Un.tsd States
Environment* Protection
Robert S Kerr Environmental Researcr fPA 600 2 79 1 77
Laboratory A jyusi 1979
Ada OK 74820
"!•»>•"*• and D«vtiopm«oi
Activated Carbon
Treatment of
Industrial Wastewaters
Selected Technical
Papers
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution-sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-79-177
August 1979
ACTIVATED CARBON TREATMENT
OF INDUSTRIAL WASTEWATERS
SELECTED TECHNICAL PAPERS
Edited By
Industrial Sources Section
Source Management Branch
Robert S. Kerr Environmental Research Laboratory
Ada, Oklahoma 74820
OFFICE OF ENERGY, MINERALS, AND INDUSTRY
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
-------
DISCLAIMER
This report has been reviewed by the Robert S. Kerr Environ-
mental Research Laboratory, U.S. Environmental Protection Agency,
and approved for publication. Mention of trade names or commer-
cial products does not constitute endorsement or recommendation
for use.
ii
-------
FOREWORD
The Environmental Protection Agency was established to
coordinate administration of the major Federal programs designed
to protect the quality of our environment.
•>
An important part of the Agency's effort involves the search
for information about environmental problems, management tech-
niques and new technologies through which optimum use of the
nation's land and water resources can be assured and the threat
pollution poses to the welfare of the American people can be
minimized.
EPA^'s Office of Research and Development conducts this
search through a nationwide network of research facilities.
As one of these facilities, the Robert S. Kerr Environmental
Research Laboratory is responsible for the management of programs
to: (a) investigate the nature, transport, fate and management of
pollutants in ground water; (b) develop and demonstrate methods
for treating wastewaters with soil and other natural systems;
(c) develop and demonstrate pollution control technologies for
irrigation return flows; (d) develop and demonstrate pollution
control technologies for animal production wastes; (e) develop
and demonstrate technologies to prevent, control, or abate pollu-
tion from the petroleum refining and petrochemical industries;
and (f) develop and demonstrate technologies to manage pollution
resulting from combinations of industrial wastewaters or indus-
trial/municipal wastewaters.
This report contributes to the knowledge essential if the
EPA is to meet the requirements of environmental laws that it
establish and enforce pollution control standards which are
reasonable, cost effective and provide adequate protection for
the American public.
.c
W. C. Galegai
Director
Robert S. Kerr Environmental Research Laboratory
ill
-------
ABSTEACT
Historically, activated carbon has been used to remove
taste and odor constituents in drinking water, or to decolorize
granular sugar. Industrial wastewater treatment employing granu-
lar activated carbon was basically, a technical curiosity until
the 1970's. During this decade, the articles as well as claims
for and against activated carbon treatment for industrial waste-
waters are increasing at a phenomenal rate.
EPA's Office of Research and Development, has been vitally
interested in this treatment technology, from the technical
standpoint as well as furnishing a technical data' base to the
Federal and State guidelines and permit program. Other prime
users of the technical data base are industry, consultants and
academia.
Because of the tremendous interest in the organic consti-
tuent removal by activated carbon, the two industrial categories
displaying the most interest are the petroleum refining and pet-
rochemical industries. EPA's Office of Research and Development
has co-sponsored two technical symposia for the petroleum refin-
ing/petrochemical industries, and activated carbon treatment as
an important section of both agendas. The technical papers pre-
sented research activities conducted by consultants, industries,
and EPA.
The presentations made at these symposia have been arranged
into the following sequence: (1) State-of-the-Art, (2) Organic
Compound Removal, (3) Granular Pilot-Scale Studies, (4) Powdered
Activated Carbon Pilot-Scale Studies, (5) Full-Scale Granular
Activated Carbon Treatment, (6) Full-Scale Powdered Activated
Carbon Treatment, and (7) Activated Carbon Regeneration.
Economics of Activated Carbon Treatment are presented in
the applicable individual technical papers and is not a separate
topic for this report.
iv
-------
CONTENTS
Foreword iii
Abstract iv
Figures vii
Tables xiv
1. Activated Carbon State of the Art 1
"Current State of the Art of Activated
Carbon Treatment." 1
2. Organic Compound Removal by Activated Carbon 51
"Cautions and Limitations on the Appli-
cation of Activated Carbon Adsorption
to Organic Chemical Wastewaters." .... 51
"Organic Reduction Through Add-On Acti-
vated Carbon at Pilot-Scale." 65
3. Granular Activated Carbon Pilot-Scale
Studies 81
"Activated Sludge Enhancement--A Viable
Alternative to Tertiary Carbon Adsorp-
tion." 81
"Pilot Plant Activated Carbon Treatment
of Petroleum Refinery Wastewater." .... Ill
4. Powdered Activated Carbon Pilot-Scale
Studies 135
"Combined Powdered Actuvated Carbon—
Biological Treatment: Theory and
Results." 135
"Powdered Activated Carbon Enhancement of
Activated Sludge for BATEA Refinery Waste-
water Treatment." 154
"Treatment of Oil Refinery Wastewaters
with Powdered Activated Carbon." 179
V
-------
5. Full-Scale Granular Activated Carbon
Treatment 199
"Activated Carbon Treatment of Combined
Storm and Process Waters." 199
"Water Pollution Abatement at BP Oil Cor-
poration's Marcus Hook Refinery." 212
6. Full-Scale Powdered Activated Carbon
Treatment. . 236
"Case History: Use of Powdered Activa-
ted Carbon in an Activated Sludge
System." 236
"Case History: Use of Powdered Activa-
ted Carbon with a Bio Disc Filtration
Process for Treatment of Refinery Wastes."252
"Granular Carbon Reactivation— State of
the Art." 265
vi
-------
FIGURES
Current State-of-the-Art--Activated Carbon Treatment
Figure
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
]
•
Adsorption of DBS to Equilibrium in ...
Continuously Stirred Flow System
Freudlich Isotherm Application
Cross-Sectional View- -Multiple Media Furnace . .
Simplified Flow Diagram, Tahoe, California . . .
Simplified Flow Diagram, Colorado Springs, Colo.
Schematic Flow Diagram
Frequency Analysis for Effluent BOD
Frequency Analysis for Effluent COD
Flow Diagram for the Garland Plant
Physical/ Chemical Pilot Plant Schematic, Pomona, CA
Sulfides and Headloss vs Time
BOD Removal Treatment Results
COD Removal Treatment Results
ARCO Flow Diagram, Watson Refinery . ...
Performance of ARCO Carbon Plant, COD Removal . .
Performance of ARCO Carbon Plant, O&G Removal . .
Carbon Treatment System, Reichhold Chemicals Co.
BP Treatment System Flow Diagram, Marcus Hook Ref .
COD Removal as a Function of Adsorber Contact Time,
Pag
39
39
40
40
41
41
42
42
43
43
44
44
45
45
46
46
47
47
48
Period 1
20 Adsorbe'r Oil Removal as a Function of Influent. . 48
Concentration, Period 1
vii
-------
igure Page
21 Full-Scale Process Performance, BP Marcus Hook . . 49
22 Effluent COD Attainable From Activated Carbon . . 49
Systems
23 Influent and Effluent Oil and Grease Distributions 50
for a Full-Scale Activated Gabon System
24 Carbon Adsorption Capacity for Various Plants . . 50
Cautions and Limitations on the Application of
Activated Carbon Adsorption to Organic' Chemical
Wastewater
1 Batch Adsorption Isotherm Tests 61
2 Breakthrough Curve for Plant a Bio-Treated Waste- 61
water
3 Effect of Molecular Weight on Amenability of Adsorp- 62
tion of Alcohols
4 Effect of Molecular Weight on Amenability of Adsorp- 62
tion of esters
5 Effect of Molecular Weight on Amenability of Adsorp- 63
tion of Organic Acids
6 Percent of Compound Adsorbed VS. Molecular Weight . 63
Functionality Effects
7 pH Effects on Adsorption Removals of Selected Organ- 64
ic Compbands
8 Breakthrough Curve for Activated Carbon Adsorption . 64
of a Four-Component Mixture in a Continuous System
Organic Reduction Through Add-On Activated Carbon
At Pilot-Scale
1 Carbon Columns 74
2 Pilot Treatment Facility . 75
3 Sampling Point-Plant Intake . 74
4 Sampling Point-DAF Effluent 76
viii
-------
Figure Page
5 Mixed Media Filters 77
6 Carbon Columns 77
7 Total Ion Chromatogram of DAF Effluent ... 78
(Neutral Fraction, Four-Day Composite)
8 Total Ion Chromatograms of DAF Effluent and 79
FC Effluent (Neutral Fraction, Four-Day
Composite)
9 Total Ion Chromatograms of Activatedi-Carbon 80
Activated Sludge Enhancement: A Viable
Alternative to Tertiary Carbon Adsorption
1 Treatment Schemes 99
2 Part I - TOG and COD Removal by Pretreatment. 100
3 Part I - Activated Sludge TOC and COD Distri- 101
bution
4 Part I - Granular Carbon Column TOC and COD- 102
Distributions
5 Part II - TOC and COD Removal by Pretreatment 103
6 Part II - Activated Sludge TOC and COD Distri- 104
butions
7 Part II - Granular Carbon TOC and COD Distri- 105
butions
8 Comparison of Estimated Carbon Costs .... 106
9 Estimated Effective Carbon Cost at 1 MM GPD 107
Pilot Plant Activated Carbon Treatment of
Petroleum Refinery Wastewatef
1 Carbon Sources 126
2 Adsorption Isotherm 127
3 Multi-Media Filter 127
4 Mini Column System 128
ix
-------
Figure Page
5 Refinery Waste Water Treatment System . . 128
6 Activated Carbon Pilot Plant Flow Diagram 129
7 Treatment Results, BOD 129
8 Treatment Results, COD 130
9 Treatment Results, TOG 13°
10 Benzene Extract of API Carbon 131
11 Methylene Chloride Extract of API Carbon . 131
12 Methanol Extract of API Carbon 131
13 Hexane Extract of API Carbon 131
14 Chloroform Extract of API Carbon .... 131
15 Chloroform Extract of FC Carbon 131
16 Chloroform Extract of New Activated Carbon 132
17 Standard Oil Mixture 132
18 APIC Extract Saturate Fraction 133
19 FCC Extract Saturate Fraction 133
20 APIC Extract Aromatic Fraction . . . . . 133
21 FCC Extract Aromatic Fraction 133
22 APIC Extract Polar Fraction 133
23 FCC Extract Polar Fraction 133
Combined Powdered Activated Carbon - Biolog-
ical Treatment: Theory and Results"
1 A Simple Illustration of PACT 147
2 Effect of Sludge Age and Carbon Dose on .
Effluent TOG and Apparent Loading 148
x
-------
Figur
3
4
5
6
7
1
2
3
4
5
6
7
1
2
3
4
e
Photomicrographs
Comparative Settling Characteristics of. . .
PACT and Activated Sludge
Settling Characteristics of PACT and ....
Activated Sludges
Specific Resistance "r" of Pact vs Biolog- .
ical Sludge
DuPont PACT Process: Chambers Works ....
Powdered Activated Carbon Enhancement of
Activated Sludge for BATEA Refinery Waste-
water Treatment
Simplified Refinery BPT Wastewater Treatment
System
Schematic of Activated Sludge Reactor Used .
in Pilot Program . .
Effect of Mixed-Liquor Carbon Concentration
On Effluent SOC
Soluble Organic Carbon - Phase 3
Soluble Chemical Oxygen Demand
Ammonia-Nitrogen - Phase 3
Phenolics - Phase 3
Treatment of Oil Refinery Wastewaters With
Powdered Activated Carbon
Effect of Powdered Carbon on BOD Removal
Effect of Powdered Carbon on Effluent COD .
Effect of Powdered Carbon on Effluent Total
Organic Carbon
Effect of Powdered Carbon on Effluent Total
Carbon
Page
149
dU*»*7
150
151
152
153
. 172
. 173
. 174
175
176
111
178
190
191
192
193
xi
-------
Figure Page
5 Effect of Powdered Carbon on Effluent Oil 194
6 Effect of Powdered Carbons on Effluent
Suspended Solids
f
7 Effect of Powdered Carbon on Effluent
Nitrogen 196
8 Effect of Powdered Carbon on Effluent
Phosphorous 197
9 Effect of Powdered Carbon on Effluent Zinc • 198
Activated Carbon Treatment of Combined Storm
and Process Waters
1 Waste Water Activated Carbon Treatment Plant 207
2 Relative Efficiency Profile of Cell After . .
208 Hours of Operation @ 259 GPM and 650 to
600 PPM COD in Feed 207
3 Typical Performance Data—Staggered Operation,
Second Rainy Season 20&
4 Performance Data—Bulk Processing Cells,
Second Rainy Season 209
5 Carbon Transfer and Regeneration System • • 1210
Water Pollution Abatement at BP Oil Cor-
poration's Marcus Hook Refinery
1 Carbon Adsorption Isotherm of Filtered Waste-
Water 225
la Schematic Flow Diagram of Proposed Sand Fil-
ter - Activated Carbon Treatment Facilities 226
Case History: Use of Powdered Activated Carbon
In An Activated Sludge System
1 Treatment Plant Flow Diagram .248
2 Effect of Powdered Carbon on Effluent TSS
Loading 249
xii
-------
Figure Page
3 Effect of Powdered Carbon on Effluent TSS . . 249
4 Comparison of COD Removal Efficiency .... 250
5 Effect of Powdered Carbon on Effluent COD . . 250
6 Effect of Powdered Carbon on Effluent BOD . . 251
7 Effect of Powdered Carbon on BOD Rempval . . 251
xiii
-------
TABLES
Gttgreftt':State-of- the-Art- -Activated Carbon Treatment
Table Page
1 Comparative Analysis of Batch Isotherm Data- 24
Refinery & Petrochemical Wastewater
2 Amenability of Typical Organic Compounds to . 25
Activated Carbon Adsorption
3 Relative Amenability to Carbon Adsorption of .26
Typical Petrochemical Wastewater Constitu-
tents
4 Influence of Molecular Structure and Other . .26
Factors of Adsorbability
5 Properties of Several Commercially Available .27
Carbons
6 Typical Properties of Powdered Activated ... 28
Carbon
7 Stages of Thermal Regeneration 28
8 Water Quality at Various Points in Process. . 28
9 Water Quality at Various Stages of Treatment 29
at South Lake Tahoe
10 Carbon Efficiency Per Regeneration Period at 29
South Lake Tahoe--November 1968 through
January 1971
11 Colorado Springs Tertiary Plant Design Data . 30
12 Tertiary Treatment Plant Data Summary-- ... 30
First Regime
13 Tertiary Treatment Plant Data Summary-- ... 31
Second Regime
14 Comparison of Wastewater Strength Battelle . 31
and CRSD Test Programs
xiv
-------
Tables Page
15 Effect of Post-Ozonation on Effluent Organic 31
Quality
16 Quality of Raw Wastewater 32
17 PCT System Design Data 32
18 Summary of Physical/Chemical Treatment . . 33
System Performance
19 Effect of Regeneration on the PCT Carbon . 33
Characteris tics
20 Tertiary Treatment Plants 34
21 Physical/Chemical Treatment Plants .... 35
22 Summary of PCT Pilot-Plant and Full-Scale . 36
Plant Performances
23 Carbon Pilot-Plant Results For Petrochemical 36
and Refining Wastewaters
24 Refinery Wastewater Treatment Results ... 37
25 Carbon Regeneration Activity Analysis ... 37
26 Pilot-Plant Results-Tertiary Carbon Appli-. 37
cation
27 Design Criteria for the Arco Carbon Plant . 38
28 Activated Carbon Adsorption Design Data . . 38
29 Thermal Regeneration Design Data 38
Cautions and Limitations on the Application of
Activated Carbon Adsorption to Organic Chemical
Was tewaters
1 Comparison of Multi-Component Isotherm Data 60
With Single Component Data
XV
-------
Organics Reduction Through Add-on Carbon
At Pilot Scale~
Table Page
1 Daily Performance for Common Wastewater . . 72
Parameters
2 Average Performance Over 4-Day Study . . .
Period For Common Wastewater Parameters 73
Activated Sludge Enhancement: A Viable
Alternative to Tertiary Carbon Adsorption
1 Effluent Summary--Part II 108
2 Sludge Data—Part II . . . . „ 109
3 Powdered Carbon (PC) Requirements 110
Pilot-Plant Activated Carbon Treatment of
Petroleum Refinery Wastewater
1 Controlled Conditions for losterm Study . . 121
2 Adsorption Data for Eleven Activated Carbons 121
3 Contact Time Adsorption 121
4 Anthrafilt Filtration Organic Carbon . . .
Removal 122
5 Mini-Column TOG Results 122
6 Comparison of Mini-Column TOG Removal . . .
Efficiency 122
7 TOG Results of Mini-Column Upflow Mode . . 123
8 Percentage Removal of TOG for Upflow Mode . 123
9 Refinery Process Data 123
10 Refinery Wastewater Treatment Results ... 124
11 Carbon Adsorption Capacity 124
xvi
-------
Table Page
12 Carbon Regeneration Activity Analysis ... 124
13 Solvent Extraction Efficiency Data .... i25
14 Composition By Types of Organics . . . . „ 125
15 Compound Types Indicated by GC-MS 125
Combined Powdered Activated Carbon - Biological
Treatment: Theory and Results
1 Synergistic Effect On DOC Removal with PACT 143
2 Reasons To Consider PACT 144
3 Economic Considerations of Powdered Activa-
ted Carbon Addition 145
4 -Chambers Works Half Full Flowbate.:Test
Test Period 3/13/77 - 3/26/77 Inclusive . . 146
5 Chambers Works Full Flowbate Test
Test Period 4/26/77 - 5/6/77 Inclusive . . 146
Powdered Activated Carbon Enhancement of
Activated Sludge For BATEA Refinery Waste-
water Treatment
1 Operating Conditions 166
2 Properties of Powdered Activated Carbons . 167
3 Phase I and II - Effect of Carbon Type and
Addition Rate on Effluent Quality 168
4 Phase III - Effect of Carbon Type and . . .
Addition Rate, Sludge Age, and Influent
Pretreatment on Effluent Quality 169
5 Phase IV - Effect of High Sludge Age, Low .
Carbon Addition Rate, and Decreased Hy-
draulic Retention Time on Effluent Quality 170
6 Phase III - BAT Guideline and Actual Varia-
bility Factors For Pilot Plant Fed 25 mg/lit-
er of Carbon Al 171
xvii
-------
Activated Carbon Treatment of Combined
Storm and Process Waters
Table Page
1 Typical Performance Data First Rainy 201
Season
2 Adsorption Data 203
3 Cost Data December, 1972 - March, 1973 . . 204
Water Pollution Abatement At BP Oil Cor-
poration's Marcus Hook Refinery
1 Sand Filtration Pilot Plant Performance . 227
2 Rapid Sand Filter Design Data 228
3 Activated Carbon Adsorption Design Data . 22.9
4 Thermal Regeneration Design Data ....." 230
5 Wastewater Analysis ..... 231
6 Analysis of Sand Filter Backwash Water . . 232
7 Analyses of Water Removed from Sand Filters
by Pressure Prior to Backwash 233
8 Wastewater Treatment Plant—Capital Cost . 234
9 Estimated Wastewater Treatment Plant--
Annual Operating Costs 235
Case History: Use of Powdered Activated
Carbon in an Activated Sludge System
1 Operating Conditions of Wastewater Treat-
ment Plant 245
2 Summary of Powdered Activated Carbon Ad-
dition Trials 245
3 Effect of Powdered Carbon on Daily Average
Effluent Suspended Solids 246
4 Effect of Powdered Carbon on COD 246
5 Effect of Powdered Activated Carbon on BOD 246
xvlii
-------
Table Page
6 Comparison of Carbon Treatment Costs and
System Performance 247
7 Effect of Powdered Carbon on Daily Max-
imum Effluent Suspended Solids 247
Case History: The Use of Powdered Activated
CarbonWith A BioDisc-Filtration Process
For Treatment of Refinery Wastes
1 Treatment of RBS Effluent With Powdered
Activated Carbon 259
2 Activated Sludge Effluent Treated with
PAC 260
3 Activated Sludge Effluent Treated with PAC 260
4 Effect of pH on Toxicity 261
5 Effect of pH on Carbon Treatment 261
6 Comparison of Two Carbons 261
7 Comparison of Two Carbons 262
i
8 Granular Carbon Treatment of Trickling
Filter Effluent 263
xix
-------
SECTION 1
ACTIVATED CARBON STATE OF THE ART
"CURRENT STATE OF THE ART OF ACTIVATED CARBON TREATMENT"
Davis L. Ford, Ph.D., P.E.
Senior Vice President
Engineering-Science, Inc., Austin, Texas
The treatment of wastewaters using activated carbon has received wide attention for
several years - more recently catalyzed by the development of effluent quality guidelines
pursuant to Public Law 92-500. These effluent quality criteria and the accompanying
development documents prominently mention carbon as an applicable and attractive
treatment concept, particularly in the Best Available Treatment Economically Achievable
(BATEA) process mode currently stated as necessary to produce the 1983 quality level
objective. Until recently, most of the literature has dealt with exploring theoretical
concepts and documenting experimental results. As more information on full-scale opera-
tions is becoming available, however, it is considered appropriate to review the current
state of the art of activated carbon treatment - both in municipal and industrial sectors.
It is the purpose of this treatise, therefore, to present pertinent and current information
relative to the activated carbon treatment of municipal and industrial wastewaters. A
brief discussion of adsorption concepts and carbon characteristics also is included.
ADSORPTION CONCEPTS AND THEORY
Molecules are held together by cohesive forces ranging from strong valence bonds to
the weaker van Der Waals forces of attraction. These attractive forces are satisfied in the
solid phase interior molecules, having the ability to capture certain fluid molecules as
they contact the surface, van Der Waals forces are the bases for the adsorption of waste-
water constituents onto carbon which has been activated to maximize this interphase
accumulation of liquid constituents at the surface or interphase of the solid phase.
The rate at which substances are removed from the liquid phase (adsorbate) to the
solid phase (adsorbent) is of paramount importance when evaluating the efficacy of acti-
vated carbon as a wastewater treatment process. Unfortunately, the task of quantifying
the many forces acting at the solid-liquid interface is a formidable one. Developing a
mathematical expression which describes the dynamic phenomenon occurring in a continuous-
flow/fixed-bed reactor has been difficult because of multi-variable influences. The overall
adsorption rate represents the combined effects of diffusion through a laminar layer of fluid
surrounding the constituent, surface diffusion, and adsorption on the internal pore surfaces.
Most mathematical solutions for equations which describe concentration/time profiles are
limited to the special case in which only one of these phenomena controls the overall rate
of adsorption (1).
One expression for the continuous-flow regime assumes the diffusion of the constituent
through the liquid phase and through the pores of the carbon (which are rate-limiting),
then combining these resistances in an overall mass coefficient term. Using this rationale:
-------
where:
q = flow rate
C = concentration of the adsorbate
s
D = adsorbent bed depth
C = equilibrium adsorbate concentration
k_r = overall mass transfer coefficient
A more convenient expression of Equation 1 is in terms of the adsorbate rate with
respect to the weight of the carbon in the columns:
q dC/dM = ly/X (C -C) (Eq- 2)
where:
M = weight of the carbon in the column
X = packed density of the carbon in the column
Another proposed model predicts four successively decreasing adsorption rates would be
observed as the adsorption proceeds to equilibrium. The initial rate would be limited by
the rate of adsorbate transfer across the film layer, film diffusion, or, if sufficient turbulence
existed, control would be exerted by the combined rate of external surface adsorption and
macropore filling. After the external surface adsorption capacity was exhausted, there would
exist three secondary adsorption rates controlled, respectively, by the filling of the macro-
pore (an effective radius of 5,000 to 20,000 A°), the transitional pore (20 to 100 A°), and
the micropore (10 to 20 A effective radius). This model is illustrated in Figure 1 (2). It
is inherent in this model that the intraparticle transport occurs as a series of adsorption/
desorption steps, each linear with respect to time, and their summation resulting in a time/
linear function.
The development of adsorbate removal kinetics on a batch basis can be used to appro-
ximate carbon effectiveness and predict organic residual levels. The adsorption.isotherm
is used for this purpose and is defined as a functional expression for the variation of adsorption
with concentration of adsorbate in bulk solution at a constant temperature. The isotherm is
expressed in terms of removal of an impurity - such as BOD, COD, and color - per unit
weight of carbon as a function of the equilibrium impurity remaining in solution. Linear
plots as shown in Figure 2 can be expressed in terms of the empirical Freundlich equation.
This expression relates the amount of impurity in the adsorbed phase to that in solution:
X _ Kr]/n
-ft ~ KC (Eq. 3)
*
where:
X = amount of impurity adsorbed
M = weight of carbon
C = equilibrium concentration of impurity in solution
K,n = constants
The Freundlich isotherm is valid within the context of a batch test for pure substances
and some dilute wastewaters. As shown in Figure 2, its application is limited in certain
-------
cases when a significant portion of the organic impurities are not amenable to sorption,
resulting in a constant residual, regardless of the carbon dosage.
The constants "n" and "K" can be used to define both the nature of the carbon and
the adsorbate. A high "K" and "n" value, for example, indicate good adsorption through-
out the concentration range studied. A low "K" and "n" value would infer low adsorption
at dilute concentrations with high adsorption at the more concentrated levels. Variations
of the constants for selected wastewaters are shown in Table 1 (3).
FACTORS WHICH INFLUENCE ADSORPTION
There are many factors which influence both the rate and magnitude of adsorption -
underscoring the difficulty in developing predictive models which would apply to all
complex wastewaters. A brief discussion of the more important factors is presented herein.
Molecular Structure. The molecular structure, or nature of the adsorbate, is partic-.
ularly important in dictating the degree of adsorption that can actually occur. As a rule,
branched-chain compounds are more sorbable than straight-chain compounds, the type and
location of the substituent (functional) group affects adsorbability, and molecules which
are low in polarity and solubility tend to be preferentially adsorbed. Unless the screening
action of the carbon pores actually impedes, large molecules are more sorbable than small
molecules of similar chemical nature. This is attributable to more solute chemical bonds
being formed, making desorption more difficult.
Inorganic compounds demonstrate a wide range of adsorbability. Disassociated salts -
such as potassium chloride and sodium sulfate - are essentially nonsorbable. Mercuric
chloride and ferric chloride are relatively sorbable, and iodine is one of the most adsorbable
substances known.- Generally, however, a significant reduction in inorganic materials is
not expected in carbon systems.
Organic compound sorbability can be classified to some extent. Primary alcohols and
sugars, for example, are resistant to adsorption, while ethers and certain organic acids
are highly sorbable. Recently published experimental data presented in Table 2 are
indicative of the sorbability of many organic compounds (4). Additional sorbability data
conducted independently are presented in Table 3.
Solubility. An increase in solubility acts to oppose the attraction of the adsorbate to
carboru Thus, polar groups which have a high affinity for water usually diminish adsorption
from aqueous solutions. Conversely, the greater adsorption of the higher aliphatic acids
and alcohols is attributed in part of their relatively lower solubility in an aqueous solution.
There are exceptions to this, as in the case of the highly soluble chloracetic acid (5).
lonization. lonization is generally adverse to adsorption by carbon as strongly-
ionized materials are poorly adsorbed. Hydrogen ions, which are significantly adsorbed
under some conditions, would be an exception to this. Some negative ions, therefore, are
more sorbable when associated with hydrogen ions. For this reason, mineral acids - such
as sulfuric acid - are sorbable at higher concentrations./
-------
A change in ionization can drastically affect adsorption. A low pH, for example,
promotes the adsorption of organic acids whereas a high pH would favor the adsorption of
organic bases. Phenol adsorbs strongly at neutral or low pH while the adsorption of the
phenolate salt at a high pH is poor. The optimum pH is therefore solute-specific and must
be determined for each wastewater.
Temperature. As adsorption reactions are generally exothermic and high temperatures
usually slow or retard the adsorption process, lower temperatures have been reported to favor
adsorption (1,5). Very little information has been presented, however, which documents
significant shifts in adsorbability within the temperature range of 65 F to 90 F (typical of
most wastewaters). Lower temperatures should increase adsorption, but the effect in aqueous
solutions is very small.
Adsorption of Mixed Solutes. Most wastewaters contain a myriad of compounds which
may mutually enhance, interfere, or act independently in the adsorption process. Factors
which affect overall adsorption of multiple adsorbates include the relative molecular size
and configuration, the relative adsorptive affinities, and the relative concentrations of the
solutes (1). Predictive models obviously require validation for complex wastewaters, as
extrapolation from investigations using synthesized wastes containing controlled concen-
trations of selected adsorbates may not reflect all of the interactions occurring in the waste.
A summary of the factors which potentially influence adsorbability is presented in
Table 4.
PROPERTIES OF ACTIVATED CARBON
Activated carbons are made from a variety of materials including weed, peat, lignin,
bituminous coal, lignite, and petroleum residues. Granular carbons produced from medium
volatile bituminous coal or lignite have been most widely applied in the treatment of
wastewater as they are relatively inexpensive and readily available. The activation of
carbon is essentially a two-phase process which includes burning off the amorphous decom-
position products and enlarging the pores in the carbonized material (6). The burn-off, or
carbonization, phase involves drying the carbon at approximately 170 C, heating the
material to 270 C to 280 C with the evolution of carbon monoxide, carbon dioxide, and
acetic acid, and, finally, completing the carbonization process at a temperature of 400°C
to 600 C. The yield following carbonization is approximately 80 percent. The intermediate
product is then activated by using carbon dioxide or steam at a temperature of 750°C to
950 C, burning off decomposition products, exposing and widening the pores in the development
of macroporous structure. In the activation process, the kind of adsorptive powers developed
are determined by:
1. the chemical nature and concentration of the oxidizing gas;
2. the extent to which the activation is conducted;
3. the temperature of the reaction; and
4. the amount and kind of mineral ingredients in the char (5).
The proper activation conditions provide an oxidation which selectively erodes the
surface so as to increase the surface area, develop greater porosity, and leave the remaining
-------
atoms arranged in configurations that have specific affinities.
Activated carbons from coke, coal, or lignite have specific properties, depending on
the material source and the mode of activation. Property standards are therefore helpful
in specifying carbons to perform a specific task.
As a rule, granular carbons made from calcined petroleum coke have the smallest
pore size, the largest surface area, and the highest bulk density. Lignite carbon has the
largest pore size, least surface area, and the lowest bulk density. Bituminous coal has a
bulk density equal to that of petroleum coke and an average pore size and surface area
somewhere between those of petroleum coke and lignite-based carbons (7). A brief
description of carbon properties follows:
Jotal Surface Area; This is the surface area of carbon expressed in square meters per
gram, normally measured by the adsorption of nitrogen gas by the BET method (8).
Carbon Density: Apparent density is the weight in grams of one ml of carbon in air.
Bulk density, backwashed and drained, is often used and is usually expressed in
pounds per cubic foot.
Particle Size Distribution: The particle size distribution is critical in terms of
hydraulic loading and backwash rates. Commonly manufactured particle size ranges
for granular activated carbons expressed in limiting U.S. Standard Sieve Sizes
include 8 x 16, 8 x 30, 10 x 30, 12 x 40, and 20 x 40. Effective sizes (sieve
opening at which 10 percent of the material passes) range from 0.55 mm to 1.30 mm.
In general, the uniformity coefficient (the millimeter opening at which 60 percent
of the material passes divided by the millimeter opening at which 10 percent of
the material passes) for granular activated carbon should not exceed 2.1.
Adsorptive Capacity: The best measure of adsorptive capacity is the effectiveness of
the carbon in removing the critical constituents (BOD, COD, color, etc.) from the
wastewater in question. Various tests, however, have been developed to give
relative removal capacities of activated carbon under specific conditions. Phenol
number is used as an index of a carbon's ability to remove taste and odor compounds;
tannin is representative of organic compounds added to water by decayed vegetation;
and iodine and molasses numbers are used to show if a carbon is activated. The
iodine number, defined as the milligrams of iodine adsorbed by one gram of carbon
(with an iodine concentration in the residual filtrate is 0.03) is probably the most
widely used method of expressing carbon capacities. It generally can be correlated
to the ability of an activated carbon to adsorb low molecular weight substances
while the molasses number correlates the carbon's ability to adsorb higher molecular
weight substances. The iodine number measures the micropores having an effective
radius of less than 20 angstroms and the molasses number measures the transitional
pores ranging from 20 to 500 angstroms.
These are the principal general parameters used in specifying carbons, the objective
being the selected carbon
-------
1. has the adsorption capability to meet the effluent requirements;
2. incurs minimal losses occurring during carbon transport and regeneration;
3. has good hydraulic characteristics with respect to head loss or pressure drop; and
4. represents the most cost-effective media to accomplish the prescribed task.
These, of course, represent general parameters of specification and should be augmented by
test data and process requirements developed from bench- or pilot-scale evaluation using
representative wastewater samples and selected carbons.
Comparative properties of the most widely used carbons in wastewater treatment -
those from lignite and bituminous coals - are shown in Table 5 (6); a broader presentation
of properties from commercial carbon manufacturers also is tabulated in this Table (9).
Typical properties of a petroleum-base powdered carbon developed but not yet marketed by
AMOCO are shown in Table 6.
REGENERATION
Because of economic and solid waste disposal considerations, it is generally more
feasible to regenerate spent carbon for subsequent reuse than to dispose of it. In the
regeneration process, the objective is to remove from the carbon porous structure the
previously-adsorbed materials, thus reinstituting its ability to adsorb impurities. There are
several modes of regeneration which can be applied including thermal, steam treatment,
solvent extraction, acid or base treatment, and chemical oxidation. Of these, only
thermal regeneration using a multiple-hearth or rotary-tube furnace is widely applied in
wastewater treatment. The discussion on regeneration therefore centers around thermal
treatment.
Thermal regeneration refers to the process of drying, thermal desorption, and high
temperature (1,200 F to 1,800 F) heat treatment in the presence of limited quantities of
ox\i\dizing gases such as water vapor, flue gas, and oxygen (10). Multiple-hearth furnaces
are the most commonly used, although rotary kilns or fluidized-bed furnaces are occasionally
applied.
The sequential stages in thermal regeneration are shown in Table 7. Of these steps,
the gasification stage is the most critical. The system should be controlled in order to
selectively gasify the sorbed organic material while minimizing the gasification of the
carbon structure. Basically, there are three major variables involved in thermal regenera-
tion. These include furnace temperature, residence time, and the carbon loading. Of
these three, furnace temperature is controllable, although it may take several hours to
adjust, the residence time can be changed by varying the rabble-arm speed in a multiple-
hearth furnace, or the rotation rate and slope of the tube in a rotary kiln. Very little can
be done to change the carbon loading, which affects the severity of regeneration required.
The multiple-hearth furnace is the most commonly used system for granular carbon
regeneration. A schematic diagram of a typically-designed multiple-hearth furnace is
shown in Figure 3. The wet spent carbon is added at the top of the furnace and drops from
hearth to hearth, being raked along by the rabble arms. The temperatures shown for the
various hearths are typical gas temperatures for granular carbon regeneration. It was
-------
found that the addition of steam on hearths four and six gave a more uniform distribution
of temperatures throughout the furnace. The effect of steam is to reduce the apparent
density and increase the iodine number of the regenerated carbon (6). Normally, about
one pound of steam per pound of carbon is used. Natural gas or fuel oil is usually added
to supply the auxiliary heat. Although the fuel requirement varies, generally about
3,000 BTU/lb of carbon and 1,300 BTU/lb of steam generated is required.
Rotary kiln furnaces are not being installed in new facilities but many are still in use.
The kilns are usually direct-fired, counter-current units with steam injection with up to
10 percent excess air. Again, natural gas or fuel oil is commonly used as auxiliary fuel.
The heat efficiency of rotary kilns is less than that of multiple-hearths.
Fluidized-bed systems for powdered carbon regeneration have been evaluated on a
pilot scale and appear to have promise (11, 12, 13). In this process a bed of inert
granular material such as sand is fluidized by the upward flow of hot gases and the wet
spent carbon is injected directly into the bed. The inert bed particles provide a reservoir
of heat which is rapidly transferred to the spent carbon particles. The heat economy of
the fluidized-bed furnace is less favorable than the multiple-hearth furnace or the rotary
kiln - providing afterburners are not required.
There is considerable debate as to the effect of regeneration on carbon capacity,
carbon losses during the regeneration cycle, and hydraulic characteristics of the regen-
erated carbon. Much of this is focused on the regeneration effects on lignite as compared
to bituminous carbon. The manufacturers of lignite claim that, after a number of regen-
eration cycles, both carbons tend to become more like each other and perform the same (9).
Specifically, lignite - while softer and lighter than bituminous - is less susceptible to
change during regeneration and can be regenerated under less severe conditions (lower
temperature and shorter residence time). The bituminous manufacturers note that lignite
carbon losses per cycle are higher than bituminous, indicating a higher operating cost,
and bituminous adsorption capacities based on molecular weights ranging from phenol at
94 to dextran at 10,000 are significantly higher than lignite (14). It is not the intent in
this writing to favor one over the other, but simply to present information from manufacturers
of both types. Based on discussions with independent evaluators of bituminous and lignite,
several trends in thinking evolve. First, few full-scale plants using lignite are currently
in operation, necessitating an evaluation based on pilot-plant studies. There appears to
be little firm evidence that bituminous carbon capacities are higher than those of lignite,
although screening tests should be conducted in any event to determine the optimum carbon
to perform a specific task. Somewhat higher losses have been noted during regeneration
of lignite as compared to bituminous (six to nine percent for lignite, and four to seven
percent for bituminous over three cycles). Lignite requires lower temperature for its
regeneration and more skilled regeneration operation and control. There has been little
evidence to show significant hydraulic differences as demonstrated by head loss charac-
teristics in the two carbons following regeneration. Carbon selection is therefore predi-
cated on process test results, capital cost requirements, and an annualized cost evaluation.
-------
EXPERIENCE IN CARBON TREATMENT OF MUNICIPAL WASTEWATERS
Treatment of municipal wastewaters using activated carbon has been practiced on a
full scale for over 10 years. For this reason, ample performance records are presently
available for a realistic evaluation of this application, both in the physical/chemical and
biological effluent polishing modes of treatment. A summary of these systems has been
tabulated elsewhere, but it is the intent in this section to document selected municipal
facilities and discuss salient features of these systems which are pertinent to this overall
discussion on carbon. The two effluent polishing systems selected for discussion are the
South Lake Tahoe and Colorado Springs facilities. The physical/chemical discussion
centers around the new plant being constructed for the Cleveland Regional Sewer District,
with the Garland, Texasj Rosemount, Minnesota; and Pomona, California plants also
included. A syllabus of these selected case histories is presented as follows.
South Tahoe Tertiary Treatment System. The first major and most widely publicized
application of activated carbon in treating domestic wastewater was the system constructed
for the South Tahoe Public Utility District in 1965. The plant was conceived to polish the
effluent from an existing activated sludge facility to a quality level which would have
little or no impact on the receivirig waters in an ecologically-sensitive area.
The facility consists of a chemical mix-coagulation, precipitation, and clarification
unit, an ammonia stripping tower (which has been used infrequently), a recarbonation and
settling basin, mixed-media filters, carbon adsorption, and final chlorination. A simplified
flow diagram is shown in Figure 4 (15). The water quality at various points in the process
following 18 months of operation is presented in Table 8 (16). The more recently reported
water quality at the various stages of treatment is presented in Table 9 (9). The carbon
(bituminous) efficiency per regeneration period at the South Tahoe Plant is cited in
Table 10 (6).
It should be recognized that this facility was a demonstration project, involving expensive
capital and annual expenditures. With the possible exception of nitrogen, the effluent
quality should represent the best level obtainable in treating a domestic wastewater effec-
tively using maximum biological treatment polished by chemical treatment, filtration, and
carbon adsorption.
Colorado Springs Tertiary Treatment System. As at Tahoe, the Colorado Springs
treatment system involves chemical treatment, filtration, and.carbon adsorption for polishing
a slipstream of biologically treated effluent. There are basically two regimes of historical
data from this facility. The first was when the slipstream was taken from an overloaded
trickling filter effluent (the flow diagram schematically illustrated in Figure 5). The second
regime follows the addition of a contact stabilization facility to the biological treating
component of the system. This indicates that the influent to the polishing portion of the
system, the design data for which are presented in Table 11, has a lower concentration of
biodegradable organics in the second regime than in the first (17). This is detected in a
general comparison of the effluent quality observed during the two periods. Effluent quality
data from a three-month period in 1972 indicate an approximate level obtained during the
first regime. The average quality values for the reactor-clarifier influent and effluent,
-------
filter effluent, lead carbon column effluent, and polishing carbon effluent are tabulated
in Table 12. The data averaged during the second regime of operation are shown in
Table 13 (17). Considerable improvement is noted in the comparison, underscoring the
need for adequate biological pretreatment if the tertiary system is to realize maximum
performance. It is noted that the Tahoe and Colorado Springs (second regime) effluents
are relatively similar - both plants have adequate biological pretreatment. The deteriora-
tion of effluent quality when this is not the case is evidenced in Table 12.
The Colorado Springs facility is using a bituminous coal activated carbon, using a
multiple-hearth furnace regeneration system. As in the other carbon systems, there has
been sulfide generation problems in the carbon vessel. This problem has been ameliorated
by adding a copper sulfate solution (100 mg/l as CuSOJ to the top of the carbon columns
and altering the backwashing frequence to remove biological solids.
Cleveland Regional Sewer District Physical/Chemical System. A physical/chemical
wastewater treatment plant of which activated carbon is an integral part is presently being
constructed for the Cleveland Regional Sewer District (CRSD) at the Cleveland Westerly
plant site. This system is designed to receive an average flow of 50 MGD with peak flows
of 100 MGD. A flow diagram of this system is shown in Figure 6. Extensive pilot-plant
treatability studies were conducted prior to finalizing the design as there is a significant
industrial contribution in the raw waste load (18). Moreover, assurance of meeting the
NPDES requirements of 20 mg/l BOD (30-day average) and 30 mg/l (seven-day average)
was required. These studies, combined with those more recently conducted by CRSD,
provided some interesting results with respect to carbon adsorption.
The raw waste load of the Westerly collection system documented during the two
pilot studies (1970-71 and 1974-75) is shown in Table 14. The strength of this wastewater,
combined with the organic nature of the industrial component, created some difficulty in
meeting the required permit levels during the pilot-plant evaluation. Although these
results could be overly-pessimistic for various reasons, CRSD evaluated several process
alterations with the objective of improving process efficiency and effluent quality. Most
of this investigation centered around the use of ozone injection - both as a post-carbon
and as a precarbon mode of operation. The postozonation step was originally conceived
as strictly a disinfection step, although some further reduction of BOD through ozone
oxidation was anticipated. However, studies indicated the ozone requirement for disin-
fection to the required level of 200 fecal coliforms per 100 ml ranged from 2.6 to 10.9
mg/l (19). The higher doses were attributed to the ozone demand from the sulfides
generated in the carbon columns and ferrous bicarbonate resulting from iron slugs in the
influent. This ferrous species was solubilized in the recarbonation step and was not
altered while passing through the carbon columns. As the higher demands of ozone are
economically prohibitive, reducing the ozone-demanding constituents or looking closer
at chlorine was merited. Moreover, postozonation did not improve organic effluent
quality - as indicated in Table 15. Ozonation prior to the carbon columns was then
applied by CRSD with encouraging results. This quality improvement in terms of a
frequency analysis of carbon column effluent BOD is shown in Figure 7. The same trend
in terms of COD is shown in Figure 8. These results infer.that some nonsorbable compounds
are converted to more sorbable intermediates through ozone transformations. This possibility
-------
is substantiated in the literature, documenting ozone oxidation of aldehydes, ketones, and
alcohols to carboxylic acids, the product being more sorbable than the reactants (20, 21).
Based on these results, the possibility of injecting ozone prior to the carbon columns is
being seriously considered because of the potential advantages of:
1. possibly transforming nonsorbable compounds to sorbable compounds, enhancing
carbon adsorption efficiency;
2. reducing the bacterial load to the carbon columns through particle sterilization; and
3. increasing the dissolved oxygen level in the carbon column influent, reducing the
possibility of sulfide production and anaerobic bacterial activity.
There may be problems concerning this approach, such as the technical problems involved
with injecting ozone in the pressurized carbon influent line, the high decomposition rate of
ozone, and enhancing corrosion potential in the carbon reactors. Moreover, disinfection
following carbon adsorption still needs to be resolved. However, the advantages presently
appear to outweigh the disadvantages and this approach is being seriously considered as a
process modification.
