It is also reported that a Wellman-Lord unit for tail gas
treatment would cost about 130-250 percent of the cost of the preceding Claus
plant (5).
In Figure A10-2, the capital investment costs for the Wellman-Lord S0a
recovery system are presented as a function of the capacity of the preceding
Claus plant. The cost curves for the incinerator—absorber train and the
regneration unit are separately plotted. This is because several Wellman-Lord
incinerator-absorber trains are often associated with only one central regen-
eration unit. These cost curves, based on the single data point provided by
Davy McKee, are plotted assuming that costs increase with the 0.6 power of
capacity ratio. As shown in Figure A10-2, there is considerable difference
between the Davy McKee cost data and the earlier EPA or Oil and Gas Journal
cost data. For a 100 Mg/D of sulfur Clans plant, for example, the Davy McKee
data show a capital investment cost of $6.9 million for the Wellman-Lord unit
(assuming single incinerator-absorber train) versus the $2.56 million indi-
cated by EPA data. For comparison purposes, a cost curve for the Claus unit
A10-9
-------
o
1
o
1—I
•co-
4-1
I
OJ
c
•H
&,
O
30
2
10
7
10
^Reference 1
OReference 5
OReference 2
riPersonal communication with G. Koutelas x,"
^ -.&• -
0.6 slope
I I I I I I
J I
I l I I I I
100
7 1000
Capacity of Preceding Glaus Plant, Mg/Day Sulfur
Figure A10-2. Capital investment costs (mid-1980 dollars) for Wellman-Lord
SC>2 recovery systems
-------
Appendix A10
Wellman-Lord Process
based on two 1980 data points is also plotted in Figure A10-2. These plots
show that the Davy HcKee cost curves should be used in estimating the capital
cost for the Wellman-Lord unit because: 1) they are based on recent cost data
furnished by the process licensor, and 2) they are consistent with the capital
costs of a Claus unit based on recent cost data. For two Clans units each
with a capacity of 523 Mg/day, the total capital investment cost is $23.7
million. The capital investment cost for the corresponding Wellman-Lord
system, with two incinerator-absorber trains and a single regeneration unit,
was quoted by Davy McKee at $35 million or 1.48 times the capital investment
cost of the Claus unit. Capital costs for Wellman-Lord units based on Oil
and Gas Journal or earlier EPA data are considerably lower than the capital
costs of the preceding Claus units, and therefore appear to be unrealistic.
The operating requirements for the Wellman-Lord S02 recovery process are
presented in Table A10-2. These operating requirements are based on estimates
provided by Davy McKee for a Wellman-Lord unit treating the tail gas from a
1050 Mg/D of sulfur Claus plant in a coal liquefaction facility (1). Opera-
ting requirements for Wellman-Lord units of other sizes can be estimated from
these data.
TABLE A10-2. OPERATING REQUIREMENTS FOR THE WELLMAN-LORD
S02 RECOVERY PROCESS (1)
(Basis: 1047 Mg/D Claus unit with 95% sulfur
recovery8, 980 K incineration temperature)
Steam, 3.2 MPa net export 39 Mg/hr
Cooling water, 14 K rise 28.9 ms/min
Electric power 1,280 kW
Fuel gas for incineration 229 GJ/hr
Soda ash, 100% Na2C03 anhydrous 246 kg/hr
Operating labor 1 man/shift
-T--1-T r- I _, IL— - - . I _ _L" J Z1J --!._!__ .__f . . i ' _.I
Claus tail gas: 2,830 mVmin. 410 K, 97 kPa
Composition of Claus tail gas: 48.00% Nz, 30.00% C0a, 0.30% SO,, 0.60% H,S,
20.85% H20, 0.03% COS, 0.20% CO, and 0.02% S.
A10-11
-------
Appendix A10
Wellman-Lord Process
7. References
1. Odgen, R.J. of Davy McKee Engineers and Constructors. Lakeland, Florida
to J.F. Geick of Black and Veatch Consulting Engineers, Kansas City,
Missouri. Preliminary Technical Information, Wellman-Lord SOa Recovery
System, Claus Plant Tail Gas Treatment. October 1980.
2. U.S. Environmental Protection Agency. Standard Support and Environmental
Impact Statement. Volume I: Proposed Standards of Performance for
Petroleum Refinery Sulfur Recovery Plants. EPA-450/2-76-016a, September
1976.
3. HPI Construction Boxscore. Hydrocarbon Processing, Section 2, October
1980.
4. Osborne, W.J. and C.B. Earl. Recent Experience of the Wellman-Lord
Sulfur Dioxide Recovery Process In Sulfur Removal and Recovery from
Industrial Processes edited by J.B. Pfeiffer. Advances in Chemistry
Series 139, American Chemical Society, Washington, D.C. 1975.
5. GPA H2S Removal Panel. Processes Clean Up Tail Gas. The Oil and Gas
Journal, August 28, 1978, pp 160-166.
6. Gas Processing Handbook, Hydrocarbon Processing, 58(4): 132, April 1979.
7. Prevention of Significant Deterioration Analysis and Application for
Permit to Construct the Breckinridge Project. Report prepared by Bechtel
Group Inc., for Ashland Synthetic Fuels Inc. and Arco Energy Co. Inc.,
June 1981.
8. Rncker, J.E. of American Petroleum Institute to R.M. Statnick of U.S.
Environmental Protection Agency Reliability Data on Refinery Tail Gas
Cleanup Systems. July 30, 1981.
8. Personal Communications
Koutelas, G., J.F. Pritchard & Co. to TRW Environmental Division.
September, 1980.
A10-12
-------
APPENDIX All
CYCLONES
1. Process Description
A cyclone is an inertial separator used to remove medium-sized entrained
particulate matter (15 to 40 |im) from gases. As shown in Figure All-1, a
cyclone consists of a cylindrical shell fitted with: 1) a tangential inlet
through which dusty gas enters, 2) an axial exit pipe for discharging the
cleaned gas, and 3) a conical base with dust discharge facility.
Cyclones in common use can be classified into four types:
• the common cyclone, tangential inlet with axial dust discharge,
• tangential inlet with peripheral dust discharge,
• axial inlet through swirl vanes, with axial dust discharge, and
• axial inlet through swirl vanes, with peripheral dust discharge.
Frequently cyclones are also classified as conventional and "high efficiency."
High efficiency cyclones merely have a smaller body diameter to achieve
greater separating forces, but there is no sharp dividing line between the two
groups. High efficiency cyclones are generally considered to be those with
body diameters up to about 23 cm.
The flow pattern is complex in even the simplest cyclone. As depicted in
Figure All-2, the dust—laden gas enters tangentially and creates a descending
helical current in the body and cone. The entrained particulate matter,
because of inertial forces, tend to move toward the outside wall, from which
they are led to a receiver. Near the apex of the cone, the spinning gas flows
radially inward and then forms an ascending helical current, ultimately being
discharged from the top of the cyclone (1).
All-1
-------
CLEANED SAS
DISCHARGE
TANGENTIAL
IHLET FOB
CONTAM'HATEO GAS
CYLINDRICAL
BODY
D'JST
HOPPER
* OUST DISCHARGE
Figure All-1. Single efficiency cyclone
ZONE OF INLET
INTERFERENCE
OUTER
VORTEX
P&RllCLES
THRO&N
TO WALL OF
COLLECTOR 8r
CEHTRIFUGAL
FORCES
TOP VIEW
OUTER
VORTEX
CLEAN GAS
RISE THROUGH
VORTEX CREATED
BY CYCLONIC
ACTION
CONE
' DUST OUTLET
SIDE VIEW
Figure All-2. Flow patterns in a cyclone
All-2
-------
Appendix All
Cyclones
2. Process Applicability
Cyclones can be successfully used in applications where:
• dust must be collected in dry form,
• temperatures are high,
• dust concentrations are high,
• gas is under high pressure, and
• dust or gas becomes corrosive when wet.
Cyclones should not be specified for conditions where:
• dust will adhere to cyclone and dust hopper walls because of
its surface properties or because temperatures drop below the
gas dewpoint and/or
• dust is very fine (below 1 to 5 microns depending on the dust
density and the gas flowrate).
Cyclones are constantly used as the lowest cost collector in applications
where only a small portion of the dust to be collected is below 5 microns. In
groups, or as multicyclones, these devices are often used as a first-stage
collector in large modern plants. In case the dust concentration proves too
large for cloth filters, electrostatic precipitators, or even wet collectors,
cyclones are provided for precollection. It may also be necessary to protect
expensive wet collectors from excessive abrasion, and in such instances,
cyclones offer low cost protection.
A special area of application for cyclones is the cleaning of very hot
gases that have high dust loads. The units are built from heat—resistant
materials and are enclosed in a refractory-lined vessel. For cyclones of
larger diameter, it is possible to line internal surfaces with layers of
insulating, abrasion-resistant refractory.
All-3
-------
Appendix All
Cyclones
3. Process Performance
In general, cyclone performance increases with an increase in the
following: 1) density of the particulate matter, 2) inlet velocity into the
cyclone, 3) cyclone body length, 4) number of gas revolutions, 5) ratio of
cyclone body diameter to cyclone outlet diameter, 6) particle diameter,
7) amount of dust entrained in carrier gas, and 8) smoothness of inner cyclone
wall (1). The performance of the cyclone decreases with increases in carrier
gas viscosity, cyclone diameter, gas outlet diameter, gas inlet duct width,
inlet area, or gas density (1).
Pressure drop across a cyclone usually ranges between 1 and 4 inlet
velocity heads, corresponding to 25 to 78 mm water gauge. Pressure drop
increases with the square of the inlet velocity. Removal efficiency also
increases, but not so rapidly as pressure drop. All devices intended to
minimize pressure drop result in decreased dust removal efficiency, except
those which minimize overall pressure drop by recovering the energy in the
vortex flow leaving the gas outlet.
One of the primary parameters affecting the performance of a cyclone is
particle size. Figure All-3 illustrates comparative collection efficiency for
2 axial-entry cyclones with diameters of 15.2 and 30.5 cm, respectively, as a
function of percent of dust under 10 urn (2). If, for example, one considers a
flow stream with 50 percent of the particulates less than 10 (im, then effi-
ciencies of 85 and 73 percent would be expected for those 2 cyclones, respec-
tively. Figure All-4 shows estimated efficiencies as a function of particle
size (2). If the size distribution is available for the inlet dust, the
overall collector efficiency may be estimated from Figure All-4. Current
performance data for mechanical collectors are limited since these devices are
often used in conjunction with another control device in which only the over-
all efficiencies are given. Some test data are available, and these are
provided in Tables All-1 and All-2.
All-4
-------
Collection Efficiency, percent
H-
09
e
i-t
OQ
c
i-i
0)
h-'
I
Ul
3 n
C o
(-• M
rr I—1
H- (D
o n
*•< rr
n H-
M o
O 3
3
n> CD
t-n
CO i-n
^ H-
cn n
ft H.
re ro
3 S
to o
ho O
--- Hi
l-t
H-
O
C
to
H-
O
Collection Efficiency, percent
H-
O
Co
n
o
ro
n
rt
H-
O
ro
Mi
i-h
H-
n
H-
ro
3
n
cu
X
I
ro
n
^
n
M
o
3
ro
-------
TABLE All-1. PERFORMANCE DATA FOR CYCLONES INSTALLED ON
SMALL UTILITY BOILERS (-73 Mw) (2)
Steam Load Emission Rate
Boiler Type (10» kg/hr) ng/J
Spreader stoker water tube 50 1217
Spreader stoker water tube 29 82
Pulverized water tube 82 427
Spreader stoker water tube 54 867
Spreader stoker water tube 73 146
Spreader stoker 37 1311
Chain gate 47 133
All-6
-------
TABLE All-2. REPRESENTATIVE PERFORMANCE OF CYCLONES AND INERTIAL SEPARATORS (3)
Collector
Type
Series cyclone
Special cyclone
Special cyclone
Cyclone
Cyclone
Cyclone
Cyclone
Rotary stream
separator
Rotary stream
separator
Inertial
Mechanical
Mechanical
Proceis
Fluid catalytic
cracking
Laboratory test
Laboratory test
Abrasive cleaning
Drying
Grinding
Planning mill
Test dust
(2.65 sp. gr.)
Tar-macadam plant
Cyclone outlet
Grinding
Rubber dusting
Material
Catalyst
Fly ash
Micronized tale
Talc
Sand and gravel
Alum in on
Wood
—
—
Sand and gravel
Iron scale
Zinc stearate
Air Flow,
m'/min
1100
5.2
5.2
65
350
68
88
8.5
—
48
330
93
Pressure
Drop,
cm. wg
High
58
58
0.8
5
3
9
20
—
10
12
23
Efficiency,
wt %
99.98
91.2
83.9
93.0
86.9
89.0
97.0
100
90
90
50.0
56.3
88.0
Inlet Load,
6400
0.1
13
5
87
1.6
0.2
0.9
200
4.6
13
0.3
1.4
Inlet Mass
Median
Size, (in
37.0
3.0
2.3
—
8.2«
—
—
5
1
2b
5.3C
3.2*
1.7
Outlet mass median size, jim = 3.2
Size frequency maximum
Outlet mass median size, (im = 1.8
Outlet mass median size, pun = 2.5
-------
Appendix All
Cyclones
4. Secondary Waste Generation
Cyclones produce a significant quantity of solid waste. However, without
particulate control, solid wastes would appear as stack emissions. The method
of solid waste disposal is a function of the composition of the particulate
matter collected. Particulate matter collected from transfer, crushing, and
screening operations can often be reused. Some collected particulate matter
can be used as a component of concrete mixtures. For solid wastes which
require disposal, the primary method is by landfill ing. Liners and proper
operating procedures can minimize runoff or leaching into the water table.
5. Process Reliability
Since cyclone collectors generally operate dry, there are seldom any
corrosion problems. Breakdowns are generally due to high gas temperatures or
abrasiveness of the dust particles. The most critical factor in start-up and
shutdown of the process is the possibility of moisture condensation on the
internal surfaces of the cyclone. Condensation will invariably start corro-
sion and initiate dust agglomeration and buildup.
A 20-year life for cyclone collectors is frequently attainable. Wear
generally occurs at the entry section of tangential-entry cyclones and at the
vanes of axial-entry cyclones. With refractory-lined cyclones, there may be
spilling and occasional fallout of part of the refractory, especially on the
horizontal top section and the gas outlet tube. In general, cyclones are very
reliable and have negligible maintenance costs.
All-*
-------
Appendix All
Cyclones
6. Process Economics
Installed costs for cyclones have been reported for pulverized coal-fired
boilers in previous EPA studies (2). These costs are provided in Figure
All-5. To obtain capital investment costs, indirect costs for construction
and field expenses, engineering, contingencies, etc. have to be added to the
installed cost. The major elements of the annual operating costs are operat-
ing labor, maintenance labor, materials, and replacement parts. Generally
0.003 man-hours of operating labor per hour of operation are required for
cyclones. Supervision man-hours required are approximately 25 percent of the
man-hours required for operating labor. Maintenance labor, materials, and
replacement parts are roughly equivalent to 1 percent of the equipment pur-
chase price.
7. References
1. U.S. Environmental Protection Agency. Air Pollution Engineering Manual,
Second Edition. AP-40, May 1973.
2. D.S. Environmental Protection Agency. Technology Assessment Report for
Industrial Boiler Applications: Particulate Collection, EPA-600/7-79-
178h, December 1979.
3. Caplan, K.J. Source Control by Centrifugal Force and Gravity, Air
Pollution, Volume IV, edited by A.C. Stern, Academic Press Inc., 1977.
All-9
-------
o
o
CO
100
7
fi
•H
B
a)
P,
10
7
en
O
en
c
103
I I I I I i I
4 7 10"
Gas Flow, n3/min
Figure All-5. Installed cost for cyclones used in pulverized
coal boiler applications (2)
All-10
-------
APPENDIX A12
FABRIC FILTRATION PROCESS
1. Process Description
The basic collection mechanisms which occur in a fabric filtration pro-
cess are inertial impaction, diffusion, direct interception, and sieving. The
first three mechanisms occur only for a brief period of time immediately after
a new or just-cleaned filter is placed in service. The sieving action of the
dust layer accumulating on the fabric surface soon predominates, particularly
at high (1 g/m3) dust loadings.
Fabric filter systems typically consist of cloth bags or envelopes,
suspended or mounted in such a way that the collected particles fall into a
hopper/bin for disposal when dislodged from the fabric. Bags usually have the
conveying gas flow from inside the bag to the outside; envelopes, from the
outside in. The accumulated material is dislodged by such devices as bag
shakers, the reverse jet from a ring-slit jet (which moves along the bag), and
a pulse-jet reverse gas-flow (which reverses total air flow in a bag).
Filters are usually compartmentalized and installed in multiple units.
The reverse-jet ring-slit device works continuously to remove the collected
particulates as it travels along the bag during filtration. Because it
removes most of the accumulation, tightly felted fabric works best, and the
initial particulate layer becomes less significant.
With the pulse-jet equipment, removal occurs on a scheduled basis when
the flow in a bag or a series of bags is reversed momentarily. Only those
bags being cleaned at any moment are out of use; the rest of the unit keeps
operating. An isometric view of a two-compartment pulse-jet fabric filter is
provided in Figure A12-1.
A12-1
-------
>
M
I
Pyramidal or
Trough Hoppers
Access Plates
Solenoid Valves
Compressed Air
Manifold
Dirty Air Inlet
Baffle Plate
Access Door
Figure A12-1. Isometric view of a two-compartment pulse-jet fabric filter (1)
-------
Appendix A12
Fabric Filtration
Shakers require shutdown of gas flow to the entire unit during the
cleaning operation. Thus, continuity of operation requires the use of multi-
ple units.
Fabric filters can be categorized as to the type of service and frequency
of bag cleaning. Intermittent fabric filters are cleaned after filtering is
completed (i.e., after the process stream is secured or shutdown) usually at
the end of each day. These fabric filters operate with low dust loadings
since they cannot be cleaned while on stream. In continuous duty fabric fil-
ters, cleaning of a portion of the filters occurs at periodic intervals while
the remainder continue to process gases. These filters are more expensive
than the intermittent types due to the accessories required in the cleaning
process and the additional filter area required for continuous operations.
Efficiency of removal depends largely upon the choice of filter medium.
A variety of fabrics are available for use in filtration systems. A selected
listing of materials and their properties is presented in Table A12-1 (2).
Tabulated air permeabilities reflect fresh fabric performance and will
decrease upon use. Other fabrics which are available include asbestos, graphi-
tized fiber, and polyethylene. Cotton and nylon, although the least expensive
fabrics listed, have poor resistance to acid attack. Orion, Dacron, and
polypropylene all resist attack by acids and may be suitable at low tempera-
tures. Fiberglass and Teflon offer higher temperature acid resistance
although Teflon is more costly. A Teflon or a Teflon-Orion mixture called MTI
may be used when significant concentrations of fluorides are present. With
the exception of wool and fiberglass, most listed materials provide good
resistance to attack by alkaline substances.
A12-3
-------
TABLE A12-1. CHARACTERISTICS OF SELECTED FILTER FABRICS (2)
Operating Temperature
Exposure. I
Fiber
Cotton
fool
Nylon"
Orion
Dacron**
Polypropylene
Nomex**
Fiberglass
Teflon**
a .
P - Poor, F -
Long
360
370
370
390
410
370
490
560
500
Fair, G
Short
380
390
390
410
440
390
530
590
530
of 125 Pa.
- Good, E -
Supports
Combustion
Yes
No
Yes
Yes
Yes
Ye i
No
Yes
No
Excellent.
Air
Permeability
0.05-0
0.10-0
0.07-0
0.10-0
0.05-0
0.03-0
0.12-0
0.05-0
0.07-0
.10
.30
.15
.23
.30
.15
.27
.35
.33
Compos! tlon
Cellulose
Protein
Polyamlde
Polyacrylonltrile
Polyeater
01 ef in
Poly amide
Glass
Polyfluoro-
ethylene
Abrasion
G
G
E
G
E
E
E
P-F
F
Mineral
Acids
P
F
P
G
G
E
F
E
E
b
Organic
Acids
G
F
F
G
G
E
E
E
E
Alkali
G
P
G
F
G
E
G
P
E
Cost
KankC
1
7
2
3
4
6
8
5
9
, , , .
CCost rank, 1 - lowest cost, 9 - highest cost
••DuPont registered trademark.
-------
Appendix A12
Fabric Filtration
2. Process Applicability
The applicability of fabric filters depends npon the emission limita-
tions, temperature, moisture content, the bag replacement schedule of the
fabric filter, and other aspects.
Temperature of the gas imposes limitations upon usable materials of
construction for bags, as well as affecting system size with regard to
volumetric flow. Moisture or other condensibles can affect performance of the
fabric filter and may render it inoperative. Significant moisture content
will definitely eliminate fabric filters from consideration. The air stream
flow rate can also have considerable effect on the performance of the fabric
filter. Wide fluctuations in the flow rate will cause a shift in the collec-
tion efficiency of the fabric filter. However, if applicable, fabric filters
provide high efficiencies with the retention of very fine particulates. In
addition, the particulates are collected in the dry form at relatively low
pressure drops.
3. Process Performance
Laboratory studies with fabric filters have demonstrated a strong
correlation between outlet concentration and face velocity (air-to-cloth
ratio) for a given loading and type of fabric. Predicted and observed outlet
concentrations for bench scale tests indicate that outlet concentration
increases to a certain extent and then tapers off as air-to-cloth ratio (face
velocity or fabric loading) is increased. The results of these tests are
presented in Figure A12-2. This effect has also been substantiated by field
pilot studies (refer to Figure A12-3) although there are some inconsistencies
which could be due to control problems in field experimentation.
A12-5
-------
Outlet Concentration (Co), g/nr
>
I—1
I
09
i-!
(D
ho
I
to
H- rj
h- CD
rr C^
(D H-
>-! O
rr
rt (t>
ro P-
w
rt fU
« 3
O
a cr
n w
> ft
i-t
rt, <
P i-!
w re
cr en
e
P M
D rr
CO i-n
C O
0 if
cr
C ^J
i-i TO
'< 0
O
rt, 3-
CT W
I-i O
H* P
O h-1
fD
cr
H-
n
o
01
o
co
o
KJ
H-
n
r1
o
05
d-
H-
p
05
N)
O
a>
o
00
o
8
to
o
n
CO
i
co
CO
01
_»
.u
b
o o <
8 | 8
m
Wl -J 00
co b b
xj _» (O
_. o p
In b> '-U
M -> O
CJl IO -»
b b co
m
z
m
H
o'
O
z
o
3"
___^
5_
3
3'
3
3"
5'
-n
2?
m
m
i
^^
H
i i 1111
-------
O.I
0.08
0.06
0.04
o
•H
4-1
CO
0)
0)
PL,
0.02
0.01
0.008
0.005
" Penetration = 1 - fractional collection efficiency
O NOMEX
A DRALON T
D GORE-TEX
• TEFLON
0.91
1.83
2.74
3.66
Air-to-Cloth Ratio, m/min
Figure A12-3. Fabric filter field test results for different
bag materials (1)
A12-7
-------
Appendix A12
Fabric Filtration
Although a relationship between pressure drop and air—to—cloth ratio does
exist, it is difficult to directly relate pressure drop to the collection
efficiency of the fabric filter. The pressure drop across fabric filters
results from resistance to air flow presented by the fabric and by the
collected dust layer and may be expressed by (3):
AP = K- IL, + K, Ct U * (1)
0 G d G
where: AP is the total pressure drop or total resistance, Pa,
K- is the resistance coefficient of the clean fabric after
initial use, Pa per m3/s,
K is the dust resistance coefficient, Pa per m3/s per g of
dust per m* of fabric,
U is the superficial filtration/face velocity, m/s,
C is the inlet dust concentration, g/m3, and
t is the filtration period, seconds.
From equation (1) it is obvious that the pressure drop generally varies as the
square of the face velocity (air-to-cloth ratio), at constant dust loadings.
This is demonstrated in Figure A12-4 where the resistance value for three dif-
ferent fabric materials are illustrated for a 60 minute cleaning cycle.
The filter pressure drop is influenced by the frequency and completeness
of bag cleaning. Figure A12-5 shows the effect of continuous versus cylical
cleaning on the filter pressure drop for a reverse air—jet type of cleaning
mechanism. At the lower dust loading value of 1 g/am3, this curve indicates
that operating the cleaning mechanism only 30 percent of the time causes an
increase in pressure drop of only 10 percent. Because of the wear on most
A12-8
-------
Pressure Drops Across Bags, cm WG
09
e
NJ
I
i—1 n>
rt CD
fD co
Co
rt pu
H- l-l
o o
CO 13
CO
/•^
OJ 03
rt
ft)
H
W
rt
H-
O
H-
3
00
O
cn
fD
d
O
Ul
I I
:r on c/J c/5
t) H- H- H-
rt M I-1 M
fB H- H- H-
i-i n n n
o o o
03 H- H- H-
C N N N
OQ fD fD fD
0,0,0.
OQ
H-
O
CD
H-
O
Increase in Filter Pressure Drop, %
H-
TO
c
fD
>
M
to
I
ro
o
t-n
o
n
m
<
CO
n>
rt
fD
ro
CD
CD
<->• C
O CL
-a H
fD O
fi T3
&}
rt Hi
H- O
O i-l
n
M
(D
03
3
H-
3
OQ
•n
H
fD
fD
3
O
v;
ho
O
oo
o
O
O
-------
Appendix A12
Fabric Filtration
fabrics imposed by cleaning, these values would indicate the desirability of
accepting minimal pressure drop increase while reducing bag maintenance. How-
ever, the same figure also indicates a need for more frequent cleaning at
higher dust loadings.
4. Secondary Waste Generation
Fabric filters produce a significant quantity of solid waste. However,
it must be realized that without particulate control, solid wastes appear as
stack emissions. The method of solid waste disposal is a function of the
composition of the particulate matter collected. Particulate matter collected
from transfer, crushing, and screening operations can often be reused. Some
collected particulate matter can be used as a component of concrete mixtures.
For solid wastes which require disposal, the primary method is by landfill ing.
Liners and proper operating procedures can minimize runoff or leaching into
the water table.
5. Process Reliability
The reliability of a fabric filter system in meeting performance require-
ments depends mainly on the frequency of bag breakage, the time taken to iso-
late the broken bags, the leakage of bypass dampers, and sometimes the air-to-
cloth ratio. Conservative design and proper operation can generally minimize
the frequency of broken bags. Also periodic replacement of all bags in a com-
partment will reduce the average age of the bags and the frequency of break-
age. Minimizing excursions below the acid dew point also tends to extend bag
life.
A12-10
-------
Appendix A12
Fabric Filtration
The time to isolate a broken bag can be drastically reduced if detectors
are placed on the outlet of each compartment. These detectors actually
indicate the opacity of the gas stream from the compartment and allow
operators to isolate quickly the faulty compartment for repair.
Leaks through bypass dampers also affect performance since leakage can
approach 5 percent. Leakage can be reduced essentially to zero by including
two louvered dampers with a continuous purge of clean reverse air to block
uncleaned gas.
High air-to-cloth ratios experienced during cleaning or maintenance can
reduce performance. However, fabric filter efficiency generally is high
enough that requirements can be met during these periods. The air-to-cloth
ratio influences reliability more strongly than performance because it affects
bag life.
6. Process Economics
Costs of fabric filter baghonses depend upon: 1) type of fabric, 2) air-
to-cloth ratio; 3) intermittent or continuous duty, 4) pressure or suction
type construction, 5) standard or custom design, 6) type of cleaning
mechanism, and 7) materials of construction.
Costs for mechanical shaker, pulse-jet, reverse-air, and custom baghouses
are presented in Figures A12-6 through A12-10. The prices are based on net
cloth area which is determined from the specified air—to—cloth ratio. The net
cloth area is the average filter area seen by the air flow. This may differ
from the gross cloth area (which is the total cloth area in the baghouse) due
to the need to isolate cloth area when continuous cleaning is used. For an
intermittent system, the net and gross cloth areas are the same, since the
A12-11
-------
>
M
I
0)
1
u
o
H
•CO-
O
u
u
15-
10-
5 -
200 400 600 800 1000
Net Cloth Area, m
1200
2
1400
1600
1800
2000
Figure A12-6. Costs for pressure flow, intermittent duty, mechanical shaker
standard construction baghouses, exclusive of bags (4)
-------
I
I—'
U)
S-J
-------
140
>
I—'
I
1600
3200 4800
Net Cloth Area, m2
6400
8000
Figure A12-8.
Costs for continuous duty, pressure flow, mechanical shaker
standard construction baghouses, exclusive of bags (4)
-------
fo
I
OJ
u
01
Q
0)
O
•H
CX
cd
O
280
240
200
160
120
80
2000
4000
Net Cloth Area, m
6000
2
8000
10,000
Figure A12-9.
Costs for continuous duty, pressure flow, reverse air standard
construction baghouses, exclusive of bags (4)
-------
hJ
I
01
0)
u
0)
Q
w
JJ
CO
•H
P.
cfl
U
1400
1200 -
1000 -
300 -
600 -
400 -
200-
10
15 20 25
Net Fabric Area, 103 m2
30
35
40
Figure A12-10. Costs for continuous duty, reverse air custom construction
baghouses, exclusive of bags (4)
-------
Appendix A12
Fabric Filtration
baghouse is cleaned after a filtering cycle. For continuous filters, factors
to calculate gross cloth area can be found in Table A12-2. Dsing this calcu-
lated gross area and type of filter desired, bag costs can be determined from
Table A12-3.
TABLE A12-2. APPROXIMATE GUIDE TO ESTIMATE GROSS CLOTH AREA (4)
Net Cloth Area,
m3
1
370
1100
2200
3300
4400
5600
6700
7800
8900
10,000
12,300
16,700
- 370
- 1100
- 2200
- 3300
- 4400
- 5600
- 6700
- 7800
- 8900
- 10,000
- 12,300
- 16,700
On Dp
Gross Cloth Area,
m3
Multiply by 2
Multiply by 1.5
Multiply by 1.25
Multiply by 1.17
Multiply by 1.125
Multiply by 1.11
Multiply by 1.10
Multiply by 1.09
Multiply by 1.08
Multiply by 1.07
Multiply by 1.06
Multiply by 1.05
Multiply by 1.04
Baghouse prices are flange-to-flange, including the basic baghouse with-
out bags, 3 meter support clearance, and inlet and exhaust manifold. Pressure
baghouses are designed for 3 kPa gauge pressure and suction fabric filters are
designed for a negative pressure of 5 kPa gauge. Custom baghouse prices are
more a function of specific requirements than of pressure or suction construc-
tion, so prices do not differentiate between pressure or suction. Custom bag-
houses are designed for continuous operation and normally use reverse air
cleaning. All baghouses except custom baghouses are generally factory-
assembled. The operating requirements for fabric filters primarily consist of
labor, replacement bags, and power. Labor requirements include 2 to 4 man-
hours per shift for operations and 1 to 2 man-hours per shift for maintenance.
Bag life is estimated at 0.3 to 5 years with an average of 1.5 years. Power
A12-17
-------
TABLE A12-3. BAG PRICES ($/m2)a (4)
>
M
1
M
00
Class Type
Standard Mechanical shaker, <1900 m1
Mechanical shaker, >1900 m*
Pulse jetb
Reverse air
Custom Mechanical shaker
Reverse air
Dacron
4.30
3.80
6.50
3.80
2.70
2.70
Orion
7.00
5.40
10.20
6.50
3.80
3.80
Nylon
8.
7.
7.
4.
4.
10
50
50
80
80
Nome i
12.40
11.30
14.00
11.30
7.00
7.00
Glass Poly-
propylene
5
4
4
3
3
.40
.80
.80
.20
.20
7.00
5.90
7.50
5.90
3.80
3.80
Cotton
4
4
4
4
4
.80
.30
.30
.30
.30
December 1977 prices.
For heavy felt, multiply price by 1.5
-------
Appendix A12
Fabric Filtration
requirements for shaker and blower motors are given as approximately 4kW/1000
ma of cloth area (4). Power usage will depend on dust loading and cleaning
cycle.
7. References
1. U.S. Environmental Protection Agency. Technology Assessment Report for
Industrial Boiler Applications: Particulate Collection. EPA-600/7-79-
178h, December 1979.
2. U.S. Environmental Protection Agency. Air Pollution Control Devices for
Hazardous Waste Incineration Permit Writer Guidelines. Report prepared
under Contract No. 68-02-3174, Task 9, TSA 1-7, March 1980.
3. U.S. Environmental Protection Agency. Air Pollution Engineering Manual,
Second Edition. AP-40, May 1973.
4. U.S. Environmental Protection Agency. Capital and Operating Costs of
Selected Air Pollution Control Systems. EPA 450/5-80-002, December 1978.
A12-19
-------
-------
APPENDIX A13
ELECTROSTATIC PRECIPITATION PROCESS
1. Process Description
Electrostatic precipitators use electrical rather than mechanical forces
for the removal of suspended particulates in a gas stream. The process encom-
passes three basic functions: the charging of particles, the collection of
particles on an electrode of opposite polarity, and the removal of the col-
lected particles. Particles suspended in the gas stream are charged by pas-
sing through a high voltage direct—current corona established between a dis-
charge electrode, usually a small diameter wire which is maintained at high
voltage, and a grounded collecting surface (collecting electrode). As the
particles pass through the corona, they are bombarded by negative ions emanat-
ing from the discharge electrode and charged within a fraction of a second.
The charged particles, influenced by electric field forces, migrate toward the
grounded collecting surface where they are deposited and held by electrical,
mechanical, and molecular forces. Particulates adhering to the collecting
surface are periodically dislodged by mechanical rappers or by flushing with
water and fall to a hopper from which they are subsequently removed.
A majority of the industrial electrostatic precipitators (ESPs) used
today are the single-stage, wire and plate type. Charging and collection take
place in the same section of the ESP. The collecting surface consists of flat
parallel plates spaced from 15 to 30 cm apart with wire or rod discharge elec-
trodes located between the plates. The plates usually range from 4 to 12 m in
height and 4 to 7 m in length. Plate-type precipitators are typically used
for dry particulate collection. In tube-type ESPs, the collecting surface
consists of a cylinder with the discharge electrode centered along its longi-
tudinal axis. They are generally used for wet gas cleaning. With either
type, a complete precipitator consists of many units in parallel and in
series. A schematic arrangement of wire/plate and wire/tube ESPs is shown in
Figure A13-1.
A13-1
-------
Appendix A13
ESPs
Discharge
electrodes
Ground
electrodes
Ground
electrodes
Wire/Plate Type
Gas flow
Wire/Tube Type
Figure A13-1. Schematic arrangement of wire/plate and wire/tube precipitators
The fundamental equation for describing the collection efficiency of ESPs
is the Deutsch equation:
n = 1-exp [-(A/QG)w]
where: n is the collection efficiency (weight fraction collected),
A is the area of the collection electrode,
QG is the gas flow rate, and
w is the particle migration velocity.
Typically, the migration velocity is treated as an empirical constant and cal-
culated from measured efficiencies. Migration velocities thus calculated form
the basis for ESP sizing in a specified application. Migration velocities
increase with increasing particle size for particles larger than approximately
0.5 jim in diameter. Below 0.5 urn, migration velocities tend to increase with
A13-2
-------
Appendix A13
ESPs
decreasing particle size due to the effect of diffusion on the particle charg-
ing mechanism. For a specified collection efficiency, the required ESP size
generally increases with decreasing gas flowrate and increasing migration
velocity.
In the absence of migration velocity measurements, estimates of average
migration velocities may be based upon industrial experience. Ranges of
migration velocities encountered for various applications are available and
provided in a following section. These values may be used in conjunction with
the Deutsch equation for rough estimating purposes; however, since the Deutsch
equation does not consider such factors as dust reentrainment, gas leakage,
and poor velocity distribution, this equation represents an idealized
approach.
In both ESP types, there are four main components: electrode system,
precipitator casing, collected material removal system, and power supply.
In the single—stage wire/plate ESP the discharge electrodes may be round
wire, square twisted rods, ribbons, etc. The choice of construction material
is a function of the corrosive service. The most common collection electrode
in the wire/plate design is a smooth plate with vertical interlocking baffles.
These baffles provide strength and also produce near-zero velocity conditions
as the gas flows in a normal direction to them. Other special plate electrode
configurations are rod curtains, zigzag plates, and various hollow electrodes
with pockets on the outside surfaces for discharging the collected dust to the
hopper from quiescent gas zones. As in the case of the discharge electrodes,
the plates are fabricated from a wide choice of construction materials,
depending upon the degree of corrosive service. The high-voltage discharge
electrodes are suspended vertically between each pair of collection elec-
trodes. They are carefully centered between the collection electrodes to
ensure proper corona gaps. Typical design parameters for commercial wire/
plate precipitators are shown in Table A13-1 (1).
A13-3
-------
Appendix A13
ESPs
TABLE A13-1. TYPICAL VALUES OF DESIGN VARIABLES FOR COMMERCIAL ESPs (1)
Design Variable
Normal Range of Values
Plate spacing
Gas velocity
Vertical height of plates
Horizontal length of plates
Applied voltage
Gas temperature
Treatment time
Draft loss
Efficiency
Corona current
Field strength
0.20 to 0.28 m
0.61 to 2.4 m/s
3.7 to 7.3 m
0.5 to 1 x height
30 to 75 kV
Up to 643 K standard
810 high temperature
977 K special
2 to 10 s
25 to 125 Pa
Dp to 99.9%, although usually
90 to 98%
0.033 to 3.3 mA/m of wire
276 to 590 kV/m
The precipitator casing may be fabricated from a wide selection of
materials including mild steel, lead-lined steel, brick, and concrete. The
length of the ESP may be comprised of several fields in the direction of the
gas flow. Each field section can be treated as a separate unit module so
that additional fields can be easily tied into the system as the need for
increased collection efficiencies arise. The entire casing is supported on
a steel base that rests on support steel. Main support columns, attached to
the steel base, support structural members near the top of the precipitator
from which the grounded electrodes are suspended. The casing roof is also
supported from these structural members. The discharge electrodes are hung
from high-voltage insulators located on and supported by the roof. These
insulators, together with gas seals and rapper mechanisms, are enclosed in an
attic space above the casing roof. The storage hoppers for the wire/plate ESP
are located under the collection electrodes and are a structural continuation
of the casing.
A13-4
-------
Appendix A13
ESPs
Collected dust from the ESP is removed by rapping and washing. Rapping
methods usually consist of mechanical actuated "hammers", which are driven
electrically or pneumatically. The impact blow from the hammer is trans-
mitted vertically to the freely suspended collection electrode so that the
dust is released in the direction of gravity. Rappers are provided for the
discharge electrodes as well. In the case of some viscid dusts the removal of
collected material can be accomplished by washing down the plate. Periodic
water sprays are used, and by proper cycling of these sprays to maintain a
wetted electrode, reentrainment can be completely eliminated.
The electrical energy requirement for an electrostatic precipitator is
that necessary to produce an effective corona. The power supply must deliver
a unidirectional negative current to the discharge electrodes at a potential
very close to that which will produce arcing across the electrodes. The value
of the potential difference used in the single-stage precipitator is usually
in the range of 20,000 to 100,000 volts. The current delivered may vary from
20 to 500 mA. Power requirements are relatively small since only the dust is
treated rather than the total gas flow. Solid state rectification is used in
most new ESPs. Selenium and silicon rectifiers provide reliable service with
long life. The ESP potential is maintained at the optimum value by a spark
counter or current-sensing feedback circuit.
The single-stage wet wire/tube precipitator design is better suited for
wet collection applications. This type of precipitator is built in a cylin-
drical shell. The collection electrodes consist of nested pipes which are
connected and sealed to header sheets attached to the shell. The discharge
electrodes are supported above the header sheet and are suspended axially in
the collection electrode pipes. Water is introduced above the header sheet
and flows over carefully leveled weirs at the tops of the pipes to form a
water film on the inner walls of the pipes. The charged particles are col-
lected in the water film, where their charge is neutralized, and drained off
A13-5
-------
Appendix A13
ESPs
from the bottom of the ESP with the water. Because wet collection is
involved, materials of construction are usually corrosion resistant.
The power supply is similar to that for the plate type precipitator.
However, it is simpler in concept because the vertical gas flow pattern elimi-
nates the need for sectionalization.
2. Process Applicability
Electrostatic precipitators are used extensively on large volume applica-
tions where the fine dust and particulate is less than 10 to 20 pm in size
with a predominant portion in the sub jim range. The precipitators can achieve
high efficiencies (in excess of 99 percent) depending on the resistivity of
the particulate matter and the characteristics of the gas stream. Wet or dry
particulate can be collected including highly corrosive materials if the units
are suitably constructed. Precipitators can be used at temperatures up to
810 K but are normally operated at temperatures below 650 K. The static
pressure drop through the units is low, usually up to 12 kPa for units operat-
ing at gas velocities of 0.6 to 2.4 m/sec. Safety precautions are always
required since the operating voltages are as high as 100,000 volts. The
overall size of electrostatic precipitators is comparable to fabric filters
(baghouses). Space requirements are an important factor in the layout and
design of the facilities.
By far the greatest application for ESPs has been in the electric power
industry. ESPs are the established type of control equipment for the elimina-
tion of flue gas particulate emissions from coal combustion. However, the
application of ESPs is not limited to the electric power industry. Table
A13-2 lists some of the industries in which ESPs have found application.
A13-6
-------
TABLE A13-2. TYPICAL APPLICATION DATA FOR ELECTROSTATIC PRECIPITATORS (6)
U)
i
Industry
Electric Power
Portland Cement
Steel
Application
Fly ash-Pulverized
Coal Boiler
Dust from Kilns
Dust from Dryers
Mill Ventilation
Blast Furnace Gas
Cleaning
Tar Collection
Gas Flow,
D>/S
24-380
24-470
14-47
0.9-4.7
9.4-47
24-94
Temperature,
K
400-590
420-670
320-450
280-320
310-340
310-340
Dust
Concentration
g/««
09. -11
1.1-34
2.3-34
11-57
0.05-1.1
0.23-2.3
Dust Size,
* <10|i»
25-75
35-75
10-60
35-75
100
100
Collection
Efficiency
%
95-98
85-99+
95-99
95-99
95-99
95-99
Nonferrous Metals
Palp and Paper
Chemical
Petroleum
Gas
From Coke Oven
Gases
Fume Collection 14-35
From Open Hearth
and Electric
Furnace
Fume From Kilns, 2.4-470
Roasters, and
Sinter Machine
Kraft Soda Fume 24-94
Acid Mist 1.2-9.4
Gas Cleaning, SO, 2.4-9.4
CO,, etc.
Powdered Catalyst 24-70
Tar Removal 0.9-24
From Gas
420-640
340-870
410-450
310-370
290-370
450-560
280-340
0.11-6.9
0.11-114
1.1-4.6
0.05-2.3
0.02-2.3
0.23-57
0.02-0.5
95
10-100
99
100
100
50-75
100
90-99
90-98
90-95
95-99
90-99
99-99.9
90-98
-------
Appendix A13
ESPs
3. Process Performance
The three most important design criteria for ESPs are the precipitation
rate (W> , the specific collection area (SCA), and the gas velocity
e
Because precipitation rate can vary with resistivity, particle size distri-
bution, gas velocity distribution, rapping, and electrical factors, an
effective rate parameter or migration velocity is usually adopted. Variation
of this parameter with fly ash resistivity and coal sulfur content is shown in
Figures A13-2 and A13-3. Particulate resistivities of over 10» to 1010 ohm-cm
were found to decrease collection efficiency. This loss results since some
high resistivity particles are not able to dissipate their charge as they
attach to the collector electrode. Eventually, an electric potential can
build up at the surface of the collected particulates giving rise to tuft-like
discharges of polarity opposite to that of the collector. Subsequently,
arriving particles are repelled rather than collected at these sites (1,2).
As explained earlier, the Deutsch equation is generally used to estimate
collection efficiency when one solely relies upon plate area, gas flow rate,
and average electrical migration velocity. In most cases, however, field data
indicate lower efficiencies than predicted by the Deutsch relationship. To
account for the observed particle collection levels, White (2) and Peters (4)
use modified Deutsch equations as more realistic predictors of particulate
collection efficiency. White uses "effective" migration velocity computed
from experimental measurements. Alternatively, Peters determines the effec-
tive migration velocity using the Cunningham correction factor for particle
slip. Equipment suppliers rely on experience in similar applications to size
commercial precipitators. Typical application data for precipitators are
shown in Table A13-2.
A13-8
-------
0.6
0.5
LU
0.4
O °'3
0.2
oc
0.
0.1
10'
10
ID
«s
6
Q.
O
3 LU
a:
a.
0
12
Figure A13-2.
RESISTIVITY, ohm-cm
Drop in precipitation rate We with increasing fly ash
resistivity for a representative group of precipitators (5)
u. /
£ 0.6
»^
~
-------
Appendix A13
ESPs
The Deutsch equation indicates that increased migration velocities at a
constant gas flow rate will yield greater efficiency values. Conversely, for
any single migration velocity value, the collection efficiency is inversely
proportional to the gas flow. The relationship of gas flows to theoretical
collection efficiency as a function of migration velocity is represented in
Figure A13-4. Some suggested values of migration velocities for various
applications are tabulated in Table A13-3.
It should be noted that migration velocity is directly proportional to
particle size. Therefore, the larger the particle size, the larger the migra-
tion velocity and the higher the particulate collection efficiency. Gas velo-
city in the ESP is extremely important since collection is highly sensitive to
velocity variations. The gas flow velocity at which maximum efficiency can be
attained depends on such factors as plate configuration, precipitator size,
and the judicious use of flow distributors required to minimize velocity
gradients. The design velocity limit for high efficiency fly ash precipita-
tors is about 1.5 to 1.8 m/s.
In general, the performance of a given ESP unit is a function of "the
size of the box" (plate area and depth), the particle resistivity and size
distribution, the electrical parameters defining particle charge and field
strength, and proper operation and maintenance of equipment. Typical collec-
tion efficiencies that can be achieved by ESPs in various applications are
listed in Table A13-2.
4. Secondary Waste Generation
The collected particulate is the only secondary waste generated by ESPs.
However, it must be realized that without particulate control, these particu-
lates would appear as stack emissions. Since the collected particulates can
be removed from the collector surface in the dry form or washed off with
A13-10
-------
100
g 90
w
c
o
o
0)
o
o
80
70
cm/s
I I 1
2 46
Gas Velocity, m/s
Figure A13-4. Theoretical precipitator collection efficiencies
at different migration velocities (6)
TABLE A13-3.
TYPICAL AVERAGE MIGRATION VELOCITIES ENCOUNTERED
IN COMMERCIAL ESP SYSTEMS (6)
Application
Migration Velocity (w),
m/s
Pulverized coal (fly ash)
Paper mills
Open-hearth furnace
Secondary blast furnace (80% foundry iron)
Gypsum
Hot phosphorous
Acid mist (H2S04)
Acid mist (Ti02)
Flash roaster
Multiple hearth roaster
Portland cement manufacturing (wet process)
Portland cement manufacturing (dry process)
Catalyst dust
Grey iron cupola (iron to coke ratio - 10)
0.10 to 0.13
0.076
0.057
0.12
0.16 to 0.19
0.027
0.057 to 0.076
0.057 to 0.076
0.076
0.079
0.10 to 0.11
0.057 to 0.070
0.076
0.030 to 0.036
A13-11
-------
Appendix A13
ESPs
water, the resultant waste can be in the form of dust or water slurry. In the
dry form, some precaution would be required to minimize resuspension of the
particulates as fugitives into the ambient air.
The primary method of particulate disposal is by landfilling. Liners and
proper operating procedures can minimize runoff or leaching into the water
table. Some collected particulate matter can be used as a component of con-
crete mixtures. However, such application is very limited.
5. Process Reliability
There are two types of reliability problems that are encountered with
ESPs: mechanical and operational. Mechanical problems include, for example,
misalignment of electrodes, breakage of corona wires by fatigue or by elec-
trical burning, and air in-leakage into the hoppers. Operational problems
commonly encountered include poor electrical set adjustments, shorted corona
sections (e.g., caused by broken wires), overloading the precipitator by
excessive gas flow, and failure to empty hoppers of collected dust. A more
complete list of the most frequently encountered problems is given in Table
A13-4.
ESP reliability can be improved by increased sectionalization; precipita-
tors can have as many as 100 independent electrical sections. To minimize the
effect of ash-valve failures or plugged hoppers, the trend is toward locating
a hopper under each section. A well-designed ESP may have 5 to 25 percent of
the bus sections out of service during the operational year. The loss in
efficiency must be offset by additional collection areas beyond the plate area
added for performance efficiency contingencies. Commonly, this extra area is
expressed in terms of extra fields. For instance, a precipitator that could
meet guarantees with four fields may have a fifth field installed to account
for normal deterioration and contingency.
A13-12
-------
TABLE A13-4. COMMONLY ENCOUNTERED PRECIPITATOR PROBLEMS (3)
Fundamental Problems
1. High resistivity particles
2. Reentrainment of collected particles
3. Poor gas flow
4. Insufficient or unstable rectifier equipment
5. Insufficient number of corona sections
6. Improper or incompatible rapping
7. Gas velocity too high
8. Aspect ratio too small
9. Precipitator size too small
Mechanical Problems
1. Poor electrode alignment
2. Distorted or skewed collecting plates
3. Vibrating or swinging corona wires
4. Excessive dust deposits on corona electrodes and/or collecting
plates (sometimes cemented on)
5. Formation of dust mountains in precipitator inlet and outlet
ducts
6. Gas turning vanes and/or gas distribution screens plugged with
dust
7. Air inleakage into hoppers, shells, or gas ducts
8. Gas leakage around precipitation zones and/or through hoppers
Operational Problems
1. Full or overflowing hoppers
2. Shorted corona sections (e.g., broken wires)
3. Rectifier sets or controls poorly adjusted
4. Precipitator overloaded by excessive gas flow
5. Precipitator overloaded by excessive dust concentration
6. Process upsets (e.g., poor combustion, steam leaks)
A13-13
-------
Appendix A13
ESPs
6. Process Economics
Several EPA reports (5,7) give ESP cost estimates. However recent infor-
mation received from vendors (8) has indicated that these cost estimates are
much lower than current market costs. Costs for ESPs are given in Figure
A13-5 (9). This figure illustrates capital investment costs for precipitators
used on 500 MW pulverized coal fired power plants. These costs are based on
20 different designs and estimates. Included in the estimates are material
and labor costs for the installation of the collectors and associated
equipment. Added to these costs are differential and indirect field costs,
engineering costs, fee at 3 percent, and contingency and miscellaneous costs
at 10 percent.
Costs for ESPs increase significantly as emission limits become more
restrictive. Note that capital investment increases about 20 percent when the
outlet emission is halved. Costs also vary depending upon the type of coal
and properties of the fly ash. To meet NSPS standards for electric utilities
of 13 ng/J would require capital investment from $42/kW to j57/kW (1978
dollars) depending upon the type of coal being fired. Costs also increase
with increase in size as shown in Figure A13-6 for a dry ESP. Although these
costs (no date given) are indicated to be substantially lower than current
market costs, they should provide a relative cost vs. size factor. Many items
influence costs, and for wet precipitators, the cost may be 2 to 10 times
higher than for dry ESPs (7).
Annual operating and maintenance costs for ESPs are generally composed of
maintenance labor and materials, maintenance supervision, plant general and
administration, and electrical power costs. Maintenance labor and material
costs are approximately 2 percent of the total capital investment. Mainte-
nance supervision and plant general and administrative costs are approximately
A13-14
-------
10
001 002 0.05
PARTICULATE EMISSION LIMIT. lb/106 BTU
002 005 0.1
PARTICULATE EMISSION LIMIT, lb/106 BTU
10
001 002 005 01
PARTICULATE EMISSION LIMIT, lb/106 BTU
10
001 002 005 0.1
PARTICULATE EMISSION LIMIT, lb/106 BTU
•1 20-COMPARTMENT FABRIC FILTER WITH 2-YEAR BAG-REPLACEMENT
•2 20-COMPARTMENT FABRIC FILTER WITH 4-YEAR BAG-REPLACEMENT
•3 40-COMPARTMENT FABRIC FILTER WITH 2-YEAR BAG-REPLACEMENT
Notes: 1 lb/10b Btu = 430 ng/J.
in 1978 dollars.
Costs are
Figure A13-5.
Capital investment for collectors on 500MW
(net) power plants (9)
A13-15
-------
10'
en
o
u
03
4-1
•H
P.
0)
u
10C
10=
100
1000
10,000
100,000
Plate Area, m
Figure A13-6. Purchase price of dry-type electrostatic
precipitators (7)
A13-16
-------
Appendix A13
ESPs
5 percent and IS percent of maintenance labor and material costs, respec-
tively. Electrical power requirements average 0.012 kW/m3 for high efficiency
ESPs and 0.009 kW/m» for other ESPs. High efficiency ESPs are those units
which have collection efficiencies of over 99 percent.
7. References
1. Tomany, J.P. Air Pollution: the Emissions, the Regulations, and the
Controls. American Elsevier Publishing Co., Inc., 1975.
2. Dennis, R., ed., Handbook on Aerosols. TID-26608. Technical Information
Center, Energy Research and Development Administration, Tennessee, 1976.
3. White, J.J. Role of Electrostatic Precipitators in Particle Control: A
Retrospective and Prospective View. Jour. Air Poll. Control Assoc. Vol.
25, No. 2., February 1975.
4. Peters, J.M. Predicting Efficiency of Fine-Particle Collectors.
Chemical Engineering, Volume 80, No. 9, 99-102, 1973.
5. U.S. Environmental Protection Agency. Technology Assessment Report for
Industrial Boiler Applications: Particulate Collection. EPA-600/7-79-
178h. December 1979.
6. U.S. Environmental Protection Agency. Air Pollution Engineering Manual,
Second Edition. AP-40, May 1973.
7. U.S. Environmental Protection Agency. Capital and Operating Costs of
Selected Air Pollution Control Systems. EPA 450/5-80-002, December 1978.
8. Correspondence from K.E. Buttke, Southport Equipment Corporation to Bob
Bakshi of TRW. January 26, 1982.
9. Severson, S.D., F.A. Homey, D.S. Ensor, and G.R. Markowski. Economic
Evaluation of Fabric Filtration Versus Electrostatic Precipitation for
Ultrahigh Particulate Collection Efficiency. FP-775, June 1978.
AL3-17
-------
-------
APPENDIX A14
VENTDRI SCRUBBING PROCESS
1. Process Description
A Venturi scrubber is one of several types of wet scrubbers and is the
most commonly used for large scale control of particulates. It uses the
venturi effect to create water droplets in the high velocity gas stream.
These droplets coalesce and collide with and scrub the particulates. Use of
baffles and mist eliminators in conjunction with the velocity decrease returns
the droplets with entrained particulates to the liquid phase.
In venturi type scrubbers (Figure A14-1) mixing is achieved by using fans
to accelerate the incoming gas stream to velocities in the range of 46 to 120m
per second in the converging venturi section. As high velocity gas enters the
throat of the venturi, it encounters a stream of scrubbing liquor flowing down
the walls of the chamber. The liquor is introduced by flooding, sprays, or
weirs, and a spray of atomized liquid is directed into the gas stream ahead of
the venturi throat. As the thin sheet of liquid reaches the venturi throat,
it is sheared off by the gas stream to become a mass of droplets entrained in
the atomized sprays. Typically these droplets are 25 to 100 \im in size. In
the turbulent zone just beyond the entrance to the throat, these relatively
massive droplets move much more slowly than dry particulates in the gas
stream. As a result, collisions between droplets are frequent, so most parti-
cles are captured by the scrubbing liquor.
As droplets with their burden of particulates move through the venturi,
the gas/droplet mixture decelerates, and the droplets collide with each other
to form even larger and heavier drops. When they enter the next stage, these
are separated from the gas in a cyclonic (or other) separation device.
A14-1
-------
Appendix A14
Venturi Scrubber
DIRTY GAS
INLET
O
LIQUOR
INLETS
ALTERNATE
LIQUOR INLETS
FLOODED
ELBOW
TANGENTIAL
INLET
CLEAN GAS
OUTLET
CYCLONIC
SEPARATOR
LIQUOR
OUTLET
Figure A14-1. Standard venturi scrubber
Inertial impaction is the most important mechanism for collection of par-
ticulates larger than 0.1 — 0.3 um in diameter. The performance of most
inertial impaction devices may be expressed by the following relationship (1):
C
= exp (-A dpaB)
(1)
where:
P (Penetration) is the fractional amount of particle not
removed (i.e., 1 - efficiency),
C is the outlet particle concentration, mg/m3,
C. is the inlet particle concentration, mg/m3,
A and B are empirical constants (B ~ 2 for venturi scrubbers), and
d is the aerodynamic particle size and can be related to the
pa
Stokes diameter by using Stokes law, pm.
A14-2
-------
Appendix A14
Venturi Scrubber
Hesketh (2) developed an empirical relationship between penetration of
all particles 5 |im or less in diameter and the pressure drop across Venturis
based on data from the collection of a variety of industrial dusts. Assuming
that particles larger than 5 |im are collected with 100 percent efficiency,
this relationship may be utilized with size distribution data to estimate
overall penetration:
Pt = 0.475W(AP) ~1<43 (2)
where:
W is the weight fraction of inlet particles 5 jim or less in
diameter, and
AP is pressure drop in kPa.
From equation 2 it is obvious that with an increase in small «5 (im) par-
ticles, a higher pressure drop is required to maintain a given particle pene-
tration. For a given dust, as the pressure drop is increased, finer droplets
are atomized to interact with the dust particles through impingement and
agglomeration, with the consequent increase in collection efficiency.
Increasing the pressure drop can be accomplished by either increasing the gas
stream throat velocity, increasing the scrubber liquor flow rate, or both.
The relationship between particle size and pressure drop is shown in Figure
A14-2. The relationship between pressure drop and collection efficiency is
the same for all types of venturi scrubbers irrespective of the size, shape,
or general configuration of the scrubber. Venturi scrubbers are normally
operated at pressure drops of between 1.5 and 20 kPa, depending on the
characteristics of the dust, and at liquor flow rates of 0.4 to 2.7 L/min
A14-3
-------
Appendix A14
Venturi Scrubber
per gas flow of 1 mVmin. The collection efficiencies range from 99+ percent
for one (im (micron) or larger sized particles to 90 to 99 percent for parti-
cles below one pm size.
T3
0)
>
o
£
0)
0)
N
(-1
CO
0-.
2.0
1.5
1.0
0.5
0
Figure
0 50 100 150 200 250 300
Pressure Drop Required, cm WG
lifficiency performance of venturi scrubbers (3)
A separator for removal of the agglomerates from the gas stream is pro-
vided downstream of the scrubber. These separators are usually of the cyclone
type in which the gas stream and agglomerates are given a cyclonic motion
which forces the liquid and particles to impinge on the walls of the separator
by centrifugal force. The separator normally consists of a cylindrical tank
with a tangential inlet located at the lower side of the tank and an exhaust
outlet located at the top of the tank on the centerline axis. A cone bottom
with outlet is provided to collect the liquid slurry. The collected particles
A14-4
-------
Appendix A14
Venturi Scrubber
settle to the bottom of the cone and are sent to the water treatment facility
while the cleaner liquid above the sediment is removed and recycled to the
scrubber.
For hot processes, a considerable amount of water is vaporized in the
scrubber and upstream equipment (e.g., quencher) which must be handled by the
fan. Although the gas volume is reduced, a large portion remains as water
vapor which results in higher horsepower requirements and in higher operating
costs. To alleviate this condition, a gas cooler can be incorporated into the
separator to cool and dehumidify the gas stream. Several types of gas coolers
are used for this purpose; one type employs spray banks of cooling water fol-
lowed by impingement baffles while a second type utilizes flooded plates or
trays with either perforated holes or bubble caps to permit passage of the gas
stream through the cooling water bath. Several plates or trays can be used
in sequential stages to provide the necessary cooling and contact time.
2. Process Applicability
Scrubber selection for particulate emission control depends upon the par-
ticulate loading, particle size distribution, and required removal efficiency.
Generally, venturi scrubbers are utilized in those process areas where wet
collection is necessary or appears desirable - after consideration of such
factors as capital investment, power requirements, slurry disposal, and heat
transfer or absorption duties. Table A14-1 gives typical values of pressure
drops for selected processes and equipment where venturi scrubbers are
utilized for particulate control.
A14-5
-------
Appendix A14
Venturi Scrubber
TABLE A14-1. TYPICAL PRESSURE DROPS FOR VENTDRI SCRUBBING SYSTEMS (3)
Process Pressure Drop, kPa
Basic Orygen furnaces 10.0-15.0
Brick Manufacturing 0.7-8.7
Clay refractories 2.7
Coal-fired Boilers 3.7
Detergent Manufacturers 2.5-10.0
Ferroalloy plants 10.0-19.9
Glass Manufacturing 16.2
Gray iron foundaries 6.2—14.9
Kraft recovery furnaces 3.7-7.47
Lime kilns 3.0-10.0
Petroleum catalytic cracking 10.0
Phosphate fertilizer 3.7-7.47
Phosphate rock crushing 2.5-5.0
Secondary aluminum 7.5
3. Process Performance
Collection performance of venturi scrubbers is affected by the following
parameters: 1) pressure drop, 2) liquid-to-gas ratio (L/G), 3) gas velocity,
and 4) particle size distribution.
Particle cut diameter is the diameter of the particle that it will col-
lect at 50 percent efficiency. Particle cut diameter is a frequently used
parameter for describing the particle collection efficiency of venturi scrub-
bers. One reason for this is because plots of collection efficiency versus
particle diameter tend to be rather steep in the region where inertial impac-
tion is the predominant collection mechanism. A plot of cut diameter versus
pressure drop for a gas—atomized spray scrubber is provided in Figure A14-3.
A14-6
-------
OJ
iJ
e
•r-t
Q
3
CJ
0.5
0.2
0.1
0.2
0. 5
10
20
Gas-Phase Pressure Drop, kPa
Figure A14-3. Cut diameter for gas-atomized spray scrubbers (4)
A14-7
-------
Appendix A14
Venturi Scrubber
This plot is based on experimental data from large Venturis, orifices, and rod-
type units, plus mathematical models. A plot of particle size vs pressure
drop has been illustrated earlier in Figure A14-2.
A relationship between pressure drop and L/G ratios for Venturis has
been developed (4) assuming that all energy from the imposed pressure drop is
used to accelerate the liquid droplets to the throat velocity of the gas:
- > \
AP = 10 (D ) _ (3)
G Q
G
where:
AP is the pressure drop, kPa,
U is the throat velocity of the gas, m/sec, and
Q./Q-, is the L/G ratio, L/m3 .
L G
An alternative approach by Hesketh (2) indicates that the pressure drop for
Venturis is proportional to D 2 and (GL/Q ) 0.78. Hence, for a specified
gas velocity, the pressure drop across Venturis is approximately proportional
to the L/G ratio.
The relationship between 1iquid-to—gas ratio and cut diameter for
Venturis has been plotted (4) with a liquid density of 1000 kg/m3 and a gas
_j
viscosity of 1.8 x 10 Pa«s (equal to the viscosity of air at one atmosphere
and 298 K). Figure A14-4 presents plots of liquid-to-gas-ratio versus cut
diameter with gas velocity and pressure drop as parameters. The factor f, the
ratio of relative velocity between gas and liquid droplet to gas velocity, is
approximately 0.25 for fly ash and hydrophobic aerosols and significantly
A1A-8
-------
Appendix A14
Venturi Scrubber
higher for hydrophilic (water attracted) aerosols. This figure may be used to
determine approximate operating conditions once the required cut diameter has
been estimated. Care must be used to ensure that adequate liquid is supplied
to provide good gas sweeping; a minimum L/G ratio of approximately 1 L/m3 is
recommended.
It has been shown that the pressure drop across a venturi is proportional
to the square of gas velocity and directly proportional to the L/G ratio.
Therefore, within limits, increasing gas velocity will result in increasing
pressure drop and decreasing the device cut diameter, other things being
equal. Velocity data are not available for survey sites equipped with
Venturis, although typical gas velocities employed commercially are 30 to 120
m/s. The low end of this range, 30 to 45 m/s, is typical of power plant
applications while the upper end of the range has been applied to lime kilns
and blast furnaces.
3.0
2.0
1.0
3
CJ
0.5
0.1
t — \J *
relative gas-particle velocity
•
gas velocity
I I I I I I I 111
0.2
1.0
4.0
Liquid-to-Gas Ratio, L/m3
Figure A14-4. Venturi cut diameter vs liquid-to-gas ratio (4)
A14-9
-------
Appendix A14
Venturi Scrubber
4. Secondary Waste Generation
Venturi scrubbers produce a significant quantity of liquid waste which
may be discharged to a settling pond or piped to a water treatment plant after
solids removal. The quantities of wastewater discharged are difficult to
predict since systems use different L/G ratios as well as different degrees
of recirculation. A venturi scrubber on a pulverized coal boiler operating at
an L/G ratio of 0.9 L/mJ with no recircnlation will discharge about 2000
L/min. Usually this discharge is pumped to a settling pond where the fly ash
settles to the bottom and the liquid is either discharged, evaporated, or
recycled. Pond liners may be used to prevent leaching of any metals or chemi-
cals into the soil and surrounding water table. Although intrusion upon a
local water body or supply is always possible, good operating procedures can
minimize this potential impact. Since the properties of pond discharge waters
are dependent upon the process particulate and gas composition, it is not
possible to specify average values.
5. Process Reliability
Maintenance is critical in venturi scrubbing systems when corrosive gases
(such as sulfur oxides) are present. In addition, handling of large amounts
of water containing potentially corrosive or abrasive materials causes more
problems than those found in dry systems. Hence, frequent and thorough
inspection of equipment is essential for reliable operation.
6. Process Economics
Costs of venturi scrubbers depend upon: 1) volumetric flow, 2) oper-
ating pressure, and 3) materials of construction. For venturi scrubbers
constructed of 3.2 mm carbon steel and gas capacities up to about 5700 actual
A14-10
-------
Appendix A14
Venturi Scrubber
m3/min, venturi scrubber costs can be estimated with the following equation
(5):
C = 7117 + 14.4 V- 0.0011 V2 (4)
where:
C is the venturi scrubber cost, $10* and
V is the inlet gas flow rate, actual m3/min.
These costs are flange-to-flange costs (December 1977 dollars) and include the
venturi, elbow, separator, pumps, and controls. Correlations for adjusting
these costs for different vessel thicknesses, materials of construction, and
internal coolers are presented in Reference 5.
7. References
Calvert, S. How to Choose a Particulate Scrubber. Chemical Engineering.
84(18): 54-68, August 29, 1977.
Hesketh, H.E. Fine Particle Collection Efficiency Related to Pressure
Drop, Scrubbant and Particle Properties, and Contact Mechanism. Journ.
Air Poll. Control Assoc. 24(10): 939-942, October 1974.
Capital and Operating Costs of Selected Air Pollution Control Systems.
Prepared by Gard, Inc. EPA-450/5-80-002, December 1978.
Calvert, S. , J. Goldschmid, D. Leith, and D. Metha. Wet Scrubber
System Study, Volume I - Scrubber Handbook. EPA-R2-72-1180, Report
prepared by A.P.T., Inc. for the U.S. EPA, August 1972.
Vatavik, W.M. and R.B. Neveril. Part X: Estimating Size and Cost of
Venturi Scrubbers. Chemical Engineering, November 30, 1981, pp 93-96.
A14-11
-------
-------
APPENDIX A15
FLARES
1. Process Description
Flares are special burners used to dispose of relatively nigh volume
waste streams consisting of hydrocarbons and other combustible gases. Combus-
tion of these gases in the flare is intended to destroy most of the objection-
able components present. Basically, a flare system consists of piping to col-
lect the gases, devices to remove entrained liquid, and a burner open to the
atmosphere. No provision is made to treat the combustion products or to re-
cover heat. An ignition source is provided by a continuous pilot or pilots.
Depending on the composition of the waste gas being flared, auxiliary fuel
could be required to sustain a stable flame. Combustion air is supplied from
the surrounding atmosphere.
The elevated flare is the most common type of flare system in use today.
In this flare, gas is discharged without substantial premizing and ignited
and burned at the point of discharge. Combustion of the discharged gases
takes place in the ambient air by means of a diffusion flame. This type of
combustion often results in an insufficient supply of air and thus a smokey
flame. A smokeless flame can be obtained when an adequate amount of combus-
tion air is mixed sufficiently with the gas so that it burns completely.
Smokeless burning is usually accomolished by injecting steam into the flame.
Elevated flares can be free standing, guyed, or structurally supported by a
derrick. Free standing flares provide ideal structural support. However, for
high standing units, fabrication costs and the nature of the soil (for founda-
tions) have to be carefully evaluated. Guyed units require considerable
amounts of land for the wires to be spread apart.
A second common type of flare is the ground flare, which consists of a
burner and auxiliaries located at or near ground level. The burner may be
constructed with or without shielding but it must allow for the free escape of
the flame and combustion products. The advantage of ground flares is the
capability of shielding the flame. Compared to elevated flares, they either
A15-1
-------
Appendix A15
Flares
require more land (if unshielded) or the burners, controls, and shielding may
be more expensive than a stack. Also, if the pilot ignition system fails, the
ground flare cannot disperse the gases as well as an elevated flare.
A third system which has been recently developed and is being employed
with increasing frequency, particularly where noise, luminosity, and smoke
formation are severely criticized by local residents, is an enclosed "low-
level* ground flare used in conjunction with an elevated flare. The system
has sometimes been referred to as the controlled combustor/low-level flare
system.
The elements of a ground flare consist of the stack, burner seal, pilot,
ignition system, and controls. Many different configurations are available
with differences due primarily to the method of location of the air or steam
injection and the design of the burners. Included in the stack of most flares
is a seal to prevent air entering the stack due to wind or the thermal con-
traction of stack gases, and to reduce the amount of purge gas that may be
needed to expel combustibles when the flared gas is lighter than air. Water
seals may also be used both to seal and to direct the gas flow to alternate
flares. In addition, a knock-out drum is provided to separate any liquid that
may be entrained in the gas stream. The knock-out drum and water seal are
usually incorporated into the same vessel.
Ground flares may also consist of multiple burners enclosed within a
refractory shell as the recently developed "low-level/controlled combustor*
flares. The essential purpose of a low-level flare is complete concealment of
the flare flame as well as smokeless burning at a low noise level. The feed
gases are connected by a manifold to a series of burner heads which discharge
the gas into a refractory enclosure. Mixing of the gas and air is accom-
plished by a series of multi-jet nozzles. Combustion air is provided by the
natural draft of the enclosure. Smokeless burning is obtained with little or
A15-2
-------
Appendix A1S
Flares
no steam because of the turbulence and temperature of the burning zone due to
the natural draft and the enclosure. The size of the enclosure depends upon
the capacity of the flare but can be quite large. Figure A15-1 illustrates
most of the major components of an enclosed-type ground flare.
Main Combustion
Chamber
Flare Pilots
With Thermocouples
Acoustical/Wind-Fence
Second-Stage Burners
First-Stage Burner
and Combustion Chamber
Figure A15-1. Typical ground flare
A15-3
-------
Appendix A15
Flares
Figure A15-2 depicts an installation that employs a low-level flare sys-
tem. As material is released into the various laterals leading to the flare
system through the safety and pressure control devices, an increased pressure
in the flare header results. When the pressure increases to that equivalent
to the hydraulic seal leg in the low rate seal drums, the seal is blown and
the gases flow to the low rate burners in the ground flare (controlled cora-
bustor). A liquid knock-out drum (or drums) upstream of the low pressure seal
drum allows for a disengagement of entrained liquid which is returned to the
process. If the flow through the flare header increases, increased back pres-
sure results which, in turn, blows the water seal in the high rate seal drum
and the high rate burners in the ground flare are activated. When the capa-
city of the ground flare is exceeded, the elevated flare water seal is blown
and allows gases to flow to the elevated flare.
For detailed information regarding the variations in flare structural
design, EPA publication AP-40 (1) should be consulted.
2. Process Applicability
There are three main considerations in deciding whether to flare a waste
gas. These are: 1) the variability of the flow of the waste stream, 2) the
expected maximum volume of the stream to be flared, and 3) the heat content
of the waste stream.
High variability of flow of the waste stream is probably the most impor-
tant factor. A flare is designed to operate for practically an infinite "turn-
down" range of flows. Alternate waste gas disposal systems (such as catalytic
or thermal incinerators) need adequate control of the flow of waste gas and
can only be used for continuous or at least fairly continuous gas flows.
A15-4
-------
FLUE
FUEL GAS
AIR
Ln
I
Ul
VENT PURGE
COMBUSTOR
LOW RATE
BURNERS
HIGH RATE
BURNERS
PLANT FLARE
HEADER
CONTROLLED
COMBUSTOR
ELEVATED
FLARE
ELEVATED FLARE
SEAL DRUM
SEAL
WATER
LIQUID
KNOCKOUT
DRUM
SEAL
WATER
SEAL
WATER
HIGH RATE SEAL
DRUM FOR
COMBUSTOR
LOW RATE SEAL
DRUM FOR
COMBUSTOR
STEAM
PURGE
GAS
•CONDENSATE
BACK TO
PROCESS
Figure A15-2. Low-level flare system
-------
Appendix A15
Flares
The volume of the waste stream to be disposed of is also an important
factor. With very large volumes of gas, thermal or catalytic incinerators
become impractical due to the size of the equipment needed. However, the
capacity of an elevated flare can be increased easily by increasing the dia-
neter of the stack. A typical small flare with a 10-cm diameter stack has a
capacity of 0.24 mj/s. A normal refinery flare with a capacity of 40 mj/s
would need only a 0.91-m diameter flare stack.
The heat content of a waste gas falls into two classes, depending on
whether or not the gases are capable of maintaining their own combustion. In
general, a waste gas with a heating value greater than 7.5 MJ/m» (lower heat-
ing value) can be flared successfully. It may be necessary to enrich the
waste gas by injecting a gas with a high heating value if the heating value of
the waste gas is below 7.5 MJ/m». The addition of such a rich gas is called
endothermic flaring. Gases with a heating value as low as 2.2 MJ/m3 have been
flared but at a significant fuel demand.
Flares are primarily used in the oil and petrochemical industry from ini-
tial field production through transportation, storage, refining, processing,
and distribution. Flares are seen on offshore drilling and production plat-
forms, at LPG and LNG terminals, along hydrocarbon pipelines, in petroleum
refineries, at underground storage terminals, and at petrochemical plants.
However, their use and importance is not limited to those applications.
Flares are used for anaerobic sewage digesters, coal gasification, rocket
engine testing, nuclear power plants, sodium/water heat exchangers, steel pro-
duction, heavy water plants, and ammonia-fertilizer terminals.
3. Process Performance
The successful performance of a flare system depends upon achieving com-
plete combustion in the burning zone of the flare. This is a function of
A15-6
-------
Appendix A15
Flares
the quantity and distribution of oxygen in the combustion zone and the
type of gas being burned.
The most critical determinant of complete combustion is the amount and
distribution of oxygen in the burning zone. It is essential that a stoichio-
metric quantity of air be evenly distributed in the primary mixing zone. This
primary air must be well mixed with the gas prior to flame ignition, or incom-
plete oxidation will take place. The remaining air required to complete the
combustion process is induced into the flame through aspiration and thermal
draft effects.
In order to get good mixing of oxygen or air with a combustible material,
an expenditure of energy is required. This energy may be provided by the
flare gas stream through pressure reduction and/or thermal draft. It may also
be provided from an external source, such as steam injection, power gas as-
sist, or a blower fan. The most commonly used method is steam injection.
This promotes combustion in five ways: 1) the injection of the steam can be
used to aspirate air, 2) the injection of steam can provide turbulence to aid
in the mixing of the fuel and air, 3) the steam reacts with the fuel to form
oxygenated compounds that burn readily, 4) the steam lowers the temperature of
the unbnrnt gases, retarding thermal cracking, and 5) the steam reduces the
partial pressure of the fuel which in turn reduces the formation of carbon in
diffusion flames.
The composition of the gas being burned affects the rate of burning, the
flame temperature, flame buoyancy, and the mixing of air into the flame.
While the flared gas caloric value is considered to be a primary parameter
in determining flare efficiency (mainly through its effect on flame tempera-
ture), the gas composition itself, as characterized by the hydrogen-to-carbon
ratio, is important in the formation of soot.
A15-7
-------
Appendix A15
Flares
The efficiency of a combustion flare is difficult to measure, and con-
sequently estimates of pollutant emission indices vary. Table A15-1 sum-
marizes flare efficiencies from the most recent studies. Although these
studies are valuable contributions to the knowledge of flare performance, none
allow an accurate determination of pollutant emissions nor do they provide ad-
equate information on the effects of either scale or flared gas composition.
Part of the reason is that measurements are difficult to make on large ele-
vated flares. In addition, the data are extremely difficult to interpret
because of the intermittent nature of the flame, limited accuracy of measuring
low concentrations of incompletely burned hydrocarbons, and the difficulty of
measuring and summing all small losses from the flame. The West German gov-
ernmental regulatory agency has assumed all elevated flares to be 75 percent
efficient for the purpose of regulating flare installation and use (7).
4. Secondary Waste Generation
A secondary pollutant associated with flaring is combustion noise. To
suppress smoke formation in process flaring, it is customary to inject steam
into the flared gas, which produces noise in two ways. The first is jet noise
associated with the high velocity steam jets, and the second is a more
indirect effect of the steam injection process on the acoustics of the entire
flare system, including the stack, the liquid-seal drum in its base, and the
flare headers. Combustion noise is not generally a problem in process
flaring. However, in emergency flaring conditions when the gas discharges in
some 100 m/s the flame could be noisy. It is manifest in a subaudible vibra-
tion at low flaring rates and high steam flows, which becomes an environmental
noise problem.
A15-i
-------
TABLE A15-1. SUMMARY OF FLARE COMBUSTION EFFICIENCY ESTIMATES
Source
Ref. 2
Ref. 3
Ref. 4
Ref. 5
Ref. 6
Flare Tip Design
1.3 cm dia
Discrete holes in 5.1 cm
dia. cap
Commercial design 70 cm
dia. , steam added
Commerical design 15 cm
dia. , air assist
Commerical design 3 tips
Throughput, Flare
Flared Gas MJ/s Efficiency, %
Ethylene 0.12 - 0.62
Propane 0.09
-50% Ha plus light 14 - 52
hydrocarbons
Propane 13
Natural gas 8.2 (per tip)
>97.8
96 - 100
97 - >99
91 - 100
>99
at 10 cm dia.
-------
Appendix A15
Flares
Careful orifice design can greatly reduce the jet noise from steam
flares. However, in practice the minimum size of suppressant orifices is
limited due to potential plugging problems. Additional gains in jet noise
reduction have been made by improving the efficiency of steam usage through
the use of a highly responsive optical smoke suppressant control system
developed by the Shell Development Co. (6). This systems uses a ground level
optical sensor to detect a combined measure of flow, composition, and sup-
pressant effectiveness (6). This eliminates waste of steam due to delays in
response to gas flow cessation.
5. Process Reliability
Many critical factors must be considered and resolved in flare design.
Thermal radiation, liquid carry-over, noise, pressure drop, visible light,
condensate vacuum, and explosion hazard (resulting from air entry into the
flare stack) are important factors influencing safe operation. In addition,
operational design factors such as flame stability, positive piloting, relia-
ble ignition, effective mixing, and air entrainment for the steam, or assist-
gas injection and control of the injection medium are important to the relia-
ble operation of the flare system.
Flare systems must be operable during emergencies such as power failure,
so pneumatic control systems are becoming increasingly popular. For example,
fluidic control of steam for smokeless burning is now available. In addition,
new ignition systems using a flint to generate sparks to ignite the pilot
flame have been developed. Automatic thermocouple alarm systems which can
detect pilot flame extinguishment are available. Information received from
vendors indicates that a reliability of 99+ percent is achievable by ground
and elevated flare systems (Personal communication with R.A. Ferraro, March
1982). It has been shown that pilot lights on a hydrogen flare have provided
ignition in wind speeds of up to 134 km/h. Once the flame was ignited, winds
up to 176 km/h and simulated rain of 144 mm/hr failed to extinguish it (8).