Garland Physical/Chemical System. A physical/chemical treatment plant is presently
being constructed at Garland, Texas. This 30 MGD plant is being designed to produce an
effluent having less than 10 mg/l BOD and TSS. The raw wastewater quality used for design
is presented in Table 16 (22). The flow diagram for this facility is shown in Figure 9. The
carbon basins are common-wall concrete facilities, each having a surface area of 950 square
feet and a depth of 10 feet. Nine basins can treat,the 30 MGD, while the remaining basin
is off-line for the backwashing or regeneration. The wastewater is in .contact with the carbon
for a minimum of 30 minutes and the upflow rate is 2.5 gpm/ft at design flow. A 14.5 foot
multiple-hearth furnace is designed to regenerate 80,000 pounds of carbon daily. Based on
the treatability studies, an effluent quality of 15 mg/l COD and 10 mg/l BOD is predicted
(22).
Rosemount Physical/Chemical System. The Rosemount, Minnesota physical/chemical
facility has been in operation since 1974. The system receives between 0.3 and 0.6 MGD
and consists of screening, chemical clarification using either lime or ferric chloride,
prefiltration, upflow activated carbon adsorption, postfiltration, clinoptilolite ammonia
exchangers, and disinfection. During the first year of operation, the average BOD removal
was 90 percent, the average effluent BOD concentration being 23 mg/l. More recently,
however, operating problems have been reduced and the reported effluent BOD has been in
the range of five to 10 mg/l (23).
Pomona Physical/Chemical Pilot Plant. A pilot-scale physical/chemical system has
been operating for 27 months in Pomona, California, under the auspices of the Los Angeles
County Sanitation District and the Environmental Protection Agency. The objectives of
this study included an evaluation of the long-term effectiveness of granular activated
carbon in the removal of soluble organic matter from chemically-clarified municipal waste-
water, controlling hydrogen sulfide generation in the carbon columns, and determining the
effects of repeated thermal generations on carbon characteristics and performance.
A flow diagram of the pilot system is shown in Figure 10, the design data for which are
presented in Table 17 (24). The primary organic control parameters used in these studies
10
-------
were total chemical oxygen demand (TCOD) and the dissolved chemical oxygen demand
(DCOD). A summary of the overall performance of the physical/chemical system is
presented in Table 18. These quality levels represent average values observed during the
27 months of study. For comparative purposes, the effluent quality from an 8.0 MGD
activated sludge system treating the same wastewater is also included. As noted in Table
18, the major portion of pollutants, with the exception of DCOD, was removed in the
clarification phase.
The problems of excessive biological growths in the columns, particularly the
biochemical reduction of sulfates, has persisted in many of the case histories cited. A
review of this problem was therefore an integral part of the Pomona study. As micro-
organisms have the hydrogen acceptor preference of molecular oxygen, nitrate, sulfate,
and oxidized organics, it follows that an environment having the presence of sulfates and
organics and the absence of molecular oxygen and nitrates would favor the anaerobic
sulfate reducers. The sulfide levels in the effluent approached as high as six mg/l. This
was a major concern as sulfide production was correlated directly to net head loss in the
column (head loss before the daily backwash minus the head loss after backwash). The
fact that the column head loss was caused primarily by abundant biological growths in the
column is underscored by the graphical relationship depicted in Figure 11 (24). Several
methods of ameliorating the biological proliferation and sulfide production were under-
taken. This concluded oxygen addition, intensive air/water backwash, chlorination, and
sodium nitrate addition to the carbon column at an average dosage of 5.4 mg/l (as N) was
determined to be the most effective in terms of retarding sulfide generation in the carbon
column.
The spent carbon used in the regeneration studies was backwashed, dewatered to
approximately 50 percent moisture, conveyed to a six-hearth furnace, and regenerated at
temperatures ranging from 1,650 F to 1,790 F. Steam in the amount of 0.6 Ib/lb of
carbon was added to the lower two hearths to enhance the regeneration. The regeneration
cycle normally took from 53 to 66 hours to complete. The effects of regeneration on the
physical properties of the carbon are shown in Table 19 (24). A reduction in the iodine
number (a measure of the extent to which the micropores have been cleared) is noted,
although the molasses number (a measure of the extent of macropore clearing) is essentially
unchanged. The ash content of the carbon, which measures the amount of calcium and
other inorganic residues picked up by the carbon during service increased over 60 percent
from the virgin level during the first regeneration. The ash increase during the second
regeneration was less, however, with a decrease during the third regeneration. The over-
all carbon regeneration loss (bituminous coal) ranged from 2.5 to 6 percent with an average
loss of 4.3 percent over three regenerations.
Summary. A summary of the municipal wastewater treatment using the tertiary appli-
cations of activated carbon has been published recently and is shown in Table 20 (6). A
similar summary for a straight physical/chemical application of carbon is shown in Table
21 (6) and in Table 22 (25).
EXPERIENCE IN CARBON TREATMENT OF INDUSTRIAL WASTEWATERS
Although ample carbon treatment data from full-scale facilities are becoming available
11
-------
from the municipal sector, much of the performance data in industrial applications evolves
from pilot-plant studies. There are several full-scale carbon systems treating industrial
wastewaters, however, and this section includes results from pilot studies and selected
operational systems. Specifically, results from pilot-plant studies - primarily in the refinery
and petrochemical sector - will be cited, as well as the full-scale systems at the ARCO
Refinery near Wilmington, California, the Reichhold Chemicals system in Tuscaloosa,
Alabama, and the British Petroleum Refinery in Marcus Hook, Pennsylvania.
Pilot Carbon Studies. There have been pilot studies conducted in the industrial sector
evaluating activated carbon as both a total (physical/chemical) process and as an effluent
polishing (tertiary) unit. Nine of these studies have been conducted by the author for
various petrochemical and petroleum refining facilities (26). The efficiency ranged from
50 to 86 percent COD removal, as noted in Table 23. It should be recognized that these
results were obtained using virgin carbon with controlled hydraulic and feed rate regimes.
The attendant problems of full-scale operation and the effect of using regenerated carbon
on process efficiency should be considered when translating these results to what might
occur in an operating carbon treatment system.
A comprehensive pilot-plant study treating petroleum refinery effluents by activated
carbon was conducted recently by the Environmental Protection Agency (27). Both API
separator effluent and biologically-treated effluent were charged to the columns in order
to obtain a comparative evaluation. The results of this study in terms of BOD and COD
removal are plotted in Figures 12 and 13, respectively. It is appareDt that, when operated
in parallel, the biological system was more effective in removing BOD and COD -
particularly the former. This is consistent with the results observed in pilot studies
conducted by the author. It is also noted that carbon adsorption following biological
treatment was particularly effective in reducing both the BOD and COD to low levels. The
residual COD is in the same range as that cited in Table 23 when the carbon application
mode and influent COD levels were similar. The complete results of this study are shown
in Table 24. It is noted that there is no removal of cyanides or ammonia although there
was a reduction of the cited organic constituents, particularly phenols. There was surpris-
ingly good removal of chromium, copper, iron, and zinc although the exact mechanisms
of removal were not determined. The carbon capacity observed in the columns was 0.31
Ib TOC removed/lb of carbon, while the isotherm determined capacity was 0.12 Ib/lb.
This difference was attributed to biological activity observed in the column. The carbon
regeneration activity analysis is reported in Table 25, the change in iodine and molasses
numbers showing trends similar to those observed in previous studies. The investigators
emphasized that these data were generated using virgin carbon in the columns, and cautioned
that the presence of iron or aluminum salts present in the effluent could have a deleterious
effect on the carbon through the regeneration cycle. They hypothesized that aluminum
salts can remain on the surface of the carbon during regeneration, reducing the effective
surface area of the carbon and reducing its adsorption capacity. Moreover, iron salts can
catalyze oxidation reactions of the carbon and the gases during regeneration, thus perma-
nently damaging the carbon structure (27).
Extensive pilot-plant studies have been conducted recently by Union Carbide evaluating
activated carbon as a tertiary process treating effluent from an activated sludge plant
12
-------
receiving petrochemical wastewaters (28). The objectives of this study were both to
establish a basis of justification for applying activated carbon and to optimize its design.
Controlled organic and hydraulic application rates were applied to pilot-scale carbon
columns. The results of the various throughput rates are listed in Table 26. The first
verity which can be deducted from this Table is the fact that, as the activated sludge
improves in quality, the efficiency of the adsorber in terms of COD or soluble organic
carbon (SOC) improves. The improved efficiency with decreasing hydraulic throughput
rates is also apparent. There was still a 30 to 50 percent COD residual in the adsorber
influent attributable to organic compounds which are neither biologically degradable nor
sorbable on activated carbon. There was a high BOD residual at the same throughput
rate, although this residual was reduced to 35 to 50 percent of the adsorber influent at the
uneconomical throughput rate of 0.15 bed volumes per hour. This BOD residual includes
low molecular weight-oxygenated organics such as aldehydes, ketones, alcohols, glycols,
and other polar organics which are adsorbed to a very limited extent. Other observations
from this study are cited as follows:
1. breakthrough curves for multi-component wastewaters tend to be poorly defined
and sporadic rather than sharp sigmoidal wavefronts;
2. in multi-bed series adsorbers, the lead bed removes the more readily-adsorbable
organics and has the highest adsorptive capacity - typical capacities through both
beds ranged from 0.2 to 0.4 Ib COD/lb of carbon and 0.1 to 0.3 Ib BOD/lb of
carbon (the observed capacities for the first bed generally were twice as high as
those observed in the second); and
3. the maximum hydraulic application rate that appeared to be technically justifiable
was 0.5 bed volumes per hour (through both columns of the series).
Full-scale Carbon Studies. There are currently several full-scale operational carbon
treatment systems in the United States. Three systems from which data are available are
cited here. These include the Atlantic Richfield (ARCO) system in California, the
Reichhold system in Alabama, and the British Petroleum (BP) system in Pennsylvania.
ARCO Carbon Treatment System (Watson Refinery)
The first full-scale carbon treatment system in the petroleum refining industry was
installed by ARCO at their Watson Refinery near Wilmington, California. This facility
was necessitated by a resolution from the Los Angeles Regional Water Quality Control
Board which limited the amount of COD that could be discharged by industry into the
Dominguez Channel in Los Angeles County. At the time, process wastewaters could be
treated in the County's primary treatment unit, but they could not accommodate storm
runoff from the Refinery. This presented a problem as the storm water collection system
and the process sewers were interconnected. Therefore, during periods of rainfall, the
combined storm runoff and process flow could not be sent to the County nor could it be
discharged to Dominguez Channel because of its COD concentration. A holding basin
and a carbon treatment plant were therefore deemed to be the most efficacious way of
treating this combined flow on an intermittent basis (29).
The treatment system, completed in 1971, is illustrated schematically in Figure 14.
The process wastewater flows through an API separator and then through pH adjustment and
chemical flocculation tanks. Following the addition of both coagulants and polyelectro-
lytes, the flow is routed to one of two circular dissolved air flotation units. At this
13
-------
point/the oil and grease (O & G) levels are normally below 15 to 20 mg/l. Under dry
weather conditions, the flow goes to the County for further treatment. During periods of
rainfall, it is diverted to a 50 million gallon holding basin where it is surged and conveyed
"as required" to the carbon treatment plant. The plant consists of 12 adsorber cells, the
flow to which is controlled by a handwheel-operated slide gate. Each bed can be back-
washed with treated water from a backwash sump, the backwash water being returned to the
holding basin. The carbon handling and regeneration system includes storage hoppers for
spent and regenerated carbon, a pump, eductor, piping, and centrals to convey the carbon
from any cell to the regeneration furnace and back, and the multiple-hearth furnace with
gas scrubber. It is gas-fired and supplemental steam and air are added on two hearths. An
afterburner section is separately gas-fired to raise the off-gas temperatures to approximately
1,450 F (necessary to combust the organic vapors in the exit gas).
The actual design criteria for the system are presented in Table 27. Based on a
probability distribution of two typical runs presented in Figure 15, the effluent COD was in
the range of the predicted level if the influent concentration did not exceed the design
basis. Chlorination was attempted once the effluent COD exceeded the design level on
the premise that the halogenated compounds would be more effectively adsorbed and the
overall performance would be enhanced. There was no noticeable improvement in effluent
COD when this step was concluded, however (29). Although oil and grease was not cited
as a design parameter, the observed effluent concentration and its dependence on the influent
level is shown in Figure 16. The performance of this full-scale system in terms of COD
removal approximates the design basis and the observed reduction of COD and O & G is
consistent with pilot studies using similar wastewaters. It was noted, however, that algal
proliferation in the holding basin adversely affected the carbon plant performance. When
excessive algal growths developed, a deterioration of effluent quality and more frequent
backwashing was characteristic of the system. At one time, the algae became so concen-
trated that the carbon plant had to be shut down. Copper sulfate was added to the holding
basin to minimize algal growth and ease this problem.
The most significant variance of observed performance from the basis of design was the
carbon capacity. Although a precise determination was not possible, the loading based on
several runs ranged from 0.30 to 0.35 Ib COD removed/lb of carbon, rather than the 1.75
Ib COD/lb carbon prediction cited in Table 27.
With the exception of carbon capacity, the carbon plant generally performed as expected.
It is no longer in operation, however, primarily because the treatment requirements imposed
by Los Angeles County have been changed.
Reichhold Carbon Treatment System (Tuscaloosa Plant)
The Reichhold Chemical Plant produces sulfuric acid, formaldehyde, pentaerythritot,
sodium sulfate, sodium sulfite, orthophenylphenol, and a number of synthetic resins. A
carbon adsorption system was designed to treat an effluent flow of 500,000 gpd having an
average BOD concentration of 390 mg/l, a COD level of 650 mg/l, and a pH ranging from
5.4 to 12.3 (30). An agreement between the Company and the Alabama Water Improvement
Commission established an organic removal objective of 90 percent.
14
-------
A physical/chemical.treatment plant using activated carbon was designed and
constructed and Figure 17 presents a simplified flow diagram. The process streams are
routed to a 2.5-day retention equalization basin where the water is pumped to an acid
mixing chamber. Concentrated sulfuric acid is added to maintain a pH range of 6.5 to
8.5. A nonionic polymer is added prior to the flocculation chamber. Following gravity
sedimentation, the water is pumped to one of two moving-bed adsorbers. Each adsorber
contains 124,000 Ib of granular activated carbon. At the design flow rate of 175 gpm/
adsorber, the empty bed contact time is approximately three hours. Treated water is
collected in a trough at the top of each adsorber where it flows to a final retention tank
and then to the river. A conventional regeneration furnace with an afterburner and wet
scrubber system is used for reactivation of the carbon.
In the first few months of operation, the results have been reported as meeting the
effluent quality requirements. The effluent BOD has been in the 35 to 40 mg/l range,
while a 90 percent removal of COD infers an effluent at the 65 to 70 mg/l level (30).
British Petroleum Carbon Treatment System (Marcus Hook, Pennsylvania)
The British Petroleum (BP) Refinery is a 105,000 barrel/day Class "B" refinery located
in Marcus Hook, Pennsylvania. In order to comply with discharge standards prescribed by
the Delaware River Basin Commission (DRBC), a preliminary engineering program and
treatability study was undertaken by the Company. Based on these results, the decision to
treat the effluent from the existing API separator to the required level using a filtration/
carbon adsorption system was made.
The treatment system, placed into operation in March 1973, is shown in Figure 18
(31). The API separator effluent flows to an intermediate surge basin where it is pumped to
three downflow filters, operated in parallel. The flow rate ranges from seven to 12 gpm/
ft , depending on the operation, and the media consists of 2.5 ft of anthracite and 4.5 ft
of sand. The fJlters are backwashed using filtered water, with provisions for air scouring
at 7.1 scfm/ft . The design filter pressure is 47.5 psi and the maximum allowable pressure
drop through the media is 6.5 psi. The backwash interval is 12 hours.
Three carbon adsorbers (10 ft diameter and 65 ft high) are operated in parallel. Each
adsorber contains 92,000 pounds of granular activated carbon in a bed depth of 45 feet.
An additional 8,000 pounds occupy the lower and upper cone areas. Flow to the three
adsorbers is controlled by the level in the filtered water holding tank. The upward flow
rate is 8.5 gpm/ft , which gives an empty bed contact time of 40 minutes. Spent carbon
at the bottom of the adsorber is pulsed out at the rate of approximately 1,000 Ib/day.
Fresh carbon is then added to the top of the column from a feed hopper. The design
criteria and activated carbon properties for the system are shown in Table 28 (32).
The regeneration facility is a five-foot diameter multiple-hearth furnace. The design
feed rate to the six-hearth furnace is 125 Ib/hr. In an atmosphere controlled by the addition
of steam, the adsorbed organics are volatized and oxidized. In order to insure complete
oxidation, all flue gases pass through an afterburner fired by refinery fuel gas and maintained
at a temperature of 1,350°F. A wet scrubber is included for gas cooling and the removal
of particulate matter. The design data for the thermal regeneration system is shown in
15
-------
Table 29.
As the full-scale system has been in operation for over two years, performance data are
available (31, 32). During the first months of operation while the adsorber still contained
virgin carbon, the COD removals were determined to be independent of influent concentration
but dependent on contact time. This relationship is shown in Figure 19. The oil and grease
removal during this period, shown in Figure 20, followed two different regimes. Good removal
was observed initially prior to the start of carbon pulsing. However, once the pulsing began,
taking the pulsed bed out of operation, the adverse effect on the remaining two columns is
reflected in terms of deteriorating effluent quality. The reduction of oil removal with increasing
increasing influent concentrations is attributed to both the oil removal mechanism and pulse
bed "off-line" mode of operation.
There have been four distinct phases of operation since the carbon system came on line.
The first period of data reported herein occurred when virgin carbon was in the adsorbers and
the foul water condensate stream from the Fluidized Catalytic Cracking Unit (FCCU) was not
yet included in the raw wastewater stream. The second period included the FCCU stream,
but virgin carbon was still present in the adsorbers as the wave front had not yet reached the
top of the carbon bed. The third period encompassed a time following a complete turnover of
the carbon bed and still included the FCCU stream. The fourth period excluded this stream
and incorporated a modification to the column septum design, still following a complete carbon
turnover. The average and maximum adsorber effluent concentrations for several parameters
observed during each period is tabulated in Table 30. A significant deterioration in quality
following the inclusion of the FCCU foul water condensate and complete carbon turnover in
the reactor is noted in Figure 21. This has been attributed to inadequate pretreatment of the
API separator effluent in terms of O & G and soluble organic removal as well as significant
buildups of anaerobic biological growths and oily materials in the carbon media. A 40
percent decrease in adsorptive capacity of the regenerated carbon also has been observed.
The iodine number of the regenerated carbon is 560 to 680, while the virgin carbon was in
the range of 950 to 1,000 and the molasses number of the regenerated carbon was 280, or
approximately 50 higher than the virgin carbon. This indicates u decrease in micropores and
an increase in macropores during the regeneration cycle.
The Company is currently reviewing methods of improving process performance over that
observed in the last three periods of operation, or possibly changing the system concept
altogether. Particular emphasis is being placed on more adequate pretreatment, including
biological oxidation, as steps necessary to effectively use the existing carbon system.
SUMMARY
Activated carbon treatment of industrial wastes, while promising, must be carefully
evaluated before process decisions are made and capital funds are committed. As noted in
the pilot- and full-scale case histories presented herein, breakthrough geometry and adsorption
kinetics of multi-component wastewaters are difficult to define, many organic compounds
are not amenable to carbon adsorption, and the effects of regeneration on carbon capacities
are variable and unpredictable. For these and other reasons, comprehensive testing and
technical reviews are a necessary prerequisite to process commitment.
16
-------
Unfortunately, most of the literature regarding carbon treatment of industrial waste-
waters centers around pilot-plant results. There are, however, sufficient data from pilot
studies and the two full-scale systems in the petroleum refinery industry to draw general
conclusions, at least within this industrial category. For example, a distribution of long-
term average COD concentrations in effluents from carbon adsorbers treating refinery
wastewaters is presented in Figure 22 (33). These levels are relatively consistent with
those residuals reported in the petrochemical industry (28). A similar presentation of oil
and grease effluent levels from a full-scale refinery carbon system is shown in Figure 23.
Carbon adsorption capacities in terms of Ibs of COD removed/lb of carbon exhausted have
ranges from 0.2 to 0.4 in the petrochemical industry and from less than 0.1 to 0.55 in the
petroleum refining industry. These are lower than reported carbon capacities for municipal
wastewaters, as shown in Figure 24, emphasizing the inaccuracies which can occur by
extrapolating results from the treatment of one wastewater and using them as the basis for
predicting another.
SYNOPSIS
A review of the current state of the art of activated carbon treatment has been pre-
sented. Basic concepts of activated carbon treatment have been included, as well as
pertinent municipal and industrial case histories with which the author is familiar. It is
recognized that new truths pertaining to this subject become known on a continuing basis.
However, in evaluating process concepts, developing design bases, predicting effluent
quality, and finalizing management decisions in terms of constructing control systems with
attendant capital commitments, one must base these judgments on the current state of the
art. It is toward the objective of defining the art of activated carbon treatment that this
information is presented.
In the pursuit of this definition, certain apparent verities emerge. Some of the more
significant include the following:
1. Adsorption theory is rigorous for single solutes, but becomes less definitive when
applied to wastewaters containing multiple components with varying molecular
weights and chemical characteristics. A good example is the poorly-defined
and erratic breakthrough geometry observed in the carbon treatment of complex
industrial wastewaters.
2. Many classes of organic compounds are not amenable to carbon adsorption -
particularly oxygenated organics - and show up as residual BOD, COD, or TOC
in carbon column effluents. This limits the overall process efficiency of pure
physical/chemical treatment systems. As many of these residual compounds are
biodegradable, activated carbon as a polishing process is generally capable of
producing a better quality of effluent than is the strict physical/chemical appli-
cation.
3. Anaerobic conditions which prevail in carbon reactors have caused difficulty
through the proliferation of biological growths on the carbon media. Sulfide
generation has been particularly troublesome. Although biological activity in
the column has the potential of increasing carbon capacities and is hypothesized
to offer regeneration possibilities, full-scale experience has indicated that
uncontrolled biological growth in carbon reactors has more negative features than
17
-------
positive.
4. Ozonation of wastewaters prior to carbon adsorption has been demonstrated on a
pilot scale to enhance adsorption in the columns. This is attributed to the ozone
products being more sorbable than the reactants. Other possible attractions of
this application are controlling biological densities on the carbon media through
partial disinfection and maintaining an aerobic environment in the reactor column.
5. Carbon capacities used as design criteria generally have been overstated and have
not been realized in full-scale experience. Breakthrough concentrations which
force premature regeneration cycles in single-operated columns, and desorption
phenomena which cause low capacities in series-operated polishing columns, are
partially responsible for this.
6. Design criteria for carbon adsorption systems should be sensitive to:
(a) wastewater constituents and their classification in terms of adsorbability, and
the effluent residual potential; and
(b) effect of the selected carbon media with respect to capacity, resistance to
abrasion, regeneration impacts, and hydraulic characteristics; and
(c) the biological growth potential with the associated effects on carbon capacity,
backwash requirements, and hydraulic characteristics of the flow through
the column; and
(d) the necessary pretreatment requirements for control of suspended solids
(organic and inorganic), oil and grease, dissolved oxygen, biological
population, and other constituents which affect carbon adsorber performance.
REFERENCES
(1) Weber, Walter J., Physicochemical Processes for Water Quality Control, John Wiley
and Sons, Inc., New York, (1972).
(2) Bell, Bruce A. and Molof, Alan H., "A New Model of Granular Activated Carbon
Adsorption Kinetics," Water Research, Vol. 9, Pergamon Press, London, (1975).
(3) Ford, D. L., "The Applicability of Carbon Adsorption in the Treatment of Petro-
chemical Wastewaters," Proceedings, The Application of New Concepts of Physical-
Chemical Wastewater Treatment, Sponsored by the International Association of Water
Pollution Research and the American Institute of Chemical Engineers, Vanderbilt
University, Nashville, (September, 1972).
(4) Giusti, D. M., Conway, R. A., and Lawson, C. T., "Activated Carbon Adsorption
of Petrochemicals," Journal, Water Pollution Control Federation, (May, 1974).
(5) Hassler, John W., Purification With Activated Carbon, Chemical Publishing Co.,
Inc., New York, (1974).
(6) Environmental Protection Agency, Process Design Manual for Carbon Adsorption,
Technology Transfer Manual, (October, 1971).
(7) American Water Works Association, "AWWA Standard for Granular Activated Carbon,"
Journal, AWWA, (November, 1974).
(8) Smisek, M., and Cerny, S., Active Carbon, Elsevier Publishing Co., New York, (1970).
(9) De John, P. B., "Carbon from Lignite or Coal, Which is Better?", Chemical Engineering,
(April, 1975).
(10) Loven, A. W., "Perspectives on Carbon Regeneration," Chemical Engineering Progress,
(November, 1973).
18
-------
(11) Federal Water Quality Administration, "Development of a Fluidized Bed Technique
of the Regeneration of Powdered Activated Carbon," Water Pollution Control Res.
Series ORD-17020 FBD, (March, 1970).
(12) Environmental Protection Agency, "Powdered Activated Carbon Treatment of
Combined and Municipal Sewage," Water Pollution Control Res. Series 11020 DSO,
(November, 1972).
(13) Berg, E. L., etal, Chemical Engineering Proceedings 67, No. 107, 154, (1970).
(14) Fuchs, John L., Private Communication to Engineering-Science, Inc., (June, 1975).
(15) Culp, R. L., and Roderick, R. E., "The Lake Tahoe Water Reclamation Plant,"
Journal, Water Pollution Control Federation, (February, 1966).
(16) Slechta, A. F., and Culp, G. L., "Water Reclamation Studies at the South Tahoe
Public Utility District," Journal, Water Pollution Control Federation, (May, 1967).
(17) Ford, D. L., personal communication with the City of Colorado Springs, Colorado,
(July, 1975).
(18) Batelle - Northwest and Zurn Environmental Engineers, "Westerly Advanced Waste-
water Treatment Facility, Process Design Report and Appendices," Cleveland, Ohio,
(1971).
(19) Guirguis, W. A., et al, "Ozonation Studies at the Westerly Wastewater Treatment
Center," Second International Ozone Symposium, Montreal (May, 1975).
(20) Evans, F. L., Ozone In Water and Wastewater Treatment, Ann Arbor Science, Inc.,
(1972). ;
(21) Snoeyink, V. L., Weber, W. J. and Mark, H. B., "Sorption of Phenol and Nitro-
phenol by Active Carbon," Environmental Science and Technology, (October, 1969).
(22) McDuff, D. P., and Chiang, W. J., "Physical Chemical Design for Garland, Texas,"
The Applications of New Concepts of Physical - Chemical Wastewater Treatment,
Sponsored by the International Association of Water Pollution Research and the
American Institute of Chemical Engineers, Vanderbilt University, Nashville,
(September, 1972).
(23) Engineering-Science, Inc., verbal communication with the Metropolitan Sewer
Board, Rosemount, Minnesota, (1975).
(24) Directo, L. S., and Chen, C. L., "Pilot Plant Study of Physical Chemical Treat-
ment," 47th Annual Water Pollution Control Federation Conference, Denver,
Colorado, (October, 1974).
(25) Engineering-Science, Inc. and Cleveland Regional Sewer District, Report on
Evaluation of Continuing Westerly Pilot-Plant Studies, prepared for CRSD,(June,
1975).
(26) Ford, D. L., and Buercklin, M. A., "The Interrelationship of Biological-Carbon
Adsorption Systems for the Treatment of Refinery and Petrochemical Wastewaters,"
6th International Association of Water Pollution Research Conference, Jerusalem,
(June, 1972).
(27) Short, T. E., and Myers, L. A., "Pilot Plant Activated Carbon Treatment of
Petroleum Refinery Wastewaters," Robert S. Kerr Environmental Research Laboratory,
Ada, Oklahoma, (1975).
(28) Lawson, C. T., "Activated Carbon Adsorption for Tertiary Treatment of Activated
Sludge Effluents from Organic Chemicals and Plastics Manufacturing Plants - Appli-
cation Studies and Concepts," Research and Development Department, Union Carbide
Corporation, South Charleston, West Virginia, (August, 1975).
19
-------
(29) Environmental Protection Agency, Refinery Effluent Water Treatment Plant Using_
Activated Carbon, Envirpnmental Protection Technology Series, EPA 660/2-75-020,
(June, 1975).
(30) Shumaker, T. P., "Granular Carbon Process Removes 99 - 99.2% Phenols," Chemical
Processing, (May, 1973).
(31) McCrodden, B. A., "Treatment of Refinery Wastewater Using Filtration and Carbon
Adsorption," Advanced Petroleum Refinery Short Course, Principles and Practice in
Refinery Wastewater Treatment, University of Tulsa, (June, 1973).
(32) McCrodden, B. A., "Treatment of Refinery Wastewater Using Filtration and Carbon
Adsorption," presented at Technology Transfer Conference, Activated Carbon in Water
Pollution Control, Sponsored by the Pollution Control Association of Ontario, and the
Canadian Society for Chemical Engineering, (October, 1974).
(33) Engineering-Science, Inc., Report to the National Commission on Water Quality,
Petroleum Refinery Industry - Technology and Cost of Wastewater Control, (June, 1975).
DISCUSSION
Ed Sebesta: Were these pilot plant data taken with regenerated carbon and equilibrated
carbon ?
Davis Ford: No. It was virgin carbon. That's a good point. Many of the pilot plants
studies that we have run have been with virgin carbon. It is not that we don't recognize
that it would be more applicable with regenerated carbon, it is just sometimes that it is
hard to get that much regenerated carbon from the vendor and that has been our problem.
One would think it would be more practical to run all these pilot studies with regenerated
carbon.
Bob Huddleston: Do you have any data that indicates whether or not the effects of ozone
were anything other than simply killing the microbes?
/
Davis Ford: Yes. They ran some GC work to identify the nature of the compounds after
organization and I don't have the results in this paper. Just in talking with these people,
it's what you really have done is gone from the ketones, aldehydes to carboxylic acids,
there has been a shift and the carboxylic acids being more ameniable to carbon absorption,
so its a combination of the biological growth but also the transformation of these oxygenated
organics to carboxylic acids.
M_. K. Hutton: Discuss if you will the cost of ozone. Do you have any figures on this?
I expect it is quite expensive.
Davis Ford: I left cost completely out of the paper because that is almost another paper
itself; but considering the concentration of ozone that we think we have to have, it is
probably about 4 or 5 milligrams per liter is within the cost effective parameters that we
have set to shis point. So we think it is cost effective.
Anonymous: Do you have some comments on phenol removal ?
20
-------
Davis Ford: No. Phenol is highly sorbable, most of the data we have seen on phenol
removal has been excellent'and the only exceptions to this was when we just had so many
physical problems, the biological growths and oil and grease, that the phenol removal
has gone down. I think that is shown in one table. If you don't have this constraint,
phenol removal should be good, carbon is a good phenol remover.
Morris Wiley: The little village where I live has a new sewer system and biological
treatment plant. It is an extended aeration without any primary treatment, then a second
stage nitrification and finally a sand filter. They are able to do less than one part per
million BOD and less than one part per million ammonia in the effluent. Now I should
caution that it is only operating at one quarter of rated capacity. Thus, it appears that
if one has to go to these very low organic and nitrogen contents in the effluent, that one
can do it by a conventional biological system simply by spending a lot of money. I
estimate that this is about $1,000 per family per year, total annualized cost for this
system including the sewer connections which were put in. Have you done any economic
studies to try to figure out which would really be preferable, an advanced biological
reactor system versus an activated carbon where you need a very low concentration of
pollutants in the effluent?
Davis Ford: Let me speak in general terms here. First, you have a good analytical
chemist at that plant.
Morris Wiley: Well, I have some reservations about the precision of the tests. The plant
operator is a Texaco employee running the plant in his spare time with his wife and his
son, but it is a good quality effluent and the county is accepting their test reports.
Davis Ford: Let me say, from the cost studies that we have run, we have seen biological
treatment is more cost effective than carbon, I don't think there is any question about
that. When you get down to the fact that a biological system can't make a permit and
you have to go to carbon, you know some people are not sympathetic with cost effectives,
but if you can get to those levels biologically, it just intuitively has to be more cost
effective than carbon. Carbon has gone up, it is 50-60$ now for bituminous. When we
started running these studies three or four years ago, it was 32$ or 30$ and just,.because
of these lower carbon capacities I have stated, biological treatment has to be more cost
effective. If we can make it with biological treatment, that is certainly the theology that
I would like to see.
Morris Wiley: Then it looks as though our research should be concentrated on enhancing
the effectiveness this biological treatment as Jim Grutsch pointed out earlier?
Davis Ford: 1 do not disagree with that. I also think we have got to be realistic, there
are certain limitations to biological treatment. When you look at BAT, if and when it
comes, it is going to take some real engineering to make those numbers biologically.
Neale Fugate: Do you have any data on the possible adsorption of heavy metals, or
conversely desorption of heavy metals on carbon?
21
-------
Davis Ford: Yes, sir. 1 have it, but it is not presented in the paper. I have just seen some
data on that. There is slight reductions of some of the metallic ions probably as metallic
hydroxides. I think in some literature absorption of some metallic ions has been reported.
It has not been significant, something like 10-15% reduction of zinc, mercury, cadmium,
some of these substances but percent removal is quite low, but some removal has been noted.
I am not sure of the mechanism as a physical removal of zinc or sorption of zinc iron, but
one of those two mechanisms.
Robert L. Wortman: It is my understanding that ozone is a very effective phenol remover,
it removes the phenol quite effectively, it is probably even better than activated carbon
(through oxidation). It would seem to me that may be secondary conventional treatment
followed by ozonation may be as effective as secondary treatment, foil owed by carbonation.
My question is, have you had any experience on this in combination with carbon filtration
and I am interested in treating each one individually, that is considering ozonization as a
treatment method?
Davis Ford: Let me answer that two ways. First of all, I think I mentioned in the talk
that we tried preozonation after carbon with bad results. However, there is a treatment
plant, financed in part by EPA which goes on line in about two months and that is at Estes
Park, Colorado, which is exactly the system you described. Other than that plant, I am
not personally familiar with a real good biological treatment system including nitrification
followed by ozone but that system is designed that way. That is the upper sanitation district
in Estes Park, Colorado, the plant is completed and is probably going through shake down
right now and we will get some good full scale information from that facility in the next
few months.
Anonymous: I think there is also some work being done in Pomona, I don't know exactly
what it was, but on a different combination of ozonation and various schemes; but to further
explore the Estes Park, if you say you had a multimedia filter and then go in ozonation, this
may be quite effective to work with.
Fred Gowdy: Davis, I wonder if you would give us some of the reasons you believe why the
carbon capacities that we have realized in some of our full scale plants, haven't been any-
where close to those predicted by various pilot studies?
Davis Ford: I am not really sure that I know the answer to that, I can just hypothesize.
First of all, I think Union Carbide has presented carbon capacities in lead columns as far
as polishing columns were concerned, so that would indicate the operation has a lot to do
with carbon capacity, in other words, how well you can operate the carbon utilization in
a series mode of operation to get the operable capacity of the overall system. You can
pretty well do that in a pilot scale. I am not sure you can do it in a full scale quite so
easily. It is easier to operate that pilot scale, I assure you, than a full-scale plant. So
I think part of the answer lies in the operation. Secondly, it is possible that you had more
physical problems in terms of biological growth, influent constituents in the full scale that
you don't perceive in a pilot scale. You might control the pilot scale a little better and
for this reason are able to realize a better carbon capacity. Those are the two main reasons
I think, I just can't prove it.
22
-------
Ed Sager: I wondered if you would like to elaborate on the back washing features of these
columns. Why is that required if you have adequate pretreatment, are you removing
suspended solids build-up?
Davis Ford; I think the best case history to point to there is the BP, they have got some
data on their filters, good filters which work quite well; but the effluent suspended solids
are probably in the 10-15-20 milligram per liter range and these influent suspended solids
to the carbon column are going to accumulate after a certain period of time, combined
with the biological proliferation in the columns. I think you always have to have the flexi-
bility operational flexibility of back washing columns as well as the filters. I think that
has been the experience to date. You may not have to do it, but it is an operational
flexibility that you should have. It seems like even if you go in with zero suspended
solids, some point in time you are going to have to back wash the filters or the carbon
columns.
Pat DeJohn: One of the things that one of the gentlemen asked about the loadings, most
of the loadings that are reported are reported with virgin carbon and the properties of
activated carbon change quite substantially when you regenerate carbon and I think that is
part of the BP problem, they are not able to maintain their phenol standards because of
the loss in capacity, adsorptive capacity and consequently, their system has been sub-
stantially under design. The other thing that you mentioned about why you should be back
washing, that BP plant is a pulse bed unit and there are fines that are created due to the
regeneration of the carbon and they have no way to remove those fines; consequently they
recycle them back into the absorber and they pulse that unit a lot of times because of
pressure drop problems and that is another thing about that particular plant.
Davis Ford: I agree with that. In the past, Pat, when we have had to use virgin carbon,
we have just,applied a factor on the capacity, reduced capacity, just an empirical factor
and that is a less favorable way of doing it than using regenerated carbon; but we just
haven't had that luxury a lot of times.
J. J. Chavez: I was just wondering if you could elaborate on the regeneration phase
From the economic side by which a refinery could put in one regulation facility and the
options open to him if he can't afford one?
Davis Ford: There again I didn't really get into cost. I believe the reported regeneration
cost now is something like 15-20$ per pound. Some manufacturers, I believe, are offering
a regeneration service. If you have a small system, it doesn't make sense because of
economy in scale to put in your own regeneration facility, to have a contractor regenerate
it, if that's possible, but I think if you go into regeneration which in most areas you would
have to for most size of plants, you have to count on about a 15-20$ per pound regeneration
cost.
23
-------
BIOGRAPHY
Davis L. Ford holds a B.S. in Civil Engineering
from Texas A & M University and M.S. and Ph.D.
degrees in Environmental Health Engineering from the
University of Texas at Austin. Dr. Ford is currently
Senior Vice President and Member of the Board of
Directors of Engineering Science, Inc., in Austin,
Texas. Dr. Ford has written 4 books, 20 reports, 60
publications in the field of environmental engineering
and has consulted for over 50 industries, the United
Nations (WHO and PAHO), the EPA and various state
and municipal agencies.
TABLE 1 "COMPARATIVE ANALYSIS OF BATCH ISOTHERM DATA -
REFINERY & PETROCHEMICAL WASTEWATERS"
K n
OIL SEPARATOR (PRIMARY) EFFLUENT
Refinery- Petrochemical Complex No. 1 0.0290 0.77
Refinery - Petrochemical Complex No. 2 0.0036 0.80
Refinery No. 3 0.0140 0.36
SECONDARY (ACTIVATED SLUDGE) EFFLUENT
Refinery - Petrochemical Complex No. 1 0.0062 0.60
Refinery No. 3 0.0043 1.00
Refinery Secondary Effluent 0.0051 0.96
Refinery Secondary Effluent 0.0038 1.08
Refinery Secondary Effluent 0.0020 0.69
SINGLE ADSORBATE
Phenol 0.1110 5.80
Dichlorethane (pH 4) 0.0045 1.82
(pH6) 0.0038 1.67
(pH 10) 0.0041 1.49
24
-------
TAffiE 2
AMENABILITY OF TYPICAL ORGANIC COMPOUNDS
TO ACTIVATED CARBON ADSORPTION [4]
Compound
Alcohols
Methanol
Ethanol
Propanol
Butanol
n-Amyl alcohol
n-Hexanol
Isopropanol
Allyl alcohol
Isobutanol
t Butanol
2 Ethyl butanol
2 Ethyl neianot
Aldehydes
Formaldehyde
Acetaldehyde
Propionalriehyde
Butyr aldehyde
Acrolem
Crotonaldehyde
Benialdehyde
Paraldehyde
Amines
Di N Propytamine
Butylamme
Di N Butylamme
AHylamme
Ethylenediamme
OiethySenelnamtne
Mcnethanolamme
Oiethanolamme
Tntihanclamine
Monoisopropanolamine
Diisopropanolamine
Pyridines & Morpholines
Pyndme
2-Melnyl 5 Ethyl pyridme
N Methyl morphohne
N Ethyl morpholme
Aroma tics
Benzene
Toluene
Ethyl benzene
Phenol
Hydroqumone
Aniline
Styrene
Nitrobenzene
Esters
Methyl acetate
Ethyl acetate
propyl aceiate
Buiyl acetate
Ptirnaryamyi acetate
Concentration (mjfl)
Molecular
Weigh!
32.0
46.1
60.1
74.1
88.2
1022
601
58.1
74.1
74.1
102.2
130.2
300
44.1
581
72.1
56.1
701
106.1
132.2
101.2
73.1
129.3
57.1
60.1
103.2
61.1
105.1
1491
751
133.2
791
121.2
101.2
115.2
78.1
92.1
106.2
94
110.1
931
1042
1231
74 1
881
1021
1162
1302
ra|ucvuai
Solubility
<*>
OO
OO
oo
7.7
1.7
0.58
OO
oo
8.5
oo
0.43
0.07
00
00
22
7.1
20.6
15.5
0.33
10.5
OO
oo
oo
oo
OO
OO
oo
95.4
OO
oo
87
OO
si. sol
00
OO
0.07
0047
002
6.7
6.0
3.4
0.03
0.19
31.9
87
2
068
02
Initial (C0)
1,000
1,000
1.000
1.000
1.000
1.000
1.000
1.010
1.000
1,000
1,000
700
1,000
1,000
1.000
1.000
1.000
1.000
1.000
1,000
1.000
1.000
1,000
1,000
1.000
1.000
1.012
996
1.000
1.000
1.000
1,000
1,000
1.000
1.000
416
317
115
1.000
1,000
1.000
180
1.023
1.030
1.000
1.000
1.000
985
Final (Cf)
964
901
811
466
282
45
874
789
581
705
145
10
908
881
723
472
694
544
60
261
198
480
130
686
893
706
939
722
670
800
543
527
107
575
467
21
66
18
194
167
251
18
44
760
495
248
154
119
Absorbability
1 compound/
I carbon
0.007
0020
0.038
0107
0155
0191
0025
0.024
0.084
0.059
0.170
0138
0.018
0022
0057
0.106
0061
0.092
0.188
0.148
0.174
0103
0.174
0063
0021
0062
0.015
0.057
0067
0.040
0.091
0.095
0 179
0.085
0.107
0080
0050
0.019
0161
0.167
0.150
0028
0196
0054
0100
0149
0169
0.175
Percent
Reduction
3.6
10.0
18.9
53.4
71.8
95.5
12.6
21 9 .
41.9
29.5
85.5
98.5
9.2
11.9
27.7
52.8
306
45.6
94.0
73.9
80.2
52,0
87.0
314
10.7
29.4
7.2
275
33.0
20.0
45.7
47.3
89.3
425
53.3
95.0
792
84.3
806
83.3
74.9
88.8
956
26.2
505
752
846
880
Compound
Esters
I sop ropy I acetate
Isobutyl acetate
Vinyl acetate
Ethylene ilycol monoethyl ethei acetate
Ethyl acrylale
Butyl acrylate
Ethers
Isopropyl ether
Butyl ether
Dichloroisopfopyl ether
Glycols & Glycol Ethers
Ethyltjne glycol
Diethylene glycol
Trielhylene glycol
Tetraethylene glycol
Propylene glycol
Dipropylene glycol
Hexyfene glycol
Ethylene glycot monomethyl ether
Ethylene glycol monoethyl ether
Ethylene glycol monobuty! ether
Ethylene glycol monohexyl ether
Diethylene glycol monoethyl ether
Diethylene glycot monobutyl ether
Ethoxytnglycol
Halogenated
Ethylene dichloride
Prupylenedichloride
Ketones
Acetone
Methylethyt ketone
Methyl propyl ketone
Methyl butyl ketone
Methyl isobutyl ketone
Methyl isoamyl ketone
Diisobutyl ketone
Cydoheianone
Acetophenone
Isophorone
Organic Acids
Formic acid
Acetic acid
Propionic acid
Butyric acid
Valeric acid
Caprotc acid
Acrylic acid
BeniQIC acid
Oxides
PfOpylene onde
Styrene onde
Dosage 5 g Carbon C/l solution.
Continuation (m|/l)
lolecullr
Weight
102.1
116.2
86.1
132.2
100.1
128.2
102.2
130.2
171.1
62.1
106.1
150.2
194.2
76.1
1342
118.2
76.1
90.1
118.2
1462
134.2
162.2
1782
990
113.0
58.1
72.1
86.1
100.2
100.2
1142
1422
982
120.1
138.2
46.0
601
74.1
881
1021
116.2
721
122.1
581
120.2
M|UlWlf»
Solubility
(*)
2.9 .