A15-10
-------
Appendix A15
Flares
6. Process Economics (9)
The capital cost of flares depends upon the following:
• the degree of sophistication desired and the auxiliaries
selected such as knock-out drums, seals, controls, ladders,
platform, etc.,
• the basic structure and support of the flare,
• the height and capacity,
• availability of utilities such as steam and natural gas,
• variability in the composition of the waste gases, and
• the frequency of flaring.
Typical costs of elevated flares range between $30,000 to $100,000.
For the same capacity range, ground flare costs can be ten times as high.
Blower assisted flares (forced draft flares) are between two to three times
the cost of air-equivalent conventional elevated flares. Typical costs of
elevated and ground flares are given in Figure A15-3. These costs are in
December 1977 dollars. Costs vary substantially between plants due to the
degree of automation of the system, operating periods, and characteristics of
the gas stream. Additionally, flares may be used intermittently or only under
emergency conditions. Some are provided with auxiliary fuel while others
depend on the fuel value of waste gases. There are no typical ranges of
annual operating costs; however, it is estimated that less than 1 man-hour
maintenance per shift will be expended and that the major operating costs may
be in the use of steam as a smoke suppressant.
A15-11
-------
Capital Costs of Elevated Flares, $103 (December 1977)
H-
OQ
C
HI
L/l
I
n
o
en
(D
to
s
ex
OQ
i-l
O
3
a.
O
to
en
NJ
O
o
Tl
I — i
O
^ o
OQ
K)
Ul
O
00
o
o
o
O
O
00
00
O
o
o
Capital Cost of Ground Flares, $103 (December 1977)
ro
o
~~\
00
o
O
o
ho
-P-
O
ho
CO
o
LO
ho
O
U)
ON
O
C-
o
o
tt
en
OQ
ro
O
O
Ml
OJ
O
Ln
O
-------
Appendix A15
Flares
7. References
1. U.S. Environmental Protection Agency. Air Pollution Engineering Manual,
Second Edition. EPA Publication AP-40, May 1973.
2. Palmer, P.A. A Tracer Technique for Determining Efficiency of Our
Elevated Flare. E.I. du Pont de Nemours and Co., Wilmington, Delaware
(1972).
3. Lee, K.C. and G.M. Whipple. Waste Gas Hydrocarbon Combustion in a Flare.
Union Carbide Corporation, South Charleston, West Virginia (1981).
4. Siegel, K.D. Degree of Conversion of Flare Gas in Refinery High Flares.
Ph.D. Dissertation, University of Karlsruhe (German), February 1980.
5. U.S. Environmental Protection Agency. Development of Flare Emission
Measurement Methodology, Draft Report. EPA Contract No. 68-02-2682
1981.
6. Schwartz, R. and M. Keller. Environmental Factors vs. Flare Application.
CEP, September 1977.
7. Government of the Land of North Rhine/Westphalia, Reine Lift fur Morgen -
Utopie oder Vinklichkeit.
8. Brzustowski, T.A. Flaring in the Energy Industry. Prog. Energy Combust.
Sci., Vol. 2, pp 129-141. Pergamon Press, 1976.
9. U.S. Environmental Protection Agency. Capital and Operating- Costs of
Selected Air Pollution Control Systems. EPA 450/5-80-002, December 1978.
8. Personal Communications
Telephone conversation between Bob Bakshi, TRW, and Robert Ferraro, John
Zink Co., Long Beach, California, March, 1982.
A15-13
-------
-------
APPENDIX A16
THERMAL INCINERATION PROCESSES
1. Process Description
Thermal incinerators (also called direct-flame afterburners) consist of
refractory lined chambers, which may vary in cross-sectional area; one or more
burners; temperature indicator-controllers; safety equipment; and sometimes
heat-recovery equipment such as heat exchangers. The operating principle of
thermal incinerators is the oxidation of combustible compounds (e.g., hydro-
carbons, CO, H2S) from waste streams by direct combustion. In practice,
thermal oxidizers are generally used for the destruction of residual combus-
tible pollutants. The typically low concentration of combustibles in waste
gases usually requires that supplemental fuel be used.
The feed containing the gaseous pollutants enters the firing chamber of
the thermal incinerator and is intensely mixed with the fuel combustion gases.
The gas mixture, at a reaction temperature of between 840 K and 1100 K, tra-
vels through the firing chamber where the oxidation of the pollutant gases
take place. The chamber is of sufficient length to provide the retention
period necessary for the required degree of conversion. The combustion gases
are discharged to the atmosphere.
Equipment for thermal oxidation of gaseous wastes varies widely depending
upon the manufacturer, but several basic types of equipment are used. The
first is the line burner which is used with waste gases containing sufficient
oxygen for their combustion. Line burners may be no more than a gas pipe with
a number of holes which inject a fuel, such as natural gas, into the waste gas
stream at the point of ignition. The waste gas therefore passes through the
flame of the line burner and is heated to a temperature above the autoignition
temperature of the organic constituents. More sophisticated arrangements of
the line burner employ a series of baffle plates over the gas jets which
A16-1
-------
Appendix 16
Thermal Incineration
promote better contact between the flames and the waste gas (see Figure
A16-1). The line burner is usually installed in the waste gas duct or an
extension of the duct which is refractory lined or insulated beyond the point
of ignition.
Other systems utilize an external burner which can be either a natural,
forced-draft, or aspirating type. In this system the burner is usually
located at one end of a cylindrical or rectangular duct, or tunnel, and the
waste gas is passed into the duct through the burner flame or around it (see
Figure A16-2). Baffles or tangential entry of the waste gas give sufficient
turbulence to mix the high—temperature products of combustion from a conven-
tional burner with the waste gas so that a final temperature of the desired
magnitude is achieved. Such units can be constructed of refractory-lined
carbon steel or stainless steel if the temperatures are not too high, and can
be oriented either horizontally or vertically depending upon the most desir-
able physical arrangement of the system.
The third type of system is called the jet burner. Here, a high-velocity
discharge burner is fired into the throat of a refractory-lined venturi sec-
tion and the venturi action creates a slight suction capable of pulling the
waste gas into the incinerator (see Figure A16-3). Basically this is the same
principle as the tunnel-type burner arrangement except that a waste gas fan is
not required.
Generally it is more desirable to use a forced draft fan to push the
waste gas into a thermal incinerator than it is to use an induced draft type.
This is because the latter must be capable of handling high temperatures
unless some intermediate heat transfer device is used.
Heat exchangers are usually furnished as optional equipment, but in many
industrial designs they are integral units inside the combustion chamber
A16-2
-------
STACK
I
OJ
PROFILE
PLATE\
GAS SUPPLY r
MANIFOLO <1
BURNER
EXHAUST GASES
FROM PROCESS
Figure A16-2. Tunnel incinerator
BURNER
- COMBUSTION
CHAMBER
¥—fi
Figure A16-1. Line Burner
Figure A16-3. Jet incinerator
-------
Appendix A16
Thermal Incineration
shell. Heat recovery offers a way to reduce the energy consumption of incin-
erators. The simplest method is to use the hot gases exiting the incinerator
to preheat the incoming gases. Design is usually for 35 to 90 percent heat
recovery efficiency. Typical incinerator heat exchanger configurations are
shown in Figure A16-4.
1
u
Incinerator
E
k
TEx(
From ov€
xhoust Exh(
) stock to s
Fresh
air
1
From -, ^_ 1
oven -" 1
just Exhaust
tack to stock
Fresh '
<*•
:hanger j |
To oven To oven
n From oven
Scheme A
Self-Recuperative
Scheme 8
Makeup Air
Scheme C
Self-Recuperative + Makeup
Figure A16-4. Incinerator/heat exchanger configurations
For small—volume applications direct—fired thermal incinerators can be
provided as packaged units which include the basic chamber, fan, and controls.
Larger units are usually custom designed or are modifications of standard com-
ponents integrated into a complete unit. The control and instrumentation com-
plexity required by these systems will depend on the characteristics of the
process gas stream. Steady—state processes require the least control, whereas
fluctuating gas streams would require modulating controllers for the burners,
recycled gas, and other components. Most units only require basic controls
such as safety pilots, flame failure shut-offs, high temperature shut-offs (in
case of fan failure), temperature monitors, and recorders. The fans used for
these units are usually axial flow or low pressure centrifugal types since
pressure drops for the incinerator alone are usually less than 0.5 kPa. If a
heat exchanger is included, the pressure drop may increase up to 1.5 kPa
depending on the configuration of the heat exchanger and the number of passes.
A16-4
-------
Appendix A16
Thermal Incineration
Fireboxes of boilers and fixed heaters can also be used as thermal incin-
erators for combustible air contaminants. Boiler firebox conditions approxi-
mate those of a well designed thermal incinerator (afterburner), provided
there are adequate temperature, retention time, and turbulence available.
Oxidizable contaminants can be essentially converted to carbon dioxide, sulfur
dioxide, and water in boiler fireboxes.
2. Process Applicability
Thermal incinerators are widely used in various industrial applications
to control odors, smoke, hydrocarbons, and carbon monoxide. This process has
been used successfully for years as an air pollution control device to remove
combustibles from waste streams in the petroleum, surface coating, and
chemical industries. The most important control point variable is chamber
temperature, which is readily controllable by varying the fuel flow rate. The
performance of the system therefore remains quite stable and does not
deteriorate with time.
In most cases supplemental fuel is required to raise the raw gas tempera-
ture. At lower operating temperatures, incomplete oxidation may occur,
resulting in increased levels of organics and CO in the combustion flue gases.
Oxidation or partial oxidation of contaminants containing halogens could
create an extremely hazardous effluent (e.g., hydrogen chloride, phosgene).
Thermal incineration also does not oxidize/remove contaminants which are
already in the oxidized form (e.g., SOJ§ S03).
3. Process Performance
The performance of a thermal incinerator is a function of retention time,
operating temperature, flame contact, and velocity. There is no quantitative
mathematical relationship between efficiency and these variables because the
A16-5
-------
Appendix A16
Thermal Incineration
kinetics of the combustion process are complex and the flow in the thermal
incinerator is not easily defined. Tomany (1) provides an equation that
demonstrates the relationship of conversion efficiency to the pollutant
reactivity and retention period:
100-C = e~kt (1)
100
where:
C = reaction conversion, percent,
k = reaction rate constant,
t = retention period, sec.
k = reaction rate constant, sec , and
Chemical reaction rates increase with temperature as derived from the
empirical Arrhenius equation. Thus, for any desired conversion of a pollutant
to CO and HO, the incineration temperature determines the reaction rate and
equation 1 predicts the required retention period, or dwell time. Theoreti-
cally, an infinite retention period would be required for 100 percent conver-
sion.
Assuming good design, some generalizations can be made with respect to
thermal incinerator efficiency. Overall efficiency increases with 1) increas-
ing operating temperature as illustrated by Figure A16-5, 2) increasing con-
tact between the contaminated gases and the burner flame, and 3) retention
time (for retention times of less than 1 second) as illustrated by Figure
A16-5.
Organics destruction efficiency is dependent upon the incinerator design
and the inlet concentration of organic materials. No direct predictions can
be made from one design to another. Tables A16-1 and A16-2 show typical data
from tests on large and small incinerators.
A16-6
-------
H-
OQ
l-i
n>
RESIDENCE TIME, SECOND
>
M
I
I
Ui
rt O
0) (D
3 w
X! 'ft
ft) i-j
i-l C
CD O
rt rt
C H-
f| O
n> 3
£13 (B
3 i-ti
P* Ml
H-
i-f O
ro H-
t/5 ft
H- 3
CL o
CD ^
3
n {a
rt £U
H'
3 i-n
n> c
'-N O
ho rt
•^s p.
O
3
01 '-j b> lo b
-------
TABLE A16-1. TYPICAL ANALYSIS OF EMISSIONS ENTERING AND LEAVING
LARGE DIRECT-FIRED THERMAL INCINERATOR (3)
CO , ppm
CO, ppm
Organ! cs as CO , ppm
1030
In
6,300
59
1,568
Volume (dry basis), m'/s 5.64
Organics (as carbon)
Organics destruction
Includes increase of
TABLE A16-2.
CO , ppm
i
CO, ppm
Organics as CO , ppm
Volume (dry basis), n
Organics (as carbon)
Organics destruction
kg/min. 16.1
efficiency, % 85
CO across afterburner.
Temperature
K 1090 K
Out In Out
22,000 6,600 27,000
230 65 21
235a 1,591 70
5.57 5.6 5.57
2.39 16.4 0.72
96
TYPICAL ANALYSIS OF EMISSIONS ENTERING AND LEAVING
SMALL DIRECT-FIRED THERMAL INCINERATOR (3)
980 K
In
1,950
8
521
I»/S 1.1
kg/min 1 .0
efficiency, % 77
Temperature
1030 K
Out In Out
19,000 2,000 23,500
110 9 24
122* 408 33a
1.0 1.1 1.0
0.23 0.79 0.06
92
Q
Includes increase of CO across afterburner.
A16-8
-------
Appendix A16
Thermal Incineration
Tests have been conducted on several boilers that have been used as
thermal incinerators. Table A16-3 summarizes these test results and shows the
apparent efficiencies of boilers in controlling organic acids and aldehydes.
A list of expected residual concentrations for various components when
incinerated is provided below:
Boiler Incinerator
Constituent Effluent Effluent
Reduced sulfur species 5 ppm 5 ppm
Volatile organic compounds (VOC) 100 ppm 100 ppm
CO 100 ppm 300 ppm
HCN 1 ppm 1 ppm
NHj 0.5 ppm 0.5 ppm
The basis for the boiler performance estimates for incineration control of
volatile organic compounds (VOC) and carbon monoxide emissions is presented in
Table A16-4. The data show effluent CO levels of 300 ppm or less and VOC
levels of 100 ppm or less for a variety of incineration sources.
4. Secondary Waste Generation
With thermal incineration, the pollutants are usually converted to S02,
COJt HjO, and N2> which are discharged to the atmosphere. No secondary waste
is generated.
5. Process Reliability
Data on process reliability in the form of stream factors are not gen-
erally available. This is largely due to the fact that thermal incineration
of waste gases is performed on an intermittent basis. Operating problems may
A16-9
-------
TABLE A16-3. TEST DATA ON BOILERS USED AS THERMAL INCINERATORS (3)
317 k« Boiler
liter-Tube Type,
Git-Fired
Tiro 200 kl Boilers
Couon Stack
liter-Tube Type,
Git-Fired
ISO kl Boiler*
liter-Tube Type,
Git-Fired
Two 84 k* Boileri.
Couon Stick
Locomotive Type,
Git-Fired
112 kl Boiler
H8T Type.
On-Fired
i
M
C
Volume of guest «*/»
Sttck
Boiler inlet
Orginic iclds, kg/hr
Inlet
Outlet
Efficiency, %d
Aldehydes, kg/hr
4.11
0.76
0.68
0.25
60
4.86
1.38
1.22
0.29
78
2.22
1.13
1.00
0.64
36
1.79
0.15
0.05
0
100
I" He it smokehoute effluent wit idmltted into boiler firebox through the nnltijet burner.
Heit taokehonie effluent wit admitted into boiler firebox through diffueer looted it front of firebox floor.
CBenderlng cooker effluent wit idaitted into boiler firebox through diffuter located it reir of firebox floor.
Two tettt were run. The ute of thii boiler it in ifterburner hn been ditcontinued, prinirily beciuie the
• inlnuB firing rite of the boiler wit insufficient to incinerite lir contuinintt.
Efficiency shown it Appirent Efficiency. Boilert could not be tested unless sir contuinintt were vented to it.
1.60
0.22
0.16
0.06
60
1.70
0.35
0.20
0.17
14
Inlet
Outlet
Efficiency, %d
0.10
0.04
59
0.18
0.18
0
0.18
0.14
23
0.01
0
100
0.005
0.04
0
0.01
0.08
0
-------
TABLE A16-4. SUMMAEY OF PERFORMANCE DATA FOR HYDROCARBON AND
CARBON MONOXIDE FOR COMBUSTION/INCINERATION
CO
Emissions, Total Hydrocarbons,
Facility ppmv (2% 0 ) ppmv (2% 0 )
________ _ i i _
ft
Bituminous coal-fired utilities (4)
• pulverized, dry bottom 50 23
• pulverized, wet bottom 253 23
• cyclone 241 49
Oil-fired utilities* (4) 175 24
Gas-fired utilities* (4) 109 14
Gas-fired commercial/ insti tutional
combustion source sa (5) 26 17
Residual and distillate oil-fired
commercial/ insti tutional
combustion sources* (5) 59 31
Bituminous coal-fired commercial/
institutional combustion sources
(pulverized dry bottom)* (5) 59 31
Carbon black CO boiler (6) 124b 77°
Carbon black CO boiler (6) 49b
a
Industrial gas-fired boil ers (7) 26 6
Industrial oil-fired boilers* (7) 47 15
Industrial bituminous coal-fired
boilers (pulverized, dry
bottom)4 (7) 59 31
Tabulated emissions data and averages of data from several emissions
.sources
Flue gas 0 and CO data are not available
1 2
Non-methane hydrocarbons
A16-11
-------
Appendix A16
Thermal Incineration
occur in the burner control system, especially when the unit is operated
intermittently. Manufacturers say that an on-stream reliability in excess of
98 percent is achievable (Personal communication with J. Kirkland, March 9,
1982).
6. Process Economics
Costs for thermal incinerators are based on the inlet gas flow rate,
design residence time, package or custom design, and whether a heat exchanger
is used for heat recovery. The basic cost of the incinerators includes the
incinerator and base, fan, motor, starter, integral ductwork, controls,
instrumentation, and heat exchanger (where applicable). Equipment costs
increase with increased retention time due to the longer and larger retention
chambers.
A minimum of auxiliary equipment is required for thermal incinerators
since the units are normally self-contained. Usually some duct work, a fan,
and a capture device are required to transport the process gas stream from the
source to the control device if the distance is appreciable. A separate fan,
however, is supplied with the incinerator to ensure proper distribution and
mixing of the gases in the incinerator. An exhaust stack is also required to
disperse the exhaust gases above the level of the surrounding buildings.
Costs for thermal incinerators (December 1976 dollars) are presented in
Figure A16-6. The installed costs shown are for incinerators designed for
1) no heat recovery, 2) primary heat recovery, and 3) primary and secondary
heat recovery. The average installed cost of incinerators with primary heat
recovery is roughly 25 to 30 percent greater than incinerators without heat
recovery. Incinerators with secondary heat recovery have roughly a 50 to 60
percent higher installed cost than incinerators with only primary heat
recovery.
A16-12
-------
I
M
OJ
0)
O
01
O-
TO
O
280
240
200
160
120
80
0.25
0.5
0.75
1.0
1.25
Process Flow, 103 Nm3/min (approximate)
Figure A16-6. Capital costs for direct flame thermal incinerators (8)
-------
Appendix A16
Thermal Incineration
Annualized costs for thermal incinerators (5840 hours of operation per
year) are given in Figure A16-7. These costs are for incineration of two
different concentrations of combustibles (~100 ppmv and 25 percent of the
lower explosion limit concentration). Additional assumptions used to develop
these costs are:
• Incinerators designed for both oil and natural gas operation,
• Costs based on outdoor location,
• Fuel costs at il.60/GJ,
• Electricity at io.03 kW-hr,
• Depreciation and interest equal to 16 percent of capital
investment,
• Inlet stream temperature of 295 K, and
• Incinerators with a capability of attaining 1100 K with
0.5 second residence time.
It should be noted that designs and costs will change drastically for
different residence times. Longer residence times require longer residence
chambers and consequently, higher costs. For more detailed information on
thermal incinerator costs, reference 9 should be consulted.
A16-14
-------
E
0)
u
0)
Q
w
o
c
800
700
600 -
500 -
A - No heat recovery
B - With primary heat recovery
C - With primary and secondary
— heat recovery
LEL - Lower explosion limit
400 —
300 -
200
100 -
0 0.25 0.50 0.75 1.0 1.25 1.5
Flow, 103 Mm3/m
Figure A16-7. Annual costs of direct flame thermal incinerators (8)
A16-15
-------
Appendix A16
Thermal Incineration
7. References
1. Tomany, J.P. Air Pollution: the Emissions, the Regulations, and the
Controls. American Elsevier Publishing Co., Inc., 1975.
2. Cegielski, J.M. Hazardous Waste Disposal by Thermal Oxidation. John
Zink Company, 3rd Printing.
3. D.S. Environmental Protection Agency. Air Pollution Engineering Manual,
Second Edition. AP-40, May 1973.
4. U.S. Environmental Protection Agency. Emissions Assessment of
Conventional Stationary Combustion Systems; Volume VI. External
Combustion Sources for Electricity Generation, November 1980.
5. U.S. Environmental Protection Agency. Emissions Assessment of
Conventional Stationary Combustion Systems; Volume IV. Commercial/
Institution Combustion Sources. October 1980.
6. U.S. Environmental Protection Agency. Control of Emissions from Lurgi
Coal Gasification Plants. EPA-450/2-FG-012, March 1978.
7. U.S. Environmental Protection Agency. Emissions Assessment of
Conventional Stationary Combustion Systems; Volume V. Industrial
Combustion Sources, October 1980.
8. Taback, J.J., T.W. Sonnichsen, N. Brunetz, andJ.L. Stredler. Control of
Hydrocarbon Emissions from Stationary Sources in the California South
Coast Air Basin. Final Report, Volume 1 prepared by KVB, Inc., for
California Air Resources Board, ARB Contract No. 5-1323, June 1978.
9. Rolke, R.W., et al. Afterburner Systems Study. EPA EHSD 71-3, August
1972.
8. Personal Communications
Kirkland, J., Hirt Combustion Engineers, Montebello, California to TRW
Environmental Division. March 9, 1982.
A16-16
-------
APPENDIX A17
CATALYTIC INCINERATION PROCESSES
1. Process Description
Catalytic combustors are devices which are used to dispose of relatively
clean gaseous wastes containing low concentrations of combustible materials.
Combustible components of the waste are destroyed by catalytic oxidation over
a catalyst bed at temperatures significantly lower than those required for non-
catalytic thermal oxidation.
A typical flow diagram for a catalytic incineration system is shown in
Figure A17-1. Contaminated gas is first fed into a preheat zone. Here the
gas is heated to the required temperature and some of the contaminants are
destroyed by thermal oxidation. The gas then flows through the catalytic
element where the remaining contaminants are destroyed by catalytic oxidation.
This element may be a fixed or fluidized bed of catalyst material.
Several variations of the process configuration shown in Figure A17-1 are
possible. In one of these, the fan is located between the preheat burners and
the catalyst bed. Placement of the fan in this position provides better gas
mixing and a more even distribution of gases. It also enables the use of an
atmospheric burner for preheat, since a negative pressure exists in the pre-
heat section. (Otherwise a premix gas burner is used in the preheat section
because of the positive pressure existing in the chamber.) Fans used in this
variation must be constructed of materials that can withstand the maximum
temperature of the gas being handled.
Process variations for heat recovery are possible. When a sufficient
temperature rise occurs across the catalyst bed, it is often desirable to re-
cycle a portion of the hot purified gas and combine it with low temperature
influent. This reduces preheat requirements and, therefore, fuel usage.
A17-1
-------
i
N)
TEMPERATURE
PREHEATER BURNERS
TEMPERATURE
CONTROL
LEGEND:
1. Contaminated Gas
2. Purified Gas
Figure A17-1. Catalytic incinerator with fixed catalyst bed
-------
Appendix A17
Catalytic Incineration
Another way to reduce the preheat required, even when the combustible content
of the waste (and the resulting temperature rise across the catalyst) is rela-
tively low, is to incorporate a heat exchanger. The influent is passed
through the cold side of the heat exchanger and then to the catalyst bed. The
effluent from the bed is passed through the hot side of the exchanger and
recycle may also be used (1). Sufficient waste heat may be available in the
recycle stream to produce steam. In some cases the waste heat recovered by
these methods can eliminate the need for preheat burners (except during
start-up).
Various catalysts have been used to promote low temperature oxidation of
combustible waste gas components. Metals such as platinum, palladium, and
rhodium as well as copper chromite and oxides of copper, chromium, manganese,
nickel, and cobalt have been used successfully. While the catalyst can be
selected to maximize combustion efficiency for a particular waste stream or
component, in general, the goal of commercial manufacturers has been to make
catalysts available which are effective for a wide range of materials over an
extended period with minimal maintenance and replacement.
Catalytic incineration equipment is commercially available from a number
of suppliers. Systems generally incorporate proprietary features and catalyst
formal ations. Catalytic incineration systems have a number of industrial
applications, particularly in food processing, coating and other solvent hand-
ling processes, and chemical manufacturing (2,3). They also find application
in the petrochemical and synthetic fuels industries. Catalytic oxidation
units have been used to control emissions from coke ovens and catalytic
crackers and have been employed in a large number of miscellaneous industrial
gas cleaning operations (4). A catalytic incineration system is being
installed to dispose of Stretford offgases at the Tennessee Eastman coal gas-
ification facility in Kingsport, Tennessee.
A17-3
-------
Appendix A17
Catalytic Incineration
2. Process Applicability
Catalytic incineration may be feasible for treatment of certain gaseous
waste streams from synthetic fuels plants. The actual applicability will be
determined by catalyst and waste stream characteristics.
Waste streams suitable for catalytic incineration contain combustibles
such as hydrocarbons, CO, Ha, and possibly small amounts of reduced sulfur
species. The normal effective range for catalytic oxidation extends from a
few parts per million (ppm) of combustibles up to a heating value of 0.74
MJ/m* (2). The concentration of combustibles in the waste should always be
below the lower flammability limit to guard against explosion and fire, and is
generally less than 25 percent of the lower explosive limit. Sufficient oxy-
gen must be present in the waste (or must be added in the incinerator) to
ensure oxidation of combustibles.
The compatibility of the waste stream with the catalyst is of critical
importance. All catalysts are susceptible to poisoning agents, activity sup-
pressants, and fouling agents. For example, most platinum-family catalysts
are permanently poisoned by heavy metals, phosphates, and arsenic compounds
and fouled by alumina and silica dusts, iron oxides, silicones, organic
liquids, and tars. Their activity is temporarily suppressed by halogens and
sulfur compounds (2). While catalyst manufacturers have developed catalysts
which can resist some of the effects of poisons and suppressants (especially
non-precious metal formulations), fouling of the catalyst with particulate
matter from the waste gas can be a major problem. Therefore, streams with
high particulate loadings are unsuitable for treatment by catalytic inciner-
ation.
A17-4
-------
Appendix A17
Catalytic Incineration
3. Process Performance
The operating efficiency (percent of the waste organics combusted to car-
bon dioxide and water) of a catalytic system is strongly dependent upon the
catalyst temperature. A tabulation of removal efficiencies of various solvent
vapors as a function of catalyst temperature is presented in Table A17-1 (5).
Increased catalyst temperatures generally result in increased removal effi-
ciencies. However, most catalysts are permanently deactivated at temperatures
above 920 K. Thus, the temperature is considered a maximum catalyst outlet
temperature.
Space velocity is also an important determinant of efficiency. Space
velocity is generally expressed as a gas hourly space velocity (GHSV) in terms
of unit volumetric flow rate per unit volume of catalyst. Therefore, GHSV is
related to residence time. Values of the GHSV are typically around 20,000
hr , but may range from about 5,000 to 50,000 hr"1. Hydrocarbon destruction
efficiencies would be expected to increase as GHSV decreases.
The concentration of combustibles and oxygen in the waste gas has a
marked effect on combustion efficiency. It is desirable to operate with as
high a concentration of combustible (but still below the lower flammability
limit) as possible since conversion efficiencies are generally proportional to
combustible compound concentrations when other operating variables are held
constant. It has also been found that increased oxygen concentration results
in increased efficiency when other parameters are held constant. There is,
however, a trade-off, because as oxygen concentration is increased in a gas
stream (through air injection), the contaminant concentration is decreased.
That is, the efficiency of the system will be increased by oxygen addition
while simultaneously being decreased by contaminant dilution. There is
generally an optimum which must be determined for individual disposal systems.
A17-5
-------
TABLE A17-1. CATALYTIC OXIDATION OF SOLVENT VAPORS EVOLVED FROM A COATING OVEN (5)
>
I—
I
Coating Type
Vinyl
Vinyl
Epoxy
Phenolic
Oleoreainona
Alkyd
Solvent Vapor Average
Quantity of Quantity Oven Catalyat
Coating, Evolved, Temperature, Temperature,
L/hr Evolved Solvent kg/hr K E
72 lylol and laopborone 34 4)0 700
770
840
920
163 Methyl iaobutyl 123 440 720
So tone 750
770
801
70 Xylol and Butyl 39 450 700
Celloaolve 755
70 Mineral Spirit* 40 485 660
66 Mineral Spirita 25 490 700
770
840
920
30 Mineral Spirita 14 415 640
645
700
770
840
920
«
Solvent
KeBOval
28
54
77
93
77
79
85
88
79
91
65
80
89
94
95
41
52
80
89
94
95
The catalyat diacharge temperature wat held conatant for each teat period and anperficial gaa velocitiea were
constant within 15 percent for all data polnta.
-------
Appendix A17
Catalytic Incineration
Catalyst properties also affect process performance. In addition to the
problems with poisons, suppressants, and fouling agents noted above, all cat-
alysts suffer from a gradual loss of activity with use. This loss of activity
is a result of the combined effects of contamination, thermal stresses, and
mechanical forces (e.g., erosion and abrasion) to which the catalyst is
subj ected.
Emissions from a number of catalytic incineration systems recently re-
ported by Retallick (6) are presented in Table A17-2. Data were also supplied
by UOP (Personal communication with J. Brewer, January 22, 1982) for one of
their fixed bed precious metal catalytic incinerators. This unit burned a
waste gas consisting of a dilute stream of cyclohexanone, toluene, xylene, and
other organics (concentration not specified). Inlet temperature was 590 K
and GHSV was about 40,000 to 45,000 hr x. Destruction of 97 percent of the
organics was reported, and organic emissions were approximately 50-100 ppm
(expressed as methane).
Recent tests were also performed by EPA (7) on the exhaust of a Formox
formaldehyde plant. The catalyst used was a precious metal formula evenly
distributed over a high-surface area aluminum oxide support material. The
inlet contained 3000-8000 ppm of CO, 100-900 ppm of methanol, 2500-4500 ppm of
dimethyl ether, 50-500 ppm of formaldehyde, 90-92 percent of nitrogen, and 6-7
percent of oxygen. The overall removal efficiency of the catalytic incinera-
tion system ranged from 97.9 to 98.5 percent. Test data did not indicate any
trend which would predict a maximum catalyst life. A minimum of 3 to 5 years
was indicated. A brief test data summary is provided in Table A17-3.
Theoretically, destruction efficiency of carbon monoxide in catalytic
incinerators is expected to be greater than for hydrocarbons because the
initiation temperature for catalytic combustion of CO is less than for hydro-
carbons (Personal communications with J. Brewer, January 22, 1982 and
A17-7
-------
TABLE A17-2. EMISSIONS OF VARIOUS HYDROCARBONS FROM INDUSTRIAL CATALYTIC
INCINERATION (6)
Incinerator Inlet
Temperature, Concentration (ppm)
Industry Flue-Gas Components K Inlet Outlet
Rotogravure Butyl acetate 530-700
printing Ethyl acetate
Isopropanol
n-Hexane
Toluene
Magnetic Methylethyl ketone 530-700
printing Methylisobutyl ketone
Toluene
Metal Butanol 530-700
printing Methylisobutyl ketone
Toluene
Xylene
Synthetic Acetone 475-590
Formalin
Methanol
Phenol
Toluene
300-1000
7-10
1200-2500
500-1500
1200-1500
8-10
7-10
1-20
TABLE A17-3.
CATALYTIC INCINERATION REMOVAL EFFICIENCY
FROM A FORMOX FORMALDEHYDE PLANT (7)
Removal Efficiencv (%)
Test Date
11/30/79
12/18/79
1/11/80
3/13/80
3/18/80
3/25/80
Flow Rate,
m»/hr
4945
5149
2957
—
4124
4590
Outlet
Temp, K
813
705
813
—
778
783
CO
98.9
99.1
99.3
99.0
99.0
99.0
CH,OH
97.2
98.6
99.2
93.2
98.9
97.1
CHjOCH,
96.8
97.2
96.6
96.6
96.7
96.5
HCHO
95.0
98.6
97.0
98.6
99.1
98.7
Overall
97.9
98.5
98.3
97.9
98.3
98.0
A17-8
-------
Appendix A17
Catalytic Incineration
B. Hernquist, January 15, 1982). A direct relationship is believed to exist
between the autoignition temperature of an organic vapor and the catalytic
initiation temperature (5). In general, small molecules with saturated bonds
(particularly methane) and aromatic compounds are harder to destroy catalyt-
ically than long straight-chain alkanes or unsaturated hydrocarbons (Personal
communication with J. Brewer, January 22, 1982). However, a reliable rule of
thumb does not exist for predicting the relative destruction efficiencies of
complex organic molecules.
The behavior of reduced sulfur compounds (such as H2S) during catalytic
oxidation is a function of temperature and the catalyst material. At low
temperatures, sulfur compounds are oxidized to S02, but as the temperature
increases, production of S03 becomes significant. For nonprecious metal
catalysts, the formation of S03 appears to be negligible below 810 K (1).
However, the onset temperature for emissions of sulfur as S03 is lower for
precious metal catalysts. On monolithic platinum/gamma alumina catalysts, S03
begins to form at about 750 K, and with platinum/palladium catalysts, temper-
atures must be kept below 670 K. (Personal communication with J. Brewer,
January 22, 1982).
4. Secondary Waste Generation
Catalytic incineration processes generate a secondary solid waste in the
form of spent catalyst. The type, quantity, and characteristics of the waste
will be very specific to the particular catalyst used and, to some extent, the
waste stream treated.
Waste disposal requirements for a 140 mVmin fluidized bed system employ-
ing a nonprecious metal catalyst have been estimated by Air Resources Co. to
be about 380 to 410 kg of spent catalyst every two years (Personal communica-
tion with B. Hernquist, January 15, 1982). In many cases it may be desirable
A17-9
-------
Appendix A17
Catalytic Incineration
to withdraw and replace only part of the catalyst charge periodically or con-
tinuously (e.g., to enhance catalytic activity which has been lost due to cata-
lyst degradation) rather than to replace the entire charge at once. Thus, the
spent catalyst from a fluidized bed system may be disposed of in small or
large batches.
When utilizing precious metal catalysts, it is often cost-effective to
recover the metal values by treating the spent catalyst, rather than discard-
ing it. Characteristics of the secondary wastes from precious metal recovery
processed were not determined.
5. Process Reliability
Data on process reliability in the form of stream factors are not gener-
ally available. This is largely due to the fact that catalytic incineration
systems are often used to treat waste gases on an "on call" or intermittent
basis. Operating problems which commonly cause downtime are often due to such
cyclic operation or to changes in influent composition which can adversely
affect the catalyst. Problems can also occur in the burner control system,
especially when the unit is operated in a noncontinuous mode or when influent
conditions change. Lack of a proper preventive maintenance program can also
lead to problems which result in downtime.
A datum point for a fixed bed precious metal catalytic combustor was
obtained from a manufacturer (Personal communication with J. Brewer, January
22, 1982) for an installation that is operated continuously at a Mystic Tape
plant in Illinois. Operating logs showed that during one year, nine shutdowns
occurred, in most cases for instrument maintenance. Total downtime was about
20-30 hours.
A17-10
-------
Appendix A17
Catalytic Incineration
6. Process Economics
Standard-size units are generally available for fluidized bed catalytic
combustors in the range of 14 to 850 m*/min of influent. Data from Air
Resources Co. (Personal communication with. B. Hernquist, January 15, 1982)
indicate that the capital cost of a standard 140 m*/min fluidized bed unit,
f.o.b. Chicago, is $66,000. Cost of the initial charge of catalyst is about
i3,000. These costs are derived from a recent quote to the State of
California. They include the basic catalytic combustor, a 4.2 GJ/hr gas pre-
heat burner, catalyst loading hopper/unloading cyclone, associated ductwork
and controls, and a small stack. Costs do not include an inlet fan or addi-
tional ductwork for recycle or bypass. A recuperative heat exchanger for pre-
heating incoming waste is usually installed with the unit. The cost of such a
heat exchanger of this size unit would be $35,000 to $40,000. Installation of
a 140 m'/min unit with heat exchanger on a prepared foundation is estimated at
$7,000 to $10,000 (less freight). For larger or smaller units, capital costs
can be reasonably estimated by assuming that f.o.b. costs increase with the
0.6 power of the capacity ratio (Personal communication with B. Hernquist,
January 15, 1982). All costs reported here are based on early 1982 dollars.
Utility requirements for fluidized bed combustors are mainly fuel gas for
the preheat burners and electric power for the inlet fan. On installations
which utilize a recuperative heat exchanger, preheat requirements normally can
be met entirely with waste heat from the process, and fuel gas is only needed
for the startup. Otherwise, for a 140 mj/min (5000 scfm) unit, the preheat
burner is normally fired at about a 2.6 GJ/hr rate depending on influent
temperature. Fan power requirements can be calculated for specific influent
streams using standard methods by assuming the fan must develop a static head
of 4.0 kPa (Personal communication with B. Hernquist, January 15, 1982).
Operating labor is minimal for the combustor and may be established at about
one hour per shift.
A17-11
-------
Appendix A17
Catalytic Incineration
Fixed bed precious metal catalytic combustors are available in sizes from
about 3 to 3000 ms/min. Most units are customdesigned for specific installa-
tions, particularly with regard to the catalyst formulation. Capital costs
for these units can run from $350 to $1400 per m'/min plus $105 per m'/min for
the initial catalyst charge (Personal communication with J. Brewer, January
22, 1982). This includes the basic combustor, preheat burners, recuperative
heat exchanger, and all associated ductwork and controls. Based on a mid-1981
installation at an Ampex plant in Redwood City, California, the capital cost
(f.o.b. Chicago) of a 170 mVmin unit was $100,000. Installed costs for a
grass roots installation can be reasonably estimated to be twice the f.o.b.
cost. For larger or smaller units, f.o.b. costs can be estimated as this base
cost times the capacity ratio raised to the 0.8-1.0 power (Personal communica-
tion with J. Brewer, January 22, 1982).
Operating costs are similar to the fluidized bed systems. Pressure drops
across the catalyst fixed bed (not the entire system) are generally on the
order of 0.5 kPa implying a slightly lower utilities load.
A17-12
-------
Appendix A17
Catalytic Incineration
7. References
1. Fisher, J.F. and G.R. Peterson. Control of Hydrocarbon and Carbon
Monoxide Emissions in the Tail Gases from Coal Gasification Facilities.
ORNL/TM-6229, Oak Ridge National Laboratory, Oak Ridge, Tennessee,
August 1978.
2. U.S. Environmental Protection Agency. Recommended Methods of Reduction
Neutralization, Recovery or Disposal of Hazardous Waste, Volume III,
Disposal Process Descriptions—Ultimate Disposal, Incineration, and
Pyrolysis Processes. Cincinnati, Ohio, February 1973.
3. Hardison, L.C. and E.J. Dowd. Emission Control Via Fluidized Bed
Oxidation. Chemical Engineering Progress, 73(8):31, August 1977.
4. U.S. Environmental Protection Agency. Catalytic Afterburners, In:
Pollution Engineering Manual. Second Edition, AP-40, May 1973.
5. U.S. Department of Health, Education and Welfare. Air Pollution
Engineering Manual. U.S. Government Printing Office, Washington D.C.,
1967.
6. Retallick, W.B. Design of Transfer-limited Catalytic Incinerators.
Chemical Engineering, January 12, 1981, pages 123-125.
7. U.S. Environmental Protection Agency. Catalytic Incineration of Low
Concentration Organic Vapors. EPA-600/S2-81-017, October 1981.
8. Personal Communications
Jerry Brewer, U.O.P. Automotive Products Division, Des Plaines, Illinois,
to L. Scinto, TRW, January 22, 1982.
Bob Hernquist, Air Resources Co., Palamine, Illinois to L.Scinto, TRW,
January 15, 1982.
A17-13
-------
APPENDIX A18
FUGITIVE ORGANICS CONTROL
A. LEAK DETECTION AND REPAIR
1. Process Description
Leak detection methods include individual component surveys, area (walk-
through) surveys, and fixed point monitors.
In an individual component survey each fugitive emission source (pump,
valve, compressor, etc.) is checked for VOC leakage. The source may be
checked for leakage by visual, audible, olfactory, soap bubble, or instrument
techniques. Visual methods are particularly effective in locating liquid
leaks. However, observation of a visible leak does not necessarily indicate
VOC emissions, since the leak may be composed of non-VOC compounds. Escaping
vapors from high pressure leaks can be audibly detected, and leaks of odorous
materials may be detected by smelling the odor.
Leaks from individual- components can also be detected by spraying soap
bubbles on equipment components. However this method does not distinguish
between leaks of non-VOC compounds from VOC leaks. In addition, this method
is not effective on hot sources or moving parts. With hot sources there is
the possibility of the evaporation of water in the solution, whereas the
motion of the moving parts may interfere with the motion of the bubbles.
A good method of identifying leaks of VOC from equipment components is
by using portable hydrocarbon detection instruments. The instrument is used
to sample and analyze the air in close proximity to the potential leak surface
by traversing the sampling probe tip over the entire area where leaks may
occur. The hydrocarbon concentration of the sampled air is displayed on the
instrument meter. The hydrocarbon concentration observed during monitoring of
a component is proportional to the VOC emission rate from the component. Data
from petroleum refineries have been used to develop relationships between
A18-1
-------
Appendix A18
Fugitive Organics Control
monitoring concentration and mass emission rates. The hydrocarbon concentra-
tion which defines a component needing maintenance must be chosen. Components
which have indicated concentrations higher than this "action level" are marked
for repair. Data from petroleum refineries indicate that large variations in
mass emission rate may occur over short time periods for an individual equip-
ment component. More frequent monitoring intervals tend to reduce the chance
of missing "large leaks" caused by their variable leak rates.
An area survey (also known as a walk-through survey) requires the use of
a portable hydrocarbon detector and a strip chart recorder. The procedure
involves carrying the instrument within one meter of the upwind and downwind
sides of process equipment and associated fugitive emission sources. An
increase in observed concentration indicates leaking fugitive emission
sources. The instrument is then used for an individual component survey in
the suspected leak area.
Fixed point monitors are automatic hydrocarbon sampling and analysis
instruments at various locations in the process unit. The instruments may
sample the ambient air intermittently or continuously. Elevated hydrocarbon
concentrations indicate a leaking component. As in the walk-through method,
an individual component survey is required to identify the specific leaking
component in the area. For this method, the portable hydrocarbon detector is
also required. Leaks from adjacent units and adverse meteorological condi-
tions may interfere with the method.
Many pumps have spares which can be operated while the leaking pump is
being repaired. Leaks from packed seals may be reduced by tightening the
packing gland. At some point, the packing may deteriorate to the point where
further tightening would have no effect or possibly even increase fugitive
emissions from the seal. The packing can be replaced with the pump out of
A18-2
-------
Appendix A18
Fugitive Organics Control
service. When mechanical seals are utilized, the pump must be dismantled so
the leaking seal can be repaired or replaced. Dismantling pumps will result
in spillage of some process fluid and evaporative emissions of VOC. These
temporary emissions may be greater than the continued leak from the seal, if
the seal leak is small.
Leaks from packed compressor seals may be reduced by the same repair
procedure that was described for pumps. Other types of seals require that the
compressor be out of service for repair. Since most compressors do not have
spares, repair or replacement of the seal would require a shutdown of the
process. Temporary emissions resulting from an extra shutdown may be greater
than the emissions from the seal if it were allowed to leak until the next
scheduled shutdown.
In general, relief valves which leak must be removed in order to repair
the leak. In some cases of improper reseating, manual release of the valve
may improve the seat seal. In order to remove the relief valve without shut-
ting down the process, a block valve is required upstream of the relief valve.
A spare relief valve should be attached while the faulty valve is repaired and
tested. After a relief valve has been repaired and replaced, there is no
guarantee that the next over-pressure relief will not result in another leak.
Most other valves have a packing gland which can be tightened while in
service. Although this procedure should decrease the emissions from the
valve, in some cases it may actually increase the emission rate if the packing
is old and brittle or has been overtightened. Plug type valves can be lubri-
cated with grease to reduce emissions around the plug. Some types of valves
have no means of in-service repair and must be isolated from the process and
removed for repair or replacement. Other valves, such as control valves, may
be excluded from in-service repair by operating or safety procedures. In many
A18-3
-------
Appendix A18
Fugitive Organics Control
cases, valves cannot be isolated from the process for removal. Most control
valves have a manual bypass loop which allows them to be isolated and removed.
Host block valves cannot be isolated easily although temporary changes in
process operation may allow isolation in some cases. If a process unit must
be shut down in order to isolate a leaking valve, the emissions resulting from
the extra shutdown will probably be greater than the emissions from the valve
if it were allowed to leak until the next scheduled process shutdown which
permits isolation for repair.
Depending on site specific factors, it may be possible to repair process
valves by injection of a sealing fluid into the source. This type of repair
may affect the operability of the valve such that replacement of the source
may be necessary within a short time after its repair. The emissions that
could result due to the replacement of the source should be evaluated when
considering this type of repair. It should be noted that injection of sealing
fluid has been successfully used to repair leaks from valves in petroleum
refineries in California.
In some cases, leaks from flanges can be reduced by replacing the flange
gaskets. Most flanges cannot be isolated to permit replacement of the gasket.
Data from petroleum refineries show that flanges typically emit only small
amounts of VOC.
2. Process Applicability
Leak detection and repair methods are designed to reduce the emission of
volatile organic compounds (VOC) that result from process fluid leaks from
plant equipment such as pumps, compressors, in-line process valves, pressure
relief devices, open-ended valves, sampling connectors, flanges, agitators,
and cooling towers.
A18-4
-------
Appendix A18
Fugitive Organics Control
3. Process Performance
The instrument survey of individual components is the only type of
detection method for which performance has been quantified. There are several
factors which determine the control effectiveness of individual component
surveys; these include: 1) action level or leak definition, 2) inspection
interval or monitoring frequency, 3) interval between detection and repair of
the leak, and 4) achievable emission reduction due to maintenance.
The action level is the minimum hydrocarbon concentration observed during
monitoring which defines a leaking component which requires repair. The
choice of action level for defining a leak is influenced by a number of impor-
tant considerations. First, the percent of total mass emissions which can
potentially be controlled by the monitoring and repair program can be affected
by varying the leak definition, or action level. Table A18-1 gives the per-
cent of total mass emissions affected by various action levels for a number of
equipment types. The data in this table indicate that, in general, a low
action level results in larger potential emission reductions. However, the
choice of an appropriate leak definition is limited by the ability to repair
leaking components. Test data indicate that about 50 percent of valve leaks
with initial screening values equal to or greater than 10,000 ppmv can be suc-
cessfully repaired. Similar data indicate that attempted repair of valve
leaks with initial screening values of less than 10,000 ppmv can increase
rather than decrease emissions from these valves. From these data it is con-
cluded that repairing leaks with screening values below 10,000 ppmv may not
result in a net reduction in mass emissions (1).
A18-5
-------
Appendix A18
Fugitive Organics Control
TABLE A18-1. FRACTION OF TOTAL MASS EMISSIONS FROM VARIOUS SOURCE TYPES
THAT WOULD BE CONTROLLED BY DIFFERENT ACTION LEVELS (1)
Action level (ppmv)
Safety/relief valves
Compressor seals
Flanges
Fraction of mass emissions (as %)
100,000 50,000 10,000 1,000
Source type
Pump seals
Light liquid
Heavy liquid
In-line valves
Vapor service
Light liquid
Heavy liquid
service
service
service
service
56
0
85
49
0
68
0
92
62
0
87
21
98
84
0
97
66
99
96
23
20
28
0
33
48
0
69
84
0
92
98
48
"Level of emission at which repair of the source is required.
These data show the fraction of the total emissions from a given source type
that is attributable to sources with leaks above the various action levels.
A monitoring plan may include annual, quarterly, monthly, or even weekly
inspections. The length of time between inspections should depend on the
expected occurrence and recurrence of leaks after a piece of equipment has
been checked and/or repaired. This interval can be related to the type of
equipment and service conditions, and different intervals can be specified for
different pieces of equipment after appropriate equipment histories have been
developed. According to reference 2, the recommended monitoring intervals for
refineries are: annual—pump seals, pipeline valves in liquid service, and
process drains; quarterly—compressor seals, pipeline valves in gas service,
and pressure relief valves in gas service; weekly—visual inspection of pump
A18-6
-------
Appendix A18
Fugitive Organics Control
seals; and no individual monitoring—pipeline flanges and other connections,
and pressure relief valves in liquid service. The choice of the interval
affects the emission reduction achievable since more frequent inspection will
result in leaking sources being found and fixed sooner. In order to evaluate
the effectiveness of different inspection intervals, it is necessary to esti-
mate the rate at which new leaks will occur and repaired leaks will recur.
The estimates which have been used to evalute yearly, quarterly, and monthly
inspections are shown in Table A18-2 (1).
If a leak is detected the equipment should be repaired within a certain
time period. The allowable repair time should reflect an interest in elimi-
nating a source of VOC emissions but should also allow the plant operator suf-
ficient time to obtain necessary repair parts and maintain some degree of
flexibility in overall plant maintenance scheduling. The determination of
this allowable repair time will affect emission reductions by influencing the
length of time that leaking sources are allowed to continue to emit pollu-
tants. Some of the components with concentrations in excess of the leak
definition action level may not be able to be repaired until the next sche-
duled unit shutdown, e.g., a unit turnaround.
The effects of different allowable repair intervals are shown in Table
A18-3. The percentages shown in the table are the percent of emissions from
the component which will be affected by the repair. As indicated, the emis-
sions which occur between the time the leak is detected and repair is
attempted increase with increasing allowable repair intervals.
Based on the above considerations of action level, inspection interval,
and allowable interval before repair, leak detection and repair methods are
capable of reducing refinery VOC emissions by 50 to 80 percent (1). Detailed
A18-7
-------
TABLE A18-2. ESTIMATED OCCURENCE AND RECURRENT RATE FOR VARIOUS MONITORING INTERVALS (1)
00
i
CD
Eatimated percent
of aourcea leaking
it above 10,000 ppm
Estimated percent of
repaired tonrcea which
are found leaking at .
aubaeonent inspections
Estimated percent of
aonrcea which are
found leaking at
subsequent inspections^
Source Type
Puwip aeala
Light liquid aerrice
Heavy liquid service
In-line valvea
Vapor aervice
Light liquid service
Heavy liquid service
Safety/relief valvea
Compressor aeala
Flangea
initially*
23
2
10
12
0
8
33
0
Annual
20
20
20
20
20
20
20
20
Quarterly
10
10
10
10
10
10
10
10
Monthly
5
5
5
5
5
5
5
5
Annual
4.6
0.4
2.0
2.4
0.0
1.6
6.3
0.0
Quarterly
2.3
0.2
1.0
1.2
0.0
0.8
3.3
0.0
Monthly
1.2
0.1
0.5
0.6
0.0
0.4
1.7
0.0
Approximate fraction of sources having leaks equal to or greater than 10,000 ppn prior to repair.
Approzlaate fraction of leaking aonrcea that were repaired but found to leak during subsequent inapectioni. These approximations are
baaed on engineering judgment.
°Approximate fraction of aonrcea that were found to leak during a anbaeqnent inapection. These approximations are the product of the
information presented in footnotea a and b.
-------
Appendix A18
Fugitive Organics Control
TABLE A18-3. PERCENT OF MASS EMISSIONS AFFECTED BY VARIOUS
REPAIR INTERVALS (1)
Allowable repair interval, days
Percent of emissions affected
30
95.9
15
97.9
5
99.3
1
99.9
methods of calculating achievable VOC emission reductions are presented in
reference 1. Whether these same reductions can be achieved in synthetic fuels
facilities has not been demonstrated.
4. Secondary Waste Generation
No secondary waste is generated by this control process.
5. Process Reliability
The reliability of leak detection by visual, olfactory, or soap bubble
techniques is at best marginal. For example, the soap bubble method does not
distinguish between non-VOC and VOC leaks and is totally ineffective on hot
sources or moving parts. The best method of identifying leaks from individual
components is by using portable detection instruments.
The reliability of the area survey method for locating leaks is also not
well established. It has been estimated that the walk-through survey combined
with selected individual surveys will detect about 50 percent of the leaks
identified in a complete individual survey. The area survey method also does
not detect leaks from sources such as elevated valves or relief valves. Leaks
from adjacent units and adverse meteorological conditions can also interfere
with the walk-through survey. Consequently, the walk-through survey is best
for locating only large leaks with a small resource expenditure.
A18-9
-------
Appendix A18
Fugitive Organics Control
Fixed point monitors have been estimated to locate up to 33 percent of
the number of leaks identified by a complete individual survey. However,
fixed point monitors are more expensive since multiple units are required and
the portable instrument is still essential to locate the specific leaking com-
ponent.
The effectiveness of repairs in reducing fugitive emissions has not been
well documented; however, data for valve repairs have been collected from
several refineries. In many cases, perfect repair cannot be achieved but
emissions from the component can be minimized by identifying the lowest
practically achievable emission rate.
6. Process Economics
Capital expenditures for leak detection and repair methods include costs
for monitoring instruments, caps used on open-ended lines, replacement seals,
etc. In last quarter 1978 dollars, instrument costs are estimated to be $4250
and costs for caps are estimated to be $457line based on installation of a 2.5
cm screwed valve.
The costs of replacement seals are usually combined with labor costs as
an equivalent labor requirement. Actual labor requirements will depend upon
the number of workers required to monitor a component (1 for visual, 2 for
instrument), the time required to monitor, the number of components in the
unit, and the number of times the component is monitored per year.
Typical monitoring and repair times for different components are given in
Table A18-4. For pump and compressor seals, the repair times shown in this
table include the cost of replacement seals at the equivalent of $15/labor
hour.
A18-10
-------
TABLE A18-4. MONITORING AND LEAK REPAIR TIME REQUIREMENTS (3)
Pumps
Single Mechanical Seals
Double Mechanical Seals
Valves
Safety/Relief Valves
Type of
Monitoring
Instrument
Visual
Instrument
Visual
Instrument
Instrument
Monitoring
Time,
minutes
5
0.5
5
0.5
1
8
Repair
Time,
hours
80a
80a
1.13
1.13
(gas service)
Compressor Seals Instrument 10 40
a .
Includes the cost of replacement seals at the equivalent of $15/hr.
A18-11
-------
Appendix A18
Fugitive Organics Control
B. EQUIPMENT SPECIFICATION
1. Process Description
The second method used to control fugitive organic emissions is by equip-
ment specification. Some of the specifications may be applicable to more than
one source.
Fugitive emissions from pumps and compressors occur at the junction of a
moving shaft and a stationary casing. Equipment specifications that may be
implemented for control of fugitive emissions include improvement of the seal
at the junction, and collection and control of the emissions from the junc-
tion. Diaphragm type pumps or "canned" pumps do not have a shaft/casing
junction. In normal operation they do not leak the pumped fluid, although
failure of the diaphragm may result in temporary emissions of VOC.
Double mechanical seals are a method of improving the seal at the shaft/
casing junction. These seals consist of two mechanical sealing elements with
a barrier fluid in the chamber between the seals. This chamber is either
flushed with circulating barrier fluid or flooded with static barrier fluid.
The pressure of the static barrier fluid can be monitored to detect failure of
the inner seal. Any fluid leaking through the inner seal may be dissolved or
suspended in the barrier fluid, and subsequent degassing of the barrier fluid
may result in VOC emissions. Therefore, barrier fluid degassing vents must be
controlled in order to provide maximum control effectiveness of double
mechanical seals.
Emissions of VOC from degassing vents can be controlled by a closed vent
system which consists of piping and, if necessary, flow inducing devices to
transport the degassing emissions to a control device such as a process heater
or vapor recovery system.
A18-12
-------
Appendix A18
Fugitive Organics Control
A closed vent system could also be applied to control emissions from the
seal area of pumps and compressors. The seal area would be enclosed in order
to collect the emissions and a vacuum eductor or a compressor could be used to
remove vapors from the seal area. However, normal operating practices may
require frequent visual inspection or mechanical adjustments in the seal area.
This would not be possible with a closed vent system at the seal area. A
potential problem with this approach is that explosive mixtures may be created
by enclosing the seal area, and therefore safety and operating practices may
limit the use of closed vent systems. However, this type of system has been
applied to compressor seal areas in petroleum refineries.
Fugitive emissions from in-line valves occur at the stem or gland area of
the valve body. Diaphragm and bellows seal valves do not have a stem or gland
and therefore are not prone to fugitive emissions. Diaphragm valves are
generally used where hazardous or toxic process fluids are present and fugi-
tive emissions must be eliminated.
An additional source of fugitive emissions from open-ended valves is
leakage through the seat of the valve. These emissions can be controlled by
installing a cap, plug, flange, or a second valve to the open end of the
valve. In the case of a second valve, the upstream valve should always be
closed first after use of the valves. Each time the cap, plug, flange, or
second valve is opened, any VOC which has leaked through the first valve seat
will be released. Caps, plugs, flanges, etc. for open-ended valves do not
affect emissions which may occur during use of the valve. These emissions may
be caused by line purging for sampling, draining, or venting through the open-
ended valve.
A18-13
-------
Appendix A18
Fugitive Organics Control
Fugitive emissions from pressure relief devices occur because of improper
seating or partial failure of the device. Emissions which result from normal
operation of the devices caused by over-pressure of the process or vessel
which the device protects are not normally considered fugitive emissions.
Manufacturers of relief valves state that resilient seat or "0-ring"
relief valves provide better reseat qualities compared to standard relief
valves. No test data are available to verify these statements. These
improvements would have no effect on overpressure emissions or fugitive
emissions due to seal failure or "simmering."
Although rupture disks are also pressure relief devices, they can be
installed upstream of a safety/relief valve in order to prevent fugitive emis-
sions through the relief valve seat. This procedure may require use of a
larger size relief.valve because of operating codes. The disk/valve combina-
\
tion may also require appropriate piping changes to prevent disk fragments
from lodging in and damaging the relief valve when relieving overpressure. A
block valve upstream of the rupture disk is also required in order to permit
in-service replacement of the disk after overpressuring. If the disk could
not be replaced, the first overpressure would result in the relief valve being
the same as an uncontrolled relief valve, and it may actually be worse since
disk fragments may prevent proper reseating of the relief valve. Tandem
pressure relief devices with a three—way valve can be used to avoid operation
without overpressure protection. Rupture disk/relief valve combinations must
have some provision for testing the integrity of the disk. The area between
the rupture disk and relief valve must be connected to a pressure indicator,
recorder, or alarm. If the process fluid is not hazardous or toxic, a simple
bubbler apparatus could be used to test disk integrity by connecting the
bubbler to the disk/valve area. The control efficiency of the disk/valve
combination is assumed to be 100 percent for fugitive emissions. If the disk
M8-14
-------
Appendix A18
Fugitive Organics Control
integrity is not maintained or if the disk is not replaced after overpressure
relief, the control efficiency would be lowered. The disk/valve combination
has no effect on emissions which result from overpressure relieving.
A closed vent system can be used to transport the discharge or leakage of
pressure relief devices to a control device such as a flare. Since over-
pressure discharges as well as fugitive emissions are routed to the control
device, it must be sized appropriately. A larger pressure relief device may
be required for use with a closed vent system. The control efficiency of a
closed vent system is dependent on the effectiveness of the control device.
Typical flare systems may be only 60 percent effective for fugitive emission
destruction. This efficiency reflects the fact that many flare systems are
not of optimum design. As a result, flares that are designed to handle large
volumes of vapors associated with overpressure releases are used to handle low
volumes of fugitive emissions. With such designs, optimum mixing is not
achieved because the vent gas exit velocity is low and large flares generally
cannot properly inject steam into low volume streams. A properly designed
flare system typically exhibits a 99 percent hydrocarbon destruction
efficiency.
2. Process Applicability
Equipment specifications provide a more restrictive level of control than
the leak detection and repair control process discussed earlier. Several
equipment specifications could be applied to existing equipment components. A
Table A18-5 is a listing of the types and effectiveness of equipment specifi-
cations available. The primary disadvantage of this control technique is its
high initial capital expenditure. However, operating costs are much lower
compared to the leak detection and repair control process.
A18-15
-------
TABLE A18-5. EFFECTIVENESS OF EQUIPMENT MODIFICATIONS (3)
Control efficiency
Source type/equipment modification (%)
Pumps
Sealess pumps
Double mechanical seals/closed vent system
Closed vent system on seal area -100
Compressors
Double mechanical seals/closed vent system ~100a
Closed vent system on seal area -100
Safety/relief valves .
Closed vent system 60
Rupture disks 100
Open-ended lines c
Caps, plugs, blinds, second valves 100
Sampling connections
Closed loop sampling 100
In-line valves
Diaphragm valves 100
aAlthough this control efficiency is not attained in all cases, it is
.achievable in some cases.
This control effectiveness reflects the fact that a closed vent system is
normally sized for emergency relief.
cThis control efficiency reflects the use of these devices downstream of an
initial valve with VOC on one side and atmosphere on the other.
A18-16
-------
Appendix A18
Fugitive Organics Control
3. Process Performance
Process performance is difficult to accurately determine due to many
complicating factors. For example, the efficiency of caps or plugs for open-
ended valves is dependent on 1) the frequency of removal of the cap or plug,
since this removal will result in emission of fluids trapped by the cap or
plug and 2) the emission rate through the valve seat. The estimated control
achievable by the various equipment specifications/modifications are shown in
Table A18-5. These estimates represent the maximum emission reduction possi-
ble for the equipment modifications. In some instances, the actual emission
reduction will depend on other factors such as the efficiency of control
devices attached to closed vent systems. Vapor recovery systems would
approach the 100 percent efficiency, but flares may in some cases be only 60
percent effective for hydrocarbon destruction.
4. Secondary Waste Generation
Typically no secondary waste is generated by equipment specification.
5. Process Reliability
The reliability of equipment specifications is directly related to the
effectiveness of the equipment being installed. Double mechanical seals may
develop leaks at the outer seal/shaft junction after extended periods of use.
Reliability of a double mechanical seal and closed vent system is dependent on
the effectiveness of the heater, or vapor recovery system, and the frequency
of seal failure. Failure of both the inner and outer seals can result in rela-
tively large VOC emissions at the seal area. Pressure monitoring of the
static barrier fluid may be used in order to detect failure of the seals. In
A18-17
-------
Appendix A18
Fugitive Organics Control
addition, visual inspection of the seal area can also be effective in detec-
ting failure of the outer seals. Upon seal failure, the leaking pump or
compressor would have to be shut down for repair.
Fugitive emissions from rupture disks may be caused by pinhole leaks in
the disk itself caused by corrosion or fatigue. Fugitive emissions from
relief valves may be caused by failure of the valve seating surfaces, improper
seating after overpressure relieving or process operation near the relief
valve set pressure which may cause "simmering."
Diaphragm valves are generally extremely reliable (approximately 100
percent). The applicability of diaphragm valves is limited by the strength of
the diaphragm. Diaphragm valves may not be suitable for many applications
because of process conditions or cost considerations.
As a whole, equipment specifications are probably more reliable in con-
trolling fugitive emissions compared to the leak detection and repair control
option discussed earlier.
6. Process Economics
Capital costs on a component basis are presented in Table A18-6.
Although labor costs are significantly reduced when equipment specification is
used for fugitive organics control rather than leak detection and repair,
maintenance and replacements costs may be substantial. Annual costs are
typically higher than for leak detection and repair and the difference in cost
is strongly dependent on the operations.
A18-18
-------
TABLE A18-6. CAPITAL COST DATA (3)
Capital Cost Itei
$ value/component
(1978 last quarter dollars)
Monitoring Instruments
Caps for Open-Ended Valves
Doable Mechanical Seals
• Seals
• Installation
Flush Oil System for Double Mechanical Seals
Vents for Compressor Degassing Reservoirs
Vents for Pump Degassing Reservoirs
Rupture Disks for Relief Valves
Closed Loop Sampling Connectors
$4,250
$ 45/line*
$ 575/pump
$ 240/pump
$1,500/pump
$6 ,530/compressor
$3 ,265/pump
$l,730/valvee
$ 400/sample connector
fflased on installation on a 2.5 cm screwed valve.
Costs are for a seal oil flushing system for pumps with double mechanical
seals.
Based on installation of ,a 122 m length of 5.1 cm diameter schedule 40 carbon
steel pipe at a cost of $5200; plus three 5.1 cm cast steel plug valves and
one metal gauze flame arrestor at a cost of $1330. These costs include
connection of the degassing reservoir to an existing enclosed combustion
.device or vapor recovery header.
This cost is based on the assumption that two pumps (such as a pump and its
spare) are connected to a single degassing vent.
Based on the installation of 7.6 cm stainless steel rupture disk ($195), one
7.6 cm rupture disk holder, carbon steel ($325), one 0.6 cm pressure gauge,
dial face ($15), one 0.6 cm bleed valve, carbon steel, gate ($25), installa-
tion ($240), and a block valve to facilitate disk replacement ($240).
Costs for offset mounting and installation are also included.
Based on installation of a 6 m length of 2.5 cm diameter, schedule 40, carbon
steel pipe and three 2.5 cm carbon steel ball valves.
A18-19
-------
Appendix A18
Fugitive Organics Control
7. References
1. Wetherold, R.G., and L.P. Provost. Emission Factors and Frequency of
Leak Occurrence for fittings in Refinery Process Units. Final Report.
EPA/600/2-79/044, Radian Corporation, February 1979.
2. U.S. Environmental Protection Agency. Control of Volatile Organic
Compound Leaks from Petroleum Refinery Equipment. EPA-450/2-78-036 OAQPS
No. 1.2-111, Research Triangle Park, North Carolina, June 1978.
3. U.S. Environmental Protection Agency. VOC Fugitive Emissions in
Synthetic Organic Chemicals Manufacturing Industry - Background
Information for Proposed Standards. Preliminary Draft. OAQPS, March
1980.
A18-20
-------
APPENDIX A19
FUGITIVE DOST CONTROL TECHNIQUES
Fugitive dust emissions from synfuel plants are generated from coal/shale
storage piles and coal/shale conveying and processing equipment. Emissions
from storage piles occur due to wind erosion and coal handling operations.
Different control techniques are used to effectively control fugitive emis-
sions from reserve (dead) or active (live) storage piles.
1. Dust Control for Reserve Storage Piles
Since reserve storage piles are normally subject to minimal disturbances,
the exposed pile surfaces can be sealed or stabilized for dust control. Four
basic methods are generally used: 1) vegetative stabilization, 2) chemical
stabilization, 3) capping, and 4) stacking.
Vegetative techniques have been used to control dust emissions from
tailing piles and in agricultural applications. Vegetative techniques need a
soil which supports growth. The efficiency of vegetative cover in reducing
wind erosion is dependent on the density and type of vegetation that can be
grown. For tailing piles, the use of vegetative stabilization results in a
decrease in emissions of approximately 65 percent. Combined with a chemical
stabilizer the efficiency can increase to approximately 90 percent.
Chemical wetting type stabilization agents are used to provide better
wetting of fines and longer retention of moisture. These wetting agents
reduce the water surface tension allowing the fines to be wetted with a mini-
mum amount of water. These treatments will protect stockpiled materials until
the added moisture has been removed by heat and wind. Some of these agents
remain effective for weeks or months without rewatering the piles, depending
on local conditions.
A19-1
-------
Appendix A19
Fugitive Dust Controls
Crusting procedures involve the use of bunker C crude oil, water soluble
acrylic polymers, or organic binders. The compounds are sprayed on the sur-
face of the stockpile, coating the top layer of particles with a thin film.
This film causes the particles to adhere to one another, forming a tough
durable crust resistant to wind and rain. As long as the crust remains
intact, the stockpile will be protected from wind losses. By adding coloring
agents to the solutions, protected areas can be easily identified.
Capping procedures involve the use of asphaltic compounds, earth, or
polyethylene tarpaulins to pave the surface area of the pile. Capping can be
performed by using a mixture of wood pulp and asphalt in a slurry which is
sprayed over the surface of the pile. Alternatively, asphalt or road tar can
be sprayed on the stockpile, thus sealing the pile from air and moisture. The
covering is usually about 3 mm thick. Polyethylene tarpaulins are another
effective means of dust control. However, they are not very practical because
of the difficulty in handling them due to high wind speeds and the large size
of the storage pile under consideration.
A positive and relatively inexpensive method of dust control involves the
use of well stacked material on the surface of a properly compacted pile.
Erosion can be controlled by providing the stockpile with a 150 mm layer of
fine stockpile material (6.4 mm x 0 mm) over the top and sides. The fine
material can be anchored in place with a 100 mm layer of larger size material
(24 mm x 0 mm) with better weathering characteristics. During stockout and
reclaim operations, this outer layer of material has to be kept separated from
the areas being worked. This type of seal is an economical means of reducing
wind and rain erosion.
Table A19-1 summarizes information on the performance, reliability,
costs, and secondary wastes generated by dust controls for inactive storage
piles.
A19-2
-------
TABLE A19-1. KEY FEATURES OF STORAGE PILE DUST CONTROL TECHNIQUES
Method
Control Principle
Control
Effectiveness*
Reliability/
Special Problems
Dnit Costs*
Other
Pollutants
Generated
Windbreak
Use of vegetation, berms,
or natural land features to
prevent direct wind imping-
ment on coal/shale storage
pile
4-10% redaction in dusting
losses as compared to pile
set in open on level ground.
(Effectiveness varies with
pile distance from windbreak)
Reserve Storage Pile Stabilization Methods
Vegetative Covering pile with sod
I
OJ
Ch em i c a 1
¥ettin» Agents
Modify surface tension
properties to improve
effectiveness of water
sprays
Crusting Agents
Organic binders com-
bine with coal par-
ticles to form tough
Approximately 65% reduction
over unstabilized pile; 90%
if soil stabilizer used
Dp to 90% reduction
in dusting losses
Up to 90%
1. Requires frequent &2.70/B1 Soil dust from
watering earth covering
2. Handling of sod
duration reclaim
operations is cum-
bersome and expensive
3. Upper coal layer is
contaminated
1. Piping requires heat i.33-,77/Mg Depends on
tracing to be effect- wetting agent
ive in freezing utilized
weather conditions
problems in equipment
exposed to sprays
3. Nay increase coal
degradation
4. Effects are short-term
1. May increase chances i.055 — .22/m* Depends on
of spontaneous com- crusting agent
bust ion, especially utilized
in piles subject to
stockont and reclaim
operations
2. Croat tend* to break
up during heavy raint
(Continued)
-------
TABLE A19-1. (Continued)
Method
Cappi og
Control
Control Principle Effectiveness
b
Paving with tirth or Up to 100%
asphalt or cover with
KelUbUity/
Special Problem!
1. Both coverings
nay increa se
t a neons c ambus tion
Unit Coits*
J. 49/m» for
aaphal t
Other
Pollutants
Genera ted
Soil dust from
earth covering
2. Polyethylene pre-
sents severe
handling problems
and is also not
practical in high
vind climates
41.96/K* for
polyethylene
Stacked Coal
Coating surface of com-
pacted storage pile with
layer of select, medium
slied higher grade coal
No data available
Either deliveries of
both coals must be
coordinated or both
coals must be readily
available from storage
Active Storace Pile Stabilization Methods
Water Sprsy Spray application of
210-250 L/Mg co»l to
reduce dusting
Conf i nement
Approximately 50%
reduction in lo&ses
Up to 99% reduction
in 1osse s
1. Cannot be used in
freezing weather
condi t ions.
2. May increase
degradatlon
al
3. Frequent re-treatment
necessary
1 . Approz IBS tely
1 million to J3
mil 1 ton per silo
depending on size.
Estimated price
for silo in
Kentucky: $1.5
million (silo)
+ il.35 million
(foundation) - $2.85
million (total)
*Cost and control efficiency data obtained from Reference 1
Data obtained from Reference 2
°Data obtained from Reference 3
-------
Appendix A19
Fugitive Dust Controls
2. Dust Control for Active Storage Piles
In active storage piles, dust is generated because of stockpiling opera-
tions in which free-falling distances are often up to 18 meters. In addition,
dust is generated as material is removed from active storage piles by under-
ground feeder systems or surface machines such as a bucket wheel stacker-
reclaimer.
There are two primary methods of dust control currently being used:
water spraying and enclosures. Crusting agents and chemical sprays discussed
for dead/reserve storage piles are inadequate for active storage piles,
because of the frequent stockout and reclaim opearations. In addition, it is
cumbersome and expensive since the disturbed area must be restored by
respraying.
Spraying the piles with water to suppress the dust can provide a simple
solution to the problem. Spray nozzles must be placed at strategic locations
over the stockout area. The spraying operation is simple, involving only the
starting and stopping of a pump.
Water requirements for large-volume operations vary from 210 to 250
liters/Mg of material. The dust-suppression effectiveness of water spraying
is excellent in warm weather, questionable in freezing weather; wind is also a
major factor. The added moisture may cause the stored material to stick
together, creating handling problems as the material moves through the plant
supply system.
Using enclosures (e.g., storage bins, silos, and barns) for storing
materials is generally the most effective means of reducing storage emissions.
However, enclosures can be very expensive since they have to be designed to
A19-5
-------
Appendix A19
Fugitive Dust Controls
withstand wind and snow loads and to meet requirements for interior working
conditions. An alternative to enclosure of all material is to screen the
material prior to storage, sending the oversized material to open storage and
the fines to silos.
Wind screens, or partial enclosure of storage piles, can reduce wind
erosion losses but do not permit capture of the remaining storage pile
fugitive dust emissions.
Table A19-1 summarizes information on the performance, reliability,
costs, and secondary wastes generated by dust controls for active storage
piles.
3. Dust Control for Coal/Shale Conveying and Processing
The control of coal dust from coal conveying and processing can generally
be accomplished by use of either of the following methods:
• Dust suppression by application of a wetting agent or
• Dry collection of the coal dust using a mechanical
dust containment and collection system.
Dust suppression systems using a wetting agent are presently available.
These systems consist of pre-engineered modules which incorporate both water
handling components and automatic spray controls. A typical spray solution
contains 1,000 to 4,000 parts of water to one part of a wetting agent. The
rate of spray application is about 4 to 8 L/Mg of material. This rate of
application results in an increase in total surface moisture by about 0.5 to
1.0 percent.
A19-6
-------
Appendix A19
Fugitive Dust Controls
Wet dust suppression is accomplished by the following three inherent
actions:
• Confinement of the dust by a curtain of moisture droplets,
• Wetting of the dust by contact and penetration with
moisture droplets, and
• Formation of agglomerates too heavy to remain airborne (or
become airborne) by "cementing" small dust particles to
each other with moisture droplets.
Small wet dust suppression systems may use normal water supply pressure
and a very small proportioner. Such installations requiring freeze protection
have the option of using heaters inside the proportioner cabinet and a heater
for the wetting agent drum. This option eliminates the requirement for a
separate heated room. All piping is heat traced in the normal manner as
necessary.
Adding moisture to coal is generally undesirable due to the handling and
stockpiling problems it causes. Excessive moisture can cause spontaneous
combustion by speeding up pyritic oxidation. Upon oxidation the pyrites
increase in volume, thereby breaking the coal into small particles. This
increases the exposed surface area of the bulk material and, therefore,
increases the rate of oxidation and the amount of fines. The addition of
water also increases the moisture content of the coal accentuating freezing of
the material in winter months.
Dry dust collection systems consist of an enclosure to contain the dust,
ductwork and exhaust to convey the dust laden air, and dust collectors to
separate the dust from the effluent air. Typically, hoods are used to capture
dust at transfer points. Conveyors generally have a half cover. This not
only provides dust containment but shields the conveyor from wind, rain, and
snow.
A19-7
-------
Appendix A19
Fugitive Dust Controls
Of major importance in ductwork and exhauster design is the ductwork
velocity. The velocity should be within the recommended range for the various
dust collection points. Typical velocities are listed in Table A19-2.
Dust collectors applicable to the collection of the captured dust are:
1) dry centrifugal collectors, 2) fabric filters, 3) electrostatic
precipitators, and 4) Venturi scurbbers. These are discussed in detail in
Appendices All through A14.
4. References
1. Information on control efficiency and unit costs for coal storage piles
obtained from Black & Veatch, Consulting Engineers, Kansas City, B&V
Project No. 8827-012, January 1981.
2. Power Engineering, July 1979.
3. Jutze, G.A., et al.. Technical Guidance for Control of Industrial Process
Fugitive Emissions, PB 272-288, PEDCo Environmental, Cincinnati, Ohio,
March 1977.
A19-8
-------
TABLE A19-2. DUCTWORK AND EXHAUSTER DESIGN VELOCITIES
Dust Collection Point
Recommended Velocity
Truck and rail car dump
Head pulley
Tail pulley
Tail pulley
(belt entering)
Screens
Vibration feeder and
crusher feed
Silos and bins
30 to 60 m/min of indraft through
open areas.
32 m*/min per m of belt width for
speeds up to 180 m/min, 37 to 42 m'/min
per m if over 180 m/min. Never use
less than 20 mVmin, except on inter-
mittent sampling circuit exhaust.
46 to 70 mVmin per ra of belt width for
speeds up to 180 m/min; 56 to 74 m'/min
per m above 180 m/min. Dse higher
exhaust rates if falls exceed 2.5 to
3.0 m.
20 mVmin for belts less than 1 m wide,
28 m3/m if 1 m or wider.
4.6 m3/min per m of screen deck area
for enclosed screens. Minimum 7 m3/min
per m if screen cover is not totally
enclosed.
28 to 46 m3/min per m of feeder and
crusher feed chute.
Determined by material feed rate. One-
half to one times the displaced air
should be added for entrained air plus
volumes of 60 to 75 m/min for indraft
through open areas.
A19-9
-------
APPENDIX A20
FLDE GAS DESDLFORIZATION PROCESSES
A large number of flue gas desulfurization (FGD) processes are commer-
cially available. Hie key features of several FGD processes are summarized in
Table A20-1. The following sections give brief descriptions of calcium based
wet systems (using lime/limestone as an example), sodium based wet systems
(dual alkali and Wellman-Lord process), and dry scrubbing (spray drying).
1. Calcium Based Wet Systems
Calcium based wet FGD systems are the most prevalent type of S02 control
processes used for coal-fired utility generating stations. Lime, limestone,
and alkaline ash systems represent about 84 percent of the FGD capacity cur-
rently operating, under construction, or being planned. Almost all of these
processes are nonregenerable or throwaway processes. The major impetus for
the use of nonregenerable systems is economic. The cost of limestone is such
that discarding the resulting sludge is the cheapest S02 control mechanism
available.
The history of lime and limestone FGD systems since their initial appli-
cation in the early 1970s has not always been favorable. Numerous problems
have plagued commercial systems throughout the development of this technology.
One of the major causes of problems is the relatively low solubility of most
calcium compounds present in FGD systems. The high solubility of calcium
chloride also causes problems. Because of the low solubility of calcium
carbonate and calcium sulfite, high liquid to gas (L/G) ratios are needed for
adequate S02 removal. The low solubility of calcium sulfate has often caused
scale formation in a system either due to undersized reaction tanks or chem-
istry upsets. The high solubility of calcium chloride results in high liquid
phase chloride concentrations which can cause corrosion.
A20-1
-------
TABLE A20-1. KEY FEATURES OF S02 REMOVAL PROCESSES
Process Lime/Limestone
Feature Scrubbing Double Alkali Scrubbing
Principle Liquid phase absorb- Liquid phase absorption
or limestone slurry. hydroxide, sodium sul-
fite, sodium bisulfite,
sodium sulfate, and
sodium carbona te sola—
tion. Regeneration of
the sodium sulfite/
bisulfite with line in
a reactor. A dilute
1500 ppra SO, and where
less than 25% oiidation
of sulfite to sol fate
•ode can be used for
8000 ppm SO, .
Chiyoda
Thoroughbred 121
Liquid absorption of
SO, in a single
vessel , where 1 ime-
stone addition and
dissolut ion, air
oxidation, and gyp sun
precipitation occur.
Wei Iman- Lord
of SO, in a sodium bi-
sulfite, sodium snlfite,
and sodium carbonate
solution. A rich SO,
is produced by evapora-
tion, which is then pro-
cessed in a Claus unit
sulfur or in a sulfuric
acid plant.
Dry Scrubbing
which contacts the
flue gas with an
aqueous alkal ine
material and produces
a dry product . Sys-
spray dryer; 2nd
collector which re-
reaction product
from flue gas.