0.63'
2.8
22.9
2.0
0.2
1.2
0.03
0.17
00
oo
00
00
oo
00
oo
oo
oo
00
0.99
oo
oo
oo
0.81
0.30
oo
26.8
43
v. si. sol.
1.9
0.54
0.05
2.5
0.55
1.2
OO
oo
oo
oo
24
1 1
OO
029
405
0.3
Initiil (Co)
1.000
1.000
1.000
1.000
1.015
1.000
1,023
197
1.008
1,000
1.000
1.000
1.000
1.000
1.000
1.000
1.024
1,022
1.000
975
1,010
1,000
1.000
1.000
1.000
1.000
1.000
1.000
988
1,000
986
300
1.000 "
1,000
1,000
1.000
1,000
1,000
1.000
1,000
1.000
1.000
1.000
1.000
1.000
Finil (0)
31)
180
357
342
226
43
203
nil
nil
932
738
477
419
884
835
386
886
705
441
126
570
173
303
189
71
782
532
305
191
152
146
nil
332
28
34
765
760
674
405
203
30
355
89
739
47
Adsorbtbilitr
I compound/
I carbon
0.137
0164
0.129
0.132
0.157
0.193
0.162
0.039
0.200
0.0136
0.053
0105
0116
0024
0.033
0122
0028
0.063
0.112
0.170
0.087
0.166
0.139
0.163
0.183
0043
0.094
0139
0.159
0.169
0.169
0.060
0.134
0194
0.193
0047
0048
0065
0 119
0 159
0194
0129
0183
0052
0190
Percent
Seduction
68.1
82.0
64.3
65.8
77.7
95.9
80.0
100.0
100.0
6.8
26.2
52.3
58.1
11.6
165
614
13.5
310
559
87.1
436
82.7
697
81 1
929
21 8
46.8
69.5
80.7
84.8
852
100.0
66.8
972
96.6
235
240
326
59.5
79.7
970
645
261
953
-------
TABLE 3 "RELATIVE AMENABILITY TO CARBON ADSORPTION OF TYPICAL
PETROCHEMICAL WASTEWATER CONSTITUENTS"
Compound % Adsorption Compound % Adsorption
Ethanol 10 Vinyl acetate 64
Isopropanol 13 Ethyl acrylate 78
Acetaldehyde 12 Ethylene glycol 7
Butyraldehyde 53 Propylene glycol 12
Di-N-propylamine 80 Propylene oxide 26
Monoethanolamine 7 Acetone "
Pyridine 47 Methyl ethyl ketone 47
2-Methyl 5-ethyl pyridine 89 Methyl isobutyl ketone 85
Benzene 95 Acetic acid 24
Phenol 81 Proprionic acid 33
Nitrobenzene 96 Benzoic acid 91
Ethyl acetate 50
Initial Compound Concentration = 1,000 mg/l Powdered Carbon Dosage = 5,000 mg/l
TABLE 4 "INFLUENCE OF MOLECULAR STRUCTURE AND OTHER FACTORS OF ADSORBABILITY"
1. An increasing solubility of the solute in the liquid carrier decreases its adsorbability.
2. Branched chains are usually more adsorbable than straight chains. An increasing length of the chain decreases solubility.
3. Substituent groups affect adsorbability:
Substituent Group Nature of Influence
Hydroxyl Generally reduces adsorbability; extent of decrease depends on structure of
host molecule.
Amino Effect similar to that of hydroxyl but somewhat greater. Many amino acids
are not adsorbed to any appreciable extent.
Carbonyl Effect varies according to host molecule; glyoxylic are more adsorbable
than acetic but similar increase does not occur when introduced into
higher fatty acids.
' Double Bonds Variable effect as with carbonyl.
Halogens Variable effect.
Sulfonic Usually decreases adsorbability.
Nitro Often increases adsorbability.
Aromatic Rings Greatly increases adsorbability.
4. Generally, strong ionized solutions are not as adsorbable as weakly ionized ones; i.e., undissociated molecules
are in general preferentially adsorbed.
5. The amount of hydrolytic adsorption depends on the ability of the hydrolysis to form an adsorbable acid or base.
6. Unless the screening action of the carbon pores intervene, large molecules are more sorbable than small molecules
of similar chemical nature. This is attributed to more solute carbon chemical bonds being formed, making
desorption more difficult.
7. Molecules with low polarity are more sorbable than highly polar ones.
26
-------
TABLE 5
PROPERTIES OF SEVERAL COMMERCIALLY AVAILABLE CARBONS* '6|
PHYSICAL PROPERTIES
Surface area, m2/gm (BET)
Apparent density, gm/cc
Density, backwashed and drained, Ib/cu. ft.
Real density, gm/cc
Particle density, gm/cc
Effective size, mm
Uniformity coefficient
Pore volume, cc/gm
Mean particle diameter, mm
SPECIFICATIONS
Sieve size (U.S. std. series)
Larger than No. 8 (max. %)
Larger than No. 12 (max. %)
Smaller than No. 30 (max. %)
Smaller than No. 40 (max. %•)
Iodine No.
Abrasion No., minimum
Ash (%)
Moisture as packed (max. %)
ICI
AMERICA
HYDRODARCO
3000
(LIGNITE)
600-650
0.43
22
2.0
1.4-
0.8-
1.7
0.95
1.6
8
5
650
1.5
0.9
CALGON
FILTRASORB
300
(8x30)
(BITUMINOUS)
950-1050
0.48
26
2.1
1.3 - 1.4
0.8 • 0.9
1.9 or less
0.85
1.5- 1.7
900
70
8
2
WESTVACO
NUCHAR
WV-L
(8x30)
(BITUMINOUS)
1000
0.48
26
2.1
1.4
0.85 - 1.05
1.8 or less
0.85
1.5- 1.7
5
950
70
7.5
2
WITCO
517
(12x30)
(BITUMINOUS)
1050
0.48
30
2.1
0.92
0.89
1.44
O.bO
1.2
5
5
1000
85
0.5
1
* Other sizes of carbon are available on request from the manufacturers.
** No available data from the manufacturer.
— Not applicable to this size carbon.
TYPICAL PROPERTIES OF 8 X 30-MESH CARBONS19'
Total surface area,
Iodine number, min
Bulk density, Ib/ft3 backwashed and drained
Particle density wetted in water, g/'cm3
Pore volume, cmVg
Effective size, mm
Uniformity coefficient
Mean-particle dia., mm
Pittsburgh abrasion number
Moisture as packed, max.
Molasses RE (relative efficiency)
Ash
Mean-pore radius
LIGNITE
CARBON
600-650
500
22
1.3-1.4
1.0
0.75-0.90
1.9 or less
1.5
50-60
9%
100-120
12-18%
33 A
BITUMINOUS
COAL CARBON
950-1,050
950
26
1.3-1.4
0.85
0.8-0.9
1.9 or less
1-6
70-80
2%
40-60
5-8%
14 A
27
-------
TABLE 6 "TYPICAL PROPERTIES OF POWDERED ACTIVATED CARBON (PETROLEUM BASE)1
Surface Area (BET m /gm)
Iodine No.
Methylene Blue Adsorption (mg/gm)
Phenol No.
Total Organic Carbon Index (TOCI)
Pore Distribution (Radius Angstrom)
Average Pore Size (Radius Angstrom)
Cumulative Pore Volume (cc/gm)
Bulk Density (gm/cc)
Particle Size Passes
Ash (wt%)
Water Solubles
pH of Carbon
100 mesh (wt%)
200 mesh (wt%)
325 mesh (wt%)
:,300-
!,700-
400-
10-
400-
15-
20-
0.1-
0.27-
97-
93-
85-
8-
2,600
3,300
600
12
•800
•60
•30
•0.4
0.32
100
98
95
5
0
9
TABLE 7 "STAGES OF THERMAL REGENERATION"
Stage
Drying
Thermal Desorption
Pyrolysis and Carbonization
Gasification
Approximate Temperature ( F)
Ambient to 212
212 to 500
400 to 1,200
1,200 to 1,900
Processes
Water Evaporation
Physical desorption of volatile adsorbed organics
Pyrolysis of nonvolatile organics and
carbonization of the pyrolysis residue
Gasification of pyrolytic residue through
controlled chemical reaction with water vapor,
vapor, carbon dioxide, or oxygen
TABLE 8 "WATER QUALITY AT VARIOUS POINTS IN PROCESS" (16)
Quality Parameter
BOD (mg/l)
COD (mg/l)
Total organic carbon (mg/l)
ABS (mg/l)
PO, (mg/l as PO4)
Color (units)
Turbidity (units)
Nitrogen organic N (mg/l as N)
Ammonia N (mg/l as N)
NO,, and NO
Unc
2 (mg/l as N)
lorinated: Coliforms (MPN/100 ml)
Fecal coliforms (MPN/100 ml)
Virus
Chlorinated: Coliforms (MPN/100 ml)
Fecal coliforms (MPN/100 ml)
Raw
Waste water
200-400
400-600
-
2.0-4.0*
-
-
-
10-15
25-35
0
-
-
-
-
Secondary
Effluent
20-100
80-160
-
0.4-2.9
25-30
-
30-70
4-6
25-32
0
2,400,000
150,000
-
-
Separation-
Bed Effluent
<}
20-60
8-18
0.4-2.9
0.1-1
10-30
<0.5-3
2-4
25-32
0
9,300
930
Negative
8.6
Carbon-
Column Effluent
<1
1-25
1-6
<0.01-0.5
0.1-1
<5
<0.5-1
1-2
25-32
0
11,000
930
Negative
<2 1
28
-------
TABLE 9 "WATER QUALITY AT VARIOUS STAGES OF TREATMENT AT SOUTH LAKE TAHOE"
Effluent
Quality
Parameter
BOD (mg/l)
COD (mg/l)
SS (mg/l)
Turbidity (JTU)
MBAS (mg/l)
Phosphorus (mg/l)
Col i form
(MPN/lOOml)
Raw
Wastewater
140
280
230
250
7
12 ,
50 x 10
Secondary
30
70
26
15
2.0
6 ,
2.5x 10
Chemical
Clarifier
10
10
0.7
Filter
3
25
0
0.3
0.5
0.1
50
Carbon
1
10
0
0.3
0.1
0.1
50
Chlorinated
Final
0.7
10
0
0.3
0.1
0.
<2
TABLE 10 "CARBON EFFICIENCY PER REGENERATION PERIOD AT SOUTH LAKE TAHOE"
NOVEMBER 1968 THROUGH JANUARY 1971
Apparent Density (gm/ml)
2
Percent Ash
Chemical Oxygen Demand
Parameter
Carbon Dosage ^
(Ib regenerated/million gallons treated)
Iodine Number Spent Carbon
Regenerated Carbon
Spent Carbon
Regenerated Carbon
Spent Carbon
Regenerated Carbon
Percent Removal
Ib COD applied
Ib COD applied/MG
Ib COD removed/MG
Ib COD applied/lb 1
carbon regenerated
Ib COD removed/lb
carbon regenerated
Methylene Blue Active Substances (Methylene Blue Active
Substances (MBAS)) Percent removal
Ib MBAS applied
Ib MBAS applied/MG
Ib MBAS remoyed/MC
Ib MBAS applied/lb
carbon regenerated
Ib MBAS removed/lb
carbon regenerated
1 3 3
-Based on ft of carbon fed to furnace at 30 Ib/ft
November 1968 through November 1970
Average
207
583
802
0.571
0.487
6.4
6.8
49.9
28,250
162
81
0.78
0.39
77
995
5.7
4.4
0.027
0.021
Maximum
418
633
852
0.618
0.491
7.0
7.2
63.3
54,970
254
149
1.56
0.71
93
1,675
10.7
8.2
0.045
0.039
Minimum
111
497
743
0.544
0.478
5.8
5,8
30.1
15,680
105
32
0.52
0.16
58
457
2.6
1.6
0.012
0.007
29
-------
TABLE 11 "COLORADO SPRINGS TERTIARY PLANT DESIGN DATA"
Solid Contact Clorifier 48 ft diameter, 11 ft 9 in sidewall, 2 ft cone depth
1,809 sq ft surface area 168,465 gallon capacity
2 hr detention time and 0.76 gpm/sq ft rise rate - @ 2 MGD flow
Spent Lime Tank 14 ft diameter, 10 ft sidewall depth, 8 ft cone
1,948 cu ft 14,610 gallon capacity
New Lime Holding Tank 14 ft diameter, 10 ft sidewall depth, 8 ft cone
1,948 cu ft 14,610 gallon capacity
Recorbonization Tank 14 ft diameter, 14 ft sidewal! depth, plus 2 ft freeboard - 154 sq ft _
2,156 cu ft 16,160 gallon capacity - 12 minutes detention time @ 2 MGD flow
i
Activated Carbon Adsorbers 20 ft diameter, 14 ft sidewall depth, 10 ft of carbon media
314 sq ft surface area, 4,396 cu ft 32,970 gallon capacity _
total tower detention time is 24 minutes - 4.5 gpm/sq ft (@ 2 MGD flow)
carbon bed detention time is 17 minutes - 4.5 gpm/sq ft (@ 2 MGD flow)
3,140 cu ft of carbon a 30 Ib/cu ft - 94,200 Ibs carbon
Dual Media Sand Filters 12 ft diameter, 11 ft sidewall depth, 1 ft 6 in top cone
113 sq ft surface area
1,243 cu ft capacity 9,323 gallon capacity
3 ft of e.s. 1.5 mm sand 5 ft of e.s. 2.8 mm anthrafilt (#2)
TABLE 12 "TERTIARY TREATMENT PLANT DATA SUMMARY (FIRST REGIME)"
PARAMETER R-C INF. R-C EFF. FILTER EFF. LEAD CARBON EFF. POLISH CARBON EFF.
BOD 106 46.9 48.0 44.0 33.0
COD 305 120 114 83 64
TOC 76 37.9 36.8 27.9 23.8
TSS 56 31.0 3.7 2.1 3.7
Turbidity 53 10.1 4.2 3.6 3.2
O.PO (total) 32 2.2 1.7 2.2 1.9
0.POT (soluble) 31 0.16 1.6 2.2 1.8
MBAS 5.32 3.22 - 1.36 0.51
pH 7.35 11.44 7.93 6.91 7J2
Alkalinity (total) 197 282 69 63 69
Hardness ^ 163 ; 210 250 240 253
C (as C ) 52 79 95 84 92
C°C (asaC CO ) 130 197 237 209 231
Color a J 162 44 38 25 13
Su I fates 88 80 439 439 401
Sulfides 0.15 0.14 - 0.26 041
Flow (MGD) 1.93 1.93 1.93 1.70 ]'.7Q
Lime Dose (ppm Ca 0) 349
NOTE: During this period, the reactor was operated with a deep sludge blanket of approximately six feet.
30
-------
TABLE 13 "TERTIARY TREATMENT PLANT DATA SUMMARY (SECOND REGIME)"
PARAMETER
BOD
COD
TSS
Turbidity
0. Phosphate (total)
0. Phosphate (soluble)
MBAS
Fecal Coliform
Fecal Strep
Average Flow (MGD)
Average Lime Dose
Lead Tower
Polish Tower
INFLUENT
(mg/l)*
102
258
62
54
30
26
4.2
5 x 105/K>0 m|
6.5x 10/100 ml
EFFLUENT
(mg/l)*
8.0
15.7
2.2
2,2
1.0
1.0
0.15
225/100 ml
1150/100 ml
1.5
345 mg/l
5 ft of twice-regenerated + 5 ft of virgin carbon
8 ft of twice-regenerated + 2 ft of virgin carbon
*AII units in mg/l unless indicated.
The "INFLUENT" is the secondary effluent going to the solids contact clarifier.
The "EFFLUENT" is the final polish carbon tower effluent.
Influent to tertiary plant is from a trickling filter plant with a flow of
23 to 25 MGD; trickling filter plant capacity is 13 MGD.
REMOVAL"
(percent)
92.2
93.9
96.5
96.3
96,
96,
96,
99.96
99.82
TABLE 14 "COMPARISON OF WASTEWATER STRENGTH BATTELLE AND CRSD TEST PROGRAMS"
Probability (% of Occurrences)
10 50 90
Battelle Series (1970-71) BOD
CRSD Series (1974-75) BOD
Battelle Series (1970-71) COD
CRSD Series (1974-75) COD
Battelle Series (1970-71) BOD/COD
CRSD Series (1974-75) BOD/COD
180 mg/l
115 mg/l
320 mg/l
215 mg/l
0.56
0.53
240 ma/I
170 mg/l
500 mg/l
350 mg/l
0.48
0.48
320 mg/l
250 mg/l
720 mg/l
580 mg/l
0.44
0.43
TABLE 15 "EFFECT OF POST-OZONATION ON EFFLUENT ORGANIC QUALITY"
Parameter
BOD (unfiltered)
BOD (filtered)
COD (unfiltered)
COD (filtered)
TOC
CRSD PILOT-PLANT TEST RESULTS*
Before Ozonation
32 mg/l
25 mg/l
72 mg/l
59 mg/l
31 mg/l
After Ozonation
31 mg/l
24 mg/l
66 mg/l
52 mg/l
29 mg/l
*AII data represent average values over an eleven day period. Ozone dosage ranges from 4-9 mg/l.
31
-------
TABLE 16 "QUALITY OF RAW WASTEWATER"
GARLAND PHYSICAL/CHEMICAL TREATMENT FACILITY
Parameter Results
Total BOD (mg/l) 266
Filtered BOD (mg/l) 236
Total COD (mg/l) 542
Filtered COD (mg/l) 240
Suspended Solids (mg/l) 233
Alkalinity (mg/l as CaCO.) 200
PH 3 7.2-7.7
P04 (mg/l) 15
TABLE 17 "PCT SYSTEM DESIGN DATA" (24)
CHEMICAL TREATMENT SYSTEM
1. Flocculation: Detention Time, minutes 45
Chemical Dosage Alum, mg/l Al 22
polymer, mg/l 0.25
2. Sedimentation: Detention Time, hrs. 1.5
Overflow Rate, gpd/sq ft ' 900
Underflow, % of plant flow 1.25
Underflow solids, % by weight 2
3. Gravity Thickening: Solids Loading, Ib/day-sq ft 12
Underflow solids, % by weight 4
4. Vacuum Filtration: Yield, Ib/hr-sq ft 2
Cake solids 18
5. Sludge Incineration: Solids Loading, Ib/hr-sq ft 2
CARBON TREATMENT SYSTEM
1. Carbon Contacting (8 x 30 mesh carbon) Empty-bed contact time 25
Hydraulic surface Loading, gpm/sq ft 4
Backwash volume, % of plant flow 5
Sodium Nitfate dosage, mg/l N 5.5
Carbon Dosage, Ib/MG 250
Carbon Regeneration loss, % 5
32
-------
TABLE 18 "SUMMARY OF PHYSICAL/CHEMICAL TREATMENT SYSTEM PERFORMANCE" (24)
Average Percent Removal
Parameters
Suspended Solids (mg/l)
Turbidity (JTU)
TCOD (mg/l)
DCOD (mg/l)
BOD (mg/l)
Total Phosphate (mg/l P)
Nitrate (mg/l N)
Color
pH
Notes: 1. Average alum dosage = 25 mg/l Al (275 mg/l alum)
2. Average polymer dosage = 0.3 mg/l Calgon WT-3,000
3. AS Effluent refers to Activated Sludge Plant Effluent (8 MGD, existing facility)
Raw
Sewage
199
321
49.4
11.1
7.7
Clarified
Effluent
28.3
22.9
95.8
48.6
36.2
1.3
0.9
20
6.8
Carbon
Effluent
6.7
6.3
19.3
13.5
7.8
0.9
1.3
7.8
6.8
Chemical
Treatment
85.8
70.2
1.6
88.3
Carbon
Treatment
76.3
72.5
79.9
72.2
78.5
30.8
61.0
Overall
96.6
94.0
72.7
91.9
AS Effluent*
11.6
7.7
39.5
25.7
8.0
33.1
TABLE 19 "EFFECT OF REGENERATION ON THE PCT CARBON CHARACTERISTICS"
Spent Carbon
Regenerated Carbon (Composite Sample)
Carbon
Characteristics
Iodine No. (mg/l)
Apparent Density
(g/cm cu ft)
Molasses No.
Methylene Blue No.
(mg/g)
Ash (%)
Mean Particle Dia.
(mm)
(Composite
Virgin
Carbon
1,040
0.484
222
259
6.4
1.44
1st
Reg.
402
0.629
120
147
10.3
1.46
2nd
Reg.
572
0.585
168
153
8.22
1.58
Sample)
3rd
Reg.
570
0.594
154
153
8.67
1.48
Before Quenching
1st
Reg.
805
0.528
213
223
10.7
1.57
2nd
Reg.
722
0.537
233
243
11.6
1.50
3rd
Reg.
773
0.526
230
246
7.81
1.54
After Quenching
1st
Reg.
751
0.565
189
227
12.0
1.55
2nd
Reg.
727
0.548
221
239
12.2
1.50
3rd
Reg.
721
0.535
204
245
9.0
1.43
-------
TABLE
TERTIARY TREATMENT PLANTS
Site
Arlington, Virginia
Colorado Springs, Colo.
Dallas, Texas
Fairtax County, Va.
Los Angeles, Calif.
SaJ Montgomery County, Md.
Occoquan, Va.
Orange County, Calif.
Piscataway, Md.
St. Charles, Missouri
South Lake Taho, Calif.
Windhoek, South Africa
Status
1973
Design
Operating Dec. '70
to Present
Design
Design
Design
Des i gn
Design
Construction
Operating Mar. '73
to Present
Construction
Operating Mar. "68
to Present
Operating Oct. '68
to Present
Average
Plant
Design Capacity
Engineer (MDG)
Alexander Potter/ 30
Engineering-Science
Arthur B. Chafet 3
& Assoc.
URS Forest & 100
Cotton
Alexander Potter/ 36
Engineering-Science
City of Los Angeles 53
CH2M/Hill 60
CH2M/Hill 18
Orange County Water 15
District
Roy F. Weston 5
Moran and Cooke 5.5
CH2M/Hill 7.5
National Institute 1.3
for Water Research
Pretoria, So. Africa
Total
No. of Contact Hydraulic Carbon Effluent
Contactor Contactors Time1 Loading Depth Carbon Requirements2
Type In Series (Min.) (gpm/ftz) («) Size (Oxygen Demand)
Down flow 1
Gravi ty
Down flow 2
Upflow 1
Packed
Down flow 1
Gravity
Down flow 2
Gravity
Upflow 1
Packed
Upflow 1
Packed
Upflow 1
Packed
Down flow 2
Pressure
Down flow 1
Gravity
Upf 1 ow 1
Packed
Down flow 2
Pressure
38 2.9 15 8 x 30 BOD
30 5 20 8 x 30 BOD
10 8 10 8 x 30 BOD
BOD
(by
36 3 15 8 x 30 BOD
50 4 26 8 x 30 COD
30 6.5 26 8 x 30 BOD
COD
30 5.8 24 8 x 30 BOD
COD
30 5.8 24 8 x 30 COD
37 6.5 32 8 x 30 BOD
30 3.7 15 8 x 30
17 6.2 14 8 x 30 BOD
COD
30 3.8 15 12 x 40 COD
< 3 mg/1
< 2 mg/1
< 10 mg/1
< 5 mg/1
1980)
< 3 mg/1
< 12 mg/1
< 1 mg/1
< 10 mg/1
< 1 mg/1
< 10 mg/1
< 30 mg/1
< 5 mg/1
< 5 mg/1
< 30 mg/1
< 10 mg/1
1 Empty bed (superficial) contact time
for average plant flow.
2 BOD: Biochemical oxygen demand
COD: Chemical oxygen demand
50 MGD ultimate capacity
-------
TABLE 21
cn
PHYSICAL/CHEMICAL TREATMENT PLANTS
Site
Cortland, New York
Cleveland Westerly,
Ohio
Fitchburg,
Massachusetts
Garland, Texas
LeRoy, New York
Niagara Falls,
New York
Owosso, Michigan
Rosemount,
Minnesota
Rocky River,
Ohio
Vallejo,
California
Status 1973
Design
Construction
Construction
95% Complete
Construction
Design
Construction
50% Complete
Design
Operational
Operational
Design
Average Total
Plant Contact Hydraulic Carbon
Capacity Time1 Loading Depth
Design Engineer (MGD) (Min.) (gpm/TtZ) (ft)
Stearns & Wheeler 10 30 4.3
Engineering-Science 50 30 3.7
Camp Dresser & 15 35 3.3
McKee
URS Forest & 30 30 2.5
Cotton
Lozier Engineers 1 27 7.3
Camp Dresser & 48 20 3.3
McKee
Ayres, Lewis, 6 36 6.2
Norris & May
Banister, Short, 0.6 66 4.2
Elliot, Hendrickson (max)
and Associates
Uillard Schade & 10 26 4.3
Assoc.
Kaiser Engineers 13 26 4.6
17
17
15.5
10
26.8
9
30
36
(max)
15
16
Effluent
Ca rbon Requ i remen ts
Size (oxygen Demand)
8 x 30 TOD
8 x 30 BOD
8 x 30 BOD
8 x 30 BOD
12 x 40 BOD
8 x 30 COD
12 x 40 BOD
12 x 40 BOD
8 x 30 BOD
12 x 40 BOD
(90%
30 mg/1
20 mg/1
10 mg/1
10 mg/1
10 mg/1
112 mg/1
10 mg/1
10 mg/1
15 mg/1
45 mg/1
of time)
1 Empty bed (superficial) contact time for average plant flow
2 BOD: Biochemical oxygen demand
COD: Chemical oxygen demand
-------
TABLE 22 "SUMMARY OF PCT PILOT-PLANT AND FULL-SCALE PLANT PERFORMANCES"
Effluent Effluent Effluent
Blue Plains
Pilot Plant
Owosso,
Michigan
Pomona,
California
Rosemount,
Minnesota (1st year)
Rosemount, Minnesota
(last 3 to 4 months)
Battelle Pilot Pla
at Westerly
CRSD Pilot Plant
at Westerly
Estimated based on BOD similar to COD removals across clarifier.
2
Just around carbon columns.
Raw COD
(mg/l)
320
250-350
321
sar)
>ota
»)
t 527
437
COD
(mg/l)
16
24-30
19
-
42
56
Removal
95
-91
94
-
92
87
Raw TOC
(mg/D
100
-
-
-
-
90
TOC % Raw BOD
(mg/l) Removal (mg/l)
8 92 150
140
1201
230
240
21 77 206
BOD
(mg/l)
6
8
7.8
23
26
32
Removaj
96
84
78.
90
89
84
52
TABLE 23 "CARBON PILOT-PLANT RESULTS FOR PETROCHEMICAL AND REFINING WASTEWATERS"
Type of Wastewater
Refinery
Refinery
Refinery
Petrochemical
Refinery
Refinery
Refinery
Petrochemical
Design Q
(MGD)
28
1.9
22
3
26
28
8
29
Process
Application
Physical/Chemical
Physical/Chemical
Physical/Chemical
Tertiary
Tertiary
Tertiary
Tertiary
Tertiary
Influent
COD (ma/I)
600
800
670
150
100
300
100
150
Effluent
COD (mg/l)
103
201
143
49
41
50
40
48
Percent
Removal
83
75
79
67
59
83
60
68
36
-------
TABLE 24 "REFINERY WASTEWATER TREATMENT RESULTS"
Parameter
BOD (mg/l)
COD (mg/l)
TOC (mg/l)
Oil & Grease (mg/l)
Phenols (mg/l)
Chromium (mg/l)
Copper (mg/l)
Iron (mg/l)
Lead (mg/l)
Zinc (mg/l)
Sulfide (mg/l)
Ammonia (mg/l)
Cyanides (mg/l)
Turbidity, (TTU)
Color (Std. Color Units)
API Separator
Effluent
97
234
56
29
3.4
2.2
0.5
2.2
0.2
0.7
33
28
0.25
26
30
Carbon Treated
Effluent
48
103
14
10
0.004
0.2
0.03
0.3
0.2
0.08
39
28
0.2
11
15
Biologically Treated
Effluent
B iologi cal -Carbon
Treated Effluent
Virgin Carbon
API Separator Effluent to Carbon
Biologically Treated Effluent to Carbon
7
98
30
10
0.01
0.9
0.1
3
0.2
0.4
0.2
27
0.2
17
15
/ITY ANALYSIS"
Iodine No.
1,010
906
991
3
26
7
7
0.001
0.02
0.05
0.9
0.2
0.15
0.2
27
0.2
5
1
Molasses No.
216
405
304
TABLE 26 "PILOT-PLANT RESULTS-TERTIARY CARBON APPLICATION"
COD (mg/l)
SOC (mg/l)
BOD (mg/l)
Influent
Concentration
600
500
400
300
200
300
200
100
50
250
200
150
100
50
20
^Throughput Rate = 0.6 - 1.2 Bed Volumes per hour
^Throughput Rate = 2.4 Bed Volumes per hour
Throughput Rate = 1.2 Bed Volumes per hour
Effluent
Concentration
280 '
230
175
120
65
1402
95
30
10
884
70
54
36
20
10
2503
155
55
25
21 85
186
135
68
30
12
Average
Percentage Removal
532
53
70
80
654
65
64
64
60
50
531
54
56
60
68
,73
23
45
50
135
7
10
32
40
40
throughput Rate = 0.5 Bed Volumes per hour
4Throughput Rate =0.15 Bed Volumes per hou
hour
37
-------
TABLE 27 "DESIGN CRITERIA FOR THE ARCO CARBON PLANT"
Number of Rainfall Days Per Year (max.)
Maximum Rainfall Runoff Rate (per day)
Maximum Rainfall Runoff Rate (per year)
Average Influent COD Concentration
Average Effluent COD Concentration
Carbon Capacity
30 days
12.6 Million Gallons
378 Million Gallons
250 mg/l
37 mg/l
] Ib carbon exhausted per
1000 gal. water treated (1.75
Ib COD removed/lb carbon)
TABLE 28 "ACTIVATED CARBON ADSORPTION DESIGN DATA AND ACTIVATED CARBON PROPERTIES"
Rated Flow (Each of Three Adsorbers)
Adsorber Diameter
Adsorber Bed Depth
Contact Time (Empty Bed)
Hydraulic Loading
Design Inlet Pressure
Pressure Drop Through Carbon
Carbon Inventory Carbon Bed
Adsorber Total
Theoretical Carbon Capacity
Carbon Dosage 0.86 Ib carbon/1,000 gal Throughput
Activated Carbon Properties Filtrasorb 300
(8 x 30 Bituminous Coal)
667
10
45
40
8.5
60
35
92,000
100,000
0.3
gpm
ft
ft
min _
2
gpm/ft
psi
psi
Ib
Ib
Ib TOC/lb carbon
Total Surface Area (N, BET Method)
Bulk Density
Particle Density Wetted in Water
Mean Particle Diameter
Iodine Number, minimum
Ash
Moisture
950-1050
26
1.3-1.4
1.5-1.7
950
Max 8%
Max 2%
m /g,
lb/ftJ
g/cc
mm
TABLE 29 "THERMAL REGENERATION DESIGN DATA"
Furnace
Regeneration Rate
Steam Addition Rate
Fuel
Fuel Rate
Combustion Air Rate
Design Temperatures
Hearth 4
Hearth 6
Afterburner
Hearth 4
Hearth 6
Afterburner
Hearth 4
Hearth 6
Afterburner
60" x 6 Hearth with Integral Afterburner
125 Ib/hr
125 Ib/hr
Refinery Fuel Gas
188 CFH
68CFH
310 CFH
5,000 CFH
1,800 CFH
8,120 CFH
1,725°F
1,750°F
1,250 F
38
-------
FIGURE I
ADSORPTION OF DBS TO EQUILIBRIUM
IN CONTINUOUSLY STIRRED
FLOW SYSTEMSW
160
140
120
100
80
60
40
20
k, = 1.23 mg/g-hr
k2 = 1.78 mg/g-hr
k. =2.80 mg/g-hr
-k0= 4.28 mg/g-hr
I
I
I
15 21 27 33 39
TIME (hours)
45 51
ORGANIC CONSTITUENT IS DODECYL-BENZENESULFONATE (DBS)
Figure 2
FREUDLICH ISOTHERM APPLICATION ^
' DATA CONFORM TO
- FREUNDLICH
. ISOTHERM
§
K
O
O
a
o
o
DECREASING
WASTEWATER
COMPLEXITY
FREUNDLICH ISOTHERM
APPLICABILITY
RESTRICTED TO
DEFINED LIMITS
NON-SORBABLE
RESIDUAL
57
EQUILIBRIUM CONCENTRATION, C (mg/l)
39
-------
CARBON IN
GAS OUT
FIGURE 3
Robbie Arm
Rabble Teeth
Steam
CROSS-SECTIONAL. VIEW
OF MULTIPLE-HEARTH FURNACE
HEARTH
I (200°- 300°F)
2 (300e-450°F)
3 (400"- 1000° F)
4 (IOOO°-I6000F)
5 (l600°-l800eF)
6 (1600°-1800° F)
CARBON OUT
MIXED
MEDIA
FILTRATION
MIXED
MEDIA
FILTRATION
RECARBONATION
AND
SETTLING
AMMONIA
STRIPPING
TOWER
FIGURE 4 p.,
SIMPLIFIED FLOW DIAGRAM °5J
SOUTH TAHOE
CHEMICAL
CLARIFICATION
FINAL EFFLUENT
PUMP STATION
RAPID
MIX a
FLOCCULATION
40
-------
< FLOCCULATOR
Raw
Wastewater
FIGURE 6
SCHEMATIC FLOW DIAGRAM
WESTERLY ADVANCED WASTE TREATMENT PLANT
PILOT PLANT
[ELECTROLYTIC UNIT
MFOR HYPOCHLORITE
'GENERATION
91
c!g
LU
41
-------
FIGURE 7
FREQUENCY ANALYSIS FOR EFFLUENT BOD[a
C R S D - PRECARBON OZONATION STUDY
WESTERLY PILOT PLANT
FIGURE 8
FREQUENCY ANALYSIS FOR EFFLUENT COD
N
C R S 0 PRECARBON OZONATION STUDY
WESTERLY PILOT PLANT
o>
E
100
90
80
70
60
50
40
30
20
•
UJ
ID
10
9
8
7
6
100
90
80
70
60
50
40
o
o 20
UJ
UJ
12 5 10 20 30 40 50 60 70 80 90 95
% OF THE VALUES LESS THAN STATED VALUE
• NO PRECARBON OZONATION
A OZONE <6.2 mg/l
O OZONE =6.2 mg/l
a OZONE =3.2^6.2 mg/l
98 99
10
9
8
7
6
A
O
D
2 5 10 20 30 40 50 60 70 80 90 95 98 99
% OF THE VALUES LESS THAN STATED VALUE
NO PRECARBON OZONATION
OZONE <6.2 mg/l
OZONE =6.2 mg/l
OZONE =3.2 —6.2 mg/l
-------
FIGURE 9
FLOW DIAGRAM FOR THE GARLAND PLANT
INFLUENT
\
(EQUALIZATION I
I
IPRETREATMENTI
1
* ^PRIMARY
9 CLARIFIER
1
1
4
SLUDGE
CONDITIONING "
*
„ TRICKLING
P FILTERS
FILTER
PRESS
1
— 1
1
FINAL
CLARIFIER
llNCINERATIONi
\
[LANDFILL!
CHEMICAL
CLARIFICATION
-*|RECARBONATION
|-»| FILTRATION H
h CARBON I
ADSORPTION!
, 1 ,
IDISINFECTION!
EFFLUENT
FIGURE 10
PHYSICAL/CHEMICAL PILOT PLANT JCHEMATIC
POL
hii
potr
cb cb <,
ram
FLOCCULATION
43
-------
FIGURE II
SULFIDES AND HEADLOSS vs TIME
0
345 678
STUDY PERIOD (weeks)
FIGURE 12
BOD REMOVAL TREATMENT RESULTS
0
10
CO
o>
E
CO
UJ
a
CO
_J
160
API EFFLUENT
6 BIOLOGICALLY
TREATED
EFFLUENT
O CARBON TREATED
EFFLUENT
D CARBON TREATED
BIOLOGICAL
EFFLUENT
67 8 9 10
STUDY PERIOD (days)
44
12
13
14
-------
FIGURE 13
COD REMOVAL TREATMENT RESULTS
API EFFLUENT
& BIOLOGICALLY
TREATED
EFFLUENT
O CARBON TREATED
EFFLUENT
O CARBON TREATED
BIOLOGICAL
EFFLUENT
"0123
FIGURE 14
ARCO FLOW DIAGRAM
WATSON REFINERY
PROCESS
WASTEWATER
567 8 9 10 II
STUDY PERIOD (days)
13 14
1
1
API SEPARATOR
I
NEUTRALIZATION
J
CHEMICAL
FLOCCULATION
I
DISSOLVED AIR
FLOTATION
T
To LACSD^_
HOLDING BASIN
so SG
DOWN FLOW
BEDS
BACKWASH
PUMP
SUMP
HOLDING BASIN FLOW TO
LACSD OR TREATMENT
SYSTEM DEPENDING
ON WATER QUALITY
z
tu
a.
CARBON
HOPPERS
— i
35—1
iff FINAL I
CHLORINE
FINAL
EFFLUENT
SUMP
~l
V
V
L 1
REGENERATION
FURNACE
FINAL EFFLUENT
FLOW DIVERTED TO HOLDING BASIN
WHEN RAINFALL EXCEEDS
O.I INCHES
CARBON
45
-------
FIGURE IS
PERFORMANCE OF ARCO CARBON PLANT
COD REMOVAL •
400
360
320 —
o> 280
Q
i
<
O
240
200
160
120
I 10 20 30 40 50 60 70 80 90
% OF THE VALUES LESS THAN STATED VALUE
99
D INFLUENT I
O EFFLUENT I
FIGURE 16
PERFORMANCE OF ARCO CARBON PLANT
080 REMOVAL
99
10 20 30 40 SO 60 70 80 90
% OF THE VALUES LESS THAN STATED VALUE
INFLUENT* EFFLUENT
-------
FIGURE 17
CARBON TREATMENT SYSTEM
REICHHOLD CHEMICALS CO.
Influent
Nonionic
Polv
Acid.
EQUALIZATION
ACID MIXING
TANK
FLOCCULATION
TANK
CARBON
ADSORBER
FEED SUMP
MULTIPLE-
HEARTH
FURNACE
Sludge To
Disposal
t
''-"': ' '
•:';::::'::.'
9
. _«. .^
JL
ift;
i;;^:.::;Wg
5S?::;:'^
SiSsjii
CARBON
ADSORBER.
%
1
i
FINAL
EFFLUENT
HOLDING
BASIN
V
I
I
I
I
T
To
River
CARBON--"-
FIGURE 18
BP TREATMENT SYSTEM FLOW DIAGRAM
MARCUS HOOK REFINERY
Backwash
_AP_L Bottoms
API SEPARATOR
— 1
1 L
5LUDG
fANK
1
4-
SURGE BASIN
Treated
Process
Effluent
Ones - through
Cooling Water
IU u \>
f
w,
mK
m
yftW::
'x--:'r:-.:
:'?Sr
:';:;:¥•::
Zl
ii
i
jar
Sm>
«
m
'•'£•'$'.
if
I
55|i
|
to
a»
.•/.-.'•.'..•;
«
mm
m
»s*
ii
•
•
1
^
I
«. ___ 4d---^_— - J
. t CARBON ADSORBERS
Air
3
DOWNFLOW FILTERS
MULTIPLE-
HEARTH
FURNACE
- Fuel
- Steam
QUENCH
TANK
47
-------
FIGURE 19
COD REMOVAL AS A FUNCTION r i
OF ADSORBER CONTACT TIMEL32!
PERIOD I
FIGURE 20
ADSORBER OIL REMOVAL AS A r ,
FUNCTION OF INFLUENT CONCENTRATION L32J
PERIOD I
100
90
80
— 70
-J 6O
UJ
tt
Q
O
O
50
40
30
20
10
0
100
90
80
70
< 60
O
Uj 50
o:
o 40
30
20
10
0
40 42 44 46 48 5O 52 54 56 58 60
CONTACT TIME (minutes)
0 10 20 30 40 50 60 70 80 90 100
INFLUENT OIL(mgXI)
INITIAL REMOVALS
REMOVALS AFTER START OF BED PULSING
-------
FIGURE 21
FULL-SCALE PROCESS PERFORMANCE
BP MARCUS HOOK
FIGURE 22
EFFLUENT COD ATTAINABLE FROM
ACTIVATED CARBON SYSTEMS
100
90
80
70
60
^o
£ 50
O 40
UJ
a:
30
20
10
PERIOD
• PHENOLS
• COD
en
E
O
O
O
UJ
III
z
<
UJ
a:
UJ
o
z
O
PERIOD I VIRGIN CARBON, NO FCCU FOUL CONDENSATE
2 CARBON BEDS NOT TURNED OVER, FCCU
CONDENSATE INCLUDED
3 CARBON BEDS TURNED OVER, FCCU
CONDENSATE INCLUDED
4 EFFLUENT SEPTUMS BENT, NO FCCU
CONDENSATE, ONE YEAR OPERATING
450
40O
350
300
250
200
150
100 -
50 -
D Systems treating
biological treatment
system effluent
A Only full-scale refinery
system treating total
waste
10
20 30
50
70 80 90
99
PERCENT OF PILOT-PLANT OR FULL-SCALE SYSTEMS
TREATING REFINERY OR RELATED WASTEWATERS
WITH LONG TERM MEAN COD LESS THAN
STATED VALUE
-------
FIGURE 23
INFLUENT AND EFFLUENT OIL AND
GREASE DISTRIBUTIONS FOR A
FULL-SCALE ACTIVATED
CARBON SYSTEM
,-.
^
o>
E
GREASE
(0
_i
0
180
160
140
120
100
80
60
40
20
0
Average Influent = 30.2 mg/l
S.D. = 36.3 mg/l
Coef. Vor. = 1.20
No. of Data Points = 55 over 5 months
Average Effluent = 7.4 mg/l
S.D. = 11.1 mg/l
Coef. Var. = 1.5
No. of Data Points = 50 over 5 months
FIGURE 24
CARBON ADSORPTION CAPACITY
FOR VARIOUS PLANTS
— 0.8
0
_n
o 0.7
-O
^ 0.6
TJ
-
r-
^0.2
2 O.I
A A
/
-
— Blue
A-/
\
Z-j i i-Ji
^r^oe, C(
(.
/
x°
-Owosso, Mich.
Plains (D.C.)
A /'
A — Ww*
'osemount, Minn.
•/if.
) ^^
^
-Colorado Sprir,
land Westerly
^
gst Colo.
,-.
^-"
0
--— -
c
E
o
^^-
]
3
100 200 300 400 500
INFLUENT COD TO CARBON COLUMN (mg/l)
PERCENT OF THE VALUES LESS THAN
STATED VALUE
A Municipal Wastewater
O Refinery Wastewater
D Petrochemicals Wastewater
-------
SECTIOH 2
Organic Compound Removal by Activated Carbon
CAUTIONS AND LIMITATIONS ON THE APPLICATION OF ACTIVATED
CARBON ADSORPTION TO ORGANIC CHEMICAL WASTEWATERS
C. T. Lcwson
and
J. C. Hovious
Union Carbide Corporation, South Charleston, West Virginia
OBSERVATIONS ON ACTUAL WASTEWATERS
Normal practice in assessing the feasibility of activated carbon adsorption is to first
conduct batch adsorption isotherms with powdered activated carbon. These tests provide
a relative indication of the amount of organic removal achievable by adsorption and the
ultimate adsorptive capacity of the carbon. Unfortunately, these data are useful only in
a relative sense - for comparing the relative merits of two different carbons, or for
comparing the relative amenability of different wastewaters to adsorption. Isotherm data
are not suitable for designing continuous granular carbon adsorption columns, since the
dynamic effects and interactions in a continuous bed differ too greatly from the batch
equilibrium situation in an isotherm.
Batch adsorption isotherms have been performed on many organic chemical bio-
treated effluents and raw wastewaters of widely varying composition and overall organic
concentration (COD or TOC basis). Figure 1 shows curves of percent organic removed vs.
carbon dosage. It is obvious the amenability of these wastes to adsorption is rather wide -
ultimate percentage organic removals vary from 45% to 90%. Even for those cases where
90% organic removal can be achieved by massive carbon dosing in batch tests, the
organic removals achievable in continuous adsorbers are lower, often significantly so.
Comparing batch and continuous data for Plant A bio-treated effluent:
Ultimate % removal of TOC in isotherm test 90%
%TOC removal in continuous bed adsorber prior to breakthrough
(1 bed vol/hr) 52%
%TOC removal before breakthrough (0.5 bed vol/hr) 64%
Not only was the organic removal predicted by isotherms not achieved, even in early
stages of the column run with fresh carbon, but doubling the bed depth (halving the
throughput rate) only raised the amount of TOC removal from 52 to 64%.
When batch adsorption data are plotted as log (wt. organic adsorbed/wt. carbon) vs.
log (organic concentration remaining in solution), the Freundlich isotherm results - X/M
vs_. C . When the logarithmic plot is extrapolated to C = CQ (the initial solution concen-
tration), the resulting (X/M)- is called the "adsorptive capacity" of the carbon. These
C «
isotherm capacities are usually°greater, often very much greater, than the ultimate
adsorptive capacity exhibited by a granular adsorption bed at exhaustion. Typical examples
from prior publications (1):
51
-------
Adsorptive Capacity, gm COD/gm carbon
Wastewater Isotherm Column at Exhaustion
Plant A bio-effluent TT~ ~ 0.26
Plant D bio-effluent 0.65 0.24
Plant G raw wastewater 0.075 0.05
Note that carbon had a very low adsorptive capacity for the constituents in Plant G
wastewater.
It was further noted that increasing the carbon bed depth yielded not only the afore-
mentioned limited increase in percentage organic removal, but also a diminished adsorptive
capacity. Operating two beds in series until each bed was exhausted at an overall through-
put of 0.5 bed vol/hr (1 BV/hr/column) showed:
gm COD adsorbed/gm carbon
CoTTI Col. 2
Plant A 0.26 0.17
Plant D 0.23 0.13
One may postulate that selective adsorption of the "more adsorbable" organics occurs in the
lead bed. Not only that, when the first bed is exhausted, the second bed has little
remaining capacity - implying that the "less readily adsorbed" species, which are initially
removed in the second bed, block its further usefulness in removing "readily adsorbed"
species (available when the first bed breaks through). This phenomenon is shown by the
breakthrough curves of Figure 2. Just how "easily adsorbed organics" may be differentiated
from "less readily adsorbed" species is a difficult question. The rate of adsorption at the
carbon surface, the rate of diffusion within the carbon pores, and the ultimate capacity of
the carbon for the adsorbates all play a part in affecting "adsorbability." In isotherm tests
these factors are reflected in the measured adsorptive capacity and limiting percentage
organic removal. In continuous columns they are reflected in the organic removal achieved
and the rapidity with which organic breakthrough and exhaustion occur. The absence, or at
least minimization, of diffusion limitations in powdered carbon isotherms can give a markedly
different picture of "adsorbability," in terms of both extent and capacity, compared to the
complex inter and intra particle diffusion phenomena which occur in a granular carbon bed.
PURE COMPOUND ADSORPTION STUDIES
Because of these observations of low percentage removals of TOC and COD, limited
benefits of increasing carbon bed depth, and rather limited adsorptive capacities (in
continuous columns), an extensive investigation was conducted (2) to quantify some aspects
of "adsorbability" of specific organics. The effects of functionality, molecular weight and
structure, aqueous phase pH, solubility, polarity, carbon surface structure, and adsorbate
interactions were investigated. Column vs. isotherm comparisons were also made for several
pure chemicals and simple mixtures.
In the first phase, 93 organic chemicals from 11 functional group families were
subjected to single dosage adsorption tests.
Functional group families:
Alcohols
52
-------
Aldehydes
Amines and alkanolamines
Aromatics
Pyridines and morpholines
Esters
Glycols and glycol ethers
Ketones
Organic acids
Ethers and oxides
Halogenated aliphatics
Test conditions:
Initial compound concentration = 1000 mg/l (or solubility, if 1000 mg/l)
Carbon: 5 gm/l of powdered Westvaco Nuchar WV-G (surface area = 1100 m /gm)
Contact time = 2 hours
Typical graphs of adsorbate loading vs. molecular weight are shown in Figures 3, 4,
and 5 for the alcohols, esters, and organic acids tested. The unfavorable effect of branching
and the favorable effect of unsaturation can also be seen. Figure 6 is a "map" of percent
of compound adsorbed (by 5 gm/l carbon) vs. molecular weight showing the regions which
encompass all the experimental data.
Region A Aromatics (including ring-substituted compounds)
Region B General trend for aliphatic mono-functional oxygen and nitrogen-
containing compounds
Line C Saturated organic acid line, shown for comparison
Region D Glycols, glycol ethers
Note that the aromatics, even those containing polar groups (-OH, -NO«, -Cl) are quite
amenable to adsorption, while the poly-functional oxygenated compounds (glycols, glycol
ethers) are especially difficult to adsorb.
Qualitatively, it was concluded that adsorbability, as reflected in constant dosage
tests, is favored by increasing molecular weight, aromaticity, and degree of unsaturation.
Increasing polarity, solubility, branching, and degree of dissociation (for amines and
organic acids) tended to severely limit the extent of adsorption.
The relative ease of adsorption for simple oxygenated organic compounds may be
summarized as follows:
Number of carbon atoms 4
acids aldehydes esters ketones alcohols glycols
Number of carbon atoms 4 (straight-chain compounds)
acids aldehydes alcohols esters ketones glycols
The relatively low adsorption, in general, of compounds containing 4 carbon atoms in
these tests is particularly noteworthy. The carbon dosage of 5 gm/l was a rather massive
dosage - in a continuous column adsorber it would be equivalent to treating only about
24 gallons of wastewater per pound of carbon. This poor organic removal at high carbon
dosage implies that carbon adsorption may have limited application to many organic
chemical manufacturing wastewaters, from a cost-effectiveness viewpoint. Low molecular
53
-------
weight oxygenated organic cpmpounds represent a very large fraction of the production
volume in the industry.
In the next phase of the study, Freundlich adsorption isotherms were obtained on 1000
mg/l solutions of five different organic compounds (all containing four carbon atoms) using
four different carbons at different aqueous-phase pH levels. Figure 7 shows that pH had a
marked effect on the isotherm adsorptive capacity, (X/M)_ , for butyraldehyde and ethyl
acetate. Data for Carbon A, Witco 517, are shown; similar trends were observed for other
carbons. Gas chromatographic analyses indicated that, at acidic or basic pH levels/ ethyl
acetate was hydrolyzed into its less adsorbable (lower molecular weight) components. At
elevated pH, butyraldehyde underwent an aldol condensation into a higher molecular weight
(more adsorbable) component. More detailed discussion of differences between the carbons
tested and the effects of carbon surface properties are contained in Ref. (2).
A potential problem in successfully applying activated carbon to multi-component
wastewaters is indicated. It is not possible to select a pH level that assures maximal
adsorption of all wastewater constituents. An acidic pH is required to facilitate adsorption
of organic acids (to avoid ionization), near neutral pH favors adsorption of esters, while a
highly alkaline pH is necessary to maximize adsorption of aldehydes.
MULTI-COMPONENT ADSORPTION STUDIES
Interactions in multi-component systems were examined in the third phase of the study
by adsorption isotherm tests on binary and four-component mixtures. The results are shown
in Table 1.
In a test (at pH = 7) of a butanol/ethyl acetate binary mixture (500 mg/l of each
compound), the total adsorptive capacity was observed to be 0.237 gm cpds/gm carbon,
while the sum of the two adsorptive capacities (at 500 mg/l initial concentration) from
pure-component isotherms was 0.27 gm/gm. By gas chromatographic analysis of the residual
solutions, the butanol adsorbate loading was 1.05 times the loading in a pure component test
while the ethyl acetate loading was 0.76 times the pure component loading. Even though
more ethyl acetate was adsorbed from the mixture, on a weight basis, than butanol (as
expected from Figure 7), the interaction had a greater deleterious effect on ethyl acetate, in
terms of using pure component data to predict mixture adsorbate loadings.
By contrast, in the butyraldehyde/MEK binary test, the total adsorptive capacity from
the mixture (0.275 gm cpds/gm) was greater than the sum of the pure component values
(0.270 gm/gm), more butyraldehyde (the "more adsorbable component") was adsorbed on a
weight basis, and the ratio of mixture loading to pure component loading was greater for
butyraldehyde, 1.28 vs. 0.60.
In both four-component mixture isotherms (Table 1), the total adsorptive capacity (0.29
and 0.225 gm/gm for Carbons A and C, respectively) was significantly less than the sum
of the pure component capacities (0.575 and 0.312 gm/gm for Carbons A and C, respectively).
Also, in the Carbon C test, less butyric acid was adsorbed than butyraldehyde or ethyl
acetate, despite the previous pure component conclusion that acids are more adsorbable than
54
-------
other organic compounds.
Thus, while pure component adsorption data are useful in determining which waste
streams are potential candidates for activated carbon treatment, the data cannot be used
to quantitatively predict adsorption from multi-component systems.
CONTINUOUS vs. BATCH ADSORPTION STUDIES
In the fourth and final phase of the study, continuous column, granular carbon
adsorbers were operated to exhaustion for comparison of ultimate capacities with isotherm
capacities. Each column was operated at a throughput rate of 1 bed volume/hour. The
feed concentration was 1000 mg/l in the pure component tests, 2000 mg/l (500 mg/l each
compound) in the four component test (feed TOC = 1200 mg/l). The four component
mixture was butyraldehyde, ethyl acetate, butanol, and butyric acid. In all cases, the
ultimate capacity (at exhaustion) was less than that predicted by isotherms (83 to 89%).
Adsorptive Capacity, gm/gm carbon
Column Column
Organic Isotherm at Exhaustion at Breakthrough
MEK 0.160 0.132 OTTO!
Butyraldehyde 0.220 0.192 0.141
4-Component Mixture 0.170 (TOC) 0.151 (TOC) 0.124 (TOC)
Note that the usable capacity in a single-column system - the capacity achieved at
"breakthrough," where significant organic concentrations appear in the treated effluent -
is significantly less than indicated by the isotherms (64 to 72%). Figure 8 shows the experi-
mental breakthrough curve for the four-component mixture.
A second important criterion when applying activated carbon in continuous beds is the
amount of wastewater that can be treated before breakthrough (often called the "carbon
dosage"). Breakthrough in full-scale practice (where the carbon is removed from service)
would be set at the maximum organic concentration allowed by effluent standards. For
purposes of discussion, breakthrough was set at the 95% organic removal level in these
lab tests. Observed dosages were:
Test Breakthrough "dosage" gal WW treated/lb carbon
MEK 12.7
Butyraldehyde 19.4
4-Component Mixture 12.4
Treating large volumes of wastewater at these inlet concentrations will, thus, lead to
rather short service life (rapid breakthrough), even with the fairly ideal adsorption wave-
front shape in Figure 8.
Hydraulic effects (channeling and axial dispersion) are a partial explanation for the
failure of column adsorbers to achieve adsorbate loadings predicted by isotherms. The
inherent differences between the static equilibrium attained in batch isotherms and the
dynamic situation in a continuous column are also limiting factors, particularly when
treating complex multi-component ( 4) wastewaters.
A recent paper (3) by Keinath has shown that competitive adsorption, even in binary
55
-------
solutions, can lead to chromatographic displacement effects in column adsorbers treating
wastewaters of varying composition. In this situation, the weakly-adsorbed solute is retained
by the carbon for a time and then is eluted as a rather concentrated "peak" upon prolonged
contact with the feed wastewater. Keinath further showed that fluidized bed adsorbers are
much less susceptible to this effect than more conventional packed beds - a finding that could
have significant impact on adsorber design. Chromatographic displacement and related
interaction effects are almost certainly key contributors to the diffused, spread-out adsorption
wavefronts and fluctuating organic removals observed when treating multi-component organic
chemical wastewaters of high component and concentration variability. The question of
dynamic interactions and effects In multi-sorbate systems is an area where further research and
development is sorely needed, not only to shed further light on which wastes can be efficiently
treated by activated carbon adsorption, but also to ascertain how best to design and operate
continuous adsorption systems to minimize the undesirable effects of interactions.
CONCLUSIONS
1. In pure component studies, specific organic chemicals have been shown to differ widely
in their amenability to adsorption, depending on molecular weight, structure, polarity
and solubility. Low molecular weight oxygenated organics are particularly difficult to
adsorb efficiently.
2. The relative ease of adsorption of different functional group compounds can vary strongly
with pH, depending on the chemical nature of the adsorbates. An optimum pH cannot
be predicted for a multi-component wastewater of unknown or varying composition.
3. While pure component data could be used to predict binary adsorption capacity in
isotherms fairly closely, a four-component mixture isotherm showed only about 60% of
the adsorptive capacity predicted. Mutual solubility effects competition for adsorption
sites, and inability to maintain a pH level optimum for all components contributed,to
this inability to extend pure component data to more complex mixtures.
4. While isotherm capacities were somewhat extrapolatible to continuous column behavior
in pure component adsorption and simple mixture studies, such extrapolations have not
proved to be possible with real wastewaters (complex mixtures), particularly with bio-
treated effluents.
5. More importantly, the key parameters of interest in real wastewater treatment situations -
percentage organic removal achievable and water volume treated per pound of carbon
before breakthrough - cannot be predicted from isotherm tests.
6. The physical differences between an equilibrium adsorption situation in a powdered
carbon isotherm and the dynamic, multi-component interactions in a continuous granular
carbon bed are too great to permit prediction of column performance from isotherms or
pure component data. The dynamic interplay of adsorption rates, pore diffusion rates,
hydraulic effects, and pure component chemical properties in continuous columns is an
area where further research efforts must definitely be applied.
REFERENCES
(1) Lawson, C. T., and J. A. Fisher, "Limitations of Activated Carbon Adsorption for
Upgrading Petrochemical Effluents," Water - 1973, AlChE Symposium Series No. 136
(1974)
56
-------
(2) Giusti, D. M., R. A. Conway, and C. T. Lawson, "Activated Carbon Adsorption
of Petrochemicals," JWPCF, 46, No. 5, May (1974)
(3) Keinath, T. M., "Design and Operation of Activated Carbon Adsorbers Used for
Industrial Wastewater Decontamination," paper presented at 68th Annual Meeting -
AlChE, Los Angeles, Nov. (1974)
DISCUSSION
Frank Manning: Would you care to comment on how you would personally combine
activated carbon with biological treatment together, which one is put first?
C. T. Lawson: Our normal thinking is that we would put carbon before biological treatment
on specific streams where it has an application. In other words, on a stream say that's
noxious for some reason but not biodegradable or perhaps biologically inhibitory, then we
consider that a very good application for consideration of carbon. As a post-biological
polishing step, 5t depends on how close we are to meeting a permit. I guess, if a little
more treatment would get us to the permit then I would say a tertiary carbon system might
be justified. I don't think, and we are convinced of this fact, that you can simply add on
a carbon system and make up for deficiencies in your biological system. If you are far
away from meeting a permit, the first think to do is make the bio-system work. Then if you
need carbon to close the gap and it is effective in doing so, put it on; but I don't think that
carbon can ever be considered as a prescription for a poorly operated treatment system.
Sterling Burks: Have you performed any studies on absorption of trace heavy metals? I
know this question was asked of Davis Ford.
C. T. Lawson: No, we haven't.
Milton Beychok: As you know, I have congratulated you once before because I think it is
long overdue that someone brought some science to this field. If we look at the data you
have given us this morning, the isotherms in thermodynami-c terms, the isotherms are really
equilibrium. Your column data telling us you can't approach that equilibrium the way you
think you should and that involves kinetics, and if we look again and make an analogy
between that and other refining processes and what has been done in improving kinetics and
the approach to equilibrium through catalysis. Do you intend or do you know of any research
going on this pretreatment of carbon, trying to do something to improve the kinetics as well
as perhaps making kinetic models and seeing what you can learn?
C. T. Lawson: Well the question of pretreating carbon, I really don't know of any work
that is going on. It seems to be the kind of thing that the carbon vendors would be
interested in and they are pretty competitive and closed-mouth fellows; they don't tell you
much about what they are doing. There is some potential here, I would say, for tailoring
carbons through pretreatment or through activation processes; but it is not being really
studied much to my knowledge. On the question of developing a diffusion model or a
kinetic model, there is quite a bit of academic interest I think in doing this. The only ^
papers that I have seen are really for binary systems and that is not of too much interest in
a practical application sense. I know Dr. Keinath at Clemson University and Dr. Weber
at the University of Michigan have done some work in this modelling area.
57
-------
Anonymous: It just seems to me from what we all know that the regenerated carbon behaves
differently than the fresh carbon and you showed us today percentage wise some remarkable
differences between two different carbons, that it would be a fruitful area for research work
whether it be on the part of Union Carbide or the carbon manufacturer.
C. T. Lawson: Yes, I don't really want to give a commercial for anybody, but I will say
that Witco's carbon in the studies we looked at performed quite well, very good carbon. I
think there has been a lot of reluctance over the years to use anything but bituminous coal
carbons because its "sturdier, more rugged, easier to regenerate, not as friable as the other
carbons." Looking at waste treatment with some of the speciality carbons, coconut-shell
carbons, has shown some pretty good results, - a little bit better than bituminous coal carbons,
But when you talk about 75-85$ per pound for carbon, it is really out of the question for
large volume waste water treatment.
Nick Sylvester: You mentioned at the end of your talk that you have done some more research
that you have not reported on and you also mentioned this apparent chromatographic type
phenomena in the columns. Have you done more work on that and are you going to report
that?
C. T. Lawson: That is what we have tried to do, we have tried to take a column and take
samples down through the column at different times and see just where specific compounds end
up in the column on a dynamic basis, to see if there is some way that you can predict when
something is going to be spit out as a big eluted peak, and I hope that data will be in shape
for the Cincinnati paper.
Dave Skamanca: Davis Ford mentioned that if you are going to put in an activated carbon
system, you are really better off with biological system in front of it to decrease the loading
and apparently improve the performance of the carbon. But as you decrease the loading,
the pounds in the concentration to the activated carbon system, you also decrease the isotherm
data, you don't get as good a loading on the activated carbon. Are you getting ahead of
the game in most of these systems by trying to remove morex>f these pounds up stream with
a carbon, are you gaining on the problem, are you just staying even so to speak?
>
C. T. Lawson: My own opinion is that the kind of capacities that you see in isotherm tests
are really an interesting qualitative observation, but they have no practical significance.
Anything you can take out before the carbon column by reasonable treatment means, I think
is worthwhile. From what we have seen, running continuous column systems, this 25% or
so by weight absorptive capacity (on a COD basis) to petrochemicals seems to be a pretty
real number and relatively independent of the feed concentration in continuous systems for
the wastes we have examined thus far. The fact remains that isotherm capacities obviously
do vary strongly with concentration, you can go up and down. We really see no way to
relate that to a continuous system. We use our isotherm data just for relative comparison
of one waste water to another and one carbon to another; we don't use it in a quantitative
sense at all.
58
-------
BIOGRAPHY
Cyron T. Lawson is a Project Scientist in the
Water Quality Development Group of Union
Carbide's Chemicals and Plastics Division,
Research and Development Department. He holds
the B.Ch.E. and M.S.Ch.E. degrees from Georgia
Institute of Technology. He has taught graduate
courses in water and wastewater treatment as an
Adjunct Instructor at the West Virginia College of
Graduate Studies. Cyron is a Registered
Professional Engineer (Chemical) in West Virginia
and Is a member of the AlChE and its Environmental
Division.
Joseph C. Hovious is a Group Leader/
Technology Manager in the Research and
Development Department of Union Carbide's
Chemicals and Plastics Division. Mr. Hovious
is an M.S. Environmental Engineer from the
University of Illinois. He teaches graduate
courses at the West Virginia University
College of Graduate Studies. He is a member
of the WPCF, AWWA, and a Registered
Professional Engineer in West Virginia.
-------
TABLE 1 "COMPARISON OF MULTI-COMPONENT ISOTHERM DATA WITH SINGLE COMPONENT DATA"
Mixture 1 - Carbon A
Butanol
Ethyl Acetate
Mixture 2 - Carbon A
Methyl Ethyl Ketone
Butyraldehyde
Mixture 3 - Carbon A
Butyraldehyde
Ethyl Acetate
Butanol
Butyric Acid
Mixture 3 - Carbon C
Butyraldehyde
Ethyl Acetate
Butanol
Butyric Acid
(a)
Equilibrium Carbon Loading Equilibrium Carbon Loading Fraction of Single Component
In Mixture Test in Single Component Tests, Equilibrium Loading
g cpd/g carbon (a) g cpd/g carbon (b) from Mixture (c)
0.116
0.121
0.237
0.063
0.212
0.275
0.110
0.160
0.270
0.105
0.165
O70
1.05
0.76
0.60
1.28
0.072
0.075
0.031
0.112
0.290
0.072
0.066
0.023
0.064
0.225
0.080
0.080
0.068
0.084
0.312
0.44
0.47
0.28
0.80
0.90
0.83
0.34
0.76
Compounds each initially present at 500 mg/l. Mixtures 1 and 3 dosed at 1 gm carbon/liter and Mixture 2 at
2 gm carbon/ liter. Equilibrium loadings (X/M)_ are calculated from total TOC loadings and chromatographically
determined fractional compositions of removed material.
(b) Equilibrium loadings measured at C of 500 mg/l.
(c) Calculated by dividing the equilibrium loading of the individual components in the mixture by the single-component
equilibrium loading.
-------
I*
»
•H
d
«
a
I
h
>H
T3
41
§
§
§
O
«R
FIGURE 1
ACTIVATED CARBON TREATMENT OF BIOLOGICAL EFFLUENTS
BATCH ADSORPTION ISOTHERM TESTS
C - initial COD (or TOO concentration before carbon treatment
Plant E, C0 = 288 mg COD/1
a Plant F, Cn = 2300 mg TOC/1
Plant A, CQ = 1030 rag COD/1*
x
Plant C
1363 mg COD/1
*Plant A tests run at different times,
several months apart
20
Carbon Dosage, gm/1
30
FIGURE 2
BREAKTHROUGH CURVE FOR PLANT A BIO-TREATED WASTEWATER
U)
§
U
1000
800
600
400
200
Feed - C,
Col. 2
J_
JL
J_
0.5 BV/hr
0.88 liters/BV
(2 col. )
I
50 100 150 200
Wastewater Throughput, bed volumes (BV)
40
61
-------
FIGURE 3
EFFECT OF MOLECULAR WEIGHT ON AMENABILITY OF ADSORPTION OF ALCOHOLS
FIGURE 4
EFFECT OF MOLECULAR WEIGHT ON AMENABILITY OF ADSORPTION OF ESTERS
0 200
I
z"
0.150
0.100
0.050
Co* 1000mg/ I as Alcohol
5gm/l Carbon Ootagi
*C0« 700 mg/l
Fn-Hexonol
O 2 Ethyl -Butanol
n-Amyl Alcohol
0 *
2-Ethyl Hexanol
Butanol
' e Isobutanol
O t- Butanol
SPropanol
/Q Oltopropanol
^ Alcohol
Ethonol
'Methanol
40 80
Molecular Weight
120
0.200. _
O.ISO
s
E
Z
v
X
O.IOO
0.050
0,,= 1000 mg cpd/l
5 am carbon/I
Butyl
Acrylate
Butyl
Acttate
. , , Isobutyl
Ethyl Acrylate o/// Ace,at,
Propyl Acetate
/ /
Isopropyl Acetate
n-Amyl
Acetate
Ethyl Acetate
• Methyl Acetate
40 80
Molecular weight
A 1
120
-------
FIGURE 5
EFFECT OF MOLECULAR WEIGHT OH AMENABILITY OF ADSORPTION OF ORGANIC ACIDS
0.200
0.190. _
§
I 0.100
X
Tl
'i
E
0.090
C0« 1000 m« cpd/ I
S«m carbon/I
O Bwiioic
Proplonlc
'Acttle
Formic
80
W*I«M
120
FIGURE 6
PERCENT OF COMPOUND ADSORBED VS. MOLECULAR WEIGHT FUNCTIONALITY EFFECTS
100-
.Aromatic
40 60 80
100 120
MOL. WT.
140 160 180 200
63
-------
FIGURE 7
pH EFFECTS ON ADSORPTION REMOVALS OF SELECTED ORGANIC COMPOUNDS
WITH CARBON A
0.5
0.4
0.3
I
?
0.2
O.I
Butyraldthyd* •
Ethyl Acetate
Methyl Ethyl
10
12
pH
FIGURE 8
U
o
1200 '
1000
800
600 .
400
200..
BREAKTHROUGH CURVE FOR ACTIVATED CARBON ADSORPTION OF A
FOUR-COMPONENT MIXTURE IN A CONTINUOUS SYSTEM
Feed
TOC adsorbed* 0.151 gmTOC/gm carbon
Efflumt
Breakthrough
Throughput , b*d volumes (at I BV/H)
64
-------
ORGANICS REDUCTION THROUGH ADD-ON ACTIVATED CARBON AT PILOT SCALE
Fred M. Pfeffer
U. S. Environmental Protection Agency
Robert S. Kerr Environmental Research Laboratory
Ada, Oklahoma 74820
Wyman Harrison and Leo Raphaelian
Argonne National Laboratory
Argonne, Illinois 60439
ABSTRACT
The current wastewater BATEA model for the petroleum refining industry is
the treatment sequence: activated sludge, mixed-media filtration, activated
carbon. In an effort to develop data to assist in evaluating the model for
specific organic compounds, the EPA (Ada, Oklahoma) entered into an Interagency
Agreement with ERDA (Argonne National Laboratory) in January 1975. In cooper-
ation with API, a .25 GPM pilot test was conducted at the SOHIO Refinery in
Toledo, Ohio. Argonne followed with GC/MS analysis of samples collected across
the treatment system to identify specific organics which are treatable versus
those which pass-through (refractories).
The EPA's involvement included: the mobile pilot plant, refinery selec-
tion, conduct of the field study, sample preparation, and reporting. Argonne1 s
analytical results showing a small overall reduction in organics by mixed-media
filtration and a large reduction by carbon adsorption are discussed.
INTRODUCTION
In January 1975, the EPA (Robert S. Kerr Environmental Research Laboratory,
Ada, Oklahoma) entered into an Interagency Agreement with ERDA (Argonne Nation-
al Laboratory, Chicago) to develop data to assist in evaluating the performance
of the BATEA model in the Development Document of 1974 (1). Since that time
the BATEA regulations (and hence the BATEA model) have been remanded by a
ruling of the 10th Circuit Court on the petition for revision of the guidelines
by the API (2). However, the requirements for reconsideration and reissuing
of guidelines as stipulated in the ruling, together with the mandates in PL-
92-500 (Sec. 301. d.) (3), and the Settlement Agreement between EPA and NRDC
(4), are added incentive to complete the work funded through this Interagency
Agreement.
The proposed BATEA model was fixed bed carbon adsorption added onto the
BPT model, which is biological treatment followed by granular media filtra-
tion. The specific treatment train selected for study was activated sludge,
65
-------
mixed-media filtration, and granular activated carbon. The performance cri-
terion was to be the reduction of major organic compounds identified in the
influent to the biological treatment system. Pursuant to the agreement,
Argonne would perform qualitative organics analyses on samples provided by
EPA. Pilot-scale filtration and carbon adsorption would be applied to the final
effluent from a full-scale refinery treatment system. The results would serve
as guidance for determining the need for larger-scale study and would not be
used in predicting the performance of a full-scale add-on carbon system.
Refinery Selection
Considerable time was allocated to refinery selection, as there was suf-
ficient funding to study only one refinery. Repeated discussions and meetings
were held with members of the API's W-20 Task Group to arrive at a "represent-
ative" refinery. It was agreed to acquire permission from a Class B refinery
whose final effluent quality met BPT, with the possible exception of suspended
solids. Other criteria would include intake water quality and variability,
refinery turnaround plans, and final effluent quality, raw waste loading, and
hydraulic detention times typifying the activated sludge process at a Class B
refinery.
Agreement was reached in September 1976, to conduct the study at SOHIO's
Toledo refinery. This is a Class B refinery (crude topping and catalytic
cracking) with coking, having a crude capacity of 120,000 BPSD. The treatment
train at that time consisted of the API Separator, dissolved air flotation
(DAF), activated sludge (extended aeration) having 16-18 hours detention, and
final clarification. The final effluent quality routinely satisfied BPT re-
quirements with the exception of suspended solids. The refinery treatment
system returned to steady state in November 1976, following a 1-month turna-
round period.
The Pilot Study
A 30' EPA mobile trailer was transported to Toledo and positioned near
the final clarifier. Facilities aboard the trailer included 6" I.D. glass
columns for filtration and carbon adsorption (Figure 1), a TOC analyzer for
monitoring organic carbon breakthrough, pumping and distribution capability,
and sampling gear. The sampling equipment, pumps, and distribution lines were
fabricated and installed such that the only materials in contact with water
moving through the pilot treatment system were stainless steel, glass, Teflon,
and polypropylene (Figure 1). Sampling points aboard the trailer were: 1)
SOHIO's final clarifier effluent, 2) pilot mixed-media filter effluent, and
3) pilot carbon column effluent (Figure 2). The two remaining sample points
were SOHIO's intake water and DAF effluent (Figures 3 & 4). These five points
were sampled and iced on 4-hour intervals for 24-hour compositing over a con-
secutive 4-day period. During the study, there were no significant changes
in recorded flows through the full-scale treatment system, as measured by the
hourly biofeed pumping rates.
Two parallel down-flow mixed-media filters were utilized such that while
one was operating for 24 hours, the second, having been backwashed, was ready
for use the next day (Figure 5). Figure 2 shows the configuration of the
66
-------
filtering bed: anthrafilt, sand, and gravel. The sand used has an Effective
Size = 0.2 mm and a Uniformity Coefficient = 4.5. Backwashing was accomplished
by alternately pulsing with air and pumping carbon column effluent.
Two up-flow carbon columns (Figure 6) were packed as shown in Figure 2
and operated in series to achieve a total bed depth of 6 feet. A constant
flow rate of 0.25 gpm was maintained, giving a residence time in the carbon
bed of 36 minutes. The carbon used was Calgon's Adsorption Service Carbon.
Calgon's analyses of a sample from the lot used at Toledo gave these results:
Apparent Density (gcc): 0.51
Molasses Number: 282
Iodine Number: 821
Sieve Result (mesh): 8x40
Attention was given to decontaminating material coming in contact with
water samples. All glassware was cleaned by firing, maintaining 550°C for
1-hour. Sample bottle caps contained Teflon liners which had been cleaned by
Soxlet extraction with methylene chloride—the solvent later used in the
laboratory for extracting the organics from the water samples.
Each daily composited sample set was transported in ice chests to Detroit
for air shipment. The samples arrived at RSKERL in Ada within 9 hours of
final compositing in Toledo.
Laboratory Phase
Performance of the full-scale biosystem and the add-on filtration/carbon
train for the common wastewater parameters is shown in Tables 1 & 2. Some
values are reported as less-than (<), reflecting lower limits of detectability
as a function of the sampling and analytical protocol.
Following the field study, the remaining responsibility of EPA was the
preparation of the composited water samples for organics analysis by Argonne.
This involved a tedious liquid-liquid extraction sequence using methylene
chloride. Again, all glassware was fired for organics decontamination. A
major problem was emulsion formation, requiring emulsion breaking and phase
separation by various techniques. Each organic extract was dried by passing
through anhydrous sodium sulfate and the solvent was stripped, resulting in
1-ml of concentrated extract which was sealed in a glass ampul. A period of
9 man-hours was involved in preparing each sample to the ampul stage; there
were 20 samples requiring this preparation.
Gas-Chromatography/Mass-Spectrometry (GC/MS)
The samples supplied to Argonne for analysis consisted of 1 mililiter
methylene chloride solutions of the acid, base, and neutral fractions com-
posited over the 4-day sampling interval.
Analysis of the specific organics in these fractions was performed on a
Hewlett-Packard Gas Chromatograph/Mass Spectrometer equipped with a data system
67
-------
and such peripheral equipment as a Zeta plotter and hard copy unit. Capillary
columns were used in the gas chromatograph. These columns allow considerably
greater separation and resolution of the organic components in a sample than
do standard packed columns. Capillary columns also provide increased sensi-
tivity and drastically reduced background from column bleed in the mass spec-
tra. Also as opposed to typical GC/MS operation, no separator was used to
remove the carrier gas. The outlet of the capillary column was connected
directly to the source of the mass spectrometer and, therefore, there could
be no discrimination in the amount of each component reaching the mass spec-
trometer. That is, assuming that the individual components in the mixture are
not lost in the column, the effluent of the column and the amount of these
components reaching the source of the mass spectrometer is a true representa-
tion of the quantities of compounds injected on the column. Finally, a Grob-
type injection system was used in place of the inlet splitters typically used
with capillary columns. The Grob system avoids the loss of large amounts of
samples and the discrimination, typically found in split systems, of compon-
ents of the mixture. It permits the analysis of minute concentrations of the
specific organics present.
Figure 7 is a capillary column GC/MS total ion chromatogram of the neutral
fraction of the dissolved-air-flotation effluent. It can be seen that there
are over one hundred peaks or components in this fraction and that many of the
components are present in minute quantities; that is, of the order of 200 ppf-
of the original water sample, assuming 100% extraction efficiencies. It was
found that the organics in this neutral fraction of the DAF effluent were
predominantly n-alkanes, alkyl benzenes, alkyl naphthalenes and polynuclear-
aromatic hydrocarbons.
The activated sludge treatment system reduced the concentration of the
organics in the DAF effluent by nearly 98%, as shown in Figure 8. It can be
seen from the graphs that the peak height of several of the peaks in the FC
effluent is approximately one-twentieth those in the DAF effluent, indicating
there is approximately a twenty-fold reduction in pollutants by the activated
sludge process.
Further reductions in organics were accomplished by the multi-media fil-
ter and the activated carbon column as shown in Figure 9. The concentration
of the largest peaks of the compounds refractory to the add-on treatment
system is of the order of 10 ppb. The percent reduction of the major classes
of organics by the multi-media filter and the activated carbon column is as
follows:
Compound % Reduction
Alkanes
Alkyl Benzenes
Indenes
Indanes
Naphthalene
Alkyl Naphthalenes
Anthracene/Phenanthrene
Alkyl Anthracenes /Phenanthrenes
Other PNAs
70-98
35-90
50-60
76-96
66
65-90
86-93
89-98
96-98
68
-------
It can be seen that there is generally greater than fifty percent reduc-
tion in these classes of organic compounds.
Work has not yet been completed on the acid and base fractions. Results
will be available in a few months. The results of this study will be published
as EPA and Argonne reports and as such will be available to the public.
It is expected that sufficient funds will be forthcoming in the next
fiscal year to search and manipulate the data stored on disc for those consent-
decree organics that may be present in these samples.
REFERENCES
(1) Development Document for Effluent Limitations, Guidelines and New Source
Performance Standards for the Petroleum Refining Point Source Category,
EPA-440/l-74-014-a, April 1974.
(2) Ruling of the 10th Circuit Court, Denver, on the suit of EPA by API,
handed down August 11, 1976.
(3)' Public Law 92-500, 92nd Congress, S.2770, October 18, 1972.
(4) Settlement Agreement in the U.S. District Court for the District of
Columbia between the Natural Resources Defense Council and the U.S."
EPA, Civil Action No. 2153-73, June 7, 1976.
ACKNOWLEDGEMENT
The authors wish to acknowledge the assistance of the Calgon Corporation
relating to activated carbon and the efforts of the API's Water Quality Com-
mittee and W-20 Task Group in arriving at a suitable refinery. We wish to
thank Messrs. C. Tome, L.S. Van Loon, and J.H. Walters for assistance in the
wastewater sampling program and Mrs. C.S. Chow for help with the GC/MS
analyses. Most important, the study would not have been possible without the
cooperation of SOHIO personnel at the refinery in Toledo and in the Department
of Environmental Affairs in Cleveland.
DISCUSSION
Peter J. Foley, Mobil Oil; Would you comment on the contributions of the
filter and the carbon columns in reducing organics?
Raphael ion: Although I do not have the exact numbers at hand, it appears to me that the
multi-media filter had little or no effect on the concentration of organics whereas the
activated carbon removed appreciable amounts of organics.
T. A. McConomy, Calgon Corp.: The naphthalene removal was only 66% as compared
to more than~90% for other PNA's, why?
Raphaelian: These numbers are approximate figures based on the average of peak areas of
individual components. Generally, one can say that those organics that have a long
69
-------
alkyl chain such as an alkane or an alkylated PNA are adsorbed on the carbon whereas
parent PNA's are not. This is, of course, a crude approximation. We are still trying to
get better gas chromatograms because all of these results are dependent upon how well you
separate the compounds.
N. F. Seppi, Marathon Oil: Please comment on methylene chloride purification - also
what about decomposition products from methylene chloride under basic conditions?
Pfeffer; Regarding the purity of methylene chloride, we relied upon the glass
distillation procedures of the manufacturer (Burdick and Jackson). In addition,
Argonne received a blank extract obtained by taking a high purity water and
performing the acids, neutrals and basic extractions. This blank would also
account for laboratory contaminations. I cannot offer any information about
alkaline decomposition products.
Anonymous: Is there additional data from industry on the study at Toledo?
Pfeffer; I do not know. Both Exxon and SOHIO conducted parallel work to our
own, presumably into the realm of GC mass-spec. We would entertain comparing
notes with Exxon and SOHIO at some later date in order to validate what actu-
ally took place in Toledo.
Judith Thatcher. API; I noticed that the TOC of the influent water is very
close to the TOC of the final effluent. Have you done any identification in
the organics in the influent water to the refinery?
Pfeffer; We are looking at it, but haven't identified all the components yet.
Arthur J. Raymond, Sun Oil Co.; What phase was used on the capillary columns,
and did you notice that your highly branched-chains were degraded much faster
than the less branched? Also, was benzene degraded much faster than toluene
and xylene?
Raphaelian^ I don't understand what you mean by degraded.
Raymond: Did they decompose faster or disappear or reduce? Not in the column
but in your system as you went from the influent water to the final clarifier.
If you had percent reductions, which compounds went faster than others?
Raphaelian; I am still putting all this data together. However, I can say
that it appears that the branched-chained alkanes, which were present in smaller
quantities than the straight-chained alkanes, were not removed as well by the
treatment system as the straight-chained alkanes. Because of the minute
quantities of pollutants present, I am presently doing single-ion monitoring
to try to get a better idea of the percent reduction across the treatment
system.
Raymond; What phase did you use on a capillary column? What coating?
Raphaelian; For the work presented in this talk, OV-101 was the liquid phase
and the columns were wall coated open tubular (WCOT) and not support coated
70
-------
open tubular (SCOT) capillary columns. It is difficult with OV-101 to get
symmetrical peaks with polar compounds, that is, peaks without tailing. I
used FFAP capillary columns for the acid and base fractions. By the way, we
see a variety of alkylated phenols in the acid fraction.
Ed.Sebesta. Brown & Root; I noticed that for TSS there is no decrease across
the filter. Do you have any explanation for that?
Pfeffer; My only explanation is that considering the flow rate and sand spec-
ifications, the TSS coming from the final clarifier were such that the filter
was ineffective. Also/ at the 10 mg/l level, differences are probably within the experi-
mental error of the test procedure.
BIOGRAPHIES
Fred M. Pfeffer holds the BA and MS degrees in Chemistry
from the University of Cincinnati. He is currently a
Research Chemist at the EPA's Robert S. Kerr Environmental
Research Laboratory at Ada, Oklahoma.
Wyman Harrison holds the SB, SM, and a PhD degree in
Geophysics from the University of Chicago. He is
currently the Assistant Director of Applied Geoscience
and Engineering, Energy and Environmental Systems
Division at the Argonne National Laboratory, Argonne,
m* •
inois.
Leo Raphael ian holds the AB degree from Harvard
University and the MA and a PhD degree in Chemistry
from Yale University. He is currently Manager of
Environmental Sciences, Energy and Environmental
Systems Division at the Argonne National Laboratory,
Argonne, Illinois.
71
-------
Table 1. DAILY PERFORMANCE FOR COMMON WASTEWATER PARAMETERS
Oil & Grease
Cyanide
Phenol
COD
BOD
TOC
TSS
Oil & Grease
Cyanide
Phenol
COD
BOD
TOC
TSS
MG/L INTAKE MG/L DAF EFFLUENT MG/L FC EFFLUENT
Day 1 Day 2 Day 3 Day 4
<10 <10 <10 <10
<0.02 <0.02 <0.02 <0.02
0.03 <0.01 0.03 0.01
<15 18 <15 <15
<10 <10 14 <10
19 19 17 15
35 29 11 <10
Day 1 Day 2 Day 3 Day 4
22 33 21 22
0.19 0.25 0.31
320 260 520 450
122 172 154 154
82 127 108 96
39 56 72 60
31 56 37 30
Day 1 Day 2 Day 3 Day 4
<10 <10 <10 <10
0.16 0.12 0.20 0.10
0.02 0.01 0.04 0.02
49 50 51 44
<10 15 21 24
22 29 27 17
12 <10 <10 <10
MG/L FILTER EFFLUENT MG/L CARBON EFFLUENT
Day 1 Day 2 Day 3 Day 4
<10 <10 <10 <10
0.16 0.15 0.20 0.10
0.02 0.01 0.02 0.02
42 38 51 44
<10 11 22 27
19 26 23 18
<10 <10 12 12
Day 1 Day 2 Day 3 Day 4
<10 <10 <10 <10
<0.02 <0.02 <0.02 <0.02
<0.01 <0.01 <0.01 <0.01
<15 <15 <15 <15
<10 <10 <10 <10
10 12 11 <5
<10 <10 <10 <10
-------
Table 2. AVERAGE PERFORMANCE OVER 4-DAY STUDY PERIOD
FOR COMMON WASTEWATER PARAMETERS
Oil & Grease
Cyanide
Phenol
COD
BOD
TOC
TSS
MG/L INTAKE
<0.02
0.02
18
21
MG/L DAF
24
0.25
390
150
103
57
38
MG/L FC MG/L FILTER MG/L CARBON
0.14
0.02
48
17
24
0.15
0.02
44
17
22
<0.02
<0.01
73
-------
Figure 1. CARBON COLUMNS
Figure 3. SAMPLING POINT: PLANT INTAKE
-------
WASTE
«n
FINAL CLARIFIER
FIGURE 2 - PILOT TREATMENT FACILITY
-------
Figure It. SAMPLING POINT: DAF EFFLUENT
76
-------
Figure 5. MIXED MEDIA FILTERS
Figure 6. CARBON COLUMNS
-------
oo
JkLbul
OAF EFFLUENT
JUJL
0_, , ^ 40 60
'
80
' idO ' ' 120
TIME (min)
Figure 7. Total Ion Chromatogram of DAF Effluent (Neutral Fraction, Four-Day Composite)
-------
ll
I
Jl,
1,
DAF EFFLUENT
V
FC EFFLUENT
(IOX)
o
I I I
20
40
60
TIME (min.)
80 100 120
Figure 8. Total Ion Chromatograms of DAF Effluent and FC Effluent (Neutral Fraction
Four-Day Composite) v-i-xuu,
-------
ACTIVATED-CARBON
EFFLUENT (IOX)
/s^fj^^
FINAL CLARIFIER
EFFLUENT
120
TIME (mm.)
Figure 9- Total Ion Chromatograms of the Activated-Carbon and Final-Clarifier Effluents (Neutral Fraction
Four-Day Composite) '
-------
SECTION 3
Granular Activated Carbon Pilot-Scale Studies
"ACTIVATED SLUDGE ENHANCEMENT: A VIABLE
ALTERNATIVE TO TERTIARY CARBON ADSORPTION"
Leonard W. Crame
Senior Chemical Engineer, Texaco Inc.
INTRODUCTION
In view of the possibility of more stringent 1983 BATEA (Best
Available Technology Economically Achievable) effluent guidelines,1'2'3'^
petroleum refiners are faced with the dilemma of an insufficient data base
to determine the proper approach for making cost-effective improvements.
The EPA previously proposed granular activated carbon adsorption after acti-
vated sludge treatment as BATEA technology; however, the current emphasis is
to consider both effluent quality and the cost effectiveness of attaining the
desired results. Two proposed approaches to BATEA technology are (1) in-
creasing the sludge age (or mean cell residence time) of the activated sludge
biomass to develop a more diverse population capable of assimilating biore-
fractory organics or (2) adding powdered carbon directly to activated sludge
aeration basins. Both alternatives to tertiary carbon adsorption would re-
quire little capital investment and would lower operating costs.
Grutsch and Mallatt5'6'7'8'9'10'11 have proposed that the best
refinery end-of-pipe treatment for soluble organic removal should include pH
control, equalization, optimized dissolved air flotation (DAF), and high
sludge age (20-50 days) activated sludge treatment. High sludge ages (SA)
require mixed liquor solids levels above conventional levels (5-10 days SA).
These higher levels increase solids flux and must be considered in secondary
clarifier solids loadings. Also high effluent TSS, despite less frequent
sludge wasting, can result in a loss of mixed liquor solids.