Fly Aah
Alkalinity Scrubbing
ing the fly ash alka-
linity for SO,
removal . Hydra ted
dolomitic 1 ine,
(Mg(OH), and Ca(OH),)
achieve an outlet SO,
of 43 ng/J.
O
f
Feed Stream
Requirement s
Absorbent
Product/
Waste
Particulate can be
removed in an EPS or
fabric filter to
achieve 99-Hfc at
1 ow e s t e ne r gy c o n-
sumption. Fly ash
nay be removed in a
venturi where the
fly ash contains
significant alka-
linity. A scrubber
can be used for
both high particn-
late and SO, re-
moval .
Slaked lime or 200-
300 mesh limestone
6-12% slurry circu-
lated.
Gypsum can be pro-
duced with forced
oxidation. Calcium
anlfate/snlfite can
be produced with
50-70% solids.
Excessive particulates
should be removed in an
ESP, fabric filter, or
venturi. 0, should be
less than 7% for con-
centrated mode.
Sodium hydroxide, sodium
sulfite/bisulfite, and
a smal1 amount of sodium
sulfate.
Filter cake with 60-70%
sol ids, primarily cal-
cium sulfite and calcium
snlfate.
Particulates and
chlorides should be
removed from inlet
flue gas if byproduct
gypsum is to be sold.
Limestone slurry.
Gypsum (CaS04-2H,0)
with less than 20%
moisture content.
Particulates and chlo-
rides must be removed
from flue gas.
Concentrated sodium
snlfite/bisulfite.
Concentrated SO, purge
atream (90% SO,).
Inlet SO, concentra-
tion should not
exceed 1000 ppmv.
Venturi is used to re-
move particulates and
a portion of SO,
Line slurry or so dine Fly ash alkalinity and
carbonate solution. hydra ted lime (calcium
and magnesium hydrox-
ide).
Sodium sulf He-sodium
sulfate; calcium aul-
f ite /snlf ate.
Sludge consists of fly
ash, gypsum
(CaS04 -2H,0), (Mf (OH)a,
small amount of cal-
cium snlfite.
(Continued)
-------
TABLE A20-1. (Continued)
NJ
o
I
U)
Process Line /Lime stone
Feature Scrubbing
Efficiency 90% r en oval can be
obtained generally
for low and high
tained with higher
L/6 and pressure
extent scrubber type.
renoval for high
sulfur coals when
adiplc acid li used.
Coomercial ly demon-
strated in over 30
PGD units.
and OIH coata. SO,
lively simpl e
process.
approximately 2
aa a«h collected.
Sludge can be
thizo tropic.
Sludge quantities
forced oxidation.
Double Alkali Scrubbing
90-99% removal can be
obtained for low and
high sulfur coals. Con-
Louisville Gas 1
Eiectrlc's 200 Mi coal-
industrial units (General
costs. SO, can be
equipment.
atructlon aiay be re-
quired to prevent
eorroaion and pitting.
Chlvoda
Thoroughbred 121 fellaan-Lord
300 ppaiv. Proceaa 11SM coal-fired boiler.
in Gulf Power 'a Scholi SO,
elation - 20MI proto-
type.
Capital and OIH coata Coanercially deaion-
to prototype eiperl- sulfur with a Clans unit
saleable gypana by- potential for acaling
product. than calcium ayeteau
quired. Syateai required
to process SO, to sulfur
or aulfurio acid.
Dry Scrubbing
701 SO, removal using
Lower projected
1982. Product dia-
poaal could be a
problem when aodlun
aalta are need aa
abaorbenta.
Fly Aah
Alkalinity Scrubbing
S5-9M removal of SO,
2. 93*4 expected in
Col strip units 314.
Coanercially demon-
taina little calclnm
aulflte which la>-
Less potential for
acal lag.
boiler* which burn
high alkalinity coala.
-------
Appendix A20
FGD Processes
The lime/limestone system described here is a generic process which is
offered by many different vendors. Each vendor may have unique features
incorporated in their design; however, the overall process chemistry and equip-
ment are generally very similar. Both lime and limestone produce the same
reaction products, although differences exist in the operating parameters. A
general flow diagram is shown in Figure A20-1.
In most coal-fired lime-based FGD systems, hydrated lime (CaO) is slaked
onsite to form calcium hydroxide slurry. This slurry reacts with S02 to form
calcium sulfite and calcium sulfate. In a limestone system, the chemistry is
similar; however, carbon dioxide is also generated. The limestone, often
gravel size or larger, is crushed in ball mills to produce a reagent slurry.
In the typical system, the flue gas is contacted with a slurry of calcium
sulfur salts and the reagent. The gas may or may not contain fly ash, depend-
ing on the absorber design. Venturis and Turbulent Contact Absorbers have
been used for combined S0a and particulate control while spray towers, perfor-
ated trays, and other low pressure drop contactors are often used only for
S02 removal. A large amount of slurry (relative to the gas volume) is sprayed
or dispersed in some manner in the contactor, saturating the flue gas and
removing the S02. Typical L/G ratios are 5.3-16 m3 liquid per thousand actual
m3 of gas. The scrubbed gas is then passed through mist eliminators and is
often reheated to restore bouyancy before being discharged.
The S02 rich liquor typically drains into a large tank where neutraliza-
tion and precipitation reactions occur. The alkaline reagent is added to this
tank to maintain the desired system pH. When lime is used as the reagent, the
feed liquor pH is typically 8-9 and the absorber effluent is 5-7. In a lime-
stone system, because it is buffered by the carbonate species, the pH usually
ranges between 5 and 6. Operating at this lower pH tends to increase the
A20-4
-------
O
I
Ln
S02 ABSORBER
^ r — >
REHEATER
FAN
FLUE GAS-
7
STEAM
<—MAKE-UP WATER
LIME
OR
LIME
STONE"
LIME
SLAKER
CRUSHING SLURRY
AND
GRINDING
EFFLUENT HOLD TANK
TO STACK
SECOND STAGE
SOLID-LIQUID
SEPARATOR
OR
SETTLING POND
SOLID-LIQUID
SEPARATOR
SOLID WASTE
Figure A20-1. Typical process flow diagram for lime/limestone scrubbing
-------
Appendix A20
FGD Processes
sulfite oxidation rate. As the sulfate concentration rises, the gypsum rela-
tive saturation increases. If a super—saturated condition is reached without
adequate gypsum seed crystals present in the slurry, hard scale can form on
the absorber internals, disrupting operation. In order to obtain scale-free
operation, the reaction tank is either sized to maintain the gypsum relative
saturation low enough to prevent scale formation, or the oxidation is forced
to near completion, which results in adequate crystal precipitation sites to
prevent scaling.
A small portion of the recirculating slurry is removed to control the
suspended solids concentration. This stream can be thickened, filtered, or
centrifuged before it is ultimately disposed.
2. Sodium Based Wet Systems
Approximately ten percent of commercial FGD systems are based on sodium
processes, of which there are three types, dual alkali, sodium carbonate, and
Wellman-Lord. All of these processes absorb S02 in the same manner. The
makeup sodium reagent is typically either caustic (sodium hydroxide) or soda
ash (sodium carbonate). Trona, a mineral form of sodium carbonate, is also
used.
The sodium compounds used in these processes are relatively expensive
compared to reagents such as lime or limestone. However, sodium compounds
have certain advantages. The primary advantage is that they are much more
soluble than their respective calcium salts. As a result, the liquid phase
alkalinity of the absorbing liquid permits very efficient S02 removal at low
L/G ratios. The predominant alkaline species is the sulfite ion which neu-
tralizes the sulfurous acid formed. Absorber effluent is treated with makeup
reagent to convert the bisulfites back to sulfites.
A20-6
-------
Appendix A20
FGD Processes
The high solubility of sodium salts, although good for S02 removal,
creates a disposal problem. Since the salts are soluble, direct disposal can
only be economically accomplished in arid areas where evaporation ponds can be
employed. The waste, if it is not oxidized to sulfate, could also place a
large chemical oxygen demand on any receiving waters.
The dual alkali and Wellman-Lord systems solve the soluble effluent dis-
posal problem in two different manners. In the dual alkali system, lime is
reacted with the absorber effluent to produce a sludge which is discarded.
This reaction also converts the sodium sulfite back into caustic, which is
returned to the system. In the Wellman-Lord system, the bisulfite containing
effluent is steam stripped to form sodium sulfite and gaseous S02. The S02 is
then processed into sulfur or sulfuric acid.
The dual alkali process uses alkaline sodium compounds to achieve
efficient S02 removal, and lime to regenerate the sodium-based scrubbing
liquor. The flue gas is normally treated for particulate removal before it
reaches the system. A presaturator can be included to prevent chloride build-
up in the recirculating liquor when treating high chloride coals. The dual
alkali process has two major chemistry steps, absorption and regeneration.
Sodium sulfite is the major alkaline species present in the liquid phase.
In the absorption step, caustic and soda ash react with S02 to form
sodium sulfite which, upon additional sorption of S02, forms sodium bisulfite.
Oxidation of the sulfite ion to sulfate also occurs.
A portion of this recirculating clear liquor stream is sent to a reaction
tank where lime is added to precipitate the sulfur species. The solids are
separated in a thickener and concentrated in a vacuum filter before disposal.
A20-7
-------
Appendix A20
FGD Processes
The regenerated sodium compounds are then recycled to the absorber. Alkali
losses due to the filter cake moisture are replaced by the makeup sodium rea-
gent. Most systems include some type of filter cake washing system to mini-
mize the sodium loss.
The oxidation reaction of sulfite to sulfate requires a treatment scheme
to remove the inactive sulfate from the clear liquor. Two mechanisms exist,
the selection of which depends on the relative sulfate to sulfite concentra-
tions. Both calcium sulfate and calcium sulfite are relatively insoluble
compounds, although gypsum (a form of calcium sulfate) is about an order of
magnitude more soluble than calcium sulfite. Consequently, unless the sulfate
to sulfite ion ratio in the solution is the same as the ratio of solubility
products of the two compounds, only one species should precipitate. When this
ratio is high, gypsum will precipitate as lime is added. This is called a
"dilute" system because the concentration of the active sodium sulfite is
fairly low. In a "concentrated" system, both calcium sulfite and sulfate pre-
cipitate and the liquor is subsaturated with respect to gypsum by a copreci-
pitation mechanism. The sulfite and sulfate form a solid solution in the
crystal lattice structure. This mechanism maintains the subsaturated gypsum
condition.
The Wellman—Lord process, shown in Figure A20—2, uses an aqueous solution
of Na.SO, to remove SO,. Oxidation of the sulfite ion also occurs resulting
Z 3 Z
in the formation of sulfate. The NaHS03 solution from the absorber is
thermally decomposed in a steam-heated evaporator to regenerate sodium
sulfite.
Most of the water vapor from the overhead S02-H20 mixture is condensed,
and the S02-rich gas is further concentrated to about 85 percent S02 by steam
stripping of the condensate. The SO^-rich stream is suitable for processing
into elemental sulfur or sulfuric acid. A slurry of Na2S03 crystals is formed
in the evaporator, redissolved with the condensate from the evaporator over-
head, and recycled to the absorber as regenerated scrubbing solution.
A20-8
-------
ABSORBER
O
Figure A20-2. Process flow diagram Wellman-Lord process
-------
Appendix A20
FGD Processes
Accumulation of sulfate in the system is prevented by a continuous
purge treatment of the NaHSO, solution from the absorber. Earlier Wellman-
Lord installations used a refrigerated purge treatment section to crystallize
and remove NajS04. More recent developments indicate that a high temperature
purge treatment is less energy and capital intensive. Sodium losses from the
the purge treatment are made up by addition of NaOH or Na2C03 to the recircu-
lating scrubbing solution.
3. Dry Scrubbing
Wet reagent dry systems use a solution or slurry, which is evaporated to
dryness by the flue gas, to remove SO,. This category includes the fastest
growing FGD process, spray drying. There are several advantages of wet rea-
gent dry systems over wet processes. First, as the title indicates, the flue
gas leaving the S02 absorber is not saturated with water vapor. This usually
prevents moisture condensation and subsequent corrosion in downstream equip-
ment. Another benefit is that the waste material is handled dry, rather than
in a slurry or a liquor. For throwaway applications, the dry solids can
easily be disposed of. One current disadvantage of these systems is the some-
what lower SO, removal capabilities on high sulfur coals. Commercial operat-
ing experience is also somewhat limited.
Spray drying is a relatively new FGD process rapidly gaining utility
acceptance. In spray dryer systems, hot flue gas from the boiler or air
preheater passes into a reaction vessel with a residence time of 5 to 10
seconds. Here it is contacted with a slurry, paste, or solution of alkaline
material. The flue gas is adiabatically humidified to within 30 K of its
saturation temperature by evaporation of the slurry solution. The slurry is
dried to generally less than one percent free moisture. These salts, along
with fly ash, are entrained in the flue gas and carried to the particulate
control section. This is typically a baghouse, although ESPs can also be
A20-10
-------
Appendix A20
FGD Processes
used. Some of the solids drop to the floor of the dryer and are combined with
the baghouse catch. Collected solids may be recycled through the reagent
preparation system along with solids from the particulate collector. The SO
capture reactions take place both during and following the drying process,
continuing even on the particulate collector surfaces.
There are several types of spray dryer system designs being offered. All
involve the following five steps:
• reagent preparation,
• atomization,
• droplet-gas contact,
• S02 absorption-water evaporation, and
• particulate removal.
In addition, solids recycle is integrated into most of the current system
designs. During reagent preparation, the alkaline material is dissolved or
slurried in water. Limestone has not been shown to be sufficiently reactive
for this application. Both lime and soda ash are reactive and cost effective
for spray dryer applications. At this time, however, most spray drying
systems are lime based.
Atomization is accomplished in either a rotary atomizer or with nozzles.
In a rotary atomizer, liquid is fed into a rotating wheel. The liquid is
accelerated, and is atomized at the wheel's edge forming a spray of droplets.
The spray leaves the wheel horizontally at an angle of about 180°. Droplet
size is dependent on the wheel speed, fluid viscosity, and feed rate, with
wheel speed being the most important variable.
A20-11
-------
Appendix A20
FGD Processes
In a nozzle atomizer, either the slurry or solution is fed under high
pressure (hydraulic) or a separate high pressure (two-fluid) medium is sup-
plied. In a nozzle, the feed is ejected from the orifice as a high speed film
which disintegrates into droplets. In a large module, multiple nozzles can be
installed.
S02 absorption and water evaporation occurs simultaneously, generally
within a second or two after the droplets and gas enter the dryer. In a
properly designed system, the dried product must be free flowing while the
chamber itself remains dry and free of deposits. This is achieved by con-
trolling 1) the particle size of the atomized feed, and 2) the dryer outlet
temperature. The dry salt mixture produced is usually about 70 percent
anhydrous sulfite and 30 percent anhydrous sulfate.
Particulate removal is accomplished either in a baghouse or precipitator.
Bag collectors offer sites for additional S02 removal with unreacted alkalin-
ity. Dp to 10 percent of the total S02 removal has occurred in the baghouse
in some cases. Bag fabrics are susceptible to wetting, therefore operation
must remain well above the saturation temperature. On the other hand, ESP
collectors do not achieve significant S02 removal, but they are less sensitive
to condensation. Therefore, the spray dryer can be operated closer to satura-
tion (about 11 K approach) which causes higher S02 removal in the reactor.
Some spray drying systems also incorporate solids recycle to increase
reactant utilization and in some cases also take advantage of fly ash alka-
linity. Spent solids removed from the bottom of the spray dryer are sent to
the sorbent preparation area (e.g., slurry tank). Most system designs
reslurry the product solids in a loop separate from the fresh sorbent prepara-
tion loop, and then combine the two slurries just upstream of the atomizer.
This allows closer control of the recycle slurry pH, which has been found to
impact the overall system performance.
A20-12
-------
Appendix A20
FGD Processes
4. Process Economics
Two commercially available systems, limestone and Wellman—Lord, were se-
lected as being representative of FGD processes for PCTM purposes. It is rec-
ognized that other FGD processes may offer certain technical or cost advan-
tages, but the two selected processes are believed to be "state-of-the-art" as
far as achievable levels of SO, control are concerned.
FGD costs for boilers in synfuels plants will depend upon the amount of
sulfur emissions control required to meet NSPS standards. This may vary
depending upon the amount of sulfur in the coal. For an Illinois No. 6 type
bituminous coal having a sulfur content of 3 to 3.5 percent approximately 90
percent removal is generally required. However, for low sulfur coal a smaller
amount of removal may be necessary.
FGD cost data have been developed by the EPA for electric utility steam
generating units ranging in size from 25 MWe to 1000 MWe (2). The cost vari-
ations were principally governed by 1) size of the boiler, 2) coal used, 3)
averaging time over which the plant must meet the S02 limitation, and 4) level
of control maintained.
Table A20-2 presents estimated capital investments and annual operating
costs for the limestone and Wellman-Lord FGD processes applied to a 500 MWe
unit burning a 3.5 percent sulfur coal and achieving 90 percent SO, control.
These estimates are in 1980 dollars. Capital investment estimates include
direct and indirect costs, such as purchase of equipment, installation labor
and material, engineering costs, land required for sludge, etc. Annual
operating costs include fuel costs, operating labor, and overhead expenses for
safety.
A20-13
-------
Appendix A20
FGD Processes
TABLE A20-2. COSTS OF S04 CONTROL ALTERNATIVES FOR 90% S02 REMOVAL* (2)
Capital Investment, Annual Operating Costs,
J/kWh
Limestone Scrubber 160 0.0049
Wellman-Lord Scrubber 155 0.0045
n
500 MWe unit burning a 3.5 percent sulfur eastern coal.
In order to estimate costs for synfuel steam and power generating units
it is more useful to use costs in terms of dollars per flue gas flow rate.
The NSPS costs were converted to $ per kmol/hr based on information provided
in Reference 2. For the limestone and Wellman-Lord FGD processes, the capital
costs for a 500 MWe electric unit were estimated to be Jl070 per kmol/hr and
$1040 per kmol/hr, respectively. Similarly, annual operating costs were esti-
mated to be $245 per kmol/hr and $225 per kmol/hr, respectively. When apply-
ing these costs to synfuel plants, adjustments for size can be made based upon
the six—tenths rule. Costs for FGD units requiring less than 90 percent con-
trol can be estimated by bypassing part of the flue gas and treating the
remainder for 90 percent removal so as to achieve the required S02 removal.
5. References
1. U.S. Environmental Protection Agency. EPA Dtility FGD Survey: April-June
1980. EPA 600/7-80-029c, July 1980.
2. U.S. Environmental Protection Agency. Electric Utility Steam Generating
Units - Background Information for Proposed S02 Emission Standards. EPA
450/2-78-007a, July 1978.
A20-14
-------
APPENDIX A21
NO CONTROL TECHNOLOGIES
x
1. Process Description
The techniques for controlling manmade nitrogen oxides (NO ) emissions
can vary widely depending on the type of source. The term NO is used gener-
ally to represent the gaseous pollutants nitric oxide (NO) and nitrogen
dioxide (N02). NO emission sources range from large utility boilers for
electrical power generation to small residential furnaces, from large station-
ary internal combustion (1C) and gas turbine (GT) engines for peaking electri-
cal power to automobiles and other motor vehicles, and from indirect-fired
process furnaces and heaters in the chemical manufacturing industry to nitric
acid plants. The focus of this discussion will be limited to NO controls for
stationary combustion sources in which the fuel is burned to generate heat
rather than for NO used as a reactant in the manufacture of a commercial
x
chemical product.
The developers of NO control techniques usually are the combustion
equipment manufacturers, such as boiler and engine manufacturers, and vendors
of specialty products, such as catalysts. Often such development has been
promoted by outside support (e.g., EPA, DOE, or EPRI funded contracts) and has
had an objective of meeting a regulatory requirement. Specific examples,
giving numerous developers/D.S. licensees, are included in Table A21—1.
The purpose of NO control is to reduce the emission of these oxidants
into the atmosphere. This should be accomplished without increasing other
potentially hazardous pollutants and with minimal effects on process efficien-
cy, operability, and economics. Two general approaches are available to
achieve these goals for combustion systems. One involves combustion modifi-
cations, and the other involves flue gas treatment.
A21-1
-------
TABLE A21-1. DEVELOPERS/U.S. LICENSEES OF VARIOUS LOW NOX TECHNIQUES
Name/Appl ication
Developers/0. S. Licensees
Comments
Combustion Modification Techniques
N>
M
I
OtilltY Boilers
Controlled flow/split I lute vtrliblo velocity
burner/coil will-fired
Low N0i concentric firing burner/coal
tangent!ally-fired
Premlx burner (PM)/all fuels tangentially-fIred
Dual staged burner/coal wall-fired
Primary |it (dual register oil and |t<)
bnrner/coal or oil wall-fired
Low NO combustor tyateai (LNC8)/coal wall-fired
Planetary burner with LCNS/coal wall-fired
Dual flow pulverized coal
Potter Vheeler
EPA and Coabnatlon
Engineering
Can be anbatituted for up to 320 GJ/hr heat releaae
circular or intervene burners
Baaed on delayed >lzing, field test looki promising; not yet
considered demonstrated
Mitsubishi Heavy Industries/ Utlllies one fuel rich and one fuel lean nozzle with inter-
Combustion Engineering venlng FOB
Bibcock I Illcoz
Babcock- Hitachi/Babcock
I »11 cox
Electric Power Research
Institlte and Babcock I
Illcoi0
Babcock - Bitachi/Babcock
I lilcoi
Based on PGB into prlsiary zone of dual register burner
Based on deep ataging by special furnace design
Flues of fuel-lean lower level pass through extreaiely fuel-rich
reducing flaae
Ishlksawaj ima-Harlaa Heavy Based on boundary layer system
Industrlea/Foater Wheeler
Distributed ailing burner/coal wall-fired
In-furnace NO removal process/wide possible
application (Including industrial process
coatbustors)
Larte Industrial Boilers (water-lute)
Fluldized bed combustion/coal
EPA, Energy I Environaontal Based on rich fire-ball concept for volatilizing fuel, secondary
Research Corp., Foster zone still fuel-rich and tertiary zone for completing combustion
fheeler and Babcock I Vilcox
Mitsubishi Heavy Induatries Based on secondary fuel injection, likely to require new
combustor design for fuel optimization, still in RID stage
Above utility boiler concepts are applicable to this design
Conoco, Rlley-Stoker, and
others
(Continued)
-------
TABLE A21-1. (Continued)
Name/Application
Developera/D.S. Licenaeea
Coaunonta
latertnbe and Plretube Pactaaed Bollera and Inda
Haul 11
I
U>
Modified dealgn/dietlllate oil and natural (»
LNV/dlatlllate and reaidual oil, and natural gai
LNB olld and |»
LNB/dlitilltte and reaidual oil
Varioui techniquea/coal, reaidual and dlttillate
oil, natural gaa, and other fnela
Pol it combuator/natural gaa
1,C, Englnea
Diatlllate oil, natural (ii, or deal fuel engines
Natural |ii and dlatlllatt oil
G>» Tnrblnea (GT1
Lean burn
Klch burn/quick qaency
EPA and Keeler Daea FOB, commercially available, can be retrofit on aome nnlta
Nippon and Ti»/Clviltech Controlled mixing
Coen Staged combuatlon, commercially available
EPA and Energy and Environ- Staged combustion, atlll in BID stage
mental Beaearch Corp.
— Befer to Table A21-2 for variona techniqnea, applicable concepta
in varioua atagea of BID or commercially available
Gaa Beaearch Inatltute and High efficiency condensing flue gas systema, atlll in BID
Battelle Columbus Laboratoriea
Befer to Table A21-4, applicable conceptt In varioui ataiei of
BID or commercially available
EPA and A.D. Little, Inc.
Variona OT manufactnrera
External FOB and catalytic treatment, combined vlth torch
Ignition for gaa and water fuel emnlaiona for oil. in RID etage
Deeful on clean fnela only
EPA, Pratt I Whitney, and Can obtain JO ppmv NO at 15% air with high nitrogen oila,
NASA KID atage *
Flue Oaa Treatment Ttchnlauea
Utility Boiler and Lane Industrial Boilert (water-tube)
Thermal De-NO Bxion
x
Selective catalytic removal
t.C. Engtnet
Catalytic reduction
Commercially available
Widely demonstrated In Japan
Engelhard and other catalyat Bequirea low oxygen level in exhauat
manufacturera
Nhlle EPA aaalated development ia available without licenae, application la tangential deaign
Demonatrated in Japan, conault D.3. aupporter for availablity
^Poaalbly itill In BID atage, new furnace deaign
n>r«anmably diatillate and realdual oil, but not atated
-------
Appendix A21
NO Controls
x
Combustion modification (CM) is a means of changing flows, flow patterns,
or temperature of the fuel and oxidant or of combustion product intermediar-
ies, generally prior to free radical recombination, but always prior to con-
vective heat extraction. These changes, leading to lower temperatures and
reduced availability of oxygen at critical stages of the combustion process,
result in less NO being formed. Table A21-2 lists and briefly describes a
number of presently available CM techniques. Flue gas treatment (FGT) is a
post combustion process in which one or more species of the flue or exhaust
gas react, often with an added component, to achieve chemical reduction of the
nitrogen species. For NO FGT control, the flue gas NO is usually reacted
x x
with added ammonia. Table A21-3 includes some presently available FGT
techniques along with fuel changes and fuel additives.
The industrial usage of these NO control techniques varies widely
because of the many regulations affecting NO emissions and the costs asso-
ciated with their development and implementation. Because low excess air
(LEA) operation saves fuel and is usually easy to implement, its use is wide-
spread. On the other hand, the application of FGT to large boilers and
furnaces is generally only found in Japan and California, both of which have
stringent NO emission limits, and the cost of this approach is relatively
high. Combustion modification is generally more cost effective.
2. Process Applicability
FGT may be applied to almost any source, and is restricted only by pos-
sible space limitations and economic considerations. While some CM techniques
may be applied with relative ease, others should not be used on certain
sources. This is especially true for retrofit applications, where space
limitations or combustion system design make their application and performance
A21-4
-------
TABLE A21-2.
COMBUSTION MODIFICATION TECHNIQUES FOR N0x CONTROL
Description of
Control Technique Technique
Low Excess Air Combustion air is
(LEA) reduced to the mini-
mum amount required
for complete combus-
tion while maintain-
ing proper stream
temperature
Efficiency
(as % of
NOX Reduction)
0 - 25
5 - 25
0-28
0 - 24
5 - 35
Type of
Fuel Fired
Pulverized
coal
Stoker coal
Residual oil
Distillate
oil
Natural gas
Range of
Application
Excess Oj lowered
to 5.2* On the
average
Excess 02 limited
to 5-6* minimum
Excess Oy can be
reduced to 2.5i
Excess 02 can be
reduced to 3.0%
Stage of Development
Available but imple-
mented on a limited
basis only
Available now but
need R&D on lower
limit of air
Available
Available
Limitations and Comments
Limited by increase in CO,
HC and parti cul ate emis-
sions. Increase in boiler
efficiency may be achieved
as a benefit
Danger of overheating grate,
clinker formation, corrosion
and high CO emissions.
Added benefits include in-
crease in boiler efficiency.
Limited by increase in CO,
HC, and TSP emissions
Generally practical because
of Increased boiler effi-
ciency. Best NOX reductions
reported for large multi-
burner units
ho
M
I
Staged Combustion
Overfire air
Injection (OFA)
Staged Combustion
Air (LEA + OFA)
Injection of air above 5-30 Pulverized
the top burner level coal
through OFA ports to-
gether with a reduc-
tion in air flow to
the burners (staged
combustion)
Reduction of under 5-25 Stoker coal
grate air flow and
increase of overfire
air flow
Fuel rich firing 20 - 50 Residual oil
burners with secon- 17 - 44 Distillate
dary combustion oil
air ports
Injection of secondary 5-46 Gas
air downstream of the
burner(s) in the direc-
tion of the flue gas
path
Burner stoichio-
metry as low as
100%
Excess 02 limited
to 5% minimum
70-90% burner
stoichiometry can
be used with proper
burner installation
70-90X burner
stoichiometries can
be maintained.
Applicable to all
units; however,
requires extensive
equipment modifi-
cation
Available but not
demonstrated
Most stokers have
OFA ports as smoke
control but may need
better air flow con-
trol devices
Technique 1s appli-
cable on packaged
and field-erected
units. However, not
commercially avail-
able for all design
types
Technique is still
experimental espec-
ially for small
firetube and water-
tube units
Limited by possible increase
in slagging and corrosion.
Excess air may be required
to ensure complete combus-
tion, thereby decreasing
efficiency
Overheating grate, corrosion,
and high CO emissions can
occur 1f under grate air
flow is reduced below accept-
able level as in LEA
Best implemented on new
units. Retrofit is probably
not feasible for most units
especially packaged ones.
Found to be less effective
on firetube boilers than
watertube boilers. Gen-
erally less effective for
gas-fired units
(Continued)
-------
TABLE A21-2. (Continued)
>
I—"
I
Efficiency
Description of (as % of
Control Technique Technique NOx Reduction)
Air and Fuel One or more burners 27 - 39
Rich Firing fired on air only.
Remainder of burners
firing fuel rich.
Flue Gas Recirculation of the 0-20
RecTrculation flue gas to the burner
(FGR) windbox
15 - 30
58 - 73
48 - 86
Reduced J_oaji Reduction of fuel and Up to 45%
TRL) air flow to the burner
Type of
Fuel Fired
Pulverized
coal
Pulverized
coal
Residual oil
Distillate
oil
Natural gas
Pulverized
coal
Range of
Application
Boilers must have
a minimum of 4
burners or
designed with
excess burners
A maximum of 25%
of the flue gas
can be redrcu-
lated
Up to 25-30% of
flue gas recycled.
Can be implemented
on all design
types .
Flue gas recircu-
lation rates pos-
sible up to 45%.
Technique is
applicable to all
boiler types
except ones equip-
ped with ring
burners.
Load may be re-
duced to 2531 of
capacity
Stage of Development
Available, but engi-
neering refinement 1s
needed prior to
Implementation
Not offered because
the method is com-
paratively ineffec-
tive
Available. Requires
extensive modifica-
tions to the burner
and windbox.
Available now. Best
suited for new boilers.
Retrofit application
would result 1n exten-
sive burner modifica-
tions.
Available, but not
Implemented because
of negative opera-
tional impacts
Limitations and Comments
Load reduction required in
most cases. Possible in-
creased slagging and corro-
sion. New boiler design
will incorporate the re-
quired number of burners.
Flue gas recirculation
lowers the bulk furnace gas
temperature and reduces 0~
concentration in the com-
bustion zone. Requires
installation of flue gas
recirculation ducts and
fans. Hay cause combustion
instability.
Best suited for new units.
Costly to retrofit. Pos-
sible flame Instability
at high FGR rates.
Flame instability problem
1s not severe except for
ring burners. Minor burner
modifications can guarantee
stable flames. Most effec-
tive on watertube units.
Best used with Increase in
firebox size for new
boilers. Load reduction
may not be effective be-
Average
Stoker coal
Load may be re-
duced to 25%
Available
cause of Increase 1n
excess Oj.
Only for stokers that can
reduce load without in-
creasing excess air. Not
a desirable technique be-
cause of loss in boiler
efficiency.
(Continued)
-------
TABLE A21-2. (Continued)
Control Technique
Low NO Burners
X
(LNB)
Efficiency
Description of (as % of
Technique NO* Reduction)
33* decrease to
25% increase
31% decrease to
17% increase
32% decrease to
82% increase
New burner designed 45 - 60
to utilize controlled
air-fuel mixture
Type of
Fuel Fired
Residual oil
Distillate
oil
Natural gas
Pulverized
coal
Range of
Application
Applicable to all
boiler types and
sizes . Load can
be reduced to 25%
of maximum
Tests to 20% of
rated capacity.
Applicable to all
units.
Prototypes are
limited to size
ranges >30 Mw
Stage of Development
Available now as a
retrofit applica-
tion. Better imple-
mentation with im-
proved firebox
design.
Technique available.
However, retrofit
application is not
feasible due to
initial low load
factor of industrial
units.
Development stage pro-
totypes are available
from major boiler
Limitations and Comments
Technique not effective when
it necessitates an increase
in excess 02 levels, RL
possible implemented in new
designs as reduced combus-
tion intensity (enlarged
furnace plan area)
Least effective on firetube
boilers because of lower
combustion Intensity.
Applicable for new watertube
units with Increased firebox
size.
Low NOX burners could main-
tain the furnace 1n an oxi-
dizing environment to mini-
manufacturers
>
I—"
I
mize slagging and reducing
the potential for furnace
corrosion. More complete
carbon utilization results
because of better coal/air
mixing 1n the furnace.
Lower 02 requirements may
be obtained with all the
combustion air admitted
through the burners.
20 -
20 -
20 -
NH3 Injection Injection of NH3 into 40 -
convective section of
the boiler
40 -
40 -
40 -
50
50
50
60
70
70
70
Residual oil
Distillate
oil
Natural gas
Pulverized
oil
Residual oil
Distillate
oil
Natural gas
New burners
described gener-
ally applicable
to all boilers.
More specific
information needed.
NH3 injection rate
limited to
£"••
Applicable for
large package and
field-erected
watertube boilers.
Not feasible for
firetube boilers.
Commercially offered
but not demonstrated
Commercially offered
but not demonstrated
Commercially offered
but not demonstrated
Specific emissions data
from Industrial boilers
equipped with LNB are
lacking
Limited by furnace geometry.
Performance Is sensitive
to flue gas temperature
and residence time at
optimum temperatures. By-
product emissions such as
ammonium bisulfate could
cause operational problems.
Some increased maintenance
of air heater/economizer
parts might be necessary
when burning high sulfur
oil . Technique is very
costly. Should have fewer
problems when firing
natural gas.
-------
TABLE A21-3.
NO FLUE GAS TREATMENT CONTROL ALTERNATIVES FOR BOILERS
x
I
oo
Control
Technique
Selective
Catalytic
Reduction (SCR)
- Fixed Packed
Bed Reactors
Selective
Catalytic
Reduction (SCR)
- Moving Bed
Reactors
Selective
Catalytic
Reduction (SCR)
- Parallel Flow
Reactor
Absorption-
Oxidation
Selective
Catalytic
Reduction
- N0x/S0x
Removal
Description
of
Technique
Utilizes NH3 to
selectively reduce
NOX to N2
Utilizes NH3 to
selectively reduce
NO- to N2
X C
Utilizes NH3 to
selectively reduce
NO, to N?
X t
Removes NOX from
flue gas by absorb-
ing the NO or NOX
into a solution
containing an oxi-
dant which converts
the NOX to a
nitrate salt
Utilizes NHj to
catalytically reduce
NOX after SOX is
absorbed and reacted
with catalyst
Principle
of
Operation
Reactor contains
ring shaped catalyst
pellets packed in
fixed bed
Reactor contains
catalyst (rings or
pellets) gravity-bed,
mechanically- screened.
and returned to
reactor
Reactor contains a
special catalyst
arrangement (honey-
comb, parallel plate
or tubes)
Use of gas/1 iqind
contactors. Per-
forated plate and
packed towers accom-
plish NOX absorption
by generating high
gas/liquid inter-
facial areas
Reactor and catalyst
removes both NOX and
SO?, uniquely designed
parallel flow reactor
used to avoid parti-
culate problems
Efficiency
(% as
NOX Reduction)
Up to 90%
Up to 90%
Up to 90%
The relative
insolubility
of NO in water
will prohibit
a high effi-
ciency
80% NOX reduc-
tion, 90% SOX
reduction
(theoretical )
Applicability
Applicable only
to flue gas
streams contain-
ing particulate
emissions of
less than 20
mg/Nm3 (hot side
ESP required)
Applicable only
to flue gas
streams contain-
ing less than
1 g/Nm3 parti -
culates
Testing currently
under way for
high particulate
flue gas
No published
information
available
Should be appli-
cable to high
particulate flue
gas
Stage of
Development
Commercially avail-
able for natural
gas-fired boilers
only at this time
Has been applied
in Japan to several
oil-fired indus-
trial and utility
boilers only
Had been applied in
Japan to several oil-
fired industrial and
utility boilers only.
Applicability to coal
fired boi lers cur-
rently being tested
by EPA
No coal-fired tests
have been made
No continuous coal-
fired NOX removal
test data for NOX/
SOX systems are
available
Limitations and Comments
Although it is possible to
install a hot ESP to reduce
the particulate level to 20
mg/Nm3, this is expensive
and not always effective.
Fixed bed SCR systems are
not considered for applica-
tion to coal-fired boilers.
Although it is possible to
install a hot ESP to reduce
the particulate level to
1 g/Nm3 this 1s expensive
and not always effective.
Moving bed SCR systems are
not considered for appli-
cation to coal-fired
boilers.
Greatly reduces particulate
impaction as gas flow is
parallel to catalyst surface.
Unreacted NH3 downstream can
react with SOg or S03 to
form airmonium bisulfate or
the NH3 could enter FGD and
ESP equipment.
The presence of particulates
requires a prescrubber.
The presence of SO,, requires
FGD pretreatment. Increased
NOx concentration requires
a larger column height and
increased oxidant concentra-
tion. Nitrate salts formed
as a secondary pollutant.
System 1s not affected by
changes 1n the boiler gas
flow rate or particulate con-
centrations. Changes in NOX
concentration because of
boiler load changes may be
compensated for by conven-
tional control system used
with the NHa Injection
equipment.
(Continued)
-------
TABLE A21-3. (Continued)
Control
Technique
Adsorption
NOx/SO,
Removal
Electron Beam
Radiation
NOx/SOv
Removal
Adsorption-
Reduction
NOx/SO.
removal
Oxidation-
Absorption-
Reduction
NOX/SOX
Oxidation-
Absorption
Description
of
Technique
Adsorbed NOX is re-
duced to N2 while
50-2 is reduced and
condensed to ele-
mental S
A dry process that
utilizes an electron
beam to bombard the
flue gas, thereby
removing NOX and
S02
Simultaneously re-
moves NOX and S02
from flue gas by
absorbing them into
a scrubbing solution
Simultaneously re-
moves NOX and S02
from flue gas by
oxidizing NO to N0.2
and then absorbing
N02 and S02 into a
scrubbing solution
Excess 03 is used to
selectively oxidize
NOX to NgOs
Principle
of
Operation
The adsorption process
removes NOX and S02
from flue gas by
adsorbing them onto
a special activated
char
Flue gas is taken from
the boiler air pre-
heater and passed
through a cold ESP to
remove parti culate.
NH3 is added and the
gas is then bombarded
with an electron beam.
Based on the use of
chelating compounds
complexed with iron
to "catalyze" the
absorption of NOX
Based on the use of
gas-phase oxidants,
either OT or C102,
to selectively oxi-
dize NO to N02
N205 formed by oxi-
dation is absorbed
into aqueous solu-
tion and concentrated
to form a 60* HN03
by-product
Efficiency
(% as
NOX Reduction)
40-601 NOX
reduction.
80-95* SOX
reduction
Removal effi-
ciency will
decrease as
NOX and S02
Increase
60-70% NOX
reduction,
90% SO?
reduction
90% NOX reduc-
tion,
95% S02 reduc-
tion for oil-
fired tests
Not available
Applicability
Hay be applicable
to handle coal
By-product treat-
ment technology
needs to be more
fully developed
before commer-
cialization
Applicable only
to high sulfur
coals
Not applicable to
low sulfur coals
Hay be applicable
to handle high
particulate flue
gas
Stage of
Development
Presently in the
prototype unit stage
of development
No coal-fired tests
have been performed
at this time
Preliminary testing
stage of development
Prototype stage of
development. No
coal-fired flue gas
tests have been per-
formed at this time.
One coal -fired test
has been performed
with no published
information
Limitations and Comments
Very complex process. Numer-
ous process steps involve
hot solids handling with
numerous mechanical problems
possible
NOX/S02 removal will drop
off drastically at low radia-
tion doses based on oil-fired
pilot tests. Sulfate and
nitrate salts as well as
other ionic species formed
as by-products
Requires large adsorbers with
high liquid rates. Absorbing
solution is highly corrosive
sulfate and nitrate salts
formed as secondary pollutants
Costly gas-phase oxidants
create secondary wastewater
pollution problems. The use
of C102 Introduces a chloride
pollutant problem.
Production of nitrate salts
poses a potential secondary
pollution problem. Corrosion
problems.
-------
Appendix A21
NO Controls
x
difficult or impossible. While flue gas recirculation (FG8) and other tech-
niques aimed primarily at reducing peak flame temperature may provide signif-
icant NO^ reductions with clean fuels, their use alone is not effective for
high-nitrogen fuels. Generally, use of staged combustion, especially advanced
low NO^ burner (LNB) designs, can be extremely effective with high nitrogen
fuels. For these reasons, the application of CM should be considered almost
on a case-by-case basis. In addition, increased or alternate pollutant for-
mation caused by the application of NO controls may be possible. For
example, the degree of NO reduction in certain FGT processes is partially
determined by ammonia break-through. This possibility of other emissions due
to newly emerging CM techniques is an important consideration. Previous
environmental assessment studies showed that, in general, current state-of-the-
art CM techniques caused minimal increases in other pollutants as long as CO
was less than 200 ppmv. However, the application of new techniques may gener-
ate other emissions such as polycyclic organic mater (POM) and nitrated poly-
cyclic aromatic hydrocarbons (PAHs). Similarly, although concern over poten-
tial operating problems, such as corrosion with LEA and staging, has lessened
with recent testing of present techniques, such problems should be evaluated
for new techniques. A summary of NO control techniques and their applica-
bility is provided in Table A21-4.
3. Process Performance
The operating principles of CM may be defined in terms of four variables:
time, temperature, turbulence, and concentration. Yet the application of
these variables must be made for a wide range of combustors and fuels (gas-
eous, liquid, solid, clean, and dirty, with varying levels of fuel nitrogen
and other contaminants). An added complication is that the fuel nitrogen has
A21-10
-------
TABLE A21-4. APPLICABILITY OF NOX CONTROL TECHNIQUES
Control Technique
Application and Discussion
Techniques Reducing Overall or Peak Combustion Temperatures
Derate or load reduction
Flue gas recirculation (FGR)
- Exhaust gas recirculation (EGH)
Reduced air preheat (RAP)
- Manifold air temperature
(reduction)
Air-to-fnel ratio increase (A/F)
Retard
Water or steam injection
Catalytic combustion
Techniques Controlling Oxygen Availability
Low excess air (LEA)
Derate may result in increased capital cost.
Fans and controlling operation often increase costs.
External or internal EGR applied to 1C engines.
Efficiency loss associated with this technique can
largely be recovered by economizer in new designs.
This is widely used on large turbo-charged 1C engines
with slight decrease in efficiency.
This is used on 1C engines - more successfully on gas
and dual-fired than diesel.
Retard is used on 1C engines and provides cooling via
increase in integrated combustion volume and surface
area.
Water or steam injection is widely used on GTs; water
injection used on 1C engines.
This technique offers extremely low NOX levels due to
low combustion temperatures (still in RID stage),
presently limited to clean fuels.
LEA increases efficiency, but may be restricted with
existing burners and combustors by CO, smoke, and
possible POM increases.
(Continued)
-------
TABLE A21-4. (Continued)
Control Technique
Application and Discussion
Off-stoichiometric combustion (OSC)
- Bias burner firing (BBF)
- Stratified combustion
- Various LNB and LNCS designs
Reburning (or secondary fuel
injection)
Staging
- Burners out of service (BOOS)
- Overfire air (OFA)
- Air lances
- Various LNB and LNCS designs
OSC has both fuel rich and fuel lean zone(s), allowing
possible homogenous NO reduction.
Lower burners are operated fuel rich, and upper burners
are operated fuel lean.
Technique demonstrated on small 1C engines.
See PM burner and Planetary burner/LNCS in Table A21-1
for examples.
This is done by addition of fuel to make a secondary
fuel-rich zone downstream of primary combustion process
followed by air for burnout (still in RflD), likely
requires new combustors for optimization. See
in-furnace NO reduction in Table A21-1 for example.
Staging has primary and sometimes secondary (if
tertiary air used) combustion zones which are fuel
rich, followed by additional air for burnout.8
Staging achieved by injecting air rather than fuel/air
in selected upper burners. BOOS may result in load
reduction due to pulverizer capacity, or may
intentionally be a technique for load reductions.
OFA can be applied to new or retrofit boilers,
including stokers.
This technique has resulted in significant NOZ
decreases in process heaters and furnaces.
See Table A21-1 under utility boilers for numerous
examples.
(Continued)
-------
TABLE A21-4. (Continued)
Control Technique
Application and Discussion
I
M
U>
Fluidized bed combustion (FBC)
Techniques Involving Flue Gas Treatment
Catalytic reduction
Selective catalytic reduction (SCR)
Non-catalytic reduction
Techniques Involving Other Processes
Fuel additives
Fuel changes
Bed mixing causes variation in oxygen availability,
heterogenous NOX reduction, and reduced combustion
temperature.
Technique is limited to 1C engines with low oxygen
content in exhaust.
SCR utilizing catalyst normal with ammonia injection
alone to reduce NO ; significant NO reductions
possible. x x
See Thermal De-NO process described in text.
Technique is not effective for directly reducing NO
and may result in byproduct emissions. However,
additives for corrosion, smoke, etc., may increase
flexibility of certain CM and FGT techniques.
Technique generally is not cost effective. Possible
exceptions include fuel nitrogen decreases associated
with desulfurization, and gasification and indirect
liquification of coal.
Turbulence and mixing make this separation between OSC and staging not fully accurate.
-------
Appendix A21
NO Controls
x
varying degrees of volatility in different fuels. Thus a control technique
developed for a fuel containing highly volatile nitrogen species may not be
nearly as successful on a fuel containing significant fuel nitrogen which is
preferentially char-bound. Techniques that reduce peak combustion temper-
ature, while generally successful with low-nitrogen fuels, usually do not work
as effectively when applied to high-nitrogen fuels. This is due to the low
activation temperature of fuel nitrogen fragments when compared to nitrogen
fixation. Consequently, techniques that control oxygen availability and
mixing are applied to high-nitrogen fuels and can be applied to low-nitrogen
fuels, if cost or other factors justify it.
In FGT, NO can be reduced to nitrogen (N2) with or without a catalyst.
In noncatalytic reduction processes, a reducing agent (normally ammonia) is
added. For example, the Thermal De-NO process utilizes ammonia which is
injected into the flue gas to reduce most of the NO to nitrogen. In some
cases varying amounts of hydrogen are also added to allow for a wider tempera-
ture range within which significant NO reductions are possible with minimal
ammonia emissions. However, in most FGT processes, a catalyst is used (often
required) to promote this NO reduction. Nonselective reduction catalysts,
requiring no added component, can be used with exhaust gas having very little
oxygen (e.g., a rich-running 1C engine). Usually nonselective catalytic reduc-
tion is more beneficial in significantly decreasing carbon monoxide and hydro-
carbon emissions than in reducing NO . Most flue and exhaust gases would
require that a reducing agent (e.g., ammonia) be added for selective catalytic
reduction (SCR). SCR has been developed and widely demonstrated in Japan.
Typical exhaust (flue) stream emissions along with NO variations for
uncontrolled processes and for varying degrees of control are given in Tables
A21—5 and A21-6. The major stream components of nitrogen, carbon dioxide, and
oxygen depend primarily on excess air levels and partially on fuel compo-
sition.
A21-14
-------
TABLE A21-5. AVERAGE CRITERIA POLLUTANT EMISSIONS, INCLUDING NO^ RANGE
Equipment /Fuel
Utility and Larae
Tangential/
Bituminous Coal
Wall-Fired/
Bituminous Coal
Wall-Fired/
Residual Oil
Cyclone/
Bituminous Coal
Spreader Stoker/
Bituminous Coal
Package Water-tube
NO/
Emission
SO/
Factors in ng/J
d
Particulate CO
HC
Industrial (Water-Tube) Boilers
285
(100-502)
430
315
(190-350)
680
(up to 1170)
265
Boiler
Chain Grate Stoker/ 150
Bituminous Coal
Residual Oil
Distillate Oil
Natural Gas
Package Fire-tube
Residual Oil
115
(95-170)
55/900®
(45-102)
45/1106
(30-190)
Boilers
115
(95-170)
702S
602S
482 S
679S
679S
757S
482 S
434S
482 S
195A 11.2
186A 21.9
30. 5S 13.0
+8.6
35. 7A 18.1
233A 35.7
233A 25
30. 5S 3.4
+8.6
8.2 1.6
3.4 8.6
30. 5S 3.4
+8.6
0.9
0.9
0.9
6.4
5.6
4.3
0.9
0.4
1.7
0.9
(Continued)
A21-15
-------
TABLE A21-5. (Continued)
b .
Emission Factors in ng/J
Equipment/Fuel
NO
SO
Particulate
CO
HC
Package Fire-tube Boilers (continued)
Distillate Oil
Natural Gas
Gas Turbines
(15 MW)
Distillate Oil
Natural Gas
70 434S
40 0.3
(28-55)
365 10.7
195 2.2
(170-220)
7.3 1.6 0.4
2.6 8.6 1.7
16.0 47.0 8.6
6.0 49.0 8.6
1C Engines
Distillate Oil 1700 95.9
Natural Gas 1500 0.2
Dual 1100 —
Process Heating
Oil 154.8 627S
Gas 70.1 860S
103
78.4
8.6
313
177
nil
nil
115
555
13.1
12.9
Very abbreviated, see Ref. (1) for comprehensive list of equipment sizes and
.fuel types.
NO^ as N0a, SO^ as S02, and HC as methane. See Appendix F of Ref. (2) for
conversion factors to ppmv at 3% 02 for typical fuels.
Q
Values cited are updated and from Ref. (3) except for process heating
emissions. Values in parentheses are the range of NO emissions reported in
the listed references. While other sources also have a range of values
around the average, no convenient basis for comparison exists.
S represents weight percent sulfur in fuel; A represents weight percent ash
in fuel. For example S02 emissions from a tangentially-fired, water-tube
boiler (bituminous coal with 3% sulfur) are 702 x 3 equals 2106 ng/J.
First number cited is average emission for units without air preheater;
second, with air preheater.
Fuels not specified, natural draft refinery process heater data given.
A21-L6
-------
TABLE A21-6. NO REDUCTIONS ACHIEVABLE THROUGH VARIOUS CONTROLS
>
I—'
I
Equipment Type Fuel
Utility Boiler/Industrial Boiler
Tangential Coal
Tangential, >73 MW
Opposed Wall and Coal
Wall Firing (dry)
Cyclone Coal
Opposed Wall and Coal
Wall Firing (wet)
Tangential Oil
Dtilitv Boiler
Opposed Wall and Oil
Wall Firing
Application
N,R
N,R
N,R
N,R
N,R
N,R
N,R
N
R
R
R
R
R
R
R
Control Level
(ng/J)
280
215
129
430
258
215
172
129
86
60
680
610
790
550
215
129
86
43
215
129
86
43
Control Strategy
Control Approach
Baseline
OFA
OF A + TDN
Baseline
OFA
LNB
LNB + OFA
LNB + OFA + TDN
Advanced LNB
Advanced Burner/
Furnace Concepts
Baseline
LEA
Baseline
BOOS
Baseline
BOOS, OFA
BOOS, OFA + FGR
Above MC + TDN
Baseline
BOOS, OFA or LNG
(BOOS, OFA, or LNB)
+ FGR
Above CM + TDN
Data
Available
Now
1983
Now
Now
Now
1983
1985
1985
Now
Now
Now
Now
Now
Now
Now
(Continued)
-------
TABLE A21-6. (Continued)
I
M
CO
Equipment Type Fuel Application*
Industrial
Stoker Watertube Coal
- Spreader
- Chain Grade and
Underfeed
1C Engine
Spark Ignition Gas
> 75 kW
Compression Ignition Oil
>75 kl
Dual Fuel
Gas Turbine
Simple Cycle, > 15 MW Oil
N.R
N,R
N,R
N.R
N.R
N,R
N,R
N.R
N,R
N.R
N.R
N.R
N.R
N.R
N
N
Control Level
(ng/J)
265
215
172
129
150
129
1500
1200
1100
900
1700
1400
1200
1000
1100
900
800
700
310
130
130
10
Control Strategy
Control Approach
Basel ine
LEA + OFA
TON
LEA + OFA + TON
Baseline
LEA
Baseline
EGR
Retard
A/F Increase
Baseline
EGR
Retard
A/F Increase + Retard
Baseline
EGR
Retard
A/F Increase
Baseline
Water Injection
Dry Control
Catalytic Combustion
Data
Available
1983
1983
1983
Now
Now
Now
Now
Now
Now
Now
Now
Now
Now
Now
1985
1990
(Continued)
-------
TABLE A21-6. (Continued)
Equipment Type
Fuel
Application,
Control Level
(ng/J)
Control Strategy
Control Approach
Data
Available
Industrial Processes
Glass Melting Furnaces Various
Cement Kilns
Refinery Heaters
Varies
(15% reduction
typical)
Baseline
LEA
Now
N is new; R is retrofit
TON stands for Thermal De-Nox; other abbreviations spelled out in Table A21-4
-------
Appendix A21
NO Controls
x
4. Secondary Waste Generation
One of the major advantages of CM is that no secondary wastes are gen-
erated. Catalytic F6T does involve disposal of used catalyst or possible
generation of wastes and other emissions during catalyst regeneration. How-
ever, problems associated with catalyst regeneration and disposal, usually
required once every two years, are expected to be minor.
5. Process Reliability
The implementation of CM or FGT is usually characterized by uniform emis-
sion reductions on a particular unit and by few operating problems as long as
the units and controls are properly operated and maintained. This experience
is based on having skilled and experienced personnel make any necessary equip-
ment modifications. For example, all EPA work involving the implementation of
major CM on large boilers has been done in conjunction with the boiler manu-
facturer, and the boilers have been operated during tests by experienced plant
operators.
6. Process Economics
With the possible exception of LEA operation, the implementation of CM or
FGT will involve capital costs due to retrofit or to new or added equipment.
In some cases additional operating costs are also involved. It may be pos-
sible to partially offset or more than recover these costs through increased
efficiency or the use of less expensive fuels, e.g., coal in place of fuel
oil. Information on the capital and operating cost of the various CM and FGT
techniques is given in References 4 through 12. Tables A21-7 and A21-8 give
examples of cost data for selected utility boiler applications.
A21-20
-------
TABLE A21-7. CAPITAL AND OPERATING COSTS FOR UTILITY BOILER RETROFIT
N0x CONTROL
Annualized Annualized
Initial Indirect Direct Total Cost
Investment, Operating Cost, Operating Cost, to Control,
Boiler/Fuel Type */kW i/kW-yr i/kW-yr J/kW-yr
Tangential/Coal-Fired:
OFA 0.90
0.21
0.32
0.53
Opposed Wall/Coal-Fired:
OFA
LNB
BOOS
Sinele Wall/Oil-
BOOS
FGR/OFA
0.62
2.03
0.08
and Gas-Fired:
0.30
5.71
0.16
0.34
5.34
0.05
1.14
0.52
0.06
24.78
0.44
1.91
0.69
0.40
30.12
0.49
3.09
TABLE A21-8. NO LEVELS AND TOTAL COSTS FOR WALL-FIRED UTILITY BOILERS
Fuel/NO Emission Level, Recommended Control
Cost to Control,
t/kW-yr
Coal
Oil:
Gas :
ng/J
301
258
215
172
129
86
129
86
43
Requirement
OFA
OFA
LNB
OFA + LNB
BOOS
FGR + OFA
BOOS
FGR + OFA
FGR + OFA
Retrofit
0.50 to 0.70
0.50 to 0.70
0.40 to 0.50
0.95 to 1.20
0.50 to 0.60
3.00
0.50 to 0.60
3.00
3.00
New Boiler
0.10 to 0.20
0.10 to 0.20
0.30 to 0.40
0.40 to 0.50
A21-21
-------
Appendix A21
NO Controls
7. References
1. K.G. Salvensen. et al. Emission Characterization of Stationary NOX
Sources: Volume 1. Results. EPA-600/7-78-120a, June 1978.
2. E.H. Manny. Guidelines for NOi Control by Combustion Modification for
Coal-fired Utility Boilers. EPA-600/8-80-027, May 1980.
3. L.R. Waterland, et al. Environmental Assessment of Stationary NOX
Control Technologies (Final Report). EPA-600/7-82-034, May 1982.
4. K.J. Lim, et al. Environmental Assessment of Utility Boiler Combustion
Modification NO Controls: Volume 1. Technical Results. EPA-600/7-80-
075a, April 198(5.
5. K.J. Lim, et al. Industrial Boiler Combustion Modification NOZ
Controls: Volume 1. Environmental Assessment. EPA-600/7-81-126a, July
1981.
6. U.S. Environmental Protection Agency. Technology Assessment Report for
Industrial Boiler Applications: NO Combustion Modification.
EPA-600/7-79-178f, December 1979. *
7. U.S. Environmental Protection Agency. Combustion Modification Controls
for Stationary Gas Turbine: Volume 1. Environmental Assessment. EPA-
600/ 7-81 -122a, July 1981.
8. U.S. Environmental Protection Agency. Environmental Assessment of
Combustion Modification Controls for Stationary Internal Combustion
Engines. EPA-600/7-81-127, July 1981.
9. U.S. Environmental Protection Agency. Emissions from Refinery Process
Heaters Equipped with Low-NO Burners. EPA-600/7-81-169, March 1981.
10. U.S. Environmental Protection Agency. Assessment of the Need for Flue
Gas Treatment Technology. EPA-600/7-78-215, November 1978.
11. U.S. Environmental Protection Agency. Evaluation of the Advanced Low-
NO Burner, Exxon, and Hitachi Zosen DeNOx Processes. EPA-600/7-81-
125, July 1981.
12. Mobley, J.D. Assessment of NO Flue Gas Treatment Technology.
IERL-RTP-1084. x
A21-22
-------
APPENDIX Bl
GBAVITY SEPARATION
1. Process Description
Gravity separation is a wastewater treatment process which relies upon
the different densities of immiscible oil, water, and solids for successful
operation. The wastewater stream is fed to a vessel which provides a quies-
cent zone where the various phases separate. Oils and solids with specific
gravities less than that of water float to the top of the aqueous phase while
heavy sludges and solids sink to the bottom of the vessel.
The design of oil—water separators used in refineries has been well de-
fined in studies by the American Petroleum Institute (API)(1). In API separa-
tors, the influent wastewater passes through trash bars and a skimmer to re-
move floating oil before entering the quiescent zone of the separator. In
this quiescent zone, the wastewater velocity is kept very low to prevent any
turbulent mixing. An adjustable weir at the end of the separator divides the
waste into aqueous and organic phases. A rotating skimmer is used to remove
the organic phase from the surface. The bottom of the separator contains
slowly moving panels which convey settled solids to a pump.
The other commonly used gravity separator design is the corrugated plate
interceptor (CPI). A CPI consists of a number of parallel corrugated plates
mounted from 2 to 4 cm apart at an angle to the horizontal. Between 12 and 48
plates are typically used. Wastewater flows downward between the plates, with
the lighter oil droplets floating upward into the tops of the corrugations
where they coalesce. The oil droplets move up the plates to form a floating
layer that is skimmed from the surface of the process tank. Laminar flow
between the plates enhances separation of oil and water (2).
Bl-1
-------
Appendix Bl
Gravity Separation
Gravity separators operate best when the densities of the materials to be
separated are very different. Separation is poor when the densities are about
equal or if emulsions are present which do not break down easily.
2. Process Applicability
Gravity separation is generally used as an initial treatment step for the
removal of nonemulsified free oils and grease from wastewaters. API and CPI
separators are widely used in processing oily wastes and wastewaters from pet-
roleum refining, metal, and food processing. Petroleum refining produces
large quantities of wastewater contaminated with oil from leaks in heat
exchangers and storm water runoff from process areas. In the metals indus-
try, cold rolling rinse and coolant waters and metal-working rinse waters may
contain several thousand mg/L of free oil. Wastewaters from meat, fish and
poultry slaughtering, cleaning, and byproduct processing (such as stream ren-
dering) may also contain several thousand mg/L of grease and/or oil (3). Gra-
vity separation is also used in all currently operating commerical Lurgi gasi-
fication facilities for the removal of tars and oils from process condensate.
3. Process Performance
Typical removal efficiencies for oil/water separators range from 60 to
99 percent for oils and 10 to 50 percent for suspended solids (4). The
removal efficiency is dependent upon many factors, including the relative den-
sities of the oil and sludge phases, the configuration of the separator,
retention time, and the size of the oil and tar droplets. Other parameters
such as wastewater temperature or the type of solids in the wastewater can
also affect the separation efficiency. Removal of specific contaminants will
vary from plant to plant depending on these factors. Variations in wastewater
flow rate, temperature, and pH can also have adverse affects on the perform-
ance of a gravity separator (5).
Bl-2
-------
Appendix Bl
Gravity Separation
4. Secondary Waste Generation
Two liquid waste streams can be generated during gravity separation. An
organic liquid whose density is lighter than that of water is generated by
skimming the surface of the wastewater. This liquid will consist primarily of
light oils. The other waste stream is comprised of the material whose density
is greater than that of water. It will consist primarily of heavy tars and
oils with entrained solid sediments.
The process wastewater may contain significant quantities of dissolved
gases. The separator can be covered to contain these gases and prevent their
emission to the atmosphere. Offgases from the enclosed gravity separator can
be combined with similar waste gases for further processing.
5. Process Reliability
Gravity separation as a method for separating nonemulsified oil from
wastewater is highly dependable, providing the equipment is properly control-
led and regularly maintained (5). Gravity separators have been widely used in
many different types of industries for many years. From this it may be
inferred that their reliability is satisfactory. Most of the problems seen in
gravity separators are the result of process upsets which greatly change the
nature of the wastewater feed to the system.
6. Process Economics
Purchased equipment costs for API separators and associated oil and water
pumps were obtained from Reference 6; for a unit with a capacity of about 90
m3/hr the equipment cost was $38.300. Costs for units treating other flow
rates were calculated using an exponential cost scaling relationship with an
exponent of 0.6. Installed equipment costs were calculated on the basis that
equipment costs comprise 70 percent of the installed equipment costs (7).
Bl-3
-------
Appendix Bl
Gravity Separation
Estimates of operating and maintenance requirements for electricity,
maintenance materials, and labor were made as follows. Electricity require-
ments were estimated by assuming that 74.7 kWh were required per 1000 m3
(8). Annual maintenance materials and labor requirements were estimated as 2
and 3 percent respectively, of the purchased equipment cost (6).
Figure Bl-1 presents the installed equipment cost per mVhr of waste-
water flow, updated to first quarter 1980 dollars, as a function of wastewater
flow. A CE index ratio of 1.18 was used to update the costs. Figure Bl-2
presents the operating and maintenance costs as a function of wastewater flow
rate.
7. References
1. American Petroleum Institute, Division of Refining. Manual on
Disposal of Refinery Wastes. Volume on Liquid Wastes. 1st Ed.,
Washington, D.C., 1979.
2. Ford, D.L. and R.L. Elton. Removal of Oil and Grease from Industrial
Wastewaters. Chemical Engineering, October 17, 1977. pp. 49-56.
3 . Patterson, J.W. Wastewater Treatment Technology. Ann Arbor Science
Publishers, Inc., Ann Arbor, MI, 1975, pp. 175-189.
4. Bush, K.E. Refinery Wastewater Treatment and Reuse. Chemical
Engineering, 83(8): 113-118, 1976.
5. U.S. Environmental Protection Agency. Treatability Manual, Volume III,
Technologies for Control/ Removal of Pollutants. EPA-600/8-80-042c, July
1980, Section 4.1.
6. The Richardson Rapid System — Process Plant Construction Estimating
Standards, 1978-79 Edition. Section 100-425.
7. Peters, M.S. and K.D. Timmerhaus. Plant Design and Economics for
Chemical Engineers, 2nd Edition. McGraw-Hill, 1968, p. 134.
8. Prather, B.V. and E.P. Young. Energy for Wastewater Treatment.
Hydrocarbon Processing, May 1976, pp. 88-91.
Bl-4
-------
103
First Quarter 1980 $
,02
7
2 -
10
I I ' I I
J L__J—i
10
7 100
7 1000
Wastewater Flow Rate, m Vhr
Figure Bi-1. Installed equipment cost for API separator systems (6)
Bl-5
-------
Operating Costs, 5/1000 m'
to *.
to
'l
c
>\
n>
Cd
o
•a
TO
i-i
BJ
rr
H-
0
00
n
o
en
hd
en
ft
CJ
i-l
rr
O
en
rf
ft)
er-
as
-------
APPENDIX B2
COAGOLATION/FLOCCDLATION
1. Process Description
Coagulation and flocculation are terms which describe techniques for en-
hancing the gravity separation of suspended tars, oils, and solids from a
wastewater through the addition of chemicals. Each of these terms is commonly
used to describe the overall process. However, coagulation and flocculation
are actually two distinct chemical processes which are typically used together
for the removal of suspended solids and colloidal particles from wastewaters.
Coagulation is the process of destabilizing colloidal particles by the
addition of a chemical. The particles in a colloidal dispersion tend to col-
lect surface charges which repel similarly charged particles. The coagulant
chemical neutralizes the charges built up on the particle surfaces, destabi-
lizing the particles and allowing them to combine with other particles.
Flocculation is the formation of settleable particles from destabilized
colloidal-sized particles. Flocculation occurs by the coalescence and physi-
cal enmeshment of the colloidal particles into suspended particles. These
suspended particles then form a visible layer or floe of material which en-
meshes more colloidal particles and may also trap other noncolloidal parti-
cles. Often a flocculation aid is added to the suspension to enhance floe
formation.
Common coagulation chemicals are alum, lime, and ferric sulfate, while
some common flocculation aids are polymers, activated silica, and polyelectro-
lytes.
B2-1
-------
Appendix B2
Coagulation/Flocculation
The equipment generally required to effectively remove fine suspended
solids from wastewater consists of a flash mixing tank and a sedimentation
basin or clarifier. The waste stream enters an agitated tank where the appro-
priate coagulants, polymers, and pH adjustment chemicals are added to and
thoroughly mixed with the wastewater. The sedimentation basin must provide
adequate retention time for the coagulated solids to settle. Facilities must
also be provided for storing, diluting (or dissolving), and metering the chem-
icals.
2. Process Applicability
Coagulation and flocculation are widely used in the treatment of potable
water as well as wastewater. They are applied for the removal of fine sus-
pended solids (including colloids) and have been applied in the last 15-20
years in the removal of a number of metals from wastewater. Coagulation and
flocculation are also often used to enhance the performance of other treatment
process. Examples of such processes include softening, chemical precipita-
tion, filtration, and air flotation.
The use and dosages of specific coagulants and flocculants for treating a
given wastewater stream are generally determined by the use of jar tests and
field trials.
3. Process Performance
The performance of coagulation/flocculation systems varies widely depend-
ing upon the composition and concentration of impurities in the water being
treated. As indicated above, this technique is often combined with softening
and/or filtration in an overall treatment scheme. Suspended solids can be re-
duced to as low as 10 mg/L in a properly designed system. Removal efficien-
cies for oils range from 61 to 95 percent.
B2-2
-------
Appendix B2
Coagulation/Flocculation
4. Secondary Waste Generation
The major secondary waste stream produced by coagnlation/flocculation
processes is the sludge stream. This sludge is typically quite gelatinous
and may be difficult to dewater. The solids content is typically 1 to 2 per-
cent by weight. The sludge will contain insoluble metal hydroxides and en-
trapped solids. Depending on the chemicals used and the composition of the
wastewater being treated, it may also contain calcium carbonate, sulfates,
phosphates, and/or chlorides. Adjustment of the effluent pH may be required
prior to subsequent treatment or disposal. Some emissions of volatile species
present in the influent wastewater may also occur with these processes. Gen-
erally, these emissions would be controlled by upstream wastewater pretreat-
ment processes (e.g., stripping) but covered vessels and vapor collection sys-
tems can also be used.
5. Process Reliability
In general, coagulation and flocculation have good reliability both from
a performance and mechanical standpoint. Their reliability has been proven
through a number of years of successful operation.
The key to performance reliability is to keep the influent wastewater
composition and flow rate as constant as possible. Changes in these parame-
ters require frequent jar tests and adjustments to chemical feed rates and may
lead to reduced performance. In some cases, for example, addition of too much
polymer may actually work against the desired waste flocculation.
Mechanical reliability depends heavily on the use of proper materials of
construction, since some of the chemicals used are extremely corrosive. Exam-
ples are alum solutions and ferric chloride.
B2-3
-------
Appendix B2
Coagulation/Floccnlation
6. Process Economics
Estimated costs from the literature for the addition of alum, ferric
chloride, and polymer to wastewater for coagulation and flocculation are given
in this section. Costs for filtration and clarification equipment are given
in other appendices (B4 and B19, respectively).
Estimated construction costs for an alum addition system, including
liquid alum storage, chemical feed equipment, building, and rapid-mix tank
with stainless steel agitator for a dosage rate of 200 mg/L were found in
Reference 1. Annual operating and maintenance requirements for electric power
(to operate pumps, mixers, and feeders), maintenance materials, chemicals, and
labor manhours were taken from the same source. The construction costs were
converted to installed equipment costs by adding 28 percent to account for
piping, electrical, instrumentation, and site preparation.
Costs for ferric chloride addition at a dosage rate of 100 mg/L were
taken from Reference 2. Equipment and annual costs items similar to those for
the alum addition system are included (2).
Costs for a polymer flocculant feed system were taken from Reference 3.
These are based on a polymer dosage of 1 mg/L. The system includes chemical
storage, chemical feeding, and rapid mix tank. The various sizes of waste-
water treatment systems have different specific equipment requirements and
mixing procedures.
Costs given in September 1976 dollars were updated to a first quarter
1980 basis by using a CE index ratio of 1.32. Figures B2-1 through B2-6 pre-
sent updated installed equipment and operating and maintenance costs as a
function of wastewater flow rate for the three feed systems.
B2-4
-------
Tl
H-
OQ
e
<-(
o>
w
NJ
I
Installed Equipment Cost, S/(ra3/hr)
,-~ o
o
D
cn
W
I
Ln
CD
0
O
O
en
o
1-1
§
D-
H-
rt
H-
O
3
en
rt
-------
0. 1
2 -
7 -
7 100 2 47 1000
Wastewacer Flow Race, m Vhr
Figure B2-2. Operating costs for alum addition systems (1)
B2-6
-------
7 -
'-1000
u
.c
E 7
First Quarter 1980 S
-------
0.1
7 -
10
4 7 100 2 47 1000
Wastevater Flow Rate, in Vhr
Figure B2-4. Operating costs for ferric chloride addition
systems (2)
B2-8
-------
Installed Equipment Cost, ?/(mVhr)
cn
K>
I
Tl
I—
00
CO
ro
i
Ln
X M
^ P
cn cn
re
Q-
c
H-
x)
re
D
r-i
C
h-1
"a
re
(D
-------
10
- 4
First Quarter 1980 $
Labor
0.1
10
7 100 2 47 1000
Uascewater Flow Rate, m!/hr
Electricity
Figure B2-6. Operating costs for polymer feed systems (3)
B2-10
-------
Appendix B2
Coagulation/Flocculation
7. References
1. D.S. Environmental Protection Agency. Innovative and Alternative
Technology Assessment Manual. EPA-430/9-78-009, February 1980, pp. A144-
A145.
2. Ibid, pp. A146-A147.
3. Ibid. pp. A150-A151.
B2-11
-------
-------
APPENDIX B3
AIR FLOTATION
1. Process Description
Air flotation is a process used to separate fine particles, oil and
grease from aqueous streams. Small gas bubbles rising through the wastewater
adhere to small particles and reduce their apparent density. The resulting
froth rises to the surface where it is skimmed off.
Two types of flotation units are currently in use: dissolved air flota-
tion (DAF) and induced air flotation (IAF). In dissolved air flotation the
entire wastestream, a portion thereof, or a recycle stream of the effluent
water is saturated with a gas (usually air) under pressure. Typically the
pressure is between 0.27 and 0.58 MPa (1). The gas-saturated water is then
depressurized in the flotation vessel, causing the formation of many very
small bubbles. These bubbles attach themselves to oil and suspended solids in
the wastewater. The resulting agglomerates, being lighter than water, rise to
the surface where they form a froth layer which is continuously skimmed off.
Particles which are heavier than water form a sludge which is withdrawn from
the bottom of the unit. Wastewater retention in the DAF vessel is typically
20 to 60 minutes.
The induced air flotation unit operates on the same principles as the DAF
unit. The gas is drawn into the wastewater by the use of a rotor-disperser.
The induced air intimately mixes with the wastewater. The rotor-disperser
produces very small bubbles which greatly enhance flotation. A typical IAF
unit consists of four cells, each of which has a retention time of about one
minute (2).
Both types of flotation devices are very often used in conjunction with
flocculation. The addition of chemicals such as lime, alum, and polyelectro-
lytes to the wastewater prior to flotation can significantly improve the
B3-1
-------
Appendix B3
Air Flotation
removal efficiency of the process, especially when colloidal suspensions or
emulsified oil are present.
Gases other than air can be used in connection with DAF units. Nitrogen
is sometimes used in refinery applications to reduce the chance of fire (1).
2. Process Applicability
Air flotation has been used for many years to treat a variety of indus-
trial wastewaters, including effluents from API separators, metal finishing,
pulp and paper manufacture, and cold-rolling mills (1). Flotation units have
also been used to thicken sludge. The process is effective in removing oil,
grease, and fine suspended solids from wastewater streams. It is particularly
attractive in removing particles with densities close to that of water (1),
where other treatment methods, such as gravity separation, are ineffective, or
require long settling times.
3. Process Performance
Air flotation is capable of removing a wide range of suspended solids,
oils, and grease from wastewater streams. Except in the case of dissolved
gases, air flotation will not significantly reduce the concentration of dis-
solved species. Oil removals of 75 to 85 percent are typically achieved
depending on the nature of the waste and the process configuration (3).
Solids removals are expected to be in the 20 to 75 percent range (4). When
flocculation is combined with flotation, free oil removals may be even higher,
97 percent or more (5), and suspended solids removals of 80 to 93 percent can
be achieved (6).
B3-2
-------
Appendix B3
Air Flotation
4. Secondary Wastes
Air flotation produces two major secondary waste streams. The first is
the froth skimmed from the surface of the wastewater. It contains oil and
entrapped solid particles. The second, sludge, consists of solids having den-
sities greater than that of water. With proper wastewater pretreatment, the
air released in the flotation process is not likely to produce significant
stripping of volatile species into the ambient air (1).
5. Process Reliability
Air flotation has been demonstrated over many years to be a reliable pro-
cess. Without chemical pretreatment, it is subject to variations in influent
wastewater conditions including flow rate, suspended solids, oil and grease
loadings, temperature, and pH (6).
6. Process Economics
Estimated construction costs for DAF units including flocculation, were
obtained from the literature (6). The design basis assumes air injection at
9.4 mj air per 1000 m3 wastewater, 33 percent effluent recycle, 25 minutes
detention time, and a wastewater flow rate of 4.9-7.3 m»/hr per ma of flota-
tion unit cross-sectional area. The DAF unit and chemical feed equipment are
included in the costs. In order to convert the construction costs to
installed equipment costs. Reference 6 recommends adding 28 percent to account
for piping, electrical, instrumentation, and site preparation.
Operating and maintenance requirements, including labor, energy (for
wastewater recycle pumping, air compression, and chemical feed equipment), and
chemicals (alum and polymer) were taken from Reference 6.
B3-3
-------
Appendix B3
Air Flotation
Figures B3-1 and B3-2 are graphs of these cost data updated to first
quarter 1980 dollars. A CE index ratio of 1.32 was used to update installed
equipment and chemicals costs, given in Reference 6 in September 1976 dollars.
7. References
1. U.S. Environmental Protection Agency. Treatability Manual, Volume III,
Technologies for Control/ Removal of Pollutants. EPA-600/8-80-042c, July
1980, Section 4.4.
2. Lash, L.D., and E.G. Kominek. Primary Waste-Treatment Methods. Chemical
Engineering, October 6, 1975, pp. 49-61.
3. U.S. Environmental Protection Agency. Coal Conversion Control
Technology. Volume I, Environmental Regulations: Liquid Effluents. EPA-
60077-79-228a, October 1979.
4. Bush, K.E. Refinery Wastewater Treatment and Reuse. Chemical
Engineering, 83(8): 113-118, 1976.
5. American Petroleum Institute, API Manual on Disposal of Refinery
Wastes, Liquid Wastes Volume. Washington, D.C., 1969.
6. U.S. Environmental Protection Agency. Innovative and Alternative
Technology Assessment Manual. EPA-430/9-78-009, February 1980. pp. A100-
A101.
B3-4
-------
First Quarter 1980 S
<= 1000
too
I I I I I I
1
I 1 I I I I
_J i i 1111
10
4 7 100 2 47 1000
Wastewater Flow Rate, m Vhr
Figure E3-1. Installed equipment cost for dissolved air
flotation (6)
B3-5
-------
100
o
o
S 4
First Quarter 1980 $
Operating an
Maintenance
2 -
J L
i i i i i t
J 1 1
10
7 100 2 47 1000 2
Wastewater Flow Rate, mVhr
Figure B3-2. Operating cost for dissolved air flotation (6)
B3-6
-------
APPENDIX B4
FILTRATION
1. Process Description
Granular-media filtration is one of the oldest and most widely used
methods for the removal of suspended solids, oils, and tars from wastewater
streams. The wastewater flows by either gravity or pressure through a bed of
inert material which physically retains the solids. Various materials includ-
ing sand, anthracite, resins, fibers, and garnet have been used as filter me-
dia. Combinations of materials have also been used in multimedia filters.
The filter bed is contained in a basin or tank and is supported by an under-
drain system that retains the filter medium in place while the filtered liquid
is drawn off.
As the filter bed becomes loaded with retained solids, the flow rate de-
creases and/or a higher pressure drop is sustained across the bed. Cleaning
of the bed is usually accomplished by backwashing with clean effluent. The
backwash water is sent through the bed at a velocity such that the filter bed
becomes fluidized and turbulent. Solids dislodged from the filter media are
discharged in the spent wash water (1). The frequency of backwashing depends
on the suspended solids content, the flow rate of the influent wastewater, and
the capacity of the bed. In systems where continuous flow is required, multi-
ple filters, arranged in parallel, are used so that influent flow can be di-
verted from the units being backwashed. Air may also be used prior to back-
washing to help dislodge the solids. This air scouring reduces the amount of
backwash water required.
Important factors in the design of a granular filtration system include
the filtration rate (volume of wastewater per unit filter bed cross-sectional
area), bed depth, media depth rates (for dual- and multiple—media beds), back-
wash rate (volume per unit filter bed cross-sectional area), air scour rate,
filter run length (cycle time), and terminal head loss (1).
B4-1
-------
Appendix B4
Filtration
Pretreatment of the wastewater by such processes as coagulation/floccula-
tion may be used to increase the solids removal efficiency, generally at a
penalty of reduced filter run lengths (time between backwashes) (1,2).
2. Process Applicability
Generally speaking, granular—media filtration is applicable to waste-
waters containing 5 to 250 mg/L of suspended solids and up to 200 mg/L of oil.
Granular—media filters are used to remove suspended solids and oil from waste-
waters from petroleum refineries, municipal sewage treatment plants, and pulp
and paper plants (3). Filtration is used in the wastewater treatment systems
in the SASOL gasification complex in South Africa and at Kosovo, Yugoslavia.
3. Process Performance
Removal of suspended solids by granular—media filtration ranges from 30
to 70 percent without pretreatment and between 80 to 98 percent when floccula-
tion and/or coagulation are used as pretreatment steps (3). Examples cited in
Reference 1 from a variety of industries showed total suspended solids
removals up to 99 percent with a mean of 68 percent and oil and grease
removals up to 98 percent with a mean of 30 percent. Application of sand fil-
ters to refinery wastewater effluents from API separators indicate suspended
solids removals of 33 to 93 percent, averaging 79 percent; inlet suspended
solids and oil concentrations range as high as 90 mg/L and 178 mg/L, respec-
tively (5). Performance data are not available for filters in synthetic fuels
pi ant s.