Grutsch and Mallatt emphasize that optimized chemically-assisted
DAF pretreatment (or comparable pretreatment) reduces the colloid charge
(zeta potential) to maximize particle agglomeration for efficient flotation,
and reduces the organic load on the activated sludge unit (ASU). Removing
colloids normally present in raw refinery wastewater allows better biofloc-
culation and lower effluent total suspended solids (TSS) since most refinery
colloids and biosolids have repelling negative charges. The microbial popu-
lation could then acclimate to the biorefractory organics by producing
enzymes which reduce these to simpler biodegradable substrates. Current
reports from within the petroleum industry seem to indicate some benefits
for increasing SA. Other investigators12 have reported that high SA (low
food/microorganism ratio) produces poor sludge settleability.
-------
As a result of pilot studies at the Du Pont Chambers Works, Button
and Robertaccio13 were issued a U.S. patent^ for the Du Pont PACT process. 1
The PACT process basically^ involves the addition of powdered carbon (or
fuller's earth, etc.) to an ASU, usually in a range of 50-400 mg/1 based on
influent flow. Du Pont has reported!6,17,18 a number of advantages of the
PACT process which include:
(1) color removal,
(2) stability against shock loadings,
(3) improved BOD removal,
(4) improved refractory organic removal,
(5) resistance to toxic substances,
(6) improvement in hydraulic capacity,
(7) improved nitrification (mainly in municipal wastes),
(8) foam suppression, and
(9) improved sludge settling and increased clarifier capacity.
A disadvantage of the PACT process is that the system can become very expen-
sive if powdered carbon addition rates become high (hundreds of mg/1) , even
though powdered carbon is cheaper than granular carbon.
DeJohn and Adamsl.9»20,21 nave developed a considerable amount of
pilot study data on activated -sludge-powdered carbon systems. They report
significant enhancement in studies involving refinery and petrochemical
wastewaters. DeJohn and AdamS explain the powdered carbon enhancement
mechanism as localization and concentration of oxygen and pollutant as the
result of adsorption on carbon surfaces, resulting in a more complete bio-
oxidation. The adsorption of biorefractory organics allows a longer resi-
dence time for these components in the system. Other researchers22,23 have
found similar improvements using activated sludge-powdered carbon systems
and propose analogous enhancement mechanisms.
has reported a case history of a full-scale activated
sludge-powdered carbon demonstration run at the Corpus Christi, Texas, Sun
Oil refinery. Results included better system stability, reduction of foaming,
resistance to upset conditions, lower effluent suspended solids and clearer
effluent, and improved organic removal. These improvements were achieved
by maintaining only a 450-mg/l powdered carbon reactor concentration with a
10-mg/l powdered carbon dosing requirement. The shortcoming of this inves-
tigation was that a parallel control could not be run simultaneously and most
improvements reported could possibly have been attributed to better
clarification.
The merits of powdered carbon enhancement have been further
confused with the more recent development of several types of powdered
carbons with significantly different properties.
82
-------
SCOPE OF WORK
1. Objective
The objective of this study was to determine if the relatively
simple process changes of increased sludge age or the addition of powdered
activated carbon in conventional refinery activated sludge systems can sig-
nificantly enhance the removal of organic wastewater contaminants to achieve
or approach the level of proposed BATEA (1983) technology more cost effec-
tively than the addition of granular activated carbon contactors to BPCTCA
(1977) technology.
2. Procedures
In Part I of this study, five completely-mixed (15 gal) ASU's were
operated in parallel with identical 18-hr retention times and 300-gpd/sq ft
clarifier rise rates. A sixth ASU was run as a second-stage unit with the
same 18-hr retention time (Figure 1). All biological reactors were located
in a temperature controlled room in an attempt to dampen influent wastewater
temperature variations and control biological reactions at about 85 F.
ASU's A and F served as controls, simulating conventional refinery
units with a 0.3 Ib TOC/lb MLVSS-day organic loading. Separate controls were
run to determine the effect of optimized pretreatment on activated sludge
treatment and tertiary carbon adsorption. Equalized (24-hr) and pH-control-
led refinery wastewater was pretreated by dual-media (sand-anthracite) fil-
tration (4.6 gpm/sq ft) and a chemically assisted DAF unit (1.5 gpm/sq ft)
prior to control ASU's A and F, respectively. The optimized DAF pretreatment
neutralized the negatively charged colloids, thus facilitating their removal
and producing a bio-unit feed that contained essentially only soluble
organics. Sodium phosphate (monobasic) was added to the filter and DAF unit
effluents for a minimum TOG: phosphorus ratio of 100:2 to assure a proper
nutrient level. Effluents from ASU's A and F were continuously filtered
through a dual-media (sand-anthracite) tertiary filter for TSS removal before
passing through a series of four granular activated carbon contactors to
simulate proposed 1983 BATEA technology.
ASU's B, C, and D treated optimized DAF effluent with sludge wast-
ing calculated for a 50-day biological SA. A commercially available, conven-
tional-surf ace-area, powdered carbon25 (designated PC-C) was added daily to
ASU C to maintain a 500-mg/l reactor operating level. Similarly, PC-H, a
high-surface-area powdered carbon,26 was added to ASU D to investigate its
enhancement capabilities.
ASU E was also operated at a 50-day SA treating ASU B effluent to
determine if there was any benefit to ASU staging.
In Part II of this study, the second-stage ASU E was placed in
parallel with other ASU's treating the DAF unit effluent as shown in
Figure 1 and redesignated as ASU G. PC-H was added to ASU G to maintain
a 2500-mg/l operating level while powdered carbon levels were increased in
ASU C and D to 1000 mg/1.
83
-------
Effluents were collected daily, Monday through Friday, as grab
samples. Grab samples were taken in lieu of composites for convenience since
the pilot unit treatment scheme contained significant equalization. Samples
from the equalization basin, biological reactors, and carbon columns were
filtered with Gelman type A/E glass fiber filters to give the soluble con-
taminant (TOC, COD, S=) level. Glass fiber filters were used instead of
0.45-micron filters because the solids retained on glass fiber filters
define the TSS measurement. Samples were analyzed immediately after collec-
tion or were preserved until analyzed using accepted preservation methods.27
All effluent data were compared after plotting values on log-normal
probability papers. Single straight line data fits were determined by calcu-
lating 50th and 90th percentile values. The 50th percentile values equaled
the antilog of the mean of the log values of data sets. The 90th percentile
values were calculated assuming a single-tailed log-normal data distribution
In N90 = In NSO + 1.282 In Sd
where
Sd is the standard deviation.
Engineering judgment was used to determine which data sets being compared
appeared different and required additional statistical analyses to confirm
significance of median differences. Median data values were compared to
determine if they were from the same population using a paired t-
assuming a log-normal distribution as follows:
compute t =
- (Id)2/n
n(n-l)
n
3" « mean of differences = '%~fd±/n. where d^ = In x-^-ln yi
1
for i = l,2,3,.,.n of n data pairs
Sd = standard deviation of d^.
The test hypothesis is that data are from the same population, therefore,
their true medians are equal. This hypothesis is rejected if the calculated
t-value exceeds the tabled two-tailed t-value for n-1 degrees of freedom for
the selected (90 percent) confidence level.
RESULTS
1. Part I - Pretreatment
The blended refinery wastewater contained excessive amounts of
colloids and TSS coated with oil, complicating pretreatment. The DAF unit
chemical dosages were relatively high at 40 mg/1 filter alum (Al2(S04)3'14.3
84
-------
H20) and 20-40 mg/1 Dearborn 431 cationic polymer. Chemical dosages were
initially optimized using zeta potential titrations in conjunction with jar
tests; however, only jar tests were continued since zeta potential calcula-
tions became tedious and inconsistent due to the high wastewater specific
conductivity (usually 4000-8000 micromhos/cm). This salinity was primarily
due to the brackish intake waters to the refinery and may have impeded
chemical coagulation at lower chemical dosages as reported by Grutsch and
Mallatt.8
The superiority of the optimized DAF unit (operating at 1.5 gpm/sq
ft) as a pretreatment system over simple sand filtration is clearly evident
in the three weeks of run data represented in Figure 2.
DAF unit effluent 50th percentile TOG (158 mg/1) was 55 percent
less than the 50th percentile TOG (352 mg/1) in the equalization basin
influent to the DAF unit, while the sand filter gave only a 18-percent reduc-
tion. Since, at best, only a 31-percent TOG reduction could be achieved by
vacuum filtration of equalization basin samples with glass fiber filters
(which define TSS) , the true effectiveness (55 percent reduction) of colloid
and oil coagulation and removal in the chemically assisted DAF unit can be
seen. The DAF unit effluent contained essentially only soluble organic
contaminants.
The continuous dual-media filter (operating at 4.6 gpm/sq ft) could
only manage a TOG reduction of about one-third of the DAF unit. There was
no indication that a shorter run time would improve the filter effluent
significantly. It appeared that due to the nature of the solids, chemical
addition to the filter feed would have been required for an improved system.
The purpose of the filter, however, was to produce a biological reactor feed
with characteristics comparable to DAF treatment without chemicals. It was
observed, on occasions, that DAF pretreatment was very poor when chemical
feed pumps failed.
Figure 2 illustrates COD removal by both pretreatment systems
employed and again demonstrates the effectiveness of optimized DAF treatment.
As in subsequent graphs, the data points are not shown to avoid congestion of
data.
Fiftieth percentile oil and grease values during Part I of this
study were 101, 70, and 16 mg/1 for the equalization basin, filter effluent
and DAF unit effluent, respectively. The equalization basin 50th percentile
TSS level of 78.0 mg/1 was reduced to 57.5 mg/1 by the filter and to 19.0
mg/1 by the DAF unit. A portion of the TSS in the DAF unit effluent was due
to biological growth rather than influent solids. It is possible that part
of the organics reduction through the DAF unit was the result of biological
activity which could also occur in full-scale systems.
2. Part I - Activated Sludge Performance
TOG and COD (filtered) effluent variability plots for pilot-scale
18-hr retention ASU's are compared in Figure 3. These results are from the
initial 3-week data run.
85
-------
Control ASU F produced a better effluent than control ASU A, both
operating at equal 0.3 Ib TOC/lb MLVSS-day (F/M) loadings. MLVSS levels in
ASU A averaged 1,148 mg/1, about twice that of ASU F due to a twofold
increase in feed TOG. Considering both TOC and COD removal, ASU B, C, E,
and F did not show any significant overall difference in performance.
Fiftieth percentile TOC values ranged from 53-58 mg/1 while COD values were
97-116 mg/1 as shown in Figure 3. Sludge age was not a controlling perfor-
mance variable as ASU B (50-day SA) and ASU F (about 10-day SA) differed
greatly in solids retention time with average MLSS levels of 1,621 mg/1 and
816 mg/1, respectively. Chemically assisted pretreatment for removal of
colloids and oil had the most significant effect on organics removal. The
high-surface-area powdered carbon (designated PC-H) significantly enhanced
organic removal in ASU D, with a 50-day SA and a 500-mg/l PC-H operating
level. Enhancement was not evident in ASU C containing the conventional-
surf ace-area powdered activated carbon (designated PC-C). Powdered carbon
addition increased the average ASU C MLSS level to 1,885 mg/1 with ASU D
averaging 1,976 mg/1.
Since a marginal enhancement occurred with the addition of PC-H at
a 500-mg/l level, the scope of this investigation was expanded to evaluate
powdered activated carbon addition at a 1000-mg/l level and only PC-H at
approximately 2500 mg/1. This would give a greater overview of the enhance-
ment capabilities of powdered carbon, especially the highly active PC-H. ASU
E was taken out of service since it was only succeeding in lysing biological
cells as a second stage following ASU B. The reactor was placed in parallel
with other units being fed by the DAF unit and redesignated ASU G. The SA
was maintained at 50 days and PC-H was built up to a reactor level of about
2500 mg/1 for Part II of this study.
3. Part I - Granular Carbon Adsorption
Granular carbon Series A, treating ASU A effluent, exhausted two
130-gram carbon beds during 17 days of a 3.4-gpm/sq ft hydraulic loading in
Part I. A 20-mg/l soluble (filtered) TOC and a 44-mg/l soluble COD effluent
(50th percentile, Figure 4) was produced with 0.10 and 0.09-g TOC/g carbon
accumulative loadings at exhaustion. Carbon Series F, treating ASU F efflu-
ent, reduced the 50th percentile effluent soluble TOC and COD to 23 mg/1 and
40 mg/1, respectively. Because of the .relatively few data points used to
establish Figure 4, there is little significance in the difference between
carbon series A and F 50th percentile values. A single carbon bed was ex-
hausted to a 0.12-g TOC/g carbon loading. TOC loadings of carbon columns in
Series A were comparable to an average of 0.11 g TOC/g carbon reported for
the granular carbon during the four previous exhaustions prior to each
regeneration.
t
Granular carbon effluents were of substantially better quality than
all biological unit effluents. The 50th percentile soluble TOC and COD re-
ductions in carbon Series A were 44 mg/1 and 84 mg/1, respectively, whereas
carbon Series F accounted for a 50th percentile 35-mg/l soluble TOC and 65-
mg/1 soluble COD reduction.
-------
4. Part II - Pretreatment
Pretreatment by filtration and chemically assisted DAF treatment
continued as in Part I of this study. Again, using dosages of 40-mg/l filter
alum and 20-40 mg/1 Dearborn 431, optimized DAF pretreatment reduced the
equalization basin TOC and COD by more than 50 percent as shown in Figure 5.
Filtration could only remove gross quantities of oil and solids without pre-
liminary chemical coagulation. Although equalization basin, filter, and DAF
unit effluent 50th percentile TOC and COD concentrations were approximately
equal to those experienced in Part I of this study, there existed a greater
degree of variability in Part II. A contributing variability factor was the
rainfall dilution of refinery wastewater streams as an average of 0.21 in./
day of rain fell during Part II compared with 0.06 in./day during Part I.
5. Part II - Activated Sludge Performance
The effluent quality for Part II, basis filtered TOC and COD, is
given in effluent frequency distributions, Figure 6, for the six-week run
period. Control ASU A (F/M = 0.3), without optimized pretreatment, continued
producing the most inferior effluent and experienced three upsets due to the
development of a filamentous bulking sludge. The unit was restarted on each
upset occasion with new seed and allowed to acclimate for a few days before
effluent data were used for comparison with parallel systems. ASU B, C, and
F, as in Part I, produced nearly equivalent effluents in terms of filtered
TOC and COD with neither high SA (50 days) nor 1000 mg/1 PC-C enhancing bio-
logical treatment. PC-H added to ASU D and G at levels of 1000 and 2500
mg/1, respectively, reduced TOC and COD substantially. Compared with high SA
control ASU B, 50th percentile TOC was reduced an additional 10 mg/1 and 22
mg/1 in reactors D and G, respectively. COD 50th percentile reductions below
reactor B were 22 mg/1 for ASU D (1000 mg/1 PC-H) and 39 mg/1 for ASU G
(2500 mg/1 PC-H). The ASU G run time was abbreviated, however, due to the
time required for acclimation at the higher PC-H level. As in Part I of this
study, it was observed that as powdered carbon levels were suddenly increased
in ASU C, D, and G, performance was exceptionally good for a short period of
time.
Phenols feed concentrations were higher in Part II of this investi-
gation as 90th percentile values reached 18 mg/1, compared with 8.5 mg/1 in
Part I. ASU D (see Table 1) and G, containing PC-H, provided the best
phenols removal with 50th percentile phenols levels of 0.05 mg/1 and 0.04
mg/1, respectively. This was slightly lower than the high SA control ASU B
(0.06 mg/1) and low SA control ASU F (0.07 mg/1). Although the lack of opti-
mized pretreatment produced higher 50th percentile phenols levels (0.11 mg/1)
in ASU A, even poorer reductions were experienced with ASU C as in Part I.
Similar results were obtained in Part I. An occasional high phenols value
was measured in the effluents of ASU D and G but not with the consistency
or magnitude of ASU C.
Effluent oil and grease values, included in Table 1, illustrate the
significance of removing most of the oil and grease before biological treat-
ment. The 18 mg/1 50th percentile oil and grease effluent level of ASU A
greatly exceeded the concentration of 5 mg/1, or less, discharged from
87
-------
reactors receiving optimized pretreatment. The addition of PC-H to ASU D and
G gave slight oil and grease improvement with SA alone not being an
enhancement factor.
Effluent TSS levels were high in ASU A at a 50th percentile level
of 86 mg/1. TSS increased to more than 150 mg/1 when filamentous sludge
bulking occurred. However, ASU's with DAF pretreatment produced a more
settleable sludge. The high SA control ASU B had a significantly higher
effluent TSS level than the lower SA control ASU F, but the high SA could
be maintained. ASU C and D experienced effluent TSS levels less than control
ASU B despite higher reactor solids due to powdered carbon. No effluent TSS
increase was observed in ASU G due to the higher (2500 mg/1) PC-H level.
Ammonia nitrogen removal in system A was less than the 80 percent
achieved in the systems with optimized pretreatment. The organic loading was
higher and the sludge age was less than in other systems. The factors con-
trolling the degree of nitrogen removal were not investigated. Nitrification
during Part II was not as complete as that obtained in Part I.
Sludge Characteristics and production rates are summarized in
Table 2 for all ASU's. As expected, ASU A had the highest measured oxygen
uptake averaging 0.16 mg oxygen/1-min due to a higher influent organic con-
centration. Oxygen consumption averaged 0.10-0.12 mg oxygen/1-min in other
ASU mixed liquors, but a relationship of increased oxygen demand and enhanced
biological treatment did not exist. The sludge volume index (SVI), a measure
of sludge compactability, significantly improved with SA and powdered carbon
addition. Sludge settling velocities were exceptionally high with the worst
rate (ASU A) being 0.17 ft/min corresponding to a 1830-gpd/sq ft clarifier
rise rate. Other mixed liquors settled with zone settling velocities of
0.34-0.39 ft/min. The average MLSS concentration of 745 mg/1 in ASU F was
too low for zone settling to occur. One of the most surprising results of
powdered carbon addition was that less biomass was produced than in control
systems. ASU G produced an average of 0.08 Ib biomass/lb COD removed com-
pared with control rates of 0.22 for ASU A and 0.19 for ASU F. PC-H was more
effective than PC-C at reducing biomass production rates at the same SA. The
total sludge production of activated sludge powdered carbon systems was not
much higher than controls, due to lower biological sludge production rates.
Powdered carbon inventories and makeup requirements for ASU's are
summarized in Table 3 for Parts I and II of this study. PC-C losses were
slightly higher than PC-H but still reasonably close to 2 percent per day.
Since biological sludge was wasted at a rate of 2 percent per day in high SA
reactors, it is a fairly good assumption that powdered carbon lost in efflu-
ents was in the same proportion to biological sludge as in the mixed liquor.
Thus both biological and powdered carbon SA may be assumed to be equal for
simplification of powdered carbon daily makeup requirements. The powdered
carbons must be wetted to prevent loss of floating carbon in the clarifier.
This was accomplished by boiling the carbon slurry in this study. Vacuum
degassing could also be used.
Another observation made during Part II was that activated sludge-
powdered carbon systems significantly reduced aeration basin foaming compared
-------
with control systems. Foaming in ASU aeration basins was not a problem but
did occur occasionally.
6. Part II - Granular Carbon Adsorption
A single granular carbon bed was exhausted from carbon Series A
with an accumulative organic loading of 0.15 g TOC/g carbon during Part II.
The data include three short runs. The first two carbon beds required back-
washing almost daily due to high TSS levels which could not be continuously
removed by dual-media filtration. Fiftieth percentile effluent soluble TOG
was 30 mg/1 (see Figure 7) for a reduction of 38 mg/1 from ASU A. ASU A 50th
percentile soluble COD was reduced by 84 mg/1 to 79 mg/1. Although phenols
levels were generally low (50th percentile of 0.04 mg/1) a few very high
effluent phenols levels were detected in carbon Series A giving a 90th
percentile phenols value of 4.8 mg/1 (Table 1). Phenols must have been
adsorbed, concentrated, and then eluted in slugs from the carbon beds to
achieve such a high level. Effluent oil and grease levels remained low with
50th percentile values less than 3 mg/1.
Carbon Series F exhausted a single carbon bed to an accumulative
organic loading of 0.13 g TOC/g carbon while surpassing the performance of
carbon Series A. The 50th percentile soluble TOC was significantly lower at
18 mg/1 for a 28 mg/1 reduction (Figure 7) from ASU F. Fiftieth percentile
soluble effluent COD was 64 mg/1 for a 44 mg/1 reduction. Phenols levels
were extremely low at 0.02 mg/1 (50th percentile) and no sudden loss of
adsorbed phenols was detected during most of the Part II data run (Table 1).
Oil and grease effluent levels (50th percentile) were again less than 3 mg/1.
The lower dashed lines in Figure 7 represent the performance of ASU
G, the best of the activated sludge-powdered carbon reactors. ASU G produced
an effluent superior to carbon Series A and approached the quality of carbon
Series F. The powdered carbon enhancement removed about 85 percent of the
soluble TOC adsorbed on carbon Series F and about 60 percent of the COD based
on 50th percentile effluent values. The 2,500 mg/1 PC-H operating level in
ASU G significantly reduced effluent color to a level comparable with
granular carbon effluent color.
7. Economics
Although unequal in overall performance, a high SA activated
sludge-powdered carbon system (ASU G, 72 mg/1 COD) approached the level of
granular carbon adsorption (carbon Series F, 64 mg/1 COD) to within 8 mg/1
COD at the 50th percentile point. Both systems would require extensive pre-
treatment and tertiary suspended solids removal. All other process compon-
ents being essentially equal, daily carbon usage costs were estimated for
theoretical plant flows of 1-5 MM gpd.
The cost of virgin powdered carbon (PC-H or PC-C) was estimated at
$0.30/lb and it was assumed that wasted carbon would be thrown away. To
calculate the equivalent powdered carbon dosage required for an 18-hr reten-
tion ASU it was assumed a 50-day SA would be maintained, giving an average
2 percent powdered carbon makeup. This was the equivalent of a 37.1 mg/1
89
-------
powdered carbon addition rate based on influent flow. A powdered carbon
feeder and storage facilities were included in powdered carbon costs using
Du Pont economicsl6 and applying the 0.6 rule. It was assumed that the
powdered carbon feeder could handle a 50 mg/1 addition rate.
Regenerated granular carbon adsorption costs were estimated, using
Brown and Root, Inc. economics,29 and converted to 1977 dollars. Daily gran-
ular carbon costs were estimated using 17.2 percent of the fixed investment
for operating and maintenance cost and 17.7 percent for depreciation. The
total daily costs for powdered carbon were estimated using the same per-
centage allowances.
Daily estimated carbon costs are shown in Figure 8 for theoretical
flows of 1-5 MM gpd by scaling up ASU G and carbon Series F carbon require-
ments. The cost effectiveness of the relatively simple process change of
adding powdered activated carbon to the activated sludge process can be
clearly seen. Estimated daily cost savings would range from $987/day at 1 MM
gpd flow to $2750/day at 5 MM gpd using high--surface-area powdered carbon
(PC-H) addition rather than granular carbon adsorption. The incremental cost
would be about $14.73 per pound of COD at 1 MM pgd (see Figure 9).
DISCUSSION
1. Increasing Sludge Age (SA)
Contrary to conventional activated sludge design techniques, the
increased SA did not result in sludge deflocculation, higher SVI, and high
effluent TSS. With the exception of a few days, the high SA control ASU B
easily achieved a high SA as a result of good pretreatment as proposed by
Grutsch and Mallatt.lO>H However, no enhanced performance was measured at
the high SA nor in the two-stage system with both reactors at a high SA in
Part I. Possibly more emphasis should be placed on the benefits of optimized
pretreatment than on increased SA as parallel activated sludge systems at
about the same SA (F/M = 0.3) were vastly different in performance at
contrasting degrees of pretreatment.
Increasing SA gave a reduction in biological solids production as
the conventional F/M control ASU F produced 0.19 Ib VSS/lb COD removed com-
pared with 0.16 for high SA control ASU B. This sludge production was not
quite as significant as it would have been if ASU F had operated at 5-10 days
SA where many conventional ASU's operate instead of about 14 days. Any re-
duction of biological solids production would help lower sludge treatment and
disposal costs.
The high effluent 50th percentile ammonia level of ASU A (11.5
mg/1) in Part II was probably the result of upset conditions which resulted
from filamentous sludge bulking causing the loss of biomass. The calculated
10-day SA of ASU A was only slightly less than ASU F (14 days) which produced
a 50th percentile ammonia level of 2.3 mg/1 in Part II. Although increasing
SA generally does improve nitrification, no conclusions could be drawn as to
its effect in this study. The conventional ASU F had already produced an
effluent that was about 90 percent nitrified.
90
-------
2. Activated Carbon Enhancement Mechanism
A 0.59-g TOC/g carbon loading was calculated in ASU G while operat-
ing with a 2500 mg/1 PC-H level. This extremely high "apparent" TOC loading,
as explained by Flynn,16 may be the result of continuous adsorption of slowly
biodegraded organics which are "biologically regenerated" from the carbon
many times over the biological and carbon SA. "Apparent" TOC loadings there-
fore increase with higher SA, optimizing the use of powdered carbon, until
the carbon becomes loaded with completely biorefractory organics. This
explanation of an "apparent" loading or enhancement mechanism appears logi-
cal; however, oxygen uptake and biological sludge production data presented
here negate biological regeneration in this study. ASU G not only had com-
parable oxygen uptake measurements with control ASU B, but it produced about
50 percent less biological solids. This implies that the actual enhancement
may have been predominantly due to adsorption on the high PC-H surface area.
Considering that PC-H had approximately five times the surface area of con-
ventional powdered carbons, such as PC-C, the expression of the TOC loading
as 0.12 g TOG/500 sq m of surface area would be more reasonable.
DeJohnl9,30 explains that granular carbon columns are sometimes
undersized because the designer uses virgin carbon and assumes that regen-
erated carbon will have the same activity. The thermal regeneration process
will enlarge some carbon pores reducing the surface area and decreasing the
adsorption of small molecules which are not so strongly adsorbed on larger
pores. Assuming that many small molecules require small powdered carbon
pores for moderately strong adsorption, PC-H may have been more effective
than PC-C because of pore size distribution, provided that the normally
biorefractory refinery organics were small molecules.
The mechanism of powdered carbon enhancement of the activaged
sludge process was not defined in this study and needs further investigation
in Phase II. Target SA's of the activated sludge-powdered carbon systems
were 50 days. Ideally, systems should be operated for periods of several
SA's to insure that equilibrium conditions have been reached and that the low
(2 percent) daily powdered carbon makeup rate will continue to give consis-
tent results.
The selection of the best powdered carbon for a particular acti-
vated sludge enhancement is not a simple task since powdered carbons vary in
their adsorptivity. Carbon isotherms performed on a refinery wastewater
would exhibit a wide variability and require a statistical analysis to select
the best powdered carbon. Isotherms would have to be performed on the acti-
vated sludge effluent (as in Phase II of this study) to determine enhancement
strictly due to adsorption.
The powdered activated carbon (PC-H) utilized with very good
enhancement results is not, as of yet, commercially available. Because of
the relatively high cost of granular carbon adsorption, other powdered
carbons at similar and higher operating levels would probably offer a signi-
ficant improvement in activated sludge performance and remain more cost
effective than granular carbon.
91
-------
3. Granular Carbon Adsorption
Granular carbon adsorption data indicate that the quality of end-
of-pipe refinery wastewater treatment depends on optimization of each treat-
ment step from primary to tertiary treatment. The use of equilibrium or re-
generated granular carbon in pilot studies will provide a more realistic data
base, recognizing that economics would favor regeneration for many potential
users.
The classical approach^1>32»33,34 for handling carbon adsorption
data to establish breakthrough curves was virtually useless in this study
because it assumes the carbon column influent has a single adsorbate. In the
calculation of accumulated TOC loadings at apparent breakthrough of carbon
columns there were several instances where organics were eluted in slugs from
carbons in both series. At times, phenols were two orders of magnitude high-
er than normal. This phenomenon is a very real problem and must be consid-
ered when establishing stringent effluent discharge guidelines for industry.
Even the best available technology, disregarding economics, has its
limitations.
SUMMARY
The EPA 1983 guidelines for the petroleum refining industry have
assumed that 1977-type technology must be upgraded by the addition of costly
systems, such as granular activated carbon adsorption. The results of this
API study indicate that, should the EPA adhere to the granular carbon techno-
logy originally proposed, it may be possible to achieve this level of treat-
ment technology by the much more cost-effective method of adding powdered
activated carbon to the 1977 activated sludge system.
Process modifications including optimized pretreatment and the
addition of a high-surface-area powdered activated carbon can be used to pro-
duce an effluent which is comparable in quality to that obtained by granular
carbon adsorption. Increasing activated sludge age from the conventional
mode of operation (about 10 days) to about 50 days did not give a significant
system improvement; however, in conjunction with powdered carbon addition,
high sludge age allowed higher equilibrium reactor concentrations (2500 mg/1)
at low (2 percent) carbon makeup rates. This benefit has been demonstrated
with the high-surface-area carbon and it is possible that it can also be
obtained with increased levels of conventional powdered carbon. The cost-
effectiveness of any powdered carbon will depend on the wastewater charac-
teristics and powdered carbon adsorptivity, which was greater for the high-
surface-area carbon (2462 sq m/g) than for the conventional-surface-area
carbon (550 sq m/g) investigated here. Even granular carbon adsorption was
found to have limitations as slugs of phenols were eluted, on occasion,
into the effluent.
ACKNOWLEDGEMENTS
This study was funded in part by the American Petroleum Institute,
Divison of Refining, CREC Liquid Waste Subcommittee.
92
-------
REFERENCES
1. Environmental Protection Agency, "Petroleum Refining Point Source
Category Effluent Guidelines and Standards," Federal Register Vol 38
(240) 34542 (December 14, 1973). '
2. Environmental Protection Agency, "Petroleum Refinery Point Source
Category Effluent Guidelines and Standards," Federal Register Vol 39
(91) 16560 (May 9, 1975). '
3. Environmental Protection Agency, "Petroleum Refining Point Source
Category Effluent Guidelines and Standards," Federal Register, Vol 40
(98) 21951 (May 20, 1975).
4. Jones Associates, "Effluent Limitations in the Petroleum Refining
Industry," Vol IB, Prepared for the Office of General Counsel, American
Petroleum Institute (January, 1976)
5. J. F. Grutsch and R. C. Mallatt, "Optimize the Effluent System - Part 1:
Activated Sludge Process," Hydrocarbon Processing, Vol 55 (3) 105 (1976).
6. J. F. Grutsch and R. C. Mallatt, "Optimize the Effluent System - Part 2:
Intermediate Treatment," Hydrocarbon Processing, Vol 55 (4) 213 (1976).
7. J. F. Grutsch and R. C. Mallatt, "Optimize the Effluent System - Part 3:
Electrochemistry of Destabilization," Hydrocarbon Processing, Vol 55 (5)
221 (1976).
8. J. F. Grutsch and R. C. Mallatt, "Optimize the Effluent System - Part 4:
Approach to Chemical Treatment," Hydrocarbon Processing, Vol 55 (6) 115
(1976).
9. J. F. Grutsch and R. C. Mallatt, "Optimize the Effluent System - Part 5:
Multimedia Filters," Hydrocarbon Processing, Vol 55 (7) 113 (1976).
10. J. F. Grutsch and R. C. Mallatt, "Optimize the Effluent System - Part 6:
Biochemistry of Activated Sludge Process," Hydrocarbon Processing,
Vol 55 (8) 137 (1976).
11. J. F. Grutsch and R. C. Mallatt, "A New Perspective on the Role of the
Activated Sludge Process and Ancillary Facilities," Presented at Joint
EPA-API-NPRA-TU Open Forum on Management of Petroleum Refinery Waste-
waters, Tulsa, Oklahoma (January 26-29, 1976).
93
-------
12. D. L. Ford and W. W. Eckenfelder, Jr., "Effect of Process Variables on
Sludge Floe Formation and Settling Characteristics," Journal Water
Pollution Control Federation, Vol 39 (11) 1850 (1969).
13. G. Grulich, D. G. Button, F. L. Robertaccio, and H. L. Glotzer, "Treat-
ment of Organic Chemicals Plant Wastewater with the Du Pont PACT
Process," Presented at AIChE National Meeting (February, 1972).
14. D. G. Button and F. L. Robertaccio, U. S. Patent 3,904,518 (September 9,
1975).
15. E. I. Du Pont DeNemours and Company, "Du Pont PACT Process," Technical
Bulletin.
16. B. P. Flynn and L. T. Barry, "Finding a Borne for the Carbon: Aerator
(Powdered) or Column (Granular)," Proceedings of the 31st Annual Purdue
Waste Conference (May 5, 1976).
17. B. P. Flynn "A Methodology for Comparing Powdered Activated Carbons for
Activated Sludge," Presented at 168th National Meeting, ACS, Div. of
Petroleum Chemistry, Symposium on Disposal of Wastes from Petroleum and
Petrochemical Refineries,(September 13, 1974).
18. B. P. Flynn, F. L. Robertaccio, and L. T. Barry, "Truth or Consequences:
Biological Fouling and Other Considerations in the Powdered Activated
Carbon - Activated Sludge System," Presented at 31st Annual Purdue Waste
Conference (May 5, 1976).
19. P. B. DeJohn and A. D. Adams, "Treatment of Oil Refining Wastewaters
with Granular and Powdered Activated Carbon," Proceedings of 30th Annual
Fardue Industrial Waste Conference (May 6, 1975).
20. A. D. Adams, "Powdered Carbon: Is It Really That Good?," Water and
Wastes Engineering, Vol 11 (3) B-8 (1974).
21. P. B. DeJohn and A. D. Adams, "Activated Carbon Improves Wastewater
Treatment," Bydrocarbon Processing, Vol 54 (10) 104 (1975).
22. A. B. Scaramelli and F. A. DiGiano, "Upgrading the Activated Sludge
System by Addition of Powdered Carbon," Water and Sewage Works, Vol 120
(9) 90 (1973).
23. A. E. Perrotti and C. A. Rodman, "Enhancement of Biological Waste Treat-
ment by Activated Carbon," Chemical Engineering Progress, Vol 69 (11)
63 (1973).
24. J. A. Rizzo, "Case Bistory: Use of Powdered Activated Carbon in an
Activated Sludge System," Presented at Joint EPA-API-NPRA-TU Open Forum
on Management of Petroleum Refinery Wastewaters (January 26-29, 1976).
25. ICI United States Inc., "Powdered Bydrodarco Activated Carbons Improve
Activated Sludge Treatment," Product Bulletin PC-4 (October, 1974).
94
-------
26. Amoco Research Corporation, Amoco Active Carbon Grade PX-21 Product
Information Sheet (May, 1976).
27. APHA, Standard Methods for the Examination of Water and Wastewater,
13th Ed., New York, New York (1971).
28. H. E. Klugh, Statistics; The Essentials for Research. 2nd Ed. , John
Wiley & Sons, Inc., New York, New York (1974).
29. Brown and Root, Inc., "Economics of Refinery Wastewater Treatment,"
American Petroleum Institute Publication No. 4199 (1973).
30. P. B. DeJohn, "Carbon from Lignite or Coal: Which is Better?,"
Chemical Engineering. Vol 82 (9) 113 (1975).
31. Metcalf and Eddy, Inc., Wastewater Engineering, McGraw-Hill, New York
New York (1972).
32. W. W. Eckenfelder, Jr., Industrial Water Pollution Control, McGraw-
Hill, New York, New York (1966).
33. H. J. Fornwalt and R. A. Hutchins, "Purifying Liquids with Activated
Carbon," Chemical Engineering, Vol 73 (8) 1976 (1966).
34. J. L. Rizzo and A. R. Shepherd, "Treating Industrial Wastewater With
Activated Carbon," Chemical Engineering, Vol 84 (1) 95 (1977).
BIOGRAPHY Leonard W. Crame
Leonard W. Crame is a Senior Chemical
Engineer in the Air and Water Conservation
Section of Texaco's Prot Arthur, Texas, Research
Laboratories. He has a B.S. degree in Engineer-
ing Technology (Chemical) and an M.S. in Thermal
and Environmental Engineering from Southern
Illinois University at Carbondale. Len has been
involved in several refinery wastewater treat-
ment pilot studies and conceptual designs since
joining Texaco in 1973. He is a member of
the Texas Water Pollution Control Association
and has recently authored several papers on
wastewater treatment.
95
-------
DISCUSSION
E. A. Buckley, Lion Oil Co.; What were the levels of alum and polyelectro-
lyte used in the DAF pretreatment to the system?
Len Crame: In the dissolved air flotation unit we were using, we found
by the use of the zeta meter and jar tests that it required 40 milligrams
per liter of filter alum, 20 milligrams to 40 milligrams per liter of catonic
polymer Dearborn 431. It was not the intent of this study to try to zero in
on the best and most economical chemical dosage but mainly to get the soluble
feed for the bio units. We found from experience that wide fluctuations in
the feed characteristics did not affect these two chemical doses.
Ed Sebesta, Brown & Root: In the slide (Figure 2) comparing effluent and
COD concentrations from the various pretreatment systems, were the samples
filtered or unfiltered?
Len Crame; They were filtered COD's for our bio effluents and carbon
effluents. In Figure 2, I did not identify them, but it is total COD. It
does include suspended solids because we were looking for the contribution
of solids in this case.
I also would like to make the comment that I do pretty much agree
with everyone else's presentations as far as what work has been done with
carbon on the enhancement mechanism and we will continue to look at this
throughout the second phase of our pilot study. I think that you have to be
very careful in using powdered activated carbon. In a short term study I
agree with the other gentlemen (Amoco)(DuPont), that when you first put in
activated carbon you have to allow time for this matrix to form, which we
did. And you don't get the same settling effect as when you initially add
carbon. When you allow the system to come to an equilibrium and the bio-mass
starts adhering to the powdered carbon, it does greatly improve the sludge
settling characteristics, but it takes a little time. I believe that when
you initially add powdered carbon your results are going to be very good
because you're going to get a tremendous amount of adsorption. We followed
this and have seen it. I'm very hesitant about including data right after
you start running an enhanced bio-system. You will see a sharp decrease in
the effluent organic levels. You have got to wait until an equilibrium is
reached.
J. Dewell, Phillips Petroleum Co.; In your cost comparison between enhanced
activated sludge and the granular carbon, I wasn't sure if the enhanced
activated sludge assumed that the conventional activated sludge was already
in place or not. Would these comparisons still be valid on a grass-roots
treating system?
Len Crame; We were assuming that activated sludge was already in place and
actually we were only comparing the cost of carbon contactors and regenera-
tion equipment against the additional equipment you have to put in to add
powdered carbon. We were not including filters. We would believe the filter
would have to be a part of both treatment systems and would have no effect
on this comparison.
96
-------
J. Dewell; Do you have any feel for what would be most cost effective on a
grass-roots basis, assuming no treatment system at all exists?
Len Crame: I think it is a safe argument that activated sludge is probably
the best system for driving the effluent organic levels down for 1977 and
1983 but not necessarily the most cost effective. At this time I do not see
any other system that perhaps can be enhanced cost effectively with powdered
carbon for 1983.
J. Dewell; I was referring to a situation where one does not have activated
sludge at this time and is meeting 1977 standards so one is going to come to
1983 without any activated sludge.
Len Crame; I think this must be determined on an individual basis. As you
know when you calculate out guidelines for '77 or '83 you will find that in
some cases you are stuck with very tight guidelines for a certain parameter
and I don't think it is appropriate to say which type of treatment would be
best. We would definitely not put in any powdered-carbon enhanced system
until we piloted it and you would be taking a risk if you did. All treat-
ment systems are unique, including activated sludge systems and enhanced
biological systems. We think that powdered carbon addition has a lot of
merit, but still you should determine it on a case-by-case basis.
F. L. Robertaccio, DuPont: I think that the easiest way to look at it is
that the activated sludge system in this case is common to both the powdered-
carbon addition and the granular-carbon addition system so the difference in
cost here would have added to it the cost of the activated sludge system if
you were starting out with a brand new plant. You can use that as a first
estimate, but what we have found at the plant I talked about yesterday is
that with a grass-roots plant you have additional savings that you can accrue
to take full benefit of the system. We talked about having smaller secondary
clarifiers, higher upflow rates through the clarifiers, smaller dewatering
equipment; and having no secondary solids disposal if you go through regenera-
tion. So our experiences have been that with the grass-roots system you can
put in powdered carbon systems with regeneration for the same capital costs
and essentially the same operating costs as a conventional secondary waste-
water treatment system. If you want further reference on this there was a
paper I referred to yesterday that had details of those cost estimates.
J. E. Rucker, API; Please comment on why your COD values were greater than
those we looked at earlier this morning from the Argonne work?
Len Crame; The refinery where we were located is a very complex refinery
and I am quite sure that the refractory COD that remains is going to differ
from plant to plant. We did try to exclude everything from the chemical
plant, but I am not surprised really that we have a different refractory
COD level and I don't think you can compare the refractory COD's out of
these carbon systems from plant to plant and find a great consistency as far
as concentration goes.
97
-------
Jeffrey Chen, Dravo; What would you propose to use to treat the sludge
generated from the pretreatment unit? Will the cost associated with the
treatment be cost effective when compared to the improvement of the following
bio system?
Len Crame; We were assuming that for our best case here that when we were
comparing granular carbon with powdered carbon you would have a primary
sludge treatment and disposal problem in both cases so that it really
doesn't affect our economics here. Sludge disposal is another problem and
again it does depend upon the availability of land and other considerations
and it is just something totally different; but actually we're comparing the
two systems here and assuming that primary sludge is going 'to be a problem
in both cases. You would have to do a cost effectiveness study on the pre-
treatment and sludge disposal cost vs the benefit obtained from it. But
from an operational standpoint, once you get the colloids and oil out it is
much easier to operate the activated sludge process, since the oil and solids
interfere with flocculation and sludge settling.
'\
Tom McConomy, Calgon Corp.; During the period you were operating the granu-
lar carbon columns, was the carbon changed or was the same carbon used during
the entire test?
Len Crame; As you will see in the paper we had four carbon columns and we
would measure TOG at intermediate points and whenever we found a breakthrough
on the first carbon column, we would shift the carbon columns and put a fresh
regenerated column on the talil-end of the system. This is why we are confi-
dent that the final column effluent is representative of what carbon adsorp-
tion can do with the regenerated carbon. We did try to determine how much
we had in those columns and we were running about 0.12 to 0.15 pounds TOG
per pound of carbon, which I think is fairly typical. But because of the
activated sludge enhancement you get with powdered carbon, where I won't
necessarily say "biological regeneration" occurs, you are effectively regen-
erating it somehow by .desoprtion, or whatever, within the process and that
is what makes it economical.
-------
FILTER
PH
CONTROL
ACTIVATED
SLUDGE
B
CONV.
0.3 F/M
HIGH SA
+
PC-C
HIGH SA
PC-H
CONV.
0.3 F/M
HIGH SA
PC-H
FILTER
CARBON COLUMNS
HIGH
SA
HIGH
SA
FILTER
CARBON COLUMNS
FIGURE 1 - TREATMENT SCHEMES
-------
600
500
I 40°
S
,: 300
o
H
w
PL,
W
200
100
1000
900
800
700
i 600
i 500
4
- 400
I r
I I
!S^^/QJ2^
Q
O
w
w
300
200
100
*DESIGNED FOR LESS THAN OPTIMUM PRETREATMENT
I I i i i
i i i
0 70 80
95
5 10 20 30 40 50 60 70 80 0
PERCENT OF TIME LESS THAN INDICATED VALUE
FIGURE 2 - PART I - TOG AND COD REMOVAL BY PRETREATMENT
98
100
-------
go
wo
W
Q
W
Pd
W
H
H -
SO
WO
3°
w
H
iJ
M
fn
100
90 -
80 -
70 _
60 -
50 -
40 -
30 -
300
.j
u 200
100
90
80
70
60
50
50TH PERCENTILE VALUES, MG/L
A, FILTER - LOW SA
B,
C,
D,
E,
F,
DAF
DAF
DAF
DAF
DAF
HIGH SA
HIGH SA-500 MG/L PC-C
HIGH SA-500 MG/L PC-H
HIGH SA-STAGE 2
LOW SA
TOC
6T~
56
55
48
53
58
COD
128
108
97
89
116
105
- F
J_
10
20
30 40 50 60 70
80
90
I
95
98
PERCENT OF TIME LESS THAN INDICATED VALUE
FIGURE 3 - PART I - ACTIVATED SLUDGE TOC AND COD DISTRIBUTIONS .
-------
200
©
8lOO
O 90
w 80
w 70
5 60
E 50
S 40
w
30
20
50
40
30
§8 20
o
o
H
7
6
5
4
3
I
I
I
I
I
I
I
I
95
2 5 10 20 30 40 50 60 70 80 90
PERCENT OF TIME LESS THAN INDICATED VALUE
FIGURE 4 - PART I - GRANULAR CARBON COLUMN TOG AND COD DISTRIBUTIONS
93
102
-------
u
g
w
a
700,
600
500
400
300
200
80
70
60
T P
334
*DESIGNED FOR LESS THAN OPTIMUM
il III i i I
1000
90C,
800L
PERCENT OF TIME LESS THAN INDICATED VALUE
±—d
FIGURE 5 - PART II - TOC AND COD REMOVAL BY PREtREATMEST
103
-------
100
90
•J 80
o 70
s 60
50
w 30
5
M
ft.
,. 20
10
^300
52
,200
Q
O
I
^ 90
L 80
^ 70
w
60
3 50
fc
40
1 f
50TH PERCENTILE VALUES, MG/L
A, FILTER - LOW SA
TOC
DAF
DAF
DAF
DAF
DAF
HIGH SA
HIGH SA-1000 MG/L PC-C
HIGH SA-1000 MG/L PC-H
LOW SA*
HIGH SA-2500 MG/L PC-H
4A
41
34
46
22
COD
T6T
111
119
89
108
72
10 20 30 40 50 60 70 80 9
PERCENT OF TIME LESS THAN INDICATED VALUE
FIGURE 6 - PART II - ACTIVATED SLUDGE TOC AND COD DISTRIBUTIONS
98
104
-------
100
90!
80
70
^ eo
S 50
so 40
wo
feQ 30
w
20
10
200
| §100
£§80
wg 70
H 60
£ 50
40
r I i i
2 5 IB 2153154^"
PERCENT OF TIME LESS THAN INDICATED VALUE
FIGURE 7 - PART II - GRANULAR CARBON TOG AND COD DISTRIBUTIONS
105
-------
W
H
O
H
3500
3000
2500
2000
£ 1500
o
1000
500
0
\
GRANULAR CARBON
POWDERED CARBON
I
12345
FLOW MM GPD
FIGURE 8 - COMPARISON OF ESTIMATED CARBON COSTS
106
-------
50
40 __
30 —
20 _
10 —
0
i
—
36
$0.61
T "D Jf"*flyr\
JjJJ L»UJJ
POWDERED
CARBON
44
$3.19
LB COD
GRANULAR
CARBON
8
$14.73
LB COD
INCREMENTAL
IMPROVEMENT
FIGURE 9 - ESTIMATED EFFECTIVE CARBON. COST AT 1 MM GPD
107
-------
PHENOLS
TABLE 1- PART II - EFFLUENT SUMMARY
(ALL VALUES MG/L)
OIL & GREASE TSS
AMMONIA
EFFLUENT
SAMPLE
EQ BASIN
FILTER
DAF UNIT
ASU A
ASU B
ASU C
ASU D
ASU G
ASU F v
CARBON COL.
(SERIES A)
CARBON COL.
(SERIES F)
PERCENTILE
50TH 90TH
7.3
—
—
0.11
0.06
0.15
0.05
0.04
0.07
0.04
0.02
18.0
—
—
0.16
0.14
0.38
0.20
0.13
0.17
4.8
0.08
PERCENTILE
50TH 90TH
108
69
14
18
5
3
*3
< 3
4
<3
<3
191
130
19
38
7
9
5
3
7
6
4
PERCENTILE
50TH 90TH
64.0
34.0
13.0
86.0
27.5
23.0
18.0
22.0
8.4
—
—
119
74.0
21.0
149
41.0
77.0
57.0
44.0
28.0
—
—
PERCENTILE
50TH 90TH
20.9 27.2
—
—
11.5 20.0
3.9 9.2
3.3 5.4
3.1 4.8
3.1 4.4
2.3 5.4
—
—
-------
TABLE 2 -PART II - SLUDGE DATA
AVERAGE VALUE
ACTIVATED SLUDGE UNIT
-
NOMINAL LOADING.
ACTUAL LOADING
MLSS, MC/L
HLVSS, MG/L
PC, MG/L
% V3S
OXYGEN UPTAKE, LS 02/LB COD REM
MG/L-MIN
SVI, ML/G
SETTLING VELOCITY, FT/MIN
BIOMASS PRODUCTION RATEd
TOTAL PRODUCTION RATES
A
F/M =0.3
F/M =0.3
1,487
1,302
0
88
0.40
0.16
95
0.17
0.22
0.25
(J
50-DAY SA
39-DAY SA
1,892
X.562
0
83
0.68
0.12
64
0.34
0.16
0.19
C
50-DAY SA
42-DAY SA
2,728
2,269
l,000a
83
0.71
0.12
41
0.38
0.12
0.17
D
50-DAY SA
44-DAY SA
2,720
2j416
l,000b
89
0.49
0.11
43
0.38
0.11
0.14
P
F/M = 0.3
F/M * 0.3
745
689
0
92
0.61
0.11
91
N/AC
0.19
0.21
G
50-DAY SA
56-DAY SA
4,096
3,898
2,500b
95
0.47
0.10
30
0.39
0.08
0.09
aCONVENTIONAL-SURFACE-AREA.
bHIGH-SURFACE-AREA.
CDISCRETE SETTLING
dLB VSS/LB COD REMOVED.
eLB TSS/LB COD REMOVED (INCLUDES CARBON)
-------
TABLE 3 - POWDERED CARBON (PC) REQUIREMENTS
ASU-PART
C-I
D-I
C-II
D-II
G-II
PC
LEVEL
(MG/L)
500
500
1,000
1,000
2,500
PC
TYPE
PC-C
PC-H
PC-C
PC-H
PC-H
PC
INVENTORY
(G)
28.4
28.4
56.7
56.7
141.9
AVG PC
LOSS
(G/DAY)
0.68
0.56
1.50
1.12
2.21
PC
MAKEUP
(%)
2.4
2.0
2.6
2.0
1.6
110
-------
"PILOT PLANT ACTIVATED CARBON TREATMENT OF
PETROLEUM REFINERY WASTEWATER"
Jack H. Hale, Leon H. Myers, and Thomas E. Short, Jr.
Robert S. Kerr Environmental Research Laboratory
INTRODUCTION
One of the first documented uses of carbon for environmental control was in 1793
when a physician used charcoal to remove the odor associated with gangrene. About
sixty years later a scientist by the name of Stenhouse recommended the use of charcoal to
remove the odors from sewers. Potable water was "purified" by carbon in 1862 J Moon-
shiners in the hills of Tennessee were using charcoal, not to mellow their product, but to
remove "hog track" odors.
Since the mid-1960's, there has been an increasing effort to utilize activated carbon
as a secondary or tertiary treatment system to treat wastewater. Two of the most success-
ful uses of activated carbon were demonstrated at Lake Tahoe and in Pomona, California,
to treat domestic municipal wastewaters. It is, of course, a natural progression to treat
industrial wastewaters with systems that appear successful in treating municipal waste-
waters.
The first domestic petroleum refinery to use activated carbon treatment was the
Atlantic Richfield Refinery in Carson, California. The system was designed for inter-
mittent use to treat rainfall runoff and process wastewater during storm periods. The
second application was designed for the BP Refinery, Marcus Hook, Pennsylvania, to
treat process wastewaters. Unlike the Arco system, the BP system was designed to operate
in a continuous mode. Neither system relied on biological treatment preceding the
activated carbon system.
During the late 1960's, activated carbon treatment of industrial wastewater was
gaining momentum as the treatment system of the future. Statements such as "organic
removal" or "removal of dissolved organics from wastewater has been demonstrated" and
"activated carbon treatment systems will remove dissolved organic contaminants" were
prevalent. These innocent gross statements were translated rather rapidly to imply that if
you had an organic waste treatment problem, it could be solved with granular activated
carbon.
With the complete cooperation of Kerr-McGee Petroleum Refinery, Wynnewood,
Oklahoma, a study program was devised to investigate the adsorptive capacity of
activated carbon.
ISOTHERM STUDIES
Adsorption isotherms are used to indicate the effectiveness of an activated carbon
for a specific wastewater under controlled conditions. When the isotherm is to be used as
111
-------
a predictive tool, the wastewater should be as representative as possible, pH and
temperature adjustments should not be made, and sample collections need to be made in
glass containers.
CARBON EVALUATION
Eleven commercially available carbons were evaluated for this petroleum refinery's
wastewater. It is necessary to point out that adsorption varies with the carbon being
evaluated and the water sample. The data presented is based on specific activated carbon
for one petroleum refinery. Comparable results may or may not be achieved using the
same carbon at another refinery.
The controlled conditions used for comparison purposes of the eleven carbons are
shown in Table 1. Both pulverized and granular modes were evaluated.
The waste sample to be evaluated was collected from the API separator discharge,
settled for four hours, and the candidate water was drawn from the middle of the vessel.
The adsorption capacity of one gram sample of the candidate carbons, both pulverized
and granular, is shown in Table 2.
Graphically expressed, the comparison of pulverized and granular adsorption
capacities appears in Figure 1 .
CONTACT TIME
A major factor relating to adsorption is the amount of time the water to be treated is
in contact with the activated carbon. A sample of wastewater from another refinery was
obtained for these studies. Pulverized activated carbon (Filtrasorb 400) was weighed to
obtain 0.1, 0.5, 1.0, 2.0, and 5.0 grams of sample, and the isotherm procedure was
followed for three hours. Samples were obtained at 20, 40, 60, and 180 minutes,
filtered, and analyzed for total organic carbon. Table 3 represents the typical results.
ISOTHERM TESTING
Figure 2 shows the results of an isotherm carried out on wastewater from the refinery
used in this study and a candidate activated carbon. Obviously, this curve does not
follow the typical "Freundlich" isotherm. Instead of the usual straight line, a curve
resulted. At the lower end of the curve the amount of carbon added to the wastewater is
increased but the concentration of TOC does not decrease as much as would be expected.
In fact, the concentration of TOC changes only very slightly at high carbon dosage.
Refinery wastewater is composed of a rather complex mixture of materials. Some of these
materials are readily adsorbable while others are not. The initial amount of carbon picks
up the easily adsorbed materials and leaves behind the others. Therefore, increasing the
carbon dosage does not reduce the concentration of TOC to the same degree. In fact,
there is a possibility that this wastewater contains non-adsorbable components that cannot
be removed by carbon treatment.
112
-------
From the "isotherm'/1 the adsorptive capacity of the carbon, tested with this
specific wastewater, projects to approximately 0.1 Ib of TOC per ib of carbon. This was
considered to be a reasonable value and indicated that further investigation of activated
carbon treatment was justified.
Results from isotherm tests are considered to be valuable only from an initial
screening standpoint. Their results are generally not sufficient for predicting results of
full-scale granular activated carbon treatment. For a better evaluation of activated
carbon treatment, continuous column pilot plant investigation at the waste stream site
yields much more reliable results.
ANTHRAFILT FILTRATION
When treating oily wastes by activated carbon, it appears that a pretreatment system
is needed to assist in the removal of insoluble oils and suspended matter. An anthrafilt
downflow filter system was evaluated for effectiveness, again using filtrable TOC as the
prime parameter. The glass column used for this exercise was five feet in length and 1.5
inches in diameter. A diagram of the system is shown in Figure 3. Total organic carbon
results obtained at these time intervals are shown in Table 4.
This particular study lasted 2.5 hours before a pressure drop was noted. The charge
water contained an appreciable quantity of oil, and the oil percolated into the anthrafilt
layer during the study period; When the column was backwashed with distilled water,
the percolated oil was easily removed from the anthrafilt, leaving an apparently clean
bed.
MINI-COLUMN STUDY
A downflow mini-column system was designed using stock one inch ID glass tubing
that was six feet in length. The system was designed with an electronic sampler to
composite hourly samples. A diagram of this system is shown in Figure 4.
The primary study involved dividing the six columns into two sets of three columns
each to evaluate comparison of treatment effectiveness. Each column was packed with
1,000 cc of Filtrasorb 400 using potable water was the wetting liquid. The flow rate was
set at 400 ml/minute and a column pressure of 6-7 psig was used. Total organic carbon
was the primary parameter used for treatment effectiveness. Analytical results of the
primary study are recorded in Table 5.
Comparing each column for removal efficiency, using the anthrafilt effluent TOC as
the base, the percentages obtained are shown in Table 6. Columns 1 and 2 compare very
well for the five-day averages, although they do exhibit considerable daily variances.
Columns A-3 and B-3 are not agreeable, and no reason for this deviation is projected.
A second study was conducted by connecting five columns in an upflow mode. Each
column contained 1,000 cc of Filtrasorb 400 packed in the same manner as the previous
study. TOC results of this study are shown in Table 7.
113
-------
Percentage reductions of each column based on the anthrafilt effluent are shown in
Table 8. The average column trend indicates there is an increase of treatment effective-
ness with each succeeding column, as one would expect.
PETROLEUM REFINERY DATA
Listed in Table 9 is the pertinent information needed to characterize the refinery
where the study was conducted. This refinery was a 30,000 barrel, Class "B" refinery
(using the API classification system); processes included fluid cat cracking, HF alkylation,
catalytic reforming, and asphalt production.
The existing refinery eastewater treatment system is shown in Figure 5. Wastewater
from the refinery includes cooling tower blowdown, boiler blowdown, oily process water,
stripper effluent, and contaminated runoff. These wastes flow to an API separator for oil
removal. The effluent from the separator is treated in a "Pasveer Ditch" activated sludge
treatment system. The biologically treated effluent then flows through a series of holding
ponds and the effluent from the final pond is discharged to a small stream.
Two complete pilot plants were installed and operated simultaneously, one on the
refinery's API separator effluent (secondary) and the other on the clarifier effluent from
the biological treatment system (tertiary),
ACTIVATED CARBON TREATMENT PILOT PLANT
Figure 6 contains a flow diagram of the pilot activated carbon treatment systems.
The wastewater to be treated first flows through a dual-media filter constructed of 4-inch
PVC pipe. This filter consisted of an 18-inch layer of sand over pea gravel, topped with
a 6-inch layer of anthrafilt.
Dual-media filtration pretreatment was chosen because of its reliability and effective-
ness, and because it does not require the use of iron or aluminum salts as coagulants.
This latter point is particularly important insofar as regeneration of activated carbon is
concerned, since aluminum salts during regeneration can remain on the surface of the
carbon at high temperatures. These salts can become permanently attached to the surface.
Thus, the effective surface area of the carbon is reduced and its adsorption capacity is
seriously reduced. Iron salts present a similar problem. In addition, these salts can
catalyze oxidation reactions of the carbon and the gases in the regenerator. Thus, the
structure of the carbon becomes permanently damaged.
After pretreatment, the wastewater entered a "Calgon" activated carbon pilot plant.
This plant was set up so that the wastewater flowed down three of the 5-inch ID columns.
The first column contained an 18-inch layer of granular activated carbon while the
remaining columns had a 36-inch layer of carbon.
The flow rate through each pilot plant was adjusted to 1/4 gpm. During the
operation of the pilot plant, samples of the API separator effluent, biological treatment
effluent, and both pilot plants' effluents were taken every two hours. These samples were
114
-------
composited and preserved according to recommended EPA methods. Twenty-four hour
composite samples were analyzed daily for a spectrum of water pollution control
parameters using EPA analytical methodology and analytical quality control techniques.
The dual-media filter and carbon columns were backwashed whenever the pressure in
the first column exceeded 20 psi. The pilot plants were operated over a 10-day period,
at the end of which time the first columns in both plants were near exhaustion.
Figure 7 shows the BOD5 daily composite analysis for the API separator effluent
before and after treatment by the various schemes studied. Activated carbon treatment
was not able to achieve the same level of BOD5 reduction as bio-treatment. In fact,
bio-treatment did a considerably better job. However, carbon treatment following bio-
treatment did show further reductions in BOD.
Figure 8 shows the COD daily composite results for the same wastewater streams.
Apparently, both bio-treatment and carbon treatment effect about the same COD
reductions. Carbon treatment following bio-treatment yielded the best reduction of COD.
The TOC results for the daily composites are shown in Figure 9. Unlike the BOD5
and COD results, the TOC indicates increased removal can be obtained by the API-
carbon combination over the bio-treatment system alone. There is an erratic behavior
pattern which indicates unadsorbable organics passing through the carbon columns.
From the BOD5 and COD plots, it may be concluded that for secondary treatment
alone, on the wastewater studied, the biological treatment system is preferable because
it gives the greatest BOD^ reduction and gives COD reduction equivalent to carbon
treatment. The best levels of reduction were obtained with biological treatment followed
by carbon adsorption.
Table 10 gives the results of other parameters evaluated in this study. As would be
expected, both biological and activated carbon treatments are able to produce signifi-
cant reduction in the organic parameters, such as BOD, COD, TOC, oil and grease, and
phenols. It should be noted that carbon has a particular affinity for the removal of
phenols. The color and turbidity are also improved. Cyanides and ammonia, for all
practical purposes, were not removed by either of these treatments.
Sulfides are a peculiar problem for activated carbon treatment. Apparently, the
sulfide content increases as it goes through the carbon system. This is probably due to
biological activity occurring at the surface of the carbon. The sulfur containing
compounds which are adsorbed upon the carbon are reduced anaerobically to r^S giving
rise to the increase in sulfide content. Bio-treatment, on the other hand, decreases the
sulfide content very significantly.
One of the least expected removals by activated carbon was observed for the metals
—chromium, copper, iron, and zinc. These rnetals are significantly reduced by the
cqrbon. Whether these removals are due to adsorption, filtration, or other phenomena
has not been evaluated. However, it must be kept in mind that these removals were
115
-------
accomplished with virgin carbon. The effect of/some of these metals upon carbon
regeneration has not been determined. Also, whether regenerated carbon can remove
equivalent amounts of metals has not been established.
As mentioned previously, the isotherm study indicated that the carbon could adsorb
0.1 Ib TOC/lb of carbon. The columns on the API separator e/fluent indicated a capacity
of 0.31 Ib TOC/lb of carbon, as shown in Table 11. The most probable explanation for
the difference in the capacity as determined in the isotherm test and the pilot plant study
is biological activity in the columns.
In the case of the API separator effluent carbon column, this biological activity
manifested itself by the anaerobic production of h^S. While the increased capacity is a
desirable situation, the production of sulfides is not. Although anaerobic activity was
not apparent in the clarifier effluent columns, it was observed that algae started growing
in the carbon beds. In fact, it was necessary to cover the columns to prevent sunlight
from making this growth possible.
Because of the considerable cost of granular activated carbon, regeneration of the
spent carbon is essential if the cost of treatment is to be kept at a moderate level.
Samples of spent carbon were taken from the first column of each pilot plant and sent to
Calgon Corporation for regeneration studies. Calgon regenerated these two carbons using
their standard evaluation techniques in a muffle furnace at 1750° F and with a normal air,
flue gas, and steam atmosphere.
As shown in Table 12, both carbon samples were regenerated to a good activity as
indicated by the iodine numbers obtained. The regenerated clarifier carbon appeared to
have a little better activity than the API separator carbon. In general, the iodine number
can be related to the surface area of pores larger than 10 A in diameter. The molasses
number, likewise, is related to the surface of pores larger than about 30 A. As can be
observed in Table 15, there is an increase in the molasses number with a slight decrease
in iodine number. Thus indicates a shift in pore size distribution. Apparently, the
smaller pores are being destroyed while larger pores are being created. Final clarifier
carbon appeared to have less destruction of the pore matrix than did the API separator
carbon.
GAS CHROMATOGRAPHY STUDIES
Studies by Dr. T. C. Dorris and Dr. S. L. Burks at Oklahoma State University have
shown that some organics present in refinery wastewaters are toxic to fish, and some
organics are refractory to biological degradation. These organic chemicals persist in
lakes and streams for long periods of time. It is also noted that a refractory organic can-
not be detected by the TOD test that has been used in the past to evaluate the efficiency
of treatment systems. Tests such as TOD, TOC, and COD are much more realistic for the
indication of organics in a water sample. Of these tests, TOC is the one that measures
only organics.
The laboratory researchers must go further and use sophisticated methods for actually
identifying and characterizing the refractory organics so that engineers will know the
116
-------
type^of organics that must be removed, thus designing treatment systems to improve the
quality of water discharged to the receiving streams.
Carbon treatment is a relatively efficient method for removing organics from refinery
wastewater and may be used as a secondary or tertiary treatment system to reduce the BOD
of effluent water to a very low level. In studies to learn more about carbon treatment,,
carbon columns were installed at a petroleum refinery wastewater treatment plant at a
point immediately after the API separator and at a point after the'final clarifier for the
activated sludge treatment unit. The carbon removed from these columns was extracted
with several organic solvents to determine which solvents were most efficient and to
provide samples of the organics for silica gel, gas chromatographic (GC), infrared, and
GC-mass spectrometric analyses.
v
About 100 ml of an air dried carbon sample were weighed and placed directly into
the soxhlet extraction apparatus with a glass wool plug at the bottom to prevent the
carbon from getting into the siphon tube. Three hundred ml of solvent were used in the
flask. The rate of solvent reflux was adjusted to give four cycles per hour, and the
extraction was continued for 22 hours. After 22 hours, the solvent-oil mixture was
cooled to room temperature and filtered through a 10 cm depth of anhydrous sodium
sulfate crystals contained in a 20 mm by 25 cm glass tube to dry the mixture and remove
any carbon particles that might be present. The dried solvent-oil mixture was then
distilled through a 20 mm diameter column containing 12 cm of glass beads until the
volume remaining in the flask was slightly less than 50 ml. The concenfrated solvent-oil
mixture was measured into a sample vial, and the volume adjusted to exactly 50 ml with
rinsings from the flask.
A gas chromatograph, with a flame detector and an electronic integrator, was used
to determine the amount of oil extracted with each solvent. A 1/8-inch by 20-foot
stainless steel column, packed with 70/80 mesh Chromosorb G, AW-DMCS, with 5%
OV-101 stationary phase, was programmed to hold at room temperature for four minutes
with the oven lid open, then the lid was closed and the program proceeded from 35° to
65° C at 8 degrees per minute and from 65° to 350° C at 6 degrees per minute. The
temperature was held at 350° C for up to 20 minutes, if it appeared necessary to elute all
of the sample. Injector and detector temperatures were about 370° C.
A solution of 3 gm of diesel fuel diluted to 50 ml with solvent was used as a standard
since it was available, although an oil with a higher aromatic content would have been
more desirable. The total area of the chromatogram of the standard, excluding the
solvent peak and correcting for baseline drift near the end of the chromatogram, was
compared with the total areas obtained from the 50 ml samples of extract. Data obtained
from the carbon extractions and GC analyses are shown in Table 13.
Chloroform, benzene, and methylene chloride showed about the same efficiency for
extracting oil from the carbon, while hexane and methanol were less efficient. Figures
10 through 14 show the chromatograms obtained on the extracts from API carbon using the
five different solvents. Comparison of the benzene, methylene chloride, and chloroform
extracts (Figures 10, 11, and 14) shows that the oils extracted with these solvents were
i
117
-------
almost identical in composition. However, chloroform showed an advantage in the
recovery of one low boiling compound which was an appreciable amount of the sample as
indicated from the size of the peak. This peak had a retention time of 11.3 minutes
which corresponds to the retention time of benzene. This peak could not be seen when
benzene was used as the solvent, and shows at a much lower concentration in the
methylene chloride extract.
Chloroform was used as the solvent for extracting a sample of carbon that had been
used to treat effluent water from the final clarifier of the activated sludge treatment
system. This carbon is referred to as FC carbon, and the resultant chromatogram is shown
in Figure 15. The FC carbon contained about 1/3 as much oil as the API carbon. A
sample of new carbon also was extracted with chloroform to determine whether the new
carbon contained any extractable material. This extract contained no measurable oil, as
illustrated by Figure 16. This figure also serves as a good illustration of the purity of the
chloroform used as the solvent. Figure 17 is a chromatogram of the standard mixture used
to quantitate area data in terms of grams of oil extracted.
The API carbon extract and FC carbon extract were each separated into three
fractions by selective elution and desorption from silica gel. The columns were pre-wet
with hexane and as the last drop of hexane disappeared into the surface of the gel, the
sample-gel mixtures in the 10 ml beakers were emptied into their respective columns.
The samples were then fractionated by eluting the saturates with 10 ml of hexane, the
aromatics with 10 ml of benzene, and finally desorbing the polar fraction with methanol.
Each of the fractions was then reduced in volume to 0.5 ml by evaporating the solvents
with a gentle stream of air.
Gas chromatography was used to obtain an estimate of the amount of saturates,
aromatics, and polar material in each of the samples. The oil extracted from the FC
carbon contained such small amounts of saturates and aromatics that it was necessary to
increase the sensitivity of the GC by a factor of 10 to obtain measurable areas. These
areas were then divided by 10 to get back to the same basis as the area measurements on
the polar fraction. The results are tabulated in Table 14.
These data indicate that the activated sludge treatment system has reduced the
saturates and aromatics to very low levels in the final clarifier effluent but has been
relatively ineffective in removing some of the polar type organic compounds.
Figures 18, 20, and 22 are chromatograms of the saturate, aromatic, and polar
fractions obtained by the silica gel separation of the oil from API carbon. Figures 19, 21,
and 23 are chromatograms of the saturate, aromatic, and polar fractions obtained by the
silica gel separation of the oil from FC carbon.
There was insufficient time for complete GC-MS identifications, but Table 15 lists
some of the compounds and types that have been found to date in the saturate, aromatic,
and polar fractions. The organic structures shown in the table are those normally found in
petroleum and petroleum products, but it is noted that three sulfur compounds were found
in the aromatic fraction of the FC carbon extract. It may be that these compounds are not
118
-------
biologically degraded as much as the aromatics and are a more significant part of the
aromatic fraction after biological treatment.
The polar fractions could not be analyzed by GC-MS because of the inability of the
gas chromatograph to separate the complex mixture. The complexity of the mixture is
illustrated in Figures 22 and 23 by the lack of baseline separation. Differences in the
two chromatograms indicate that some of the components have been degraded in the
activated sludge treatment system. It is probable that the reduction in phenolic com-
pounds, that js known to occur, would account for some of the differences noted in the
chromatograms.
Infrared scans on the polar fractions were almost identical. Functional groups noted
were OH and 0=0, and CH2 and CH3 adsorption bands indicate aliphatic type structures.
Further work must be done on the polar fractions to identify a number of the individual
compounds in the mixture so that an assessment of their toxicity can be made and treat-
ment methods developed to remove these compounds.
REFERENCES
1. Hansler, John W. Activated Carbon. Chemical Publishing Company, Inc.,
New York, New York, 1963.
2. Burks, S. L., T. C. Dorris, and G. L. Walker. Identification of Toxic Organic
Chemical Compounds in Oil Refinery Wastewaters. Technical Completion Report to
U.S.D.I. Office of Water Resources Research, Project BO-17, Oklahoma, 1970.
Unpublished.
3. Hale, J. H. and L. H. Myers. The Organics Removed by Carbon Treatment of
Refinery Wastewater. Presented at the Oklahoma Industrial Waste and Advanced
Water Conference, Stillwater, Oklahoma, April 1973. Roberts. Kerr Environmental
Research Laboratory, Ada, Oklahoma. Unpublished.
BIOGRAPHIES
Jack H. Hale holds a BS in Chemical Engineering from
Oklahoma State University. He is currently a research
chemist in the Industrial Section of the Source Management
Branch at the Robert S. Kerr Environmental Research
Laboratory, Ada, Oklahoma.
Leon H. Myers holds a BS in chemistry/biology from
Southwestern Oklahoma State University and a MS in
sanitary science from Oklahoma University. He is
currently Chief, Industrial Section of the Source Manage-
ment Branch at the Robert S. Kerr Environmental Research
Laboratory, Ada, Oklahoma.
119
-------
Thomas E. Short holds a BS in Chemical Engineering from
Lamar University and a MS and Ph.D. in Chemical
Engineering from Oklahoma State University. He is currently
a Chemical Engineer at the Robert S. Kerr Environmental
Research Laboratory at Ada, Oklahoma.
DISCUSSION
Bill McCarthy: When you used solvent extraction for recovering your adsorbent components,
could you give us an idea of what efficiency or recovery you got?
Leon H. Myers: No I can't, because we did not attempt a mass balance; however, the quan-
titative value of three solvents produced similar results. The quantity recovered agrees with
the projected isotherm loading.
Bill McCarthy: Do you think solvent regeneration might be a viable process?
Leon H. Myers: Yes, I do believe it is both feasible and viable, and this is determinant on
the mixtures of hydrocarbons to be removed. In some cases, particularly with light hydro-
carbons, steam regeneration might be the most viable regeneration process. Neither of these
alternate regeneration modes have been thoroughly proven.
120
-------
Table 1. CONTROLLED CONDITIONS FOR ISOTHERM STUDY
1. 325 mesh (pulverized) 5. 100 rpm agitation
2. 200ml sample 6. 23° C temperature
3. 1 hour contact 7. 75 mg/l TOC
4. 0.05, 0.20, 0.50, 1.0 gm 8. pH 7.4
carbon
Table 2. ADSORPTION DATA FOR ELEVEN ACTIVATED CARBONS
Carbon Sample mg/l TOC Adsorbed on 1 gm Carbon
Pulverized Granular
1
2
3
4
5
6
7
8
9
10
11
50
60
56
57
56
52
47
51
49
35
46
38
37
37
48
29
31
41
47
38
13
34
TableS. CONTACT TIME ADSORPTION
Time in Minutes
Gm Carbon
(Blank)
0.1
0.5
1 0
2.0
5.0
20
(206 mg/l)*
124
63
57
55
56
40
(260 mg/l)*
96
45
47
52
54
60
(260 mg/l)*
82
51
106
77
50
180
(260 mg/l)*
101
73
45
34
24
* Results are expressed as mg/l filtrable total organic carbon.
121
-------
Table 4. ANTHRAFILT FILTRATION ORGANIC CARBON REMOVAL
Charge
After 5 minutes (0.3 gal.)
After 1 hour (3. 2 gal.)
After 2. 5 hours (7. 9 gal.)
TOC
255 mg/l
220 mg/l
2 10 mg/l
206 mg/l
Turbidity
11.0
2.75
Tab.le 5. MINI-COLUMN TOC RESULTS
Date
3/17
3/18
3/19
3/20
3/21
Final Clarifier
Effluent*
35
33
48
37
35
Anthrafilt
Effluent*
31
34
40
48
38
Columns
A-l*
13
11
13
17
18
B-l*
9
11
16
23
19
A-2*
9
10
14
22
17
B-2*
10
12
14
16
25
A -3*
7
10
8
11
16
B-3*
12
18
12
22
32
All concentrations are mg/l TOC.
Table 6. COMPARISON OF MINI-COLUMN TOC REMOVAL EFFICIENCY
Date
3/17
3/18
3/19
3/20
3/21
Average
A-l
58
68
68
65
53
62
B-l
71
68
60
52
50
60
Percentage Removals
A-2
71
71
65
54
55
63
B-2
68
65
65
67
34
60
77
71
80
77
58
73
62
47
70
54
16
50
122
-------
Table 7. TOC RESULTS OF MINI-COLUMN UPFLOW MODE
Date
3/24
3/25
3/26
3/27
Final Clarifier
Effluent*
50
35
41
38
Anthrafilt
Effluent*
38
33
33
35
Columns
1*
19
17
20
17
2*
12
13
24
19
3*
11
10
18
17
4*
13
10
13
13
5*
13
9
12
10
All concentrations are mg/l TOC.
Table 8. PERCENTAGE REMOVAL OF TOC FOR UPFLOW MO|DE
Columns
Date
3/24
3/25
3/26
3/27
1
50
48
39
51
2
68
61
27
46
3
71
70
45
51
4
66
70
61
62
5
66
73
64
71
Average
47
51
59
65
69
Table 9. REFINERY PROCESS DATA
Capacity
API Class
Wastewater Volume
Refinery Processes:
30,OOOBPSD
B
18 gal. per minute
1 . Crude Distillation
2. Crude Desalting
3. Vacuum Crude Distillation
4. Fluid Cat Cracking
5. HFAIkylation
6. Hydro Cracking
7. Catalytic Reforming
8. Asphalt Production
123
-------
Table 10. REFINERY WASTEWATER TREATMENT RESULTS
Median Values
Parameter
BOD5
COD
TOC
Oil and Grease
Phenols
Chromium
Copper
Iron
Lead
Zinc
Sulfide
Ammonia
Cyanides
Turbidity*
Color**
API Separator
(mg/l)
97
234
56
29
3.4
2.2
0.5
2.2
0.2
0.7
33
28
0.25
26
30
Bio-Treated
(mg/D
7
98
30
10
0.01
0.9
0.1
3.0
0.2
0.4
0.2
27
0.2
17
15
Carbon-Treated
(mg/l)
48
103
14
10
0.004
0.2
0.03
0.3
0.2
0.08
39
28
0.2
11
15
Bio-Carbon
Treated
(mg/l)
3
26
7
7
0.001
0.2
0.05
0.9
0.2
0.15
0.2
27
0.2
5
1
* Turbidity given in Jackson Turbidity Units.
** Color given in color units.
Table 11. CARBON ADSORPTION CAPACITY
Capacity Lbs TOC Adsorbed Per Lbs Carbon
Isotherm Study
API Separator Effluent
Carbon Column
0.12
0.31
Table 12. CARBON REGENERATION ACTIVITY ANALYSIS
Iodine No.
Molasses No.
Virgin Carbon
API Separator Carbon
Final Clarifier Carbon
1010
906
991
216
405
304
124
-------
Table 13. SOLVENT EXTRACTION EFFICIENCY DATA
Carbon
Identification
API
API
API
API
API
FC
NEW
Solvent
Chloroform
Benzene
Meth/lene Chloride
Methanol
Hexane
Chloroform
Chloroform
Grams
Carbon
49.3008
50.6917
54.5660
56.3179
55.3762
56.6447
44.6169
Grams Oil
Extracted
5.04
5.01
5.44
3.71
4.03
1.95
0.00
Gram Oil/
Gram Carbon
0.102
0.099
0.100
0.066
0.073
0.034
0.000
Table 14. COMPOSITION BY TYPES OF ORGANICS
Organic Type
Saturates
Aromatics
Polar Material
Oil from
API Carbon
11.1%
24.7
64.2
100.0
Oil from
FC Carbon
0.2%
1.3
98.5
100.0
Table 15. COMPOUND TYPES INDICATED BY GC-MS
API Carbon Extract
FC Carbon Extract
Saturate Fraction
Aromatic Fraction
C9H16
C9H18
5.10^20
C12H24
Cg through
n-paraffins
Polar
Ethylbenzene
Xylenes
Co Benzenes
C]Q Benzenes
Naphthalene
Cj j Benzenes
C^ Naphthalene
Cj2 Naphthalene
C 13 Naphthalene
Phenol
Hydroxy toluene
Ethylbenzene
Xylenes
C9 Benzenes
C10 Benzenes
C7H8 S
C7H1QS
C9H8 S
125
-------
PULVERIZED
GRANULAR
FIGURE - I
4567
CARBON SOURCES
8
10
II
-------
o
UJ
m
o:
o
<
o
o
LJL
O
CO
s
<
a:
o
Ij
_i
5
OUUp
z
m ' 200ml /min
< -J<- or 4.4g/m/ft.2
0 lOOh /
o I /
5 /
< 50h /
cc DUr /
0 0
^ /
°- I WASTE - API SEPARATOR, EFFLUENT
1 I CARBON-FILTRASORB 400
"1
I
1
1
5! 1 1 , !
/
L
Y//.
W
Ix
K^
1
1
r.
K
p«t
l.*"4
kV
LPUMP b
^ 1
y
fc^1
API SEPARATOR M
WATER SAMPLE m
H
m
m
m
10 - 50 100 500
TOC-MILLIGRAMS PER LITER
5i5"ANTHRAFILT
tL30"SAND (20 - 30 Mesh)
H2" LIMESTONE (0.5"Dio.)
r|.5"GLASSWOOL
TO RESERVOIR
FIGURE 2 - ADSORPTION ISOTHERM
FIG. 3-MULTI-MEDIA FILTER
-------
COLUMN
234
s~\
\
ELECTRONIC
SAMPLER
FINAL
CLARIFIER
FINAL
CLARIFIER
SAMPLE
DO
2345
COLUMN
EFFLUENT
1ULTI-MEDIA
FILTER
SAMPLE
FIG.4-MINI COLUMN SYSTEM
SAMPLE
POINT-I
SAMPLE
POINT-2
COOLING TOWER
BLOWDOWN
BOII FR BLOWDOWN
OILY PROCESS WATER '
STRIPPER EFFLUENT
CONTAMINATED RUNOFF
\
TO HOLDING
POND
API SEPARATOR PASVEER DITCH " CLARIFIER
FIGURE 5 - REFINERY WASTE WATER TREATMENT SYSTEM
128
-------
FEED
SAMPLE
Pea
fGravel
DUAL-MEDIA FEED "CALGON" ACTIVATED
FILTER PUMP CARBON PILOT PLANT
FIGURE 6 - ACTIVATED CARBON PILOT PLANT
FLOW DIAGRAM
Of
UJ
cc
u
a
cc
o
If)
a
o
CO
200-
160-
APISEPARATOR
EFFLUENT
40 -
8 10
DAYS INTO STUDY
12
CARBON TREATED
EFFLUENT
BIO-TREATED EFFLUENT
BIO-CARBON EFFLUENT
16
FIGURE 7- TREATMENT RESULTS, BOD
129
-------
280
240
ui
-. 200
cc
u
a
s
cc
160
120
5
§ 80
u
40
70
60
50
DC
UJ
a.
API SEPARATOR
EFFLUENT
BIO-TREATED
EFFLUENT
CARBON TREATED
EFFLUENT
BIO-CARBON
TREATED
EFFLUENT
12
! 4 6 8 10
DAYS INTO STUDY
FIGURE 8- TREATMENT RESULTS, COD
14
16
API SEPARATOR
EFFLUENT
40
cc
o
I
o
o
30
20
10
I
I
2 4 6 8 10
DAYS INTO STUDY
FIGURE 9-TREATMENT RESULTS, TOC
130
I2
BIO-TREATED
EFFLUENT
CARBON TREATED
EFFLUENT
BIO-CARBON
EFFLUENT
14
-------
TEMPERATURE, °F
P5°F 260 380 490 610 720 828 895 I3S°F 260 380 490 610 720 825 895
UJ
-------
Temperoture , °F
CO
z
o
0.
CO
Ul
te
I35°F200 320 440 550 665 780 865
1 1
1 1 1 1 1 1 1 I l i i 1 1 1
i
\^ ^^— •* —
1 1 1 1 i t i i i 1 I l i i
0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 64
TIME, min.
FIGURE 16 - CHLOROFORM EXTRACT OF NEW ACTIVATED CARBON
780 865
Temperoture °F
I35°F 200 320 440 550 665
ui
CO
z
o
0.
CO
UJ
tr
4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 64
TIME, min.
FIGURE 17 - STANDARD OIL MIXTURE
132
-------
UI
ui
a:
TEMPERATURE, °F
135°F 260 380 490 610 720 825 895 I35°F 260 380 490 610 720 825 895
I |^H^^MBBBp^^H^I^ 1""*""""""""""""""""^ ["- f- • w —
I I I I I I I I 1 I T I 1
SENSITIVITY x 10
TIME, min
FIGURE 18 -APIC EXTRACT SATURATE FIGURE 19-FCC EXTRACT SATURATE
FRACTION FRACTION
TEMPERATURE, °F
i i i i i t I
SENSITIVITY x 10
i i i i i i i
16 24 32 40 48 56 64
0 8 16 24 32 40 48 56 64 0 8
TIME, min.
FIGURE 20-APIC EXTRACT AROMATIC FIGURE 21-FCC EXTRACT AROMATIC
FRACTION FRACTION
V)
Z
O
a.
V)
ui
ac
TEMPERATURE, °F
I35°F 260 380 490 610 720 825 895 I35°F 260 380 490 610 720 825 895
0 8 16 24 32 40 48 56 64 08 16 24 32 40 48 56 64
TIME, min.
FIGURE 22- APIC EXTRACT POLAR FIGURE 23- FCC EXTRACT POLAR
FRACTION FRACTION
133
-------
SECTION 4
Powdered Activated Carbon Pilot-Scale Studies
"COMBINED POWDERED ACTIVATED CARBON - BIOLOGICAL TREATMENT:
THEORY AND RESULTS"
Francis L. Robertaccio, Senior Engineer
E.I. duPont de Nemours & Co., Inc.
Wilmington, Delaware
The purposes of this paper are: to acquaint you with an
overview of some of the theoretical aspects of the combined pow-
dered activated carbon-biological treatment process, and to pre-
sent recent start-up experiences and results from the 1.5 x 10
M /day installation based on the process at the Du Pont Chambers
Works Plant in Deepwater, New Jersey.
Pact is Du Font's name for a patented process for purifica-
tion of sewage and/or industrial wastewater which comprises sub-
jecting the wastewater to an aerobic biological treatment process
in the presence of powdered activated carbon (^-' (Figure 1). The
aerobic biological treatment vessel (s) can have many geometric
configurations. Single or multiple reactors can be used. The
reactors can be plug flow, completely mixed, or somewhere in be-
tween. Powdered activated carbon addition is compatible with
activated sludge, contact stabilization, or aerated lagoon sys-
tems; that is, any process in which the carbon can be suspended.
The rate of powdered activated carbon addition for a given waste-
water is a function of the effluent quality desired. When the
rate of addition is expressed in terms of weight of carbon added
per unit volume of incoming wastewater, the rate becomes a func-
tion of the type of carbon used.
In addition, certain internal process controls such as the
solids retention time, or sludge age, can be changed to influence
the rate of application of a given type of carbon to produce a
desired result. (2). some of these relationships are illustrated
in Figure 2. Here sludge age arid carbon dose are shown as vari-
ables affecting effluent quality as measured by the total organ-
ic carbon (TOG) test. All data points represent treatment con-
ditions by which the effluent BOD of the industrial wastewater
tested was reduced to negligible concentration. The effluent
TOG is shown to be reduced by an independent increase in either
the sludge age or the carbon dose.N Note that the improvement in
effluent quality, by increasing sludge age, is less apparent
when carbon is absent. We have long postulated this phenomena
results because the adsorbed microorganisms have the sludge age
rather than the relatively shorter hydraulic detention time to
biodegrade adsorbed and difficult-to-degrade molecules. It is
important to recognize the economic advantage associated with the
ability to biodegrade these materials in the biological reactor
as an inherent advantage of the process.
135
-------
The alternative for this type of effluent quality improvement is
a more expensive add-on granular carbon adsorption step.
Table 1 illustrates the effectiveness of using carbon as
an adsorbent in a biological reactor.(3) The table shows the
dissolved organic carbon (DOC) removal from a PACT unit and
comparable data from a biological unit. Also shown is the
combined removal obtained from a separate adsorption of the
biological unit effluent using the same type and quantity of
carbon used in the PACT unit. The same carbon combined with
bacteria in the PACT unit removed more DOC and exceeded the
quantity expected from separate isotherm determinations.
Of course, as more molecules are biodegraded, adsorption
sites are filled with molecules that are more biorefractory and
a more rigorous form of regeneration is needed if the spent
carbon is to be reused. Alternatively, more active carbons
(i. e., higher surface area) can be used in throw away doses.
Economic considerations grouped as various capital and operating
expenses dictate the choice. To some extent the economics are
strongly influenced by the carbon usage rate, however, site
specific factors such as the local costs of alternative sludge
disposal methods must be considered. At the PACT treatment
facility for Chambers Works we will thermally regenerate carbon
from PACT sludge but wet oxidation can also be used.
The heart of the PACT system is a matrix of microorganisms
and powdered activated carbon. Figure 3 shows the matrix. ' '
The photo on the left is powdered activated carbon in the water;
the photo on the right the PACT matrix. The PACT matrix has
some interesting properties.
First, the carbon acts as a weighting agent. Sludge
settling rates are vastly improved as illustrated in the series
of pictures in Figure 4. Activated sludge and PACT mixed liquor
were taken from treatability units operating on the same waste-
water at the same sludge age. The series of photographs are
taken at different elapsed settling times shown on the timer in
the background. Note that the PACT sludge settles better and
has a clearer supernatant. The PACT sludge also compacts very
well. The PACT sludge had a mixed liquor suspended solids
concentration of 7700 mg/1 (about 65% carbon) and a sludge
volume index of 20 cc/g. The activated sludge had a mixed
liquor concentration of 2400 mg/1 and a sludge volume index of
46. We feel that the improved sludge settling, achieved by
simple carbon addition can result in the processing of more
wastewater through existing hydraulically overloaded treatment
plants. This is often a viable alternative to rather expensive
capital equipment expansion programs to accomplish similar
results. Of course, carbon addition will improve effluent
quality at the same time. A new treatment plant can incorporate
136
-------
this advantage in its design by substantially reducing the size
of secondary clarification. Another illustration of the im-
proved settling of PACT sludge is shown in Figure 5. Here the
initial settling velocity of activated sludge is compared at
various concentrations to a family of PACT sludge curves at two
carbon levels and two temperatures. The numbers on the PACT
labeled curves are the application rates for carbon in mg/1.