4. Secondary Waste Generation
Granular media filtration produces one secondary waste stream intermit-
tently (backwash effluent) and, periodically, spent filter media. The volume
B4-2
-------
Appendix B4
Filtration
of the backwash stream is generally between 2 to 10 percent of the filtered
wastewater volume (1,2,3). Backwash water contains significantly higher
levels of suspended solids and oil than the original wastewater. It can be
returned to treatment units proceeding the filter, or it may be disposed of by
incineration or landfill ing.
After a period of time, the filter media may become clogged such that it
cannot be sufficiently regenerated by backwashing. In such case, the filter
media would be discharged and replaced with new filter media.
5. Process Reliability
Granular media filtration is very reliable both from a performance and
mechanical standpoint (1). Ordinarily pilot tests are conducted before
attempting to apply the process to the treatment of new wastewater streams,
since no generalized approach exists for the design of full-scale filters (6).
The experience gained in such pilot studies adds significantly to the reli-
ability of the filtration system. The process has been successfully applied
in several coal gasification complexes.
Reference 6 presents a discussion of a number of commonly encountered
problems in the filtration of wastewater and control measures which have been
applied to solve them. Most of the problems arise from such operational fac-
tors as insufficient cleaning of the bed or excessive backwash flow rates.
Air and/or water scouring, water surface washing, and special media cleaning
solvents are commonly applied solutions to these problems.
6. Process Economics
Estimated installed equipment costs for a package gravity filtration sys-
tem were obtained from the literature (6). The costs are based on gravity
B4-3
-------
Appendix B4
Filtration
filtration at 12 m»/hr-ma and are for complete facilities including prefiltra-
tion detention basins, steel filtration vessels, media, piping, valves, con-
trols, electrical system, backwash system, surface wash system, chemical fil-
tration aid feed systems, raw water pumps, backwash/clear well storage basins,
and other ancillary items. The costs are valid for systems with flow rates of
45-320 m*/hr. Typical operating and maintenance requirements for granular
media filtration systems as a function of wastewater throughput were also
taken from Reference 6.
Figure B4-1 presents the installed equipment costs updated from January
1978 to first quarter 1980 dollars, using CE cost index ratio of 1.23. Figure
B4-2 shows the variation in operating and maintenance costs as a function of
wastewater flow rate. Maintenance materials costs were also updated using the
same CE index ratio.
7. References
1. U.S. Environmental Protection Agency. Treatability Manual, Volume III,
Technologies for Control/ Removal of Pollutants. July 1980, Section 4.6.
2. D.S. Environmental Protection Agency. Innovative and Alternative
Technology Assessment Manual. EPA-430/9-78-009, February 1980, pp. A102-
A103.
3. Lash, L.D. and E.G. Kominek. Primary Waste Treatment Methods.
Chemical Engineering, October 6, 1975, pp. 60-61.
4. Lanouette, K.H. Heavy Metals Removal. Chemical Engineering,
October 17, 1977, p.75.
5. Peoples, R.F., P. Krishnan, and R.N. Simonsen. Non-Biological Treat-
ment of Refinery Wastewater. Journal of the Water Pollution Control
Federation, 44(11). 1972, pp. 2120-2128.
6. D.S. Environmental Protection Agency. Estimated Costs for Water
Treatment as a Function of Size and Treatment Efficiency.
EPA-600/2-78-182, Santa Ana, CA, 1978.
B4-4
-------
7
4
2
10,000
7
4
2
1,000
7
4
2
100
-
-
-
-
-
\
-
-
1 I 1 I 1 I I 1
^-_
1 1 1 1 1 1 I I
First Quarter 1980 S
1 < iitili
10 2 4 7 100 2 47 1000 2 47
Wastewater Flow Race, m Vhr
Figure B4-1. Installed equipment cost for granular
media filtration systems (6)
B4-5
-------
sLabor
first Quarter 1980 $
Materials
0. 1
Electricity
I I I I I I I I
I i i i ill
10
7 100 2 4 7 1000
Wastewater Flow Rate, m Vhr
Figure B4-2. Operating costs for granular media
filtration systems (6)
B4-6
-------
APPENDIX B5
SOLVENT EXTRACTION
1. Process Description
Solvent extraction processes are often used to remove dissolved polar or-
ganics such as phenols from process wastewater. A number of processes based
on the use of different solvents have been developed over the years. Such
solvents as benzene and sodium hydroxide, tri-cresyl phosphate, n-butyl ace-
tate, diisopropyl ether, methyl isobutyl ketone, and various proprietary sol-
vents have been used (1).
A commonly used process for removing dissolved organics from synfuels
process wastewaters is the Phenosolvan process; this is a proprietary solvent
extraction process developed by Lurgi for the removal of phenols from coke
oven and gasification plant aqueous waste streams. The process uses an organ-
ic solvent such as n-butyl acetate, diisopropyl ether, or, in a newer version,
a proprietary solvent called Phenisol (2). The remainder of this appendix
deals with the Phenosolvan process.
Figure B5-1 is a simplified flow diagram for the Phenosolvan process.
The wastewater to be treated is first passed through sand filters to remove
suspended materials and then fed to a multi-stage mixer-settler (extractor),
where it contacts a counter-flowing lean extracting solvent (such as butyl
acetate, diisopropyl ether, or Phenisol). The wastewater and the solvent are
mixed and pumped by a submerged pump into the settling portion of each stage
of the mixer-settler. From the settling portion, the gas liquor and the sol-
vent flow into subsequent mixing stages in a counter-current mode. After its
last separation stage, the rich solvent is sent to a distillation column where
the solvent is recovered overhead and recycled to the extractor. The column
bottoms, consisting of crude phenols with residual solvent, are stripped for
solvent recovery.
B5-1
-------
GAS
LIQUOR
MAKEUP
SOLVENT
FILTER
BACKWASH
LEAN
SOLVENT
SPENT
FILTER
MEDIA
Cd
I
M
RICH
SOLVENT
PHENOL
RECOVERY
SCRUBBER
RA
T
CTOR
~^N
ENT
i/ERY
3BER
^
— 4
SOLVENT
DISTIL-
^ LATION
COLUMN
V_ x
, ta
STEAM
-i— ~ N'
) -
i i
I i
i t-
i
MAKEUPl
N2
^
^
SOLVENT
RECOVERY
SCRUBBER
V_ _s
CRUDE STEAM
PHENOL f
STRIPPER ^.J^
x^_ y »
' CRUDE
^ PHENOLS
_ EXTRACTED
~~ GAS LIQUOR
^
Figure B5-1. Phenosolvan solvent extraction process
-------
Appendix B5
Solvent Extraction
The extracted wastewater from the mixer-settler is stripped with nitrogen
in the solvent recovery scrubber to recover residual solvent and then
discharged. The solvent-rich nitrogen gas is contacted with a crude phenol
slipstream from the crude phenol stripper to recover most of the solvent.
This phenolic stream is returned to the crude phenol stripper. The nitrogen
gas stream, which still contains some phenolic vapors, is then washed with a
portion of the raw wastewater. Washed nitrogen gas returns to the solvent
recovery scrubber while the raw wastewater proceeds to the mixer-settler.
2. Process Applicability
The Phenosolvan process has been incorporated into every major Lurgi type
gasification facility built to date, including the SASOL and KOSOVO coal gasi-
fication complexes. It has also been installed in over 30 other industrial
plants (3). Theoretically this process is applicable to any aqueous phenolic
wastestream. In practice it is only economical in treating high strength phe-
nolic wastes.
3. Process Performance
The Phenosolvan process will typically remove more than 99 percent of the
monohydric phenols, about 60 percent of the polybydric phenols and 15 percent
of the other organics (e.g., organic acids) present in the wastewater. Over-
all organics removals as high as 95 percent have been reported at SASOL (3,4).
Table B5-1 presents data from the Phenosolvan plants operating in two coal
gasification facilities. In both plants a steam stripper/ammonia recovery
unit is operated as an integral part of the process. The removal efficiencies
obtained in a specific installation will depend on a number of design factors
such as the solvent used, the number of extraction stages, and the ratio of
wastewater to solvent flowrates.
B5-3
-------
TABLE B5-1. PROPERTIES OF FEED AND EFFLUENT GAS LIQUORS FOR TWO DIFFERENT PHENOSOLVAN PLANTS
tfl
Ul
SASOL Plant (5)
Feed Stripped Effluent
Total Phenols 3,250-4,000 160
Steta Volatile Phenol I — 1
Fatty Acidi 300
Aronatic Aainea — —
TOC — —
COD — 1,126
BOD, — —
Tan and Oil 5,000 —
Total MB, 10,800 215
Fixed NH, — —
Total Sulfide 228 12
Total Dissolved Solids — 875
Total Suspended Solids — 21
Cyanide 6 1
PH — 8.4
PNAs
Benz ( a )anthracene — —
7,12-Dinethylbenz(a)antkraoene — —
Benzo(b) flnoranthrene — —
Benzo(a)pyrene — —
3-Me thylcholanthrene — —
Dibenz(a,h)anthracene — —
252 Group (as Benzo(a)pyrene) — —
All values except pH are in mg/L.
SASOL Plant (6) Kosovo
Pic no sol v§o Phc&osol VAC/ Ph6nocol v&&
Feed Stripped Effluent Feed
2.840 210 2,120
1,250 <3.2 —
226 194 —
231 0.45 159
4,190 — 4,970
12,500 1,330 18,900
— — 9,030
— 400
11,200 150 3,7<0
— — 250
_ -_ _
2,460 596 2,170
— — 150
— — <1
8.9 8.2 9.2
— — 0.92
— — 0.23
— — 0.68
— — 0.19
— — <0.004
— — 0.02
— — 1.26
is operated as an integral part of the process.
Plant (7)
Phecokolvan/.
Stripped Effluent
230
130
—
—
1,470
7,910
2.350
<200
-205
205
—
1,160
190
0.02
9.6
<0.008
<0.008
<0.008
<0.008
<0.008
<0.008
0.19
—No data.
-------
Appendix B5
Solvent Extraction
Some of the major criteria for selecting a solvent for the recovery of
phenol include (8):
• The solvent should have a high distribution factor for phenol.
• Its boiling point should be below that of phenol (454 K) .
• Its solubility in water should be as low as possible.
• It should be chemically and thermally stable.
Table B5-2 shows the properties of some solvents which have been used in the
recovery of phenol (9,10,11,12).
The extraction efficiency can be estimated with the following equation
(4):
X E-l
n
K S
where : E = D
W
K = distribution coefficient, unique for each solvent-waste
component-water system,
S = solvent weight flow rate,
W = wastewater weight flow rate,
X^ = waste component (phenol) wt. frac. in entering wastewater,
X = waste component (phenol) wt. frac. in existing wastewater,
and
n = number of stages.
B5-5
-------
Appendix B5
Solvent Extraction
TABLE B5-2. PROPERTIES OF SOME PHENOL EXTRACTING SOLVENTS (9,10,11.12)
Solvent
n-Butyl Acetate
Diisopropyl Ether
Aromatic Oil
Methyl Isobutyl Eetone
„ wt fraction of
Phenol
Distribution
Coefficient, K*
49
36.5
22
60
phenol in solvent chase
Solubility
in Water,
wt. %
1.0 at 308 K
0.8 at 308 K
0.1 at 275 K
1.9
Boiling Point,
E
398
338
353 +
392
„<• f *.<,„<-;„„ „* \ i'• ^ measured at high dilution
wt traction of phenol in aqueous phase
The distribution coefficient, Kn, must be found for each component pre-
sent in the wastewater in order to predict the overall extraction efficiency.
KJJ has been found to be a function of a number of wastewater parameters such
as temperature and pH (in the case of phenols), so these parameters must also
be known in order to accurately predict component removals by solvent extrac-
tion (13). K decreases with increasing temperature (14).
4. Secondary Waste Generation
The major secondary waste from the Phenosolvan process is represented by
the recovered crude phenol. Low volume air emissions from various unit opera-
tions (e.g., solvent storage tank vents) and spent filter media from the sand
filters are additional waste stream sources. There are no data on the charac-
teristics of the phenols recovered from gasification processes. Based on
typical Lurgi gas liquor characteristics and extraction efficiencies, it has
been estimated that the recovered phenol would consist of 85 percent monohy—
dric phenols, 10 percent polyhydric phenols, and 5 percent other organics (4).
The Phenosolvan recovered crude phenol stream from a coke oven plant was
reported to contain 80 percent total phenols, 7 percent pyridine bases, 10
percent natural oils and phenol pitch, and 3 percent water. The total phenol
fraction was composed of 70 percent phenol, 27 percent cresols, and 3 percent
xylenols.
B5-6
-------
Appendix BS
Solvent Extraction
The lower heating value of the recovered crude phenol in the Kosovo
gasification plant was reported to be 7790 kcal/kg (7). This is about 10
percent higher than the theoretical heat of combustion of phenol (15), or more
than twice the heating value of the Kosovo coal. Thus, recovered phenol
should be saleable as a byproduct or, if there is no market, suitable for use
as a fuel to recover its heating value.
A minor secondary waste from the Phenosolvan process is spent sand from
the initial filtering step. These filters are periodically backflushed with
dephenolized wastewater, with the backflush water sent to the preceding tar
and oil separation unit. No data are publicly available on the quantity of
spent sand generated.
5. Process Reliability
The Phenosolvan process was first commercialized in about 1940; since
then there have been more than 30 commercial plants installed worldwide,
ranging in capacity from 0.5 to 230 m3/hr. There are no publicly available
data on the operating histories of these plants, but the continued use of this
process tends to indicate adequate reliability.
6. Process Economics
The capital cost for a Phenosolvan plant treating 700 m*/hr of stripped
gas liquor from a Lurgi gasification plant was estimated at $14 million in
1978 (16). Cost estimates for four smaller plants, 11.4 rnVhr, 22.7 m*/hr,
45.4 m3/hr, and 136 m3/hr have also been made (2). These data were updated to
first quarter 1980 dollars using the CE index. The reported capital costs
were converted to installed equipment costs by assuming that engineering and
construction, fees, and contingency (included in the capital costs) amount to
48 percent of the installed equipment cost. Figure B5-2 presents installed
equipment costs versus wastewater flow rate.
B5-7
-------
7 -
4 -
100,000
First Quarter 1980 S
O Reference 2
A Reference 16
10,000
7
1,000
I I I I I I I I
10
7 100 2 47 1000
Uasteuater Flow Rate, raVhr
Figure B5-2. Installed equipment cost for Phenosolvan
extraction systems
B5-8
-------
Appendix B5
Solvent Extraction
Estimated utility and chemical requirements (5,16,17) include: cooling
water (1.5 - 5.2 ms/m*), low pressure steam (10-75 kg/m3), electricity (1.0-
1.7 kWh/m3), and solvent makeup (0.01-0.03 kg/m1). Utility and chemical
requirements are based on the volume (m3) of stripped gas liquor feed.
7. References
1. U.S. Environmental Protection Agency. Technology Status Report: Low/
Medium-Btu Coal Gasification and Related Environmental Controls, Vol. 1.
June 1977.
2. Singh, S.P.N., et al. Costs and Technical Characteristics of
Environmental Control Processes for Low-Btu Coal Gasification Plant, Oak
Ridge National Laboratory Report. ORNL-5425, June 1980.
3. U.S. Environmental Protection Agency. Environmental Assessment Data Base
for Low/Medium-Btu Gasification Technology, Vol. 2. September 1977.
4. Beychok, M.R. Coal Gasification and the Phenosolvan Process. Paper pre-
sented at the ACS 168th National Meeting, Atlantic City, New Jersey,
September 1974.
5. Information provided by South African Coal, Oil and Gas Corp., Ltd. to
EPA's Industrial Environmental Research Laboratory (Research Triangle
Park), November 1974.
6. Bromel, M.C. and J.R. Fleeker. Biotreating and Chemistry of Waste Waters
from the South African Coal, Oil and Gas Corporation (SASOL) Coal
Gasification Plant. Report prepared for National Gas Pipeline Company of
America, Bacteriology Department, North Dakota State University,
December 1976.
7. U.S. Environmental Protection Agency. Environmental Assessment: Source
Test and Evaluation Report — Lurgi (Kosovo) Medium—Btu Gasification,
Final Report. EPA-600/7-81-142, August 1981.
8. Wohler, F. Removal and Recovery of Phenol and Ammonia from Gas Liquor.
Chemsa, 72-74, May 1979.
9. Wurm, H.J. Treatment of Phenolic Wastes, Eng. Bull., Purdue University
Eng. Ext. Serv., 132 (11) 1054-73, 1969.
B5-9
-------
Appendix B5
Solvent Extraction
10. Earhart, J.P. et al. Recovery of Organic Pollutants via Solvent
Extraction. Chemical Engineering Progress, May 1977, p. 67.
11. Burns, G.P. M.S. Thesis, University of California, Berkeley, 1979.
12. Morrison, R.T. and Boyd, R.N. Organic Chemistry, Third Edition. Allyn
and Bacon, Inc., Boston, Massachusetts, 1973.
13. Greminger, D.C., et al. Solvent Extraction of Phenols from Water, AI/ChE
Conference, Philadelphia, Pennsylvania, June 8-12, 1980.
14. U.S. Department of Energy. Solvent Extraction of Phenols, in Processing
Needs and Methodology for Wastewaters from the Conversion of Coal, Oil
Shale and Biomass to Synfuels. DOE/EV-0081, May 1980.
15. Lanouette, K.H. Treatment of Phenolic Wastes. Chemical Engineering,
October 17, 1977, p. 99-106.
16. U.S. Department of Energy. Research Guidance Studies to Assess Gasoline
from Coal by Methanol-to-Gasoline and SASOL-type Fischer-Tropsch
Technologies. FE-2447-13, August 1978.
17. American Lurgi Corporation, Dephenolization of Effluents by the
Phenosolvan Process, company brochure, undated.
B5-10
-------
APPENDIX B6
WET AIR OXIDATION
1. Process Description
The wet air oxidation process, also known as the Zimmeraann process,
achieves the aqueous phase oxidation of dissolved or suspended organics at
elevated temperatures and pressures. The oxidation process is carried out
under pressure (up to 20.8 MPa) in order to avoid boiling and to promote in-
creased oxygen solubility. Before entering the reactor, the wastewater to be
treated is preheated by heat exchange and supplemental heating (if necessary)
to about 593 K. Compressed air or oxygen is added to the wastewater after
preheating. The degree of oxidation achieved depends on a number of factors,
including the reactor temperature and pressure, the nature of the material
being oxidized, oxygen addition rate, and the retention time in the reactor.
For example, one application of wet air oxidation is in the thermal condition-
ing of sewage sludges. Relatively mild conditions (417-483 K, 1.1-2.9 MPa)
and short retention times (15 minutes) are required in this application to
destroy pathogens, break down the sludge cell structures and reduce the water
affinity of the sludge solids (1). On the other hand, destruction of certain
toxics may require temperatures in excess of 593 K and retention times of one
hour or more (2). Catalysts such as copper salts can be added to allow opera-
tion at lower temperatures.
When oxygen uptake exceeds about 10,000 mg/L, the process is autothermic
and requires no additional energy input (3). When very high dissolved organic
concentrations are present, the process can be adapted to produce steam.
2. Process Applicability
Wet air oxidation is used commercially for the treatment of both non-
hazardous and hazardous industrial wastes. As noted previously, it is used to
condition sewage sludge and to destroy toxic organics (3). It has also been
B6-1
-------
Appendix B6
Wet Air Oxidation
used to regenerate spent powdered activated carbon from the treatment of night
soil (concentrated human waste) and municipal sewage (4). It has been consi-
dered for use in destruction of biological sludge and regeneration of spent
powdered activated carbon (4) in the design of two proposed commercial coal
gasification plants.
In general, wet air oxidation will be limited by economics to small
volume streams having relatively high organic concentrations, especially of
biologically refractory or toxic compounds. The largest treatment capacity
for a single wet air oxidation reactor is 68 m*/hr. The largest stream to be
treated is roughly 600 m3/hr (10 units at 60 m3/hr each) (4). There are
approximately 185 wet air oxidation installations in the world treating a wide
variety of waste types (4). The process is applicable to hazardous wastes
which contain too much water to be incinerated economically but which are too
toxic for biological treatment (5). The wet air oxidation process breaks many
toxic and refractory wastes into biodegradable intermediates.
3. Process Performance
Wet air oxidation is capable of reducing the levels of both organic and
inorganic COD in wastewater streams. The reduction achieved for a given waste
depends on the operating temperature and pressure, presence of catalyst, oxy-
gen addition rate, and the nature of the waste itself. COD removals of
greater than 90 percent have been reported (6) while greater than 99 percent
destruction of phenolic compounds has been achieved (2). Whether these levels
of destruction can be achieved with synfuels wastewaters is not known. Bench-
or pilot-scale tests would need to be performed on actual wastewaters.
The percentage reduction of organics is not drastically affected by the
organic loading of the waste. Wet air oxidation can break down many large
B6-2
-------
Appendix B6
Vet Air Oxidation
organic molecules which resist biological treatment into species which are
more amendable to biological oxidation. The products of wet air oxidation of
organics are generally low molecular weight compounds such as formic and
acetic acid (2). Sulfides are oxidized to sulfates (7), while cyanides and
nitriles are converted to ammonia (3).
4. Secondary Waste Generation
The major secondary waste stream produced by wet air oxidation consists
of the offgas released from the reactor. Depending on the nature of the in-
fluent waste, the offgas may contain ammonia, volatile organics, carbon
monoxide, and carbon dioxide (3). The flow rate of this waste stream is sub-
ject to wide variation depending on the flow rate of air (or oxygen) required
for the oxidation of the waste. When wet air oxidation is used to regenerate
spent carbon or destroy biological sludge, the offgas is treated by liquid
scrubbing or dispersion within the aeration basin of the biological process to
remove trace organics and odors.
The oxidized waste stream may have a residual organic concentration mea-
sured as BOD. Removal of this BOD may require further treatment to remove
residual weak acids. If wet air oxidation is used in conjunction with biolog-
ical oxidation, this stream is recycled to the aeration basin.
5. Process Reliability
Only limited data are available on the operation of wet air oxidation fa-
cilities. Some initial operational problems have been encountered, but me-
chanical and process reliability are now reported to be adequate. No details
of any operating problems were found. Careful operator attention is report-
edly required (1). Automation is possible to the degree that programmed
startup and shutdown is achievable.
B6-3
-------
Appendix B6
Wet Air Oxidation
6. Process Economics
Wet air oxidation is a capital intensive process. The most significant
capital items are the reactor and heat exchanger. The capital costs depend on
the flow rate and oxygen demand of the waste, the severity of the oxidation
conditions, and the required materials of construction (3). Higher reactor
pressures require heavier wall construction. The presence of certain species
in the influent may require special materials of construction; for example,
the presence of chlorides may necessitate titanium or Hastelloy clad system
internals to avoid corrosion problems. Wet air oxidation systems have been
limited by economics to the treatment of wastewaters with flow rates of
individual units ranging from 1.6 to 68 m3/hr (4).
Installed equipment costs for wet air oxidation systems were estimated in
Reference 8 assuming Hastelloy construction. Installed costs for a wet air
oxidation system treating 4.5 mj/hr of a high chloride (32,000 mg/L) acidic
waste containing 70,000 mg/L COD were estimated at il.8 million in August 1979
(3). In this instance titanium materials of construction were utilized. The
latter estimate is somewhat higher than that given in reference 8; the reason
for this difference is not known.
Operation and maintenance requirements were estimated based on the fol-
lowing guidelines (3):
• Energy requirements (for pumping and air compression) were
taken as 410,000 kWh per year per m3/hr.
• Maintenance material costs were estimated at 2.5 percent of the
installed equipment cost of the system.
B6-4
-------
Appendix B6
Wet Air Oxidation
• Labor requirements were estimated at 4 hours per shift, inde-
pendent of wastewater flow rate.
Figures B6-1 and B6-2 summarize these cost data, updated to first quar-
ter 1980 dollars by the CE index ratio of 1.17, as a function of flow rate.
7. References
1. U.S. Environmental Protection Agency. Treatability Manual, Volume III,
Technologies for Control/ Removal of Pollutants. EPA-600/8-80-042c, July
1980, Section 7.7.
2. Randall, T.L. and P.V. Knopp. Detoxification of Specific Organic Sub-
stances by Wet Oxidation. Journal Water Pollution Control Federation,
August 1980, pp. 2117-2130.
3. Wilhelmi, A.R. and P.V. Knopp. Wet Air Oxidation - An Alternate to
Incineration. Chemical Engineering Progress, August 1979, pp. 46-52.
4. Review Comments from Mr. Charles Soukup, Zimpro, Inc., September 8,
1982.
5. Worthy, W. Hazardous Waste: Treatment Technology Grows. Chemical and
Engineering News, March 8, 1982, pp. 10-16.
6. Hulswitt, C.E. Purifying Wastewater by Wet Air Oxidation. Plant
Engineering, August 17, 1978.
7. Teletzke, G.H. Wet Air Oxidation. Chemical Engineering Progress,
January 1964.
8. Hulswitt, C.E. New Tool for Waste Water Treatment-Flameless Combus-
tion. September 23, 1978.
B6-5
-------
10b
First Quarter 1980 $
t. 7
5 2
10 J
10"
J I I I I I I I
I I I I I I I I
I 1 I till I
1 2 4 7 10
Wastewater Flow Rate, m''hr
Figure B6-1. Installed equipment cost for wet air
oxidation svstems (8)
B6-6
-------
H-
OQ
c
l-(
0)
03
Operating Costs, 5/1000 m3
O
id
l-t
tu
H-
O
00
n
o
Ml
o
P
H-
X
H-
O-
Cu
rt
H-
O
3
3-
cn
u>
-------
-------
APPENDIX B7
STEAM STRIPPING
Steam stripping is used to remove ammonia, hydrogen sulfide, carbon diox-
ide, phenols, cyanides, organics, and other volatile compounds from waste-
water. The stripped gases and vapors can be further processed, recovered, or
destroyed. The stripped wastewater passes to further treatment or reuse. If
large amounts of ammonia are in the wastewater, the basic steam stripping pro-
cess may be expanded to include the recovery of ammonia. (See Appendix B8,
Steam Stripping Ammonia Recovery: PHOSAM-W Process.) Alternately, two stages
of steam stripping may be used to recover ammonia and produce a high purity
acid gas stream. (See Appendix B9, Steam Stripping Ammonia Recovery: Chevron
WWT Process.)
1. Process Description
A typical flow diagram of basic steam stripping is shown in Figure B7-1.
Steam stripping is usually conducted as a continuous operation in a pack-
ed tower or conventional fractionating column (bubble cap or sieve tray) with
more than one stage of vapor/liquid contact. The wastewater feed is first
preheated to its bubble point by heat exchange with stripped wastewater. The
preheated water enters near the top of the water stripper and then flows by
gravity countercnrrently to steam and vapor rising up from the bottom of the
stripper. As the wastewater flows downward, it is heated and volatile com-
pounds and gases are stripped by the steam and vapors rising from the bottom
of the stripper.
Stripping steam may be live steam or wastewater evaporated by a steam
reboiler. If live steam is used, steam condensate enters the stripped waste-
water and is not recovered. If a reboiler is used, reboiler steam is typi-
cally supplied at 440 to 520 kPa (1,2).
B7-1
-------
Wastewater Feed
-±1
-«-XH<
Stripped
Wastewater
Stripped Gases/
Vapors
To Recovery or
Disposal
rxj^-i
. Y
STM
Figure B7-1. Steam stripping (1)
B7-2
-------
Appendix B7
Steam Stripping
The key design characteristics of water strippers in various applications
are shown in Tables B7-1 and B7-2. Tray tower strippers consist of between
1 and 18 theoretical stages which correspond to 4 to 52 actual trays. Packed
tower strippers contain up to 6 meters of packing which corresponds to about
seven theoretical stages. The stripping steam requirement ranges from about
0.01 to 1.0 kg steam per kg wastewater feed. The stripper normally operates
at 130 to 170 kPa.
Another design feature of strippers is refluxing. Refluxing, not shown
in Figure B7-1, is the condensing of a portion of the steam and vapors leav-
ing the stripper and returning it to the stripper. Refluxing may be used to
reduce the amount of water taken overhead and thus increase the concentration
of volatile components in the overhead gases. Refluxing would also reduce the
amount of phenols and other less volatile components taken overhead. However,
refluxing requires a partial condenser and increases the water stripper cost
and the requirement for stripping steam and condenser cooling water. Tables
B7-1 and B7-2 show that refluxing is common in strippers used in petroleum
refining applications. However, current conceptual designs of strippers in
synfuels applications do not feature refluxing.
2. Process Applicability
No specific limits of process applicability are reported in the open lit-
erature. Steam stripping is commonly used in both industrial chemical produc-
tion and in industrial waste treatment. The effectiveness of steam stripping
is limited when the volatile compounds are chemically fixed in the wastewater.
3. Process Performance
The performance of steam stripping depends on the characteristics of the
wastewater feed and water stripper design. The performance of typical
B7-3
-------
TABLE B7-1. PERFORMANCE AND KEY DESIGN CHARACTERISTICS
OF SOUR WATER STRIPPER IN VARIOUS APPLICATIONS
Cd
-P-
Type of
Facility
Synfu s/Bygask
Synfu
Synfu
Synfu
Synfu
Synfu
Synfu
Synfu
Synfu
Synfu
Synfu
Synfu
InSit
roce
ef In
efln
• fin
efln
efln
efln
efln
efln
efln
efln
efln
efln
efin
efln
efln
efin
efln
efin
efin
efln
efin
efln
efln
./Byga."
•/Lnrai
./SRC?
./SRC
• /SEC
•/BIG.."
l/Synthue
./Paraho"
I /TOSCO*
./TOSCO"
•/Modified
Oil. Shale
.in,b
ry.
ryc
r^
"I
rye
ry"
c
ry
c
ry
tjc
ryC
<
ry
c
ry
ry°
ry
c
ry
G
ry
c
ry
c
ry
ryc
ryc
ry°
ry°
Character! t tics
KB,
4370
1500
17000
7280
12130
4850
4860
9700
17990
5100
41000
21330
5000
5000
1000
1480
JOOO
3800
1500
9800
5000
1430
1560
1310
1310
3700
3620
2640
—
5500
2640
4620
5000
—
1740
H,S
0
100
100
3830
11500
0
0
400
32400
900
30200
118
10000
8000
1500
1876
5000
4600
3000
7350
2000
2180
3120
330
340
6550
12010
5600
3830
5500
600
2940
7200
5000
1810
00 Phenol t
4000
16500
33000
6270
0
9400
9400
18800
206
6200
—
41800
—
—
—
—
—
—
—
—
—
—
—
--
—
—
—
—
—
—
—
—
—
—
~~
8600
500
300
130
400
0
0
2600
--
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
--
—
—
—
—
—
—
—
—
—
—
Characterlitlci Renoval Efficiency
Concentration. Bl/L Percent
NH.
440
45
200
75
120
300
Ho
Is
0
6
0
20
30
0
300 0
340d 8
60
50
32
46
280
210
300
194
600
227
—
790
200
580
490
124
64
47
78
128
—
750
34
120
—
—
30
11
5
9
5
50
10
2
0
50
39
5
trace
5
32
36
0
0
16
19
18
380
50
trace
0.3
238
206
1
Phenol a NB,
7140 90
410 99.5
215 98.9
110 99
312 99
0 93.8
0 94
2000 96.5
99.7
99.2
99.9
99.9
94.0
95.0
69.0
86.)
87.3
93.0
—
89.6
96.0
57.3
67.0
89.0
95.0
98.3
97.6
95.0
—
86.4
98.)
97.0
—
—
98.0
B,S Phenols
..
94
100
99.5
99.7
—
—
98
95
99.0
100
97.7
99.4
99.9
99.9
100
98.9
99.0
99.8
100
99.75
98.3
99.0
100
100
99.7
99.8
99.7
89.0
99.0
100
100
96.2
95.8
99.9
17
18
33
18
22
—
—
18
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
--
—
—
—
—
—
—
—
—
—
—
—
Stein Rate Actual
kg feed
0.114
0.129
0.109
0.129
0.105
0.105
0.117
0.130
--
—
—
—
0.21
0.17
0.047
0.081
0.048
0.081
0.11
0.22
0.084
0.044
0.074
0.055
0.14
0.39
0.32
0.16
0.012
0.11
0.13
0.075
0.042
0.034
0.14
Packing
12 trays
40 traya
—
34 traya
29 trays
18 trays
18 traya
37 trays
—
—
—
—
13 traya
13 tr ys
3.6
6.3
6 tr y
6 tr y
12 tr y
12 tr y
10 tr y
4 tr y
4 tr y
45
45
45
4)
45
8 t aya
6.0 »
45
45
6 t ays
8 t ays
8 t ays
No. of
Stages
4
14
—
12
10
6
6
13
—
—
—
—
.5.
.5!
«,
as'
aSl
«j'
«3*
a3*
Reflux
No
No
No
No
No
No
No
No
No
No
No
No
Yes
Tea
No
No
Tea
Ye.
Ye.
lea
No
No
No
No
No
Tes
Tes
Tea
No
No
No
No
No
No
Tes
Kef.
1
1
1
1
1
1
1
1
3
3
3
3
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
to negligible reaidual levels.
Date are foil-scale testing results.
Stripper design Is for 99 percent removal of free ammonia; assua.es 250 aig/L NH, i* fixed as NH4C1.
'Estimated by assuming that actual trays are 35% efficient (1). Traya in refinery sour water strippers
are between 30 and 50% efficient In most cases (5).
Estimated by assuming that the height equivalent to a theoretical plate (HETP) la 0.9 meter. Reference
5 reported that the HETP for 7.6 cm Raschig ringi in sour water strippers averaged 0.76 to 1.1 meter.
-------
TABLE B7-2. PERFORMANCE AND KEY DESIGN CHARACTERISTICS OF SOUR WATER
STRIPPERS ACHIEVING HIGH AMMONIA REMOVALS IN REFINERIES (5)
A*.r..t fm*<
Feed
2500
1200
1720
430
74
4000
1600
5410
3550
2000
1400
19000
2000
32200
1600
»«0
1850
1200
2600
5450
2623
213
4400
ilia Coaeeatra tioaa. mm/L
Stripper Bottoms
7»
25
68
64
63
100
63
45
37m
200
80
80
15
56
23
50
96
65
200
56
10
76
11
Strippinf. Steaai late.
k| steaai per x| feed
0.17
0.12
0.13
0.30
0.072
0.19
0.94
0.16
0.22
0.18
-
—
0.14
-_
0.1*
0.036
0.072
0.024
0.23
0
0.041
0.32
10 valve traya
8 valve traya
30 sieve trays
30 sieve trays
24 sieve traya
23 sieve tray a
5 flitch traya
6 ar-3* Raaohil rings
10 flex traya
4.5 ar-31 Raaohif rU|S
18 traya
20 bubble cap traya
12 Socony traya
20 sieve trays
8 babble cap traya
6 shower traya
4.1 m-3- Eaaoaif rings
4.5 ar-3' saddlea
8 valve traya
28 bubble cap traya
5 valve traya
8 flex trays
Reflux
Tea
Tea
Tea
Tea
Tea
Tea
Tea
Tea
Tea
Tea
Tea
Tea
Tea
Tea
Tea
No
No
No
No
No
No
No
No.
atM bottctts concentration.
B7-5
-------
Appendix B7
Steam Stripping
strippers in removing ammonia, hydrogen sulfide, carbon dioxide, and phenols
is shown in Tables B7-1 and B7-2. The performance of strippers in removing
COD, TOC, and specific organic compounds is reported in Table B7-3.
The most efficient strippers listed in Tables B7-1 and B7-2 are gener-
ally shown to strip ammonia down to 50 to 100 mg/L, although three strippers
achieve levels of less than 15 mg/L. Strippers achieving ammonia levels of
less than 100 mg/L remove hydrogen sulfide to 0 to 30 mg/L depending on the
initial ammonia, hydrogen sulfide, and carbon dioxide concentrations. Carbon
dioxide is totally stripped. The ability of steam strippers to achieve low
residual levels of ammonia (e.g., 100 mg/L) is adversely impacted if the
wastewater contains appreciable quantities of "fixed" ammonia. Some synfuels
wastewaters will contain appreciable quantities of organic acids. As a
result, the level of fixed ammonia may be high and the degree of ammonia
removal may be limited unless the ammonia is "freed" by pH adjustment.
For several types of wastewaters from synfuels, typically 17 to
22 percent of the residual phenols are stripped. For refinery type waste-
waters, unrefluxed strippers on the average remove 55 percent of the phenols
and refluxed strippers on the average remove 39 percent of the phenols (5).
Cyanide removal efficiencies are reported to vary widely but average
37 percent for stripping refinery-type wastewaters.
The pilot strippers described in Table B7-3 removed COD and TOC on the
average down to 170 and 118 mg/L respectively, corresponding to removal effi-
ciencies of 59 and 57 percent. The removal of toxic pollutants varied consi-
derably with average removals ranging from 9 to over 99 percent, and average
concentrations in the stripped wastewater ranging from less than 0.05 to
340 mg/L.
B7-6
-------
Cd
-J
TABLE B7-3. PERFORMANCE OF PILOT SCALE STRIPPERS IN REMOVING COD, TOC, AND
SPECIFIC TOXIC POLLUTANTS (6)a
Strloper Bottoms Concentration.
Pollutant
Conventional Pollutants:
COD
TOC
Toxic Pollutants:
Chloroform
1 ,2-Di chl oroe thane
1 ,2-Trana-dichloroethylene
Methylene chloride
1,1 ,2 ,2-Tetrachloroe thane
Te trachl oroe thy lene
1 ,1 ,1-Trichl oroe than*
1 ,1 ,2-Trichloroe thane
Tri chl oroe thy lene
Mininum
118
14
99
>99
>99
87
>99
>99
>99
>99
62
72
>99
>99
99
81
0
>99
>99
>99
Mean Data Pointa
59
57
89
97
76
75
40
78
9
>99
>99
6
40
5
45
5
v 5
5
3
1
5
74
The pilot teata covered the range of steam to feed ratio* of 0.14 to 0.36 mL (tea* per ML waatewater feed.
Insufficient data are available to report the atripping steaa ratio in aiore typical Bolts. If the mL of steam
are *iL of vapor, the stripping it earn ratio ia very uall (on the order of 0.0004 to 0.001 kg »tea» per kg feed).
'if the tiL of atean are »L of conducted §tei», the stripping steaa ratio ia the aiore typical 0.13 to 0.33 kg ateu
per kg feed. The pilot teats covered both nonreflaxed and refluxed colunni. The pilot colnans were 3.67 • (12 ft)
.high and packed with polypropylene pall ringa.
Reported as not detected, asmed to be <10 g/L.
-------
Appendix B7
Steam Stripping
As illustrated in Figures B7-2 and B7-3, for a specific type of waste-
water and desired pollutant removal, the principal design variables are the
number of theoretical stages and the ratio of stripping steam to wastewater
feed. For example, as shown in Figure B7-2, 99 percent removal of ammonia
from a Lurgi type condensate could be achieved by a stripper of three theore-
tical stages with a steam stripping rate of 0.5 kg steam per kg of wastewater
feed. Alternatively, the same removal could be achieved by a stripper of 8
theoretical stages with a steam stripping rate of about 0.12 kg steam per kg
of wastewater feed. The selected design will be an optimum balance of strip-
per costs (strippers with more theoretical stages cost more) and steam costs.
The effectiveness of stripping is limited when the volatile compounds are
chemically fixed in the wastewater. Typical examples are the fixing of ammo-
nia by hydrochloric and naphthenic acids (7). Fixed compounds may be freed by
the addition of chemicals (e.g., caustic) when practicable.
4. Secondary Waste Generation
The only secondary waste stream from steam stripping is the stripped gas
or vapor which exits the stripper. This stream is saturated with water and
contains all of the stripped gases and vapors. It may be processed to recover
specific components or alternately disposed of by incineration when possible.
5. Process Reliability
No specific reliability measures are reported in the open literature.
The process is widely used and employs standard equipment and, theoretically,
reliability should be high. In refinery applications steam stripping has pro-
ven to be highly dependable (6) .
B7-8
-------
tn 6
-------
en
-------
Appendix B7
Steam Stripping
6. Process Economics
The unit installed equipment costs of steam stripping are shown in Figure
B7-4. The bases for the installed equipment costs and for operating costs
are given in Tables B7-4 and B7-5, respectively. The scope of the steam
stripping system includes: a stripper, product cooler, feed/product heat
exchanger, reboiler, and feed and product pumps.
The installed equipment cost correlation in Figure B7-4 is a correlation
of updated costs for eight steam stripping applications as reported in
References 1 and 8. The costs vary mainly with the wastewater flowrate but
variations due to different design and feed characteristics (such as the num-
ber of theoretical transfer units) are also evident. Although these other
differences can significantly affect costs, the available cost data base is
insufficient to consider their precise impact on costs. The correlation in
Figure B7-2 is, thus, simply a correlation of cost versus wastewater flowrate.
Despite this simplification, a "good" fit of data results with an explained
variance of 81 percent. The correlation in Figure B7-4 is:
Unit Installed Equipment Cost, J/(m»/hr) = 20,378 Q~°*41
where: Q is the wastewater flowrate in m*/hr.
The unit installed equipment costs of actual applications can differ from
costs estimated with this correlation due to differences in feed and stripper
characteristics and in cost scope. The reported correlation is thus probably
accurate only to +50 percent.
B7-11
-------
10,000
1st Quarter 1980 Dollars
Note: See Table B7-4 for
data sources.
Numbers on data
points are number of
theoretical transfer
units.
1,000
2 -
J L
J - 1
' i
J L
J_L
10
100
Wastewater Feed, raVhr
Figure B7-4. Unit installed equipment cost of steam stripping
B7-12
-------
TABLE B7-4. BASIS FOR INSTALLED EQUIPMENT COSTS OF STEAM STRIPPING
w
I
M
U>
Type of
Fsclllty
Hy,asC
Hy,ssC
SRC°
SRCC
SRC°
BlGssC
Syathane
TOSCO'
NB,
4,370
8,500
7,280
12,150
4.850
4,860
9,700
5.100
Unspecified' 23,000
"CO, Is always totally
Concentration. mt/L
H,S
__
100
3,830
11,500
0
0
400
900
23.000
CO, Phenols
4,000 8.600
16,500 500
6,270 130
0 400
9,400 0
9,400 0
18,800 2,600
6,200
-
Percent
at'/hr NB, B,S Phenols
27 90 — 17
49 99.5 94 18
11 99 99. 5 18
3.6 99 99.7 22
7.4 93.8
113 94
54.1 96.5* 98 18
106 99.2 99.0
45 —
kg steal Theoretical
per kg feed Stagea
0.114 4
0.129 14
0.129 12
0.105 10
0.105 6
0.117 6
0.130 13
__
O.J4 81
Estimated
Cost Data EqulpsMBt
(1978) Coat
Il24,000d
l283,000d
»153,750d
l50,630d
i83,250d
1393. 750d
1364, 000d
1776,000'
1240 .OOO1
183,800
1193,000
1104,000
134.200
156.300
1266.000
1246,000
1524,000
1162.000
Dpda ted
(1980)
Eqaipaient
Cost
196,000
1221,000
1119.000
139.200
164,500
1305.000
1282,000
1600. 000k
11(6.000
stripped.
Installed equipment cost estimated by aaanaing
^Stripper
fCapital
design is for 99» of
cost.
free aaunonie; ass
thst filed capital Investment or
times 250 «|/L NB, is filed as NB
capital cost
4C1.
the other cost data.
FrOB Refereoce 8.
Twenty-fonr actat 1 trays with assmned tray efficiency of 35%.
-------
Appendix B7
Steam Stripping
TABLE B7-5. BASIS FOR OPERATING COSTS OF STEAM STRIPPING
Steam - 520 kPa (1)
• range 0.01-0.94 kg/kg feed*
• typical 0.105-0.130 kg/kg feed (1)
Cooling Water (1) - no reflux
• range 0.00124-0.00178 m3/kg feed
• typical 0.00135 m»/kg feed
Electricity Negligible
Operating Labor 2 man-hours/shift (1,8)
Maintenance 5.9 % of installed equipment cost
fFrom Tables B7-1 and B7-2.
Four percent of total capital cost (1,3,8).
7. References
U.S. Department of Energy. Conceptual Designs for Water Treatment in
Demonstration Plants. Contract No. EF-77-C-01-2635, March 1979.
U.S. Environmental Protection Agency. Wastewater Treatment in Coal
Conversion. EPA-600/7-79-133, June 1979.
U.S. Department of Energy. Predicted Costs of Environmental Controls for
a Commercial Oil Shale Industry. Contract No. EP-78-S-02-5107, July 1979.
American Petroleum Institute. Chapter 10-Stripping, Extraction,
Adsorption, and Ion Exchange. In: Manual on Disposal of Refinery
Wastes, Volume on Liquid Wastes, 1973.
Gantz, R.G. Sour Water Stripper Operations. Hydrocarbon Processing,
May 1975, pp. 85-88.
U.S. Environmental Protection Agency. Treatability Manual. Volume III:
Technologies for Control/Removal of Pollutants. EPA-600/8-80-042c, July
1980.
Mel in, G.A. et al. Optimum Design of Sour Water Strippers. Chemical
Engineering Progress, 71(6): 78-82, June 1975.
U.S. Environmental Protection Agency. Treatability Manual. Volume IV:
Cost Estimating. EPA-600/8-80-042d, 1980.
B7-14
-------
APPENDIX B8
STEAM STRIPPING AMMONIA RECOVERY: PHOSAM-W PROCESS
The PHOSAM-W process is a proprietary process of DSS Engineers and
Consultants Incorporated. It is a development of the original PHOSAM process
which has been used to recover ammonia from coke oven gases.
1. Process Description
A typical flow diagram of the PHOSAM-W process is shown in Figure B8-1.
The process basically consists of:
• steam stripping ammonia, hydrogen sulfide, carbon dioxide,
phenols, cyanides, organics, and other volatile compounds from
wastewater;
• absorbing ammonia out of the stripped vapors into a phosphoric
acid solution;
• regenerating the absorbing solution by steam stripping ammonia
and water; and
• condensing and distilling the ammonia-water mixture to
anhydrous ammonia.
The process produces anhydrous ammonia as a product, stripped wastewater which
passes to further treatment or use, and acid gases and other stripped vapors
which pass to further treatment or disposal. Key features of the process are
highlighted below.
Water Stripping
Water stripping in the PHOSAM-W process is essentially the same as strip-
ping described in Appendix B7, Steam Stripping.
B8-1
-------
Acid
Gases
CO
00
I
Wastewater
Feed
Aqua
Ammonia
i
s^
I
^
S
T
R
I
P
P
E
R
•4-
r
<
|J
<
T
cw
F
R
A
C
T
I
0
N
A
T
0
R
Product
-*- Ammonia
Steam
Condensate
Stripped
Wastewater
Figure. B8-1. The PHOSAM-W process
-------
Appendix B8
PHOSAM-W
The wastewater feed to the stripper is first preheated to its bubble
point partly by heat exchange with stripped wastewater and partly by heat
exchange with lean phosphoric acid solution. (Steam preheat may also be
necessary.) The preheated wastewater enters near the top of the water strip-
per and then flows by gravity countercurrently to steam and vapors rising up
from the bottom of the stripper. As the wastewater flows downward, it is
heated and volatile compounds and gases are stripped by the steam and vapors
rising from the bottom of the stripper.
Stripping steam may be live steam or wastewater evaporated by a steam
reboiler. If live steam is used, steam condensate enters the stripped waste-
water and is not recovered. If a reboiler is used, reboiler steam is usually
supplied at 440 to 520 kPa (1,2). As shown in Figure B8-1, some of the
required heat duty may be provided by flashed reboiler condensate from the two
downstream towers and by the bottoms of the ammonia fractionator. The bottoms
stream would typically be flashed directly into the water stripper.
The water stripper used in the PHOSAM-W process typically consists of
between 20 to 40 actual trays which correspond to 7 to 15 theoretical stages.
The stripping steam requirement is typically 0.11 to 0.13 kg steam per kg
wastewater feed (1). The water stripper normally operates at 130 to 170 kPa.
Phenols stripped from the wastewater may be absorbed in the ammonia
absorber and subsequently carried over to the ammonia fractionator. In the
ammonia fractionator, the phenols would be "fixed" by caustic. Refluxing
(i.e., condensing a portion of the steam and vapors leaving the stripper and
returning it to the stripper) would hold down the phenols in the stripper
overhead and consequently minimize the consumption of caustic in the ammonia
fractionator. Refluxing would also increase the concentration of ammonia in
the stripper overhead and reduce the temperature in the ammonia absorber,
thus decreasing absorber size and cost. However, refluxing increases the
water stripper cost and the requirement for stripping steam and condenser
B8-3
-------
Appendix B-8
PHOSAM-W
cooling water. Hicks, et al. expect that refluxing will rarely be used to
hold down phenols. USS Engineers and Consultants have found that refluxing to
increase the overhead ammonia concentration is beneficial only if the
wastewater feed has an ammonia concentration of less than 6000 mg/L (1).
Ammonia Absorption
The lean phosphoric acid solution used to absorb ammonia from the strip-
ped vapors contains slightly more than one ammonium ion per phosphate ion.
After absorbing ammonia, the rich solution contains slightly less than two
ammonium ions per phosphate ion.
Absorption temperatures are usually maintained below 377 K. Since the
absorption is exothermic, solution is continually pulled out of the absorber,
cooled in an exchanger, and returned to the absorber as spray. The cooled
solution cools the incoming stripped vapors; water condenses out of the
saturated stripped vapors as they are cooled. The condensed water is carried
with the phosphoric acid solution until it is stripped during solution regen-
eration (1) .
The ammonia absorber operates at about the same pressure as the water
stripper. The upper portion of the absorber where lean phosphoric acid con-
tacts the stripped vapors consists of trays. The lower portion of the absor-
ber consists of sprays for returning the cooled, recirculating phosphoric acid
solution. Other details of the absorber design are proprietary (1).
Solution Regeneration
Water and ammonia are steam stripped from the rich phosphoric acid solu-
tion at about 1.5 MPa (3).
B8-4
-------
Appendix B8
PHOSAM-W
Stripping steam can be supplied by live steam or by a steam reboiler. If
a reboiler is used, steam is usually supplied at 2.2 MPa or above. Live steam
dilutes the stripped ammonia-water mixture (1).
The ammonia-water mixture, known as aqua ammonia, is recovered as 10 to
20 percent ammonia and condensed by heat exchange with rich solution and
cooling water (1).
Ammonia Distillation
The aqua ammonia is distilled in the ammonia fractionator to produce
liquid anhydrous ammonia consisting of up to 99.99 percent ammonia. The
anhydrous ammonia typically contains less than 100 ppm water, 2 ppm oils, 3
ppm carbon dioxide, 2 ppm chlorine, and no detectable amounts of hydrogen
snlfide (3). A fraction of the liquid ammonia product is recycled to the
fractionator as a reflux.
The fractionator operates at about 1.5 MPa (3). Although the steam rates
depend slightly on the concentration of the aqua ammonia, they will generally
range from 0.2 to 0.4 kg steam per kg of feed (1).
A small amount of caustic is added to the fractionator to prevent the
accumulation of acid gases and organic compounds. The fixed compounds exit
with the fractionator bottoms and eventually with the stripped wastewater.
2. Process Applicability
No specific limits of process applicability are reported in the open
literature. The PHOSAM-W process is applicable to recovering ammonia from
wastewaters on any scale of application but it may be uneconomical for applica-
tions recovering less than 0.11 to 0.16 kg/s of ammonia (2,3). No steam
stripping process may be suitable for recovering chemically-fixed ammonia.
B8-5
-------
Appendix B8
PHOSAJh-W
3. Process Performance
The performance of the PHOSAM-W process in removing ammonia, hydrogen
sulfide, and other volatile compounds from wastewaters depends on the
characteristics of the wastewater feed and water stripper design. Thus, the
performance of the PHOSAM-W process is generally the same as the performance
of the water strippers described in Appendix B7, Steam Stripping. The perfor-
mance of strippers in removing ammonia, hydrogen sulfide, carbon dioxide, and
phenols is shown in Tables B7-1 and B7-2. The performance of strippers in
removing COD, TOC, and specific organic compounds is reported in Table B7-3.
The most efficient strippers of Tables B7-1 and B7-2 are generally shown
to strip ammonia down to 50 to 100 mg/L, although three strippers achieve
levels of less than 15 mg/L. Strippers achieving ammonia levels of less than
100 mg/L remove hydrogen sulfide to 0 to 30 mg/L depending on the initial
ammonia, hydrogen sulfide, and carbon dioxide concentrations. Carbon dioxide
is totally stripped. The effectiveness of stripping is limited when the
volatile compounds are chemically fixed in the wastewater, e.g., the fixing of
ammonia by hydrochloric and naphthenic acids (4). Thus, the ability to
achieve low levels of ammonia (e.g., 100 mg/L) will depend on the amount of
fixed ammonia present in the wastewater; pH adjustment may improve the removal
of ammonia.
For several types of wastewaters from synfuels facilities typically 17 to
22 percent of the residual phenols are stripped. For refinery-type waste-
waters, unrefluxed strippers on the average remove 55 percent of the phenols
and refluxed strippers on the average remove 39 percent of the phenols (5).
Cyanide removal efficiencies are reported to vary widely but average
37 percent for stripping refinery-type wastewaters (5).
B8-6
-------
Appendix B8
PHOSAM-W
4. Secondary Waste Generation
The only secondary waste stream from PHOSAM-W plants is the acid gas or
vapor which exits the ammonia absorber. This stream is saturated with water
and contains acid gases such as carbon dioxide and hydrogen sulfide and other
volatile compounds. This stream also contains a small amount of ammonia.
typically comprising 0.1 to 0.5 percent of the total gas volume (6).
5. Process Reliability
The original PHOSAM process is widely used by DSS and licensors to treat
coke-oven gas. The derivative PHOSAM-W process was in the design and con-
struction phase at four plants in 1979 including three synfuels applications.
The reliability of the basic PHOSAM process has been established in 14 operat-
ing plants which have been operating for up to 10 years. The service factor
for these plants has been about 95 percent (3).
6. Process Economics
The unit installed equipment costs of the PHOSAM-W process are shown in
Figure B8-2. The bases for the installed equipment costs and for operating
costs are listed in Tables B8-1 and B8-2.
The installed equipment cost correlation in Figure B8-2 is a correlation
of updated costs for eleven distinct PHOSAM-W applications as reported in
References 2, 3, 6, 7, and 8. The costs vary mainly with the wastewater flow-
rate but variations due to different feed characteristics (such as ammonia,
hydrogen sulfide, and carbon dioxide concentrations) are also evident. Al-
though these other differences can significantly affect costs, the available
cost data base is insufficient to consider their precise impact on costs.
B8-7
-------
106
7
10 =
~ 7
4
• 2
10'
1st Quarter 1980 Dollars
Note: See Table B8-1 for
data sources.
Numbers on data
points are ammonia
concentration in
feed, mg/L
6000
000
5100
4000
' I I I I i
J - 1 - 1
i I
10
100
Wastewater Feed, m /hr
1000
Figure B8-2. Unit installed equipment costs of PHOSAM-W® process
B8-8
-------
TABLE B8-1. BASIS FOR INSTALLED EQUIPMENT COSTS OF PHOSAM-W PROCESS
oo
VO
Type of
Facility
Lnrgi
SBC II
Hygaa/
Lnrgi
SBC
Low-Btn
Low-Btn
Low-Btn
P.r.10
Toaco II
Modified
In-Sitn
Oil Shale
Lnrgi
laatewater Feed Characteriattoa
Concentration. »t/L Flowrate,
NH, B,S CO, »«/hr
12,000 510 16,800
20,000 19,300 24,900
3,500- low 13.000
4.500
12.600 14,600 4,000
6000 — —
6000 — —
6000
17,990 32,400 206
5100 900 6200
21,330 118 41,800
17,000 100 33,000
*IBL - Inatalled Battery Li»lta;
** A — ___ I _ _ *t._* J_^*...11_.J U_ *• *•»»
570
170
710
570
23
45
115
293
105
468
436
DIC - Direct
1 I _1 * a MA * *• e
Date of
Original
Coat Data
1978
1978
1976
1976
1978
1978
1978
1978
1978
1978
1978
Inatalled
Original
Coat Data
IBL-il4.000.000*
IBL-i9.000.000*
Capital-iS, 000 ,000
Capital-iS, 200, 000
DI Oil, 800, 000*
DIC-i2,720,000*
DIC-i4,720,000*
Capital-i6,306,000
Capltal-i6,040,000
Capltal-iS. 600, 000
Capital-ill, 300. 000
Coat.
t*r\ * + iwt«i1wi*4**> **IBT|«****
BitUated
Inatalled
Equipment
Coat
i9,460.000b
i6.080,000b
i5,410,000b
i5,540,000b
il, 800, 000
$2,720,000
i4, 720. 000
i4,260.000b
i4,080.000b
i5,810,000b
*7,640,000b
I M • • 1\A
Updated
Inatalled
Equipment
Coat Beferenc*
iio
i«.
i7.
i7.
12,
13,
15,
14,
i4.
i6.
18,
,840,000
970,000
050,000
230.000
100,000
170.000
500.000
880.000
680.000
660 ,000
750,000
7
7
8
8
3
3
3
6
6
6
2
conatrnction at 25% of ioatalled equipment, contingency at 20% of installed eqnlpaent, and
feea at 3% of installed eqnipaent.
clat Quarter 1980 dollar*.
-------
Appendix B8
PHOSAM-W
Thus, the correlation in Figure B8-2 is simply a correlation of cost versus
wastewater flowrate. Despite this simplification, a "good" fit of data
results, with an explained variance of about 93 percent. The correlation in
Figure B8-2 is:
Unit Installed Equipment Cost, $/(ms/hr) = 795,090 Q
-0.63
where: Q is the wastewater flowrate in mj/hr.
The unit installed equipment costs of actual PHOSAM-W applications can differ
from costs estimated with this correlation due to differences in feed and
stripper characteristics and in cost scope. The reported correlation is
probably accurate only to +50 percent.
TABLE B8-2. BASIS FOR OPERATING COSTS OF PHOSAM-W PROCESS
Steam (7)
3.9 MPa
0.28 MPa
Cooling Water (7)
Electricity (7)
Chemicals (7)
HjP04 (100%)
NaOH (100%)
Operating Labor (3)
Maintenance
12 kg/kg NHs
8 kg/kg NHa
0.33 mj/kg NHs
0.066 kWh/kg NHs
0.001 kg/kg NHs
0.002 kg/kg NHs
2 man-hours/shift
7.40 percent of installed
equipment cost
aFive percent of the total capital cost (3).
B8-10
-------
Appendix B8
PHOSAM-W
7. References
1. U.S. Environmental Protection Agency. Wastewater Treatment in Coal
Conversion. EPA-600/7-79-133, June 1979.
2. U.S. Department of Energy. Conceptual Designs for Water Treatment in
Demonstration Plants. No. EF-77-C-01-2635, March 1979.
3. U.S. Department of Energy. Costs and Technical Characteristics of
Environmental Control Processes for Low-Btu Coal Gasification Plants.
No. W-7405-eng-26. June 1980.
4. Melin, G.A. et al. Optimum Design of Sour Water Strippers. Chemical
Engineering Progress, 71(6):78-82, June 1975.
5. Gantz, R.G. Sour Water Stripper Operations. Hydrocarbon Processing.
May 1975, pp. 85-88.
6. U.S. Department of Energy. Predicted Costs of Environmental Controls for
a Commercial Oil Shale Industry. No. EP-78-S-02-5107, July 1979.
7. U.S. Environmental Protection Agency. Coal Conversion Technology:
Volume I. Environmental Regulations; Liquid Effluents. EPA-600/7-79-
228a, October 1979.
8. U.S. Environmental Protection Agency. Water Conservation and Pollution
Control in Coal Conversion Processes. EPA-600/7-77-065, June 1977.
B8-11
-------
-------
APPENDIX B9
STEAM STRIPPING AMMONIA RECOVERY: CHEVRON WWT PROCESS
The Chevron Wastewater Treatment (WT) process is a proprietary process
developed by Chevron Research Company. Although the process was originally
developed to treat sour refinery wastewaters, its developers claim it to be
capable of treating sour wastewaters from coal processing and synthetic fuel
plants.
1. Process Description
A typical flow diagram of the Chevron WWT process is shown in Figure
B9-1. The process basically consists of sequentially stripping first acid
gases such as hydrogen sulfide and next ammonia from a foul water. The
recovered ammonia is concentrated and washed to yield a high-purity anhydrous
ammonia product. In refinery wastewater applications, a high-purity hydrogen
sulfide stream is produced and may be fed to a sulfur recovery plant. In
other applications, the hydrogen sulfide stream may be diluted by other acid
gases such as carbon dioxide but sulfur recovery may still be feasible.
The wastewater feed is first taken to a degasser where hydrogen, methane,
and other light gases are flashed from the wastewater. The flashed gases may
be further processed, recovered, or incinerated. The wastewater feed is then
combined with a recycled aqueous ammonia stream and fed to a purge tank. The
tank is used to even out variations in the wastewater flow and also allows the
skimming of any floating hydrocarbons (1,2).
The wastewater stream is heated by exchange with stripped water and en-
ters the acid gas (hydrogen sulfide) stripper. The stripper operates at about
366 K and 790 kPa (1,2). Heat for stripping is provided by a steam reboiler.
Hydrogen sulfide and other acid gases are stripped upwards while ammonia and
B9-1
-------
HYDROGEN SULFIDE
COOLING
WATER
^ACCUMULATOR^
STRIPPED WATER REFLUX
FLASH GAS HYOROGEN-
SULFIOE
STRIPPER
I
ro
FOUL WATER,
OEGASSER
SURGE
TANK
AMMONIA
STRIPPER
STEAM
(
STEAM
HYDROGEN-SULFIOE/AMMONIA RECYCLE
AMMONIA
SCRUBBERS
LIQUtO
'AMMONIA
COMPRESSOR
STRIPPED WATER
Figure B9-1. The Chevron WWT process (1)
-------
Appendix B9
Chevron WWT
water flow downward conntercnrrently to the stripped vapors. As shown in
Figure B9-1, stripped wastewater is typically used to provide reflux to the
acid gas stripper (2). The purpose of the reflux is to hold ammonia in the
wastewater.
The acid gas taken from the stripper typically contains less than SO ppmw
ammonia and less than 5000 ppmw water (2). The purity of this acid gas may be
affected by other contaminants in the wastewater feed, but these effects are
not discussed in the open literature.
The bottoms from the acid gas stripper are taken to the ammonia stripper.
The ammonia stripper operates at about 394 K and 450 kPa (2). Heat for strip-
ping is provided by a steam reboiler. The overhead stream, comprising mainly
ammonia, is cooled and partially condensed. Condensate is separated in an
accumulator and used as reflux to the ammonia column and as a recycle to the
surge tank. The vapor stream from the accumulator is still crude, comprising
about 98 % ammonia with water and hydrogen sulfide and other gases (1). The
stream is scrubbed, compressed, and condensed to yield anhydrous ammonia pro-
duct. This product is available at about 311 K and 1.5 MPa (1). It typically
contains less than 1000 ppmw water and less than 5 ppmw hydrogen sulfide (2).
The stripped bottoms of the ammonia stripper typically contains less
than 5 ppmw hydrogen sulfide and less than 50 ppmw ammonia (2). The purity of
this stream and the acid gas and ammonia streams can be adjusted by changing
operating conditions.
2. Process Applicability
No specific limits of process applicability are reported in the open
literature. Like the PHOSAM-W process, the Chevron WWT process may be
uneconomic for smaller applications recovering less than about 0.16 kg/s of
B9-3
-------
Appendix B9
Chevron WWT
ammonia. Because of its operating principle (physical separation), the
Chevron WWT process could be adversely affected by contaminants in the waste-
water. Although not discussed in the open literature, the degree of separa-
tion and product purities may be affected by contaminants that change the
relative volatilities of ammonia and hydrogen sulfide.
3. Process Performance
As stated above, the stripped wastewater from the Chevron WWT process
typically contains less than 50 ppmw ammonia and less than 5 ppmw hydrogen
sulfide. These levels have been routinely achieved in a number of refinery
applications. Any carbon dioxide in the wastewater is totally stripped in the
acid gas column (2). Removals of other compounds by the Chevron WWT process
have not been reported in the open literature. Whether low levels of ammonia
can be achieved with synfuels wastewaters will depend on the level of fixed
ammonia; pH adjustment may be necessary prior to ammonia stripping in order to
increase ammonia removal.
4. Secondary Waste Generation
Secondary wastes from the Chevron WWT process include the acid gas which
exits the first fractionating column, a light flash gas from the degasser, and
unspecified waste products from the ammonia scrubbers. As stated above, the
acid gas contains only small amounts of ammonia and water and subsequent
removal of hydrogen sulfide in sulfur plants is usually technically feasible.
Flaring or incineration is another disposal alternative. Light hydrocarbons
and other volatile compounds in the light flash gas can be recovered, further
processed, or incinerated.
B9-4
-------
Appendix B9
Chevron WT
5. Process Reliability
As of 1978, the Chevron WT process had been employed in ten commercial
facilities, mainly at refineries. No specific reliability measures are
reported in the open literature. Since the process uses standard equipment
and operations, reliability should be high.
6. Process Economics
The unit installed equipment costs of the Chevron WWT process are shown
in Figure B9-2. The bases for the unit installed equipment costs and for
operating costs are listed in Tables B9-1 and B9-2. The cost curve for
installed equipment cost is a correlation of updated costs for five Chevron
WT applications as reported in References 2 and 3. These costs vary mainly
with the wastewater flowrate but variations in costs due to differences in
feed characteristics may be significant. The available cost data base, how-
ever, is inadequate to consider these additional cost variations. The cost
curve shown is thus simply a correlation of cost versus wastewater flowrate.
Despite this simplification, a "good" fit of data results with an explained
variance of about 98 percent. The correlation of Figure B9-2 is:
Unit Installed Equipment Cost, i/(mj/hr) = 911,250 x Q
where: Q is the wastewater flowrate in mj/hr.
The unit installed equipment costs of actual Chevron WT applications differ
from the costs estimated from the correlation due to differences in feed and
stripper characteristics and in cost scope. The reported correlations are
thus probably accurate only to +50 percent.
B9-5
-------
106
7
10s
7
1st Quarter 1980 Dollars
Note: See Table B9-1 for
data sources.
Numbers on data
points are ammonia
concentration in
feed, mg/L
•37,000
.37,000
20,000
•
,12,000
2 -
J 1 1 I I I I
J 1
10
7 100
2
1000
Wascewater Feed, tn'/hr
10,000
Figure B9-2. Unit installed equipment cost of Chevron WWT process
B9-6
-------
TABLE B9-1. BASIS FOR INSTALLED EQUIPMENT COSTS OF CHEVRON WWT PROCESS
Cd
^D
I
Type of
Facility
Lnrgi
SRC II
Low-Btn
Low-Btu
Low-Btu
Refinery
Refinery
laatewater Feed Characterit tict
Concentration. att/L Flowrata.
NH,
12 ,000
20,000
37,000
37 ,000
37 .000
38.000
76,000
H.S CO,
510 16,800
19.300 24,900
63.000
63 .000 —
63,000 —
38,000 —
76,000
• >/hr
570
170
23
45
115
30
15
Date of
Original
Coat Data
1978
1978
1978
1978
1978
1970
1970
Original
Coat Data
TIOil2,000,000*
TIOl6.000.000*
DI Oil, 410, 000*
DI 013,100, 000*
DIOi4.320.000*
Invett»ent-i597,000
Invet t»ent-i434 ,000
Batiawted
Inatalled
Equipment
Coat
i8,no,ooob
i4,050,000b
82,410,000
i3, 100, 000
t4. 320. 000
i403.000b
i293.000b
Updated
Inatalled
Equipment
Coat
i9, 290, 000
44,650,000
82.810,000
i3, 610, 000
is, 030, ooo
i805,000d
i585,000d
Reference
3
3
2
2
2
1
1
'TIC - Total Inatalled Colt; DIC » Direct Inatalled Coat.
Attumlng that Total Inatalled Coat or Inveataent include! engineering and conatrnction at
25% of initalled equipment, contingency at 20% of inttalled equipment, and feea at 3% of
inttalled equipaient.
jl*t quarter 1980 dollart.
Theae ettimatet are not need in developing a correlation aince they are Markedly inconaiatent
with the other coat data.
-------
Appendix B9
Chevron WWT
TABLE B9-2. BASIS FOR OPERATING COSTS
OF CHEVRON WWT PROCESS
Steam (3)
- 1.1 MPa 13 kg/kg NHs
- 0.45 MPa 11 kg/kg NH»
Cooling Water (3) 0.18 m3/kg NHs
Electricity (3) 0.29 kWh/kg NHab
Chemicals (3)
- Corrosion inhibitor 0.11 i/m'c
- NaOH (100%) 0.40 kg/m»
Operating Labor (2) 2 man-hours/shiftd
Maintenance 7.40 percent of installed
equipment cost6
aAverage of 0.13 and 0.23 rnVkg.
bAverage of 0.26 and 0.32 kWh/kg.
°Average of 0.095 and 0.12 J/m»; escalated by 15 % over original
1978 data.
^Reference 3 estimates labor as 4 man-hours/shift.
Equivalent to 5 % of the total capital cost (2). Reference 3
estimates maintenance as 3 % of the total capital cost.
7. References
1. Annessen, R.J. and G.D. Gould. Sour Water Processing Turns Problem into
Payout. Chemical Engineering. March 22, 1971, pp. 67-69.
2. U.S. Department of Energy. Costs and Technical Characteristics of
Environmental Control Processes for Low-Btu Coal Gasification Plants.
No. W-7405-eng-26, June 1980.
3. U.S. Environmental Protection Agency. Coal Conversion Control
Technology: Volume 1. Environmental Regulations; Liquid Effluents. EPA-
600/7-79-228a. October 1979.
B9-8
-------
APPENDIX BIO
ACTIVATED SLUDGE PROCESS
1. Process Description
Activated sludge processes are widely used for the treatment of municipal
and industrial wastewaters which contain organic and inorganic pollutants
amenable to biological degradation. A complete activated sludge system con-
sists of two key components: a reactor and a clarifier as shown in Figure
B10-1. Organic-laden wastewater is introduced into the reactor where an
aerobic bacterial culture is maintained in a suspension called the mixed
liquor. The bacterial culture or sludge accomplishes the conversion of or-
ganic materials to carbon dioxide, water, metabolic intermediates, and
ammonia. The sludge concentration that is maintained in the reactor is depen-
dent upon the desired treatment efficiency and other application-specific fac-
tors. Oxygen is supplied to the reactor by aeration with air or an oxygen
enriched stream. The aeration is accomplished in a manner that keeps the bio-
logical sludge suspended in the mixed liquor. This insures that wastewater
and sludge are in intimate contact and that a high oxygen transfer efficiency
is maintained. In a well mixed system, a portion of the mixed liquor is con-
tinuously passed into a settling tank or clarifier where the sludge is
separated from the treated wastewater. A portion of the settled sludge is
recycled to the reactor in order to maintain the proper concentration of
microorganisms in the reactor. The remainder of the sludge is removed from
the system. The amount of waste sludge generated at this point corresponds to
the net growth rate of sludge in the reactor.
A modification of the activated sludge process involves the addition of
powdered activated carbon to achieve simultaneous biological and physical-
chemical treatment of wastewaters. In the activated carbon enhanced acti-
vated sludge (ACEAS) process, the aeration basin is charged with wastewater,
B10-1
-------
o
I
to
Air or
Oxygen
Influent
1
Recycle Sludge
Treated
Effluent
VLARIFIER
Waste Sludge
Figure B10-1. Activated sludge system
-------
Appendix BIO
Activated Sludge Process
makeup carbon, and recirculated sludge. The microorganisms grow on the carbon
particles in a similar fashion to the fixed media processes. The carbon helps
retain organic molecules near the organisms for faster metabolism (1). In the
clarifier, the carbon-weighted sludge generally has better settling character-
istics than sludge from conventional activated sludge processes (2). As a re-
sult of the physical adsorption properties of activated carbon, ACEAS systems
are more resistant to shock loadings of inhibitory compounds than are conven-
tional air activated sludge (AAS) systems. The activated carbon could also
have the added benefit of adsorbing material which is not normally removed in
biological systems. In the treatment of high strength wastewaters with
complex organics, ACEAS appears to have many advantages over conventional
activated sludge processes in terms of both performance and reliability.
An activated sludge system must be designed and operated to reflect the
microorganism growth kinetics associated with the subject waste stream. With
some restraints, the activated sludge can be maintained in the reactor for any
desired period. This is accomplished by varying the relative rates of sludge
recycle and wastage. The mean cell residence time or sludge age is the aver-
age retention time of active microorganisms in the treatment system. Sludge
age is directly related to the organic removal efficiency and the settling
characteristics of the sludge. It is therefore a critical factor in the de-
sign and operation of an activated sludge process. Through the control of the
sludge wastage rate, sludge age can be varied to optimize characteristics of
the effluent and sludge leaving the activated sludge treatment system.
2. Process Applicability
An activated sludge process may be used as a part of an overall "end-of-
pipe" treatment system for synfuel plant wastewater streams containing high
concentrations of biodegradable organic pollutants. Waste streams that have
high concentrations of suspended solids, inorganic gases, and dissolved
BIO-3
-------
Appendix B-10
Activated Sludge Process
organics may require pretreatment upstream of the activated sludge process.
This pretreatment will reduce the concentration of potentially toxic or inhi-
bitory materials entering the activated sludge system. Activated sludge pro-
cesses are sensitive to variations in flow and composition. Therefore,
another essential part of the pretreatment scheme is the equalization basin
normally used to achieve complete mixing of the input wastestreams. The use
of surge capacity to equalize flow and promote a uniform composition is cri-
tical where several wastestreams are combined for treatment in an activated
sludge system. In addition to the operational benefits derived from flow
equalization, the activated sludge treatment system can be sized to treat an
average as opposed to a peak flowrate. This may result in reduced capital and
operating costs for the activated sludge system.
Application of the activated sludge process to synfuels plant wastewaters
has been demonstrated on a bench scale only. Air and powdered activated car-
bon enhanced activated sludge systems have been successfully used in a variety
of industrial applications including the treatment of byproduct coking, refi-
nery, and organic chemical industry wastewaters. The application of activated
sludge processes to a given wastewater or combined wastewater streams requires
careful consideration of many complex factors that effect the performance of
the system and the effluent quality.
The application of activated sludge processes to coal conversion waste-
waters is highly dependent on the dissolved organic concentration in the in-
fluent wastewater. Activated sludge processes operate most efficiently on
wastewaters which contain readily biodegradable dissolved organic materials.
When BOD/COD ratios are low, there is a high probability that the effluent
from the activated sludge process will contain appreciable concentrations of
complex organics. When these wastewater conditions exist, the addition of
powdered activated carbon to the reactor or polishing of the clarifier ef-
fluent may be required to obtain an acceptable effluent quality.
BIO-4
-------
Appendix BIO
Activated Sludge Process
Another critical consideration in the use of activated sludge technology
is the ability to transfer oxygen to the microorganisms in the reactor. Be-
cause of the high organic concentrations, large quantities of oxygen (400-600
mg/L-hr) must be transferred in comparison to municipal sewage treatment
systems (5). Oxygen transfer can pose an upper limit on the number of micro-
organisms that can be supported and the degree of organic destruction that
occurs. Oxygen transfer limitation will also impact the required hydraulic
retention time (reactor volume) and subsequently the cost of the system.
It is difficult to predict the performance of an activated sludge unit
based only on influent wastewater quality data. For this reason, bench scale
or pilot testing of the activated sludge process to determine its applicabil-
ity to a synfuel plant wastewater stream is essential.
3. Process Performance
The activated sludge process is used to treat a wide range of industrial
wastewaters. Some of these wastewaters, such as weak ammonia liquor from by-
product coking operations, are similar in composition to many synfuel plant
wastewaters. Because there are no synfuel facilities operating in the United
States at this time which use this process, knowledge about the performance of
an activated sludge process in treating these wastewaters is currently limited
to foreign experience and bench scale treatability data. Comparisons of per-
formance data for the activated sludge treatment of similar wastewaters from
other industrial processes with synfuel process wastewater treatability data
is essential in predicting the performance of a full scale synfuel plant
wastewater treatment system. Tables B10-1 and B10-2 summarize some activa-
ted sludge process performance data from studies involving numerous byproduct
coking and synfuel plant wastewater treatability studies. These data were de-
veloped for the most part from single stage completely mixed systems.
BIO-5
-------
TABLE B10-1.
w
o
I
COMPARATIVE PERFORMANCE DATA FOR ACTIVATED SLUDGE SYSTEMS TREATING
WASTEWATERS FROM BYPRODUCT COKING AND VARIOUS SYNFUELS PROCESSES
Characterization
Proceaa
Cok. Plant*
OFETC*
Hy»it°
Synthane
•ETC*
H-Coalf
Synthetic*
Vaatewatar
SBC-Ik
•aatewater
faatawatar Source ODD BOO
Ammonia Still Effluent 3900-4600 1700-2(00
Biox Efflaent 300-410 5-15
Pretreated Biox Influent 5380-7130 3000-6700
Biox Efflneit 1110-1340 73-230
Baw Proceat Effluent 3400-5300 29(0
Pretreated Blox Influent 3540-4190 2570-3090
Biox Effluent 660-890 270-450
Pretreated Biox Influe.t 4800 2290
Biox Bfflueut 820-1600 (0-180
•aw Proceai Effluent 88,600 52.700
Pretreated Blox Influent 3070-41(0 1890-1600
Biox Effluent 310-3(0 24-36
Biox Effluent 190-1800 5-1000
Baw Recycle Proceea 60.000 16.900
Pretreated Influent 2,000 1200
Blox Effluent 250 t
Pkenol lot
750-1000
0.25-1.4
3500-6500
1090-1730
0.6-1.6
600-900
620-940
0.1-1.9
3000
173-1205
920
2.5-7.7
6(00
750-1450
0.3-9.7
4650
1900
280
1
. M/L
ICN-
280-510
3-36
(0-200
14-27
1-3
17-45
12-24
2-4
31
NB.-N Ori-N
35-92 21-27
23-163 4-8
4000-7300 70-140
14-54 17-21
136-148 10
•7-122 1-8
250
14.400 51
68-140 10-24
40-140 5-13
NO.-N TOC
<1
90-280
2290
11 1000
2-7 45-200
'Effluent BOD data are for a reactor allowing a well nitrified effluent ({).
ol» (ailfier effluent from Indian Bead li|nlte. Pretreated for ammonia removal; dilated to 33% atrength (7).
.Cyclone and quench efflnenta fro* Illinois eubbi tuminoua. Pretreated for aaaonla removal; no dilution (8).
,Pretreated for anon!* removal and diluted to 36-40% atrength.
Foul water from coal liquefaction. Pretreated for aulfide and
'Pretreated by dilution to 25% atreofth (12).
Pretreatod for auoni* and tar told removal before being dilated to 20% (13).
Ineee data are fro» teat eeriea c with producer Bun 90 waatewater (10).
uonla removal and diluted to 22% atrength (11).
-------
TABLE B10-2.
KINETIC AND OPERATIONAL PARAMETERS CORRESPONDING TO THE ACTIVATED
SLUDE PROCESS PERFORMANCE RESULTS PRESENTED IN TABLE B10-1
W
O
I
Coke Flint
flwrrcb
»7g«
Syn thane
MBTC*
H-Coal
Synthetic Coal
CoavereioB.B
SBC-Ik
retreated Influent
Concentration, (percentage)
••actor pB
Loading (mg COD (BOD) applied/a^
MLVSS-day)
Substrate leaoval Rate
(•I COD [BOD] re.or.d/.,
NLVSS-day)
Nean Cell Residence Tine
(9 .days)
Rydraullc Residence Time
(8. days)
100
7.0-7.5
0.18-0.82
[0.11-0.551
0.16-0.It
10-40
1.8-5.7
7.0-7.1
100
7.0-7.5
15
36-40
«.2-7.4
25
4.0-7.5
10-35
0.25-0.51 0.43-1.29
[0.15-0.341 [0.32-0.95]
0.21-0.41 0.37-1.0*
(0.14-0.321 [0.29-0.Ml (0.2-0.61
10-40
10-40
>.1-21.< 2.05-2.91
3.5-31
1.0
[0.38-0.64] [0.06-0.22]
0.29-1.13
[0.35-0.601 (0.04-0.241 [0.11-0.71]
4.4-11 ll-MOO 5-20
0.44
0.6-2.2
5-20
Yield Coefficient. I
COD basis (BOD bssis]
Decay Coefficient, k
days , COD basis
[BOD basis]
Oxygen Utilization Coefficient
(•I 0 /•« COD [BOD] removed)
Oxygen Dtilizatiou Coefficient
(mg 0>/mg MLVSS-day)
COD baais [BOD basis]
0.13
0.002
0.24
0.11
0.29
(0.43)
0.03*
(0.0371
0.77
(1.01
0.01
(0.051
0.10
(0.141
0.01
[0.02]
0.30
(0.361
0.06
(0.051
[3.71
[0.0331
0.56
[0.751
0.09
[0.111
(0.541
(0.09]
[0.30]
[0.0(1
!f fluent BOD data are for a reactor showing a well nitrified effluent (6).
law gaslfier effluent from Indian Bead lignite. Pretreated for ammonia removal; diluted to 33% strength (7)
?yoloue and quenoh effluents from Illinois sabbitnmlnoos. Pretreated for ammonia removal; no dilution (8).