While on the topic of sludge handling, it should be
mentioned that PACT sludge dewaters much more readily than
conventional activated sludge. The manifestation of this
property is reduced size of sludge dewatering equipment. The
need to dewater more sludge as a result of the presence of
carbon is offset by reduced cycle times. Figure 6 compares the
specific resistance of activated sludge to two PACT sludges
at different carbon feed doses. (•*)
A second property of the PACT sludge matrix is that it
contains an effective adsorbent. We have already explored one
aspect of the role of the adsorbent - removal of biorefractory
organic compounds - in the discussion of carbon dose and its
relationship to sludge age. In that discussion the effluent
total organic carbon consent was a gross measure of biore-
fractory material. More specific measures of biorefractory
materials which might require control in specific instances
include materials contributing to final effluent color, oil
and grease, surfactants, chlorinated hydrocarbons, phenols and
toxicity to fish or other trophic level measures of toxicity
in receiving waters. No matter how you care to measure, or are
told to measure these biorefractory materials,it is apparent
PACT can control these substances to levels beyond the
capability of conventional biological systems. In complex
waste situations control of these substances at the PACT treat-
ment plant is often a more viable alternative than biological
systems followed by granular activated carbon columns or source
treatment.
Sometimes biorefractory materials are as much, or more of a
problem w '-hia a biological system as they are in its final
effluent. Examples include materials toxic or inhibitory to
biological reactor microorganisms, materials that are periodic-
ally present in high concentrations (shock loads), or materials
that cause severe foam, odor, or bulking sludge. Unlike post-
biological separate granular activated carbon treatment, the
presence of carbon in the aerator often controls these in-
process problems as well as those normally associated with the
biologically treated effluent. Over the years, we have had a
difficult time sustaining bench scale biological treatability
units on Chambers Works wastewaters due to the periodic
presence of toxic or inhibitory materials.1 However, PACT
treatability units operated in parallel did not experience
137
-------
similar upsets. As a result of this property of the PACT
process the full scale PACT facility at Chambers Works has no
upstream equalization. We have even spiked Chambers Works
wastewater with various toxic substances including pentachloro-
phenol (190 1/min. pilot plant test) and found the PACT system
can sustain efficient operation when the conventional biological
system completely fails. Ferguson, et al , reported similar
findings in shock loading tests involving trichlorophenol. (5)
There are two intertwining reasons to consider any waste
treatment process: technical merits and economics. We have
touched upon the technical merits of the PACT process and
summarized them in Table II. In most instances any one of the
technical merits may be sufficient reason to consider PACT. In
most instances a number of technical merits must be simulta-
neously applied to the consideration of the process at a
specific site. The resulting matrix of reasons results in a
difficult appraisal of the full value of the use of the PACT
process versus alternative processes. Some of the economic
considerations are shown in Table III. At Du Pont we are con-
vinced PACT is a versatile, economic and technically viable
process. We have about 100 man years experience in PACT process
research and development. At the Chambers Works facility which
will be described next, we feel PACT represents a $7 million
capital savings and a $5 million/year operating cost savings
(1972 dollars) over the next best alternative which was granular
carbon treatment followed by activated sludge.(6)
The f,ull scale PACT facility at Chambers Works has been in
a start-up phase since mid-November 1976, and it proceeded
smoothly through the coldest winter in decades. The liquid
train is on line and the solids handling train is expected to
be fully operational fairly soon. During March 1977 a half full
flowrate test, and during early May a full flowrate test were
conducted. This portion of this paper will highlight the start-
up operation and describe results of the tests.
Figure 7 shows the major components of the PACT portion
of the Chambers Works treatment plant. Construction of the PACT
facility started in February 1974 and was completed in December
1976 at an estimated capital cost of $22.5MM. Primary effluent
is split equally to each of three 15MM liter aeration tanks as
is the recycled PACT sludge. Five 1,000 hp blowers supply air
to static mixers in the aerators. Effluent from the aerators
is conveyed to the clarifier flowsplitter and then to two
secondary clarifiers. Treated effluent (overflow) is dis-
charged to a basin and then to the Delaware River. Secondary
clarifier underflow is returned to the aeration tanks via two
2 meter screw pumps. This part of the system constitutes the
"liquid train" and includes feed and unloading facilities for
carbon, phosphoric acid and polymer.
138
-------
Waste sludge can be removed from either the mixed liquor
or recycle lines and is pumped to a sludge thickener where it
is settled to 7 + 2 weight percent solids. The thickener
underflow serves as a feed stream for the filter press which is
designed to produce a 35 weight percent dry solids PACT cake.
The cake is mechanically conveyed to a five hearth furnace
where it is dryed, the biomass incinerated, adsorbed organics
pyrolyzed and the powdered carbon in the sludge regenerated for
reuse. The furnace off-gases pass through two scrubbers and
an afterburner before being discharged to atmosphere. The dry,
regenerated carbon recovered off the bottom hearth is slurried,
acid washed, and returned to the carbon feed tanks for recycle
to the process. No waste sludge is produced. This part of the
system is called the "solid train".
The liquid train startup became evident in November 1976
when 2.7 x 105lkg of powdered carbon and 2.0 x 105 kg of
bacterial solids were added to one aerator. Water temperature
at the time was 11-15°C. During January a second aerator and
clarifier were brought on line. During February the carbon
regeneration startup began and during March the sludge press
was brought on line. The more important operating problems
encountered and solved during startup have been presented in
a recent paper by Flynn. (7' The problems were of the type
found with the startup of a conventional activated sludge pro-
cess, that is, they were not at all related to the uniqueness
of the PACT process. These problems went through the classic
problem solving stages - initial definition, questioning of
assumptions, hypothesis forming, reobservation of" the problem
in some cases, implementation of a solution and feedback on the
success of the solution.
In March a half-full flowrate test was conducted. Table IV
compares operating conditions, feed and effluent quality for
the full scale PACT facility and various bench scale controls.
The effluent color and dissolved organic carbon (DOC) are
important control parameters. During this test, flowsheet
dosages of virgin carbon (regenerated carbon not available at
this date) reduced effluent DOC to 20 ppm (43 ppm goal) for the
last seven days of the test and an average of 36 ppm for the
entire test. Effluent color was 310 (540 goal) despite the
feed color being 42% over design. The full scale, half-flow
test results compare favorably with the PACT bench scale control.
This table also presents insight into the improvement in
effluent quality offered by PACT. Note the marked decrease in
effluent DOC and color in the PACT full scale or bench scale
units versus the biological bench scale unit.
In early May a full flowrate test was conducted. Table V
presents operating conditions, feed and effluent quality for
the test.' The effluent color and dissolved organic carbon
139
-------
again met goals despite the feed color being 108% over design.
Once again this test was conducted using virgin carbon.
In summary, this paper has presented an overview of some
of the theoretical aspects of the combined powdered activated
carbon-biological treatment process (PACT) and updated recent
startup experiences from the 1.5 x 105 M3/day installation
based on this process at the Du Pont Chambers Works. This
process is a versatile, viable wastewater treatment technology;
we expect its use to become an accepted solution to a variety
of existing and future wastewater treatment problems.
REFERENCES CITED
1. U. S. Patent 3,904,518
2. Flynn, B. P., F. L. Robertaccio and L. T. Barry,
"Truth or Consequences: Biological Fouling and Other
Considerations in the Powdered Carbon - Activated Sludge
System". Presented at the 31st Annual Purdue Industrial
Waste Conference, West Lafayette, Indiana, May 5, 1976
3. Heath, H. W., "Combined Activated Carbon-Biological
("PACT") Treatment of 40 MGD Industrial Waste" presented
to Symposium on Industrial Waste Pollution Control at
ACS National Meeting, New Orleans, LA., March 24, 1977
4. Robertaccio, F. L., D. G- Button, G, Grulich and H. L.
Glotzer, "Treatment of Organic Chemical Wastewater with
the Du Pont PACT Process". Presented at A. I. Ch. E.
National Meeting, Dallas, Texas, February 1972
5. Ferguson, J. F., G. F. P. Keoy, M. S. Merrill and
A. H. Benedict "Powdered Activated Carbon - Biological
Treatment: Low Detention Time Process" presented at
the 31st Annual Purdue Industrial Waste Conference, West
Lafayette, Indiana, May 4, 1976
6. Flynn, B. P. and L. T. Barry "Finding a Home for the
Carbon: Aerator (Powdered) or Column (Granular)".
Presented at the 31st Annual Purdue Industrial Waste
Conference, West Lafayette, Indiana, May 1976
7. Flynn, B. P. "Operating Problem Solving During a
Secondary-Tertiary Treatment Plant Start-Up". Presented
at llth Mid-Atlantic Regional ACS Meeting, University
of Delaware, Newark, Delaware, 'April 22, 1977
140
-------
DISCUSSION
Ed Sebesta, Brown & Root; What is the design hydraulic loading
for secondary clarifiers at the Chambers Works wastewater
treatment facility?
Robertaccio and B. P. Flynn, Du Pont; The solid flux rate is
designed at 250 Ibs. per day per square foot which is kind
of high. At that solid flux rate, you should be able to get
an underflow concentration of 3-1/2 weight percent. In
full scale testing we have been able to generate 7-1/2
weight percent solids which means we could operate without
a waste sludge thickener and could feed our' filter press
directly from our return sludge line. The hydraulic over-
flow rate is in excess of about 1000 gallons/day/ft.2. We
have two secondary clarifiers but could send full flow
through one secondary clarifier.
Leonard W. Crame, Texaco; What does the Du Pont PACT process
patent mean to the refining industry iii terms of using this
process?
Robertaccio; Du Pont will license any user of the process.
The royalty rate will be reasonable in order to encourage
use of the process.
J. E. Rucker, API; Please comment on economics and feasibility
of regeneration of powdered carbon from PACT sludge.
Robertaccio; Economics first. We think that powdered carbon
can be regenerated for an operating cost of about 5£ a
pound. Capital costs would depend on the size of the
facility and the method used to annualize capital costs." At
Chambers Works capital costs would add another 5C a pound.
Now feasibility. We put as much effort into the regenera-
tion part of the Chambers Works facility as we did to the
PACT process. The regeneration system is being brought on
line. We have had some mechanical problems but we don't
expect to have any more trouble solving these as we had
solving other problems. Of course, a number of thermal and
wet oxidation regeneration equipment manufacturers will tell
you they think regeneration of powdered activated carbon
from PACT sludge is no problem.
Dave Skamenca, Envirotech; Did you pilot test mixing the
powdered activated carbon - biomass mixture with static
mixers?
Robertaccio: Yes.
141
-------
BIOGRAPHY Francis L. Robertaccio
Fran is a senior engineer for the DuPont Company
in Wilmington, Delaware. He has 12 years experience
in various industrial pollution control positions. He
holds a B.S. and M.S. in Chemical Engineering from
Clarkson College and a Ph.D. in Environmental
Engineering from the University of Delaware. He has
authored a dozen papers on industrial pollution control,
holds several U.S. and foreign patents and is a member
of AlChEand WPCF.
142
-------
TABLE I - SYNERGISTIC EFFECT ON DOC REMOVAL WITH "PACT"
DOC, mg/1
BIO UNIT
FEED TO BIO UNIT EFFLUENT + BATCH "PACT"(2)
TRIAL UNITS EFFLUENT CARBON ADSORPTION*1) EFFLUENT
1 183 80 59 44
2 178 70 42 18
3 167 79 55 25
(1) Take 500 cc filtered Bio Unit effluent, add 150 ppm
virgin carbon, stir 3 hours at room temperature, filter,
and analyze filtrate for DOC
(2) "PACT" unit operating at 20°C with 8.0 day sludge age
and 160 ppm carbon addition
143
-------
TABLE II - REASONS TO CONSIDER PACT
I Existing Biological Treatment Plants
A. Need to improve effluent quality
treatment plant hydraulically overloaded
biological treatment unstable
odor, foam
new, more restrictive effluent limits
biological treatment inherently incapable of
removing some control parameters
• water quality limitations not met in geographic
area
B. Need to r.educe cost
• rising sludge disposal costs
• expensive chemicals used to aid biological treatment
C. Need to expand treatment plant
• want to accept new customers
• currently overloaded
® want to accept new product's wastewater; afraid
biological process will become unstable or in-
capable of removing new waste constituents
• can no longer use off plant sludge disposal site
D- This treatment p-lant will eventually be abandoned
(i. e., to join regional plant) but I have to get the
most out of what I have
II New (potentially biological) Treatment Plants
A. Want cost effective process
B. Concern about efficiency of biological treatment
have potentially toxic waste
face strict effluent limits
want stable process
have wastewater from changing product line
future regulations might outdate biological
treatment capabilities
C. Have components in waste not currently regulated, but
want them removed now.
D. Limited amount of land available for treatment
E. Want to minimize sludge disposal problems
• have undesirable components in waste that will
be concentrated in sludge; don't want these
released to environment
• don't have land, or availability of ocean disposal
III New Advanced Waste Treatment Plants
A. Want cost effective process
B. Want flexibility to alter treatment plant
• as regulations change
• as product mix dictates different treatment
need -
- over short intervals
- over a long period
C. Concerned about stability of alternate advanced
treatment processes
144
-------
D. Want to minimize investment
IV Miscellaneous
A. Pretreatment plus municipal disposal route too
expensive
B. PACT is a low risk process compatible with many
existing waste processing schemes, changes in
product mix, or changes in regulations
TABLE III - ECONOMIC CONSIDERATIONS OF POWDERED ACTIVATED
CARBON ADDITION
PRO CON
Eliminates granular carbon • Powdered Carbon Cost
adsorption equipment needs, - virgin
including initial GAG about 0.5-0.8C/1000 liters/
inventory lOppm
using 55 to SOC/kg carbon
• Regenerated (full cost)
- about 0.1-0.2^/1000
liters/lOppm
Minimizes need for equaliza-
tion facilities to control
wastewater variability
Eliminates separate second-
ary sludge disposal if re-
generation is used
Reduces or eliminates need • May require use of flocculant
for antifoam, odor control
Protects biological system
from inhibition or toxic
upset
Reduces size requirements
for secondary sludge
settling, thickening,
dewatering
Carbon addition rate
readily changed for changes
in wastewater character-
istics or regulations
145
-------
TABLE IV - CHAMBERS WORKS HALF FULL FLOWKAan u.^-.
TEST PERIOD
Operating Conditions
Carbon Dose (ppm)
Aeration Temp (°C)
Hydraulic Residence
Time (hrs)
Sludge Age (days)
Feed Quality
BOD-Soluble (mg/1)
DOC (mg/1)
Color (APHA)
Effluent Quality
BOD-Soluble (mg/1)
DOC (mg/1)
Color (APHA)
3/13/77-3/26/7
FULL SCALE
PACT PLANT
182
22
14.6
*
304
214
1416
15.2
35.7
311
7 INCLUSIVE
BENCH
PACT
150
22
7.5
8
304
214
1416
19.3
28.4
369 •
SCALE CONTROLS
CONVENTIONAL
BIOLOGICAL
0
22
7.5
8
304
214
1416
13.8
67 . 3
1900
*no steady state material balance available
TABLE V - CHAMBERS WORKS FULL FLOWRATE TEST
TEST PERIOD 4/26/77-5/6/77 INCLUSIVE
Operating Conditions
Carbon Dose (ppm) 189
Aeration Temp (°C) 28,
Hydraulic Residence Time (hrs) 7,
Feed Quality
BOD - Soluble (mg/1) 300
DOC - (mg/1) 214
Color (APHA) 2080
Effluent Quality
BOD - Soluble (mg/1) 9
DOC - (mg/1) 28
Color (APHA) 490
146
-------
Powdered Activated
Carbon
Sewage
and/or
Industrial
Wastewater
Aerobic
Biological
Treatment
Product
Water
Gas Containing
Oxygen
Fig. 1 A SIMPLE ILLUSTRATION OF PACT
-------
00
O>
o»
.20
-> ^ & ^
V/ tf^o
&
*'
* Numbers shown in the bottom face of the exploded cube are
applied carbon doses, mg/liter
Fig. 2 EFFECT OF SLUDGE AGE AND CARBON DOSE ON
EFFLUENT TOC AND APPARENT LOADING
-------
Left:
Virgin Carbon in Water
I
Right:
PACT Sludge
V
Fig. 3 PHOTOMICROGRAPHS
-------
PACT
SLUDGE
INITIAL
SOLIDS. MG/L
PACT 7708 ± 896
ACTIVATED SLUDGE 2412 ± 150
ACTIVATED PACT
SLUDGE
••*•£•
ACTIVATED PACT
' "f SLUDGE
SLUDGE VOLUME
INDEX, CC/GRAM
20 ± 2
46 1 3
ACTIVATED PACT
SLUDGE SLUDGE
f •*
ACTIVATED
SLUDGE
PACT
SLUDGE
ACTIVATED pACT
SLUDGE SLUOGE
ACTIVATED
SLUDGE
PACT
SLUDGE
ACTIVATED PACT
SLUDGE SLUDGE
ACTIVATED
SLUDGE
NOTE: TIMER HANDS MOVE COUNTERCLOCKWISE. READ ELAPSED TIME USING SMALL NUMBERS.
LONG HAND INDICATES MINUTES, SHORT HAND INDICATES SECONDS.
FOR EXAMPLE, LOWER LEFT TIMER READS I MINUTE, 21 SECONDS.
BOTH SLUDGES FROM TREATABILITY UNITS AT SAME SLUDGEAGE, HYDRAULIC DETENTION TIME.
Fig. 4 COMPARATIVE SETTLING CHARACTERISTICS
OF PACT AND ACTIVATED SLUDGE
-------
1.0
Initial Settling Velocity, Q -| _
Ft/Min
en
0.01
PACT-150 (winter, 10°C)
0.1 1.0
Suspended Solids Concentration, %
Fig. 5 SETTLING CHARACTERISTICS OF PACT
& ACTIVATED SLUDGES
10
-------
7.0
6.0
5.0
Bio Unit
C = .0047
M = .137
"r" = 128x 1012 —
gm
2PA2M
VC
M = Slope t/v vs v
C = gm solids Ice filtrate
P = Filtration Pressure, 1.5 x 107
A = Filter Area, 38.3 cm2
V = Filtrate Viscosity, .010 poise
2 PA2
*™ = 4.40 X 1012
cnrr
en
4.0
3.0
2.0
1.0
55 ppm "PACT"
C = .0167
M = .022
"r" = 5.8 x 1012
cm
50
100 150
Filtrate Volume in Cm3
110 ppm "PACT"
C = .028
M = .0093
"r" = 1.5x 1012
cm
gm
200
250
Fig. 6 SPECIFIC RESISTANCE, "r" OF PACT vs
BIOLOGICAL SLUDGE
-------
Carbon and H3PO,
Unloading, Storage
and Feed
Primary
Effluent
Five 1,000 hp
Blowers
Air
Three
Parallel
Aerators
Polymer Unloading,
Storage and Feed
Two Parallel
Screw Pumps
V
r
He!
Two
Parallel
Clarifiers
Return __
Sludge
Waste Sludge
"Liquid Train"
To Basin
and River
I Waste
I Sludge
(Thickener
To Primary
Treatment
To Carbon Slurry
Storage Tanks "*
Fuel Oil
oading, Storage
and Feed
J
[W
To
~Atm.~
Afterburner
Filter Press
Scrubbers
"Solid Train"
HCI Unloading,
Storage and Feed
Fig. 7 DUPONT PACT PROCESS: CHAMBERS WORKS
-------
POWDERED ACTIVATED CARBON ENHANCEMENT OF ACTIVATED
SLUDGE FOR BATEA REFINERY WASTEWATER TREATMENT
Colin G. Grieves
Research Engineer, Amoco Oil Company
Michael K. Stenstrom
Research Engineer, Amoco Oil Company
Joe D. Walk
Project Director, Standard Oil Company (Indiana)
James F. Grutsch
Coordinator of Environmental Projects, Standard Oil Company (Indiana)
ABSTRACT
Pilot plant studies show that powdered activated carbon enhancement of
activated sludge is a viable alternate to and less costly substitute for
granular carbon tertiary .treatment of refinery wastewaters. Effluent quality
depends upon both the equilibrium concentration and the surface area of the
powdered carbon in the activated sludge mixed-liquor.
Operation at very high sludge ages—60 days or more—allows the carbon to
accumulate to high concentrations in the mixed-liquor even though only small
make-up amounts are added to the system. Also, carbons with a high surface
area are especially efficient in adsorbing contaminants. Consequently,
costly regeneration may be unnecessary because the spent carbon can simply be
discarded with the waste sludge'. Powdered carbons may thus eliminate the
need for the add-on granular carbon adsorption process that the Environmental
Protection Agency has recommended for meeting proposed 1983 standards for
Best Available Technology Economically Achievable (BATEA).
INTRODUCTION
According to the EPA guidelines for treating refinery wastewaters , the
sequence shown in Figure 1 is recommended for meeting 1977 standards for Best
Practical Technology Currently Available (BPTCA). For meeting 1983 goals for
Best Available Technology Economically Achievable (BATEA), the guidelines
recommend an add-on process using granular carbon adsorption. However, this
approach may be both inefficient and very costly. So far as is known, its
effectiveness has never been adequately demonstrated. Moreover, preliminary
estimates indicate that capital and operating costs for the granular carbon
adsorption and regeneration facilities may equal or exceed those of the
entire current activated sludge process.
2-25
By contrast, both patents and research studies indicate that powdered
activated carbon may be a practical and economical substitute for granular
carbon. For example, powdered carbon costs only about one-half as much as
# References inserted at end of text.
154
-------
granular--$0 65/kg versus $1.20/kg15. In addition, recent studies have shown
that powdered carbon can be added directly to the mixed-liquor in activated
sludge aeration tanks ^> 22, 23, 24. Thus> approprlate alteratiOns in
operating procedures may eliminate the need for regeneration by making it
economically feasible to discard the spent carbon with the waste sludge.
In general, the cost effectiveness of a powdered carbon process increases
with the concentration of carbon maintained in the mixed-liquor. A mass bal-
ance of such a process is represented by the following equation:
ci 9c (1)
C = Qh
where
C = Equilibrium mixed-liquor carbon concentration (mg/1)
Ci = Influent carbon concentration (mg/1)
9C = Sludge age (days)
0^ = Hydraulic retention time in the aeration tank (days)
Equation 1 reveals that the equilibrium mixed-liquor carbon concentration is
proportional to the product of the influent carbon concentration (carbon
dose) and the sludge age. Thus, equilibrium carbon concentration can be
increased by increasing the carbon dose, or the sludge age, or both. There-
fore, to keep carbon costs to a minimum, it is desirable to operate at as high
a sludge age as possible and not at an excessively long hydraulic retention
time.
A possible drawback to operation at a high sludge age is the increased risk
that toxic, inhibitory, or inert materials will build up in the aeration
tank. For example, a build-up of oily solids could reduce the oxygen trans-
fer efficiency and inhibit both the nitrifying and organic carbon utilizing
organisms. The dissolved oxygen concentration in the mixed-liquor could
also become too low for effective nitrification, and the final clarifiers
could become overloaded. Therefore, it is desirable in the pretreatment step
to remove as much solid material as possible from the wastewater before it
enters the aeration tank.
To evaluate the effects of such variables in a process using powdered carbon,
an extensive 15 month four-phase pilot plant study was carried out at Amoco
Oil Company's Texas City refinery. Pilot plants operating in parallel with
the refinery activated sludge process facility were fed the same wastewater
for treatment. Specific variables investigated were:
Carbon type, including surface area and pore volume
Carbon addition rate
Sludge Age
Pretreatment of feed to remove oil and solids
155
-------
EXPERIMENTAL EQUIPMENT
Figure 2 shows the configuration of the pilot plants. Each had a volume of
42 liters, and as many as eight units were operated in parallel during por-
tions of the study. They were housed in a rain-tight enclosure but were
neither heated nor cooled. Thus, the temperature of the mixed-liquor varied
from 4°C to 31°C.
Operating conditions and analytical procedures are summarized in Table 1.
The pH was checked daily and controlled by addition of caustic at a constant
rate. Dibasic potassium phosphate, K2HP04, was added to satisy the phos-
phorus requirement of the microorganisms.
The wastewater feed, a slipstream from the pressure filters of the refinery
treatment plant, was passed through a pilot gravity sand filter before being
fed to the pilot plants.
Table 2 summarizes the characteristics of the five powdered carbons evaluated.
Amoco's experimental high-surface-area carbons are designated as Al and A2,
PX-21 and PX-23, respectively. Those designated as B, C, and D are commer-
cially available carbons having a much lower surface area. Carbon A2 (PX-23)
has the highest pore volume.
Effectiveness was judged on the basis of the following effluent standards
proposed for a BATEA facility-*-:
Concentration,
mg/liter
Total Organic Carbon (TOG) 15
Chemical Oxygen Demand (COD) 24
Ammonia (NH3-N) 6.3
Phenolics 0.02
These standards are for a Class "C" refinery and are based on the guideline
effluent flow rate of 0.46 m3/m3 of crude throughput per stream day (19 gal/
bbl). Because the BATEA treatment sequence will undoubtedly result in very
low concentrations of effluent suspended solids, only the soluble components
of the effluent were measured.
To obtain high sludge ages, effluent suspended solids were allowed to settle
in 30-gallon plastic containers and then were returned to the pilot plants
periodically. At any given sludge age, all plants were allowed to reach
steady-state operation over an extended period of time. Then performance
data were taken over a 30-day period.
RESULTS AND DISCUSSION
The four phases of the study were carried out in sequence, with the design of
succeeding phases based on the results of the preceding ones. In summary,
they examined:
156
-------
Objective
I Effect of carbon type at an addition rate of
100 mg/liter and a sludge age of 20 days with
prefiltered feed.
11 Effect of carbon type at an addition rate of
200 mg/liter and a sludge age of 20 days with
prefiltered feed. v
111 Effect of increasing sludge age to 60 days and
reducing carbon addition rate to 25 mg/liter
with unfiltered and prefiltered feed.
IV Effect of further increasing sludge age to 150
days while reducing carbon addition rate to
10 mg/liter.
Phases I and II
The results of Phases I and II, summarized in Table 3, indicate that powdered
activated carbon significantly enhances the performance of a refinery
activated sludge process. Improvement in the quality of the effluents from
carbon-fed plants ranged from 65% for soluble organic carbon up to 95% for
phenolics. At the 200 mg/liter addition rate, the results usually satisfied
the BATEA effluent quality goals. The high surface area carbon Al was
significantly more effective than the other three. The commercially avail-
able carbon B produced slightly better effluent than carbon C, which would be
expected if efficiency is proportional to surface area. Because nitrifica-
tion was essentially complete in the control unit, carbon addition could not
improve ammonia conversion. Carbon D, which is derived from wood charcoal
and has a significantly lower pore volume than the others, performed so
poorly in Phase I that it was dropped from further consideration. The per-
formance of carbon Al at 100 mg/liter dose was about as effective as carbon
B at 200 mg/liter, or about twice as effective as the best commercially
available carbon tested.
Phase III
Table 4 shows the effects of sludge age and/feed filtration upon performance.
The plant with filtered feed performed better than one with unfiltered feed,
and a sludge age of 60 days was better than one of 20 days. No deterioration
in the settling characteristics of the mixed-liquor suspended solids was
observed at this higher sludge age.
At a sludge age of 20 days the plant with filtered feed performed marginally
better than the one with unfiltered feed. Undoubtedly, greater differences
in effluent quality would have been observed in a plant operated at a sludge
age of 60 days with unfiltered feed. (Not recorded in these data, however,
is the complete failure of the plant fed unfiltered feed shortly after
cessation of data gathering for this steady-state period.)
157
-------
Table 4 also shows how pore size and surface area affect the performance of
the carbons. Carbons Al and A2 have approximately the same surface area, but
carbon A2 has much larger pores. Yet, at an equivalent addition rate of 50
mg/liter, both carbons showed about the same performance. Thus, large pore
diameters are not required for effective treatment of this refinery waste-
water. Moreover, plants fed 50 mg/liter of either Al or A2 performed much
better than the plant fed 100 mg/liter of carbon B. In fact, these high-
surface-area carbons are between two and four times more effective than
carbon B in enhancing SOC and soluble COD removal.
A comparison of the data in Tables 3 and 4 shows that a low carbon dose and a
high sludge age enhance an activated sludge process almost as much as do a
high carbon dose and a low sludge age.
It is possible that the difference in performance is solely due to difference
in temperature between the phases—mean operating temperature during Phase
III was only 14°C, whereas during Phases I and II temperature averaged 31°C
and 25°C, respectively.
Also observed during the lower operating temperature of Phase III was an
increase in the ammonia removal efficiency of the carbon-fed pilot plants.
This phenomenon was unexpected because activated carbon, does not normally
adsorb ammonia. Possibly, the increased removal rate is due to the adsorp-
tion of potentially toxic or inhibitory organic materials which would reduce
the rate of nitrification if left in solution. The control plant in Phases I
and II had little difficulty in achieving full nitrification, perhaps because
of the higher temperature.
Phase IV
As shown in Table 5, Phase IV was designed to push the activated sludge sys-
tem to the limit by increasing sludge age to 150 days and decreasing carbon
addition to 10 mg/liter. Further, in one of the plants, hydraulic retention
time was reduced to 7.5 hours, compared with 15 hours in the other plants.
Despite similarities in influent quality during all four phases, during
Phase IV the effluent SOC and COD of the control increased by about 30-35%
over that observed during the first three phases, despite a mean temperature
of 27°C (c.f. 14°C during Phase III). All pilot plants essentially nitrified
completely.
Remarkably, however, the plant with 10 mg/liter of high surface area carbon
Al at a sludge age of 150 days produced an effluent whose soluble organic
carbon concentration was 50% lower than that of the control reactor and
slightly lower than that of all of the other pilot plants. The plant dosed
with 25 mg/liter carbon Al, with one-half the hydraulic capacity of the other
plants, produced the second best effluent.
The outstanding performance at a sludge age of 150 days indicates that
refinery activated sludge processes can be operated with very little added
158
-------
carbon. The dose may be low enough so that the carbon need not be regen-
erated but be discarded with the waste activated sludge. At a very high
sludge age, there will be smaller quantities of waste sludge to be disposed
of.
The data in Table 5 also indicates that powdered carbon can be used to
increase the hydraulic capacity of an activated sludge plant, as proposed by
others13, or to increase the effluent quality of an overloaded plant. The
carbon-fed plant that operated at one-half the hydraulic retention time of
the control produced an effluent 50% better than that of the control. Exper-
ience with pilot activated sludge plants operated at several of Amocofs other
refineries has shown that conventional activated sludge processes cannot be
operated successfully with a hydraulic residence time of only 7% hours.
Status of Powdered Carbon Enhancement of Activated Sludge
The data from Phase IV indicate that the limits of the powdered carbon
enhanced activated sludge process have not been reached. In addition, more
data are .needed before economic studies can be made to weigh the possible
options for achieving a given effluent quality: high fresh carbon dose at
moderate sludge age (20-60 days) with regeneration of spent carbon; low
fresh carbon dose at high sludge age (60-150 days) with no regeneration of
spent carbon. Cost analyses should be made for each of these extreme options,
and several intermediate ones, and compared with those for tertiary treatment
with granular carbon technology.
Figure 3 shows the qualitative curves this pilot study has generated. Of
course, the one for the 150-day sludge age is purely speculative because
only one data point exists. However, the trend of the data does show that
effluent quality is a function of mixed-liquor carbon concentration. The
curves are probably asymptotic to a residual organic carbon concentration,
but over the range investigated an increase in mixed-liquor carbon concentra-
tion causes a decrease in effluent soluble organic carbon. Furthermore, the
relationship between effluent quality, sludge age, and carbon dose is clearly
non-linear. For example, to achieve an effluent quality of 12.5 mg/liter of
soluble organic carbon, the three options are: 100 mg/liter of carbon at a
sludge age of 20 days; 47 mg/liter of carbon at a sludge age of 60 days; 24
mg/liter of carbon at a sludge age of 150 days. If the relationship were
linear, the values calculated from a base case of 100 mg/liter at a 20-day
sludge age would be 33 mg/liter and 13 mg/liter at 60 days and 150 days,
respectively.
Apparently, the process loses effectiveness because of incomplete microbial
regeneration. Microbial regeneration of the spent carbon is probably not as
effective as using fresh carbon; some materials adsorbed by the carbon are
undoubtedly non-biodegradable, even after 150 days of contact with micro-
organisms in the pilot plant. The ability to retain significant effective-
ness even at 150 days is the key to cost effective high sludge age operation
with powdered activated carbon. Of course, there may be other reasons why
carbon loses effectiveness at high sludge age, such as production of cell
lysis products which are then adsorbed by the carbon.
159
-------
Effluent Variability
Variation in effluent quality over a 30-day (or longer) period is extremely
important. The EPA* has set the daily maximum variability equivalent to the
99% probability value and the 30-day maximum variability to the 98% level.
For BATEA the daily maximum variability factors for TOG, COD, NH^-N, and
phenolics are proposed at 1.6, 2.0, 2.0, and 2.4, respectively. The 30-day
maximum values are 1.3, 1.6, 1.5, and 1.7, respectively.
Figures 4, 5, 6, and 7 show probability data for the 30-day operating periods
during Phase III. Table 6 shows the daily maximum (99% probability) and
30-day maximum (98% probability) variability factors calculated from these
figures for the plant fed with 25 mg/liter of Carbon Al. The EPA guideline
values are also given. The actual variability factor was calculated as the
99% (or 98%) probability value divided by the target quality value. In
general, the variability in effluent quality was higher than the guideline
values.
It is important to note that the proposed guideline variability factors are
unrealistic.The data base used by EPA* for their production was obtained
from limited pilot studies. In addition, BPTCA 30-day maximum (98% prob-
ability) values were used as the BATEA 30-day maximum values. Variability
factors will undoubtedly have to be amended before BATEA goals become BATEA
standards.
SUMMARY AND CONCLUSIONS
A viable alternative to granular activated carbon tertiary treatment of
refinery activated sludge effluent for meeting proposed 1983 BATEA effluent
quality standards has been demonstrated. The proposed process involves add-
ing powdered activated carbon to the aeration tank of the activated sludge
process, achieving cost effectiveness by operating at a very high sludge age
and a low carbon dose. Effective removal of oil and colloidal solids in
the pretreatment step is necessary for successful operation.
Effluent quality depends upon both the equilibrium mixed-liquor carbon
concentration and the surface area of the carbon. . An experimental carbon
with a high surface area appears to be several times more effective than the'
best commercial carbons in achieving an effluent quality standard. Pore
size of the activated carbon had no apparent effect upon effluent quality.
In general, the process can be used to meet only the long-term average
effluent quality proposed for BATEA. Daily maximum and 30-day maximum var-
iability goals, as presently defined cannot be met.
The proposed process also enhances nitrification at low temperatures and
dampens effects of increased hydraulic flow rate on the activated sludge
factors. Both phenomena will help to decrease effluent variability.
160
-------
REFERENCES
1. U.S. Environmental Protection Agency, (Draft) Development Document for
Effluent Limitations, Guidelines and New Source Performance Standards
for the Petroleum Refining Point Source Category. USEPA, Washington,
B.C., 20460 (April, 1974).
2. C. P. Derleth, U.S. Patent 1,617,014, February 8, 1927.
3. N. Statham, U.S. Patent 2,059,286, November 3, 1936.
4. F. L. Robertaccio, D. .G. Button, G. Grulich, and H. L. Glotzer, "Treat-
ment of Organic Chemicals Plant Wastewater with the DuPont PACT
Process." Presented at AIChE National Meeting, Dallas, Texas, February
20-23, 1972.
5. A. D. Adams, "Improving Activated Sludge Treatment with Powdered Acti-
vated Carbon." Proc. 28th Annual Purdue Industrial Waste Conference,
1972. ~ ' ~~~
*i
6. F. L. Robertaccio. "Powdered Activated Carbon Addition to Biological
Reactors." Proc. 6th Mid Atlantic Industrial Waste Conference,
University of Delaware, Newark 1973.
7. B. P. Flynn, "Finding a Home for the Carbon Aerator (Powdered) or Column
(Granular)." Proc. 31st Annual Purdue Industrial Waste Conference,
1976.
8. A. B. Scaramelli and F. A. DiGiano, "Upgrading the Activated Sludge
System by Addition of Powdered Activated Carbon." Water and Sewage
Works, 120: 9, 90, 1970.
9. 0. Hals and A. Benedek. "Simultaneous Biological Treatment and Acti-
vated Carbon Adsorption." Pres. 46th Annual Water Pollution Control
Federation Conference, Cleveland, 1973.
10. A. A. Kalinske, "Enhancement of Biological Oxidation of Organic Waste
Using Activated Carbon in Microbial Suspensions." Water and Sewage
Works. 115; 7, 62, 1972.
11. P. Koppe, et al, "The Biochemical Oxidation of a Slowly Degradable
Substance in the Presence of Activated Carbon: Biocarbon Unit."
Gesundeits Ingenieur (Ger) j>5_: 247, 1974.
12. A. D. Adams, "Improving Activated Sludge Treatment with Powdered
Activated Carbon." Proc. 6th Mid Atlantic Industrial Waste Conference.
University of Delaware, Newark, 1973.
13. J. F. Ferguson, G. F. P. Keay, M. S. Merrill, A. H. Benedict, "Powdered
Activated Carbon-Biological Treatment: Low Detention Time Process."
Paper presented at the 31st Annual Industrial Waste Conference: Purdue
University, Lafayette, Indiana. 1976.
161.
-------
14. F. B. DeWalle and E. S. K. Chian, "Biological Regeneration of Powdered
Activated Carbon Added to Sludge Units." Water Research 11; 439, 1977.
•
15. F. D. DeWalle, E. S. K. Chian, E. M. Small, "Organic Matter Removal by
Powdered Activated Carbon Added to Activated Sludge." Journal Water
Pollution Control Fed. 49: 593. 1977.
16. Anon, "Chemenator" Chemical Engineering 84: 1, 35. 1977.
17. P. B. DeJohn, "Carbon from Lignite or Coal: Which is Better?" Chemical
Engineering. 82:9, 13. 1975.
18. D. G. Button and F. L. Robertaccio, U.S. Patent 3,904,518,
September 9, 1975.
19. P. B. DeJohn and A. D. Adams, "Treatment of Oil Refinery Wastewaters
with Powdered Activated Carbon." Pres. at the 30th Annual Purdue
Industrial Waste Conference. 1975.
20. J. A. Rizzo, "Use of Powdered Activated Carbon in an Activated Sludge
System." Proceedings of the Open Forum on Management of Petroleum
Refinery Wastewaters, January 26-29, 1976, EPA, API, NPRA, University
of Tulsa, at the University of Tulsa, Tulsa, Oklahoma.
21. M. K. Stenstrom and C.' G. Grieves, "Enhancement of Oil Refinery Acti-
vated Sludge by Additon of Powdered Activated Carbon." Pres. at 32nd
Annual Purdue Industrial Waste Conference, 1977.
22. C. G. Grieves, M. K. Stenstrom, J. D. Walk, and J. F. Grutsch, "Effluent
Quality Improvement by Powdered Activated Carbon in Refining Activated
Sludge Processes." Pres. at the 42nd Mid-year Refining Meeting, API.
Chicago, Illinois, May 9-12, 1977.
23. G. T. Thibault, K. D. Tracy, J. B. Wilkinson, "Evaluation of Powdered
Activated Carbon Treatment for Improving Activated Sludge Performance."
Pres. at the 42nd Mid-year Refining Meeting, API. Chicagoj Illinois,
May 9-12, 1977.
24. L. W. Crame, "Pilot Studies on Enhancement of the Refinery Activated
Sludge Process." Pres. at the 42nd Mid-year Refining Meeting, API.
Chicago, Illinois, May 9-12, 1977.
25. B. R. Kim, V. L. Snoeyink, F. M. Saunders, "Influence of Activated
Sludge CRT on Adsorption." Environ. Engr. Div., ASCE, 102; 55. 1976.
26. J. F. Grutsch and R. C. Mallatt, "Optimize the Effluent System."
Hydrocarbon Processing. 55; 3: 105-112.
27. U.S. Environmental Protection Agency, Methods for Chemical Analysis
of Water and Wastes, USEPA, Washington, D.C., 1974.
162
-------
DISCUSSION
Piysuch Shah, Exxon Research and Engineering Co.: Would you please comment on the
effects of toxicant build up and on the performance of aged activated sludge units,
especially 100-150 days? Also, what is the maximum concentration that can be allowed
in the feed?
Grieves; Assume a very high sludge age of 150 days, a hydraulic residence time of 12
hours, 1 mg/liter of a toxicant (for example, chromium) in the feed, and 100% removal
of it by the activated sludge. At steady state, chromium concentration would build up to
300 mg/liter, which, in all likelihood, would be toxic to the microorganisms. However,
we have data to indicate that even at 150 days sludge age, chromium does not accumu-
late to more than 30-50 mg/liter in the sludge. We certainly have not observed any
effects of toxicant build-up — on the contrary, the 150-day sludge age reactor is the
most effective unit.
As for other toxicants — for example oil and grease and inert suspended solids
— if they are not effectively removed by prefiltration, or air flotation, they could very
well accumulate to toxic or inhibitory concentrations in the mixed-liquor. As well as
being toxic or inhibitory to microorganisms, especially nitrifiers, oxygen transfer problems
will be encountered. High inert solids concentrations may also cause overloading
problems in the final clarifies
i
Ed Sebesta, Brown & Root, Inc.: The data indicates that nitrification occurred during
some phases of the experiments while nitrification did not occur during other phases. Do
you have any comments about why this occurred?
i
Grieves: If you have ever operated a refinery waste'water treatment facility, you will
know that frequently there are excursions with nitrification. We achieved good nitrifi-
cation during the warm operating periods, phases I and II. During phase III operation, it
was relatively cool — we recorded a mixed-liquor temperature of 2°C on one occasion,
quite a severe winter for this part of Texas — and, as expected, nitrification in the
control activated sludge plant was poor. This is reflected in the probability plot (Figure
6) of the data. However, in the activated sludge pilot plants to which carbon was added,
almost complete nitrification was observed. This was unexpected because, as you know,
carbon does not normally adsorb ammonia.
Bob Smith, Carborundum Co.: Have you compared the cost effectiveness of the high
capacity Amoco carbon vs. the lower capacity carbons?
Grieves: No, we have not made this comparison yet. We have not decided whether to
go commercial with our.carbon or not. However, if and when we do decide to
commercialize our product, you can rest assured that it will be cost effective with other
commercially available carbons.
163
-------
BIOGRAPHIES
Colin G. Grieves is a Research Engineer in the
Water Conservation group at Amoco Oil Company's
Research and Development department in Naperville
Illinois. He has M.S. and Ph.D. degrees in
Environmental Systems Engineering from Clemson
University, Clemson, South Carolina, and a B.Sc.
degree in Civil Engineering from the University
of Newcastle Upon Tyne, England. Previous
employment was in the Public Health Engineering
Division of Babtie Shaw & Morton, Consulting
Civil and Structural Engineering in Glasgow,
Scotland. Colin is the author of several papers
in the wastewater treatment field.
Michael K. Stenstrom is a Research Engineer in
the Water Conservation group at Amoco Oil
Company's Research and Development department in
Naperville, Illinois. He has a B.S. degree in
Electrical Engineering and M.S. and Ph.D. degrees
in Environmental Systems Engineering from Clemson
University, Clemson, South Carolina. Mike is the
author of several papers dealing with various
aspects of municipal and industrial wastewater
treatment.
1R4
-------
BIOGRAPHIES
Joe D. Walk is Active Carbon Project Director in
the Corporate Development Department of Standard
Oil (Indiana), Chicago, Illinois. He has a B.S.
degree in Chemical Engineering from the University
of Texas. Previously he served as Process
Coordinator in air/water conservation, crude
running and product treating for Amoco Oil's ten
refineries. Prior assignments include Manager
of Technical department at Texas City refinery,
as well as positions in New York City, New Orleans,
and El Dorado, Arkansas, during his 31 years
service with Standard Oil.