0.1«
[0.4(1 (0.271
0.005
(0.031 [0.0051
[0.52]
and -25% (9).
0.56
0.02
0.61
0.02
Byproduct water from Montsns Rosebud subbltomlnous. Pretreated for ammonia removal and dilated to ~5 to 6% and -25% (9).
'Pretreated for ammonia removal and dilated to 36-40% strength. These data are from test aeries c with producer Run 90 wastewater
Foal water from coal liquefaction. Pretreeted for sulflde and ammonia removal and dilated to 22% strength (11).
JPretreited by dilation to 25% strength (12).
(10).
.rEvirvnieu oy dilution to 421* strengin ii£j.
Pretreated for ammonia and tar acid removal before being dilated to 20% (13).
MLVSS - mixed liquor volatile suspended solids.
-------
Appendix BIO
Activated Sludge Process
The data presented in Table B10-1 indicate that the activated sludge
process is effective in reducing phenolic compounds. Where synfuel process
and coking wastewaters are comparable in composition, the performance of the
process is similar. Overall performance of the activated sludge process in
terms of COD, BOD, and TOC removals varies greatly with the wastewater being
treated. This variability in performance precludes the development of mean-
ingful industry-wide performance criteria based solely upon the use of gross
wastewater characterization criteria.
Table B10-3 presents relatiye conventional pollutant removal effective-
ness data for activated sludge systems operating with and without powdered ac-
tivated carbon. The data presented represent experience obtained in
industries other than the synthetic fuels industry. Since these data repre-
sent a range of applications of the technology, they are valid for overall
comparison purposes only.
In general, data on powdered activated carbon enhanced systems are
limited in comparison to those for conventional activated sludge systems.
Furthermore, the streams that can be treated by the two processes are not
necessarily comparable. However, even when these limitations are considered,
it appears that the median and mean removal efficiencies for dissolved organic
materials and nitrogen are consistently higher in the systems with activated
carbon added. Conventional activated sludge techniques are more effective in
the removal of suspended solids, oils, and grease. There are cases in the
literature where oil and grease removal by carbon enhanced systems exceeded
that of the conventional activated sludge process (IS). Application of these
removal efficiencies directly to synfuel plant wastewater treatment is not
possible due to the diversity of wastewater compositions and system configura-
tions represented by these data.
BIO-8
-------
TABLE B10-3. CONTROL TECHNOLOGY SUMMARY FOR ACTIVATED SLUDGE PROCESS (14)
Pollutant
Number of
diti pointi
Effluent Concentration Removal Efficiency. *
Minimum Ma i In mi Median Mean Minium Maxima Median Mean
O
I
Conventional Pollntanta.
•I/L:
Activated Sludge
BOD
COD*
TOC
TSS
Oil and greaie
Total phenol
Total Kjeldahl Nitrogen
Total Phoiphornt
Activated Slndfe with
Powdered Activated Carbon
92
84
14
74
7
31
6
27
5
45
33
6
<5
0.007
J7
0.14
4,640
7,420
1,700
4,050
303
500
322
46.8
49
425
280
92
25
0.028
174
3.46
184
890
427
283
70.8
18.7
174
6.7
17
0
8
0*
6
0
26
0
>99
96
95
96
>98
>99
63
97
91
67
69
25
92
64
44
27
86
63
63
34
74
60
43
32
BOD
COD*
TOC
TSS
Oil and frease
Total phenol
Total IjeldihJ Nitrogen
24
26
25
4
4
4
1
4
33
9
17
11
<0.010
54
563
387
83
57
0.058
13
98
38
54
13
0.013
17
160
67
52
23
<0.023
28
<90
60
64
o'
8
99
>99
98
97
96
96
>99
96
91
90
o'
54
>99
96
87
86
24
53
>99
96
Actual data indicate a negative removal
-------
Appendix BIO
Activated Slndge Process
Performance Models
Although the mechanisms by which pollutants are consumed in activated
sludge processes are very complex and not completely understood, numerous
activated sludge models have been developed to predict the performance of ac-
tivated sludge processes under different operating conditions. The degrada-
tion mechanism for each component in the wastewater will depend on the concen-
tration of the component and microorganisms, aeration time, and temperature.
To develop a model for each component in a wastewater with the diverse composi-
tion of a synfuel process wastewater would be neither practical nor economical.
Therefore, process performance models are generally based on gross parameters
such as COD or BOD. These parameters do not necessarily correlate well with
many of the fundamental component interactions. But they can be valuable in
predicting the overall performance of a system over a wide range of operating
conditions and stream compositions.
The most widely used activated sludge models have been developed by
McKinney and Eckenfelder (4) and by Lawrence and NcCarty (5). The Lawrence
and NcCarty model will be presented in part to demonstrate the key factors in-
volved in designing and predicting the performance of an activated sludge sys-
tem.
The Lawrence and McCarty (5) model assumes the following basic relation-
ships between the net cell growth rate (dX/dt), the rate of substrate utiliza-
tion, and the concentrations of substrate microorganisms in the reactor:
dX /dt - Y(dS/dt)-bX (1)
dS/dt - kSX /(K + S) (2)
A S
B10-10
-------
Appendix BIO
Activated Sludge Process
where: X = concentration of active cells, mg/L,
Y = yield, g VSS/g substrate,
S = substrate concentration around the microorganisms, mg/L,
k = maximum substrate utilization rate, day ,
K = half-velocity constant, i.e. substrate concentration at one
half the maximum substrate utilization rate, mg/L, and
b - cell decay rate, (day ).
These equations demonstrate the dynamic relationships between substrate utili-
zation (leading to a decrease in substrate concentration) and the growth of
active microorganisms. These kinetic parameters can also be used to define
the mean cell residence time (0 ) or sludge age which is an important design
c
and control parameter for the activated sludge process as shown below:
-L - Y*S -b
0c K + S (3)
s
and
S = (4)
6 0 (Yk-b)-l
c
where: S = effluent substrate concentration, mg/L.
6
Equation 4 indicates that the effluent substrate (BOD, COD, and other
te constituents) concentrations are a function o:
and the system kinetic parameters, Ks, b, Y, and k.
waste constituents) concentrations are a function of design parameters, 0 ,
C
The mean cell residence time controls design of all other aspects of the
activated sludge system. The critical value of c at which washout of or-
ganisms will occur, is called the minimum mean cell residence time (0 ). 0
0 c c
can be obtained by letting S approach infinity in equation 3:
9 " = Yk - b (5)
c
B10-11
-------
Appendix BIO
Activated Sludge Process
To ensure adequate waste treatment, activated sludge systems are usually de-
signed and operated with a sludge age ranging from 2 to 20 times the minimum
mean cell residence time (16).
One other important parameter in sizing an activated sludge system is the
hydraulic detention time, 0. This can be determined by balancing the sus-
pended solids around the system and the substitution of the kinetic parameters
as shown below:
0 =
-S
o e
(6)
x (i + be )
a c
The four kinetic parameters, Y, k, K , and b, used in the Lawrence and
S
McCarty model can be evaluated by laboratory studies of the subject waste-
water. Since samples of the actual wastewater will not exist before the acti-
vated sludge system is designed and operated, lab studies using effluents from
similar processes, pilot plants, or synthetically prepared materials can pro-
vide necessary information for initial designs. Kinetic and other parameters
from certain of these studies are summarized in Table B10-2. Summarized be-
low are the ranges and averages of the values reported in Table B10-2 for all
studies along with corresponding values for municipal wastewater treatment
systems.
Kinetic
Parameter
k
Y
K.
Range (Average) for
Listed Studies
0.16-1.06 (0.75)
0.10-0.56 (0.30)
0.002-0.038 (0.02)
Range (Typical) for
Municipal Systems (16)
2-10 (5)
0.25-0.4 (0.4)
15-70 (40)
0.04-0.075 (0.06)
Units
mg COD/mg MLVSS-day
mg MLVSS/mg COD
mg COD/L
day
-1
B10-12
-------
Appendix BIO
Activated Sludge Process
Limitations of the Model
The Lawrence and HcCarty model assumes that substrate utilization occurs
according to the Honod relationship. The Monod relationship defines specific
growth rate as dictated by the presence of a growth limiting substrate or
nutrient:
/ S \
(7)
_
where: n = specific growth rate, time ,
_i
LL = maximum specific growth rate, time ,
m
S = concentration of growth-limiting substrate in solution
mass/unit volume, and
K = half-velocity constant, substrate concentration at
s one-half the maximum growth rate, mass/unit volume.
For complex wastes and heterogeneous microorganism populations that exist
in the activated sludge process, the growth-limiting nutrient cannot be iden-
tified easily and gross parameters such as COD are less accurate substitutes
in the model.
The Lawrence and McCarty model also predicts that effluent substrate
concentration is independent of influent concentration. If sludge age is
maintained constant, an increase in influent concentration should stimulate an
increase in concentration of microorganisms in the reactor so that the efflu-
ent concentration is not independent of influent concentration (17) . Several
authors have suggested modifications to existing models to include the effect
of influent concentration (14,17). However, none of these modifications is
widely used at this time.
B10-12
-------
Appendix BIO
Activated Sludge Process
The model of Lawrence and HcCarty for activated sludge remains widely
used, usually giving representative results. Other models produce comparable
results. A limitation of all these models, including the Lawrence and McCarty
model, is an assumption of steady state conditions. In practice steady state
conditions are seldom maintained due to the changes in influent flow and
composition and the dynamics of the activated sludge system.
The performance of a powdered activated carbon enhanced activated sludge
system can also be predicted by the aforementioned models. The actual re-
movals obtained are a function of the mixed liquor powdered carbon concentra-
tion. Higher mixed liquor carbon levels generally provide higher removal
rates. The use of carbon in the reactor allows a greater concentration of
microbial mass to be maintained in the reactor with subsequent increases in
the sludge age. The longer sludge age allows a greater reduction in effluent
BOD concentration as shown in equation 4. The longer sludge age also enhances
the growth of nitrifying bacteria. The presence of these bacteria will in-
crease the removal of ammonia from the wastewater resulting in lower ammonia
residuals in the treated effluent. The buffering effect of the carbon in the
reactor with respect to shock loads will also help to ensure that reliable
performance is maintained.
Clarifier performance is not easily modeled. No clear relationships
exist between sludge characteristics and settleability. A frequent design
approach uses laboratory studies on samples of the actual sludge. However,
since no full scale activated sludge systems treating synthetic fuel plant
waters exist, no sludge samples are available for laboratory study. Settling
studies using sludge from pilot-scale tests do not always correlate well with
performance later realized in the full-scale system. Thus, even when appro-
priate samples of sludge do exist, settling is an uncertain enough phenomenon
that designs frequently rely heavily on empirical parameters drawn from past
experience with similar wastewaters. These parameters include overflow rates,
B10-14
-------
Appendix BIO
Activated Sludge Process
typically ranging from 16 to 32 mj/m*-day, and solids loadings, ranging from
3.0 to 6.0 kg/m»-hr (16).
Specific Pollutant Removal
During the activated sludge treatment process, both inorganic and organic
materials are removed from the wastewaters. Certain trace elements are re-
quired by the microorganisms to synthesize new cellular material. Other
compounds are adsorbed on the surface of the biological solids in the reactor
and are removed with the solids in the clarifier. These compounds are subse-
quently removed from the system with the wasted sludge. There are limited
data available on specific pollutant removals during activated sludge treat-
ment of synfuel process wastewaters. In order to estimate the pollutant
reductions that are possible in an activated sludge process, data from other
industries must be consulted. Tables B10-4 and B10-5 present available data
on specific compound removals for air activated sludge and carbon enhanced
activated sludge processes obtained in the treatment of various industrial
wastewaters (18). The industries represented by the data in these tables are
organic chemicals production, municipal wastewater, Pharmaceuticals, and
byproduct coking among others. The statistical analyses for compounds with
several data points should be indicative of removals that can be obtained.
However, application of these data to a specific synfuels process wastewater
would require evaluation of all wastewater, system design, and operating
characteristics of the system from which the data are derived. This evalua-
tion is essential in determining the applicability of performance data for a
proposed application.
4. Secondary Waste Generation
The major secondary waste stream from the activated sludge process is the
excess sludge produced as a byproduct of the biological degradation process.
B10-15
-------
TABLE B10-4.
REMOVAL OF TRACE AND TOXIC POLLUTANTS BY AIR ACTIVATED
SLUDGE PROCESSES (18)
O
Pollutant, ug/L
Antimony
Arsenic
CadaiuB
Chroniua
Copper
Cyanide
Lead
Mercury
Nickel
Selenium
Silver
Thalliua
Zinc
Bis(chloroaethyl) ether
Bis(2-chloroethyl) ether
4-Bronophenyl phenyl ether
Bis(2-ethylhexyl) phthalate
Butyl benzyl phthalate
Di-n-butyl phthalate
Diethyl phthalate
Dimethyl phthalate
Di-n-octyl phthalate
Benzidine
1 , 2-Diphenylhydracine
N-nitrosodiphenylanine
N-nitroso-di-n-propylaMine
2-Chlorophenol
2,4-Dichlorophenol
2 , 4-Dime thy Iphenol
2-Nitrophenol
4-Nitrophenol
Pentachlorophenol
Number of
data point*
18
8
17
34
37
24
26
9
32
1
17
1
36
1
1
1
38
1
9
17
9
1
1
1
2
2
2
2
3
1
1
IS
Effluent concentration
Minima
0.3
<5
<0.5
<0.2
<0.2
<1
0.6
<0.5
4
<5
48
<0.04
<0.02
<0.03
<0.03
<0.07
2
0.9
<10
8
<0.4
Maximum
670
160
13
20.000
170
38,000
160
1.6
400
95
150.000
1.300
58
69
200
1.6
19
10
<10
<10
3.100
Median
3.5
13
4
28
30
28
61
0.7
40
33
180
13
3.6
<0.03
<0.03
9
<0.4
Mean
46
48
4
910
43
2.000
38
<0.98
78
41
32
29
5,800
^10
^10
18
64
11
9.2
6.6
24
5.000
4
340
0.84
11
5.5
<10
<27
O.4
<0.9
1.600
Removal efficiency.
H in in UB
0*
0
0*
o*
o'
o'
o*
o"
0*
0*
0*
0*
o'
oa
69
0
>o.
0*
0*
Maximum
90
96
>99
99
>99
>90
99
>97
92
>96
92
>99
>99
>99
>99
92
>50
>95
>99
Median
15
39
0
48
56
o'
50
>29
7
20
27
>84
>99
>99
o"
>98
%
Mean
30
>43
>31
45
52
>18
>49
>31
>29
o*
31
38
35
>83
>47
95
0
>60
>56
>60
0
0
0*
>84
0
46
>25
>32
>99
>99
>70
(continued)
-------
TABLE B10-4. (Continued)
o
i
Pollutant, ug/L
Phenol
2,4. 6-Trichlorophenol
g-Chloro-B-cresol
Benzene
Chlorobenzene
1 , 2-Dichlorobenzene
1,4-Dichlorobenzene
2 , 6-Dinitrotoluene
Bthylbenzenc
Hexachlorobenzene
Toluene
1 , 2 ,4-Tr ichlorobenzene
Acenaphthene
Anthracene/Phenanthrene
Fluoranthene
Fluorene
Indeno( 1 , 2 , 3-cd)pyrene
Naphthalene
Pyrene
2-Chloronaphthalene
BroBofom
Carbon tetrachloride
Chloroform
Dichlorobronone thane
1 , 1-Dichloroethane
1 , 2-Dichloropropane
1,3-Dichloropropane
Me thylene chloride
1.1.2. 2-Tetrachloroe thane
Te trachloroe thylene
1,1,1-Trichloroe thane
1.1,2-Trichloroe thane
Trichloroe thylene
Trichlorofluor one thane
Heptachlor
Isophorone
MuBber of
data points
30
10
4
9
6
12
8
1
24
4
31
11
10
7
1
2
1
26
5
1
1
2
16
2
2
2
1
5
2
11
6
1
12
5
1
2
Effluent concentration
HiniauB
<0.07
<0.2
<0.2
<0.2
47
0*
0*
0*
0*
»
0*
•
0*
98
0*
o"
>0
>53
ft
0*
>0a
ft
0*
•
0*
>0
Maxiaua
>99
98
>98
>99
>99
>99
>99
>99
>97
>99
>99
>99
>98
>99
78
>99
>99
>0
>18
>82
99
>44
>99
>99
>99
96
>0
i Median
99
0
>49
>96
>99
>99 .
95
>95
>49
18
>99
>99
83
>99
0*
>2
».
0*
ft
0*
>85
>98
0*
Mean
>77
>36
>65
>60
>67
>74
>82«
0
>83
>47
>52
>67
>84
>60
0
>99
>99
>64
16
50
0
>99
>61
0
>9
>68
0*
34
>22
>75
>74
>9
>68
19
76
>0
Note: Blank* indicate data not applicable.
"Actual data indicate negative removal.
-------
TABLE B10-5.
REMOVAL OF TOXIC POLLUTANTS BY POWERED ACTIVATED CARBON
ENHANCED ACTIVATED SLUDGE PROCESSES (18)
O
I
Pollutant, (ij/L
AntiBony
Cadaina
ChroBinB
ChroaiiuB (+6)
Copper
Cyanide
Lead
Mercury
Nickel
Selenium
Zinc
Bit (2-chloroethyl)
ether
Bi«(2-ethylheiyl)
ph thai ate
2-Chlorophenol
Phenol0
Benzene
Ethylbenzene
Toluene0
Naphthalene
1 ,2-Dichloroe thane
1 ,2-Dichloropropane
Acrolein
laophorone
NoBber of
Data Points
2
1
4
3
3
3
2
1
3
2
4
1
1
3
6
4
3
3
1
1
1
1
1
Effluent Concentration
Miniaua
41
24
<20
7
<20
<18
<10
<20
78
0
0.001
0.00|
0°
<0.002
Maxiaoa
ISO
90
20
29
45
38
22
40
140
190,000
190,000
21,000
18,000
67 ,000
Median Mean
96
10
S3 55
<20 <20
14 17
20 <28
28
0.6
<10 <14
<30
95 100
44
64
96
69
>78
>S8
13
98
100
>99
>99
100
>99
Median
88
>60
61
>67
>0
38
95
94
99
95
99
Mean
"V
0*
86
41
52
>62
39
o*
19
6
44
53
>97
92
>83
95
93
93
>96
81
93
30
97
Actual data indicate negative reaoval.
Bel on detection liaiit, tainted to be <10 |ig/L.
'Additional data added froai Keference 15.
-------
Appendix BIO
Activated Sludge Process
The quantity of sludge generated is related to the parameters discussed pre-
viously. The sludge waste stream is concentrated to between 0.5 to 1.2 wt %
solids in the clarifier (16). Thickening of the sludge can increase the
solids concentration to 2.5-3.3 wt % (16). Extensive conditioning and further
dewatering are required to significantly reduce the volume of the sludge
further.
When powdered activated carbon is used in the activated sludge process,
the volume of the solid waste stream from this process can be reduced substan-
tially. This reduction will occur when the spent carbon in the wasted sludge
is regenerated in a multiple hearth furnace or by the wet air oxidation pro-
cess. During the regeneration process, most of the biological sludge will be
oxidized to carbon dioxide and water. The volume of ash that results from
this regeneration is extremely small. It is likely that most of this material
can be disposed of with other ashes generated in the subject facilities. Any
soluble BOD remaining after wet air oxidation regeneration can be recycled to
the activated sludge reactor. On a large system, makeup carbon costs can be
substantial. The regeneration of the carbon on-site, while expensive, can
significantly reduce makeup carbon requirements and eliminate the need for
complex sludge handling and solid waste disposal facilities.
Another potential waste stream generated by an activated sludge process
is the emission of volatile species from the surface of the reactor. These
emissions would be of concern where there are significant quantities of vola-
tile organic and inorganic species in the wastewater. Where there is a high
concentration of volatiles in the wastewater, pretreatment for their removal
might be desirable. Volatile components can also be generated in the acti-
vated sludge process itself, as a result of incomplete oxidation of wastewater
contaminants. The extent of these emissions has not been quantified to date.
BIO-19
-------
Appendix BIO
Activated Sludge Process
5. Process Reliability
No full-scale experience has been gained in treating synfuels process
wastewaters with activated sludge processes in this country. However, there
is much experience using activated sludge to treat a wide variety of indus-
trial and municipal wastewaters. This experience has been used to estimate
the reliability of the activated sludge process.
Most activated sludge system operating problems evolve from sudden varia-
tions in the waste strength or flow or from settling problems in the clari-
fier. Many substances normally toxic to microorganisms can be degraded if the
microorganisms are properly acclimated. Sudden changes in the concentration
of toxic substances in the influent can debilitate or destroy the microbial
population. Even sudden changes in the concentration or nature of highly
degradable material can have a detrimental affect on the microorganisms.
Many activated sludge systems are preceded by equalization basins designed to
dampen sudden changes thus reducing their impact on system operation.
Various nutrients are needed by the activated sludge microorganisms to
encourage their growth via consumption of the organics in a wastewater stream.
Many of these nutrients will be present in the wastewater while others must be
added to the system. The principal nutrients required are nitrogen and phos-
phorus. The concentration of other nutrients required by the microorganisms
are relatively insignificant compared to the nitrogen and phosphorus require-
ments (16). Synfuels plant wastewaters will typically provide adequate nitro-
gen in the form of ammonia. Phosphorus, on the other hand, may not be present
in these wastewaters and might have to be added to the system. Adequate
nutrient concentrations are essential for reliable performance and maintenance
of the proper microorganism population.
BIO-20
-------
Appendix BIO
Activated Sludge Process
Host activated sludge systems consist of at least two reactors and clari-
fier trains in parallel. This arrangement permits one train to be removed
from service for maintenance while maintaining complete or partial treatment
capability. Downtime of a train can be minimized by proper maintenance of
equipment, particularly pumping and aeration equipment, and by monitoring the
operation of major equipment components and the overall plant. Monitoring of
the process by well trained operators is also required to ensure that satis-
factory performance is maintained.
Maintenance of the activated sludge system effluent quality is often
closely related to the performance of the clarifiers. This aspect of the
activated sludge process is frequently the most difficult to control. Even
subtle environmental changes in the reactor resulting from changes in the
influent characteristics can affect sludge settleability and result in solids
carryover in the effluent. Other conditions typically responsible for poor
clarifier performance include the following.
• If anaerobic conditions develop in the clarifier, nitrates can
be reduced to nitrogen gas. The rising bubbles of nitrogen
cause the sludge to rise instead of settle out of the waste-
water.
• Sludge age correlates somewhat with settleability in the clari-
fier. When sludge ages are too low, the sludge will not floc-
culate properly, resulting in poor settling characteristics.
When sludge ages are high, conditions favoring the growth of
undesirable organisms can be established. Predators including
worms and protozoa can thrive in these environments and break
up settling floes in the clarifier.
• A condition known as sludge bulking can inhibit sludge settle-
ability. Bulking results from the growth of filamentous
organisms in the reactor which have poor settling characteris-
tics. Reactor conditions responsible for bulking sludge
include low nutrient concentrations (including nitrogen, phos-
phorus, and iron), low pH, and low oxygen transfer.
B10-21
-------
Appendix BIO
Activated Sludge Process
Sometimes the cause of sludge bulking or other settling problems can be
identified and subsequently corrected. However, in practice, the exact causes
of observed settling problems are usually difficult to determine. Temporary
measures can be taken including the use of coagulants in the clarifier. How-
ever, these remedial measures are usually expensive and should only be used
until the source of the problem can be identified and corrected.
The use of powdered activated carbon in the activated sludge process re-
sults in a carbon weighted sludge with better settling characteristics than
sludge from a conventional treatment process. These characteristics will al-
low higher clarifier solids loading rates while maintaining an acceptable ef-
fluent suspended solids level (19). Enhanced sludge settling characteristics
derived from the use of powdered carbon will contribute to the maintenance of
the treated effluent quality.
6. Process Economics
Figure B10-2 shows the installed equipment costs for the activated
sludge process. The lower curve is based on conceptual estimates developed by
EPA using their own publications, literature values, construction grant files,
and manufacturers information (20). All values were originally indexed to
September 1979 but have been escalated to first quarter 1980 by the CE cost
index ratio of 1.06. The curve includes costs of the aeration system only;
the costs of solids handling and dewatering equipment are not included in
these estimates.
Figure B10-2 also presents the data points from which a cost curve spe-
cific to synfuels wastewaters was developed. These costs are for complete
systems including sludge dewatering, clarifiers, recycle pumping as well as
aeration basin construction and aeration equipment (2,21,22). Installed
B10-22
-------
1000
First Quarter 1980 $
100
7
otnplete System fo
Synfuels Waste1
ater (2,21,22)
Aeration Only (20)
10
I I I I I I
J L
10J
7 10" 2 47 10s
Wastewater BOD Load, kg/day
Figure B10-2. Installed equipment costs of air activated
sludge systems
310-23
-------
Appendix BIO
Activated Sludge Process
equipment costs for activated sludge treatment of vastewaters from synthetic
fuels plants are expected to be higher than for activated sludge systems
treating other wastewaters. These higher costs result primarily from the use
of multiple stage, as opposed to single stage, treatment systems and overde-
sign to provide adequate capacity to treat variable organic loads in the
wastewater.
Figure B10-3 presents the annual operating costs for power, labor, and
materials required to operate an air activated sludge system. These costs
correspond to the installed cost curve presented in Figure B10-2 for an
aeration basin only. Table BIO-6 presents the bases for the computation of
power requirements for a system including aeration basin mixing, by recycle
pumping, aeration, and sludge dewatering. Supplemental addition of nitrogen
and phosphorus were computed using a BOD: nitrogen: phosphorus ratio of
100:5:1 (23). These nutrient additions are essential to maintain a viable
bacterial population. The operating labor requirements for the system were
estimated to be 1.5 man years. Operating labor, supervision, and maintenance
labor were calculated as a percentage of operating labor and total installed
equipment cost, respectively.
Cost data on powdered activated carbon enhanced activated sludge systems
are not readily available for industrial wastewater applications. However,
the limited data available indicate that these systems have lower capital and
annual costs than comparable conventional activated sludge systems in synfuels
process applications (19). Therefore, the cost curves in Figure B10-2 and
B10-3 should represent conservative estimates for the purpose of comparisons
with other technologies.
B10-24
-------
100
Q
o
a 4
First Quarter 1980 $
en
10
• Electric:
• Labor
=0
c
H 7
OJ
c.
o
•Material:
I I I I I I i
I I i I I I
10'
7 10" 2 47
Wastewater BOD Load, kg/day
Figure B10-3. Operating costs for air activated sludge systems-
aeration only (20)
B10-25
-------
Appendix BIO
Activated Sludge Process
TABLE B10-6. BASES FOR ESTIMATING ACTIVATED SLUDGE UTILITY REQUIREMENTS
Component
Estimated Basis
Reference
Pumping
Aeration basin mixing
Aeration
Sludge dewatering
12 kWh/1000 m*
30 kWh/1000 m1
0.42 kWh/kg BOD destroyed
58 kWh/100 kg dry sludge
23
15
15
24
7. References
1. Adams, A.D. Powdered Carbon: Is It Really That Good? Water and Waste
Engineering, March 1974.
2. U.S. Environmental Protection Agency. Coal Conversion Control
Technology, Volume I. Environmental Regulations; Liquid Effluents,
Report EPA-600/7-79-228a, 1979.
3. Ellis, K.V."The Biological Treatment of Organic Industrial Wastewaters".
Effluent and Water Treatment Journal, March 1979.
4. Goodman, B. and A.J. Englande. "A Unified Approach to Activated Sludge
Design", Journal of the Water Pollution Control Federation, 46, 312 1974.
5. Lawrence, Alonzo W. and Perry L. McCarty. *A Unified Basis for
Biological Treatment Design and Operation", Journal of the Sanitary
Engineering Division, ASCE, 96, SA3, 757, 1970.
6. Luthy, R.G. and L.D. Jones. Biological Oxidation of Coke Plant
Effluents. J. Environmental Engineering Division, ASCE, Vol. 106, No.
EE4, August 1980.
7. Luthy, R.G., et al. Biological Treatment of a Synthetic Fuel Wastewater,
J. Environmental Engineering Division, ASCE, Vol. 106, No. EE3, June
1980.
B10-26
-------
Appendix BIO
Activated Sludge Process
8. Luthy, R.G. and J.T. Tallon. Biological Treatment of a Coal Gasification
Process Wastewater, Water Research, 1980.
9. Johnson, G.E., et al. Treatability Studies of Condensate Water from
Synthane Coal Gasification. Pittsburgh Energy Technology Center Report
No. PERC/RI-77/13, Pittsburgh, Pennsylvania, November 1977.
10. Sack, W.A. Biological Treatability of Gasifier Wastewater. Morgantown
Energy Technology Center Report No. METC/CR-79/24, Morgantown, West
Virginia, June 1979.
11. Reap, E.J., et al. Wastewater Characteristics and Treatment Technology
for Liquefaction of Coal Dsing the B-Coal Process. Proceedings of the
32nd Purdue Industrial Waste Conference, Ann Arbor Science, Ann Arbor,
Michigan, 1977.
12. Singer, P.C., et al. Assessment of Coal Conversion Wastewaters:
Characterization and Preliminary Biotreatability. EPA-600/7-78-181, U.S.
Environmental Protection Agency, Washington, D.C., 1978.
13. Drummond. C.J., et al. Treatment of Solvent Refined Coal (SRC-I)
Wastewater: A Laboratory Evaluation. Pittsburgh Energy Technology
Center, Pittsburgh, Pennsylvania, 1981.
14. Sykes, Robert H. Microbial Product Formation and Variable Yield,
Journal of the Water Pollution Control Federation, 48, 8, 1976.
15. Removal of Priority Pollutants PACT /WET AIR Regeneration System, Zimpro,
Inc.
16. Metcalf and Eddy, Inc. Wastewater Engineering: Treatment, Disposal,
Reuse, New York: McGraw-Hill, 1979.
17. Grady. C.P.L. and D.R. Williams. Effects of Influent Substrate
Concentration on the Kinetics of Natural Microbial Populations in
Continuous Culture, Water Research, 9, 171, 1975.
18. U.S. Environmental Protection Agency. Treatability Manual, Volume III
Technologies for Control/Removal of Pollutants, EPA-600/8-80-042c, 1980.
19. Castaldi, F.J. Application of Combined Powdered Carbon/Activated Sludge
Treatment to Lurgi Process Coal Gasification Wastewaters.
20. D.S. Environmental Protection Agency. Treatability Manual, Volume IV
Cost Estimating, EPA-600/8-80-042d, 1980.
B10-27
-------
Appendix BIO
Activated Sludge Process
21. U.S. Environmental Protection Agency. Water Conservation and Pollution
Control in Coal Conversion Processes, EPA 600/7-77-065, 1977.
22. U.S. Environmental Protection Agency. Water Treatment in Coal
Conversion. EPA-600/17-79-133, 1979.
23. Adams, C.E. , Jr. Treatment of High Strength Phenolic and Ammonia Waste
Stream by Single and Multi-Stage Activated Sludge Process. Purdue Univ.
Eng. Bull. Eng. Ext. Ser. 145. (II): 617-630. 1974.
24. U.S. Environmental Protection Agency. Process Design Manual for Sludge
Treatment and Disposal, EPA 625/1-74-006, October 1974.
B10-28
-------
APPENDIX Bll
TRICKLING FILTERS
1. Process Description
A trickling filter is a fixed-film biological treatment unit in which
wastewater is percolated or trickled over a bed of a highly permeable media to
which a biological slime is attached. At the surface of the slime layer, both
oxygen and organic matter are adsorbed and the organic material is degraded by
aerobic microorganisms. To ensure a healthy slime layer, good air circulation
throughout the filter is required. As the microorganisms grow, the slime lay-
er increases in thickness to the point that diffusion of oxygen and nutrients
to the media surface become rate limiting steps. As this occurs, the micro-
organisms enter a phase of endogenous growth and loose their ability to cling
to the media surface. The shear force of the wastewater washes off the slime
and a new layer begins to grow (1). This "sloughing" of the slime layer is
primarily a function of hydraulic and organic loading considerations. Control
of these parameters is usually provided by recirculating a portion of the
treated effluent. This recirculation also ensures continued wetting of the
media and serves to dilute high strength wastewaters. If desired, sloughed
off slime can be separated from the effluent in a downstream clarifier.
Early trickling filters used primarily porous rock or similar material.
However, the "state of the art" now favors the use of synthetic filter media,
particularly for industrial wastewater treatment. Plastic media can be
designed to provide better oxygen transfer and a much greater specific surface
area than naturally occurring materials and, therefore, greater removals per
unit volume of reactor. Additionally, the much lighter plastic media can be
contained in structures that are less expensive to construct.
Bll-1
-------
Appendix Bll
Trickling Filters
Trickling filters are classified as low, high, or super rate depending on
hydraulic and organic loading considerations. Low rate filters (usually rock
media) can handle hydraulic loadings up to 3.7 m'/day-m* and organic loadings
of 0.16-0.32 BOD kg/day-ml of filter media. The corresponding hydraulic and
organic loadings for high rate filters are 9.4 to 28 m'/day-m* and 1.4 kg
BOD/day-m* of media. Super rate trickling filters with synthetic media are
capable of handling hydraulic loadings of 140 m'/day-m* and can treat high
strength industrial wastewaters (2).
2. Process Applicability
In treating high strength wastewaters associated with synthetic fuels
facilities, the main use for trickling filters will probably be as a pretreat-
ment or roughing unit in conjunction with other biological processes. The
combined use of trickling filters and activated sludge units has been applied
in treating wastewaters from meat processing, petrochemical, and organic
chemical manufacturing (3,4).
One major advantage of the trickling filter is its ability to handle a
wide range of concentrations for any waste stream determined to be biodegrad-
able. Varying the recycle rates to dilute the influent stream can extend the
range of treatable waste strengths. There is also an economic advantage in
using the trickling filter prior to activated sludge treatment. The BOD re-
duction through the trickling filter translates to a reduced volume and 0,
requirement in the activated sludge unit and, therefore, provides cost savings
(6).
Trickling filter media are also well suited for use as a roughing treat-
ment for high temperature wastewaters (2,7). Activated sludge units are
Bll-2
-------
Appendix Bll
Trickling Filters
sensitive to high temperature. The increased surface area associated with the
trickling filter, particularly synthetic media filters, facilitates heat
transfer and substantial cooling.
A more recent application of trickling filters has been in the area of
biological nitrification. Reported results indicate plastic media filters are
capable of achieving consistently high levels of nitrification when operating
on a low BOD waste stream (8). Thus, trickling filters may be applicable to
the treatment of synfuel wastewaters both as a pretreatment process and as a
followup to an activated sludge unit.
3. Process Performance
Typically, tricking filters will remove from 60-85 percent of the
influent BOD from refinery wastewaters. With a high degree of treatment,
above 90 percent BOD removal has been obtained in phenolic wastewaters (2,5).
However, with high organic loadings, the fraction of BOD removed can be as low
as 40 percent (6). The use of trickling filters on wastewaters as concen-
trated as those from some synfuel facilities has not been reported and must be
regarded as experimental.
There are many variables which affect trickling filter performance in-
cluding wastewater characteristics, media depth and type, hydraulic and or-
ganic loading, and temperature. Eckenfelder has suggested the following per-
formance model assuming plug flow through the filter and first order reaction
kinetics (9):
S /S = e "^v* (1)
e o
where: S = effluent concentration of soluble substrate, mg/L,
e
S = influent concentration of soluble substrate, mg/L,
Bll-3
-------
Appendix Bll
Trickling Filters
K = organic removal velocity constant, ft/day,
AV = specific surface of the medium, sq ft/cu ft, and
t = residence time in the filter, day.
Eckenfelder modifies this equation by assuming that flow velocity is indepen-
dent of depth and redefining time as a flow per unit cross section area filter
media:
t - C (D/qn) (2)
where: C,n = constants which characterize the media,
D = media depth, ft, and
q = hydraulic surface loading, gpm/sq ft.
Equation 1 is further modified to yield
-K A (D/q°)
Se/So - . * * (3)
where: Kx = K (C) = modified constant.
The constant, K, varies with temperature and a correction must be applied
when predicting performance. The traditional correction expression is:
K, = K. 0(T* " T*
(4)
where: K4 = organic removal velocity constant at temperature Tx
K, = organic removal velocity constant at temperature Ta,
Ti'T* " water temperature, K, and
0 = temperature coefficient.
Bll-4
-------
Appendix Bll
Trickling Filters
The temperature coefficient, 9, has been reported to range from 1.035 to 1.07
with the higher values more applicable at higher organic loadings (8).
K, the treatability rate constant, is effectively a measure of the amena-
bility of the wastewater to biological treatment; the value of n is determined
from media characteristics. Experience indicates that K generally lies in the
range of 0.01 - 0.045 for wastewaters similar to those expected in synthetic
fuels facilities. The factor n is generally close to 0.5.
Eckenfelder proposes a graphical approach to determining the treatability
factor, K (9). If equation 3 is rewritten as:
In (Se/SQ) = KjA^D/q11) (5)
a semi-logarithmic plot of S /S vs A (D/qfl) should yield a straight line
with a slope of Kx. A summary of K values for various wastewaters is pre-
sented in Table Bll-1.
TABLE Bll-1. SUMMARY OF K VALUES FOR VARIOUS WASTEWATERS (9)
Type of Wastewater
Domestic
Fruit canning
Boxboard
Steel coke plant
Textile
Pharmaceutical
Slaughterhouse
LOB K
0.079
0.0177
0.0197
0.0211
0.0273
0.0292
0.0246
n
0.5
0.5
0.5
0.5
0.5
0.5
0.5
*In all cases, media used was Surpak having a specific surface area
of 91.9 m»/m».
Bll-5
-------
Appendix Bll
Trickling Filters
4. Secondary Waste Generation
There are two secondary waste streams associated with the operation of a
trickling filter. The first is the air stream which continually flows across
the bed (normally countercurrent to the liquid flow). This stream will con-
tain volatile gases stripped from the wastewater and volatile products of the
biological degradation process. The second waste stream is the excess slime
growth which must be continuously removed from the reactor. The slime can be
separated from the effluent in a conventional clarifier. Because the slime
from a trickling filter will normally settle quite well, small heavily loaded
clarifiers can be used. However, when a trickling filter is used as a rough-
ing unit prior to conventional activated sludge treatment, this unit may be
eliminated (i.e., the slime can pass to the activated sludge unit where it
will eventually be removed).
5. Process Reliability
Trickling filters are highly reliable treatment processes due to their
mechanical simplicity and high concentrations of microorganisms that provide
resistance to shock loadings. Further resistance to shock loads is possible
with careful control of the recycle rate used to dilute the influent. Even
when the influent needs no dilution, some recycle may be necessary to provide
the minimum hydraulic loading needed to ensure sloughing of the biofilm.
Most fixed-film reactors suffer reliability problems during cold
weather. The high surface area needed to maximize contact opportunity for
diffusion also facilitates heat transfer with the surrounding environment.
Decreased temperatures will decrease the rate of microbial activity. Benzie
et al. (8) report a decrease in efficiency of about 20 percent between warm
weather (294 K) and cold weather (272 K) conditions.
Bll-6
-------
Appendix Bll
Trickling Filters
6. Process Economics
Construction and operating and maintenance costs for a trickling filter
using plastic media were obtained from Reference 10. The construction costs
were converted to installed equipment costs by adding 28 percent for piping,
electrical, instrumentation, and site preparation. They were escalated from
1978 to first quarter 1980 dollars by using the CE index ratio of 1.15.
Figure Bll-1 presents installed equipment costs as a function of wastewater
flow rate for the system design assumed in the reference. Figure Bll-2 shows
how labor and maintenance materials costs vary with flow rate. The costs are
for carbonaceous BOD oxidation (vs nitrification) and are based on an organic
loading of 1.32 kg BOD/day-m3 and a recirculation ratio of 2:1 (dilution of 1
part of feed with 2 parts of effluent). A summary of additional design
assumptions is presented in Table Bll-2. The organic loading assumed in the
reference is lower than that which would be expected for use of a trickling
filter as a roughing unit in a synfuel facility. The wastewater flow used to
enter these curves should be increased proportionally to the increase in
organic loading over 0.32 kg BODj/day-m*. Additionally, wastewater flow rates
used with the curves will need to be increased if recirculation ratios higher
than 2:1 are used.
It should be noted that costs associated with clarifiers and recircula-
tion equipment are not included with these figures. When used as a roughing
unit, trickling filters may or may not require a secondary clarifier. Also,
the 0 fi M cost curves do not include an energy line. Pumping energy require-
ments may be approximated by using the following equation assuming a wire-to-
water efficiency of 60 percent (10):
kWh/yr = 2.2 x 10* x mVday x discharge head (m) (6)
Bll-7
-------
7
4
2
^10"
"s
~ 7
LJ
VI
u 4
tqui pinent
ro
-o
Sio3
c
4
2
102
-
t 1 1 ! t I 1 1
1 I t 1 1 I 1 I
First Quarter 1980 S
I i i i i i i i
10 2 4 7 100 2 4, 7 1000 2 47
Wastewater Flow Rate, mj/hr
Figure Bll-1. Installed equipment cost for trickling filters (10)
Bll-8
-------
Operating Costs, 5/1000 m3
Tl
H-
TO
i-i
ro
i
to
O
T3
ft)
rr
H-
3
cro
o
o
CQ
O
1-1
f-i
H-
O
?-T*
I—'
H-
00
rt
fD
i-i
en
-------
Appendix Bll
Trickling Filters
If the trickling filter is used as a single stage nitrification unit as
opposed to a carbonaceous BOD oxidation unit, construction costs are expected
to be 23 to 25 percent higher and 0 & M costs will increase 12 to 15 percent.
TABLE Bll-2. DESIGN ASSUMPTIONS INHERENT IN COST CURVES (10)
Bed Depth - 0.73 m
Hydraulic Loading = 0.026 m»/m»-min
Recirculation Ratio = 2:1, recycle to influent
Organic Loading = 0.32 kg BOD4/ms day
Construction Costs Include Allowances for: plastic media,
underdrains,
distributors, and
tower containment structures
Labor Including Fringes = ill/hr
7. References
1. Metcalf and Eddy, Inc. Wastewater Engineering: Treatment, Disposal
Reuse. McGraw Hill Book Co., New York, 1979.
2. Liptak, Bel a G. Environmental Engineers Handbook, Vol. 1: Water
Pollution. Chelton Book Co., Radnar, PA, 1974.
3. Bryan, E.H. Two-Stage Biological Treatment Industrial Experience.
Proceedings of the Eleventh Southern Municipal and Industrial Waste
Conference, North Carolina State University, 1962.
4. Smith, R.M. Some Systems for the Biological Oxidation of Phenol Bearing
Wastewaters. Biotechnology and Bioengineering, 5, 1963.
5. Bush, Kenneth E. Refinery Wastewater Treatment and Reuse. Chemical
Engineering, 83(8): 113-118, 1976.
Bll-10
-------
Appendix Bll
Trickling Filters
6. D.S. Environmental Protection Agency. Wastewater Treatment in Coal
Conversion, EPA-600/7-79-133, June 1979.
7. Water Purification Associates. Conceptual Designs for Water Treatment in
Demonstration Plants. Office of Fossil Energy, Division of Coal
Conversion, D.S. Department of Energy, Washington. DC, March, 1979.
8. Adams, Carl E., Davis L. Ford and W. Wesley Eckenfelder. Development of
Design and Operational Criteria for Wastewater Treatment. CBI Publishing
Co.. Inc., Boston. MA, 1981.
9. Eckenfelder. W.W.. Jr., Water Quality Engineering for Practicing
Engineers. New York: Barnes and Noble, 1970.
10. U.S. Environmental Protection Agency. Innovative and Alternative
Technology Assessment Manual. EPA-430/9-78-009, Office of Water
Program Operations, Washington, DC, February 1980.
Bll-11
-------
-------
APPENDIX B12
ROTATING BIOLOGICAL CONTACTORS
1. Process Description
Rotating biological contactors (RBCs) are a form of fixed film biologi-
cal treatment similar to trickling filters. In RBCs however, the slime layer
of microorganisms grows on the surfaces of disks, 1.8-3.7 m in diameter, which
are usually constructed of polystyrene or polyvinyl chloride (1). In aerobic
treatment applications, the disks are attached to a horizontal shaft and
slowly rotated through the wastewater approximately 40 percent submerged (2).
For anaerobic or denitrification treatment applications the shaft assemblies
are totally submerged in the wastewater. In either case, the entire system
would be contained in a treatment tank. Rotation rates are generally 1.3-3.0
rpm (3). A complete RBC system usually consists of two or more trains of
disks with each train consisting of multiple stages.
The process of organic adsorption in an RBC is identical to that describ-
ed for trickling filters. The rotation of the disks provides for alternating
contact with the organic material in the wastewater and with the atmosphere
for oxygen adsorption. Rotation oxygenates not only the slime layer and ad-
herent water but also the contents of the tank. Shearing forces created by
the rotation are sufficient to provide for sloughing of excess growth. Pro-
duct solids can subsequently be separated from the effluent in a clarifier.
2. Process Applicability
Like trickling filters, RBCs would generally be used either as a front
end roughing unit or as a second stage (denitrification) unit in the biologi-
cal treatment of synfuels process wastewaters. However, in some applications
it may be used as the principal treatment process. RBCs have not been used
commercially to treat synfuels process wastewaters but they have been
installed recently in several U.S. refineries (2).
B12-1
-------
Appendix B-12
Rotating Biological Contactors
RBCs have several inherent advantages over other aerobic biological
treatment methods, such as resistance to upsets, low energy needs, compact-
ness, modular construction, easy staging, and weather protection. The holding
tank in each RBC stage provides a diluting and surge-absorbing reservoir which
contributes to its stability. Host other fixed film bio-reactors do not have
a comparable reservoir. Weather protection, provided by a cover, not only en-
hances low temperature operations and prevents ice damage but also collects
aerosols and mists generated by splashing and permits easy ventilation control
(2). Additionally, the concentration of active microorganisms on the disk is
very high, permitting the system to handle more concentrated wastewaters and
to better tolerate shock loads than some other types of systems.
3. Process Performance
As stated previously, KBCs have been installed and evaluated at several
U.S. refineries. Comparison of data generated at these facilities with
McAiley's design equation indicates that equation to be useful in predicting
BOD removal. McAiley's method assumes that BOD removal is proportional to BOD
concentration in each stage and that each stage acts as a back-mixed reactor
(2):
R - M/A (BOD - BOD ) (1)
influent effluent
= P (BOD )/(K + BOD ... )
effluent effluent
where: R = specific BOD removal rate per unit area of medium,
kg/(day-m2),
BOD = BOD, mg/L,
M = wastewater flow,, million kg/day,
A = area of an RBC stage, ma,
P = a constant = maximum value of R at infinite BOD, and
K = a constant = BOD at which R = P/2.
B12-2
-------
Appendix B12
Rotating Biological Contactors
A study performed by Hormel for Vancouver B.C. tested RBC and activated
sludge processes with three wastes: oily, high soluble organic, and phenol
bearing (4). Both units reduced the oily waste BOD level (from 47 and 375
mg/L in the influent to 15 and 31 mg/L in the effluent). Activated sludge was
more efficient than the BBC in removing the soluble organics and both units
removed over 99 percent of the phenol in a stream containing 100 mg/L.
BOD removals of 0.01-0.015 kg/day-m* and COD removals of 0.005-0.02
kg/day—m2 have been reported by refineries using RBC treatment. These same
refineries reported an increase in removal efficiencies with an increase in
influent concentration. One refinery, however, reported that high influent
hydraulic and organic loadings resulted in an increase in COD in the effluent
even though soluble BOD removal remained high. Additionally, high hydraulic
loadings were found to inhibit nitrification; whereas, at low loadings, nitri-
fication resulted in up to 99 percent ammonia removal (2).
4. Secondary Waste Generation
Like the trickling filter, the only secondary wastes from an RBC are the
volatile gases released from the surface of the wastewater and slime layer and
the underflow solids from the downstream clarifier. When an RBC is used as a
roughing unit prior to activated sludge treatment, this latter waste stream
might be eliminated, i.e., the excess slime would pass on to the activated
sludge unit where it would eventually be removed.
5. Process Reliability
RBC units are considered to be moderately reliable in the absence of high
organic loadings and extreme cold temperatures. They are subject to limita-
tions involving mechanical considerations, shock loadings, toxic substances.
B12-3
-------
Appendix B12
Rotating Biological Contactors
and oil. Mechanical problems can arise from weight shifts during sloughing or
if the level in the treatment tank drops too low.
6. Process Economics
Estimated construction and operating and maintenance costs for RBCs were
obtained from Reference 5 for carbonaceous oxidation with a hydraulic loading
of 0.04 ms/day-ma. Construction costs include RBC shafts constructed of
standard media with 9300 m»/shaft, motor drives of 3.7 kW, molded fiberglass
covers, and reinforced concrete basins. The costs associated with a secondary
clarifier are not included; however, these units will not be required for all
RBCs as discussed previously. The impacts of variable organic loadings on
these costs were not specified in any of the available references. The costs
shown are based on treating wastewater with an inlet BODS of approximately 150
mg/L. The construction costs were converted to installed equipment costs by
adding 28 percent for piping, electrical, instrumentation, and site prepara-
tion. Figure B12-1 shows these costs as a function of wastewater flow rate
updated to first quarter 1980 dollars from September 1976 dollars by using a
CE index ratio of 1.32.
Figure B12-2 shows annual operating costs for electric power, mainte-
nance materials, and labor. Approximate drive energies for operating RBCs can
be determined from the following relationship (5):
kWh/yr = K x (Effective surface area of the RBC) (2)
where: K = 0.3 for standard media and
=0.2 for dense media
B12-4
-------
no"
First Quarter 1980 $
o 4
a 10J
I 1 I I I
\ i i i i i i i
I i i i i i i i
10
100
Wastevater Flow Rate, mVhr
Figure B12-1. Installed equipment cost for rotating biological
contactors - loading rate 0.04 m3/(day-m2) (5)
B12-5
-------
First Quarter 1980 $
Z 2
en
0
Electricity
fcLabor
4 -
2 -
0.1
J 1 I I
I 1—' ' ' U
Materials
J 1 1 I I I I
10 2 4 7 100 2 47 iQOO
Wastewater Flow Rate, rnVhr
Figure B12-2. Operating costs for rotating biological contactors (5)
B12-6
-------
Appendix B12
Rotating Biological Contactors
7. References
1. Metcalf and Eddy, Inc. Wastewater Engineering: Treatment, Disposal.
Reuse. McGraw Hill Book Co., New York, 1979.
2. U.S. Environmental Protection Agency. Coal Conversion Control
Technology, Vol. 1. Environmental Regulations; Liquid Effluents. EPA
600/7-79-228a, October 1979.
3. Adams, Carl E. , Davis L. Ford and W. Wesley Eckenfelder, Development of
Design and Operational Criteria for Wastewater Treatment. CBI Publishing
Co., Inc., Boston, MA 1981.
4. Flam, G.E. and R.E. Gerhard. Dse of the Rotating Biological Surface for
Refinery Wastewater Treatment. Presented at the 69th Annual AIChE
Meeting, Chicago, IL, Nov-Dec, 1976.
5. U.S. Environmental Protection Agency. Innovative and Alternative
Technology Assessment Manual, EPA 430/9-78-009, February 1980.
B12-7
-------
-------
APPENDIX B13
LAGOONS
1. Process Description
The term "lagoon" or "pond" can be applied to any body of water which is
contained in an earthen dike and designed for wastewater treatment. Lagoons
are also known as settling, oxidation, or stabilization ponds.
In an undisturbed lagoon the long residence time of the wastewater can
contribute to a reduction in the suspended solids content due to settling.
Also, the concentrations of various dissolved inorganic or organic pollutants
may be reduced through evaporation, stripping, chemical reactions, and/or the
action of one or more different types of microorganisms. Lagoons may be
classified, according to the type or types of biological activity occurring in
them, as aerobic, anaerobic, or facultative.
Aerobic lagoons stabilize wastewater by the action of aerobic bacteria.
There are two main groups of aerobic lagoons, differentiated according to the
primary mechanism by which oxygen is supplied to the bacteria. These are
known as aerobic algal lagoons and aerated lagoons. In aerobic algal lagoons,
oxygen is supplied both by diffusion of oxygen from the air and as a product
of photosynthesis by algae, while in aerated lagoons it is supplied by mechan-
ical mixing of the wastewater with the air. In both types of systems aerobic
bacteria convert degradable organics in the wastewater to carbon dioxide,
water, and cell tissue. Since aerobic algal lagoons depend on sunlight to pro-
vide energy for photosynthesis, they are typically quite shallow: depths may
be up to 0.5 m (1). Aerobic algal lagoons can be operated most successfully
in warm climates having long periods of sunlight, but, because of their
limited depth, they require relatively large land areas.
B13-1
-------
Appendix B13
Lagoons
Aerated lagoons must be built to accommodate the mechanical aeration
equipment which promotes both oxygen transfer and mixing of the pond contents;
depths of 2 to 6 m are commonly used (2). Turbulence levels in aerated
lagoons must be high enough to provide sufficient oxygen transfer and pond
mixing. Otherwise, suspended solids will settle to the bottom of the pond and
undergo anaerobic decomposition.
Anaerobic lagoons are generally 2.5 to 6 m deep and are characterized
by heavy organic loadings. Under these conditions the primary waste conver-
sion mechanism is by anaerobic digestion. In most anaerobic systems, some
bacteria convert the waste to organic acids; others then convert the acids to
methane, water, cell matter, and other end products such as hydrogen sulfide
and nitrogen. Typically, wastewater is introduced near the bottom of an
anaerobic pond, mixing with the microbial mass in the sludge blanket. The
discharge is submerged and is located near one of the sides of the lagoon.
Excess sludge is entrained in the effluent (3).
Facultative lagoons have depths between aerobic and anaerobic lagoons
(typically 1 to 2.5 m). In a facultative lagoon, the wastewater is allowed
to stratify to form three zones. These zones are an anaerobic bottom layer,
an aerobic algal surface layer, and an intermediate zone. Oxygen is main-
tained in the upper level by algae or by the use of aerators which do not com-
pletely mix the pond contents. Facultative ponds are the most common type of
wastewater stabilization lagoon (4). To achieve optimum performance, faculta-
tive lagoons are often operated in series with effluent recycle (3).
2. Process Applicability
Lagoon treatment is applied to the reduction of dissolved organics in a
wide variety of domestic and industrial wastewaters. Lagoons can be very
cost-effective when the relatively large areas of land required are available
B13-2
-------
Appendix B13
Lagoons
and inexpensive (5). Lagoons lose their effectiveness when water temperatures
are low and in cooler climates. Water temperatures above about 310 K are also
harmful to biological systems.
Aerated lagoons are useful in treating domestic and industrial waste-
waters of low to medium organic strength. They are susceptible to upsets
caused by changes in flow rate, organic loading, and temperature. Equaliza-
tion ponds prior to the lagoons may help alleviate these problems.
Anaerobic lagoons are most effectively used as roughing units prior to
further treatment of high strength wastes. Due to the possibility of odor
formation, their use is generally restricted to remote areas (1). They are
relatively insensitive to process upsets.
Facultative lagoons are commonly applied to the treatment of raw,
screened, or primary settled domestic wastewater and weak biodegradable indus-
trial wastewaters. They are susceptible to process upsets, and odor problems
may arise if insufficient oxygen is produced in the aerobic zone.
3. Process Performance
Lagoon treatment is capable of removing the majority of the BOD5 content
from wastewaters. The suspended solids level of the effluent relative to the
influent depends, in some cases, on the process configuration. In practice,
actual removals depend on such design and operational variables as pond depth,
retention time, organic loading, temperature, wastewater pH, flow and compo-
sition variations, and nutrient levels. Lagoon treatment requires long resi-
dence times relative to other biological treatment techniques. Important
variables in the design of several lagoon types are summarized in Table B13-1.
B13-3
-------
Appendix B13
Lagoons
TABLE B13-1. LAGOON DESIGN CRITERIA (2)
Criteria/Factors
Units
Aerated
Lagoon
Anaerobic
Lagoon
Facultative
Lagoon
Detention time
Depth
pH
Water temperature,
(optimum)
Oxygen requirement
Organic loading
days
m
K
kg 02/kg BOD, 0.7-1.4
removed
kg BOD,/ 0.001-0.03
ma-day
3-10
2-6
6.5-8
273-313
(293)
20-50
2.5-6
6.8-7.2
279-322
(303)
20-180
1-2.5
6.5-9.0
276-306
(293)
0.02-0.2
0.001-0.01
Operation
— One or
more cells
Parallel
At least 3
cells in
series
Removals of BOD£ from wastewaters having an influent range of 200 to 500
mg/L by aerated lagoons are reported to be from 60 to 90 percent with COD
removals of 70 to 90 percent. Total suspended solids removals of 70 to 90
percent are also reported (2). Anaerobic lagoons can typically remove from 50
to 70 percent of influent BOD, depending on loading and retention time. Total
suspended solids may actually increase in the effluent due to the presence of
excess sludge. In general, effluent from anaerobic lagoons is sent to addi-
tional treatment process(es) (2). Facultative lagoons are capable of BOD,
reductions of 75 to 95 percent; the efficiency depends on depth, detention
time, and temperature. Suspended solids concentrations depend on the extent of
algae separation in the last cell of the lagoon system (2). Anaerobic and
facultative lagoons may require the addition of nutrients, primarily nitrogen
and phosphorous, to assure proper operation. The amounts required depend on
the characteristics of the wastewater.
B13-4
-------
Appendix B13
Lagoons
4. Secondary Waste Generation
All types of lagoons produce a sludge residue. Aerated and facultative
lagoons must be cleaned out periodically to remove accumulated solids; this is
generally required every 10 to 20 years, but may be necessary more often. An-
aerobic lagoons produce less cell matter than other types of lagoons, and the
excess sludge is continually removed in the effluent.
All types of lagoons may produce seepage which may ultimately impact
groundwater. The amount depends on the nature and condition of the liner used
to contain the wastewater.
Aerated lagoons present a potential air impact due to the stripping
action of the air injected into the lagoon. Any of the volatile species
present in the wastewater may be released into the air from such lagoons (2).
Anaerobic lagoons and aerated facultative lagoons may produce odors due to the
release of such species as hydrogen sulfide.
5. Process Reliability
Aerated lagoons typically experience very high reliabilities. They re-
quire little operator expertise and have a service life of 30 years or more
(2). However, they are susceptible to upsets caused by changes in influent
temperature, flow, or composition, and to foaming caused by the aerators.
Anaerobic lagoons are generally fairly resistant to upsets. They are highly
reliable if controlled within the relatively narrow optimum pH limits of 6.8
to 7.2 (2). Facultative lagoons are highly reliable overall, requiring little
operator expertise. They have a service life of up to 50 years (2).
B13-5
-------
Appendix B13
Lagoons
6. Process Economics
Estimated construction costs and operating and maintenance requirements
for lagoons were obtained from reference 2. Construction costs were converted
to installed equipment costs by adding 28 percent for piping, electrical,
instrumentation, and site preparation. Hie design basis for aerated lagoons
was 7 days detention time, 4.6 m water depth, BOD loading of 0.016 kg/m*-day,
inlet BOD of 210 mg/L, floating low speed mechanical aerators producing 7.1
W/m3 of capacity, and three cell construction (2). For anaerobic lagoons the
costs were based on a detention time of 35 days, 3 m pond depth, a BOD loading
of 0.052 kg/m»-day, and inlet BOD of 600 mg/L (2). For facultative ponds two
loadings were assumed depending on climate: 0.0045 kg/m*-day for warm cli-
mates and 0.0022 for cool climates. A water depth of 1.2 m and an inlet BOD
of 210 mg/L were also assumed (2). Liner and land costs were not included for
any of the lagoon types.
The installed equipment and maintenance materials were updated from
September 1976 to first quarter 1980 dollars using a CE index ratio of 1.32.
Figures B13-1 through B13-6 summarize the estimated costs.
7. References
1. American Petroleum Institute, Manual on Disposal of Refinery Wastes,
Volume on Liquid Wastes, 1st edition, Washington, DC, 1969.
2. U.S. Environmental Protection Agency. Innovative and Alternative Technology
Assessment Manual, EPA 430/9-78-009, February 1980, pp. A62-A67.
3. U.S. Environmental Protection Agency. Treatability Manual, Volume III:
Technologies for Control/ Removal of Pollutants, EPA-600/8-80-042c, July
1980.
4. Metcalf and Eddy, Inc. Wastewater Engineering, McGraw-Hill, New York, 1972.
5. Ramalho, R.S. Design of Aerobic Treatment Units. Part 2: Aerated Lagoons
and Wastewater Stabilization Ponds, Hydrocarbon Processing, November 1979,
pp. 285-292.
B13-6
-------
5, io,ooc
First Quarter 1980 $
cu 4
1,000
100
I I I I I I 1 I
I I i i i
10
4 7 100 2 47 1000
Wastewater Flow Rate, m Vhr
Figure B13-1. Installed equipment cost for aerated lagoons (2)
B13-7
-------
Operating Costs, $/1000 m3
h-'
UO
CO
Tl
H-
OQ
0>
to
N3
o
T3
0)
O
O
cn
O
l-i
fu
rf
fti
OQ
O
O
3
cn
1
I rt
1—I—I I I I I"
y
-------
7 -
4 -
2 -
10,000
First Quarter 1980 S
7 -
-1,000
100
I I I I I I I
I I I I I I I
10
7 100 2 4 7 1000
Wastewater Flow Rate, rnVhr
Figure B13-3. Installed equipment cost for anaerobic lagoons (2)
B13-9
-------
7 -
0. 1
4 7 100 2 47 1000
Wastewater Flow Rate, mVhr
Figure B13-4. Operating cost for anaerobic lagoons (2)
B13-10
-------
7 -
H 1,000
100
7 100 2 47 1000 '
Wastewater Flow Race, mVhr
Figure B13-5. Installed equipment costs for facultative
lagoons (2)
B13-11
-------
First Quarter 1980 $
I 7
2 2
Derating 6> Maintenance
0.1
J I I I I I I I
I i i I i i i i
10
7 100 2 47 1000
Waatewater Flow Rate, m'/hr
Figure B13-6. Operating cost for facultative lagoons (2)
B13-12
-------
APPENDIX B14
ANAEROBIC DIGESTION
1. Process Description
Anaerobic digestion is a biological treatment process in which organic
molecules in the wastewater are converted to methane and carbon dioxide in the
absence of oxygen by special bacteria. The process can be viewed as occurring
in two steps, each involving a different group of microorganisms. In the
first step the organic matter is converted to simple organic acids (e.g., ace-
tic and propionic) by a group of microorganisms known as acid formers. In the
second step the acids are converted to methane and carbon dioxide by bacteria
known as methane formers. The latter step appears to be the slower and,
therefore, the rate controlling step. Other microorganisms convert sulfates
to sulfides and nitrates to nitrogen gas.
Anaerobic digestion is carried out in two basic types of digesters: stan-
dard rate and high rate. In a standard rate digestor the contents are gener-
ally unmixed and unheated. The material separates into several layers. The
layers are, from bottom to top, digested sludge, a zone where active digestion
of the waste is taking place, a relatively clear supernatant layer, a scum
layer, (consisting typically of undigested grease), and a gas layer consisting
primarily of methane and carbon dioxide. Sludge may be introduced into the
treatment vessel through a number of inlet ports; gas, supernatant liquid, and
digested sludge are removed through other ports. Residence times for standard
rate digestion are typically from 30 to 60 days.
The high rate single stage digestor is different from the standard rate
unit in two major ways: the contents are completely mixed and heated. The
effluent mixture must be separated into heated sludge and supernatant phases.
A detention time of 15 days or less is typical for high rate digestion.
B14-1
-------
Appendix B14
Anaerobic Digestion
A combination of these processes is known as the two-stage process. A
high rate digestor is followed by another vessel whose primary function is to
separate the digested solids from the supernatant. Gas is collected from both
vessels.
The gas produced by anaerobic digestion, which is roughly 2/3 methane and
1/3 carbon dioxide by volume, is generally collected. A portion is used to
provide process heating in the case of high rate digestion; the excess is
typically flared or used as a fuel.
Important parameters in the successful operation of anaerobic digestion
are the exclusion of dissolved oxygen and the assurance that inhibitors such
as heavy metals, sulfides, and toxic materials are not present in concentra-
tions which might be harmful to the bacteria. The pH should range from about
6.6 to 7.6; pH below 6.2 will cause the methane-forming bacteria to cease
functioning (1). Nutrients such as nitrogen and phosphorus must also be
available. Optimum temperatures are between 303-311 K for mesopb.il ic organ-
isms and 322-330 K for thermophilic organisms. Although the process can oper-
ate at lower temperatures, the rate of digestion decreases, making longer re-
tention times necessary to achieve equivalent levels of treatment (2).
2. Process Applicability
Anaerobic digestion is currently used primarily to stabilize sludge from
aerobic treatment processes. Sludge from both municipal sewage treatment
plants and a number of industrial processes are currently treated in this man-
ner. The use of the process as a primary treatment method for the direct
reduction of more dilute organic waste is a newer application (1). As men-
tioned previously, care must be taken to avoid wastes which contain inhibitors
such as sulfides and heavy metals. Odor problems can also arise when treating
sulfur-containing wastes due to the formation of hydrogen sulfide (3).
B14-2
-------
Appendix B14
Anaerobic Digestion
3. Process Performance
Anaerobic digestion of slndges is effective in reducing slndge volume.
since much, of the organics content of the sludge is converted to methane and
carbon dioxide. The treated sludge is often suitable for landfill ing or even
for use as a soil conditioner (1). Sludge treatment typically results in a
total solids removal ranging from 33 to 58 percent and destruction of 85 to
nearly 100 percent of pathogens.
4. Secondary Waste Generation
Anaerobic digestion produces two secondary waste streams: waste gas and
treated sludge. The gas stream, consisting primarily of methane and carbon
dioxide, can be combusted to provide process heat. Due to its flammable
nature, care must be taken to avoid mixing with air so that combustible mix-
tures are avoided. Explosions have occurred in wastewater treatment plants
due to methane from anaerobic digestion (1).
The treated sludge has often been found to be suitable, after dewatering,
for landfill disposal. The nature of the spent sludge will depend, of course,
on the characteristics of the waste being treated.
5. Process Reliability
Anaerobic digestion can be operated successfully if good control is main-
tained over pH, alkalinity, temperature, and toxics concentrations. Most
applications of the process are relatively sensitive to changes in any of
these variables.
B14-3
-------
Appendix B14
Anaerobic Digestion
6. Process Economics
Cost estimates were obtained from the literature for mesopb.ilic and
thermoph.ilic 2-stage digestion systems (2). Estimates of construction costs
for mesophilic digestion were based on operation between 303-311 K. Equipment
items included digester vessels, heat exchangers, gas collection equipment,
and control building. Construction costs were converted to installed equip-
ment costs by adding 28 percent for piping, electrical, instrumentation, and
site preparation. Operating and maintenance requirements included electric
power for pumping, maintenance materials, and labor. Estimates for thermo-
philic digestion systems operating at a temperature of 327 K included similar
equipment items and operating and maintenance requirements. Figures B14-1
through B14-4 show these costs, updated from September 1976 to first quarter
1980 dollars using a CE index ratio of 1.32, as a function of wastewater flow
rate.
7, References
Metcalf and Eddy, Inc. G. Tchobanoglous, ed. Wastewater Engineering:
Treatment, Disposal and Reuse. McGraw-Hill, New York, 1979.
U.S. Environmental Protection Agency. Innovative and Alternative
Technology Assessment Manual. EPA-430/9-78-009, February 1980. pp.
A218-A221.
Watkins, J.P. Controlling Sulfur Compounds in Wastewaters Chemical
Engineering, October 17, 1977, pp. 61-65.
B14-4
-------
10,000
First Quarter 1980 S
1,000
100
I I I I I I
I I I I I I I I
10
4 7 100 2 47 1000
Wastewater Flow Rate, raVhr
Figure B14-1. Installed equipment cost for anaerobic
mesophilic digestion systems (2)
B14-5
-------
10
First Quarter 1980 $
c 2
Labor
0.1
2 " 4 ~" 7 1000
Wastewater Flow Rate, m Vhr
Materials
Figure. BI4-2. Operating costs for anaerobic mesophilic
digestion systems (2)
B14-6
-------
10.00C _
First Quarter 1980 $
~ 0
-------
Operating Costs, $71000 m3
to
CO
H-
OQ
P- O
H. T)
OQ ra
fP >-f
en p
rr rt
H- H-
O 3
3 03
m O
l< O
cn 03
rt rt
ro en
a
en i-n
O
3
BJ
0)
>-(
O
cr
H-
n
cr
H-
i i i 11
1 1 I I 1 I I
~> 1—I—I I I I
-------
APPENDIX B15
ACTIVATED CARBON ADSORPTION
1. Process Description
Activated carbon adsorption is one of the most widely used methods for
removing residual organics from industrial wastewaters. The application of
activated carbon may be carried out by two methods: 1) the use of powdered
activated carbon in conjunction with biological treatment (discussed in
Appendix BIO) and 2) the direct contact of granular activated carbon with
contaminated wastewaters. The latter method is discussed in this appendix.
An activated carbon adsorber is simply a packed bed of the adsorbent
through which the wastewater is passed. Operation may be at atmospheric pres-
sure (gravity adsorber) or at elevated pressure (pressure adsorber). Three
types of adsorbers are commonly used in wastewater applications: downflow
mixed bed, packed moving-bed, and upflow expanded bed. Multiple adsorbers may
be operated either in series or in parallel. Spent carbon is emptied from the
downflow fixed bed and upflow expanded bed adsorbers after removing them from
service. In the moving-bed or pulsed-bed adsorber, portions of the bed are
periodically removed from the bottom of the vessel and replaced with reacti-
vated carbon added at the top (1).
Where quantities of wastewater greater than 11,400 to 18,300 m /day are
treated, or where carbon usage exceeds 450 kg/day, it is usually economical to
provide on-site regeneration of the carbon. Smaller facilities either dispose
of the spent carbon or ship it off-site for regeneration (2).
Granular activated carbon used for treating synfuel wastewaters will most
likely be regenerated by thermal means. The most common method of regenera-
tion is thermal oxidation in a multiple hearth furnace. Other furnace
configurations such as rotary kilns are sometimes used. Another thermal
regeneration process which shows promise is wet air oxidation.
B15-1
-------
Appendix B15
Activated Carbon Adsorption
2. Process Applicability
Granular activated carbon adsorption has been widely used in water treat-
ment for many years and in waste treatment for some 10 to 20 years (2). It
is used primarily to remove dissolved organics, although the process can also
remove cyanides and complezed metal ions (3,4).
Activated carbon adsorption is generally limited to streams which do not
contain high levels of suspended solids, oil, or grease. Hie presence of
these contaminants leads to clogging and coating of the carbon, greatly dimi-
nishing its adsorption capacity. Frequent backwashing and replacement of the
carbon are then required. Thus, the use of oil separators and/or filters is
required ahead of the carbon adsorption unit if excessive levels of suspended
materials are present in the wastewater.
Adsorption is generally most effective for fairly large organic molecules
which are slightly soluble and nonpolar. Activated carbon does not adsorb
light hydrocarbons such as methane or ethane, or extremely heavy, large or-
ganic molecules. Aromatic compounds are generally removable by adsorption, as
as are many organic molecules in the size range of the pores in the carbon
(10 ' to 10~Tm) (1). Carbon adsorption is often used to remove some compounds
which are nondegradable by biological treatment (5). While the statements
made above are accurate as general guidelines, ultimately, the applicability
of activated carbon adsorption must be determined by examining the adsorption
isotherms for each component in the waste, by conducting column tests, and, in
some cases, by conducting pilot-scale tests (1.6,7). These tests can help
establish adsorption limits and economics for the waste of interest.
B15-2
-------
Appendix B15
Activated Carbon Adsorption
3. Process Performance
The ability of carbon adsorption to remove commonly measured wastewater
parameters such as BOD, COD, and IOC depends very strongly on the nature of
the specific compounds in the wastewater contributing to those parameters.
Other variables also influence contaminant removals, including pH, tempera-
ture, pollutant concentrations, and contact time. Lowering the pH will
enhance the removal of organic acids, while raising the pH will enhance the
removal of bases. Increasing the contact time generally improves removal
efficiency. Higher percentage removals are seen for wastes with higher
contaminant concentrations.
Removals of 99 percent for phenol, 80 percent for COD, and 98 percent for
cyanide have been reported (2,3). Removal efficiencies for tertiary treatment
of refinery wastewaters from Reference 2 are summarized in Table B15-1 Other
removal data, primarily for wastewaters having compositions closer to those
expected for synfuel facilities are presented in Table B15-2.
4. Secondary Waste Generation
The most important secondary waste streams generated by carbon adsorp-
tion are related to the backwashing and regeneration steps. Backwashing of
downflow fixed bed adsorbers produces a stream containing high levels of
suspended solids. Thermal regeneration results in a gas stream which is
combusted prior to release to the atmosphere, but may contain some unburned
organics, carbon monoxide, and other pollutants derived from furnace fuel and
waste material combustion, such as S0a, NO , as well as carbon dioxide.
Scrubbers can be used to control the release of these contaminants. Solid
wastes include ash and carbon fines. If the carbon is not regenerated, the
spent carbon must be disposed of.
B15-3
-------
TABLE B15-1.
ACTIVATED CARBON ADSORPTION EFFICIENCIES
FOR REFINERY WASTEWATERS (2)
Pollutant/Parameter
Influent
Concentration
Range, mg/L
Effluent
Concentration
Range, mg/L
Percent Removal
Range
BOD
COD5
TOC
TSS
Oil and Grease
Total Phenol
13-57
29-127
17-45
3.8-21
6.3-12.3
0.002-2.7
<8.3-<12
11-79
6.2-33
<1.3-7.7
1.8-14
<0. 005-0. 02
>20-83
34-80
27-67
0-79
0-85
>50-99
TABLE B15-2. ACTIVATED CARBON ADSORPTION EFFICIENCIES FOR WASTEWATER
SIMILAR TO THOSE EXPECTED FROM SYNFDELS FACILITIES
Component
Removal
Reference
Tars
Oils
Organic Acids
COD
BOD
TOC
Phenols
99%
99%
70%
80%
60%
70%
99.9%
8
8
9
8
10
*Based on adsorption isotherms generated in the laboratory using
pretreated process condensate from Kosovo, gasification facility.
B15-4
-------
Appendix B15
Activated Carbon Adsorption
5. Process Reliability
Granular carbon adsorption is reported to be moderately reliable both
mechanically and operationally, depending on design, construction, and manufac-
tured equipment quality (2). Regeneration furnaces are subject to more
mechanical failures than other wastewater treatment processes and, thus, have
high maintenance costs (2). The wide use of activated carbon adsorption is
indirect evidence that the process has reasonable reliability.
6. Process Economics
Cost estimates for activated carbon adsorption/regeneration facilities
were obtained from the literature (5). Installed equipment costs for pressur-
ized, granular activated carbon contactors include carbon steel vessels,
valves, piping, instrumentation, and control panel. The need to use materials
of construction different than carbon steel (in order to avoid corrosion
problems) will obviously alter the costs reported. Design assumptions include
30 minutes empty bed contact time and 0.35 MPa working pressure. These esti-
mated costs were updated from 1978 to first quarter 1980 dollars using a CE
index ratio of 1.18 and plotted as a function of bed volume. Figure B15-1
summarizes the results.
Operating and maintenance requirements for activated carbon adsorption
include electric power for 10 minutes/day of backwash pumping at 8.9 m'/hr-m*.
and miscellaneous work items. Labor includes operating and maintenance labor
as well as supervision (5). The costs are presented as a function of contac-
tor volume in Figure B15-2. Not shown are electric energy costs, which were
found to be nearly constant at $0.34 per m of contactor volume per year.
B15-5
-------
10,000
First Quarter 1980 S
l,000
100
I I I I i
10
100
2 47 1000
Contactor Vessel Volume, in3
Figure B15-1. Installed equipment cost for activated carbon
adsorption contactors (5)
B15-6
-------
7 -
100 2 47 1000
Contactor Vessel Volume, m3
Figure B15-2. Operating costs for activated carbon
adsorption contactors (5)
B15-7
-------
Appendix B15
Activated Carbon Adsorption
Installed equipment costs for thermal regeneration of activated carbon in
a multiple hearth furnace are plotted in Figure B15-3 on the basis of first
quarter 1980 dollars per mass of carbon regenerated. The assumption used to
convert the literature data to this basis was that 340 kg of carbon would be
regenerated per square meter of hearth area per day. The literature data were
updated from 1978 to first quarter 1980 dollars by using a CE cost index ratio
of 1.18. Figure B15-4 shows operating and maintenance costs obtained from
Reference 5. Not shown are steam requirements, estimated in the literature at
1 kg per kg of carbon regenerated (11).
Costs for granular activated carbon for purchase, delivery, and placement
were also taken from Reference 6. These costs, as first quarter 1980 dollars
per kg, can be found in Figure B15-S. The figure can be used to estimate the
cost of the initial carbon charge, as well as that of the makeup carbon needed
to replace that lost in regeneration (from 2 to 10 percent of the carbon
regeneration rate).
B15-8
-------
First Quarter, 1980 S
100
7 -
2 -
100
7 1000
7 10,000
Carbon Regeneration Rate, kg/day
Figure B15-3. Installed equipment cost for activated carbon
regeneration furnaces (5)
B15-9
-------
10
First Quarter 1980 $
Electricity
0.1
100
Materials
j 1
J I I I I I I
I i i i i i I i
7 1,000 2 47 10,000 2
Carbon Regeneration Race, kg/day
Figure B15-4. Operating costs for activated carbon
regeneration furnaces (5)
B15-10
-------
10
First Quarter 1980 $
0.1
I i i t i
10,000
7 100,000 2 47 1,000,000 2 4
Mass of Carbon Purchased, kg
Figure B15-5. Activated carbon cost for purchase, delivery,
and placement (5)
B15-11
-------
Appendix B15
Activated Carbon Adsorption
7. References
1. Bernardin, F.E. Selecting and Specifying Activated Carbon Adsorption
Systems. Chemical Engineering, October 18, 1976, pp. 77-82.
2. U.S. Environmental Protection Agency. Treatability Manual, Volume III.
Technology for Control/ Removal of Pollutants. EPA-600/8-80-042c, July
1980.
3. Kunz, R., Casey, J., and Huff, J. A Review of Cyanide in Refinery
Wastewaters,. In proceedings of Third Annual Conference on Treatment and
Disposal of Industrial Wastewaters and Residues. Houston, Texas, April 18-
20, 1978, pp. 65-77.
4. Krickenberger, K. and Ellerbnsch, F. Best Available Technologies (BATEA)
Applicable to the Electroplating Industry. In proceedings of Third
Annual Conference on Treatment and Disposal of Industrial Wastewaters and
Residues, Houston, Texas, April 18-20, 1978, pp. 108-116.
5. D.S. Environmental Protection Agency. Estimated Cost for Water Treatment
as a Function of Size and Treatment Efficiency. EPA-600/2-78-182,
Culp/Hesner/Cnlp, Santa Ana, CA, 1978.
6. Mulligan, T.J. and R.D. Fox. Treatment of Industrial Wastewaters.
Chemical Engineering, October 18, 1976, pp. 49-66.
7. Van Stone, G.H. and G.J. Schwartz. Developments in Industrial Applica-
tions for Granular Activated Carbon, Third Annual Conference on Treatment
and Disposal of Industrial Wastewaters and Residues, Houston, TX, April
18-20. 1978.
8. Juentgen, H. and K. Juergen. Purification of Wastewater from Coking and
Coal Gasification Plants using Activated Carbon, Amer. Chem. Soc., Div.
of Fuel Chem. Prepr., 19(5): 67-84, 1974.
9. Giusti, D.M., R.A. Conway, and C.T. Lawson. Activated Carbon Adsorption
of Petrochemicals. J. WPCF, 46(5): 947-965, 1974.
10. U.S. Environmental Protection Agency. Applicability of Coke Plant
Control Technologies to Coal Conversion. EPA-600/7-79-184, PB80-108954,
Catalytic, Inc., Philadelphia, PA. August 1979.
11. Smith, S.B. Activated Carbon III - Alternative and Relative Costs of
Regeneration Processes. Proceedings of 19th Annual Pub. Water Supply
Eng. Conf., 1977, pp. 81-100.
B15-12
-------
APPENDIX B16
CHEMICAL OXIDATION
1. Process Description
Chemical oxidation, as its name implies, is simply the oxidation of
wastewater contaminants, both organic and inorganic, to more environmentally
acceptable forms, such as water, carbonates, nitrates, etc. The most widely
used chemical oxidants are chlorine, chlorine dioxide, ozone, hydrogen per-
oxide, and potassium permanganate.
Simple equipment is required for chemical oxidation: storage vessels for
the oxidizing agents, metering equipment for the wastewater and oxidant, con-
tacting vessels (e.g., stirred tank reactors, flash mixers, and plug flow con-
tactors), and instrumentation for monitoring pH and the degree of completion
of the oxidation reactions (1).
2. Process Applicability
Although chemical oxidants can be used to oxidize both organic and inor-
ganic contaminants, high stoichiometrie requirements and high costs will gen-
erally limit the use of the process to residual organic polishing applications.
Chlorine is widely used to disinfect potable water and wastewater. It
will oxidize a wide variety of substances including ammonia, cyanide, sulfide,
and mercaptans (1). In some applications (known as alkaline chlorination), so-
dium hydroxide or lime slurry is added. Chlorine is usually not recommended
for use with wastes containing high concentrations of residual organics be-
cause it can react to form chlorinated compounds (e.g., chlorinated phenols)
which are more toxic than the original materials (2). Control of pH is also
very important, both to maximize reaction rates and to avoid producing toxic
gases. For example, oxidation of cyanides should always be done at a pH of
8.5 or greater if liberating hydrogen cyanide or cyanogen is to be avoided
(3).
B16-1
-------
Appendix B16
Chemical Oxidation
Chlorine dioxide, C1Q2, has been reported to be a selective oxidant for
industrial wastewaters containing cyanide, phenol, sulfides, and mercaptans.
It does not react with a number of other organics including alcohols, glycols,
ketones, aldehydes, and organic acids. As with chlorine, pH control is very
important (4). Chlorine dioxide does not form chlorophenols. Instead, it
attacks the benzene ring to form odorless, tasteless compounds. Chlorine
dioxide will also oxidize chlorophenols formed during chlorination (2). It is
produced by the reaction of chlorine with sodium chlorite (5).
Ozone is one of the most powerful oxidants available. It is sufficiently
strong to oxidize a wide variety of organic compounds such as phenols, includ-
ing polyhydric phenols which are important color precursors, cyanides, organic
metallic complexes, organic nitrogen compounds, and other toxic organics
(4,6). Because of its nonselective nature, ozonation will probably not be
economical for wastes containing high levels of contaminants(e.g., COD, BODj,
or TOC should not be greater than 300 mg/L) (6). Ozone is generated onsite by
passing dry, oil-free air or oxygen through an electric corona discharge.
Ozone is quite unstable, breaking down to molecular oxygen with a half life of
about 20 minutes.
Hydrogen peroxide, H204, can be used to remove sulfides, mercaptans,
amines, cyanides, and lead from wastewater (1,2,7). It is a particularly
effective oxidizer of phenols over a wide range of temperatures and concentra-
tions. Acid pH (3-5) results in the most effective reaction times. Small
amounts of metal salts can be added to increase reaction rates. Iron, copper,
aluminum, or chromium salts may be used, although iron in the ferrous state is
preferred for most applications (2). Hydrogen peroxide has been used in equi-
lization basins before biological treatment systems when sudden high levels of
phenols are present due to spills or upsets. This has been done to prevent
shock loads to the organisms in downstream biological treatment systems (2).
Hydrogen peroxide has also been used to dechlorinate water after chlorine
treatment (1).
B16-2
-------
Appendix B16
Chemical Oxidation
Potassium permanganate, KMn04, has been used to destroy organic resi-
dues in wastewater as well as potable water systems. It reacts with alde-
hydes, mercaptans, phenols, and unsaturated organic acids. Generally speak-
ing, quite high stoichiometric requirements exist for permanganate oxidation.
As a result, its use is typically limited to treating low organic strength
wastewaters. For example, weight ratios of 6 to 7 pounds of KMn04 per pound
of phenol removed have been found to be effective (2). Alkaline pH (7-10)
is preferred for this application.
3. Process Performance
The extent to which a given contaminant is oxidized by a chemical oxidant
depends on the contaminant concentration, the complexity and type of the mole-
cule to be oxidized, the oxidant used, the design of the reaction vessel
(i.e., the reaction time provided and the degree of wastewater and oxidant
mixing obtained), and the pH of the reaction mixture. Wastewaters with high
organics loadings may require multiple reactors. There are limited data
available on the application of the technology to synthetic fuel plant waste-
waters. What is available is mostly from bench and pilot scale studies. Data
presented here include not only bench and pilot plant data but also results
obtained from the treatment of wastewater streams in other industries. Data
were not located for potassium permanganate oxidation of wastewater.
Reported removals of a number of pollutants from wastewaters by chlorina-
tion are summarized in Table B16-1. The wastewater sources are various in-
organic chemical and ore treating facilities (1).
Preliminary results of ozone treatment of aqueous waste produced by the
Oak Ridge National Laboratory Hydrocarbonization bench scale facility showed
reductions in phenols and polynuclear aromatic hydrocarbons of 95 to 98 per-
cent with contacting times of 15 minutes or less (8). Ozonation has also
B16-3
-------
Appendix B16
Chemical Oxidation
shown promise for the elimination of nitrogen heterocyclics, polyhydric phe-
nols (color precursors), and others (9). The effect of ozonation at various
dosage rates on COD, BOD , and TOC for two laboratory tests is shown in Table
B16-2. Contacting times were 90 minutes (6).
Reduction time tests indicate that the rate of reaction slowed signifi-
cantly after 70 percent reduction in COD. In fact, COD removal never reached
100 percent in the laboratory tests even at a high ozone dose of 300 mg/L.
TABLE B16-1. CONTROL SOMMARY FOR CBLORINATION (1)
Pollutant
COD
TSS
Copper
Cyanide
Lead
NH3-N
Effluent Concentration
Range, mg/L
441-978
33-159
320*
<2-130
2,500*
120a
Removal Efficiency
Range. %
7-39
0-97
14a
58-99+
Oa
36«
aBased on a single measurement.
TABLE B16-2. CONTROL SUMMARY FOR OZONATION (6)
Ozone
Dosage,
mg/L
50
100
200
50
100
200
325
COD
Influent,
mg/L
45
45
45
318
318
318
318
Removal,
%
40
76
88
18
23
37
50
BODs
Influent,
mg/L
13
13
13
142
142
142
142
Removal,
%
46
77
88
23
30
33
58
TOC
Influent,
mg/L
20.5
20.5
20.5
93
93
93
93
Removal,
%
24
56
76
14
17
14
46
B16-4
-------
Appendix B16
Chemical Oxidation
Phenol removals by ozone from paint stripping wastewater were reported to
be 99 percent with an initial phenol concentration of 2500 ppm; 1.7 kg of
ozone per kg of phenol was required at a pH of 11 and 60 minntes contacting
time. As an indication of the effect of pH, at a pH of 8.1 the same waste
required 5.3 kg 0, per kg of phenol and about 200 minutes contacting time to
achieve the same removal (2).
Hydrogen peroxide treatment of a phenolic waste containing 2000 mg/L of
phenol using 0.02 percent ferrous iron catalyst at a pH of 6.5 gave the
results listed in Table B16-3. Reaction times of about 5 minutes are required
for Hj02 oxidation (2).
TABLE B16-3. CONTROL SUMMARY FOR HYDROGEN PEROXIDE (2)
Peroxide dose
ke H,0,/k« COD
0.3
0.4
0.7
0.8
1.0
Phenol removal,
%
86
94
99.8
99.9
100
COD Removal.
%
28
32
44
52
69
4. Secondary Waste Generation
In general, chemical oxidation produces minimal secondary wastes. Some
sludges and evolved gases can be formed. The formation of secondary wastes as
the result of chemical oxidation depends on the oxidant used, the extent of
oxidation of certain contaminants, and pH.
Chlorination of wastes containing organics may produce effluent contain-
ing species which are more objectionable than the original contaminants due
to their increased toxicity. Likewise, if a waste containing ammonia is
B16-5
-------
Appendix B16
Chemical Oxidation
chlorinated with inadequate control, chloroamines may be produced, which are
toxic to fish (4). If excess chlorine is added to ensure more complete oxida-
tion, the residual chlorine in the effluent could be a problem (1). Dechlori-
nation using peroxide is one possible solution. Residues may be formed during
alkaline chlorination from the use of lime slurry or caustic.
The use of ozone in wastewater treatment leaves no inherent harmful resi-
due. Ozone itself is considered to be toxic, but it decomposes readily to
oxygen with a half life of about 20 to 30 minutes (1). The additional dis-
solved oxygen can be considered a beneficial residue. The ozonation of phe-
nols may produce carbon dioxide and water, although usual dosages generally
result in the formation of toxic intermediates which are readily biodegradable
by natural processes or in biological treatment units (2).
Hydrogen peroxide also readily decomposes, leaving water and oxygen. It
does not add to the dissolved solids load of the effluent. However, metal
ions used as oxidation catalysts may add dissolved solids to the effluent.
Oxidation of cyanides produce cyanates which are much less toxic than the ori-
ginal cyanides and which hydrolyze to form carbon dioxide and ammonia. The
products of snlfide oxidation may either be sulfates or elemental sulfur
depending on whether the medium is basic or acidic. The oxidation of higher
mercaptans produces insoluble disnlfides which can form an oily layer. The
final product of phenol oxidation depends strongly on the H,02 dosage; it may
range from hydroquinone and catechol at low dosages, through quinones and
carboxylic acids to carbon dioxide and water as the dosage increases (7).
Use of potassium permanganate leads to the formation of manganese dioxide
which forms an insoluble sludge (2). Cyanides are oxidized to cyanates. The
pH must be maintained above about 9 to avoid the formation of poisonous gases
(i.e., hydrogen cyanide and/or cyanogen) (1).
B16-6
-------
Appendix B16
Chemical Oxidation
5. Process Reliability
Since the equipment used in chemical oxidation is typically quite simple,
the chemical oxidation process has proven to be highly reliable in demonstra-
ted applications (1). Chlorination has been used for many years in this Coun-
try to disinfect water and has proven to be extremely reliable. Ozonation
has been used to treat a number of industrial wastewater streams and is widely
used in place of chlorine to treat potable water in Europe.
Applications of chemical oxidation to wastewaters from synthetic fuels
facilities have been confined to bench or pilot scale studies.
6. Process Economics
In general, the economics of chemical oxidation favor the treatment of
wastewater streams having low concentrations of contaminants. The process is
more favorable as a polishing step than for treatment of streams with high
pollutant concentrations. This is mainly due to the high unit cost of the
oxidants and the stoichiometric requirements for contaminant removal (9).
Installed equipment costs for chlorination systems as a function of
chlorine feed rate were obtained from Reference 5. Storage and feed system
costs are included (5). Figure B16-1 shows these installed equipment costs,
updated from October 1978 to first quarter 1980 dollars by the CE index ratio
of 1.16, as a function of chlorine feed rate. Three design options are con-
sidered: 1) ton cylinders, 2) on-site storage with bulk rail delivery, and 3)
direct feed from a rail car. The original reference can be consulted for
details. Operating and maintenance requirements were taken from the same
source. Electric power requirements are for such items as electric hoists,
evaporators, and injector pumping. Maintenance materials requirements have
been found not to vary for the three feed options, based on experience at
operating plants. Labor requirements are for routine operation. Chlorination
B16-7
-------
IOC
First Quarter 1980 S
10
7
Ton Cylinderi
I I
.Storage Tank.
Rail Car
1 I i I i I I i
100
7 1000
7 10,000
Chlorine Feed Rate, kg/day
Figure B16-1. Installed equipment costs for chlorination systems (5)
B16-8
-------
Appendix B16
Chemical Oxidation
system operating costs are found in Figure B16-2. The cost of chlorine is not
included; this cost depends on a number of factors including bulk versus
cylinder purchase.
Chlorine dioxide generating and feed system installed equipment costs
were taken from Reference 5. Items included are the C10a generator (high
strength chlorinated solution and high strength sodium chlorite solution mix-
ing vessel), sodium chlorite mixing and metering system, and chlorine feed
system. These costs were updated from October 1978 to first quarter 1980
dollars by the CE index ratio of 1.16 and are plotted in Figure B16-3 as a
function of chlorine dioxide feed rate. Operating and maintenance require-
ments for electric power, materials and labor from the same source are
summarized in Figure B16-4. Costs of sodium chlorite and chlorine, estimated
at 1.68 kg of sodium chlorite (80 percent) and 1.68 kg of chlorine per kg of
C10a generated (5), are not included but will be a major annual cost item.
Installed equipment costs for ozone generation, dissolution, and re-
cycling were taken from Reference 5. For ozone generation capacities over 45
kg/day, the costs include oxygen generation costs. If excess oxygen capacity
exists in a synthetic fuels plant, the reported costs could be lowered signi-
ficantly. Costs of contacting chambers are listed separately, since desired
contact times may vary depending on wastewater characteristics. Operating and
maintenance requirements were obtained from the same source. Ozone is gener-
ated by electrical discharge, so the electricity requirements for this process
are significant. Chemical costs for this technology are negligible, since air
is the raw material for ozone production. Figures B16-5 through B16-7 pre-
sent ozone system costs, updated to first quarter 1980 dollars as discussed
previously.
B16-9
-------
Operating Costs, S/1000 kg
I
I-1
o
OQ
e
N3
o
•a
m
3
OQ
O
O
0)
rt
0)
Ml
O
i-i
H-
H-
O
3
cn
*<
tn
en
Ul
o
o
o
\
-------
1,000
First Quarter 1980 $
100
£ -r
10
I I I I I I 1
10
7 100
7 1000
Chlorine Dioxide Feed Rate, kg/day
Figure B16-3. Installed equipment cost for chlorine dioxide
generating and feed systems (5)
B16-11
-------
Operating Costs, S/1000 kg
I
t—'
K3
OQ
C
i-t
ro
(U O
0 13
CL ro
l-i
Hi P>
ro rt
ro H-
D. 0
OQ
^ o
Cn O
rt CO
ro rt
9 en
en
K)
/-v O
o
H
H-
0
ro
H-
D-
ro
00
ro
0
ro
rt
H-
OQ
-------
10,000
First Quarter 1980 $
-1,000
100
J 1—I I I I I
J 1 1 i I I I
10
7 100
7 1000
Ozone Feed Rate, kg/day
Figure B16-5. Installed equipment cost for ozone generation,
dissolution, and recycling equipment (5)
B16-13
-------
1,000
First Quarter 1980 $
100
10
J I I I I I I I
I i i I i I I i
10
7 100 2 47 1000
Ozone Contactor Vessel Volume, m3
Figure B16-6. Installed equipment cost for ozone contactor
vessels (5)
B16-14
-------
7 -
4 -
1,000
100
10
7 100 2 47 1000
Ozone Feed Rate, kg/day
Figure B16-7. Operating costs for ozonation system (5)
B16-15
-------
Appendix B16
Chemical Oxidation
Potassium permanganate storage, mixing, and feed system installed equip-
ment costs and operating and maintenance requirements were found in Reference
5. Chemical costs are not included, nor are capital equipment costs for
sludge removal. Potassium permanganate is stored in drums, dissolved, and di-
luted on site. Figures B16-8 and B16-9 show potassium permanganate system
costs updated to first quarter 1980 dollars as discussed previously.
7. References
1. U.S. Environmental Protection Agency. Treatability Manual. Volume III,
Technologies for Control/ Removal of Pollutants. EPA 600/8-80-042c, July
1980.
2. Lanouette, K.H. Treatment of Phenolic Wastes. Chemical Engineering,
October 17, 1977, pp. 99-106.
3. Lash, L.D., and Kominek. E.G. Primary Waste Treatment Methods. Chemical
Engineering, October 6, 1975, pp. 49-61.
4. Mulligan, T.J., and Fox, R.D. Treatment of Industrial Wastewater.
Chemical Engineering, October 18, 1976, pp. 49-66.
5. U.S. Environmental Protection Agency. Estimating Water Treatment Costs.
Volume 2, Cost Curves Applicable to 1 to 200 Mgd Treatment Plants. EPA
60072-79-l62b, August 1979.
6. U.S. Environmental Protection Agency. Innovative and Alternative Techno-
logy Assessment Manual. EPA 430/9-78-009, February 1980, pp. A-136,
A-137.
7. FMC Corporation. Industrial Chemical Group, Industrial Waste Treatment
with Hydrogen Peroxide. Vendor literature, undated.
8. Klein, J.A., Assessment of Environmental Control Technology for Coal
Conversion Processes. In Processing Needs and Methodology for Waste-
waters from the Conversion of Coal, Oil Shale and Biomass to Synfuels,
DOE/EV-0081, May 1980.
9. U.S. Department of Energy. Processing Needs and Methodology for Waste-
waters from the Conversion of Coal, Oil Shale and Biomass to Synfuels,
Introduction, DOE/EV-0081, May 1980.
B16-16
-------
1,000
First Quarter 1980 S
- 100
10
J I
I I 1 1 L I I
7 10
100
Potassium Permanganate Feed Rate, kg/day
Figure B16-8. Installed equipment cost for potassium permanganate
storage, mixing, and feed systems (5)
B16-17
-------
Operating Costs, ?/kg
I
M
00
H-
oq
H
rc>
I
VO
en o
rr -d
O (D
OT rr
fC H-
- 3
00
H- n
X o
H- W
3 rt
CTQ CO
l-ti O
fD ft
m to
CL en
en
en H-
^ C
en g
rt
ro 'd
ui 3
^OQ
fu
3
Co
n-
(D
1 I I I
T 1 I I I I I
T 1—I I I I I
-------
APPENDIX B17
THERMAL OXIDATION
1. Process Description
Thermal oxidation or incineration is a high temperature process for the
destruction of a variety of wastewater contaminants. Some wastewater inciner-
ators only heat water (with no major amount of the water vaporized) and have
low energy requirements and low destruction capabilities. The wastewater in-
cinerators which have the greatest destruction capabilities, and the greatest
energy requirement, are those which introduce wastewater directly into the
combustion zone. Only the latter type will be considered in this appendix.
Wastewater incineration can be considered a combination of evaporation, pyrol-
ysis, and oxidation although oxidation is the primary process leading to
the ultimate destruction of most toxic pollutants (1). Pyrolysis, or destruc-
tive distillation, is a process in which heat breaks down the waste material
into simpler components which can either be recovered or oxidized more easily
than the original waste. Oxidation, or combustion, promotes the reaction of
waste components or pyrolysis products with oxygen to form such products as
carbon dioxide, water, and oxidized inorganics such as sulfates and nitrates.
Thermal oxidation is capable of treating a variety of organic-laden
wastes including wastewater and biological treatment sludges. Sludge inciner-
ation is discussed in Appendix C4. An important parameter to consider in a
thermal oxidation system is the energy content of the waste material. Wastes
having a high energy content are self sustaining, that is, their oxidation
liberates sufficient energy to raise the waste temperature to the temperatures
required to sustain the oxidation reactions and drive them to completion. Low
energy wastes require the addition of auxiliary fuel. Preconcentration of a
low energy aqueous waste may also be used to reduce or eliminate the need for
an auxiliary fuel.
B17-1
-------
Appendix B17
Thermal Oxidation
Figure B17-1 shows a typical wastewater incinerator system. The im-
portant elements are the liquid injection system, incinerator, and the auxi-
liary gas cooling/quenching and heat recovery units.
Gas cooling/quenching and scrubbing units are not necessarily required in
order to operate the incineration process. However, they may be used to re-
cover energy and/or reduce auxiliary fuel requirements by preheating com-
bustion air, preconcentrating the wastewater, or recovering steam. For ap-
plications in this manual, the incinerator quench system is operated such that
the offgas temperature is approximately 350 K. The quench system blowdown
(which removes wastewater IDS which is not destroyed or does not exit with the
offgas) is the treated wastewater stream. The flowrate of this stream will
depend on the inlet wastewater composition and the degree and type of treat-
ment applied to the quench offgas. For incinerating preconcentrated wastes,
the quench system blowdown will be approximately equal to the inlet wastewater
flow rate.
2. Process Applicability
Wastewater incineration is applicable to aqueous wastes containing high
concentrations of organic* which are resistant to treatment by other organics-
removal processes, such as biological or chemical oxidation (1,2,3). Incin-
eration does not destroy inorganic compounds in wastewaters. In fact, such
species often cause problems due to corrosion and reactions with refractory
liners in incineration systems (1,2).
For synthetic fuels applications, incineration can be considered for
treating concentrated, low-volume wastewater streams containing high levels of
organics. It is likely to be economical only for such streams because of the
costs of auxiliary fuel (2,4) and the equipment required.
B17-2
-------
WASTEWATER-
STREAM
QUENCH
MAKE-UP WATER
FUEL?-
~^T I -.COMBUSTION
THERMAL
OXIDATION
INCINERATOR
QUENCH OFF-GAS
INCINERATED
WASTEWATER
(QUENCH SYSTEM
SLOWDOWN)
Figure B17-1. Wastewater incinerator with quench
B17-3
-------
Appendix B17
Thermal Oxidation
3. Process Performance
Thermal oxidation is capable of achieving essentially complete destruc-
tion of a wide range of otherwise difficult to treat organics. Most organics
can be completely destroyed at 1270 K with one to two seconds residence time
depending upon design of the combustion chamber and the level of mixing.
Tests with various pesticides such as DDT, malathion, and chlordane have shown
99.96 to 99.99 percent destruction (1). Tests at a commercial hazardous waste
incinerator have shown over 99.99 percent destruction of PCBs (5). No data
are available for the incineration of synthetic fuel-derived wastewater
streams.
4. Secondary Waste Generation
Thermal oxidation produces flue gas as a secondary waste stream. The
flue gas will contain carbon dioxide, particulates (primarily inorganic salts
remaining after evaporation of the wastewater), sulfur oxides, nitrogen ox-
ides, and trace amounts of carbon monoxide and volatile organics. The concen-
trations of these latter pollutants will depend on the sulfur and nitrogen
content of the waste as well as the characteristics of the supplemental fuel
and the incinerator operating temperature. Residual organics levels will be
low for a properly operated incinerator, amounting to less than 0.01 to 0.1
percent of the influent organics.
5. Process Reliability
Thermal oxidation is widely used to dispose of a variety of industrial
organic wastes. Incinerators have been operated reliably over a period of
years without a forced shutdown (6).
B17-4
-------
Appendix B17
Thermal Oxidation
Important design parameters affecting incinerator reliability are atom-
ization of the waste, location of atomizers with respect to the primary combus-
tor, mixing of the waste with air, combustion temperature, residence time, and
excess air (7). These variables are interrelated. Rapid and complete mixing
of the waste and the combustion air is necessary to assure efficient and com-
plete combustion. Incomplete mixing may lead to the formation of soot parti-
cles which are difficult to oxidize (8).
The temperature must be high enough to achieve the desired level of
organics destruction. The higher the required temperature, the higher the
heating value of the waste plus that of the auxiliary fuel must be (7).
Higher temperature operation can be achieved either by preconcentrating the
waste so that its heating value per unit weight increases or by providing
additional auxiliary fuel. On the other hand, excessive temperatures can lead
to corrosion and attack on incinerator linings by vaporized metals (9).
Residence time requirements for the destruction of organics vary depend-
ing on the species, the incinerator design, and operating temperature. As
mentioned previously, most toxic organics can be completely destroyed at
1270 K with a minimum residence time of one second; some wastes will require
less time while some of the more resistant wastes will require more (1).
6. Process Economics
Installed equipment costs for a wastewater incineration system can be
estimated by the following equation (Personal communication with Gene Irrgang,
November 2, 1979). These costs were updated to first quarter 1980 dollars by
using the CE index ratio of 1.18.
B17-5
-------
Appendix B17
Thermal Oxidation
C - J5.19 x 10* (Duty)0*4 (1)
where: C = installed equipment cost in first quarter 1980 dollars
and
Duty » heating duty of furnace (10* kcal/hr).
The furnace heating duty is calculated by assuming 2500 kcal/kg of waste-
water incinerated. Costs include structural steel, valves, control panels,
and fuel connections. Compressor and pumps are not included. Figure B17-2
shows installed equipment costs for wastewater incineration versus incinerator
duty.
The major annual utility cost for a wastewater incinerator will be for
supplemental fuel. The fuel requirement can be estimated at 2500 kcal/kg
wastewater (1). Estimated labor costs were not obtained from any of the
available references.
B17-6
-------
010"
g 4
7 2
First Quarter 1980 $
10*
J I
I I I
I i
I
i i i i i i
7 10 2 4 7 100 2
Incinerator Duty, 106 kcal/hr
Figure B17-2. Installed equipment cost for wastewater incineration
(Personal communication with G. Irrgang, November 2,
1979)
B17-7
-------
Appendix B17
Thermal Oxidation
7. References
1. Shen, T.T. , Chen, M., and Lanber, J. Incineration of Toxic Chemical
Wastes. Pollution Engineering, October 1973, pp. 45-50.
2. Fox, R.D. Cost Assessment of Treatment Process for Toxic Wastewaters.
Third Annual Conference on Treatment and Disposal of Industrial Waste-
waters and Residues, Houston, Texas, April 18-20, 1978.
3. Paulson, E.G. How to Get Rid of Toxic Organics. Chemical Engineering,
October 17, 1977, pp. 21-27.
4. U.S. Department of Energy. Processing Needs and Methodology for Waste-
waters from the Conversion of Coal, Oil Shale and Biomass to Synfnels,
DOE/EV-0081, May. 1980.
5. Gregory, R.C. Design of Hazardous Waste Incinerators. Chemical
Engineering Progress, April 1981, pp. 43-47.
6. D.S. Environmental Protection Agency. Sludge Treatment and Disposal,
Chapter 7, Incineration-Pyrolysis of Wastewater Treatment Plant Sludges.
EPA-625/4-78-012, October 1978.
7. Kiang, T.H. Total Hazardous Waste Disposal Through Combustion.
Industrial Heating, December 1977, pp. 9-11.
8. Dunn, K.S. Incineration's Role in Ultimate Disposal of Process Wastes,
Chemical Engineering, October 6, 1975, pp. 141-150.
9. Kiang, T.H. Liquid Waste Disposal System. Chemical Engineering
Progress, January 1976, pp. 71-77.
8. Personal Communications
Irrgang, G. Trane Thermal Company, with S.L. Winton, Radian Corporation,
November 2, 1979.
B17-8
-------
APPENDIX B18
COOLING TOWER OXIDATION
1. Process Description
Cooling tower oxidation is a term which is used to describe a natural
process which occurs when wastewaters with high dissolved organic loadings are
utilized to satisfy either all or part of the makeup requirement for cooling
towers. Two beneficial processes occur when wastewaters are used for cooling
tower makeup. First, the waste stream is reduced in volume due to evaporative
concentration; this results in a smaller stream to be treated further or dis-
posed. Second, reactions with oxygen from the air and the action of aerobic
microorganisms reduce the concentration of biodegradable organics in the
water.
2. Process Applicability
Cooling tower oxidation has been observed in several petroleum refin-
eries, where biological treatment effluent and stripped foul water from cata-
lytic cracking operations have both been successfully treated (1,2,3). The
SASOL plant (Lnrgi gasifiers) has demonstrated the use of gas liquors in a
cooling tower although performance and environmental data are not publicly
available. This process is also being planned for use in the Great Plains
Gasification Project facility which is now under construction (4).
While the use of a cooling tower to oxidize organics and reduce waste-
water volume in a synfuels plant has some potentially significant advantages,
it also presents some potential problems. First, contaminant concentrations
different than those encountered in refinery wastewaters can be expected; phe-
nol concentrations, for example, will probably be higher. Associated with
this process will be an increased potential for fouling of process heat
exchangers by wastewater contaminants such as particulates, inorganics.
B18-1
-------
Appendix B18
Cooling Tower Oxidation
and/or biologically-generated solids. The potential for biological foaling of
heat transfer surfaces has not been well defined (in publicly available data),
but may be high due to the presence of organic/inorganic compounds in the
cooling water that will support biological growth. Chlorination of cooling
water for biological growth control may not be applicable because of the
potential to form toxic chlorinated organics; other biocides may need to be
used in place of chlorine. Conventional cooling water treatment chemicals for
controlling corrosion and scaling may also not be applicable to cooling towers
using process wastewaters as makeup due to interferences caused by organic
compounds.
3. Process Performance
Because of its limited application to date, very limited performance data
are publicly available for this process. In the refinery which used biologi-
cal treatment effluent as cooling tower makeup, reductions in phenolics from
12 mg/L in the influent (makeup water) to 0.09 mg/L in the cooling tower blow-
down were reported. This amounted to a reduction of 99.8 percent based on the
flow rates used in this refinery. BODS reductions of about 80 percent were
also reported (1). In applying this process to a synfuels plant, it can be
anticipated that the main variables which will affect the BOD and COD reduc-
tions achieved, will be the nature of the wastewater (i.e., source of waste-
water and extent of pretreatment) and the contact time in the cooling tower.
A major incentive for using process wastewaters in cooling towers is the
volume reduction which occurs. Conventional cooling towers operate at a mini-
mum of 4 to 5 cycles of concentration. This represents a 75 to 80+ percent
volume reduction.
B18-2
-------
Appendix B18
Cooling Tower Oxidation
4. Secondary Waste Generation
Discharge streams from cooling towers include evaporation, drift, and
cooling tower blowdown. When treated process wastewaters are used as cooling
tower makeup, evaporation will include stripped volatile gases such as ammo-
nia. Cooling tower blowdown will contain the nonvolatile nondegradable makeup
water constituents plus treatment chemicals increased in concentration by the
evaporative action of the tower. Cooling tower drift will have characteris-
tics similar to the tower blowdown.
5. Process Reliability
Cooling tower oxidation has been proven over a number of years to be
successful in handling refinery wastewater (3). Problems of biological oxida-
tion efficiency, heat transfer efficiency, cooling tower disintegration,
solids buildup, and possible odor problems have been addressed (1). Neither
performance nor reliability data for cooling towers in synfuels applications
are publicly available.
6. Process Economics
Costs of treating wastewater via cooling tower concentration and oxida-
tion were not developed. Such costs, as well as potential savings due to
reduced plant demands for raw makeup water depend strongly on the overall
plant wastewater treatment scheme and water balance considerations. Cooling
tower oxidation may, for example, require more extensive upstream treatment to
produce an acceptable quality makeup water. In addition, the impacts on heat
transfer equipment of using process wastewaters as cooling tower makeup cannot
be assessed at this time. Finally, more expensive cooling water treatment
approaches may be required to avoid the formation of toxic byproducts in the
circulating cooling water. Performance data from commerical synfuels systems
will be needed before reasonable cost estimates can be developed.
B18-3
-------
Appendix B18
Cooling Tower Oxidation
7. References
1. American Petroleum Institute. Manual on Disposal of Refinery Wastes,
Volume on Liquid Wastes. API Refining Department, Washington, DC, 1969.
2. Franco, R.J., J.C. Wykowski and J. Delaunay. Reuse of Biologically
Treated Wastewater as Cooling Tower Makeup. Industrial Waste
Engineering, Sept/Oct 1981, pp. 14-19.
3. Mohler, E.F., Jr., and L.T. Clere. Bio-oxidation Process Saves Ha.
Hydrocarbon Processing, October 1973, pp. 84-88.
4. Marion, L. Biological, Mechanical Methods Compete for Wastewater-Cleanup
Job. Chemical Engineering, June 16, 1980, pp. 82-86.
B18-4
-------
APPENDIX B19
CHEMICAL PRECIPITATION
1. Process Description
Chemical precipitation is a process which can be used to reduce the
alkalinity and hardness in a wastewater stream as well as reduce the levels of
dissolved metals. The key to this process is the addition of a chemical which
will promote the precipitation of a salt containing the desired component(s).
The precipitated solids can then be removed by settling and/or filtration.
The most common use of chemical precipitation is in lime-soda softening.
A softener can lower hardness (dissolved calcium and magnesium) and adjust al-
kalinity, by adding lime and soda ash in the correct proportions. Various ad-
ditives may also be used to improve softener performance. Coagulants and
flocculants are frequently added to aid in the agglomeration of fine solids,
and acids are often used to adjust the pH of the treated effluent water.
Lime-soda softening is widely used in industry for the treatment of both
raw waters and wastewaters. Figure B19-1 shows a typical industrial lime-
soda softening system. Influent wastewater is thoroughly mixed with lime
slurry and soda ash, which causes the pH to rise. Lime-soda softeners typi-
cally operate in a pH range of 9.8 to 10.5. Under these conditions, a variety
of compounds can exceed their solubility product constants although the pro-
cess is most often applied to reduce hardness and alkalinity by precipitating
calcium carbonate and magnesium hydroxide:
Ca++ + 2HC(>7+ CaCOH), = 2CaCO, + 2HaO (1)
Jig** + 2HC07 + 2Ca(OH)1 = 2CaCO, + Mg(OE)a + 2H,0 (2)
Ca++ + Na2CO, » CaCO, + 2Na+ (3)
Mg++ + NaaCO, + Ca(OH)1 = CaCO, + Mg(OH)a + 2Na+ (4)
B19-1
-------
INITIAL pH
ADJUSTMENT
SODA ASH
FEEDER
tri
WASTE WATER
FEED
0
LIME
FEEDER
D
MIX TANK
CLARIFICATION
0
POLYMER
FEEDER
CLARIFIER
SLUDGE
TO
DISPOSAL
FILTRATION
DEEP
BED
FILTER
FINAL pH
ADJUSTMENT
ACID
FEEDER
0
D
MIX TANK
TREATED
WATER
Figure B19-1. Process flow diagram for lime/soda softening system
-------
Appendix B19
Chemical Precipitation
Calcium sulfate, calcium phosphate, and some silica compounds can also preci-
pitate in the softener. Furthermore, some trace metals become insoluble at
high pHs and precipitate as hydroxide salts. Other trace metals can be
adsorbed on the surface of the solids.
The solids that precipitate in a chemical precipitation process are fre-
quently very fine and difficult to dewater. A clarifier is normally used for
solids removal, followed by a gravity-flow granular bed filter to remove
residual solids. The gravity filter effluent will normally need pH adjustment
before release to downstream processes. Recycle of the clarifier underflow is
used to provide the seed crystals necessary to sustain reasonable precipita-
tion rates. Clarifier underflow blowdown may be further dewatered if desired
with a device such as a rotary drum vacuum filter.
Another commonly used chemical precipitation process is one designed to
remove chromate-based corrosion inhibitors from cooling tower blowdown. In
this application, hexavalent chromium is reduced to trivalent chromium with a
reducing agent such as sulfur dioxide under low pH conditions. The pH of the
water is then raised to around 8.5 using lime or caustic. The trivalent chro-
mium precipitates as chromium hydroxide. The sludge produced is very diffi-
cult to dewater and must be disposed of as a hazardous waste.
2. Process Applicability
In principle, chemical precipitation may be used to remove any contami-
nant which will form a relatively insoluble compound upon the addition of a
reagent. Lime-soda softening is by far the most common chemical precipitation
process used in industry. In synfuels facilities another conventional pro-
cess, chromate removal, may find some limited applications in treating cooling
tower blowdown.
B19-3
-------
Appendix B19
Chemical Precipitation
Lime-soda softening may be applied to the treatment of hard makeup waters
before they are used within a plant. In this application softening is gener-
ally used to remove species which can cause chemical scaling, such as calcium
and silica. In wastewater treatment, lime-soda softening may be used to
remove trace elements and other inorganic species such as calcium, magnesium,
carbonate, silicate, phosphate, and sulfate. Potential problems related to
interferences by complezing agents or precipitation inhibitors should be care-
fully considered in evaluating new applications of chemical precipitation pro-
cesses, particularly those involving the treatment of relatively complex
wastewaters.
3. Process Performance
The performance of a lime-soda softener can vary widely, depending on
influent composition, reagent dosages, and equipment configuration. For cold
lime-soda softening, effluent calcium hardness is typically 50 mg/L (expressed
as CaCOj), and effluent nonsodium alkalinity is typically 35 mg/L (as CaC03)
(1). For cold lime-soda softening followed by sand filtration and activated
carbon polishing, typical removal levels for selected trace elements are:
Element
Cadmium
Chromium
Copper
Mercury
Nickel
Lead
Selenium
Zinc
% Removal
Reference 2
0
88
40
25
45
2
33
71
Reference 3_
Not Reported
35
50
Not Reported
40
27
Not Reported
36
B19-4
-------
Appendix B19
Chemical Precipitation
4. Secondary Waste Generation
The major waste produced by chemical precipitation is the precipitated
solids stream. For cold lime-soda softening, the waste solids consist primar-
ily of calcium carbonate and magnesium hydroxide. Smaller amounts of calcium
sulfate dihydrate (gypsum) and calcium phosphate may be present. Silica may
also be present in the magnesium hydroxide. If trace metals are removed in
the process, they will be present in the waste solids, usually as hydroxide
salts. If large amounts of trace elements are present, the solids may be con-
sidered hazardous, requiring special disposal methods.
The waste solids generated in a chemical precipitation process are fre-
quently composed of very fine particles which are difficult to dewater. Gra-
vity separation will produce a stream containing up to about 15 wt % solids.
Further concentration may be achieved using filtration, centrifngation, or
evaporation. A typical dewatered sludge would contain 45 wt % solids (4).
5. Process Reliability
Lime-soda softening is a chemical precipitation process that is widely
used in industry and is generally very reliable. Pumps and mixers are the ma-
jor pieces of moving equipment, and in these applications, they are reliable
and easily spared. Chemical scaling can occur at some points in the system,
but careful design should avoid these problems. Rotary drum vacuum filters
are reliable in most applications, as are gravity filters. Because the preci-
pitated solids are extremely fine, clarifier performance is sensitive to vari-
ations in operating parameters. Temperature fluctuations, for example, may
cause variations in particle size and floe size and upset clarifier opera-
tions. Flowrate changes and variations in reagent dosage can also cause
upsets.
B19-5
-------
Appendix B19
Chemical Precipitation
6. Process Economics
Capital, operating, and maintenance costs for a lime-soda system can be
estimated from published data on wastewater treatment. A lime-soda system for
synfuels wastewater treatment would involve the following equipment:
• Reagent addition,
• Solids precipitation and separation - a clarifier,
• Sludge dewatering - a rotary vacuum filter, and
• Final polishing - a gravity fed granular bed filter, and pH
adjust tank.
Equipment for additive feeding and initial mixing is usually combined with the
clarifier. Thus, a lime-soda softener system can be costed as the following
components:
Polymer flocculant storage mixing and feeding;
Lime storage, slaking, and feeding;
Soda ash storage, mixing, and feeding;
Softening and clarifying;
Effluent filtration (gravel bed filter);
Effluent pH adjustment (snlfuric acid addition); and
Sludge dewatering (a rotary drum vacuum filter).
Two of these cost components have been addressed in other appendices -
polymer flocculant addition in Appendix B2 and gravel bed filters in Appendix
B4. Costs for the other five components are presented below (5).
Installed equipment costs for lime equipment are shown in Figure B19-2.
The lowest feedrate data point (4.5 kg/hr) assumes the use of hydrated lime
shipped in bags, while the higher feedrate data points assume bulk quicklime
shipment with onsite slaking. In some raw water treatment systems, softening
wastes can be recalcined to reduce lime usage. This option was not considered
in the operating and maintenance cost estimates shown in Figure B19-3.
B19-6
-------
10"
First Quarter 1980 S
a 2
- 10
- 7
10'
7 10 2 4 7 100 2
Lime Feed Rate, kg/hr
Figure B19-2. Installed equipment cost for lime systems (5)
B19-7
-------
ioor~
7 -
4 -
2 -
10
First Quarter 1980 $
Labor
o
o
o
7 -
4 -
0 1
terials
0.1
I I 1 I L 1 I I
J I I 1111
lectricity
j i i I i i I i
7 10 2 4 7 100
Lime Feed Rate, kg/hr
Figure B19-3. Operating and maintenance costs for lime systems (5)
B19-8
-------
Appendix B19
Chemical Precipitation
Costs for storing, mixing, and feeding soda ash were adapted from pub-
lished data on ferrous sulfate feed systems. The cost basis assumes mild
steel bulk storage, mixing to a 6 percent solution, and feeding via a metering
pump. Figure B19-4 shows estimated installed equipment costs, and Figure
B19-5 shows estimated operating and maintenance costs.
Figures B19-6 and B19-7 show installed equipment costs and operating
and maintenance costs for circular gravity clarifiers. The power requirements
shown assume a lime sludge, and are somewhat higher than for clarifiers used
in some coagulation/clarification applications.
Figures B19-8 and B19-9 show installed equipment cost and operating and
maintenance costs for a system feeding concentrated (93 percent) sulfuric
acid. The lowest flow rate data point assumes the acid is delivered in drums,
while the higher flowrate cases assume bulk acid delivery.
Figures B19-10 and B19-11 show installed equipment costs and operating
and maintenance costs for a vacuum filter system. The design basis assumes an
inlet slurry containing 1 wt % solids, a solids loading of 8.3 kg/hr per m2,
and a filter cake containing 20 wt % solids.
Costs for a complete lime-soda softening system will be very dependent on
the water to be treated and the desired treatment level. As an example, costs
have been calculated for raw water with the following composition:
Calcium Hardness : 200 mg/L (expressed as CaCO,)
Magnesium Hardness: 100 mg/L (expressed as CaC03)
Alkalinity : 150 mg/L (as CaCO,)
Reagent dosage was calculated to achieve the following effluent composition:
Calcium Hardness : 35 mg/L (as CaCO,)
Magnesium Hardness: 90 mg/L (as CaC03)
Alkalinity : 150 mg/L (as CaCO,)
B19-9
-------
10
First Quarter 1980 $
e
.S" 2
10
1 I I I I
7 10
7 100
Soda Ash Feed Rate, kg/hr
Figure B19-4. Installed equipment cost for soda ash feed systems (5)
B19-10
-------
Operating Costs, $/1000 kg
to
M
VD
I
Tl
cro
(B
td
I
Ln
O
•d
ro
s-
OP
O
o
en
i-n
O
i-!
U)
O
(D
(D
CL
tn
rt
f
-------
10"
4
First Quarter 1980 $
103
7
2 -
102
_L
10 2 4 7 100 2 47 1000 2 47
Clarifier Area, m2
Figure B19-6. Installed equipment cost for circular clarifiers
B19-12
-------
0.1
Clanfier Area, m2
Figure B19-7. Operating costs for circular clarifiers (5)
BJ9-13
-------
10*
7
103
First Quarter 1980 S
~ 2
102
10
10
100 2 4
Acid Feed Rate, kg/hr
7 1000
Figure B19-8.
Installed equipment cost for sulfuric
acid feed systems (5)
B19-14
-------
100
7
10
o
I 4
u 2
First Quarter 1980 $
0.1
10
Labor
\
Mat
erials
Electriciqy
I i i i
7 100 2 47 1000 2
Acid Feed Rate, kg/hr
Figure B19-9. Operating costs for sulfuric acid
feed systems (5)
B19-15
-------
10-
First Quarter 1980 $
10"
I I I I I I I I
I I I I I I I I
1 t lit i L
7 10
7 100
Filter Area, m2
Figure B19-10. Installed equipment cost for vacuum filters (5)
B19-16
-------
10
100
Filter Area, m2
Figure B19-11. Operating costs for vacuum filters (5)
E19-17
-------
Appendix B19
Chemical Precipitation
For this case, an industrial water treatment manual (6) recommends the fol-
lowing reagent dosage:
Lime - 127 g/m» of 93 percent hydrated lime
or 100 g/m1 of 90 percent quicklime
Soda Ash - 65 g/ml 98 percent soda ash
In addition, 1 mg/L polymer flocculant was assumed to be added. This treat-
ment would result in a calcium carbonate/magnesium hydroxide sludge and would
produce approximately 0.35 kg (dry basis) of sludge per m* of water treated.
It was assumed the clarifer design rise rate would be 30.5 cm per hour and
that the vacuum filter design criteria would be the same as those described
earlier. The acid feed system was costed by assuming that 60 mg/L of sulfate
would be added (61 mg/L of sulfuric acid).
Figure B19-12 shows estimated installed equipment costs for:
• Softening (lime, soda ash, polymer, and clarifier systems),
• Effluent polishing (gravity filter and pH adjust or acid feed), and
• Sludge dewatering (vacuum filter system).
Figure B19-13 illustrates total operating and maintenance costs (excluding
chemical costs for lime, soda ash, polymer, and acid) for this example case.
B19-18
-------
Js 10*
E
First Quarter 1980 $
§• 2
U
-a
w
2 io3
w
c
7
jftening Equipment
J I
Polishing Equipment
Sludge Dewatering
Equipmenc
I i i i i i i
i i i i i
10
7 100
7 1000
Raw Water Flow Rate, ra'/hr
Figure B19-12. Installed equipment costs for raw water
softening example systems
B19-19
-------
Operating Costs, $/1000 m3
•- fo «. ^
I
ro
o
H-
TO
i-t
0>
VO
I
t fD
i-l
Z W
W rt
rt H-
(0 3
t-i CW
en o
O O
i-h tn
rt rt
m to
3
3' nT
OQ X
n
ID M
X C
03 P-
T3 3
I-1 CW
ft)
i-(
cn re
«< pi
w OP
rt re
ffi 3
tn en
Ml
O
1-1
e
c
I I I
-------
Appendix B19
Chemical Precipitation
7. References
1. Stearns-Roger Engineering Corporation and Radian Corporation. Design and
Operating Guidelines Manual for Cooling-Water Treatment. EPRI CS-2276,
Project 1261-1, March 1982.
2. Esmond, Steven E. and Albert C. Petrasek, Jr. Trace Metal Removal.
Ind. Water Eng, Vol. II, No. 3, May-June 1974, p. 14-17.
3. Zemansky, Gilbert M., Removal of Trace Metals During Conventional Water
Treatment. J. Am. Water Works Assoc., Vol. 66, No. 10, Part 1, October
1974, p. 606-609.
4. U.S. Environmental Protection Agency, Technology Transfer. Process
Design Manual for Sludge Treatment and Disposal. EPA 625/1-74-006,
October 1974.
5. U.S. Environmental Protection Agency. Estimating Water Treatment Costs.
Volume 2, Cost Curves Applicable to 1 to 200 Mgd Treatment Plants. EPA
600/2-79-162b, August 1979.
6. Water and Waste Treatment Data Book. Permutit, Inc., 1961.
B19-21
-------
APPENDIX B20
ION EXCHANGE
1. Process Description
The ion exchange process involves a reversible interchange of ions in
solution with other ionic species bound to a solid ion-exchange medium. A
number of natural and synthetic ion exchange media are available for the
removal of both cations (positively charged ions) and anions (negatively
charged ions). Such media are usually in the form of granules or beads,
ranging in size from about 0.3 mm to 1.0 mm in diameter (1).
An example of ion exchange is the softening of water by the use of zeo-
lites. This naturally occurring cationic exchange medium is used to replace
the hardness-producing calcium and magnesium ions with sodium ions in the
treated water. As the active sites on the ion-exchange medium become filled
with the removed species, the removal efficiency of the unit drops off. The
exchange medium is then regenerated by contacting it with a strong solution
containing the original cation; in the example, the zeolites are regenerated
with a strong brine solution replacing the sodium and displacing the calcium
and magnesium. Strong acid or caustic solutions are also frequently used to
regenerate cationic and anionic exchange resins, respectively.
Ion exchange equipment includes packed beds of ion-exchange media, hold-
ing tanks for the stream to be treated, regenerant solution handling equip-
ment, and pumps. Piping and valving are arranged so that some beds are on-
line while others are being regenerated to allow for continuous operation. In
general, ion exchange treatment of a stream containing a number of different
cations and anions will require the use of both cation-exchange and anion-
exchange resins using either separate beds in series or mixed-media beds. The
ion exchange process is a competitive process in which those species having
the highest equilibrium selectivity with respect to the given exchange medium
are removed before the other species in solution (1,2). Solution pH also can
B20-1
-------
Appendix B20
Ion Exchange
have a large impact on the ionic form various species take in solution and,
therefore, influences their uptake on a given resin. For example, chromium,
although it is a metal and might be expected to be present as a positively-
charged ion, may really be present as chromate or dichromate ions (Cr04 ,
Cra07) at a pH of 4 to 5 and would be removed by anionic resins.
2. Process Applicability
Ion exchange has been used in a number of ion-removal and recovery appli-
cations including potable water production, boiler feed water polishing,
wastewater treatment, and recovery of chromates from cooling tower blowdown
(1,3,4). Wastewater treatment applications include removal of a number of
heavy metals, chloride, ammonium, and nitrate ions and ionized sulfur species
such as sulfate, sulfite, and sulfide (4,5,6,7,8).
Ion exchange as a technique for the reduction of total dissolved solids
(TDS) is applicable for streams having TDS concentrations less than 5000 mg/L
(8). In this application, cations and anions are replaced with hydrogen and
hydroxyl ions which then form water. In order to avoid plugging or fouling of
the resin, the stream to be treated must have very low levels of suspended
solids, dissolved organics, and oil and grease. The TDS level for chromate
removal by ion exchange is cited as 2 ppm or less (1). High temperatures and
varying concentrations can contribute to resin performance problems including
physical degradation (9).
3. Process Performance
A properly designed and operated ion exchange system is capable of pro-
ducing a very high quality effluent for boiler feedwater makeup and other
systems requiring high purity water. Ion exchange is also successful in
removing chromate from cooling tower blowdown for recycle; 90 percent removal
is achievable (10).
B20-2
-------
Appendix B20
Ion Exchange
The process is reported to be less successful in industrial wastewater
treatment. Irreversible fouling and poor removal efficiency have been
reported (8). Removal efficiencies depend on the affinity of the resin for
the adsorbed ions; since there is competition for the available active sites,
removals of certain species may be high while others are low. Extremely high
affinities of some species may cause some of the resin activity to be perma-
nently lost since regeneration cannot remove them. Clogging of the bed due to
the presence of suspended solids which are filtered out by the resin bed can
cause additional problems.
4. Secondary Waste Generation
Regeneration of the ion exchange resin produces the only secondary waste
stream from the process. This stream will contain all of the species which
were reversibly adsorbed from the influent stream. Since the volume of the
regenerant stream is considerably less than that of the influent stream, the
contaminants will be present in greater concentrations. In addition, the
stream may have either a high or low pH, if the resin is regenerated with
strong base or acid. The overall volume of this stream depends on the regen-
eration frequency.
5. Process Reliability
Ion exchange has been successfully utilized in water softening and boiler
water deionization, in the treatment of metal-plating rinse waters, and in
chromate recovery from cooling tower blowdown. It has been characterized as
having a moderate degree of reliability for ammonium ion removal (5). Appen-
dix B21 discusses clinoptilolite ion exchange removal of ammonia. Quantita-
tive reliability data in terms of on-stream availability factors are not
available.
B20-3
-------
Appendix B20
Ion Exchange
6. Process Economics
Cost data were obtained for two applications of ion exchange technology:
TDS reduction and water softening with cationic resins. Capital costs for IDS
reduction were obtained from Reference 8, as a function of both waste stream
flow and TDS concentration. The capital costs were converted to installed
equipment cost by dividing by 1.48, assuming that the capital cost components
of engineering and construction, fees, and contingency together comprise 48
percent of the installed costs. The costs from Reference 8, given in 1970
dollars, were updated to first quarter 1980 dollars by multiplying by the CE
index ratio of 2.06. The installed equipment costs are presented in Figure
B20-1. Total operating and maintenance costs were available for a single
concentration, 3000 mg TDS/L, from Reference 8. They were also updated by the
same CE index ratio and are presented in Figure B20-2.
Installed costs for an ion exchange system for reducing water hardness by
300 mg/L, with brine regeneration, were obtained from Reference 6. The esti-
mated installed equipment cost included fabricated steel contact vessels with
baked phenolic lining, regeneration facilities, and pumping system. A CE
index ratio of 1.23 was used to update the costs from January 1978 to first
quarter 1980 dollars. Figure B20-3 presents these costs as a function of flow
rate. Estimated operation and maintenance requirements were taken from the
same source and do not include brine disposal costs. Electrical requirements
are for regenerant pumping, rinse pumping, and backwash pumping; feed water
pumping costs were not included. Materials costs included resin replacement
and periodic repair and replacement parts. Costs of regenerant brine were not
included. Labor requirements included both operation and maintenance. Figure
B20-4 is a plot of these operating costs versus flow rate. Electrical costs,
estimated at &0.049 per thousand m3 of feed water, are not shown.
B20-4
-------
10,000
1,000
100
1000
Wastewater Flow Rate, mVhr
Figure B20-1. Installed equipment costs for IDS
reduction by ion exchange (8)
B20-5
-------
100
First Quarter 1980 $
S 10
Operating & Maintenance
(3000 mg/L IDS)
4 -
2 -
_L
1
1 i I I i
10
100 2 47 1000
Wastewater Flow Rate, mVhr
Figure B20-2. Operating cost for IDS reduction by ion exchange (8)
B20-6
-------
1,000
First Quarter 1980 $
u
"2 100
10
I I I I I I I
J I I I I I
10
7 100 2 47 1000
Wastewater Flow Rate, aiVhr
Figure B20-3. Installed equipment cost for ion exchange
water softening systems (8)
B20-7
-------
O
I
00
TO
C
i-l
0
w
NJ
o
I
CO O
o -a
fD D
3 rt
H- H-
3 3
TO TO
CD O
•<; o
01 CO
rt rt
fD cn
W Ml
t»
(-"•
O
3
fD
X
o
I
p>
rt
fD
Operating Costs, $71000 m3
>- ro ^ ^i
I I I I I I
T
T
-------
Appendix B20
Ion Exchange
7. References
1. American Petroleum Institute. Manual on Disposal of Refinery Wastes,
Volume on Liquid Wastes. API Refining Department, Washington, DC, 1969.
Section 10.
2. Kennedy, D.C. Predict Sorption of Metals on Ion-Exchange Resins.
Chemical Engineering, June 16. 1980, pp. 106-118.
3. Blackburn, J.W. Removal of Salts from Process Wastewaters. Chemical
Engineering, October 17, 1977, p. 38.
4. Lanonette, K.H. Heavy Metals Removal. Chemical Engineering, October 17,
1977, p. 38.
5. U.S. Environmental Protection Agency. Innovative and Alternate Technol-
ogy Assessment Manual. EPA-430/9-78-009, February 1980, pp. A-124-A-125.
6. U.S. Environmental Protection Agency. Estimating Costs for Water
Treatment as a Function of Size and Treatment Plant Efficiency.
EPA-600/2-78-182, August 1978, pp. 213-232.
7. Watkins, J.P. Controlling Sulfur Compounds in lasteraters. Chemical
Engineering, October 17, 1977, p. 65.
8. Patterson, J.W. Wastewater Treatment Technology. Ann Arbor Science
Publishers, Inc., Ann Arbor, Michigan, 1975.
9. Lacey, R.E. Membrane Separation Processes. Chemical Engineering,
September 4, 1972, pp. 56-74.
10. Richardson, E.W. , E.D. Stobbe and S. Bernstein. Ion Exchange Traps
Chromates for Reuse. Environmental Science and Technology, 2 (11),
November 1968, pp. 1006-1016.
B20-9
-------
-------
APPENDIX B21
ION EXCHANGE REMOVAL OF
AMMONIA USING CLINOPTILOLITE
1. Process Description
Clinoptilolite is a naturally occurring ion exchange material that can
selectively exchange sodium or calcium ions bound to its surface for ammonium
ions in a contacting solution. The ammonia-bearing waste stream is usually
contacted with the clinoptilolite on a continuous basis in either a downflow
column or in a horizontal pressure vessel. Upon initial contact, the fresh
clinoptilolite removes nearly all ammonia from the waste stream. With contin-
uing contact, the concentration of ammonia in the effluent stream exceeds a
predetermined, acceptable level, and the column is replaced by one that has
been regenerated.
Clinoptilolite columns are regenerated by contacting the ion exchange
resin with a brine solution. Ion exchange is an equilibrium between pairs of
bound and aqueous ions. Ammonium ions are normally attached to the clinoptilo-
lite structure preferentially to sodium, but under high concentrations of
sodium ions in the brine, a large enough driving force exists to displace the
ammonium ions.
Following regeneration, the regenerant is processed for reuse and to
recover the ammonia for final disposal or use elsewhere. Various methods of
processing can be used including closed loop stripping/adsorption, steam
stripping, biological treatment, and electrolytic treatment (1). Closed loop
stripping/adsorption is used in two large scale applications (Upper Occoquan
Sewage Authority and Tahoe-Trucker Sanitation Agency) and shows promise over
the other systems. Steam stripping is usually economical only if less than
four bed volumes of regenerant are required for each cycle of the clinoptilo-
lite columns. Biological treatment is not well demonstrated on a large scale
and is sensitive to low temperature. In electrolytic treatment ammonia in
B21-1
-------
Appendix B21
Clinoptilolite Ion Ex.
the regenerant is converted to nitrogen gas by reaction with chlorine that has
been generated electrolytically from chlorides already present in the stream.
This method requires electrical power requirements approximately ten times
those needed for air stripping.
In the closed loop stripping/adsorption process, the ammonia is first
separated from the regenerant by air stripping. The offgas is routed through
an adsorption tower where the ammonia is absorbed in sulfur acid and recovered
as ammonium sulfate. The ammonia-free air is returned to the stripping sec-
tion to complete the circuit.
More energy is required for a closed loop system than for a single pass
air stripper, but this disadvantage is usually offset by two considerations
that normally limit the utility of single pass air stripping processes.
First, since no outside air is used in the stripping process, freezing during
cold weather operation is normally not a problem. Second, scaling is minimal.
The carbon dioxide in the stripping air that is responsible for scale genera-
tion is eliminated from the system within the first several cycles of opera-
tion.
2. Process Applicability
The ion exchange process may be represented by the following equation:
bNH* + B-Z = bNH*-Z' + B+b (1)
where: Z', Z = clinoptilolite exchange sites for NH4 and B ,
respectively, and
b = numerical value of the charge of ionic species B.
B21-2
-------
Appendix B21
Clinoptilolite Ion Ex.
Ions other than NH4 which may be present in the wastewater will compete with
ammonium ions for the exchange sites so that the process described by equation
1 should be viewed as selective, not specific for NH4. The selectivity of
clinoptilolite for various cations has been quantified (2), and less pre-
cisely, the order of preference of clinoptilolite for common ions is as fol-
lows (3):
Cs+ > Rb+ > K+ > NH+ > Ba++ > Sr++ > Na+ > Ca++ > Fe+++ > Al+++ >
4
Mg++ > Li+
The equilibrium capacity of clinoptilolite for ammonia is most directly
influenced by the ammonia concentration and the relative concentrations of
other cations in solution. Other cations, both those preferred and less pre-
ferred to ammonia, compete for exchange sites on the exchange material that
could otherwise be used to remove ammonium ions. The maximum exchange capaci-
ty of clinoptilolite for all ions has been measured by several investigators.
Measured capacities vary somewhat with the technique used: a list compiled by
Koon and Kaufman (3) shows an average of about 1.8 milliequivalents (meq) per
gram.
Each cationic species in the waste stream occupies a fraction of the
total exchange capacity. In addition, it has been observed in experimental
studies that the equilibrium between any cationic pair in a multicomponent
system is independent of all other ion pairs (4). One method of estimating
the ammonia capacity of clinoptilolite was developed by McLaren and Farquhar
(4).
The fraction of the total ammonia exchange capacity available under equi-
librium conditions that can actually be accessed in an operating system is
determined by the kinetics of ion exchange. For a downflow column system,
most of the exchange reactions occur in a well defined zone. This zone is a
B21-3
-------
Appendix B21
Clinoptilolite Ion Ex.
transition in the column between upper resin layers that are saturated and
lower layers having much greater capacity remaining, possibly full capacity.
With continuous loading, the portion of the column that is fully saturated
increases, forcing the transition zone to move downward through the column.
When the zone reaches the bottom of the column, ammonia breakthrough in the
effluent occurs; that is, the effluent concentration of ammonia increases
continuously from a small leakage value to the concentration in the influent.
The suddenness of breakthrough is determined by the characteristics of the
exchange zone which are determined by exchange kinetics.
The rate of ion exchange is controlled by one or more of the following
processes (5):
• film diffusion,
• fluid phase pore diffusion,
• reaction at liquid/solid interface, and
• solid phase internal diffusion.
The exchange reaction at the liquid/solid interface is usually very rapid and
therefore not rate controlling. Film diffusion may control the kinetics of
ion exchange when the media particles are small, when the superficial velocity
through the column is small, and/or when the solution phase concentration is
small. Pore diffusion is most likely rate limiting when media particles are
large and/or solid phase concentrations are near equilibrium values. Solid
phase diffusion may also be limiting when solid phase concentrations are near
equilibrium.
Various of the above rate processes may be controlling at different loca-
tions and times in the column. For most systems, the zone of most active
exchange is controlled by either film diffusion or pore diffusion (5). For
high solution concentrations, film diffusion tends to dominate pore diffusion.
B21-4
-------
Appendix B21
Clinoptilolite Ion Ex.
With increasing flow, the transfer rate increases for film diffusion whereas
pore diffusion is unaffected by flow. For both film and pore diffusion, rate
of transfer increases as the media particle size decreases.
Various expressions can be written for each of the possible controlling
rate processes to predict ion exchange performance. Usually this approach is
workable only for influents whose chemical characteristics are very well
defined. The most commonly used design tool consists of exhaustion curves gen-
erated from bench- or pilot-scale studies of the subject waste stream.
3. Process Performance
1
Ammonia removal efficiencies are only slightly affected by varying the
rate of column loading between 7.5 and 20 bed volumes (BV) per hour (6).
Decreased performance has been observed at loadings greater than 20 BV/hr for
20 x 50 mesh media grains (3).
Experimental data confirm that an improved rate of exchange is associa-
ted with smaller media grains (i.e., large surface areas per volume of media)
as expected when either film or pore diffusion controls. For 20 x 50 mesh
grains, exchange performance decreases between 20 and 30 BV/hr whereas per-
formance does not decrease until greater than 40 BV/hr with 50 x 80 mesh.
However, with smaller media grains, the pressure drop across the column is
higher. A grain size of 20 x 50 mesh is suggested for acceptable ammonia
removal without excessive pressure loss (1).
Ammonia exchange is strongly influenced by pH since the exchange reaction
occurs only with the ionized ammonia species. At high pH, less of the total
+
ammonia is available as NH4. However, at a pH that is too low, ammonia
B21-5
-------
Appendix B21
Clinoptilolite Ion Ex.
exchange is inhibited by competition with hydronium ions for exchange sites.
Optimum operating pH is about 6 with performance dropping off significantly
outside a pH range of 4-8 (3). Temperature has been found to affect ammonia
removal only slightly (1).
No known experience exists for clinoptilolits-based ion exchange except
for streams containing low level concentrations of ammonia. A limited EPA-
supported study was performed and showed promise for removal of ammonia at
concentrations of about 200 mg/L (7). Data presented herein represents the
best performance observed, but nonetheless should provide a conservative basis
for initial design estimates since a less than optimal, large grain size (20 x
35 mesh) was used in the experimental columns. Operating systems using the
standard 20 x SO mesh grain size should perform even better.
Figure B21-1 presents a partial exhaustion curve for a column loaded at 7
BV/hr. Figure B21-2 presents exhaustion curves for two columns operating in
series loaded at 12 BV/hr: the influent to the second column (3B) is the ef-
fluent from the first (3A). No significantly improved performance is realized
when the rate of loading is decreased from 12 to 7 BV/hr. Breakthrough for
column 3A, Figure B21-2 occurs at about 30 BV at the 1 mg/L level and at about
41 BV at the 20 mg/L level.
Once initial breakthrough at the 1 mg/L level occurs, the slope of the
exhaustion curve increases rapidly but then decreases gradually to zero. This
shape indicates that for applications where only low level effluent concentra-
tions can be tolerated, much unused capacity remains in the column when it is
removed from service. In some cases, greater utilization can be had by oper-
ating columns in series. When the ammonia concentration in the effluent for
the series exceeds the permissible level, the lead column is almost completely
saturated. At this point the lead column is removed from service and is
replaced by a fresh column added to the end of the series. The interval
B21-6
-------
K3
M
I
i-t
(D
6)
W
It
H-
O
0
O
M
(D
H-
S3
o
•a
It
H-
I-1
O
M
H-
It
n>
o
o
o
p
CL
fD
EFFLUENT AMMONIA CONCENTRATION, mg/L
S S § 8
to
s
-------
•n
H-
CR
c
f-i
fD
I
to
EFFLUENT AMMONIA CONCENTRATION, mg/L
CD
o
ro
O
8
C
en
o
3
O
C
1-1
(1)
H-
M
O
ft
H-
M
CD
n
o
t-"
i
w
en
fD
fD
cn
<
n
-------
Appendix B21
Clinoptilolite Ion Ex.
between removal of consecutive columns from the series is determined by the
period between equivalent breakthrough points on the exhaustion curves in
Figure B21-2. Comparable 1 mg/L breakthroughs occur at 30 and 79 BV/hr for
columns 3A and 3B respectively, and at 41 and 89 BV/hr at 20 mg/L break-
through. A period of approximately 49 BV can therefore lapse between removal
of consecutive columns in a series configuration.
Ion exchange systems with columns connected in series are relatively
complex from an operational standpoint. Intricate piping and control networks
are required to remove and replace columns in constantly changing relative
positions. Parallel combinations of columns are more commonly used. Some
increased utilization of the total ammonia capacity of the column and greatly
simplified operation are possible. However, the interval between consecutive
removals of columns from the system is generally shorter than that for a ser-
ies operation.
In a parallel configuration, each column operates at a different phase in
its exhaustion cycle, and the effluents are combined into a composite flow.
The effluent concentration of each column just before it is removed from ser-
vice can greatly exceed the level permissible in the composite stream. This
is because the ammonia contributed by this column to the composite is balanced
by others operating in initial stages of their cycle and contributing very
little ammonia. The interval between consecutive removal of columns from the
parallel system is determined by both the number of columns and the service
life of each in the system. For a parallel system of four columns loaded with
an influent containing 200 mg/L, each column would have a service life (time
between regenerations) of about 50 BVs while maintaining a composite effluent
concentration of less than 20 mg/L. The interval between consecutive removals
of columns from the system would be about 12.5 BVs.
B21-9
-------
Appendix B21
Clinoptilolite Ion Ex.
Experimental studies have shown little difference in regeneration per-
formance when columns are loaded at 4 to 20 bed volumes per hour with a 0.1
molar sodium chloride solution at pH 11.5. For loadings of 20 to 30 BV/hr,
exchange kinetics were found to be limiting. Similar results were obtained
with regenerants of 0.5 molar NaCl at pH 11.5 and 0.2 molar NaCl at pfl 12.5.
For the conditions studies, all ammonia was eluted with use of less than 30 BY
regenerant (3).
Somewhat poorer performance resulted for regeneration of clinoptilolite
that had been loaded with a wastewater containing about 200 mg/L ammonia (7).
Figure B21-3 illustrates an exhaustion curve for one regeneration determined
by the EPA study (7). The regeneration consisted of 0.35 molar NaCl at pH 12,
and the column was contacted in an upflow configuration at 18 BV/hr. After
passing 30 bed volumes of regenerant, 94 percent of the ammonia was removed;
after 47 bed volumes, 99.5 percent was removed. Essentially the same perfor-
mance was obtained with a regenerant of 0.2 molar NaCl at pH 12.
Ammonia is more easily displaced by sodium during regeneration at high pH
than at low. This results primarily from the pH dependence of the NH4/NH3
equilibrium. However, at high pH regeneration, precipitation of calcium car-
bonate or magnesium hydroxide with subsequent fouling of equipment can occur.
An additional rinse cycle is also required following regeneration to return
the pH to within the optimal operating range. Reference 1 recommends regener-
ating columns at near neutral pH with a 2 percent (0.35 molar) sodium chloride
solution. A larger volume of regenerant is required, but 30 to 40 BY should
be sufficient.
4. Secondary Waste Generation
A secondary waste stream is generated by the closed loop regenerant
recovery system. The waste stream is a blowdown to control dissolved solids.
B21-10
-------
1200 r
REGENERATION: 0.35 molar NaCI at pH 12
THROUGHPUT, BED VOLUMES (BV)
Figure B21-3. Regeneration of clinoptilotile (7)
B21-11
-------
Appendix B21
Clinoptilolite Ion Ex.
Additionally, ammonia is recovered as ammonium sulfate and can be collected
for sale as chemical fertilizer or disposed of. Many systems may also require
a backwashing step prior to regeneration to remove suspended solids trapped in
the clinoptilolite. The size of each of these streams is not certain and
would have to be determined on an individual basis.
5. Process Reliability
Clinoptilolite-based ion exchange systems have not been utilized except
to remove low level concentrations of ammonia from municipal wastewaters.
However, these applications are successful, and no major reliability problems
have been reported. Similar results are expected for applications to waste-
waters having somewhat higher concentrations of ammonia, in the few hundred
mg/L range.
6. Process Economics
Reference 1 summarizes the estimated costs for a municipal plant operated
by the Upper Occoquan Sewage Authority. The clinoptilolite system was
designed for a maximum flow of 85,200 mVday and influent ammonia concentra-
tion of 20 mg/L. Eight clinoptilolite beds operate in parallel and are loaded
at 10.8 BV/hr. Recovery of ammonia and regenerant is by a closed loop air
stripping/adsorption system. Total capital cost for the system is $4,470,000
(1974 dollars). Operation and maintenance costs are detailed in Table B21-1.
B21-12
-------
Appendix B21
Clinoptilolite Ion Ex.
TABLE B21-1. COSTS FOR UPPER OCCDQDAN SYSTEM (1)
Specifications: - flow: 85,200 m»/day
- 8 beds operating in parallel
- each bed: 10.8 BV/hr, 20 x 50 mesh
- inflnent: 20 mg/L NH3-N
- closed loop regenerant recovery
Total Capital Cost $4,470,000 (1974 dollars)
Operation and Maintenance $18.00/1000 m*
Chemicals J/1000 m»
NaOH 7.10
NaCl 1.90
HiS04 2.60
Power - 3.6 kl/1000 mj 1.80
Labor 4.70
7. References
1. U.S. Environmental Protection Agency. Process Design Manual for Nitrogen
Control. October 1975.
2. McLaren, James R. and Grahame J. Farquhar. Factors Affecting Ammonia
Removal by Clinoptilolite. American Society of Civil Engineers Journal
of the Environmental Engineering Division, 99, 429, 1973.
3. Koon, John H. and W. J. Kaufman. Optimization of Ammonia Removal by
Ion Exchange Using Clinoptilolite. Sanitary Engineering Research Labora-
tory, Univrsity of California Berkeley, SERL 71-5, 1971.
4. Ames. L. L. Some Zeolite Equilibria with Alkaline Earth Metal Cations.
American Mineralogist. 49, 1049, 1964.
5. Weber, Walter J., Jr. Physiochemical Processes for Water Qualtiy Con-
trol. Wiley-Interscience, New York, New York, 1972.
6. Battelle Northwest Laboratory. Ammonia Removal from Agricultural Runoff
and Secondary Effluents by Selective Ion Exchange. Robert A. Taft Water
Research Center Report No. TWR C-5, March 1969.
7. U. S. Environmental Protection Agency. Correspondence T320-JM-82-3,
IERL/RTP, Research Triangle Park, NC, September 30, 1982.
B21-13
-------
-------
APPENDIX B22
MEMBRANE SEPARATION
1. Process Description
There are a number of membrane separation processes, all of which have
certain features in common. These include a fluid containing two or more com-
ponents on one side of a selective membrane through which one component (or
group of like components) is more permeable than the others. On the other
side of the membrane is a fluid which receives the transferred components.
Finally, in order to cause the transfer to take place, there must be a driving
force of some kind (1).
Three membrane processes that have potential application for wastewater
treatment are ultrafiltration, reverse osmosis, and electrodialysis (2).
Dltrafiltration and reverse osmosis are very similar processes in that
they both, achieve separation by applying a pressure gradient (driving force)
across a semipermeable membrane. In ultrafiltration, which typically operates
at pressures from 0.17 to 0.79 MPa, solvent molecules are allowed to pass
through the membrane while higher molecular weight impurities are retained.
The primary factor determining separation is the size of the solute molecules.
In reverse osmosis, higher pressures are used, typically up to 7 MPa (3).
Water, light aliphatic molecules capable of hydrogen bonding (such as alco-
hols, phenol, aldehydes, acids, and amines), and other nonelectrolytes with
molecular weights less than about 200 are able to pass through the membrane
while ionic species and heavier organics are rejected (2). Reverse osmosis is
so named because it uses pressures higher than the osmotic pressure of the
solution to drive the water from an area of high solute concentration to an
area of lower concentration, or, in other words, against the direction of
normal osmotic flow. Basic types of membrane permeators include plate and
frame, tubular, helical tubes, spiral wound, and hollow fiber units (1).
B22-1
-------
Appendix B22
Membrane Separation
Electrodialysis, unlike ultrafiltration and reverse osmosis, removes the
ionic impurities, rather than the solvent, from the original solution. In
this process, cation-exchange membranes are alternated with, anion-exchange
membranes in a parallel array to form thin solution-compartments, each about
0.5-1.0 mm thick. This assembly is held between two oppositely charged elec-
trodes. The solution to be treated is circulated through the solution com-
partments. Under the influence of an electric potential, the anions in solu-
tion tend to migrate toward the anode while the cations tend to move toward
the cathode. The ions in even numbered compartments are able to transfer
through the first membrane they encounter (anions through anion-exchange mem-
branes, cations through cation-exchange membranes), but they are blocked by
the next membranes they encounter. The ions in odd-numbered compartments are
blocked from traveling in either direction. The result is the depletion of
ionic species in half of the compartments and their concentration in the other
half (1).
2. Process Applicability
In general, membrane separation processes are applicable only to streams
which are very low in suspended solids, since these impurities readily clog
the membranes' pores. The extent to which pretreatment is required depends on
the specific membrane process. Reverse osmosis and clectrodialysis membranes
require more stringent control than those used for ultrafiltration.
Ultrafiltration is capable of removing heavier molecular weight impuri-
ties from wastewater. Smaller nonionic and ionic species are not removed.
Pretreatment may be required to reduce suspended solids and concentrations of
high molecular weight impurities whose solubilities may be exceeded in or on
the membrane, leading to fouling problems (4). Ultrafiltration has been used
commercially to recover protein from whey and to recover valuable metal from
electroplating operations (1).
B22-2
-------
Appendix B22
Membrane Separation
Reverse osmosis can remove both organics and inorganics, concentrating
the contaminants in a smaller volume stream. Its applicability to a given
stream depends mainly on two factors. First, the stream must be very low in
suspended solids, and must contain little or no organics, in order to avoid
membrane plugging or fouling. Second, the osmotic pressure of the solution
must be considerably lower than the practical operating pressure of such units
(2.9-7.0 MPa) (4). The membranes are also sensitive to pH, temperature, and
oxidizing agents (such as chlorine). Chlorine concentrations must be less
than 5 ppm for cellulose acetate and polyamine membranes (5). Concentration
of the wastewater in the system must not exceed the solubility of the least
soluble compound, or plugging and fouling can result. Reverse osmosis was
developed primarily for desalting of brackish water (2), but it has been uti-
lized effectively in recovering metals from plating bathwaters (6). It can
be used to produce high quality feed water for boilers by reducing dissolved
minerals concentrations (7) and has been used to treat municipal sewage (1).
Electrodialysis is applicable to the removal of ionic species from water
streams which have been pretreated to remove suspended solids and organics.
Electrodialysis membranes are also sensitive to HaS, chlorine, and inorganic
scale-forming compounds. Electrolysis has been used to produce food grade
salt from brine obtained by concentrating seawater and to produce drinking
water from brackish water (4).
3. Process Performance
Contaminant removals using ultrafiltration depend primarily on the size
of the molecular species present in the stream to be treated; both organics
and inorganics are removed. Removal efficiencies for high molecular weight
species can be quite high. Factors affecting performance include scaling and
fouling of the membrane caused by excessive concentrations of contaminants in
or near the membrane, as well as temperature, pH, and chlorine concentration
(2).
B22-3
-------
Appendix B22
Membrane Separation
Reverse osmosis is capable of removing up to 95 percent of the IDS in
waste streams (8). Rejection rates for chlorinated hydrocarbons from 98.95 to
nearly 100 percent have been reported (9). The performance of reverse osmosis
membranes is affected by the same variables as for ultrafiltration. For com-
monly used membranes, limits on pH from slightly acidic to about 9 and an up-
per bound on temperature of about 311 K are reported (8). The water produced
by reverse osmosis is typically high quality (4).
Electrodialysis is capable of water recoveries of 80 to 90 percent from
wastewaters containing up to 100 mg/L of dissolved inorganics; the recovered
water can contain as little as 3 to 5 mg/L of dissolved solids (10). Removals
of from 50 to greater than 90 percent of ionic compounds from aqueous streams
containing from 5000 to 10,000 mg/L of such compounds are reported (4).
4. Secondary Waste Generation
All membrane processes produce a stream in which the removed species are
concentrated and a stream which has been depleted of these species. The
concentrated stream can be treated further or disposed of. The diluted stream
may be suitable for reuse elsewhere in the process due to its high quality, or
it may be discarded. Either of the streams may be considered a waste depend-
ing on the application. In wastewater treatment applications, the concen-
trated stream is generally considered to be the waste stream.
5. Process Reliability
The reliability of any of the membrane processes depends on membrane
life, which is a function of the nature and extent of fouling and plugging.
For most wastes, membrane life must be determined experimentally (4). No data
are available for reliability of membrane technologies applied to the treat-
ment of synthetic fuel process wastewaters.
B22-4
-------
Appendix B22
Membrane Separation
6. Process Economics
Estimated costs of wastewater treatment as a function of stream flow rate
for reverse osmosis and electrodialysis from the literature are presented in
this section.
Installed equipment costs for a complete reverse osmosis plant for treat-
ing waters with total dissolved solids up to about 10,000 mg/L were obtained
from Reference 11. Single pass treatment at 2.9-3.2 MPa is assumed. Water
recoveries are from 75 to 80 percent. Pretreatment and brine disposal costs
are not included. A CE cost index ratio of 1.23 was used to convert these
costs from January 1978 to first quarter 1980 dollars. Hie updated estimated
installed equipment costs as a function of wastewater flow are found in Figure
B22-1.
Operating and maintenance requirements for this system were taken from
the same source. They include electric power for feedwater pumping and other
chemical feed and pumping; materials costs for replacement of membranes and
cartridge filters, cleaning chemicals, and periodic equipment repair; and la-
bor for maintenance and performance monitoring. These costs, as a function
of flow rate, are summarized in Figure B22-2.
Capital costs for an electrodialysis system for reducing the IDS concen-
tration of an influent containing 3000 mg/L of IDS were taken from Reference
8. In order to convert these estimated costs to installed equipment costs,
the assumption was made that the capital cost components of engineering and
construction, fees, and contingency were equal to 48 percent of the installed
equipment costs. A CE index ratio of 2.06 was used to update the costs given
in 1970 dollars to first quarter 1980 dollars. Figure B22-3 presents costs
versus wastewater flow. Total operating and maintenance costs, similarly
updated, are found in Figure B22-4.
B22-5
-------
7 -
2 -
10,000
First Quarter 1980 $
7 -
-4,000
100
I I I 1 I 1 1
L J_ j II it L
I 1 i I I I
10
7 100 2 47 1000 2
Waseewater Flow Rate. mVhr
Figure B22-1. Installed equipment cost for reverse osmosis
wastewater treatment (11)
B22-6
-------
Operating Costs, S/1000 m3
ro
I
H-
CO
l-i
ro
tsj
NJ
rt O
"I X3
n> m
0) <-<
g S
n> H-
3 3
rt OQ
O
O
CD
O
i-t
i-t
ft
1-1
cn
(T)
O
cn
§
cn
(u
rt
CD
-------
Installed Equipment Cost, $/(m3/hr)
CD
I
CO
H-
05
e
I
U!
0- M
0> I—'
O M
rt (D
i-i a,
o
a. to
M* .-O
p e
i—. H-
l< 'd
3
m
03
D
rt
w n
^" o
01
Hi
o
H
G
c
o
o
~I—I—I I I I I
ro
~r
T 1 I I I I I
o
o
o
~I—I—I I I I I
-------
1,000
First Quarter 1980 $
S. 100
Operating & Maintenance
-(3000 mg/L IDS)
10
_J I I J I
10
7 100 2 47 1000
Wastewater Flow Rate, m Vhr
Figure B22-4. Operating cost for IDS reduction by electrodialysis (8)
B22-9
-------
Appendix B22
Membrane Separation
7. References
1. Lacey, R.E. Membrane Separation Processes. Chemical Engineering,
September 4, 1972, pp. 56-74.
2. Mulligan, T.J. and R.D. Fox. Treatment of Industrial Wastewaters.
Chemical Engineering, October 18, 1976, pp. 49-64.
3. Metcalf and Eddy, Inc. Wastewater Engineering. McGraw-Hill, New York,
New York. 1972.
4. Blackburn, J.W. Removal of Salts from Process Wastewaters. Chemical
Engineering, October 17, 1977, pp. 33-41.
5. U.S. Environmental Protection Agency. Coal Conversion Control
Technology, Vol. I, Environmental Regulations; Liquid Effluents.
EPA-600/7-79-228a, October 1979.
6. Lanouette, K.H. Heavy Metals Removal. Chemical Engineering, October 17,
1977, pp. 73-80.
7. Chu, T.Y., R. Ruane, and P. Krenkel. Wastewater Recycling in
Steam-Electric Power Plants. In proceedings of the Third Annual
Conference on Treatment and Disposal of Industrial Wastewaters and
Residues, Houston, Texas, April 18-20, 1978, pp. 135-145.
8. Patterson, J.W. Wastewater Treatment Technology. Ann Arbor Science
Publishers, Ann Arbor, Michigan, 1975.
9. Paulson, E.G. How to Get Rid of Toxic Organics. Chemical Engineering,
October 17, 1977, pp. 21-27.
10. Katz, W.E. Electrodialysis for Low TDS Waters. Industrial Water
Engineering,8 (6), June/July 1971.
11 U.S. Environmental Protection Agency. Estimating Costs for Water
Treatment as a Function of Size and Treatment Plant Efficiency.
EPA-600/2-78-182, August 1978, pp. 203-212.
B22-10
-------
APPENDIX B23
BIOLOGICAL PROCESSES FOR REMOVAL
OF REDUCED NITROGEN SPECIES
1. Process Description
Biological treatment systems utilize metabolic processes of microorgan-
isms to remove dissolved contaminants from a waste stream. Most of the con-
taminants are oxidized to nonobjectionable, simple inorganics as the end pro-
duct of energy reactions, while the rest are assimilated as new cell material.
A variety of physical configurations can be used to provide contact
between the microorganisms and the contaminants to be removed, but suspended
growth systems (activated sludge) are most commonly used. The cells are
dispersed throughout a reactor where they contact the incoming waste stream on
a continuous basis. Enough residence time is provided for the microorganisms
to degrade the contaminants to a desired level, and effluent from the reactor
is routed to a clarifier where the suspended cells are separated from the bulk
fluid.
Concentrated cells are drawn from the underflow of the clarifier, and a
fraction are wasted equal to their increase in mass resulting from cell growth
in the reactor. The rest, constituting the much larger fraction, are returned
to the reactor. By this process of cell wastage and recycle, a constant mass
of cells is maintained in the reactor for a controlled period longer than the
hydraulic residence time. The cell residence time, defined as the ratio of
cells in the system to their rate of removal, is the key process variable. It
controls both the overall system performance and stability. The system can be
operated over a range of residence times. The value selected reflects condi-
tions of optimal performance and stability.
B23-1
-------
Appendix B23
Biological Nitrogen Removal
The interactions occurring between microorganisms and wastewater consti-
tuents in an activated sludge system are complex and not easily quantified.
Design procedures that appeal strictly to fundamental interactions therefore
do not exist. Instead, biological wastewater treatment processes are usually
designed by a combination of empirically-based design equations and experience
compiled from operating systems.
Design equations can be developed by applying kinetic expressions des-
cribing growth of microorganisms to the particular hydraulic contraints of a
selected reactor configuration. Assuming that all growth nutrients are pre-
sent in excess, cell growth is commonly related to rate of substrate (the con-
taminant species) utilization as follows.
dl dS
dT = Y ol - bx
where: dX/dt = net microbial growth rate per volume (mass time
vol"1),
Y = growth yield (mass cells/mass substrate),
dS/dt = rate of substrate utilization per volume (mass
time vol ) ,
b = cell decay rate — including cell death, _endogenous
respiration, and cell maintenance (time ), and
X = cell mass concentration (mass vol ) .
Substrate utilization can be further expressed by the following Monod
expression:
dS kSX
dt K + S
s
B23-2
-------
Appendix B23
Biological Nitrogen Removal
where: S = substrate concentration,
k = maximum rate of substrate utilization per mass of
cells (mass substrate/mass cells-time), and
K^ = half-velocity (or saturation) coefficient.
A completely-mixed, suspended growth system, whereby the microorganisms
and influent stream are fully dispersed through the reactor, is depicted in
Figure B23-1. Equations 1 and 2 can be applied to this system by appropriate
mass balances and by assuming steady state conditions to define the key pro-
cess variable of cell residence time as follows:
1
6
YkS
K +S
s
-b
where: Q = cell residence time and
c
other terms are defined above.
(3)
Q°,S
Qe, Xe, Se
QW, XW
Figure B23-1. Completely-mixed activated sludge system
where: Q , Q , Q
xe. xw
S , S
= flow of influent, effluent, and waste streams,
respectively;
= mass concentration of cells; effluent and waste
streams, respectively; and
= substrate concentration in influent and effluent.
B23-3
-------
Appendix B23
Biological Nitrogen Removal
The substrate concentration term (S) in equation 3 refers to both its
concentration in the reactor and in the effluent since the contents of the
reactor are completely mixed. Equation 3 suggests that as the cell residence
time decreases, the concentration of substrate in the effluent will increase.
A minimum cell residence time can be defined, therefore, as the cell residence
time for an infinitely large effluent substrate concentration. The cell resi-
dence time selected for system operation exceeds the minimum value by a safety
factor to account for both uncertainties in the selection of kinetic
parameters and deviation of operating conditions from steady state. Safety
factors typically selected range from 3 to 20 or more.
Once cell residence time is determined, the reactor volume is fully
defined for a selected concentration of microorganisms in the reactor. This
concentration is constrained by practical considerations. Too large a concen-
tration may create a demand for oxygen in excess of that which can be readily
transferred to the reactor contents. The capacity of the clarifier to sepa-
rate cells from the bulk fluid may likewise be exceeded. Cell concentrations
that can be reasonably maintained generally range between 2000 and 3000 mg/1,
although significant deviations from this range are not uncommon. The hydrau-
lic residence time «J) of a reactor required to support a particular concen-
tration of microorganisms (X) having a cell residence time of 0 can be
c
expressed as follows:
o-e
(1 + 0.2 b « ) (4)
c (*'
where: S°,S = influent, effluent substrate concentration and I and
b are as defined for equation 1.
Reliable design equations for the clarifier do not exist. If a sample of
the biomass is available, bench-scale settleability tests can be used to
determine the design. However, the clarifier is frequently determined more
B23-4
-------
Appendix B23
Biological Nitrogen Removal
simply to conform with empirical specifications that summarize experience from
other systems. Typically, overflows and loading are 16-32 m»/m*-day and 3.0-
6.0 kg/m*-hr, respectively (1).
2. Process Applicability
Biological treatment processes are broadly applicable to wastewater
streams requiring removal of certain reduced nitrogen species. Two applica-
tions are considered in this section: oxidation of ammonia (nitrification)
and oxidation of thiocyante. Biological removal of nitrate (denitrification),
while not a process for removal of a reduced nitrogen species, is included
since this process frequently follows nitrification.
Any application of a biological treatment process to a particular
wastewater stream should be preceded by testing of the stream to evaluate
kinetic parameters in the design equations. In the absence of such a stream
to examine, kinetic parameters can be estimated from experience compiled from
treatment of similar wastewaters.
Biological Nitrification
Biological nitrification is designed to remove ammonia from waste streams
by oxidizing it to nitrate. The reaction occurs in a two-step process: ammo-
nia is oxidized by nitrosomonas to nitrite followed by nitrite oxidized to
nitrate by nitrobacter. The nitrosomonas reaction occurs much more slowly
and therefore controls the overall kinetics of ammonia removal (2). An over-
all expression for both processes, including cell growth may be estimated as
follows:
NH^ + 0.10 C0$ + 1.86 02 =
0.021 CsH7OiN + 0.99 H»0 + 0.98 NOT + 1.98 H+ (5)
B23-5
-------
Appendix B23
Biological Nitrogen Removal
For each gram of ammonia nitrogen removed, 4.2 grams of oxygen are
required, and 7.1 grams of alkalinity as CaCO, are consumed. The alkalinity
requirement may be decreased somewhat by carbon dioxide being stripped from
solution by the aerators. If sufficient buffer capacity is not available in
the waste stream, alkalinity must be added to maintain pH near neutral. The
dissolved oxygen concentration in the reactor should exceed 2 mg/L to ensure
sufficient driving force for cellular metabolism (2). Temperature influences
are accounted for in the kinetic parameters that follow.
Reference 3 suggests the following estimates for kinetic parameters:
yield: Y = 0.29, mg cells/mg N (6)
decay: b = 0.05, day (7)
,,-,.«. v ,.(0.051 T- 1.158) ,_.
half-velocity: K = 10 (8)
where: T = temperature, °C.
Maximum growth rate (k)' is estimated in terms of maximum specific growth rate,
» = Yk (9)
DO
and, j. = 0.47 [exp (0.098 (T-15))] [1 - 0.833 (7.2-pH)] [po + l ^ (10)
where: T = temperature, °C
DO = dissolved oxygen, mg/L, and
(i = days
If pH is outside the range of 7.2 to 8.0, a value of 7.2 should be used in the
above expression.
B23-6
-------
Appendix B23
Biological Nitrogen Removal
Biological Denitrification
Biological denitrification is designed to remove nitrate from wastewater
streams by reducing it to molecular nitrogen. In the absence of oxygen, many
microorganisms including pseudomonas. achromobacter, and bacillus can
utilize nitrate as a terminal electron acceptor. The switch from oxygen to
nitrate occurs readily because essentially identical biochemical pathways are
involved in either case (2). Thus nitrate is removed biologically not by
itself being metabolized, but by participating as the electron acceptor in the
metabolism of some other species. This species (the electron donor) must be
supplied externally, and methanol is commonly used.
Nitrate is reduced to molecular nitrogen by a two-step process summarized
as follows for methanol as the electron donor (2):
N0~ + j CHjOH - N(>7 + 3 C0a + | HaO (11)
NO^ + \ CH,OH = | N2 + J HaO + OH~ + ^ CO, (12)
An overall reaction including cell growth can be described as follows:
N0~ + 1.1 CH3OH + H+ = 0.07 C^H^N + 0.47 N, + 0.76 CO, + 2.45 H20 (13)
Any dissolved oxygen also present in the reactor combines with methanol
and nitrite. The following equations provide estimates of the methanol
required and biomass produced (2):
Cm = 2.47 N07-N + 1.53 NO^-N + 0.87 DO (14)
Cb = 0.53 N07-N + 0.32 NO^-N + 0.19 DO (15)
where: Cm = methanol concentration required, mg/L,
B23-7
-------
Appendix B23
Biological Nitrogen Removal
C = concentration of biomass, mg/L,
NO^-N, NO~-N = mg/L, and
DO = dissolved oxygen, mg/L.
The reactor configuration described for biological processes may be used
with two modifications: the main reactor is not aerated (and may even be
covered to minimize transfer of oxygen between the atmosphere and the reactor
contents), and a small, heavily aerated reactor having a residence time of 5
to 50 minutes is situated between the main reactor and clarifier. Without
aeration following the main reactor, successful operation of the entire system
would be possible only with precise stoichiometric additions of methanol (3).
With too little methanol, there would be incomplete removal of nitrate. Deni-
trification would continue in the clarifier with cells becoming the electron
donor, and the sludge would settle poorly, countercurrent to escaping bubbles
of nitrogen. With excess methanol, there would be a high BOD residual in the
effluent. Since the microorganisms responsible for denitrification can use
either oxygen or nitrate as a terminal electron acceptor, any excess methanol
is oxidized when the effluent from the main reactor is aerated. Aeration also
strips excess nitrogen from solution to further improve sludge settleability.
Reference 3 suggests the following estimates for kinetic parameters at
293 K and within the optimum pH between 7.0 and 7.5:
yield: Y = 0.6 to 1.2, g cells/g NOj-N (16)
decay: b = 0.04, day (17)
half velocity: K = 0.08 - 0.16, mg NO~-N/L (18)
maximum growth rate: k = 0.18, day (19)
B23-8
-------
Appendix B23
Biological Nitrogen Removal
Thiocyanate Biooxidation
Several autotrophic and heterotrophie bacterial species have been iso-
lated that can oxidize thiocyanate. These include: pseudomonas stutzeri.
thiobacillus thiocyanoxidans. and thiobacillus thioparus (4). Other
microorganisms will also be present in the biomass to degrade nonthiocyanate
species, possibly including nitrifying bacteria to metabolize the ammonia by-
product. The reaction stoichiometry believed to predominate is as follows
(4):
SCN~ + 2H20 + 20a = C02 + S0~ + NH* (20)
Optimal operating pH ranges between 6 and 7 (4).
Thiocyanate is one of many substances that acts as a growth substrate at
low concentration but becomes a toxic inhibitor at high concentration. The
Monod expression describing growth can be modified to account for toxic inhi-
bition as follows:
^+(^1 J (21)
O
where: K. = inhibition constant, and
n = order of inhibition.
Using semi—continuous bench—scale reactors, microorganisms were accli-
mated to a waste stream with thiocyanate as the only growth substrate (4).
The observed performance correlated well with the above kinetic expression,
and coefficents were determined by a numerical treatment of the data.
B23-9
-------
Appendix B23
Biological Nitrogen Removal
[SCN
2'24/H + T^TT + i mn I 1 (22)
where: SCN is in mg/L.
This equation can be combined with an overall yield, a . , that accounts
ob
for both growth and decay to express the cell residence time (4):
e = aob x " <23>
c
kg volatile suspended solids
where: a = 0.117
ob kg SCN ° < 28 days
aob = 4'6 V1'1 ° > 28 davs
The minimum cell residence time, [6 ] , defined by the above equation
c 1 im
when the thiocyanate concentration approaches infinity, is about 4.14 days.
3. Process Performance
Most applications of nitrification recorded in the literature are for
wastewater streams having high concentrations of both ammonia and organics.
The heterotrophic mocroorganisms that oxidize the organics have higher yields,
and, therefore, greatly outnumber the autotrophic nitrifiers, nitrosomonas
and nitrobacter. that concurrently oxidize ammonia. Certain stability is
associated with these systems over those where the principal contaminant
species is ammonia.
Four examples of nitrification of industrial strength wastewaters are
presented in Table B23-1. Each of these wastewaters also contains significant
concentrations of organics. Ammonia removals reported range from 80 to
greater than 99 percent.
B23-10
-------
TABLE B23-1. EXAMPLES OF NITRIFICATION EXPERIENCE
Case: A
Reference 5
Influent Characteristics, mg/L
COD 200-800
BOD
TOC 30-420
NHs-N 200-250
§ Org-N*
M Effluent Characteristic, mg/L (% removed)
COD
BOD
NH3-N - (80%)
Org-Na
BBC
667
950 2700 628
670 2500 149
_
125 470 58.4
_
200(68.2)
20.8(86.1)
15 (88%) 80 (83%) 0.5 (99.1)
_
D
8
11,600
3,500
-
20
780
-
-
trace
230
Organically bound nitrogen.
(A) Pilot study with a chemical processing wastewater spiked to NH
(B) Coke oven wastewater - best attainable performance reported.
(C) Coke plant wastewater - bench-scale.
(D) Pilot-scale - source of wastewater not specified.
200-250 mg/L.
-------
Appendix B23
Biological Nitrogen Removal
Most experience with denitrification is with municipal wastewaters for
which denitrification follows nitrification and over all nitrogen concentra-
tions are low. Excellent removals and stable performance have been recorded,
an example being the Central Contra Costa Sanitary District in California (3).
Nitrified effluent containing an average nitrogen concentration of 27 mg/L was
denitrified by about 98 percent during a representative three month period.
Some denitrification experience with waste streams containing large con-
centrations of nitrate is presented in Table B23-2. Most of these waste
streams also contain large organic concentrations. One exception is a case in
which potable water was spiked with nitrate and denitrified in a plug flow
reactor (11). At 293 K, removals of 99 and 94 percent were achieved for
initial nitrate nitrogen concentrations of 100 and 200 mg/L respectively.
All full-scale experience recorded for biological oxidation of thiocya-
nate is for waste streams also having high concentrations of organics. How-
ever, thiocyanate removal from synfuels waste streams free of organics appears
promising based on results of a bench-scale study using a feed of coal gasifi-
cation quench condensate from the Koppers-Totzek process (12).
The study demonstrated that more than 95 percent of the thiocyanate can
be removed from streams containing either thiocyanate only or both thiocyanate
and ammonia.
Experience with wastewaters containing thiocyanate and large concentra-
tions of organics is presented in Table B23-3. Thiocyanate removals ranging
from 83 to greater than 99 percent are reported.
B23-12
-------
TABLE B23-2. EXAMPLES OF DENITRIFICATION EXPERIENCE
Case :
Reference
A
9
A
9
B
10
C
11
ro
LO
I
Influent Characteristics, mg/L
COD
BOD
TKNa
100
Effluent Characteristic, mg/L (% removed)
NO~-N - (99%)
TKN*
NH -N
3
200
- (94%)
1300-3300
400-1100
- (72%)
1500
160
150
75
- (95%)
- (95%)
1
Ammonia plus organically bound nitrogen.
(A) Potable water spiked with N03
(B) Pilot-scale study
(C) Full-scale system - wastewater from manufacture of nylon intermediates.
-------
TABLE B23-3. EXAMPLES OF THIOCYANATE OXIDATION EXPERIENCE
Case:
Reference
Influent Characteristics, mg/L
COD
BOD
SCN~
Effluent Characteristic, mg/L (%
COD
BOD
SCN~
A
13
3900-4600
1700-2800
280-510
removed)
300-410
5-15
3-36
B C
14 15
1329 3710
650
130 12
436 710
10
3.5 (97) 2 (83)
D
16
6780
-
16
1260
-
2 (88)
(A), (B) Coke plant wastewater.
(C) Coal gasification wastewater from HYGAS.
(D) Coal gasification wastewater from slagging fixed bed.
-------
Appendix B23
Biological Nitrogen Removal
4. Secondary Waste Generation
Two secondary waste streams are associated with suspended growth
biological treatment processes. Waste sludge consisting of about 1 to 2
percent solids is intermittently drawn from the underflow of the clarifier.
The mass flow is equal to the net growth rate of microorganisms in the main
reactor which is specific to the characteristics of the influent stream.
Prior to being handled by an appropriate solids disposal technique, this
stream could be concentrated to about 30 to 40 percent solids by a filter
press or equivalent process.
In some applications, loss of volatile species from the clarifier may be
responsible for an additional secondary waste stream. Air or oxygen to sup-
port microbial reactions is introduced into the reactor by techniques that
maximize turbulence to ensure optimal mass transfer of oxygen and complete
mixing of the reactor contents. However, these same conditions are also con-
ducive to stripping of volatile species. At the near neutral pH at which the
reactor would operate, little ammonia or thiocyanate should be lost, but other
constituents possibly present, such as sulfide, may be readily stripped from
solution.
5. Process Reliability
Operating experience has been presented for removal of ammonia, nitrate,
and thiocyanate from wastewater streams having high concentrations of
organics. No known experience, and therefore measure of reliability, exists
for full-scale systems treating wastewaters containing the subject nitrogen
species exclusive of organics and at concentrations in the few hundred or more
mg/L level. However, nitrification and denitrification have been used exten-
sively to treat municipal strength wastewaters. For some of these applica-
tions, the nitrification reactor is preceded by an activated sludge system to
B23-15
-------
Appendix B23
Biological Nitrogen Removal
remove organics such that the predominant substrate for microorganisms in the
nitrification reactor is ammonia. Nitrification and denitrification are con-
sidered highly reliable in these applications (17).
6. Process Economcis
Cost data presented in the open literature consist of both cost curves
and costs reported for individual systems. Typically, the basis of costs are
not stated clearly, making comparison of data from different sources diffi-
cult. Capital and certain operation and maintenance costs for nitrification
and denitrification are presented in Table B23-4 and B23-5. Costs for thio-
cyanate oxidation can be approximated by those presented in Table B23-4.
B23-16
-------
TABLE B23-4. COST DATA REPORTED IN THE OPEN LITERATURE - NITRIFICATION1
03
W
I
System Characteristic:
Reference (18)b
Reference (19)°
Reference (20)
Flow
(m»/hr)
160
790
1600
650
Nitrogen Loading
(kg NH^-N/day)
77
380
770
390
450
2,300
4,500
23,000
45,000
230,000
Capital
Cost
(Million i)
1.5
3.9
6.2
3.3
0.56
0.88
1.2
3.2
5.2
20
0 8 M Dollars
Labor
52,000
99,000
147,000
128,000
19,000
19,000
20,000
35,000
50,000
140,000
Per Year
Power
14,000
52,000
105,000
27,000
200
600
1,300
5,500
11,000
60,000
.All data has been normalized to nitrogen loading basis.
1975 dollars - original data presented as cost curves.
*!l979 dollars
Q
1979 dollars - original data presented as cost curves.
-------
TABLE B23-5. COST DATA REPORTED IN THE OPEN LITERATURE - DENITRIFICATIONa
td
I
I—'
CO
System Characteristic:
Reference (18)b
Reference (20)
Flow Nitrogen Loading
(m'/hr) (kg NH^N/day)
160 77
790 380
1600 770
450
2,300
4,500
23,000
45,000
230,000
Capital
Cost
(Million i)
0.65
1.5
2.8
0.42
0.63
0.84
1.9
3.0
10
0 8 M Dollars
Labor
6,000
12,000
17,000
30,000
40,000
80,000
Per Year
Power
450
450
900
4,800
9,500
50,000
All data has been normalized to nitrogen loading basis.
,1975 dollars - original data presented as cost curves.
"1979 dollars - original data presented as cost curves.
-------
Appendix B23
Biological Nitrogen Removal
7. References
1. Metcalf and Eddy, Inc. Wastewater Engineering: Treatment, Disposal,
Reuse. McGraw-Hill Book Co., New York, New York, 1979.
2. McCarty, L. Nitrification-Denitrification by Biological Treatment.
Home Conference on Denitrification of Municipal Wastes, University of
Massachusetts Water Resources Research Center, 1973.
3. D. S. Environmental Protection Agency. Process Design Manual for
Nitrogen Control, October 1975.
4. Neufeld, D., et al. Thiocyanate Bio-Oxidation Kinetics. American
Society of Civil Engineers - Journal of the Environmental Engineering
Division, 107, 1035, 1981.
5. Ford, L., et al. Comprehensive Analysis of Nitrification of Chemical
Processing Wastewater. Journal Water Pollution Control Federation, 52,
2726, 1980.
6. Olthof, M. Nitrification of Coke Oven Wastewaters with High Ammonia
Concentration. Proceedings of the 34th Industrial Waste Conference,
Purdue University, 1980.
7. Bhattacharyya, A. and A. C. Middleton. Solids Retention Time a
Controlling Factor in the Successful Biological Nitrification of Coke
Plant Waste. Mid-Atlantic Industrial Waste Conference Proceedings, 1980.
8. Voets, J. P., et al. Removal of Nitrogen from Highly Nitrogenous
Wastewaters. Journal Water Pollution Control Federation, 47, 394, 1975.
9. Ericson, B. Nitrogen Removal in Pilot Plant. Journal Water Pollution
Control Federation, 47, 727, 1975.
10. Jewell, W. J. and R. J. Cummings. Denitrification of Concentrated
Nitrate Wastewaters. Journal Water Pollution Control Federation, 47,
2281, 1975.
11. Bridle, T. R. , et al. Operation of a Full-Scale Nitrification-Denitrifi-
cation Industrial Waste Treatment Plant. Journal Water Pollution Control
Federation, 51, 127, 1979.
12. Engineering Science, Inc. Biological Treatability Studies on
Koppers-Totzek Waterwaters. Austin, Texas, December 1981.
B23-19
-------
Appendix B23
Biological Nitrogen Removal
13. Luthy, R. G. and L. D. Jones. Biological Oxidation of Coke Plant
Effluent. Journal Environmental Engineering Division - American Society
of Civil Engineers, 106, EE4, 847, 1980.
14. Ely, R. B. and C. L. Bernett. Treatment of Fossil Fuel Derived
Wastewaters with Powdered Activated Carbon/Activated Sludge Technology.
In proceedings Sixth Symposium on the Environmental Aspects of Fuel
Conversion Technology, Denver, October 26-30, 1981, EPA 600/9-82-017,
September 1982.
15. U.S. Department of Energy. Experimental Analysis of Biological Oxidation
Characteristics of HYGAS Coal Gasification Wastewater. FE-2496-27, 1978.
16. U.S. Department of Energy. Biological Treatment of Grand Forks Energy
and Technology Center Slagging, Fixed-bed Coal Gasification Process
Wastewater. FE-2496-42. 1979.
17. D. S. Environmental Protection Agency. Treatability Manuals Volume III,
Technologies for Control/Removal of Pollutants. EPA 600/8-80-042c, July
1980.
18. Culp, Russell L., et al. Handbook of Advanced Wastewater Treatment. Van
Nostrand Reinhold Company, New York, New York, 1978.
19. Wilson, R. W., et al. Design and Cost Comparison of Biological Nitrogen
Removal Processes. Journal Water Pollution Control Federation, 53, 1294,
1981.
20. U. S. Environmental Protection Agency. Treatability Manual, Volume IV:
Cost Estimating. EPA-600/8-80-042d, 1980.
B23-20
-------
APPENDIX B24
CYANIDE REMOVAL BY POLYSULFIDE ADDITION
1. Process Description
Polysulfide addition is a method of removing cyanide from wastewater
streams by converting the cyanide to thiocyanate. Dnder alkaline conditions,
the reaction can be described by the following expression:
(S S)~ + xCN~ = xSCN~ + S~ (1)
x
While various sulfur species can potentially combine with cyanide to form
thiocyanate, only polysulfide appears to do so with characteristics that can
be utilized for wastewater treatment. Hydrosulfide (HS ) is not reactive with
cyanide (1). Thiosulfate (SaOj ) combines with cyanide but at a rate of at
least three orders of magnitude more slowly than does polysulfide (2). Ele-
mental sulfur is stable only under acidic conditions. Under alkaline condi-
tions, it decomposes to hydrosulfide (HS ) and thiosulfide, neither of which
combines well with cyanide (3).
Polysulfide addition is particularly useful when applied in conjunction
with an activated sludge process. In small concentrations, cyanide is norm-
ally toxic to microorganisms. In activated sludge systems fed with a waste-
water stream having strictly inorganic constituents, autotrophic micro-
organisms are involved, none of which can oxidize cyanide. In cases of cyan-
ide occurring in wastewater streams which contain significant concentrations
of biodegradable organics, the resulting, predominantly heterotrophic popula-
tion of microorganisms can sometimes be acclimated to oxidize or otherwise
remove part of the cyanide. However, incomplete removals are typical. For
either case, cyanide can be removed by first converting it to thiocyanate by
polysulfide addition, then oxidizing the readily degradable thiocyanate in
either an autotrophic or heterotrophic activated sludge system.
B24-1
-------
Appendix B24
Cyanide Removal
2. Process Applicability
The feasibility of converting cyanide to thiocyanate by polysulfide is
broadly supported by reports in the open literature. Ganczarczyk (4) suggests
that during storage and/or ammonia stripping, cyanide bearing coke-plant
effluent is converted to thiocyanate, as indicated by the following data:
raw after storage and/or
effluent _ NHs stripping
Total cyanide, mg/L 80 14
Free cyanide, mg/L 70 8
Thiocyanate, mg/L 62 200
Several authors discuss experience using polysulfide at refineries to
prevent iron corrosion. Molecular iron in conveying pipes is oxidized by
hydrosulfide (HS ) to form iron sulfide scale. This scale protects the sys-
tem from further attack, but cyanide in the waste stream dissolves it to
expose more molecular iron to attack.
Fe° + 2 HS~ = FeS + S= + H (2)
2
FeS + 6 CN~ = [Fe(CN) ] + S= (3)
Ehmke (5) describes direct injection of ammonium polysulfide at various points
into a Fluid Catalytic Cracking Unit gas plant to remove cyanide by converting
it to thiocyanate. Bonner (6) describes promoting the same reaction but by
injecting air to react with hydrosulfide to form polysulfide.
2 HS~ + 0 - 2 OH~ + 2 S° (4)
2
S= + S° = S (5)
2
B24-2
-------
Appendix B24
Cyanide Removal
Direct addition of polysulfide is probably far more reliable than air injec-
tion since the above sequence of reactions is sensitive to a variety of envi-
ronmental conditions (5).
Polysulfide addition could be made by one of several techniques, although
none have been demonstrated. Polysulfide could possibly be added to an exist-
ing wastewater treatment process such as activated sludge to accomplish the
cyanide to thiocyanate conversion concurrently with the reactions normally
associated with the treatment process. Alternatively, the conversion could be
made in a separate reactor. This reactor would closely resemble wastewater
systems such as chemical oxidation, with modifications made for the particular
requirements of the polysulfide reaction.
Perhaps the most attractive way to convert cyanide to thiocyanate is to
add the polysulfide directly to an industrial process similar to the way in
which it is done in the petroleum refining industry. This option must be
evaluated on an individual basis since its feasibility is highly specific to
the residence time, pH, and other environmental conditions associated with the
point of polysulfide addition.
3. Process Performance
A limited, EPA-supported study was conducted to evaluate the feasibility
of utilizing polysulfide addition as the basis for a wastewater treatment pro-
cess (7). The reaction was characterized by evaluating the reaction rate,
assumed to be described by the following expression:
R = :^QL~ = k[CNt]a[Sx]b (6)
B24-3
-------
Appendix B24
Cyanide Removal
where: R = rate of thiocyanate formation,
k = reaction rate constant,
a.b = order of reaction coefficients,
[S ] = concentration of polysulfide, and
[CN ] = total cyanide concentration.
Both elemental sulfur and polysulfide were used as the sulfur contributing
species, but polysulfide was shown to be far superior for this purpose.
Very little cyanide was converted at acidic or near neutral pH condi-
tions. Several factors are likely responsible. First, polysulfide is not
stable at low pH. For pH greater than 6, the dominant polysnlfide species are
= SE
S4 and S5 , but little exists at pfl less than about 8 (3).
HS + 3 S = S ~ + H+ pK = 9.86 (7)
HS~ + 4 S = S ~ + H+ pK = 9.18 (8)
It is not certain how quickly the equilibrium indicated by the above equations
develops and, therefore, what exact effect there is on the polysulfide/cyanide
reaction.
Second, it is not certain 1) how cyanide combines with the sulfur spe-
cies, 2) whether sulfur reacts with the cyanide ion (CN ), its conjugate acid
(HCN), or both, and 3) whether the reaction rates are identical or different.
If the sulfur species combines only with the cyanide ion or if it combines
with both CN and HCN but with CN at a much faster rate, then the overall
reaction rate (at constant temperature and cyanide measured as total cyanide)
should parallel the relative concentration of cyanide ion. Thus, pH and its
B24-4
-------
Appendix B24
Cyanide Removal
effect on the cyanide ion concentration should impact the reaction rate. At
nentral pH, only 0.3 percent of the total cyanide is in the form CN , possibly
providing too small a driving force to combine with polysulfide.
An additional factor influencing the rate of cyanide conversion is
consumption of polysulfide by thiosulfate according to the following (2):
HSO ~ + S - S 0 = + H+. (9)
3 2 3
This reaction however should be significant only at very high levels of pH.
The EPA study indicated best results for cyanide removal were obtained at
pH 8.7. Kinetic parameters for equation 6 are summarized in Table B24-1.
These parameters compare well with values determined at pH 8.2 by Luthy (2).
TABLE B24-1. KINETIC PARAMETERS FOR CONVERSION OF CYANIDE TO THIOCYANATE
Investigator
EPA (7)
Luthy (2)
pH
8.7
8.2
a
0.86
1.04
b
0.84
0.85
k
1.2
1.4
The rate of the cyanide reaction should be conservatively estimated by the
parameters of Table B24-1. For both the EPA (7) and Luthy (2) experiments,
polysulfide was added as sodium polysulfide. However, greatly improved kine-
tics result when ammonium polysulfide is used rather than sodium polysulfide.
This result is supported by both the EPA study and refinery experience where
corrosion problems corrected by ammonium polysulfide addition have been found
to sometimes continue when sodium polysulfide is used (5). Sodium polysulfide
B24-5
-------
Appendix B24
Cyanide Removal
is reportedly more stable than ammonium polysulfide, and therefore may dis-
sociate slowly enough to be rate limiting or at least rate influencing (6).
However, not all variables involved in the conversion of cyanide to thiocy-
anate are well understood so any application of this method would require sup-
port by extensive testing.
For the reaction kinetics defined by the EPA and Luthy studies, rela-
tively long residence times are needed to obtain high conversions unless poly-
sulfide is added in great excess (see Table B24-2). If large excesses of
polysulfide are used, residual levels will therefore result, possibly causing
problems for down-stream treatment processes.
TABLE B24-2. ESTIMATED RESIDENCE TIMES REQUIRED FOR 90% CONVERSION
OF CYANIDE TO THIOCYANATE IN A PLDG-FLOW REACTORa
Kinetic Data
Source
Luthy (2)
EPA (7)
Polysulfide Dose
pH mg-mole/L of water
8.2 0.4
0.8
2.4
4
8.7 0.4
0.8
2.4
4
Residence Time,
minutes
5800
1400
440
270
1150
300
95
60
Initial cyanide concentration 0.4 mg-mole/L of water (about 10 mg/L)
4. Secondary Waste Generation
Polysulfide must be reacted with cyanide at alkaline conditions. Fol-
lowing the reaction, the wastewater stream would probably be neutralized
either by direct acid addition or by the stream being combined with another
B24-6
-------
Appendix B24
Cyanide Removal
sufficiently acid stream. In either case, residual polysulfide may precipi-
tate and cause problems for downstream treatment processes. Depending on its
characteristics, the precipitate can be removed either by clarification or
filtration.
Petroleum refining industry experience suggests that the polysulfide
would be added as ammonium rather than as sodium polysulfide. An increase in
the concentration of ammonia would therefore result. If polysulfide is added
as sodium polysulfide, the IDS of the stream also would increase.
5. Process Reliability
There is no precedent for utilizing the reaction between polysulfide and
cyanide as the basis for, or in conjunction with, a wastewater treatment pro-
cess. However, since the process is a simple chemical addition similar to the
process of chemical oxidation, reliability is expected to be correspondingly
high.
6. Process Economics
The major cost associated with polysulfide addition is the cost of poly-
sulfide. Also contributing are chemicals required to control pH, both to
adjust pH to within the reaction range and to maintain it during the reaction.
A smaller demand for pH chemicals may be possible if the polysulfide process
is integrated with other wastewater treatment processes. For stoichiometric
additions, polysulfide is estimated to annually cost i6.3 per g/hr cyanide
removed.
B24-7
-------
Appendix B24
Cyanide Removal
7. References
1. Luthy, R. G. et al. Cyanide and Thiocyanate in Coal Gasification
Wastewaters, Journal of the Water Pollution Control Federation, 51, 2267
(1979).
2. Luthy, R. G. and Samuel G. Bruce, Jr. Kinetics of Reaction of Cyanide
and Reduced Sulfur Species in Aqueous Solution. Environmental Science
and Technology, 13, 1481, 1979.
3. Chen, Y. Chemistry of Sulfur Species and their Removal from Water
Supply, Chemistry of Water Supply, Treatment and Distribution. Ann Ar-
bor: Ann Arbor Science Publishers.
4. Ganczarczyk, J. J. Fate of Basic Pollutants in Treatment of Coke-Plant
Effluents. Proceedings of the 35th Industrial Waste Conference, Purdue
University, 1981.
5. Ehmke, E. F. Polysulfide Stops FCCD Corrosion. Hydrocarbon Processing,
July 1981.
6. Bonner, W. A. and H. D. Burnham. Air Injection for Prevention of
Hydrogen Penetration of Steel. Corrosion, October 1955.
7. D. S. Environmental Protection Agency. Correspondence T320-JM-82-3,
IERL/RTP, Research Triangle Park, N.C., September 30, 1982.
B24-8
-------
APPENDIX B25
CHEMICAL OXIDATION-REDDCEP NITROGEN AND SULFUR SPECIES
1. Process Description
Chemical oxidation is a process of transferring one or more electrons
from one chemical species to another. These reactions can be utilized in a
wastewater treatment process to change a pollutant species to a less reactive
or otherwise less objectionable form. To be feasible, the chemical oxidation
reaction must occur within a reasonable period and addition of reacting
chemicals must not be excessive. In addition to the oxidant, chemicals are
required to control reaction conditions, especially pH. The rate of most
chemical oxidations decreases sharply for conditions differing only slightly
from the optimal pH range.
Systems designed to effect chemical oxidation are generally simple (1).
Equipment is required for the following operations: chemical handling and
storage, flash mix of the oxidizing agent with the waste stream, and residence
time for the reaction to be completed.
Of the many possible oxidizing agents, few are usable for wastewater
treatment applications. Chlorine and ozone best satisfy the following
criteria for a process-applicable oxidant (2):
• treatment effectiveness,
• cost,
• ease of handling,
• compatability with preceding or subsequent treatment
steps, and
• nature of the oxidation operation.
Chlorine can be introduced to the waste stream as either sodium (or
calcium) hypochlorite or chlorine gas. Chlorine gas is usually more economi-
cal and, therefore, is used but is so reactive that it is difficult to contact
B25-1
-------
Appendix B25
Reduced Nitrogen and
Sulfur Chemical Oxidation
with large wastewater flows. For this reason the gas is contacted with a
sidestream for which tighter process control is possible. Once chlorinated,
the sidestream is returned to the reactor and mixed with the full waste
stream.
When chlorine gas transfers to the solution phase, it rapidly hydrolyzes
to hypochloric acid which then disproportionates to the hypochlorite ion.
Essentially no free chlorine (Cl,) exists at pH greater than 3.
Optimal reaction conditions for most oxidations by chlorine are at high
pH such that the active oxidizing agent is OC1 . For chlorine gas contacted
with water at high pH, two equivalents of alkalinity are destroyed per equiva-
lent chlorine added, or 1.4 grams alkalinity as CaC03 per gram chlorine as
Cla. If sufficient buffer capacity is not available in the waste stream,
alkalinity must be added for pH control.
Ozone is a more powerful oxidant than chlorine. It is thermodynamically
unstable and must, therefore, be prepared on site by passing an oxygen-
containing gas stream through a high voltage discharge gap.
Ozone is poorly soluble, and contactors must be designed carefully to
ensure optimal transfer. The most commonly used contacting systems disperse
the ozone-containing gas stream directly into the liquid. Dp to 90 percent
transfer can be achieved (2). The combined factors of poor ozone solubility
and thermodynamic instability create some difficulty in correlating lab
studies to performance of full-scale systems. Therefore, full-scale exper-
ience provides the best indicator of the feasibility of ozonating a particular
chemical species. Exact designs must be determined on an individual basis.
B25-2
-------
Appendix B25
Reduced Nitrogen and
Sulfur Chemical Oxidation
2. Process Applicability
Chlorination of Ammonia (NH3)
Numerous reactions can occur between the chlorine (as hypochlorite) and
ammonia, including the following:
Formation of: monochloramine: Nflt + HOC1 = NH2C1 + H20 + H (1)
dichloramine: NH4 + 2 HOC1 = NHClz + 2 HiO + H (2)
or.
NHzCl + HOC1 = NHCli + HiO
(3)
trichloramine: Nflt + 3 HOC1 = NCls + 3 H20 + H (4)
or,
NHiCl + 2 HOC1 = NCI3 + 2 HiO (5)
or.
NHCla + HOC1 = NCI 3 + HiO
(6)
ammonia
Which of the above reactions are involved or their sequence in oxidizing
nia completely to molecular nitrogen (Na) is not known. However, empiri-
cal observations have indicated that 1) when all ammonia is reacted, the
chlorine demand is 7.6 grams C12 per gram NH3 (as N) , and 2) the primary end
product is molecular nitrogen (N2) with small amounts of nitrogen trichloride
(NC13) and nitrate
B25-3
-------
Appendix B25
Reduced Nitrogen and
Sulfur Chemical Oxidation
The following overall reaction is in agreement with empirical
observations (3):
Nflt + 1.5 HOC1 = 0.5 N» + 1.5 H20 + 1.5 Cl~ + 2.5 H+ (7)
Formation of nitrogen trichloride probably occurs by reaction 6, while forma-
tion of nitrate probably occurs according to the following series:
NHCla + H»0 = NH(OH)C1 + H+ + Cl~ (8)
NH(OH)C1 + 2 HOC1 = N0s~ + Cl~ + 3H+ (9)
The overall reaction indicates that for each gram of NH4 (as N) removed,
8.9 grams of alkalinity are destroyed. If sufficient buffer capacity is not
available in the waste stream, alkalinity as caustic soda (NaOH) or lime
(CaC03) must be added.
Complete oxidation of ammonia to molecular nitrogen proceeds very
rapidly. Kokoropol ilos (4) suggests a first order reaction rate with a rate
_i
constant (k) of 5.1 x 10* molarity , indicating a practically instantaneous
reaction.
dNH3 = k NH, (10)
dt
Many designs provide about one minute contact time in a plug flow reactor.
Optimal pH is between 6 and 7, and there is essentially no temperature effect
(5).
The following operating procedures are usually observed for chlorination
of ammonia (5). Chemicals to control pH are mixed with the concentrated
B25-4
-------
Appendix B25
Reduced Nitrogen and
Sulfur Chemical Oxidation
hypochlorite solution prior to its being routed to the contact chamber. This
step helps prevent pockets of extreme pH that are conducive to formation of
M>7 or NC13.
The contacting vessel is enclosed with diffused air injected to mix the
contents of the reactor and strip NC13, Na, and C02 from solution. Removal of
carbon dioxide from solution helps maintain pH. The stripped gases are vented
(NClj is explosive at about 0.5% by volume) and routed to gas treatment if
necessary, possibly thermal decomposition of NCI, (at 420 K) and caustic
scrubbing of other noxious gases.
Chlorination of Cyanide
Cyanide can be oxidized completely by chlorine to molecular nitrogen
and carbon dioxide. The reaction occurs in a three step process: cyanide
(CN ) is oxidized to cyanogen (CN ), cyanogen is hydrolyzed to cyanate (CNO ),
and cyanate is oxidized to molecular nitrogen and carbon dioxide.
CN + OC1 + H»0 = CN+ + Cl + 2 OH (11)
CN+ + 2 OH~ = CNO~ + H20 (12)
2 CNO~ + 3 HOCl" + H*0 = Na + 2 HCOl + 3 Cl~ + 3 H+ (13)
An overall stoichiometry suggests that 6.8 g C12 are required per gram
complete destruction of cyanide:
CN~ + 2.5 OCL~ + 0.5 H»0 = 0.5 N2 + HCOl + 2.5 Cl~ (14)
Only free cyanide (CN or HCN) can be removed by chlorination. Cyanide
complexed with metals including zinc, copper, nickel, and iron is resistant to
oxidation.
B25-5
-------
Appendix B25
Reduced Nitrogen and
Sulfur Chemical Oxidation
At all pfl levels, cyanide is oxidized almost instantaneously to cyanogen
(6). Hydrolysis of cyanogen to cyanate occurs more slowly, but for pH greater
than 8 and for all initial cyanogen concentrations, the reaction is essenti-
ally complete in less than 10 minutes. This relationship is depicted in
Figure B25-1.
Oxidation of cyanate to molecular nitrogen is rate limiting for the
overall removal of cyanide. Cyanate combines only with the acid species HOC1
(6). Only at pH less than about 9.2 is enough of the total hypochlorite
present as the acid species to provide sufficient driving force to complete
the reaction within a reasonable period. At pH 9.2, approximately 15 minutes
residence time is required to complete the reaction for presumably any initial
cyanate concentration (6). The pH/retention time relationship is depicted in
Figure B25-2.
To oxidize cyanide, separate reactors may be used so that each can be
controlled for the different environments optimal for hydrolysis of cyanogen
and oxidation of cyanate. Alternatively, all reactions may be completed in a
single reactor with the pH maintained at a compromise value of about 9. A
plug flow reactor can provide most efficiently the long residence times
required on a continuous flow basis.
Ozonation of Ammonia
Ammonia can be oxidized by ozone according to the following equation:
NH4+ + 4 03 = NOa" + 4 0» + HiO + 2 H (15)
B25-6
-------
0.1
«et<
!«tion
35
of
°Sen
vario
(6)
B25-:
-------
10.0 r-
rc
a,
9.5
9.0
8.5
8.0,
J_
I
I
I
10 20 30 40 50 60
Retention Time, Minutes
70
80
Figure B25-2. Retention time required for oxidation
of cyanate (6)
B25-8
-------
Appendix B25
Reduced Nitrogen and
Sulfur Chemical Oxidation
Ozone demand is 13.7 g ozone per gram ammonia nitrogen. Alkalinity is
consumed and must be supplied at 7.1 grams (as CaC03) per gram ammonia nitro-
gen if the waste stream does not contain sufficient buffer capacity to prevent
wide pH fluctuations.
The mechanism of ammonia ozonation is not certain. The rate increases
from about pH 7 to pH 9. However, at high pH, ozone decomposes so quickly
that ozone solubility in accordance with Henry's Law cannot be achieved (7).
Singer and Zilli (8) suggest a first order reaction between ozone and
ammonia.
T7NH+ = K NH+ (16)
dt 4 4
where: pH K (min )
7.0 2.8 x 10~3
7.6 1.1 x 10~2
8.4 1.9 x 10~2
9.0 4.6 x 10~2
The reaction proceeds slowly and should be completed at pH 9 where it is kine-
tically most favorably. For a waste stream containing 200 mg/1 ammonia at pH
9, 95 percent removal would require a residence time of about 410 minutes and
65 minutes for a complex mix and plug flow reactor, respectively. For 90
percent removal, the residence times decrease to 195 and 50 minutes, respec-
tively.
Ozonation of Cyanide
Ozonation of cyanide occurs by a two step process whereby cyanide is oxi-
dized to the intermediate cyanate which then is oxidized to carbon dioxide and
B25-9
-------
Appendix B25
Reduced Nitrogen and
Sulfur Chemical Oxidation
molecular nitrogen (9). The oxidation of cyanate is about five times slower
and therefore rate controlling (10,11). Ozone demand for the overall reaction
is 4.6 grams ozone per gram cyanide destroyed. Neutral pH is optimal (12).
CN~ + Oj = CNO~ + Oi (17)
2 CNO~ + H»0 + 3 03 = 2 HCO,~ + N» + 3 Oi (18)
overall: 2 CN + 5 Os + H20 = Na + 5 Oi + 2 HC03~ (19)
Fundamental kinetic determinations for cyanide destruction are of limited
value since the removal rate that can be achieved in an actual system is
highly dependent upon the contactors used. Stopka (12) reports that for a gas
feed of 1 percent, single turbine contactor, and cyanide concentration of 50
mg/L with nickel catalyst, 12 minutes were required to complete the oxidation.
However, for a 2 percent ozone feed and two-stage turbine contactor with 1/3
of the total ozone dose injected into the second turbine, only four minutes
were needed to oxidize the same waste stream - all cyanide was completely
oxidized except for a cyanate residual of about 0.05 percent.
Oxidation of Reduced Sulfur Species
Little experience is reported in the open literature for chemical oxida-
tions designed specifically to remove sulfur species. Each waste stream must
be considered on an individual case basis (13).
For chlorination, thermodynamic data indicate that oxidation of sulfur
species is highly favorable. While a high level of thermodynamic favorability
does not ensure an acceptable reaction rate, for most simple electron trans-
fers between inorganic species, a reasonable reaction period is associated
with favorable thermodynamics (2).
B25-10
-------
Appendix B2S
Reduced Nitrogen and
Sulfur Chemical Oxidation
3. Process Performance
Individual chemical species vary considerably in their amenability to
oxidation by a particular oxidant. Seemingly conflicting data are often
reported in the open literature. To make a meaningful assessment of litera-
ture data, details of the reaction conditions such as pH, dose of oxidant, and
residence time need to be considered. Unfortunately, complete information is
typically not given. This information is especially needed in the case of
oxidation by ozone. Since ozone is poorly soluble in water and decomposes
rapidly, the configuration of the contactor used is critical to performance.
Not all the mass flow reported as ozone feed will necessarily be available to
oxidize wastewater constituents. The amount available depends on the dosing
technique. Despite limitations, literature data are suitable to broadly
assess the potential of chemical oxidation for a potential application.
Ammonia is very rapidly oxidized by chlorine, and in most applications,
given optimal conditions, nearly all ammonia should be removed. This perform-
ance is supported by experience obtained in disinfection of potable and other
waters. In this application, chlorine is added until there is a residual to
maintain the disinfecting condition. Before the residual develops, the chlor-
ine demand of the water is satisfied, and most of this chlorine demand is
exerted by ammonia and related species.
Oxidation of ammonia by ozone occurs much more slowly than by chlorine.
Little data have been reported in the literature.
Tables B25—1 and B25—2 present data for oxidation of cyanide by chlorine
and ozone, respectively. Cyanide removals exceeding 99 percent can be found
in both tables; however no information is given as to whether cyanide removal
implies an end product of cyanate or molecular nitrogen.
B25-11
-------
TABLE B25-1. SURVEY OF INDUSTRIAL EXPERIENCE RECORDED IN THE OPEN LITERATURE
FOR CHLORINATION OF CYANIDE
Industry Type
CN . mg/L
Influent
Effluent % Removal Reference
Inorganic chemicals
Ore Mining and dressing
Plating
Plating
Blast Furnace "A" CN
Blast Furnace "B" CN
6.8
68.3
700
32.5
3.38 + 2.5
3.56 + 4.53
0.01
0.13
0.0
0.0
99
99
99.9
99.9
0.12 + 0.24 96%
0.000 + 0.00 100%
1
1
17
18
19
19
B25-12
-------
TABLE B25-2. INDUSTRIAL EXPERIENCE OZONATING CYANIDE (1)
Ozone Feed
mg/L g/tr L/min
Ore Mining and Dressing
3200
3 9.5
3 9.5
6 4.9
Electroplating
143
143
143
143
143
143
143
143
143
143
64.4
143
64.4
64.4
64.4
64.4
173
130
195
195
195
29.7
35.2
PH
7.9
9.4
8.9
10.1
11.8
8.8
9.1
8.3
11.3
9.8
10.8
9.0
11.9
10.0
8.9
9.8
9.8
8.7
7.0-8.0
9.5
CN ,
Influent
0.90
0.355
0.163
0.195
38.4
34.2
33.0
32.8
32.5
32.1
32.0
31.5
30.2
29.5
13.0
28.5
6.5
5.6
5.3
5.3
63.0
38.0
37.5
36.3
29.0
15.2
12.9
mg/L
Effluent
<0.020
0.020
0.018
0.095
0.22
0.34
0.60
0.38
0.60
0.55
0.54
0.52
0.60
0.62
0.02
0.62
0.08
0.12
0.30
0.04
0.52
0.23
0.35
0.21
0.080
0.080
0.080
% Removed
>97
94
89
51
99
99
98
99
98
98
98
98
98
98
>99
98
99
98
94
99
99
99
99
99
>99
99
99
B25-13
-------
Appendix B25
Reduced Nitrogen and
Sulfur Chemical Oxidation
Little data explicitly addressing oxidation of sulfur species could be
located. However, each of the reduced sulfur species (sulfide, sulfite, thio-
sulfate, and thiocyanate) is believed to be very easily oxidized by either
chlorine or ozone. Sulfide is reportedly oxidized rapidly by chlorine (14).
Low level concentrations have been removed completely in conjunction with nor-
mal chlorination of secondary wastewater effluent (15). Oxidation of thiosul-
fite occurs so quickly that this species is used frequently as a scavenger to
dechlorinate chlorinated wastewaters. In this application, reference 5 sug-
gests that no separate reactor is needed because the brief residence time
needed can be provided in closed conduits or other conveying structures. No
data related to thiocyanate oxidation by chlorine could be found, but thio-
cyanate is easily oxidized by ozone (11,16).
4. Secondary Waste Generation
The only secondary waste stream directly attributable to chemical oxida-
tion is possible loss of volatile species. In particular, nitrogen tri-
chloride from chlorination of ammonia and hydrogen cyanide from oxidation of
cyanide may volatilize in sufficient quantity to require the reactors be
covered.
A variety of chemical species other than the direct oxidation products
are found in the wastewater stream following chemical oxidation. Much of this
residual results because the oxidant simultaneously attacks constituents in
the waste stream other than reduced nitrogen and sulfur species. The extent
of this concurrent activity will depend largely on the comparative amenability
of individual species to oxidation. Certain of the constituents may be
incompletely oxidized and thereby converted to an objectionable form.
B25-14
-------
Appendix B25
Reduced Nitrogen and
Sulfur Chemical Oxidation
For all oxidations, chemicals for pH control remain, increasing the IDS
of the waste stream. In the case of chlorination, residual chlorides increase
TDS. No such problem occurs with ozone since it decomposes to oxygen with a
half life of 20 to 30 minutes (1). However, until excess ozone decomposes, it
remains a powerful oxidant and, therefore, in a potentially toxic condition.
5. Process Reliability
Chemical oxidation typically requires simple equipment and has been
proven to be highly reliable in demonstrated applications (1). Disinfection
of potable and other waters by chlorination is standard practice in the Dnited
States and has proven to be reliable. Similar reliability has been experi-
enced in Europe where most disinfection of water is done by ozonation.
6. Process Economics
In general, the economics of chemical oxidation favor the treatment of
wastewater streams with low concentrations of contaminants. The process is
more favorable as a polishing step than for treatment of streams with high
pollutant concentrations. Presented in Table B25-3 are chemical requirements
for oxidizing reduced nitrogen and sulfur species and the estimated annual
cost of the oxidant.
Installed equipment costs for wastewater chlorination as a function of
chlorine feed rate were obtained from Reference 20. Storage and feed system
costs are included. Figure B25-3 shows these installed equipment costs,
updated from October 1978 to first quarter 1980 dollars by the CE index ratio
B25-15
-------
TABLE B25-3. CHEMICAL REQUIREMENTS FOR STOICHIOMETRIC ADDITION OF CBXORINE/OZONE
Cd
Stoichiometric
dose, kg ozidant
Species per kg species
Chlorination
NH4+(as N)
CN~
BS~(as S)
S0i=(as S)
Sa03=(as S)
Ozonation
NH4+(as N)
CN~
HS~ (as S)
S09=(as S)
SaOs'Us S)
7.6
6.8
8.8
2.2
4.4
13.7
4.6
6.0
1.5
3.0
Alkalinity, Annual Cost of Oxidant,
kg (as CaCOs) Optimal $ per year per mg/L
per kg species pH removed from a 1 m*/hr stream
8.9 6-7
9
1.6 c
d
1.6 c
7.1 9
7
1.6 c
c
1.6 c
11
8.4
12
3.0
6.1
82
28
36
8.9
18
Alkalinity required to maintain reaction pH - does not include alkalinity demand for
chlorine gas contacted with water.
Unit cost of chlorine: $160/Mg
No data available
jot critical
Ozone cost is for electrical power. Assumptions include: 80% transfer efficiency
all transferred Oj reacts); air feed, 22 kWhr/kg 03; and i0.025/kWhr.
-------
Installed Equipment Costs, $/(kg/day)
CO
ro
t_n
I
H-
OQ
ro
ca
N3
On
I
LO
ho 3
O cn
ft)
O-
c
H-
•a
ro
n
o
en
O
H-
0
PJ
It
H-
O
0)
•-<
CO
rt
§
CO
-------
Appendix B25
Reduced Nitrogen and
Sulfur Chemical Oxidation
of 1.16, as a function of chlorine feed rate. Three design options are con-
sidered: 1) ton cylinders, 2) on-site storage with bulk rail delivery, and
3) direct feed from a rail car. The original reference can be consulted for
details.
Ozone system installed equipment costs for ozone generation, dissolution,
and recycling were taken from Reference 20. For ozone generation capacities
of over 45 kg/day, the costs include oxygen generation costs. Figure B25-4
shows these installed equipment costs, updated from October 1978 to first
quarter 1980 dollars by the CE index ratio of 1.16. If excess oxygen capacity
exists in a synthetic fuels plant, the reported costs could be lowered signi-
ficantly. Costs of contacting chambers are listed separately (Figure B25-5)
since desired contact times may vary depending on wastewater characteristics.
B25-18
-------
10,000
First Quarter 1980 S
.21,000
100
I I I I I I
1
I I I I I
10
100 2 4
Ozone Feed Rate, kg/day
1000
Figure B25-4. Installed equipment cost for ozone generation,
dissolution, and recycling equipment (20)
B25-19
-------
1,000
7
First Quarter 1980 $
100
10
i i i i i i i
i i i i i i i
10
7 100
7 1000
Ozone Contactor Vessel Volume, m3
Fiaure B25-5. Installed equipment cost for ozone contactor
vessels (20)
B25-20
-------
Appendix B25
Reduced Nitrogen and
Sulfur Chemical Oxidation
7. References
1. D. S. Environmental Protection Agency. Treatability Manual,! Volume III,
Technologies for Control/Removal of Pollutants. EPA 600/8-80-042c, July
1980.
2. Weber, Walter J. Jr. Physiochemical Processes for Water Quality Control.
Wiley-Interscience, New York, New York, 1972.
3. Culp, Russell L., et al. Handbook of Advanced Wastewater Treatment. Van
Nostrand Reinhold Co., New York, New York, 1978.
4. Kokoropoulos, P. and G. P. Manos. Kinetics as Design Criteria for Past
Chlorination. American Society of Civil Engineers Journal of
Environmental Engineering Division, 99, 73, 1973.
5. D. S. Environmental Protection Agency. Process Design Manual for
Nitrogen Control. 1975.
6. Chamberlain, N. S. and H. B. Synder, Jr. Technology of Treating Plating
Wastes. Proceedings of the 10th Industrial Waste Conference, Purdue
University, 1956.
7. Stumm, W. Ozone as a Disinfectant for Water and Sewage. Journal Boston
Society of Civil Engineers, 45, 1958 in Singer, Philip C. and William B.
A. Silli, Ozonation of Ammonia in Wastewater, Water Research. 9, 127,
1975.
8. Singer, P., C. and W. B. Zilli. Ozonation of Ammonia in Wastewater.
Water Research, 9, 127, 1975.
9. Zeevalkink, J. A. et al. Mechanisms and Kinetics of Cyanide Ozonation in
Water. Water Research, 14, 1375. 1980.
10. Zabban, W. and R. Helwick. Cyanide Waste Treatment Technology - The
Old, The New, and The Practical, Plating and Surface Finishing. August,
1980.
11. Rice, R. G. and M. E. Browning. Ozone for Industrial Water and
Wastewater Treatment, A Literature Survey. Robert S. Kerr Environmental
Research Laboratory, D. S. EPA, EPA 600/2-80-060, 1980.
12. Stopka, K. Advanced Ozone Technology for Cyanide Destruction in
Electroplating. Plating and Surface Finishing, May 1980.
B25-21
-------
Appendix B2S
Reduced Nitrogen and
Sulfur Chemical Oxidation
13. Watkins, J. P. Controlling Sulfur Compounds in Wastewater. Chemical
Engineering, 84, 22, 61, 1977.
14. Chen, K. Y. Chemistry of Sulfur Species and Their Removal from Water
Supply, Chemistry of Water Supply, Treatment, and Distribution. Ann
Arbor Science Publishers, Inc., Ann Arbor, Michigan.
15. Nagano, J. Oxidation of Sulfides During Sewage Chlorination. Sewage and
Industrial Wastes. 22, 884, 1950.
16. Chudnov, A. P., et al. Oxidation of Thiocyanate in Aqueous Solutions by
Ozone. Koks Khim (Russ.) 5, 38, 1976, Chem. Abs., 85, 166084K, 1976.
17. Hansen, N. H. Design and Operation Problems of a Continuous Automatic
Plating Waste Treatment Plant at the Data Processing Division, IBM,
Rochester, Minnesota. Proceedings of the 14th Industrial Waste
Conference, 1960.
18. Watson, E. S. Treatment of Complex Metal Finishing Wastes. Sewage and
Industrial Wastes, 26, 182, 1954.
19. Wong-Chong, 6. M. and S. C. Caruso. Evaluation of EPA Recommended
Treatment and Control Technology for Blast Furnace Wastewater. American
Society of Civil Engineers Journal of the Environmental Engineering
Division, 104, 305, 1978.
20. U.S. Environmental Protection Agency. Estimating Water Treatment Costs,
Volume 2, Cost Curves Applicable to 1 to 200 Mgd Treatment Plants. EPA
600/2-79-162b, August 1979.
B25-22
-------
APPENDIX B26
FORCED EVAPORATION: VAPOR COMPRESSION EVAPORATORS
Forced evaporation processes are used to reduce the volume of wastewater,
and in most cases, to recover the evaporated water. The recovered water is
usually high quality and can be reused after little or no additional treat-
ment. Nonvolatile components in the wastewater are concentrated into a small
waste brine which may need further treatment. Volatile components in the
wastewater could degrade during evaporation and may be volatilized with the
water. If the evaporated water is recovered, additional treatment may be
needed before it can be reused due to the presence of the original or
degraded volatile components.
Many types and arrangements of evaporators are possible for reducing the
volume of wastewater and concentrating nonvolatile components. The two prin-
cipal classes are multi—effect evaporators and vapor compression evaporators.
Vapor compression evaporators have lower energy consumption and operating
costs than comparable multi-effect evaporators, while multi-effect evaporators
tend to have lower capital costs. Some evaporators have design features that
eliminate scaling as water is evaporated: only vapor compression evaporators
with these nonscaling features are discussed below. The use of these speci-
alized vapor compression evaporators is growing due to their lower energy
costs and absence of scaling. For a specific application though, the use of
multi-effect evaporators may be less costly if excess steam is available,
i.e., if there really is no cost burden associated with the required steam.
1. Process Description
A typical flow diagram of a nonscaling vapor compression evaporation
system, also known as a vapor compression distillation system, is shown in
Figure B26-1.
B26-L
-------
VENT
Evaporator
FEED
oa
ro
Steam
Compressor
CONCENTRATED
HASTE TO
DISPOSAL
PRODUCT
Product
Pump
Recirculation
Pump
Figure B26-1. Vapor compression evaporation system (1)
-------
Appendix B26
Forced Evaporation
The wastewater feed is first treated in a feed tank to adjust the pH to
between 5.5 and 6.5. Sulfuric acid is typically added to convert bicarbonate
to carbon dioxide for removal in a deaerator. The waste stream is then pumped
through a feed-product heat exchanger to raise its temperature to about its
atmospheric boiling point. The heated feed then enters a counterflow packed
column deaerator which strips carbon dioxide, nitrogen, and oxygen. Stripping
steam may be supplied as a bleed from the evaporator (1). Alternately, a
vacuum deaerator may be used (2).
Next, the feed enters the evaporator sump and is mixed with concentrated
slurry. The brine slurry is constantly circulated from the sump to the top of
the evaporator tubes. At the top of the vertical-tube, falling film evapora-
tor, the brine slurry is distributed to the inside wall of the tubes. The
slurry flows as a film on the inside of the tubes down to the sump. As the
slurry falls down through the tubes, part of the slurry water is vaporized by
the steam condensing on the outside of the tubes. The vapor evaporated from
the slurry is withdrawn through a mist eliminator to remove entrained droplets
before entering a compressor suction line. The vapor steam enters the com-
pressor where it is compressed to about 14 to 56 kPa. This pressure raises
the condensation temperature of the steam to about 3 to 12 K above the boiling
temperature of the recirculating brine. The steam is then fed to the shell
side of the evaporator tube bundle. As the steam condenses, its release of
heat causes more water to evaporate on the inside of the tubes. The product
water (condensed steam) is pumped from the evaporator shell through the feed
reheater to return as much heat as possible to the process before it is
recycled or sent to further treatment. The brine solids concentration in the
sump is continuously monitored. To maintain a constant slurry concentration
of about 200,000 to 400,000 mg/L total solids, a waste brine slurry is con-
tinuously removed (1, Personal communications with B.E. Heimbigner, August 20,
1979 and R.E. Hervey, September 10, 1982).
B26-3
-------
Appendix B26
Forced Evaporation
Scale formation on the evaporation side of the heat transfer surfaces is
avoided by preferential precipitation of calcium sulfate and silica on seed
crystals in the slurry. Also, the small temperature difference across the
heat transfer surfaces ensures that vaporization will occur at the brine-vapor
interface rather than at the brine film-wall interface (3). In some in-
stances, scale inhibitor must be added to the feed to avoid scaling as the
feed is preheated (1).
The vapor compression evaporation system must be constructed of corrosion-
resistant materials. The evaporator feed heat exchanger tubes and shell are
typically made of titanium. The sump, evaporator, and piping are typically
made of 316L stainless steel. Inspection of corrosion levels in existing
facilities indicates that the design lifetime of 30 years will be achieved
(Personal communication with D.E. Hervey, September 10, 1982).
2. Process Applicability
No specific limits of process applicability are reported in the open
literature. Vapor compression evaporation is most suitable for treating
wastewaters with little volatile matter since any vaporized volatiles would
contaminate the recovered water.
Bench scale testing of "typical" wastewaters from synfuels facilities
(after the wastewater has been treated by phenol extraction and ammonia
recovery) has been performed. In these tests, most of the residual ammonia
but little of the phenols were carried over into the recovered water (4). In
evaporation testing of typical process condensates, phenol and o-cresol were
carried over into the recovered water due to the formation of azeotropes with
water (5).
B26-4
-------
Appendix B26
Forced Evaporation
Of critical importance in any evaporative system is the precipitation of
scaling or foaling minerals as water is heated and evaporated. Specific
limits cannot be stated without knowledge of the specific wastewater
chemistry.
3. Process Performance
Wastewater streams containing up to 50,000 mg/L of IDS can be treated by
vapor compression evaporation to yield a high purity water product containing
less than 10 mg/L (1, Personal communication with B.E. Heimbigner, August 20,
1979). However, as discussed in the previous section, the product water may
be contaminated by volatilized components. The water produced by vapor com-
pression evaporation of cooling tower blowdown may be suitable for boiler
makeup. The recovered water from the evaporation of refinery wastewaters,
despite the presence of organics and ammonia, may be suitable as plant makeup
(1). Vapor compression evaporation systems are almost always designed to
recover over 90 percent of the feed wastewater as product (3), and recoveries
of over 99 percent are possible (1).
The residual waste brine typically contains 200,000 to 400,000 mg/L of
total solids (Personal communication with B.E. Heimbigner, August 20, 1979).
The distribution of these solids between dissolved and suspended components
depends on the specific chemical composition of the wastewater stream.
The performance of vapor compression evaporation systems in various
applications are summarized in Table B26-1. All of the systems in Table
B26-1 achieve or are capable of achieving a "clean" water product that
contains less than 10 mg/L TDS. Most of the components, including total
solids, are concentrated in the waste brine, approximately as the ratio of the
wastewater feed rate to the waste brine rate.
B26-5
-------
TABLE B26-1. PERFORMANCE OF VAPOR COMPRESSION EVAPORATION IN VARIOUS APPLICATIONS
w
NJ
C^
I
CoBDoaltion. BC/L
Sodim
CalcluB
Nagnesloa
Alkalinity
Carbona te
Bicarbonate
Sulfate
Chloride
Fluoride
Ortho Phosphate
Total Phosphate
Silica (as S10,)
TDS
SS
Oil
Phenol
TOC
Aanonla N
Ijeldahl N
Arsenic
pH
Boiling point
rise, I
(1-5 »»/!.) (1).
°Dealgn recovery of
cDesign recovery of
Coollni Tower Slowdown'
Deslan Actual Desltn Actual
55 155 2.140 5,117
528 430 980 320
275 164 10.560 12.733
12
418 268 293
2,002 1,625 45,290 51,875
62 92 2,400 6,913
3.4 58
7
55 50 200 325
3,465 2.806 61,570 77,634
67,575 108.212
8.0-8.5 8.15 7.0-7.6 7.2
0.48
product water - 99%. Design electricity oona
product water - 95%. Design electricity cona
, Zeolite Softener and Oil Shale
Predicted Predicted Pilot Tests Pilot Tests
Feed Brine Feed Brine Feed Brine
681 67.800 2,319 57,975
110 603 620 610 5 5
21 2,100 270 6.750
7,000 100,000
26 0 107 0
1.100 87,082 4.705 84.823 1,200 35,000
480 48,000 1,903 47,575
30 600
4 200 126 200 40 300
2,422 205,785 10,050 198,010 10,000 240,000
35.510 53,260 100 6,000
5 8
7.5 1
56 1,800 2.000 150
27 26 1,500 10
29 1,150
5 150
6.5 6.5-7.5 7 6.5-7.5 9 11
2.32 2.20
actual) electricity
mptlon - 26 klh/«> of
imp t Ion - 26 kfh/B* of
-------
Appendix B26
Forced Evaporation
4. Secondary Waste Generation
The vapor compression evaporation system shown in Figure B26—1
produces: 1) a "clean" water product, 2) a small waste brine of one to ten
percent of the feed volume which contains nearly all of the nonvolatile
components, 3) a gaseous vent from the deaerator of gases and vapors such as
ammonia, carbon dioxide, nitrogen, and oxygen, and 4) a small purge of
noncondensable gases (not shown in Figure B26-1) from the evaporator shell.
As stated above, the waste brine may need further treatment and can be
discharged to a solar evaporation pond or thermal drying system for final
disposal.
The composition of the gaseous vent obviously depends on the waste-
water characteristics. No data on the composition of this stream are pre-
sented in the literature, and no treatment of the stream seems to be provided.
The purge from the evaporator shell should be negligible since most gases
initially in the wastewater are stripped and vented from the deaerator.
5. Process Reliability
Small vapor compression evaporators were developed about 100 years ago
and thousands have been used for desalting seawater (typically shipboard) and
brackish water. Hundreds have been applied in the chemical industries (7).
Twenty-three large systems designed and installed since 1971 have achieved
over 40 combined years of reliable performance (8). No specific reliability
measures are reported in the open literature. The process uses standard
equipment and operations which tends to indicate adequate reliability.
B26-7
-------
Appendix B26
Forced Evaporation
6. Process Economics
The unit installed equipment costs for the vapor compression evapora-
tion process are shown in Figure B26-2. The basis for the cost correlation
is shown in Table B26-2.
The installed equipment cost correlation is a correlation of costs for
five sizes as reported by a vendor (Personal communication with B.E.
Heimbigner, August 20, 1979). The vapor compression evaporaton system costed
includes:
• feed tank and chemical feed system,
• feed/product exchanger,
• deaerator,
• evaporator,
• pumps and compressors,
• piping and valves,
• electrical and instrumentation, and
• building to house unit.
Not included are:
• wastewater feed piping,
• brine slurry piping to disposal,
• product water piping, and
• electrical lines.
The correlation of installed equipment costs is:
— 5
Installed Equipment Cost, i = 454,694 + 37.41255a - 2.168089 x 10 a* (1)
where a is the product water rate in kg/hr.
B26-8
-------
2 100
1st Quarter 1980 $
4
= 2
10
" 7
Ul
c
I I I I I I I
1 i I I I I i
_] i i I i i I
10"
10s
Product Water, !cg/hr
Figure B26-2.
Unit installed equipment cost for vapor
compression evaporation systems (personal
communication with b. neiiiiDigner, August
20, 1979)
B26-9
-------
TABLE B26-2. BASIS FOR INSTALLED EQUIPMENT COSTS OF
VAPOR COMPRESSION EVAPORATION*
Cd
NJ
ON
1
M
O
Wastewater
m'/hr
16
79
160
390
790
Recovered
Product
Water6
kg/hr
15,300
76,500
153,000
383,000
765,000
Original
Turnkey
Cost
1979 i
950,000
4,250,000
6,950,000
14,000,000
20,000,000
Estimated
1979*c
742,000
3,320,000
5,430,000
10,940,000
15,630,000
Installed Equpment Costs
1st quarter 1980$
779,000
3,490,000
5,700,000
11,490,000
16,410,000
.Personal communication with B. Heimbigner, August 20, 1979.
Based on typical recovery of 96-98 percent of wastewater feed.
Assuming that Turnkey Cost includes engineering and construction at 25%
of installed equipment and fees at 3% of installed equipment.
-------
Appendix B26
Forced Evaporation
The curve plotted in Figure B26-2 is simply the above correlation divided
by the product water rate:
Installed Equipment Cost, i/(kg/h) = 454.694 + 37.41255 - (2)
a
_5
2.168089 x 10 a.
Although the vendor reported the original cost data to be accurate to +10 per-
cent, the correlation is inherently less accurate. Above a recovered water
rate of about 150,000 kg/h, the correlation predicts costs to within one per-
cent of the original data. At the lower end of the correlation (15,000 kg/h),
the correlation overestimates costs by as much as 30 percent.
The principal operating cost item is electricity for the compressor (the
principal user), pumps, and control needs. The electricity demand typically
ranges from 18 to 24 kWh/m* of product (2) although the systems in Table
B26-2 show demands of up to 26 kWh/m3 (1). The demand obviously varies with
the extent of vaporization and the characteristics of the wastewater; the im-
pacts of these factors are shown by the increase in the boiling point of the
concentrated brine. An average electricity demand is about 22 kWh/mJ of pro-
duct (Personal communication with B.E. Heimbigner, August 20, 1979.
Operating labor requirements are estimated to be 2 man-hours per shift
(Personal communication with B.E. Heimbigner, August 20, 1979).
Maintenance costs for general maintenance and replacement materials and
chemical costs total about one percent of the capital costs (Personal communi-
cation with B.E. Heimbigner, August 20, 1979). The chemical cost for acid
addition depend highly on the water chemistry but is not a significant cost
item and is included as part of the maintenance cost.
B26-11
-------
Appendix B26
Forced Evaporation
7. References
1. Stickney, W.W. and T.M. Fosberg. Putting Evaporators to Work: Treating
Chemical Wastes by Evaporation. Chemical Engineering Progress, April
1976. pp. 41-46.
2. Rogers, A.N. et al. Treatment of Cooling Tower Slowdown. Chemical
Engineering Progress, July 1981. pp. 31-38.
3. Get Zero Discharge with Brine Concentration. Hydrocarbon Processing,
October 1973. pp. 77-80.
4. Levine, S. Disposal of Coal Gasification Wastewaters by Evaporation.
Paper presented at Advances in Coal Utilization Technology IV April 20-
24, 1981 in Denver, CO, sponsored by Institute of Gas Technology.
5. Concentration Specialists, Inc. Feasibility Analysis of the
Concentration of Coal Conversion Process Condensate. Prepared for the
U.S. Department of Energy under Contract No. DE-AC02-78EV04943.
January 1980.
6. Mukhopadhyay, D. and T.M. Fosberg. Concepts for Recycle and Reuse of
Oil Shale Wastewaters. Paper for 82nd AIChE Winter Meeting, Orlando,
Florida, March 3, 1982.
7. Beesley, A.H. and R.D. Rhinesmith. Energy Conservation by Vapor
Compression Evaporation. Chemical Engineering Progress, August 1980.
pp. 37-41.
8. Resources Conservation Company. Energy-Efficient Systems for the Pulp
and Paper Industry Brochure.
8. Personal Communications
Heimbigner, B.E. (Regional Sales Manager, Resources Conservation
Company, Seattle, Washington) to T.E. Emmel (Senior Engineer, Radian
Corporation, Austin, Texas), letter dated August 20, 1979.
Hervey, D.E. (Vice President, Business Development, Resources
Conservation Company, Seattle, Washington) to W.C. Thomas (Senior
Engineer, Radian Corporation, Austin, Texas), letter and enclosures dated
September 10, 1982).
B26-12
-------
APPENDIX B27
DEEPWELL INJECTION
1. Process Description
Deepwell injection is a liquid waste disposal method in which the waste
is pumped into a suitable subsurface formation via a well. A deepwell dispo-
sal system consists of a surface facility for handling and pumping the waste
liquid, a specially constructed well or wells, and a confined subsurface dis-
posal reservoir.
Depending on the quality constraints applicable to a given disposal sys-
tem, some wastes may need pretreatment before they are disposed of by deepwell
injection. The purpose of pretreatment may be to remove specific toxic speci-
es from the waste or it may be to prevent plugging of the disposal zone forma-
tion and damage to equipment. Pretreatment steps might include some or all of
the following:
• Oil separation. The presence of oil is undesirable because it
can reduce permeability in sands or sandstones. Also, even
minute quantities of oil can cause fouling of surface and
subsurface equipment, particularly filters.
• Suspended solids removal. Solids in the waste can cause
plugging in the injection equipment and in the subsurface for-
mation. Plugging in the formation reduces its overall capa-
city and increases the surface injection pressure required for
suitable disposal flow rates. It is usually recommended that
all solids larger than 1 to 5 micrometers be removed by filtra-
tion and/or clarification (1,2).
• Neutralization. Wastes having high or low pH may cause
problems due to corrosion. Neutralizing the waste can reduce
costs of the materials of construction for the injection
tubing, well casing, and fittings.
B27-1
-------
Appendix B27
Deepwell Injection
• Removal of heavy metals. This is frequently needed to prevent
their precipitation after injection into the receiving forma-
tion. Scaling problems are usually investigated by conducting
mixing studies using reservoir fluid under down hole conditions
(Personal communication with T. Jones, September 10, 1982).
• Organic chemicals removal. Organic chemicals may be
undesirable because they provide nutrients that support
biological growth. Additionally, some organics are extremely
viscous and others are subject to polymerization; plugging of
the receiving formation could result if these were present.
Treatment methods which may be considered for these steps are discussed else-
where in this document.
Deepwell disposal of waste fluids has a long history in Texas and
Louisiana. Over 50 years ago the petroleum industry utilized injection wells
to dispose of salt water produced with oil and gas. The successful applica-
tion of brine disposal wells and improvements in well technology led to the
use of deepwells for the disposal of chemical and industrial wastewaters.
A proper waste disposal well design is fundamental to the protection of
the environment and is controlled by regulatory agencies. In general, an
injection well will have three concentric "strings" of steel casings plus
injection tubing. Conductor casing is set to a depth of 15 to 18 m to prevent
surface caving while drilling to surface casing depth. Surface casing is set
and cemented below the deepest fresh water zones and is designed to protect
these zones from encroachment or contamination by waste fluids. The long
string casing or protection string is set through the proposed injection zone
and extends to the surface. It maintains well integrity and provides a second
protective casing through fresh water zones. The long string casing is also
cemented to the surface providing additional isolation of fresh groundwater.
To eliminate corrosion of the protective casing, wastewater is pumped down
the injection tubing. The tubing is separated from the long string casing by
B27-2
-------
Appendix B27
Deepwell Injection
a fluid-filled annulus. Waste fluids cannot move up this annulus due to a
mechanical sealing "packer" set above the injection zones. A typical injec-
tion well is shown in Figure B27-1.
Maximum injection pressures are calculated by reservoir characteristics
and set such that no fracturing of the formation or confining strata will
occur. Operating pressures are usually determined by field testing, although
in some cases maximum operating pressures are limited by regulations. Typical
injection rates for existing industrial waste injection wells range from 17 to
230 mj/hr; injection depths are generally between 300 and 1800 meters (1).
Few wells deeper than 1800 meters have been constructed due to high costs and
because satisfactory injection zones can usually be found at lesser depth.
Injection zones for Class I wells must be below the deepest underground source
of drinking water in the area (1,3).
Many types of rock formations have sufficient porosity and permeability
to accept large quantities of injected liquid wastes. In practice, most wells
have been constructed to inject into sedimentary rocks (e.g., sandstone,
limestone, or dolomite). In order for limestone and/or dolomite formations to
be good candidates they frequently must be highly fractured. Formations above
and below the injection zone must be sufficiently impermeable to assure isola-
tion of the waste from underground drinking water sources. The formation
chosen must not have hydraulic communication with areas which contain valuable
resources such as oil, gas, or commercial brines. It must be emphasized that
a detailed geological and engineering study should be conducted to determine
the suitability of a specific site. Criteria to determine the feasibility of
a disposal zone are: 1) uniformity, 2) large areal extent, 3) substantial
thickness, 4) high porosity and permeability, 5) normal pressure, 6) saline
aquifer, 7) separation from potable water horizons, 8) adequate overlying and
underlying aquicludes (formations limiting vertical mixing of the waste with
other aquifers), 9) no poorly plugged wells nearby, and 10) compatibility
between the injected waste and the mineralogy of the reservoir (1).
B27-3
-------
INPUT -51
SURFACE
BORE HOLE '.'. ''
\ \a \
__A2A.
^XX
. / / ./ /
OUTER CASING
AQUICLUDE
TT
INNER CASING-
DISPOSAL
FORMATION
f
-VALVE
SURFACE CASING
\ \
/ / / T
CEMENT -
^MECHANICAL PACKER
I I i I
OPEN HOLE
B27-1. Typical disposal well (2)
B27-4
-------
Appendix B27
Deepwell Injection
2. Process Applicability
Deepwell injection of industrial wastes has been practiced for over 25
years, with a wide variety of wastes being disposed of. Major users of injec-
tion wells include chemical, petrochemical, and pharmaceutical companies; oil
refineries and natural gas plants; and the primary metals industry (1). Over
40,000 wells have been used to dispose of oil field brine (3,4). Most of
these wells are located in the South, Southwest, and Mideast portions of the
U.S., areas of the country with favorable geologic formations for deepwell
disposal (1). Experience gained in this area has formed the basis for the
development of technology for disposing of industrial wastes.
Deepwell injection may be a viable disposal alternative for synthetic
fuels facility liquid wastes, especially if local conditions make it impracti-
cal or impossible for discharge of treated wastewaters to surface water.
Wastes containing refractory organics and/or high levels of total dissolved
solids are potential candidates for deepwell injection. Local geological
conditions, rate of injection, volume of waste, and waste characteristics
should all be considered when evaluating deepwell disposal. A detailed geo-
logic and engineering study of the area should be conducted before a determin-
ation can be made regarding site suitability.
3. Process Performance
A well designed and controlled deepwell system injecting properly pre-
treated waste into a suitable receiving formation should be capable of dis-
posing of large volumes of liquid waste over a number of years.
The major advantages of deepwell injection are: 1) it is an ultimate
disposal method, 2) it requires little land, and 3) it permanently removes the
waste from contact with air, surface water, usable groundwater, or the
B27-5
-------
Appendix B27
Deepwell Injection
surface of the ground. Disadvantages of deepwell injection include the need
to have favorable geologic conditions available in the vicinity of the waste
source (however, the area near coal deposits tend to have favorable geologic
conditions) and pretreatment adds to the expense of waste disposal.
4. Secondary Waste Generation
Secondary wastes are not generated by the injection of the wastes.
However, they may be formed as a result of the pretreatment procedures
needed before the waste can be injected. Examples of such secondary wastes
might include oily sludges from oil separation, filtered solids, and/or flue
gas from waste incineration.
5. Process Reliability
Deepwell injection has been used to dispose of industrial wastes for over
25 years. It was estimated in 1981 that there were over 400 such wells in the
U.S., with some 250 operating (Personal communication with T. Jones, September
10, 1982). In addition, related experience has been provided by a large
number of wells injecting oil field brines. Performance data for a number of
the industrial wells has shown failures as well as successes (5). Problems
have included migration of waste to usable aquifers caused by injection into
shallow aquifer faults in confining strata or defects in well casings (Per-
sonal communication with T. Jones, September 10, 1982). Properly selected
sites and well-designed and operated injection wells should be capable of
isolating the waste from any potable water sources and, therefore, avoid their
degradation.
B27-6
-------
Appendix B27
Deepwell Injection
6. Process Economics
Costs of deepwell injection are dependent upon the depth of the well, the
waste injection rate, formation characteristics, and the extent of required
pretreatment. Major cost items are for contract drilling, power, and
labor (1).
Installed equipment costs, in first quarter 1980 dollars, for waste
injection wells were estimated at $330 to $490 per meter of well depth (Per-
sonal communication with T. Jones, September 10, 1982). Mechanical integrity
testing (which is required every 5 years) was estimated at $5,000 to 135,000
per test.
Electric power requirements for wastewater pumping can be calculated
using the following formula (6):
E = H i Q x t
P
where: E •= electric power usage, kWh/yr,
H = pumping pressure, MPa,
P
Q = flow rate, L/sec, and
t = pumping time, hr/yr.
Annual maintenance costs for a well which is 1600 m deep and has an
injection rate of 32 m*/hr were estimated at i60,000/well in 1979 dollars (1).
Updating this estimate to first quarter 1980 dollars by use of the CE index
ratio 1.08 gives an estimate of $65,000 per well.
7. References
1. Smith, M.E. Solid Waste Disposal: Deepwell Injection. Chemical
Engineering, April 9, 1979, pp. 107-112.
B27-7
-------
Appendix B27
Deepwell Injection
2. Liptak, B.V. Environmental Engineers Handbook, Vol. I, Water Pollution.
Chilton Book Company, Radnor, Penn., 1974.
3. Federal Register, Vol. 45, No. 123, June 24, 1980.
4. Conrad, E.T., and N.E. Hopson. Outlook for the Future of Deepwell
Disposal. Water Resources Bulletin 11, (2), p. 370, 1975.
5. U.S. Environmental Protection Agency. Review and Assessment of Deepwell
Injection of Hazardous Waste. U.S. EPA, EPA/600/2-77/027a, June 1977.
6. Perry, R.H., C.H. Chilton, and S.D. Kirkpatrick. Perry's Chemical
Engineer's Handbook, 4th Edition, McGraw-Hill, Inc., New York, 1963,
pp. 6—2.
8. Personal Communications
Letter from T. Jones, Golden StrataServices, Inc., to W.C. Thomas, Radian
Corporation, September 10, 1982.
B27-8
-------
APPENDIX B28
SURFACE IMPOUNDMENT
1. Process Description
Surface impoundments are widely used in industry to provide for temporary
or permanent retention of waste liquids, slurries, and sludges. Other names
for surface impoundments, such as holding basins, evaporation ponds, lagoons,
and settling basins are indicative of their wide variety of uses. They are
natural depressions or artificial excavations usually constructed with earthen
dikes surrounding the impoundment area. Impoundments may be lined with a
variety of natural or man-made materials, including clay, asphalt, concrete,
and various plastics, to minimize or prevent infiltration of the impoundment
contents into the underlying strata.
The important features of a wastewater impoundment are the dike,
liner(s), and leachate collection system. Not all impoundments require a lea-
chate collection system. For impoundments containing hazardous wastes, how-
ever, it is required.
Dikes serve two purposes. First, they contain the waste within the im-
poundment. They are usually designed with at least 0.5 meter of freeboard to
accommodate storm water falling on the impoundment area. Second, they serve
to divert surface runoff away from the impoundment.
The material used to line surface impoundments is a function of the size
of the impoundment (e.g., concrete is usually used only for ponds having a
surface area less than 4000 m») , availability (e.g., naturally occurring
clays), and chemical compatibility with the waste.
B28-1
-------
Appendix B28
Surface Impoundment
If the impoundment contains a waste which is found to be hazardous under
RCRA, leachate collection is required. Leachate collection can be accom-
plished by a system of perforated pipes and sumps in a layer of permeable sand
laid at the bottom of the impoundment. The drain system needs to be graded so
that water flows toward the collection system (1). Another feature that is
often incorporated into the design of a surface impoundment is a layer of
material placed over the liner to protect its integrity; this may be especi-
ally appropriate when a plastic liner is used.
The wastewater routed to an impoundment can be discharged or recycled,
or, in the case of an evaporation pond, allowed to evaporate in place. The
residual sludges or dry solids can be periodically removed for ultimate dis-
posal elsewhere, or the impoundment can be closed with the solids in place.
2. Process Applicability
Surface impoundments are a flexible disposal alternative having a wide
variety of applications. A number of waste streams from a synfuels facility
could be managed with surface impoundments and a number of preliminary plant
designs have included them. Surface impoundments can be a relatively inexpen-
sive and an effective technique for disposing of otherwise unprocessable
wastes; impoundments also require little maintenance.
The use of surface impoundments may be constrained in some areas because
of the large amounts of land required. If the impoundment is to serve as an
evaporation pond, the climatological conditions at the plant site must be con-
sidered; generally speaking, annual evaporation rates exceeding 0.5 m/yr are
required. Care must also be taken to 1) prevent leaching of contaminants into
groundwater and 2) avoid codisposing of incompatible wastes.
B28-2
-------
Appendix B28
Surface Impoundment
3. Process Performance
Surface impoundments are widely used for the temporary or ultimate dispo-
sal of a number of different industrial wastes. If properly designed and op-
erated, they should result in minimum environmental contamination.
Their suitability for specific applications depends on site-specific fac-
tors. For example, in order to successfully serve as an evaporation pond, a
surface impoundment must be located at a site having a sufficiently high net
evaporation rate. The major drawback to their use is the need for relatively
large areas of land.
4. Secondary Waste Generation
Leachate may potentially be generated as a secondary waste stream from a
surface impoundment operation. The quantity of leachate escaping the impound-
ment is highly dependent on site-specific factors such as the net evaporation
rate and the characteristics and integrity of the material used to line the
impoundment. The characteristics of the leachate will vary depending on the
waste properties (e.g., constituent solubility, degradability) and local at-
tenuation effects (e.g., soil cation exchange capacity and pH). Under unusual
circumstances, generally caused by mismanagement of an impoundment, runoff of
contained liquid waste is possible.
If the influent wastewater to an impoundment contains volatile contami-
nants, the ambient air above the impoundment will contain some of these same
volatiles. These contaminants are likely to be present at low levels because
of upstream waste treatment processes. Therefore, additional control of these
emissions is normally not necessary. Volatile components may also be gener-
ated as a result of chemical reactions which can occur if multi-sources of
wastewater are comixed in the pond.
B28-3
-------
Appendix B28
Surface Impoundment
In the event that the impoundment is used as a temporary disposal site
for sludges and other solid wastes, a permanent disposal site will be required
and transporting the wastes could result in some spillage. This could result
in some secondary emissions.
5. Process Reliability
Surface impoundment has been widely used for disposing of various indus-
trial and municipal wastes. When properly designed and operated, potentials
for degrading groundwater and surface water should be minimized.
The major design feature for minimizing migration of leachate is the
construction of liners. The long term integrity of either natural or
artifical liners has not been established.
6. Process Economics
The costs of surface impoundments are largely capital costs associated
with the purchase of land and the construction of the impoundment. Utility
and chemical costs are negligible, as are annual maintenance costs.
Important factors influencing impoundment capital costs are land
availability and costs; impoundment surface area, depth, and configuration;
site terrain; dike configuration; type and volume of earthwork involved;
lining type and quantity; and miscellaneous construction, engineering, and
contingency costs (2).
In the absence of specific information on a pond other than its surface
area, dike height, and liner material, estimates of installed equipment costs
for ponds may be made using Figure B28-1, which is based on information con-
tained in Reference 2. The estimated costs include pond construction using
B28-4
-------
I - Unlined
II - 0.5 mm PVC limns
III - Hypalon
2.4 m dike
6.1 m dike
1000
10,000
Impoundmenc Area, m
100,000
Figure B28-1. Installed equipment costs for surface
impoundments (2)
-------
Appendix B28
Surface Impoundment
heavy equipment, rolling of the impoundment bottom surface, hand dressing of
side slopes, liner anchor ditches, liner material, liner installation, and
liner cover material. Land, pumps, piping, and other auxiliaries are not in-
cluded. Costs from the reference in mid-1975 dollars were updated to first
quarter 1980 dollars using a CE index ratio of 1.42. Figure B28-1 presents
installed equipment costs versus flow to the pond for two dike heights (2.4 m
and 6.1 m) and three liner options (unlined, 0.5 mm PVC, and 0.76 mm Hypalon).
Costs for ponds with clay liners are estimated to be close to those for PVC,
at least for smaller pond sizes.
Land costs are very site-specific; they may range from less than $0.25/m2
to greater than $12/m2 depending on location.
7. References
1. U.S. Environmental Protection Agency. Landfill-Surface Impoundment
Performance Evaluation Manual. June 1980.
2. Parker, C.L. Estimating the Cost of Wastewater Treatment Ponds.
Pollution Engineering, November 1975, pp. 32-37.
B28-6
-------
APPENDIX Cl
LANDFILL
1. Process Description
In landfill ing, waste is brought to the disposal site (usually by truck
or conveyor), spread in layers, and compacted with heavy equipment. In most
landfills the waste is covered with a thin layer of soil at the end of the
working day. The process is repeated until the desired depth is reached or
the available area is filled. A final cover of soil is then added. The fin-
ished site is either revegetated or prepared for other end-uses.
The principal methods used in landfill ing are area-fill and trench-fill.
In the area-fill method, the site is usually located in a natural depression
such as a valley but can also be an abandoned surface mine or any level area.
The waste is usually placed entirely above the original ground surface, and as
such, this method is particularly applicable in areas with shallow groundwater
table or bedrock. Since no excavation is required, one of the major landfill
construction cost items is eliminated. However, because the sidewall contain-
ment (available in trench-fills) may be lacking and landfill equipment must be
supported atop the waste in most fills, waste stability and load bearing
capacity must be good.
In the trench-fill approach, a trench (cell) for placement of waste is
constructed by excavation of native soils. The method requires that the
groundwater table and bedrock be sufficiently deep so as to allow excavation
and still maintain sufficient distance between the bottom of the trench and
the groundwater. Typical trench depths range from 3 to 10 m, and are
generally affected by the depth of the groundwater table, waste characteris-
tics (e.g., moisture content and physical strength), and capacities as well as
capabilities of landfill equipment.
Cl-1
-------
Appendix Cl
Landfill
2. Process Applicability
Landfills, which have been widely used for the disposal of municipal
refuse and a range of industrial wastes, would be suitable for disposal of
most solid wastes and sludges from a synthetic fuels facility. FGD sludges
and wastewater treatment sludges will likely require pretreatment such as
mixing with gasification/boiler ash or other commercially available fixation
methods to increase stability prior to landfilling. Area-fill is the pre-
dominant method used in existing power plants and in proposed synfuels
facilities (1,2,3,4,5,6).
3. Process Design
Factors that should be considered in designing a landfill and evaluating
landfill sites are summarized in Table Cl-1. A properly designed and operated
landfill would minimize environmental contamination. Fugitive dust emissions
can be a problem in landfilling wastes from synthetic fuel facilities. Land-
fill designs should include provisions such as water spray and daily cover for
controlling fugitive dust emissions. Leachate migration from the landfill
site presents a significant possibility of environmental degradation through
the contamination of surface or groundwater. Leachate is generated as a
result of the movement of precipitation and surface run—on through the site,
dissolving or suspending waste constituents as it passes through.
Figure Cl-1 shows the water routing through an uncontrolled landfill. As
shown, the infiltration fraction of precipitation is the major contributor to
leachate generation. The infiltration is a function of the surface conditions
and the climatological characteristics of the site and can be estimated by a
water balance (7):
Cl-2
-------
TABLE Cl-1. SITE-SPECIFIC FACTORS TO BE CONSIDERED
FOR LAND DISPOSAL OPTIONS
Climatological
• Wind conditions (e.g., speed, directional flux, dilation factors,
humidity, temperature)
• Precipitation (e.g., annual precipitation, storm intensity, snow
contributions)
• Evapotranspiration rate (e.g., season variations)
Geologic Factors
• Physiographic features (e.g., runoff coefficient, slope, drainage
patterns - dendritic trellis, annular, etc., erosional
features)
• Surface and subsurface geology (e.g., outcrops, bedrock features,
strike and dip of the bedrock, rock composition)
• Soil types (e.g., CEC capacity, texture, permeability,
stratification, homogenous vs. heterogenous deposition,
chemical composition, percent of humic material)
• Seismic factors (e.g., ground shaking or rupture)
Hydrogeologic Factors
• Drainage patterns
• Stream flow (e.g., velocity, perennial vs. intermittent, effluent or
influent source)
• Surface waters (e.g., tidal effects, recharge vs. discharge points)
• Vadose zone (e.g., depth, moisture content, hysteresis patterns,
storage capacity)
• Groundwater (e.g., depth, number of aquifers and relationships,
confined or artesian, nature of confining layer(s), capillary
fringe characteristics)
(Continued)
Cl-3
-------
TABLE Cl-1. SITE-SPECIFIC FACTORS TO BE CONSIDERED
FOR LAND DISPOSAL OPTIONS (continued)
• Piezometric surface (e.g., streamline flux patterns due to seasonal
or event related phenomena, influence of recharge/discharge
areas, streamline anomalies)
• Water quality (e.g., background vs. undersite vs. downgradient,
water uses - consumptive, irrigation, recreation, point source
contributors and their respective hydrogeologic pathways)
• Floodplain (100 year flood) (e.g., aerial flooding limits, degree of
localized streamline pattern reversal, erosional consequences)
• Wetlands (e.g., recharge vs. discharge source, wetland/groundwater
continuity and pathway)
• Recharge and discharge areas (e.g., proximity to disposal area,
volume of flow)
Land Use Factors
• Historic significance • Demographic setting
• Transportation corridor (access) • Geopolitical impact
• Beneficial uses • Ultimate land use
• Cost
Cl-4
-------
Actual Evapotranspiration
A (AET)
Precipitation (P)
Ground
Surface
Was/te
Cell
Native Soil
-Surface Runoff (R/0)
Infiltration (I)
Waste Moisture Storage (ST)
Leachate
Figure Cl-L. Water balance in a landfill
Cl-5
-------
Appendix Cl
Landfill
I = P - R/0 - AST - AET (1)
where: I = infiltration,
P = precipitation,
R/0 = runoff = C . x P,
r/o
c . = runoff coefficient (fraction of water which does not permeate
the surface and leaves the area),
AST = change in waste moisture storage, and
AET = actual evapotranspiration.
Leachate control measures consist of 1) diversion channels to prevent
surface runoff from entering the site, 2) liners to retard leachate movement,
3) a leachate collection and treatment system, and 4) final cover for the site
to limit infiltration of precipitation.
Diversion channels should be incorporated into the initial design of the
landfill and constructed before the site begins accepting waste. RCRA hazard-
ous waste regulations require that surface runoff diversions capable of
diverting 10-year 24-hr storms be constructed.
Liners for landfills can be made of clay, concrete, asphalt, or plastics.
The original Resource Conservation Recovery Act Subtitle C (8) proposed regu-
lations described two design configurations for hazardous waste landfills.
One of these required a single 1.5 m liner of compacted clay with permeability
_7
less than 10 cm/sec, the other required two liners consisting of 0.9 m com-
pacted clay and a synthetic plastic liner having a minimum thickness of
0.5 mm (see Figure Cl-2). However, these regulations are being revised, thus,
hazardous waste landfill design cannot be defined. From a technical point of
view, a properly designed landfill (for hazardous or nonhazardous waste)
should minimize leachate migration from the site. Depending on site-specific
factors and waste characteristics, the liner design may range from no liner
Cl-6
-------
30 cm
LEACHATE COLLECTION SYSTEM
NATURAL OR CONSTRUCTED SOIL LINER
K £ ^Q-7cm / sec
NATURAL SOIL == 10~4
DESIGN I
-30 cm
NATURAL SOIL^IO"4 cm/sec
LEACHATE COLLECTION
AND REMOVAL
SOIL LINERS10-7 cm/sec
MEMBRANE LINER AND
GRANULAR BEDDING
LEACHATE DETECTION
AND REMOVAL SYSTEM
DESIGN
Figure Cl-2. Hazardous waste landfill design based on
original RCRA proposed regulation (8)
-------
Appendix Cl
Landfill
for sites with deep groundwater tables, thick local clay layers, and net posi-
tive evapotranspiration rates to multilayers of liners consisting of artifi-
cial and/or clay liners of various thickness. Before choosing a liner, a
determination should be made regarding the compatibility of the liner and the
wastes to be contained. Some liners have been in place for many years; how-
ever, data are generally not extensive enough to predict liner life expec-
tancy, given its dependence on factors such as type of waste contained and
installation procedures. The choice between clay and artificial liners fre-
quently is based largely on economics.
Leachate collection can be accomplished by a system of perforated pipes
and pumps in a layer of permeable sand laid at the bottom of the fill. To
accelerate the flow of water towards the collection system, the drain system
(sand layer) should be graded to 2° or higher (see Figure Cl-3a). In addi-
tion, it is desirable to design the thickness of the drain system to exceed
the maximum depth of water mound that may accumulate. Figure Cl-3b shows the
limiting case (of slope 2°) where the shape of the water mound is given by
(9):
(2)
The maximum value of h occurs at X — D/2 and is given by:
(3)
where: h = height of water mound, m,
L = leachate generation rate, m/sec,
D = distance between drains, m, and
k = permeability of drain layer.
Cl-J
-------
,
i
t i I I ,
, v I ,
-D-
(b)
Figure Cl-3. Geometry assumed for bounding solution for
effectiveness of sand drains (9)
Cl-9
-------
Appendix Cl
Landfill
There is a wide variety of equipment available for landfill operations.
These equipment fall into two general functional categories: those used dir-
ectly in landfill operations such as excavation of fill, spreading, and com-
paction of waste or cover materials; and those that perform supportive func-
tions. Examples of equipment belonging to the first group include crawler
dozers, crawler loaders, compactors, and draglines. The second group of
equipment include water trucks and sprinklers, salt spreaders, and mobile
firefighting equipment.
4. Secondary Waste Generation
A potential secondary waste stream generated from landfill ing of waste is
leachate. The quantity will be highly dependent upon the net local evapo-
transpiration rate and other factors. Leachate characteristics will vary
depending on the waste properties and local attenuation effects. Major
factors that affect the characteristics of leachate reaching underground or
surface waters include soil cation exchange capacities, sorptive capacities,
pH, redox potential, and waste solubility/degradability. There are no data on
leachate characteristics from synthetic fuel plant waste disposal facilities.
However, some coal gasification ashes have been subjected to the EPA extrac-
tion procedure. Table Cl-2 summarizes the results of these tests.
5. Process Reliability
Landfill ing has been widely used for disposing of various industrial and
municipal wastes. When properly designed and operated, the potential for
degrading groundwater and surface water should be minimized.
Cl-10
-------
TABLE Cl-2. CHARACTERISTICS OF GASIFICATION ASH LEACHATES USING THE EPA RCRA
EXTRACTION PROCEDURE
Leachate Concentration. ui/L
Process - Coal (ref)
Lurgl ash - Roaebud (10)
Lnrgi ash - Illinois No. 5 (10)
Lnrgi ash - Illinois No. 6 (10)
lellaan-Galnsha ash (10)
I ellaian-Galnsha dust (10)
Tezsco Slag (10)
Koppers-Totzek ash - Illinois No. 6 (11)
P Koppers-Totzek ash - Greek Lignite (11)
1
1—1 Koppers-Totzek ash pond deposits (11)
Koppers-Totzek Cyclone Dust - Illinois
No. 6 (11)
VelUan-Galnsha Ash (12)
Wellnan-Galnsha Dust (12)
Hi ley ash (12)
Riley Cyclone Dust (12)
Kosovo Ash (12)
Ag As
<0.2 3
1.6 3
1.4 4
<1 19
<1 33
<2 <2
<10 <400
<10 <400
<10 <200
<10 350
NA 19
NA 33
<0.5 33
-------
Appendix Cl
Landfill
6. Process Economics
Landfill costs depend on a number of site specific factors such as land
cost, size of site, availability of local liner material (clay), depth/height
of landfill, distance of disposal site from the plant, and liner requirements.
Typical unit costs for site construction are (Personal communication with
J. Schilli):
• Clearing and scrubbing, JO.SO/m1,
• Excavation, i3.50/m*.
• Clay compaction (liner), J»3/mJ,
• Sand compaction (liner), il/m1,
• Leachate collection system (perforated pipe installed 30 m apart at
$27/m + sumps and pumps), il/m2,
• Groundwater monitoring system, J2500/well,
• Cover material, Jl.50/m»,
• Dike construction, J3.50/m3,
• Revegetation, i0.75/m»,
• Each truck requires one operator per 8-honr shift,
• The leasing agent provides backup equipment at no additional cost,
• Average speed for the off-road trucks is 20 kilometers per hour,
• Unloading and loading of the trucks requires approximately 10
minutes each,
• The following round trips were utilized in the cost calculations:
1.6 kilometers 30 minutes
4.8 kilometers 45 minutes
9.6 kilometers 70 minutes
19.2 kilometers 90 minutes
• Costs presented do not incorporate any cost for road maintenance,
supervision, or overhead, and
Cl-12
-------
Appendix Cl
Landfill
i
• Synthetic liners (installed cost):
0.51 mm nonreinforced PVC, $2.30/m*
0.76 mm nonreinforced PVC, J3.20/m*
0.91 mm reinforced hypalon, t5.10/ma
2.0 mm high density polyethylene, i7.20/m*.
Table Cl-3 summarizes the capital cost for the major landfill equipment.
As shown, depending on the size and type, the landfill equipment cost ranges
from $46,000 to $152,000. In general these machines are expected to have a
useful life of five years or 10,000 operating hours.
Based on the above unit costs for liners, the liner construction cost for
the single and double liner landfill presented in Figure Cl-2 would be
$8.25/m* and J8.90/m*, assuming no excavation is required and the 0.51 mm
nonreinforced PVC liner is used in Design II. If the 2 mm liner is used in
Design II, the cost would be increased to Jl3.8/ma. Other major capital costs
include fencing around the site, landfill equipment, and buildings for labora-
tories, sanitary facilities, and offices. Major operating costs for land-
filling are for labor, waste analysis, and waste transportation.
Figure Cl-4 presents cost estimates for transporting waste by truck as a
function of distance from the coal conversion facility to the disposal site.
As shown, the unit cost (J/Mg) more than doubles if the round trip distance is
increased from 5 km to 20 km. Figure Cl-4 is constructed based on the fol-
lowing assumptions:
• Trucks are leased.
• Up to a haul distance of 5 kilometers one way, off-road trucks are
utilized similar to a Caterpillar Model 3406.
Cl-13
-------
TABLE Cl-3. MACHINE CAPITAL COST (13)
Fully Equipped Machines
Flywheel
Machine type Horsepower*
Crawler dozer
Crawler loader
Rubber- tired
loader
<80
110-130
250-280
<70
100-130
100-130
150-190
150-190
<100
<100
120-150
120-150
Approximate
Weight Weight*
(Mg) (Mg)
<6.8
9.1-11
22-24
<9
11-15
11-15
15-20
15-20
<9
<9
10-12
10-12
8.6
14
30
10
14
14
20
21
7.7
8.2
10
12
Approximate
Cost0
($) Comment
46 ,000
83 ,000
152,000
46 ,000
65.000
70.000
100.000
107 ,000
46,000
50,000
72,000
78,000
landfill blade
landfill blade
landfill blade
GPBd— 0.76m»
GPB--1.5 m»
MPBe— 1.3 m»
GPB— 2.3 m»
MPB— 1.9 m»
GPB— 1.3 m»
MPB— 1.1 m»
GPB— 3.1 m»
MPB— 1.7 m»
Dragline
13
163,000-239,000 MPB
*1 horsepower = 0.75 kW.
Basic machine plus engine sidescreens, radiator guards, reversible fan, roll
bar, and either a landfill blade, general-purpose bucket, or multiple-purpose
bucket as noted.
^Updated to 1980 using the MAS equipment cost index of 303.3 and 659.6.
General purpose bucket.
Multiple-purpose bucket.
Cl-14
-------
400
n
M
i
6C
w
8
DC
•S
tu
a.
o
300
200
I 00
Distance, Kilometers
Figure Cl-4. Hauling cost vs. distance for nonhazardous waste
-------
Appendix Cl
Landfill
• For distances of 5 kilometers to 20 kilometers, street-legal trucks
are utilized. The trucks are tractor-trailer type dump trucks with
a capacity of 16 Hg.
• The leasing rate for off-road trucks is $8,320 per month based on an
eight-hour shift daily; for each additional eight-hour shift the
rate is increased by 50 percent (14). The operating cost was given
as $29.30 per hour which included fuel and maintenance but excluded
labor. The leasing rate for street-legal trucks is $3,923 per month
based on an eight-hour shift daily; for each additional eight-hour
shift, the rate is increased by 50 percent. The operating cost of
the truck and trailer is $20.37 per hour.
• Each off-road truck has a capacity of 27 Mg.
7. References
1. GAI Consultants, Inc. Coal Ash Disposal Manual. Electrical Power
Research Institute, FP-1257, Palo Alto, California, December 1979.
2. Coltharp, W.M., et. al. FGD Sludge Disposal Manual. Electric Power
Research Institute, FP-977, Palo Alto, California, January 1979.
3. U.S. Department of Interior. ANG Coal Gasification Company North Dakota
Project. Draft Environmental Impact Statement, March 17, 1977.
4. U.S. Department of Interior, Bureau of Reclamation. Final Environmental
Statement. Western Gasification Company (WESCO) Coal Gasification Pro-
ject, Washington, D.C., January 1976.
5. U.S. Department of Energy. Draft Environmental Impact Statement for
Solvent Refined Coal-I Demonstration Project. Newman, Daviess County,
Kentucky, January 1981.
6. Draft Environmental Impact Statement. Solvent Refined Coal-II Demonstra-
tion Project, Fort Martin, West Virginia, May 1980.
7. U.S. Environmental Protection Agency. Use of Water Balance Method for
Predicting Leachate Generation from Solid Waste Disposal Sites. SW-168,
1975.
8. Federal Register, Vol. 43, No. 243, 58947, December 18, 1978.
Cl-16
-------
Appendix Cl
Landfill
9. U.S. Environmental Protection Agency. Landfill and Surface Impoundment
Performance Evaluation Manual. Report submitted to U.S. EPA-MERL for
publication, June 1, 1980.
10. Yu, K.Y. and G. Crawford. Characterization of Coal Gasification Ash
Leachates Using the RCRA Extraction Procedure. Presented at the Sympo-
sium of Environmental Aspects of Fuel Conversion Technology-V, St. Louis,
Missouri, September 16-19, 1980.
11. Hunter, C.A. and K.Y. Yu, Characterization of Solid Wastes From Indirect
Liquefaction Facilities. Presented at the Symposium of Environmental
Aspects of Fuel Conversion Technology VI, Denver, Colorado, October 26-
30, 1981.
12. Fuchs, M.R., et al. A Comparison of RCRA Leachates of Solid Wastes from
Coal-Fired Utilities and Low- and Medium-Btu Gasification Processes.
Presented at the Symposium of Environmental Aspects of Fuel Conversion
Technology VI, Denver, Colorado, October 26-30, 1981.
13. U.S. Environmental Protection Agency. Sanitary Landfill Design and
Operation, SW-65ts, 1972.
14. Rental Rate Blue Book for Construction Equipment.
8. Personal Communications
Schilli, J., Black & Veatch, Consulting Engineers to TRW Environmental
Division. Personal communication based on in-house engineering studies
on power plant ash disposal site designs.
Cl-17
-------
-------
APPENDIX C2
SURFACE IMPOUNDMENT
1. Process Description
Surface impoundments, known also as evaporation ponds, holding basins,
lagoons, or settling basins, have been utilized widely by municipalities and
industries to process or dispose of waste liquids, sludges, and slurries.
Surface impoundments are natural depressions or artificial excavations, usu-
ally constructed with earthen dikes around the impoundment area. They are
usually lined either with in-place or compacted soils or synthetic materials.
Wastes are transported hydraulically to the impoundment in a fluid state. The
wastes deposit at the bottom of the impoundment and the supernatant is
treated, if needed, for discharge or recycle.
In some applications such as the disposal of power plant boiler ashes
where the properties of the waste are suitable, the settled wastes are dredged
and used as dike construction material to increase the height (and thus the
impoundment capacity) and strength of the dike. In applications where the
impoundment is used for storage or settling purposes, settled wastes are
dredged periodically and are transported to final disposal sites.
2. Process Applicability
Surface impoundment is a flexible disposal alternative with wide appli-
cations. Ash slurries, FGD sludges, and water and wastewater treatment
sludges generated by a synfuels facility could be managed with surface
impoundments. The impoundments can be used for dewatering purposes or as
ultimate disposal for these wastes. However, surface impoundments should not
be used for corrosive, ignitable, volatile, or incompatible waste (1).
C2-1
-------
Appendix C2
Surface Impoundment
3. Process Design
Surface impoundment design considerations are similar to those of land-
fill design; the site-specific factors listed in Appendix Cl for landfills
should be taken into account. A properly designed and operated surface
impoundment would minimize environmental contamination. The major potential
source of environmental degradation is the migration of leachate from the
site.
Figure C2-1 shows the water flow through an uncontrolled impoundment.
Water flows into the impoundment as surface inflow, precipitation, and slurry
water used in transporting the wastes and leaves the impoundment as surface
discharge, evaporation, and leachate. A water balance around the impoundment
gives:
L=SI+SW+P-E-D (1)
where: L = leachate,
SI = surface inflow,
SW = slurry water,
P = precipitation,
E = evaporation, and
D = discharge.
Leachate control measures consist of 1) diversion structures to prevent
runoff from entering the site, 2) liners to retard leachate movements, 3)
leachate collection and treatment system, and 4) final cover for the finished
site to limit infiltration of precipitation.
C2-2
-------
n
NJ
I
U)
SLURRY WATER WITH WASTE (SW)
\
SURFACE INFLOW (Si)
EVAPORATION (E)
PRECIPITATION (P)
1
LEACHATE (L)
i
DISCHARGE (D)
GROUNDWATER TABLE
Figure C2-1. Water balance for a surface impoundment
-------
Appendix C2
Surface Impoundment
Diversion structures (dikes) should be incorporated into the initial
design of the impoundment and constructed before the site begins accepting
waste. RCRA hazardous waste regulations require that surface runoff diver-
sions capable of diverting a 10-year, 24-hour storm be constructed. Also,
surface impoundments are required to maintain a freeboard capable of contain-
ing rainfall from a 24-hour, 25-year storm but shall be no less than 60 cm.
Liners for surface impoundments can be made of clay, concrete, asphalt,
or synthetic polymers. The Resource Conservation Recovery Act requires that
impoundments for hazardous wastes be designed such that no infiltration of
waste materials or leachates into the subsurface formations below the disposal
pond occurs. Liners for nonhazardous waste disposal facilities must be
designed to ensure that unacceptable levels of groundwater contamination do
not occur. Depending on the site specific factors and waste characteristics,
the liner design may range from no liner for sites with deep groundwater
tables and thick local clay layers, to multiple liners consisting of artifi-
cial and/or clay barriers of various thicknesses. In choosing a liner, the
compatibility of the liner with the waste must be evaluated. Some liners have
been in place for many years; however, data are generally not extensive enough
to predict liner life expectancy, given its dependence on factors such as
waste contained and installation procedures. The choice between artificial
liners and clay frequently is based largely on economics. A major cost item
for clay liners is transportation.
Leachate collection can be accomplished by a system of perforated pipes
and sumps in a layer of permeable sand laid at the bottom of the impoundment.
As discussed in Appendix Cl, the drain system should be graded to accelerate
C2-4
-------
Appendix C2
Surface Impoundment
the flow of water toward the collection system. It is desirable to design
the thickness of the drain system to exceed the maximum depth of water mound
that may accumulate.
4. Secondary Waste Generation
Leachate may be generated as a secondary waste stream from surface
impoundments. The quantity of leachate is highly dependent upon local site-
specific factors such as net evapotranspiration rate. The characteristics of
leachate will vary depending on the waste properties (e.g., solubility and
degradability) and local attenuation effects (e.g., soil cation exchange capa-
city and pH).
5. Process Reliability
Surface impoundment has been widely used for disposing of various indus-
trial and municipal wastes. When properly designed and operated, potential
for degrading groundwater and surface water should be minimized.
The major design feature for minimizing migration of leachate is the
construction of liners. The long term integrity of either natural or arti-
ficial liners has not been established.
6. Process Economics
Surface impoundment costs depend on a number of site—specific factors
such as land cost, availability of local liner material, depth of surface im-
poundments, etc. Reported power plant FGD sludge surface impoundment disposal
C2-5
-------
Appendix C2
Surface Impoundment
costs ranged from j5/dry Mg to $25.70/dry Mg, adjusted to 1980 dollars
(2,3,4). Typical unit costs for site construction are similar to those for
landfill construction (see Appendix Cl).
7. References
1. Federal Register, Col. 43, No. 243, 58946-59027, December 18, 1978.
2. U.S. Environmental Protection Agency. Disposal of Byproducts from Non—
Regenerable Flue Gas Desulfurization Systems; Initial Report. EPA 650/2-
74-126a, May 1974.
3. U.S. Environmental Protection Agency. Cost Comparisons of FGD Systems.
EPA Symposium on Flue Gas Desulfurization, Nov. 1974. EPA 650/2-74-126a.
4. Michael Baker, Jr., Inc., FGD Sludge Disposal Manual, Electric Power
Research Institute, FP-977, January 1979, Palo Alto, CA.
C2-6
-------
APPENDIX C3
LAND TREATMENT
1. Process Description
Land treatment (soil application, landspreading, and landfarming) refers
to the use of land or soil as a medium to treat and dispose of waste. The
major objective of land treatment is the degradation of organic wastes through
microbial action. Certain inorganic components of the waste (such as nitro-
gen, phosphorus, and potassium) thus released would increase the soil nutrient
content. This practice has been utilized successfully for the treatment/dis-
posal of municipal wastewater treatment sludges and refinery oily wastes.
The waste is generally applied to the land in liquid or semi-solid form
by spraying, flooding, or spreading. With spraying, portable or fixed irri-
gation systems can be used to apply liquid wastes. The system must be de-
signed to handle waste containing solids without clogging, or the waste must
be screened or pretreated. In the flooding method, a plot of land is flooded
with the waste and allowed to remain idle until most of the water is evap-
orated or infiltrated. Grates or perforated pipe are frequently utilized to
assure uniformity of application. For spreading, tank trucks are used to haul
waste to the site and to spread the waste on the land surface or inject it be-
low the surface. Subsurface application offers the advantages of immediate
mixing of waste and soil, elimination of potential odor problems, and control
of surface runoff.
The major advantages of land treatment are: 1) minimal disturbance of
land; 2) nutrients present in the waste tend to improve soil texture, water
retention, and overall ability to support vegetation; 3) the operation is
relatively simple and flexible; and 4) wide variations in waste character-
istics and loadings can be accommodated. The major disadvantages are 1)
wastes containing high concentrations of toxic compounds may not be success-
fully treated; 2) the method requires large land areas; 3) aerobic conditions
C3-1
-------
Appendix C3
Land Treatment
are usually maintained only within the top 15-20 cm of the soil, bioactivities
are usually high at aerobic conditions, also there may be odor problems under
anaerobic conditions; and 4) inadequate design of sites may result in release
of leachates and subsequent contamination of surface or groundwater.
2. Process Applicability
Waste added to the soil is subject to one or more of the following pro-
cesses 1) decomposition/degradation. 2) leaching, 3) volatilization, and
4) incorporation into the soil matrix (e.g., through ion-exchange adsorption
and filtration). Proposed RCRA regulations specify that the only legitimate
purpose of land treatment of hazardous wastes is for treatment of the waste to
reduce its hazardous properties. Thus, in land treatment it is essential to
maximize the degradation processes and minimize or eliminate the others.
Wastes that are suitable for land treatment are those that are biodegradable
such as sludges from a biological treatment plant.
Wastes with high concentrations of hazardous substances such as cadmium,
arsenic, lead, and mercury should not be land treated in sites where food
chain crops are grown. Proposed RCRA regulations specify acceptable limits on
application rate and cumulative loadings for cadmium and require data showing
no transfer or accumulation of arsenic, lead, and mercury in crops from
treated land.
3. Process Design
The site-specific factors to be evaluated for land-based disposal alter-
natives listed in the landfill Appendix, Cl, should also be considered in land
treatment. Topography, drainage pattern, soil type, groundwater pattern, and
quality are some of the major factors affecting the suitability of the site.
C3-2
-------
Appendix C3
Land Treatment
Preferably the site should be flat with slopes less than 0.5 percent to mini-
mize erosion and not be located in flood plains or groundwater recharge/
discharge areas. The water table should be as deep as possible and local
soils should have moderate to high cation exchange capacity (greater than 10
milliequivalents per 100 gram of soil). Liners (clay or artificial) may be
needed in sites where the groundwater contamination potential is high (e.g.,
sites with shallow local water table and permeable soil). Proposed RCRA regu-
lations require periodic sampling and analysis of soil-covers and porewater to
monitor leachate movement through unsaturated zones (1).
Local climate may affect the application rate and actual operation of
land treatment. Waste storage facilities are needed when precipitation (rain-
fall and snow) and cold climate interfere with the land treatment operation.
For example, facultative lagoons with 18 to 60 months storage capacities have
been designed for storing anaerobically digested sludge (2). Precipitation
also increases the potential of surface water contamination. Diversion chan-
nels should be built around the site to prevent surface runoff from entering
the site, and runoff from the site should be collected and treated before dis-
charge.
In addition to the site-specific hydrogeologic factors, other design/
operation factors include 1) the mode, rate, and depth of application; 2) fre-
quency of tilling; and 3) amounts of chemicals (e.g., nutrients, neutraliza-
tion agents) required since all these affect the rate of waste degradation.
Stabilized biological treatment sludges are usually transported to the
land treatment sites by tank trucks and spread with bulldozers, loaders,
graders, or box spreaders. The site is generally subdivided into several
plots which are treated in sequence. After waste application and, if neces-
sary, evaporation of any associated water, the plot is plowed and replowed
periodically until the waste has been decomposed. While the biological
C3-3
-------
Appendix C3
Land Treatment
degradation is taking place in one area, treatment may begin in another. Care
must be taken to maintain a proper ratio of waste to soil in order to maintain
an effective biodegradation rate. Optimum waste application will depend on
the meteorological conditions, waste constituents, the types and chemical
makeup of the soil, the types and quantities of microorganisms present, and
types of crops grown. This rate is best established by field tests and should
not exceed the biodegradation and plant uptake rates. Typical application
rates for refinery oily wastes and domestic wastewater treatment sludges are 5
to 10 weight percent of oil in the plow zone (top 15 to 20 cm of soil) (3) and
40 dry metric ton/10,000 m* (4), respectively. Chemicals such as nutrients
(nitrogen, phosphorous, potassium, etc.) may be added periodically and neutra-
lization agents may also be needed to maintain the proper pH level.
Land treatment relies on the ability of certain naturally occurring soil
microorganisms to decompose and utilize organic compounds under aerobic condi-
tions. Proper conditions for microbial action must be maintained, and wastes
which are toxic to the microorganisms must be avoided or pretreated.
4. Secondary Waste Generation
Land treating volatile waste may create air pollution which can be
reduced by subsurface injection or immediate tilling after application.
5. Process Reliability
Land treatment has been used for the disposal of waste sludges for many
years. With proper site selection, complete waste characterization, adequate
pretreatment to remove toxic constituents, and proper operation (e.g., adjust-
ment of application rates, periodic plowing), land treatment is a simple,
effective method for disposing of biodegradable waste.
C3-4
-------
Appendix C3
Land Treatment
6. Process Economics
Land treatment is a relatively simple disposal method requiring no com-
plex machinery or much labor. Costs depend on a number of site—specific fac-
tors such as land cost, site size, distance of disposal site from the waste
source, and storage requirements. Typical earth-moving costs listed in
Appendix Cl on landfills would also apply to land treatment. As an example of
land treatment costs, the unit cost (1978 dollars) for disposing of a 20
wt % oily waste were reported to be $18/Mg or Jl7.40/m3 (5).
7. References
1. Federal Register, Vol. 45, No. 98, May 19, 1980.
2. U.S. Environmental Protection Agency. Process Design Manual for Sludge
Treatment and Disposal. EPA-625/1-79-011, September 1979.
3. Huddleston, R.L. Solid-Waste Disposal: Landfarming. Chemical Engineer-
ing, February 26, 1979.
4. Lne-Hing, C., et al. Industrial Waste Pretreatment and EPA Cadmium Limit-
ations. J. Water Pollution Control Federation, Vol. 52, 2538, 1980.
5. Meyer, J.D. and R.L. Huddleston. Treatment of Oily Refinery Waste by
Landfarming. In Proceedings of the 34th Industrial Waste Conference, May
8-10, 1979, Purdue University, Ann Arbor Science 1980, p. 686-695.
C3-5
-------
-------
APPENDIX C4
SOLID WASTE INCINERATION
1. Process Description
Incineration or controlled combustion reduces the weight and volume of
waste by converting many solid and liquid organics into gaseous forms. The
extent of volume and weight reduction is dependent upon the waste characteris-
tics, the incineration process, and equipment used. Incineration is also a
viable detoxification process if the toxicity results from the structure of an
organic material as opposed to the properties of the elements it contains (1).
The end products of incineration include carbon dioxide, water, ash, and other
inorganic compounds.
The major advantages of incineration include: 1) complete destruction of
most toxic organics (at 1250 K with a residence time of two seconds), 2) reduc-
tion of toxic sludges and solids to inert, sterile gases and residuals, and 3)
reduction of waste volume requiring ultimate disposal. The major disadvan-
tages of incineration are: 1) generally only organic materials are suitable
for incineration, 2) incineration may generate secondary pollutants such as
particulates, sulfur oxides, nitrogen oxides, and metals such as mercury, as
well as hydrocarbons and carbon monoxide, 3) extensive scrubbing is required
for incineration wastes containing halogenated compounds, and 4) high capital
and operating cost.
The common types of incinerators used for solid waste disposal are:
1) rotary kiln, 2) multiple hearth, and 3) fluid bed reactor. Figure C4-1
is a simplified block diagram of a rotary kiln incinerator which is highly
efficient when applied to solids, liquids, and tars. A rotary kiln incinera-
tor is characterized by a cylindrical rotating furnace lined with firebrick or
other refractory in which wastes are heated by combustion of the waste or an
auxiliary fuel. The kiln is mounted at a slight angle to the ground. The
waste feed is introduced at the upper end and moves through the kiln by the
tumbling action caused by the kiln rotation. Fuel and air inlets can be
C4-1
-------
Water
o
-CN
I
Secondary air
Secondary
combustion
chamber
Plenum
^Induced
drjft f,m
Sieve tower
(demistcr)
Figure C4-1. Typical rotary kiln incinerator facility
-------
Appendix C4
Incineration
located at either end of the kiln, resulting in either countercurrent or
cocnrrent gas/waste flow. Ash is discharged from the lover end of the kiln
and quenched with water. High temperature gas seals betwean fixed and rotat-
ing parts of the discharge end of the kiln are difficult to maintain. There-
fore, rotary kilns operate at subatmospheric pressure to avoid the release of
combustion gases.
Complete combustion of the wastes in the rotary kiln is difficult to
achieve. The tumbling action in the rotary kiln results ia fine particle
entrainment in the gas stream. Therefore, an afterburner or secondary com-
bustor is almost always needed to provide increased gas mixing and addi-
tional residence time for the combustion reactions to be completed.
Multiple hearth units are built in a wide range of sizes to handle from
225 to 1150 kg/hr of dry solids. They consist of a cylindrical steel shell
with a series of refractory grates around a central shaft. Air-cooled rabble
arms are attached to the shaft and scrape the waste down to the next lower
grate. This serves to expose new surface to the hot gases and move the waste
through the drying, burning, and preheating zones of the furnace. Due to the
large amount of excess air required for cooling, these units always need
auxiliary firing.
Fluid bed incinerators use an air blower; an upright, refractory—lined
cylinder with a grid plate at the bottom; a burner in the side; and a fly ash
scrubber on the effluent gases. The bed is composed of graded silica sand
which is heated by the preheat burner and fluidized by the combustion air.
Influent air is preheated by the exit gases. Waste is fed into the fluidized
mass which is maintained at approximately 1030 to 1150 K. Twenty percent
excess air is required for combustion.
C4-3
-------
Appendix C4
Incineration
Table C4-1 summarizes the general operating characteristics of the major
commercial incineration processes. These processes operate at temperatures
ranging from 420 K to 1900 K with residence times ranging from less than one
second to more than an hour. These processes have been in operation for many
decades, incinerating various types of wastes ranging from municipal sludge to
highly toxic organic chemicals (3,4).
TABLE C4-1. OPERATING PARAMETERS FOR DIFFERENT TYPES OF INCINERATORS (2)
Incinerator Type
Temperature Range
Residence Time
Multiple hearth
Fluidized bed
Liquid incinerator
Catalytic combustor
Rotary kiln
Wet air oxidation
Molten salt
Multiple chamber
Pyrolysis
590-810 K
(drying zone)
1030-1250 K
(incineration)
1030-1250 K
920-1870 K
590-810 K
1090-1870 K
420-590 K
(1-29 MPa)
1090-1250 K
1090-1250 K
750-1090 K
0.25-1.5 hr
Liquids and
gases-seconds
Solids-longer
0.1-2 sec
1 sec
(1090 K maximum)
Liquids and
gases-seconds
Solids-hours
10-30 min
1-75 sec
Gases—seconds
Solids-minutes
12-15 min
C4-4
-------
Appendix C4
Incineration
2. Process Applicability
Table C4-2 summarizes the applicability of the incinerator process to
waste types. As shown in the table, a rotary kiln incinerator equipped with
auxiliary liquid injection nozzles is the most versatile and can be used to
incinerate almost any type of waste. Incineration has been applied to various
industrial wastes including refinery wastes, sewage sludges, paper mill waste
liquor, and pharmaceutical and organic chemical wastes (2,6). This method
would be applicable to treating/disposing of biological treatment sludges from
wastewater treatment and other organic byproducts such as oil, tar, and
phenols from synthetic fuel facilities.
Before a waste is incinerated, full characterization of the waste is
needed to provide an & priori estimate of incinerator performance, to
determine whether auxiliary fuel is needed, and to select compatible construc-
tion materials and adequate air pollution control systems. Typical char-
acterization data required include 1) heating value; 2) ash, carbon, oxygen,
hydrogen, and chlorine contents; 3) solubility; and 4) flash point.
3. Process Performance
Temperature, residence time, oxygen concentration, and the degree of
air/waste mixing achieved are the primary variables affecting combustion effi-
ciency in any incinerator design. There are two major steps involved in
evaluating these variables. The first step is to define the temperature,
residence time, and the excess air level, along with the degree of mixing
achieved in the incinerator, required for waste destruction. The second step
is to determine whether or not the required operating conditions are
achievable, since temperature, excess air, residence time, and mixing are all
interrelated.
C4-5
-------
TABLE C4-2. APPLICABILITY OF INCINERATOR PROCESSES TO WASTE TYPES (5)
lute type
Botary kiln
Multiple
he.rth1
Flnidlzed
bed*
Liquid
incinerator
Catalytic
conbnstor
Hnltiple-Chamber
incinerator
let air
oxidation
Molten-tilt
incinerator
n
j>
i
Solids
Granular hoaiogenona
Irregular bulky
(Pallets, etc.)
Low Belting point
(Tars, etc.)
Organic compounds
with fnaible aah
conatitnenta
Gases
Organic vapor
laden
Liquids
Bigh organic
strength aqneona
wastes, often
toxic
Organic liquids
Solids/liquid.
Waate contains
halogenated
aronatic componndi
(1480 K
Aqueous organic
If equipped with
auxiliary liquid
injection nozzles
If equipped with
auxiliary liquid
injection nozzles
Provided waste does
not be cone sticky
npon drying
If Baterld
can be »elted
and pnaiped
If liquid
Suitable for pyrolytlt operation
-------
Appendix C4
Incineration
The adequacy of incinerator operating conditions can only be determined
by past experience with the waste or by actual testing. If perfect mixing
conditions could be achieved and waste burnout occurred instantaneously, then
only the stoichiometric requirement of air would be needed. Neither of these
phenomena occur in real-world applications, thus, some excess air is always
needed. Since excess air acts as a diluent in the combustion process it
reduces the temperature in the incinerator. The minimum excess air require-
ment for a given incinerator depends upon the degree of mixing achieved and
waste specific factors. Approximate calculations to determine the attain-
ability of the proposed temperature and excess air rate can be made based on a
heat balance around the combustion chamber.
In an incinerator combustion time is governed by adjusting the travel
rate of waste through the combustion chambers. The chambers themselves must
be large enough to retain liberated gases for a sufficient time in order to
allow complete combustion. In practice, if an incinerator is overloaded, com-
bustion time is shortened resulting in incomplete combustion.
Turbulence is desirable and necessary and is achieved in an incinerator
by physical tumbling of waste, primarily on grate surfaces, and by blowing air
through the burning waste and the primary combustion gases created during the
initial phases of the combustion process. The tumbling of refuse on grates
accelerates the combustion of the solid materials by breaking up waste agglom-
erates, allowing easier escape of moisture and other volatiles, and easier
access of air to the heated refuse. Furthermore, turbulence prevents the pri-
mary gases created during the first phase of combustion from separating or
stratifying on the basis of differences in composition and density and pass-
ing through the incinerator unburned or incompletely burned. The passage of
the primary combustion products over baffles and through constrictions and the
addition of air through ports in secondary combustion chambers creates turbu-
lence and helps insure complete combustion of the primary gases.
C4-7
-------
Appendix C4
Incineration
As a general rule, most hazardous organic materials can be completely
destroyed at 1250 K at a residence time of two seconds. The destruction effi-
ciencies for incinerating selected organic wastes is presented in Table C4-3.
4. Secondary Waste Generation
Two secondary waste streams may be generated from the incineration pro-
cess. These are 1) flue gas and 2) residuals (ash).
The flue gas consists primarily of C02, water, and air. Incineration of
wastes containing halogens may generate gases such as chlorine and fluorine
which are relatively insoluble in water. Thus incineration warrants addition
of sufficient hydrogen (e.g., CH4) to convert those gases into HC1 or HF which
can be more readily removed by wet scrubbers. Sulfur compounds are often
found in wastes either as part of a sulfonated organic molecule or in the form
of sulfates or sulfides. Combustion of these wastes results in S02 formation
which can be removed by caustic scrubbing, lime/lime stone scrubbing, or other
S02 removal processes. Wastes containing inorganic salts and/or phosphorus
will produce oxides of the metal ions and/or phosphorus pentoxide upon combus-
tion. The oxides formed will be in a fine form and may require removal by a
high-energy venturi scrubber. The silicon content of the waste will become a
residue after combustion which will require disposal or become entrained in
the combustion gases (fly ash) and may require further control. General
guidelines for the selection of appropriate air pollution control devices or
systems of devices for hazardous waste incineration are presented in Table
C4-4.
The residual (ash) stream consists of inorganic, noncombustible mater-
ials. This is usually disposed of in landfills.
C4-8
-------
TABLE C4-3. OPERATING CONDITIONS AND DESTRUCTION EFFICIENCIES FOR VARIOUS INCINERATORS (6)
Waste
Ethylene Waste
Hexachlorocyclo—
pentadiene Waste
Phenol Waste
Methyl methacryltte
Waste
Cole Plant Waste
Amiben Waste
0 PVC Waste
1
Incinerator
Liquid Injection
Liquid Injection
Flnidized Bed
Flnidized Bed
Wet-Air Oxidation
Wet-Air Oxidation
Rotary Kiln
Temperature
1620-2020 K
1620-1650 K
1010-1030 K
1050-1060 K
550 K at 10.8 MPa
550 K at 10.8 HPa
1140 K in primary
combustion zone.
Residence Time
0.14 - 0.19 sec
0.17 - 0.18 sec
12 - 14 sec
12 sec
1.15 hr
1 hr
2-3 sec
Destruction
Efficiency (*)
> 99.999
> 99.999
> 99.999
> 99.999
- 90
~ 90
1 99.996
HamBermilled PCB
Waste
Nitrochlorobenzene
Waste
Chlorinated Hydrocarbon
and PCB Wastes
Rotary Kiln
Liquid Injection
Cement Kiln
1250-1360 K in
secondary com-
bustion zone
1530 K in kiln,
1600 K in after-
burner
1580-1600 K
3.2 sec
2.3 sec
5-10 sec
> 99.999
> 99.999
1 99.999
-------
TABLE C4-4. CONFIGURATIONS OF SELECTED GAS CLEANING DEVICES
APPLICABLE TO HAZARDOUS WASTE INCINERATION FACILITIES (7)
n
i
M
O
Gas Strom
No significant
gaseous
pollntant s
present"
Si|nif leant
gaseous
pollutants
present
Characteristics
High or low
particulate
loading
High or lor
particulite
loading
Low
particulite
loading
^Particulste control nay require
cThe tern significant gaseous pel
Control Devices*
Cyclone
Electrostatic precipitator
Fabric filter
Gas-atomized sprays0
Preformed spray
precipitator
Cyclone
Electrostatic precipitator
Gas-atomized sprays0
Preformed spray
Sorbent filled baghonse
Plate type scrubbers0
Packed bed scrubbers0
Mist
Emininator
Not usually needed
Usually needed
Not usually needed
Not usually needed
Optional
Not usually needed
Dsnally needed
Not usually needed
combinations of the Indicated devices depending on loadii
llutants is meant to include HC1 , Cl,. HF, HBr, Br,, P>0|
ration facilities generally employ these device confiture
Absorption Device
Not usually needed
Not usually needed
Not usually needed
Gas-atomized spray
Preformed spray
Plate type scrubber
Packed bed scrubber
Sorbent filled baghouse
Plate type scrubber (optional)
Packed bed scrubber (optional)
Sorbent filled baghouse (optional)
Ionizing wet scrubber (options!)
let electroststic precipitator (as
a polishing device)
Not usually needed
Not usually needed
Not usually needed
ag. size distribution, and economics.
tioni. x
Hist
Eliminator
Not usually needed
Not usually needed
Not usually needed
Usually needed
Not usually needed
Usually needed
Usually needed
Not usually needed
Not usually needed
Not usually needed
ig control.
-------
Appendix C4
Incineration
5. Process Reliability
Incineration is a commercially available process which has been proven to
be reliable. There are numerous units in worldwide operation for disposal/
treatment of municipal and industrial solid wastes and sludges. On—stream
time is believed to be high although no specific data are publicly available.
6. Process Economics
There is no typical incinerator design or facility, and the cost of
incinerating waste varies depending on the waste characteristics, the type of
incinerator, and the operating conditions. The major factor affecting equip-
ment costs is the heat input rate to the incinerator.
Figures C4-2 and C4-3 show, respectively, the estimated total capital
investment for a "grass root" installed rotary kiln and a liquid injection
incinerator as a function of heat input. The costs presented are mid-1980
dollars and are believed to be accurate to +50 % (8). The capital investment
costs include site preparation, the incinerator itself, all of the necessary
pumps and accessories, and electrical instrumentation and controls needed to
operate the unit (9). As shown, the rotary kiln incinerator is more capital-
intensive, but as discussed before, is more versatile than the liquid injec-
tion incinerator. The estimated annualized costs for these two incinerators
are presented in Figures C4-4 and C4-5, respectively. The cost curves
developed include fuel and labor costs, maintenance, debt service, and over-
head (9).
C4-11
-------
o,
-------
100
1.0
Battery Limits Energy Recovery
Battery Limits No Energy Recovery
0.1
i I i i i t
i I i I I i
1.0
10
100
1000
Incinerator Capacity, GJ/hr
Figure C4-3.
Total capital investment for a liquid injection
incinerator (8)
C4-13
-------
en
>-i
co
O
Q
O
I
(-1
O
o
CO
0)
I
en
0)
•H
t-l
•H
0.01
1.0 10
Incinerator Capacity, 107 Btu/hr
1 Btu = 1,055 GJ
100
Figure C4-4. Total annual operating costs for a liquid injection
incinerator (8)
C4-14
-------
cn
o
U
00
c
01
o.
o
3
C
Incinerator Capacity, 107 Btu/hr
1 Btu - 1 ,OS3 H.T
Figure C4-5. Total annual operating costs for a liquid injection
incinerator (8)
C4-15
-------
Appendix C4
Incineration
7. References
1. Scurlock, A.C., et al. Incineration in Hazardous Waste Management.
EPA/530/SW-141, 1975.
2. Hitchcock, D.A. Solid Waste Disposal: Incineration. Chemical
Engineering, pp. 185-194, May 21, 1979.
3. U.S. Environmental Protection Agency. Process Design Manual, Sludge
Treatment and Disposal. EPA-625/1-79-011, September 1979.
4. Becker, K.P. and C.J. Wall. Waste Treatment Advances: Fluid Bed
Incineration of Wastes. Chemical Engineering Progress, pp. 61-68,
October 1976.
5. U.S. Environmental Protection Agency. Recommended Methods of Reduction,
Neutralization, Recovery or Disposal of Hazardous Waste, Vol. IV. EPA
Contract No. 68-03-0089, February 1973.
6. TRW, Inc. Destroying Chemical Wastes in Commercial Scale Incinerators.
EPA Contract Report, Contract No. 68-01-2966, November 1977.
7. TRW, Inc. Air Pollution Control Device for Hazardous Waste Incinerators.
Permit Writer Guidelines. Report submitted by TRW, Inc., to EPA for
publication.
8. U.S. Environmental Protection Agency. Draft Engineering Handbook for
Hazardous Waste Incineration. EPA Contract No. 68-03-2550, November
1980.
9. Preliminary Incinerator Unit Cost Study for Part 264. EPA-OSW Hazardous
Industrial Waste Division, February 1981.
C4-16
-------
APPENDIX C5
CHEMICAL FIXATION AND ENCAPSULATION
1. Process Description
The terms fixation and encapsulation in solid waste management refer to
the treatment processes which stabilize or solidify the waste constituents or
enclose the waste within added agents. The fixed or encapsulated wastes are
generally disposed of in landfills. Encapsulation processes generally enclose
the waste in a coating or jacket of an inert, relatively impermeable material.
Fixation processes generally combine the concepts of solidification (the
alteration of the characteristics of a waste to attain the desired structural
characteristics) and stabilization (the immobilization of waste constituents
by chemical reactions to form insoluble compounds or by entrapping the waste
constituents in an inert polymer or stable crystal lattice). This technique
has been practiced in Europe and Japan for many years (1) and is gaining more
interest in the U.S. for handling hazardous wastes.
There are many fixation and encapsulation processes that are commercially
available. Based on the principal chemical agents used, these can be grouped
as: cement-based, lime-based, thermoplastic, organic polymer, glassification,
and encapsulation techniques (2).
Cement-Based
Most wastes slurried in water can be mixed directly with cement, and the
suspended solids will be incorporated into the rigid matrix of the hardened
concrete. This procedure is especially effective for wastes with high levels
of toxic metals, since at the pH of the cement mixture most mnltivalent
cations are converted to insoluble hydroxides or carbonates. Metal ions may
also be taken into the crystal structure of the cement minerals that form.
Materials in the waste such as sulfides, asbestos, latex, and solid plastic
may actually increase the strength and stability of the waste concrete.
C5-1
-------
Appendix C5
Fixation/Encapsulation
Impurities such as organic materials, silt, clay, coal, or lignite may delay
setting and curing. All insoluble materials passing through a No. 200 sieve
(74 micron particle size) are undesirable since they may be present as dust or
may coat the larger particles, weakening the bond between the particles and
the cement. Salts of manganese, tin, zinc, copper, and lead may cause large
variations in setting time and significant reductions in physical strength.
Other compounds which are especially active as cement setting retarders
include sodium salts of arsenate, borate, phosphate, iodate, and sulfide.
Products containing large amounts of sulfate (such as flue gas cleaning
sludges) not only retard the setting of concrete, but cause swelling and spal-
ling. A low alumina (Type V) cement is available to prevent such a reaction.
A number of additives, most of them proprietary, have been developed for use
with cement to improve the physical characteristics and decrease the leaching
losses.
Lime-Based
Fixation techniques based on lime products usually depend on the reaction
of lime with fine-grained siliceous (pozzolanic) material and water to produce
a concrete-like material. The most common pozzolanic materials used in waste
treatment are fly ash, ground blast-furnace slag, or cement-kiln dust. For
example, flue gas cleaning sludge, fly ash, lime, and other additives can be
combined to form an easily handled solid.
Thermoplastic
Thermoplastic fixation methods involve drying, heating, and mixing the
waste with heated thermoplastic or emulsified bitumen materials. The mixture
is cooled and usually buried in a container such as a steel drum. The ratio
of material to waste is generally 1:1 to 1:2. The type of waste determines
the type of material that can be used. Organic chemicals are solvents for
C5-2
-------
Appendix C5
Fixation/Encapsulation
some materials, so they cannot be used. Strongly oxidizing salts, such as
nitrates, chlorates, or perchlorates, will react with organic materials and
cause slow deterioration, hence preventing their use. At the temperatures
necessary for processing the material and oxidizing wastes, the resulting mix-
ture is extremely flammable. Testing undertaken on anhydrous salts embedded
in bitumen indicates that rehydration of the embedded salts can occur when
they are soaked in water, causing the solidified waste bitumen mixture to
split. The splitting results in an increase in waste dissolution. Some salts
(sodium sulfate, for instance) naturally dehydrate at the temperatures
required for processing; hence, these must be avoided when utilizing this
fixation method.
Organic Polymer
In the organic polymer method, a polymer is generally formed in a batch
process where the wet or dry wastes are blended with a prepolymer. When the
waste and prepolymer are thoroughly mixed, a catalyst is added. Mixing con-
tinues until the catalyst is thoroughly dispersed. The polymerized material
does not chemically combine with the waste; rather, it forms a spongy mass
that traps the waste particles. Liquid associated with the waste will remain
after polymerization. The polymer mass often is dried before disposal.
Self-Cementing
Some industrial wastes, such as flue gas cleaning or desulfurization
sludge, contain large amounts of calcium sulfate or calcium sulfite. A tech-
nology has been developed to treat these wastes so they become self-cementing.
Usually a small portion of the dewatered waste sulfite/sulfate sludge is cal-
cined to produce a partially dehydrated cementitious calcium sulfate or
C5-3
-------
Appendix CS
Fixation/Encapsulation
sulfite. The calcined material is mixed with the waste along with proprietary
additives. The product is a hard, plaster-like material with good handling
characteristics and low permeability.
Classification
Extremely hazardous wastes are sometimes combined with silica and fused
into glass. Glass is leached extremely slowly by naturally occurring water.
This approach is generally assumed to produce a safe material for disposal,
without secondary containment.
Encapsulation
For this processing alternative a discrete amount of waste is enclosed
in a coating or jacket of an inert, relatively impermeable material. The
function of the material is to prevent contact between the waste and water
and, hence, dissolution and migration of potentially harmful constituents in
the waste. A significant problem is attaining and retaining adhesion between
the coating material and the waste and also maintaining the long-term inte-
grity of the coating material.
Waste encapsulates are characterized by two elements: a stiff, weight-
supporting component and a tough, flexible, encompassing, seam-free plastic
jacket. The stiff element is geared to provide dimensional stability under
mechanical stresses and compaction in the landfill. The flexible element
insures a seal, even if the stiff element is distorted and isolates the wastes
from developing leachates.
C5-4
-------
Appendix C5
Fixation/Encapsulation
The encapsulation process includes the following steps:
• dewatering of wastes,
• coating the waste particnlates with the resin,
• evaporating the solvent carrier,
• compacting the resin-coated particles,
• consolidating by thermosetting to form a wastebinder block, and
• encapsulating the water-binder block (jacketing).
Resins used for forming the waste agglomerate and the jacket are typically
polybntadiene and polyethylene.
2. Process Applicability
Fixation and encapsulation processes have been applied to wastes from
many industries, including chemical, petrochemical, utility, and metal fin-
ishing industries (3). Specific wastes that have been treated by these pro-
cesses include FGD sludges (4), spent catalysts (1), and refinery waste (2),
among others. Table C5-1 compares the advantages and disadvantages of these
processes, and Table C5-2 presents a general matchup of waste types with these
processes (1,2). When selecting a specific process to treat wastes from syn-
thetic fuel facilities, the quantity and characteristics of the waste to be
treated, the economics involved, and the ultimate disposal requirements must
be considered and the waste must be tested for treatability potential.
3. Process Performance
The effectiveness of a process depends upon the type of process used and
on the nature of the waste being treated. The most important criteria of
C5-5
-------
TABLE C5-1. ADVANTAGES AND DISADVANTAGES OF FIXATION/ENCAPSULATION PROCESSES (J,2)
Process
Advantages
Disadvtntii.es
Cement-based 1. Additives are available at a reasonable price.
2. Cement ailing and handling techniques are
well developed.
3. Processing equipment la readily available.
4. Processing is reasonably tolerant of chemical
variations in sludges.
5. The strength and permeability of the end-product
can be varied by controlling the amount of
cement added.
Line-lined
Thermoplastic
O
i_n
Organic polymer
The additives are generally very inexpensive
and widely available.
Equipment required for processing la aimple to
to operate and widely available.
Chemistry of pozzolanic reactiona la well known.
Contaminant migration rates are generally lower
than for most other techniques.
End-product is fairly resistant to most
aqueous solutions.
Thermoplastic materiala adhere well to Incorpor-
ated materials.
Only ssiall quantities of additives are usually
required to cause the mixture to set.
Techniques can be applied to either wet or dry
sludges.
End-product has a low density as compared to
other fixation techniques.
Encapsulation 1. Very soluble contaminants are totally isolated
from the environment.
2. Usually no secondary container is required,
because the coating materials are atrong and
chemically inert.
Self-cementing 1. No additives required.
2. Product la stable, non-flammable and non-
biodegradable.
Classification 1. High degree of waste containment.
2. Additives are relatively inexpensive
(syenite and lime).
1. Low-strength cement-wsste mixtures are often
vulnerable to acidic leaching solutions.
Extreme conditions can result In decomposition
of the fixed material and accelerated leaching
of the contaminants.
2. Pretreatment, more expensive cement types, or
costly addltivea may be necessary for stabili-
zation of wastes containing Impurities that
affect the setting and curing of cement.
3. Cement and other addltivea add considerably to
weight and bulk of waste.
1. Lime and other additives add to weight and bulk
of waste.
2. Stabilized sludges are vulnerable to acidic
solutions and to curing and setting problems
1. Expensive equipment and skilled labor are gen-
erally required.
2. Sludges containing contaminants that volatilize
at low temperatures should be processed carefully.
3. Thermoplastic materiala are flammable.
4. let sludges need to be dried before they earn be
mixed with the thermoplastic material.
1. Contaminants are trapped in only a loose resin-
matrix end-product.
2. Catalysts used in the urea-formaldehyde process
are strongly acidic. Host metals are extremely
soluble at low pH and can escape in water not
trapped in the mass during the polymerisation
process.
3. Some organic polymers are biodegradable.
4. End-product la generally placed in a container
before disposal.
1. Materiala used are often expensive.
2. Techniques generally require specialized equip-
ment and heat treatment to form the jackets.
3. The aludge has to be dried before the proceaa
can be applied.
4. Certain jacket materiala are flammable.
1. Application only limited to snlfate and snlfite
sludges.
2. Additional energy ia required to produce the
calcined cementitlous material.
1. The process ia energy intenalve.
2. Some constituents such as certain metals may
be vaporized during processing.
3. Specialized equipment and trained personnel
are required.
-------
Appendix C5
Fixation/Encapsulation
TABLE C5-2. APPLICABILITY OF FIXATION/ENCAPSULATION
PROCESSES TO WASTE TYPES (1,2)
Process Treatable Waste Dntreatable Wastes
Cement-based Toxic inorganics Organics
Stack-gas-scmbber sludges Toxic anions
Lime—based Toxic inorganics Organics
Stack-gas-scrubber sludges Toxic anions
Thermoplastic Toxic inorganics Organics
Strong oxidizers
Organic polymer Toxic inorganics Acidic materials
Organics
Strong oxidizers
Encapsulation Toxic and soluble inorganics Strong oxidizers
Self-cementing High snlfate/sulfite sludge Others
Classification Toxic inorganics Organics
effectiveness are mechanical strength and resistance to chemical attack and
biodegradation. Table C5-3 presents results of a leaching study performed
with some refinery wastes stabilized by a proprietary silica cement-based fix-
ation process.
4. Secondary Waste Generation
There is no secondary waste generated.
5. Process Reliability
There is no publicly available process reliability information.
C5-7
-------
Appendix C5
Fixation/Encapsulation
TABLE C5-3. LABORATORY LEACHING RESULTS OF CHEM-FIXED REFINERY WASTES* (2)
Constituent
Total
Chromium (Cr)
Iron (Fe)
Zinc (Zn)
Nickel (Ni)
Copper (Cu)
Manganese (Mn)
Cyanide (Cn)
Cone, in the
Raw Sludge
ppm 0-62
43.5 <0.10
1310 <0.25
88.0 <0.25
8.9 <0.25
0.62 <0.25
<0.25
<0.10
Cm. of Leachate Waterb
62-125 125-188 188-250
<0.10
<0.10
<0.10
<0.10
<0.10
<0.10
<0.10
<0.10
<0.10
<0.10
<0.10
<0.10
<0.10
<0.10
<0.10
<0.10
<0.10
<0.10
<0.10
<0.10
<0.10
Concentration of the constituents in ppm in the leachate water after
application of the specified amount of distilled water.
Each 62 cm of leachate water represents approximately 80 ml of distilled
water.
6. Process Economics
The estimated cost of fixation/encapsulation varies widely and is
dependent upon the type of waste and the process utilized. Table C5-4 sum-
marizes the reported unit cost for the various processes.
C5-8
-------
Appendix C5
Fixation/Encapsulation
TABLE C5-4. UNIT COST FOR FIXATION/ENCAPSULATION PROCESSES
Processes
Cost/in* of Waste
Encapsulation
(5)
Resin fusion - unconfined particulates and sludges
Resin fusion - 0.21 m3 drum and smaller containers
Cementitious encapsulation - small containers
Plastic welding
Fixation
(1)
Cement-based
Lime-based
Organic polymer
Thermoplastic
$178
119
174
287
8-10
8-13
737
Cost available on
case—by—case
basis only
a
1979 dollars
J1978 dollars
7. References
1. Pojasek, R.B., ed. Toxic and Hazardous Waste Disposal, Vol. I. Ann Arbor
Science Publication, Inc., Ann Arbor, Michigan, 1979.
2. U.S. Environmental Protection Agency. Survey of Solidification/Stability
Technology for Hazardous Industrial Wastes. EPA-600/2-79-056, July 1979.
3. Conner, J.R. Disposal of Liquid Wastes by Chemical Fixation Waste Age.
September 1974, pp. 26-45.
4. Coltharp, W.M., et. al. State-of-the-Art of FGD Sludge Fixation.
Electric Power Research Institute Report, EPRI FP-671, January 1978.
5. U.S. Environmental Protection Agency. Laboratory and Field Evaluation of
Processes and/or Materials to Encapsulate Waste/Containers of Non-
Radioactive Hazardous Materials. EPA Contract No. 68-03-2483.
C5-9
-------
-------
APPENDIX D
COSTING METHODOLOGY
In order to provide an indication of the economic impact of pollution
controls, capital investment and operating costs were developed for most of
the pollution control processes discussed in Appendices A, B, and C. These
estimates are based primarily on cost information contained in nonproprietary
published literature. As such, they should be viewed only as general indi-
cators of expected costs and should not be construed as definitive cost esti-
mates for a specific plant. The methodologies used to develop capital invest-
ment and operating cost estimates are presented in Sections 1 and 2, respec-
tively.
There are several general factors that lead to uncertainties in the cost
estimates. Two of these factors are discussed below and are related to the
level of accuracy of the published cost data used and the general methodology
used to apply the acquired cost data to the processes addressed in this docu-
ment.
Sources of cost data used in this document are published costs for pro-
cesses applied to similar streams in related industries, costs from published
design studies for synthetic fuels plants, and vendor quotes. The accuracy of
cost data taken from published sources is influenced by the details of the
design upon which the cost was based and the cost methodology used. In addi-
tion, the accuracy of the published estimates and definition of the components
included in these estimates (e.g., contingency reserves, working capital,
land) are not always specified in the reference. Thus, extrapolation of pub-
lished costs will introduce uncertainties in the resulting estimates.
The general costing methodology used in this manual (described in this
appendix) also introduces some uncertainties. Other estimators may choose to
use different factors for components such as direct and indirect installation
D-l
-------
Appendix D
Costing Methodology
costs. In addition, available cost estimates may have been originally devel-
oped as much as 10 years ago. It is therefore possible that recent advances
in the state-of-the-art are not reflected in some of the resulting cost esti-
mates.
As a result of the above influences, the accuracy of the cost information
presented will vary. However, the cost information presented is believed to
be adequate for the use intended.
1. Capital Costs
For most of the pollution control processes discussed in this document,
capital costs are presented as installed equipment costs (IEC). TEC includes:
• Purchased and delivered equipment costs and
• Labor and materials costs for equipment installation (direct
installation costs).
The IEC may be used to develop an estimate of total capital investment (TCI)
which includes such items as:
• Installed equipment cost (IEC),
• Indirect installation costs such as engineering and construc-
tion costs, contractor fees, and contingency reserves, and
• Interest during construction.
D-2
-------
Appendix D
Costing Methodology
Items 2 and 3 are sometimes estimated based on factors applied to the IEC.
lECs are presented to enable the user to apply his own particular factors in
order to estimate total capital investment. Finally, the user can estimate a
total capital requirement by adding nondepreciable items such as working
capital and land to the total capital investment.
For cost estimates developed for this document, the major source of cost
information was the open literature although some vendor quotes were used. In
general, literature cost information is not reported as purchased and
delivered equipment costs. Some published data are installed equipment costs
(purchased equipment plus direct installation costs), some also include one or
more of the indirect installation charges listed previously, some are total
capital investment (TCI) estimates, and others are total capital requirements
(TCR). Where known, the definition of the components of the cost estimate is
given for each of the pollution control process appendices.
2. Operating Costs
Operating expenses include costs for labor (operating, supervision, and
maintenance labor), raw materials, chemicals, catalysts, utilities (steam,
electricity, cooling water, etc.), and overhead. In general, the unit costs
or factors presented in Table D-l were used to estimate operating costs for
the pollution control processes in this document. For some pollution control
processes, operating parameters such as kW/mole sulfur removed in a Claus
plant are specified so that the user can estimate operating costs based on his
particular unit costs.
D-3
-------
Appendiz D
Costing Methodology
TABLE D-l. UNIT COSTS AND FACTORS FOR OPERATING COST ESTIMATES
Operating Labor (Jll/hr)
Supervision (15% of operating labor)
Maintenance (2% of total depreciable investment)
Maintenance Supervision (5% of maintenance)
Raw Materials
Coal
Rosebud ($13.86/Mg)
Illinois No. 6 (*35.44/Mg)
Dunn County Lignite ($14.70/Mg)
Water (Jo.036/m»)
Utilities
Steam ($3.70 to $8.20/Mg depending on quality)
Electricity ($0.033/kW-hr)
Fuel gas ($1.79/GJ)
Cooling water (i0.08/ma)
Boiler feed water ($0.264/m3)
Chemicals and Catalysts
All unit costs are intended to reflect a first quarter 1980 basis.
D-4
------- |