James F. Grutsch is Coordinator-Environmental
Projects, Standard Oil Company (Indiana). He
holds undergraduate and graduate degrees in
chemistry from Indiana University. Prior to his
present assignment with Standard, Jim served
successfully as Group Leader for finishing,
blending and reclamation at the Amoco Oil
Whiting refinery, and Coordinator of Waste
Disposal for Amoco. Jim taught undergraduate
chemistry for 6 years at Indiana.
165
-------
TABLE 1 OPERATING CONDITIONS
Pilot Plant Conditions
Aeration Zone Volume 36.7 liters
Settling Zone Volume 5.7 liters
Nominal Flow Rate 2.45 liters/hr'.
Nominal Hydraulic Retention Time 15.0 hours
Nominal Settling Time 2.33 hours
Air Flow Rate 300 liters/hr.
pH 6-8.5
Caustic Addition Rate 0.12-0.30 liters/hr.
Phosphorous Added to Feed 3 mg/liter
Temperature , Ambient (4-31°C)
Analytical Work
Frequency Analysis Performed^, 28
Daily - Influent and mixed-liquor pH, temperature,
influent flow rate, caustic addition rate.
Carbon addition and sludge wastage.
3 Times Influent and effluent total and volatile sus-
a Week pended solids, soluble organic carbon, soluble
chemical oxygen demand, soluble ammonia nitro-
gen, and soluble phenolics. Mixed-liquor
suspended solids and mixed-liquor volatile
suspended solids. Sludge volume index.
Once a Material balances to calculate quantity of
Week sludge to be wasted to maintain desired sludge
age.
166
-------
TABLE 2 PROPERTIES OF POWDERED ACTIVATED CARBONS
Property
-: Carbon Designation
Experimental Amoco
High Surface Area -
Surface Area
BET, m2/g
Pore Volume, cc/g
> 15 A° Radius
< 15 A° Radius
Iodine Number
Methylene Blue Adsorption,
Phenol Number
Bulk Density, g/cc
Screen Analysis
Passes 100 Mesh, Wt.%
Passes 200 Mesh, Wt.%
Passes 325 Mesh, Wt.%
Al
Grade PX-21
3099
0. 16
1.45
3349
mg/g 586
12.8
0.298
98.4
92.7
84.1
A2
Grade PX-23
3148
0.43
1.60
3375
550
12.6
0.228
99.1
93.4
80.8
Commercially Available Conventional Surface
Area Carbons
B
717
0.28
0.51
1790
100
34.1
0.610
99.2
86.7
60.6
C
514
0.38
0.11-0.42
920
83
22.9
0.576
100.0
94.4
68.3
D
532
0.03
0.25
888
50
23.8
0.484
100.0
97.9
91.8
Molasses Number
10
205
103
85
0
-------
TABLE 3
PHASES I AND II - EFFECT OF CARBON TYPE AND ADDITION RATE ON EFFLUENT QUALITY*
50% Probability Data During 30 Days of Steady-State Operation
Sludge Age = 20 Days
Concentration, mg/liter
Filtered Pilot Plant Effluent
Component Influent No Carbon Carbon Al Carbon B Carbon C Carbon D
Phase I: Carbon Addition Rate = 100 rag/liter
Equil. Mixed-Liquor Temp = 31°C. Carbon Cone = 3200 mg/liter
soc
SCOD
NH3-N
Phenolics
Equil
SOC
COD
NH3-N
Phenolics
72.0
230
25.8
4.35
22.0
73
0.5
0.018
Phase II: Carbon
. Mixed-Liquor Temp
70.0
230
25.4
4.06
26.5
58
0.2
0.020
12.5
28.5
0.2
0.003
Addition Rate
= 25°C, Carbon
9
17
0.2
0.001
17.5
48
0.5
18.5
44
0.5
0.010 0.010
= 200
Cone
13.5
24
0.2
0.001
mg/liter
= 6400 ma/liter
15.5
28
0.1
0.003
23.0
65
0.5
0.017
* BATEA effluent standards in mg/liter are: Soluble Organic Carbon (SOC) 15
Soluble COD (SCOD) 24
Ammonia Nitrogen (NH3-N) 6.3
Phenolics 0.02
168
-------
TABLE 4
PHASE III - EFFECT OF CARBON TYPE AND ADDITION RATE, SLUDGE AGE,
AND INFLUENT PRETREATMENT ON EFFLUENT QUALITY
50% Probability Data During 30 Days of Steady-State Operation
Equil. Mixed-Liquor Temp = 14°C, Carbon Cone. = 2400 ing/liter
Influent
Pretreatment
Filtered Feed
Unfiltered
Filtered
Filtered
Filtered
Filtered
Filtered
Filtered
B
Al
Al
A2
Carbon
Addition
Type Rate, mg/liter
Influent Concentration, mg/liter
SOC COD NH3-N phenolics
73.5 294.5 19.3 3.95
Effluent Concentration, mg/liter
Sludge Age = 20 Days
32.0 103.5 12.1 0.027
29.0 83.0
Sludge Age = 60 Days
25.0 65.9
100 16.0 40.3
50 12.0 27.5
25 16.0 50.3
50 13.0 31.0
14.5
0.027
5.1
0.2
0.1
0.4
1.8
0.019
0.001
0.002
0.006
0.004
169
-------
TABLE 5
PHASE IV - EFFECT OF HIGH SLUDGE AGE, LOW CARBON ADDITION RATE,
AND DECREASED HYDRAULIC RETENTION TIME ON EFFLUENT QUALITY*
50% Probability Data During 30 Days of Steady-State Operation
Equil. Mixed-Liquor Temp = 27°C
Type
B
Al
Al
Al
Carbon
Addition
Rate, mg/liter
-
25
25
25
10
Sludge
Age, days
60
60
60
60
150
Hydraulic
Retention
Time, hr
15
15
15
7.5
15
Equil. Mixed
Liquor Carbon
Cone, mg/liter
-
2400
2400
4800
2400
Effluent Cone,
mg/liter
SOC
29
22
18
17
16
COD
99
64
52
46
49
NH3-N
0.1
0.1
0.1
0.3
0.1
Phenolics
0.018
0.010
0.010
0.010
0.010
* Filtered influent contained 78 mg/liter SOC, 270 mg/liter COD, 29 mg/liter NH-j-N, and 3.25 mg/liter
phenolics.
-------
TABLE 6
PHASE III - BAT GUIDELINE AND ACTUAL VARIABILITY FACTORS FOR PILOT
PLANT FED 25 mg/liter OF CARBON Al
BAT Guideline
Parameter
Soluble Organic Carbon
Soluble COD
NH3-N
Phenolics
Variability
Daily Max.
1.6
2.0
2.0
2.4
Factor
30 Day Max.
1.3
1.6
1.5
1.7
Actual
Variability Factor
Daily Max. 30
2.8
7.5
2.1
5
Day Max.
2.8
7.5
2.0
5
-------
FIGURE 1. 'SIMPLIFIED REFINERY BPT WASTEWATER
TREATMENT SYSTEM
Chemicals
isa
Refinery
waste- "
water
i
Slop 01
to
Treatme
Gravity
Separator
~i
i <
1 Slue
to
nt Trea
,, Dissolved Equalizat- Aeration f \ Granular Final
"" Air ion Tank ~»fclarif idl— * Media —Effluent
Flotation \^ ^/ Filter
t Sludge
Recycle
itment
-------
FIGURE 2. SCHEMATIC OF ACTIVATED SLUDGE REACTOR
USED IN PILOT PROGRAM
Air
Air — »— »
36.7 Liters
Aeration
Zone
ss
X
\/
r \~
~~° \
v /I
Side
affle
Feed -*•«=
5.7 Liters
Cp*-f-1 \ no.
Zone
— Sludge
Blanket
s=
(
^-
C
Enc
i
1
sr
^
a* Chemicals
Sampling
=="*~ Port
\Diffuser
Stone
1
Chemicals »==]
Air_»e=
Feed — ».c=
.. _
;,— -—.-_...--.--.
ffi °
HH
S| o
H 4 w
*i
1
Plexiglass
^/Construction
_a ^Effluent
Access
Hole
Top
-------
c
o
§
o
O
O
CO
0)
D
LU
FIGURE 3. EFFECTOR MIXED-LIQUOR CARBON
CONCENTRATION ON EFFLUENT SOC
Legend
20 days sludge age
60
150
carbon
dose
mg/l 100
t reatment objective
-150 days
1 2 3 A
Mixed-liquor carbon concentration
-------
o
o
FIGURE 4 - SOLUB'LE ORGANIC CARBON - PHASE 3
CONTROL, NO CARB8N. 60 DAY SBT- X
AMOCO PX-2t, 25 HO/L, 60 DAY SRT- +
AMOCO PX-21, SO UO/L. BO DAY SRT-A
CARBCH B, »00 U8/L, 60 DAY SRT - g>
LJ
CO
o
•
o
*
CO
0.013
o.t o.z o.s 0.4 O.B o.e o.r o.e 0.9
CUMULATIVE PROBABILITY
0.93
175
-------
O
O
FIGURE 5 - SOLUBLE CHEMICAL OXYGEN DEMAND
PHASE S
CONTROL, NO CARBON, 80 DAY SRT - X
AMOCO PX-JI, 25 M8/L, 80 DAY SRT- +
AMOCO PX-21, SO UO/L, 80 DAY SRT«*
CARBON B, 100 H8/L, 80 DAY 8BT - O
o:
LLJ
CD
*
O
*
O
O
*
CO
0.023
0,2 0.3 0.4 0.5 0.6 0.7 0.8
CUMULATIVE PROBABILITY
0.9
0.977
176
-------
FIGURE 6 - AMMONIA-NITROGEN - PHASE 3
CONTR81, NO CARBON, 60 DAY SRT-X
AMOCO PX-ai, IS IIO/L, 60 DAY SRT- +
AMOCO PX-21, SO M8/L, 60 DAY SftT-A
CARBON 6, 100 MC/l, 60DAYSRT-O
o:
UJ
I
to
X
2:
0.013
O.t 0.2 O.S 0,4 0.5 0.0 0.7 0.6 0,0
CUMULATIVE PROBABILITY
0.987
177
-------
FIGURE 7 - PHENOLICS - PHASE 3
to
ro
o
0,036
CONTROL, N8 CARBON, 80 DAY SRT-X
AMOCO PX-21, 25 US/L, 60 DAY SRT- +
AMOCO PX-21, 50 M6/1, 60 DAY SRT - *
CARBON B, tOO H6/L. 60 DAY 3RT - O
0.2 0.9 0.4 0.5 0.6 0.7 0.6
CUMULATIVE PROBABILITY
0.9
0.964
178
-------
"TREATMENT OF OIL REFINERY WASTEWATERS
WITH POWDERED ACTIVATED CARBON"
Paschal B. DeJohn
Manager, Purification Sales, ICI United States Inc. (Delaware)
James P. Black
Industry Coordinator, ICI United States Inc. (Delaware)
INTRODUCTION
The effectiveness of powdered carbon as an additive to improve activated
sludge treatment has been demonstrated in a variety of industrial and munici-
pal plants. This type of treatment has gained wide acceptance in the past
few years and is currently an essential part of treatment at 60-80 plants.
These plants range in size from 10,000 gpd package units located along the
Alaska Pipeline to the very sophisticated 40,000,000 gpd PACT treatment plant
at the DuPont Chambers Works in Deepwater, New Jersey. At least four petrol-
eum refiners currently use powdered carbon as an integral part of their waste
treatment scheme.
HOW THE PROCESS WORKS
The reason powdered carbon has gained such acceptance treating a wide
variety of waste streams is the extreme flexibility .which can be employed in
its usage.
• The amount of carbon used can be varied to meet the
treatment requirements as they change.
• Higher COD or BOD removal than is usually obtainable
by conventional biological treatment can be achieved.
• The combination of activated carbon in a biological
system provides more effective treatment than either
of the processes would if used singularly.
Carbon aids the biological process two ways:
1. By direct adsorption of pollutants.
2. By providing a more favorable environment for the micro-
organisms to propogate.
Adsorption is an equilibrium phenomenon. In general, carbon preferen-
tially absorbs higher molecular weight compounds. Given a related series of
organic compounds; for example, alcohols, one finds that the lower molecular
weight alcohols (methanol, ethanol) are not appreciably absorbed by carbon
while the higher molecular weight alcohols are. Fortunately, compounds which
are poorly absorbed (weakly held by the carbon) are usually compounds which
are the most amenable to biological treatment.
179
-------
We can generally classify organic compounds into three broad categories
with respect to their adsorptability onto carbon.
1. Compounds which are readily adsorbed. Th^se compounds are
usually "tightly held" by the carbon. And consequently,
they are not readily desorbed.
^
2. Compounds which are adsorbed with difficulty. These
compounds are desorbed easily.
3. Compounds which are poorly adsorbed.
Organic compounds can also be classified in terms of their susceptibili-
ty to biodegradation:
1. Compounds which are readily and rapidly biodegraded.
2. Compounds that are degraded slowly.
3. Compounds that are not biodegraded. Many of these can
function as toxicants in a biological system.
s
It is important for one to understand the interaction of carbon and the
microorganisms present in an activated sludge system. Exhibit 1 does this by
considering how a carbon-biological system handles each of the above classi-
fications of organics compounds. Relative adsorptivity and biodegradability
for organic compounds was taken from an EPA source (Reference 1).
The boxes in Exhibit 1 have been numbered from 1 through 9 and are
interpreted as follows:
Degree of
Box Biodegradability Adsorptability
1 Rapid Strong
2 Rapid Moderate
3 Rapid Weak
4 Slow Strong
5 Slow Moderate
6 v Slow Weak
7 None Strong
8 None Moderate
9 None Weak
The classes of compounds which are represented by boxes 1, 2, and 3 would be
handled quite easily by the microorganisms in a carbon-biological system.
Those compounds which are represented by boxes 1, 4, and 7 would be removed
by direct adsorption on the carbon. Compounds which fall in box 1 (both
rapidly biodegradable and strongly adsorbed) are few in number. The only
example that we could find is o-cresol.
£o Boxes 4. 5, and 6 represent compounds which are slowly biodegradable.
These compounds probably would not be removed very effectively in a conven-
tional activated sludge system. In a carbon-biological system, compounds in
180
-------
box 4 are removed by direct adsorption and are held very tightly by the
carbon. Compounds in boxes 5 and 6 are retained in the system by moderate or
weak adsorption. Because carbon is adsorbing these compounds, their concen-
tration in the liquid stream is reduced. The microorganisms are able to
degrade the organics in reduced concentrations, and as they do the equilli-
brium between the carbon and these organics in the waste is disturbed.
Because these compounds are not held very tightly by the carbon, they are
readily desorbed back into the system and a new equillibrium is established.
In this fashion, the carbon is acting as a storage area keeping the concen-
tration of slowly biodegradable organics at a level where they can be handled
by the microorganisms. Compounds which fall into the categories represented
by boxes 4, 5, and 6 are effectively handled in a carbon-biological system
because of synergistic effects. It is primarily these compounds that are
removed more effectively in a carbon-biological system as compared to either
process operating singularly.
Compounds represented by boxes 7, 8, and 9 are not biodegradable. And
in some cases, these compounds are actually toxic to microorganisms. In our
opinion, carbon performs its most beneficial action in this area. Compounds
which fall into the category represented by box 7 are removed by direct
adsorption and are held very tightly by the carbon. Compounds in box 8 are
removed by direct adsorption, and even though they are held very loosely by
the carbon, it is difficult to disturb the equillibrium. This is because
the concentration of these organics remaining in the waste stream are not
being degraded by the microorganisms. The compounds represented by box 9
are the only ones that cannot be handled very effectively in a carbon-biolo-
gical system. Fortunately, there are very few organics which are both non-
biodegradable and weakly adsorbed by carbon.
Examples of the different compounds are shown in the various boxes in
Exhibit 1.
One of the important effects of carbon in the system (not related dir-
ectly to adsorption) is the higher levels of biomass that can be used because
of the density and "weighting effect" of the carbon. (Both the use of great-
er sludge mass and the temporary retention of slowly degraded compounds by
the carbon gives more time for the compounds to be consumed biologically).
Carbon adsorbs the pollutants and oxygen, localizing them for bacterial
attack. Because the aerobic action iei dependent upon the concentration of
the reactants, this localizing effect serves to drive the reaction further
towards completion resulting in .improved BOD removal (Reference 2).
Many pollutants that are not biologically degraded in a conventional
activated sludge system would be if they were in contact with the biomass
for a longer period of time. When absorbed by the carbon, these molecules
settle into the sludge. Contact time is thereby, extended from hours to days.
This results in lower effluent COD's and TOC's. High density powdered carbons
improve solids settling in the secondary clarifiers. This results in lower
effluent suspended solids and also a reduction in BOD. Under high organic
load conditions which normally would lead to sludge bulking, the dense carbon
181
-------
will act as a weighting agent keeping the sludge in the system. When dis-
persed biofloc results due to low organic loads, carbon serves as a seed for
floe formation preventing loss of solids. Under these conditions, phosphorous
and nitrogen removal are generally enhanded.
Powdered carbon improves treatment in activated sludge process because
of its adsorptive and physical properties. Powdered carbon, can be added to
any convenient point in the activated sludge process to get it into the
aerator. Direct addition to the aerator, sludge return lines, influent
channels, or through the secondary clarifier are all possibilities. It is
not necessary to add carbon continuously in most cases. A dense, easily
wetted carbon can be added dry or in slurry form with water.
RESULTS
14 refineries have evaluated HYDRODARCO powdered activated carbons in
full scale activated sludge systems during the past three years. The first
treats a 2.2 MGD flow with an average BOD of 400 ppm in a 1.2 million gallon
aerator. Mixed liquor solids are maintained at 3600 ppm (2880 ppm volatile).
Waste activated sludge is digested aerobically, centrifuged, and hauled to
landfill. Despite a secondary clarifier overflow rate of only 423 gallons/
ft.^ and use of 22 ppm cationic polymer for secondary solids capture, efflaent
solids averaged in excess of 100 ppm. Toxic loads caused periodic loss of
aerator biosolids. Defoamer costs averaged $200/day for aerator foam control.
HYDRODARCO C, a high density, lignite based powdered carbon, was added
to the aerator over a four and one-half month period. Eventually, the equil-
librium carbon level reached 1800-2000 ppm. At the sludge solids concentra-
tion obtained and wasting rates employed, it was possible to maintain this
level with a daily average carbon dose of only 20 ppm.
Over the entire carbon test period, average BOD reduction equaled 82%
versus 23% during the post test control period (Figure 1). As carbon built
up in the system, BOD removals reached the 90-95% range, and the plant was
able to meet their 30 ppm BOD effluent standard. Effluent COD was reduced
from an average of 1180 ppm without carbon to 350 ppm with (Figure 2). Aver-
age effluent TOG decreased from 420 ppm to 100 ppm (Figure 3), and total
carbon decreased from 520 ppm to 180 ppm (Figure 4 ). The lower slope of the
carbon plots also indicate the decreased variability in effluent quality with
carbon present.
HYDRODARCO C had a dramatic effect on the reduction of oil through the
system (Figure 5). The effluent concentration was reduced by 75% (average),
and the range was narrowed as well.
Both removal of the oil by the powdered carbon and the weighting effect
of carbon resulted in lower effluent solids (Figure 6). Prior to carbon
treatment, the plant used polymer at a dosage of 20 ppm, but still experienced
poor solids settling. When carbon was added to the system, solids settling
improved, and the polymer dosage was cut in half. Since effective solids
settling could not be achieved with the use of carbon or polymer alone, it
appears that the combination of the two was required to attain the desired
results. This, of course, represents an operating cost savings for the plant.
182
-------
Improved solids settling increased sludge thickening which allowed a 65%
reduction in sludge wasting. Again, savings on the operation of the centri-
fuges, including power and labor, occurred.
Use of HYDRODARCO C eliminated the need for aerator defoamer. Removing
the foaming agents from the wastewater by adsorbing them eliminated foam
problems in the receiving stream. Defoamers only suppress foam in the aera-
tor and do not prevent its reappearance in the effluent. Carbon can reduce
operating costs by allowing surface aerators to aerate and mix the activated
sludge rather than expand energy generating foam.
Both nitrogen (Figure 7 ) and phosphorous (Figure 8 ) removals were
improved with powdered carbon. Reason: carbon adsorbs compounds toxic to
nitrifiers and allows them to operate at normal levels. In this waste,
neither nitrogen nor phosphorous were limiting for bacteria growth. Increased
nitrogen removals are attributed to the fact that the dense carbon settled
the nitrifying organisms which normally would float out of the system. The
result is a longer solids retention time which is more favorable for nitrifi- .
cation to occur. By the same token, improved solids settling is probably
the reason for decreased phosphorous levels. We suspect that the phosphorous
is precipitated with the carbon-biosolids floe and removed in the sludge
rather than degraded biologically.
While the exact reason for the bug kills prior to carbon was not known,
upsets were greatly reduced with carbon in the aerator. A possible explana-
tion is the effect carbon had on the removals of heavy metals such as zinc
(Figure 9). The ability of activated carbon to adsorb heavy metals from
wastewater has been established elsewhere (References 3, 4, and 5).
The second evaluation was conducted at a 12 MGD plant treating an
average 12 MGD flow. The TOC of the raw waste ranged from 100-1000 ppm,
averaging about 200 ppm. Major treatment problems included aerator foaming
caused by alkanolamines in the waste; high effluent TOC; oily, difficult-to-
handle sludge; and high effluent solids.
Effluent TOC's were maintained below 20 ppm during shock load periods.
This was well within the standard of 50 ppm. In a post test control phase,
a deterioration in effluent quality was observed as carbon was lost from the
system through sludge wasting.
A third evaluation was conducted at a 2.5 MGD plant treating a 550 ppm
COD refinery waste in a two stage, conventional activated sludge system.
Carbon was added to the second stage aerator over a six week period. A con-
stant daily carbon dose was maintained for each week and increased in
succeeding weeks. No sludge was wasted intentionally during this time.
Optimum treatment was found at relatively high influent dose of 200 ppm.
Effluent solids and COD removals increased 40%, and BOD removals, already
high, increased 10%.
183
-------
The major finding of this study was the increased removals of cyanide
with carbon in the aerator. In conjunction with CuSO^ treatment, the average
cyanide levels decreased from 1 ppm before carbon to 0.05 ppm. The precise
nature of this removal is not known and bears further investigation.
Results from a fourth study are summarized in Exhibit R for three
separate carbon addition periods. This plant is a current user of activated
carbon to treat 2.2 MGD flow. They have reported the following results
from carbon addition (Reference 6):
56% reduction of suspended solids.
36% reduction of COD.
76% reduction of BOD.
Foam problem elimination.
This improved plant performance is achieved at a carbon cost of
1.7-4.3
-------
BIOGRAPHY
Mr. Paschal DeJohn is Manager of
Purification Sales and Project Leader in Activated
Carbon at ICI United States Inc.; holding this
position since 1972. Mr. DeJohn holds a B.S.
degree in chemistry from Westchester State College
and an M.B.A. degree from Widener College,
Pennsylvania. Previous to his present position Mr.
DeJohn was Water & Wastewater Treatment District
Engineer with Drew Chemical, Parsippany, New
Jersey. He is 1977 Technical Conference Chairman
for WWEMI and the 1978 General Conference
Chairman for this ame organization. Mr. DeJohn
has been involved with water and wastewater treat-
ment for the past twelve years.
James P. Black is Industry Coordinator for
Water Purification at ICI United States Inc. He
has a B.S. degree in chemical engineering from
the University of Texas at Austin. Prior employ-
ment included iCl's Research & Development
Laboratory, and Corporate Planning Staff.
185
-------
DEJOHN PAPER DISCUSSION
Ed Sebesta, Brown & Root; Have you observed any situations where you
mentioned some compounds are loosely adsorbed or difficultly adsorbed and
then quite easy to be desorbed? Have you ever seen any situations or heard
of situations where because of changing influent situations you may suddenly
desorb an accumulation of adsorbed materials and effect the system in that
way?
DeJohn; Potentially that can occur. I would think something like that would
be more prone to happen in a granular carbon system. You can design a
granular carbon system around this however. In the PACT systems we've been
involved in, I'm not aware of any that have desorbed back an accumulation of
adsorbed material. But there's always the possibility that this can happen.
REFERENCES
1. EPA Contract 68-01-2926, April 1, 1975.
2. Adams, A.D., "Improving Activated Sludge Treatment with Powdered Activa-
ted Carbon - Textiles" presented at the 6th Mid-Atlantic Industrial
Waste Conference, University of Delaware, November]15, 1975.
3. Esmond, S. E. and Petrasek, A. C., "Removal of Heavy Metals by Waste-
water Treatment Plants," presented at the WWEMA Industrial Water and
Pollution Conference, Chicago, Illinois, March 14*16, 1973.
4. Sigworth, E. A. and Smith, S. B., "Adsorption of Inorganic Compounds by
Activated Carbon," Journal AWWA, Water Technology/Quality, June, 1972
(p. 306).
5. Linstedt, K. D. et. al, "Trace Element Removals in Advanced Wastewater
Treatment Processes," Journal WPCF, 43, No. 7, 1507 (July, 1971).
6. Rizzo, Joyce A., "Case History: Use of Powdered Activated Carbon in an
Activated Sludge System", presented at the Open Forum on Management of
Petroleum Refinery Wastewaters, Tulsa, Oklahoma, 1976.
186
-------
EXHIBIT 1
CARBON ADSORPTION
y
O
O
_i
O
co
TRIPHENYL
PHOSPHATE
None
(or toxic)
187
-------
EXHIBIT R
No Carbon First Carbon Period
COD
Influent 459 ppm 457 ppm
Effluent 170 ppm 135 ppm
% Removed 63 70
BOD
Influent 152 ppm 213 ppm
Effluent 15 ppm 15 ppm
% Removed 90 93
SUSPENDED SOLIDS
Effluent 115 50
COD
Influent 343 ppm 444 ppm
Effluent 266 ppm 183 ppm
% Removed 23 59
BOD
Influent 152 ppm 227 ppm
Effluent 30 ppm 14 ppm
% Removed 80 94
SUSPENDED SOLIDS
Effluent 162 72
COD
Influent 367 ppm 379 ppm
Effluent 166 ppm 112 ppm
% Removed 55 70
BOD
Influent 188 ppm 207 ppm
Effluent 12 ppm 3 ppm
% Removed 94 99
SUSPENDED SOLIDS
Effluent 79 42
188
-------
EXHIBIT S
Flow: 8 MGD design, 6 MGD actual.
Carbon Dose: 500 ppm in aerator.
MLVSS: 2900 ppm before carbon
v 1500 ppm during carbon
(A 50% loss of aerator solids resulting from continuing same
volumetric wasting rate of the more dense carbon sludge.)
2500 ppm after carbon
• BOD removal - 55% before carbon addition
70-80% during carbon addition
60% after carbon addition
• Influent oil concentrations were so high (100+ ppm) during
test that effects of carbon were overshadowed.
189
-------
% BOD
REMOVED
100
90
80
70
60
50
40
30
20
10
EFFECT OF POWDERED CARBON ON
BOD REMOVAL
HYDRODARCO C
5 10
30 50 70
90 95
99
Fig. 1
% OF VALUES LESS THAN
-------
EFFLUENT
COD, PPM
1000
900
800
700
600
500
400
300
200
100
EFFECT OF POWDERED CARBON ON
EFFLUENT COD
HYDRODARCO C
I I
I I I I I I I
I I
1
5 10
30 50 70
90 95
99
FIGo 2
% OF VALUES LESS THAN
-------
1000
900
800
700
600
IVJ
EFFLUENT
TOC, PPM
500
400
300
200
100
1
EFFECT OF POWDERED CARBON ON
EFFLUENT TOTAL ORGANIC CARBON
I
BLANK
HYDRODARCO C
I I
I I I I I I I
I I
5 10
30 50 70
90 95
99
FIG. 3
% OF VALUES LESS THAN
-------
to
CO
1000
900
800
700
600
EFFLUENT TC,
PPM 500
400
300
200
100
EFFECT OF POWDERED CARBON ON
EFFLUENT TOTAL CARBON
I I
I
I
/
BLANK
HYDRODARCO C
I I 1 I I I I
I I
5 10
30 50 70
90 95
99
FIG. 4
% OF VALUES LESS THAN
-------
100
90
80
70
60
EFFLUENT OIL
PPM
40
30
20
10
0
EFFECT OF POWDERED CARBON ON
EFFLUENT OIL
BLANK
I I I I I I I I
HYDRODARCO C
ill I
5 10
30 50 70
90 95
99
FIG. 5
% OF VALUES LESS THAN
-------
1000
900
800
700
600
EFFLUENT
SUSPENDED 500
SOLIDS, PPM
400
300
200
100
EFFECT OF POWDERED CARBON ON
EFFLUENT SUSPENDED SOLIDS
HYDRODARCO C
I I
J 1 J 1 I I
I
I
5 10
30 50 70
90 95
99
FIG. 6
% OF VALUES LESS THAN
-------
zo
1.8
1.6
1.4
EFFLUENT 1.2
NITROGEN
PPM 1-0
0.8
0.6
0.4
0.2
EFFECT OF POWDERED CARBON ON
EFFLUENT NITROGEN
BLANK
*•
I
I I I I I I I
HYDRODARCO C
I I
1
5 10
30 50 70
90 95
99
FIG. 7
% OF VALUES LESS THAN
-------
2.0
1.8
1.6
1.4
EFFLUENT 1.2
PHOSPHOROUS,
PPM 1.0
0.8
0.6
0.4
0.2
EFFECT OF POWDERED CARBON ON
EFFLUENT PHOSPHOROUS
I I
BLANK /
I
I
I
HYDRODARCO C
I I I I I I i
I I
5 10
30 50 70
90 95
99
FIG. 8
% OF VALUES LESS THAN
-------
1000
900
800
700
600
EFFLUENT 500
ZINC; PPM
400
300
200
100
EFFECT OF POWDERED CARBON ON
EFFLUENT ZINC
HYDRODARCO C
II I I I I I I I II
5 10
30 50 70
90 95
99
FIG. 9
% OF VALUES LESS THAN
-------
SECTION 5
.Full-Scale granular Activated Carbon Treatment
ACTIVAT£D.£ARa_ON TREATMENT OF COMBINED STORM AND PROCESS WATERS
M. A. Prosche
Atlantic Richfield Company, Watson Refinery, Carson, California
The Watson Refinery of the Atlantic Richfield Oil Company is located adjacent to the
Dominguez Channel in Los Angeles County. This Channel is a non-navigable dredged
tidal estuary which is lined with rip-rap and used specifically for the discharge of refinery
and chemical plant waste waters and for rain water runoff.
In 1968, the Los Angeles Regional Water Quality Control Board made a study of the
Dominguez Channel and determined that petroleum and chemical plant discharges were
causing a problem due to the oxygen demand of their waste waters entering the Channel.
The Control Board, in accordance with these findings, issued a resolution in February, •
1968, which limited the total chemical oxygen demand (COD) from all industrial discharges
into the Channel. These discharges also included any contaminated rain water runoff.
The resolution was to be complied with by February, 1971.
As defined by the resolution, the Watson Refinery was limited to 1330 pounds per day
of COD in its discharge water to the Channel. Meeting this requirement meant reducing
the COD in its discharge waters by 95 per cent.
Fortunately, the Watson Refinery as a taxpayer of Los Angeles County was able to make
arrangements with the Los Angeles County Sewer District to have its process waste water
handled in the County's primary treatment unit. However, due to limitations in the County
unit, the County was unable to handle rain water runoff. This presented a problem for
the Watson Refinery due to the fact that the rain water collection facilities were inter-
connected with the process waste water collection system. Therefore, during periods of
rainfall, the process waste water and rain water mixture could not be sent to the sewer
district facilities, nor could it be sent to the Dominguez Channel due to high COD content
of the process waste water.
To solve this problem a system was needed which would treat all the process water
plus rain water during the rainy season for the removal of COD and allow its discharge to
the Channel. A system was needed which could be started up easily when rain fell and
then shut down when no longer required. The system selected was impounding of rain plus
process water during the storm followed by activated carbon treatment to adsorb the COD
material.
ADSORPTION SECTION DESIGN AND DESCRIPTION
Since it was planned to operate the plant only during the rainy season, it was felt
there was no need to design for continuous carbon regeneration and changeout of the beds.
Therefore, the design was based on use of only the carbon beds during the rainy season,
and then regeneration of all the beds during the dry summer months in preparation for the
next rainy season. The significant design criteria for the plant and impounding basin were
as follows:
199
-------
Design Criteria
Flow Rate 100,000 BPD
3,000 GPM
Inlet COD 250 ppm
Outlet COD 37 ppm
Impounding Basin 1.2 million barrels capacity
Following the accumulation of 0.1 inches of rainfall, regulations call for diverting
process water from the sewer system. This is accomplished automatically by means of a
solenoid operated valve actuated by a rain gauge. Process water plus commingled rain
water is then switched to the 1 .2 million barrel impounding basin. Depending upon the
specific situation, the carbon plant can be started up immediately for discharge to the
Channel or it can be started up later following the rain. A flow diagram of the adsorption
section of the plant is shown in Figure 1 .
Impounded water is delivered to the plant through a 14" line from the basin. The water
is then delivered to a distribution trough where it is distributed to 12 adsorber cells by
adjustable slide gates. Each of the twelve cells is 12' by 12' square and 26' feet deep.
Each cell originally contained 13' of carbon having a dry weight of about 50,000 pounds.
The water passes down through the carbon bed where it collects in the underdrain system.
Supporting the carbon is a one-foot layer of gravel on top of a Leopold tile underdrain
system. The treated water then flows through 6" lines from each cell to a 24" collection
header leading to the effluent retention sump. Each 6" discharge line has a sample point
and an air-operated pinch valve which can be shut during backwashing.
If necessary to further control COD, a chlorine-water solution may be injected into the
incoming treated water stream. Approximately 15 minutes retention time is allowed for
chlorine contact in this sump. From the sump the water flows by gravity to the Channel.
Each carbon bed must be backwashed whenever it will not pass its share of water flow
due to buildup of solids on top of the carbon. This is indicated by the rise in the level of
the water in each carbon cell. When the level rises to the height of the backwash troughs,
the flow to and from the bed is stopped, and treated water from the backwash sump is pumped
up through the bed to expand it and flush out accumulated solids. The turbid water overflows
into the backwash troughs to the backwash effluent sump where it is pumped back to the
reservoir for settling and retreating.
EPA DEMONSTRATION GRANT
Following completion of the plant, a demonstration grant was received sponsored by the
Water Quality Research Division of Applied Science and Technology of the Environmental
Protection Agency. The specific objectives of the demonstration project were as follows:
1. Determine feasibility of activated carbon as a treatment system for storm runoff and
refinery process waters.
2. Evaluate performance of the system.
3. Determine operating costs.
4. Assess reliability of the system.
200
-------
Length of
Run, Hours
44
48
38
18
95
22
~ Table 1
Feed
Rate, GPM
3000
3000-2000
2000
2000
1000
2000
Average COD
Feed
326
360
374
310
237
147
Effluent
43
48
86
67
100
93
The report relative to this project has now been published by the EPA.
ADSORPTION SECTION OPERATION AND PERFORMANCE
The carbon plant was first placed in operation in May, 1971 for some preliminary test
work prior to the rainy-season. At that time the test water was synthesized using process
water diluted with service water. Operation with rain water and process water was not
required until later that year in December. At that time sufficient rain fell to require
placing the unit in full operation at the design rate of 3,000 gpm total to all twelve cells.
Subsequent to that a total of 6 runs were made processing impounded water from the inter-
mittent rain storms that followed. The performance data for all of these runs are shown in
Table 1.
\
Typical Performance Data First Rainy Season
Run
Number
2
3
4
5
6
During the initial runs the COD of the feed was higher than design and the plant was
unable to produce the desired removal. Therefore, the feed rate was reduced to 2,000
gpm. Even at this lower rate the effluent COD content was greater than the regulations
allowed and feed rate was again reduced to 1,000 gpm. By the end of the season the
reservoir COD had dropped to 147 ppm due to rain dilution and the rate was increased to
2000 GPM. However, performance continued poor even at this low concentration.
A good evaluation of the performance of the plant during this first season of operation
is difficult due to the significant variations in feed COD and the necessity of having to
reduce the feed rate from 3000 to 1000 gpm. However, COD data for each cell was
recorded during the course of the run and some generalizations are possible. At the end
of the season the COD loading in each cell varied from 0.2 pounds of COD per pound of
carbon to a high of 0.3 pounds of COD per pound of carbon. During the season a total
of 52,000,000 gallons of water were processed. The average feed COD concentration
was 377 ppm while the effluent averaged 67 ppm. The carbon loading averaged 0.23
pounds COD per pound of carbon. It also appeared that the potential maximum carbon
loading at a constant effluent COD concentration was quite sensitive to the COD of the
feed.
Other data collected during the first season's operation included measurement of the
carbon's adsorptive ability at various depths in the bed after a period of operation. One of
the cells was taken out of service after 208 hours of operation and the carbon was sampled
at various depths to test its relative adsorptive efficiency as compared to virgin carbon.
The data are shown in Figure 2. The data show that the top two feet of carbon were
completely exhausted while the last five feet were still 65 per cent of virgin adsorptive
201
-------
ability.
Prior to the second rainy season, we made several significant changes in operation. In
order to increase the adsorptive capacity of the unit, 15,000 pounds of carbon were added
to the existing 50,000 pounds in each of the cells. This increased the bed depth from 13 feet
to 17 feet. It was also decided to recycle effluent water back to the feed in order to
control the feed COD to a concentration of not greater than about 250 ppm. This change in
operation was made after noting during the first rainy season of operation that the effluent
COD level varied directly with the feed COD concentration and the specification COD
quantity to the channel would be exceeded during periods of high feed COD. It was also
felt that controlled feed COD would allow a higher ultimate COD loading of the carbon.
One other modification made was the aeration of the effluent recycled back to the feed to
prevent growth of anaerobic bacteria in the beds.
For the second rainy season of operation, we elected to operate part of the plant on a
continuous basis, regenerating each bed as it became spent, and returning the regenerated
carbon for use again. To accomplish this five of the cells were placed in continuous
staggered operation along with the regeneration furnace. This was the largest number of
cells operating in a progression mode that could be accommodated by the capacity of the
regeneration furnace. The remaining seven cells were used for once-through processing as
was done during the first rainy season. In all cases the flow to each cell was held constant
at 250 gpm which included the waste water feed from the reservoir plus the recycle dilution.
Typical performance data for one of the five cells operated in staggered mode are shown
in Figure 3. This figure shows the relationship between feed and effluent COD and the COD
loading on the carbon. During initial operation of this cell, it was elected to discontinue
service and regenerate after the effluent from the cell reached a COD concentration of
about 50 ppm. This corresponded to a carbon loading of 0.20 pounds of COD per pound of
carbon and a run length of 22 days.
The carbon was regenerated and returned to the same cell. In the second run the cell
was allowed to operate until the effluent reached a COD concentration of about 100 ppm.
This corresponded to a carbon loading of 0.40 pounds COD per pound of carbon and a run
length of 38 days.
The performance of the seven cells that were operated on a once through basis with no
regeneration is shown in Figure 4. Effluent COD concentration was allowed to reach 50
ppm before all the cells were shut down for later regeneration. As shown in the figure, the
corresponding carbon loading attained in all of these cells was 0.30 pounds COD per pound
of carbon and the run length was approximately 30 days.
The evaluation of performance of the carbon plant for this second rainy season was again
based on the combined operation of all the cells. During the season a total of 102,000,000
gallons of water were processed. The average diluted feed COD concentration was 233 ppm
while the effluent averaged 48 ppm. The carbon loading averaged 0.26 pounds COD per
pound of carbon.
202
-------
A summary of the performance of the plant during the two rainy seasons of operation is
shown in Table 2.
Adsorption Data
Table2
Commercial Feed COD, PPM Effluent COD,PPM Carbon Loading Lb. COD/
Operation Average Average Lb. Carbon Average
First Rains 377 67 (5723
Second Rains 233 48 0.26
As noted previously, problems with varying feed COD and algae growth were experienced
during some of the runs. These problems undoubtedly limited the performance to poorer
levels than could have been attained under ideal steady state conditions. However, these
are the real problems that exist when treating refinery waste waters and we therefore feel
the carbon loadings attained are indicative of the true performance of the commercial plant.
CARBON HANDLING AND REGENERATION
This system is shown in Figure 5. The spent carbon is removed as a water slurry through
a carbon removal trough and nozzle even with the top of the gravel layer. The slurry flows
by gravity and with the aid of water jets in the carbon removal trough to the spent carbon
transfer pump sump. From here a rubber lined pump transports the carbon slurry into the
spent carbon tank. The tank has capacity to contain the contents of one cell. From this
tank the carbon is educted with high pressure water to the dewatering screw above the
regeneration furnace. A timer valve on the spent carbon line feeding the eductor opens
and closes at timed intervals to allow close control of the carbon flow to the regeneration
furnace. The dewatering screw separates most of the water from the spent carbon slurry
and the drained water is returned to the reservoir. The carbon discharging into the furnace
contains approximately 50 per cent by weight of water.
The regeneration furnace is a 56" I.D. multiple hearth type with a total of six hearths.
It is gas fired and internal temperatures are controlled to about 1600 F. A center shaft with
rabble arms moves the carbon across each hearth and downward through the furnace. An
elaborate flue gas quench and scrubber system is installed to meet local Air Pollution
Control District requirements. The furnace is designed to regenerate 8,000 pounds of carbon
per day or one bed in about 6 to 7 days.
The regenerated carbon drops into a quench tank just below the furnace from where it is
educted into the regenerated carbon tank. From this tank the carbon is then educted and
transported through a hose back to the same cell where it originated.
The carbon handling system has worked very well. The original 1" cast iron eductors
were too small and soon lost capacity due to erosion. Larger stainless steel eductors and
a reduction of motive water pressure from 110 psig to 50 psig have remedied this. Also
originally we had some problem with displaced support gravel from the bottom of the cells
plugging up the eductors. This was remedied by constructing a gravel screen in the top
of the spent carbon tank to remove the gravel. The time required to transfer the carbon
out of the cell to the spent carbon tank is about 14 hours while transfer back to the cell
from the regenerated carbon tank takes about 7 hours.
203
-------
Operation of the regeneration furnace has been excellent. The major effort relative to
this operation has been to develop a reliable method for determination of the quality of the
regenerated carbon. After operational guidelines were developed, the furnace consistently
regenerated the spent carbon to near virgin activity.
Our experience indicates that about a 5 per cent loss is incurred during carbon handling
and the regeneration. We feel the majority of this is due to attrition of the carbon and
resulting production of fines. These fines are backwashed out of the cells prior to being
placed in operation and end up back in the reservoir where they settle out.
OPERATING COSTS
Actual operating costs for the second year of operation are shown in Table 3.
Cost Data
December, 1972 - March, 1973
Table 3
Cents Per Cents Per
Thousand Gallons Pound of COD
Cost Areas of Water Treated Removed
Utilities 9 4
Repair Labor 3 1
Operating Labor 15 7
Carbon 11 5
Miscellaneous 2 1
Total 4CT TF
Updating these costs to 1976 levels would give about 56 cents per thousand gallons of water
treated, or about 25 cents per pound of COD removed.
CONCLUSIONS
The plant has proved to be easy to maintain, easy and quick to start up, simple to shut
down and leave in a standby stage, and very reliable in its performance.
For the unique intermittent type operation required at the Watson Refinery, we feel the
Carbon Adsorption Process has been successful. However, we do not feel that this conclusion
would necessarily be the same for a continuous operation requirement or some other unique
situation.
DISCUSSION
Gantz: Did you experience any solids handling problems?
M_._A. Prose he: We had no solid handling problems. We had some concern about the silt,
sand, and carbon fines building up in the bottom of the reservoir. But during the summer months
we fust moved in a front-end loader and removed all solids.
204
-------
Ben Buchanan: I was wondering what control methods you used on your reactivation furnace
to control the reactivation.
M. A. Prose her What we finally did was measure the ABD, apparent bulk density, of the
carbon being fed to the furnace and the regenerated carbon from the furnace. We found
that a reduction of gm/cc in ABD was sufficient to bring the carbon back up to 100%
adsorption efficiency. In actual practice the operators had a pre-weighed cube and a
balance at the unit. The cube was then used to weigh a fixed volume of feed and product
carbon and the difference used as the guide to furnace operation. As a generality, spent
carbon had an ABD of 0.56 gm/cc and the regenerated 0.50 gm/cc.
Ben Buchanan; How did you determine how much of your carbon was lost by attrition, and
how much was lost through burning in your regeneration furnace?
/
M_. A^ Prosche: The'only insight we had was to roughly know how much carbon we put
back into the cells after regeneration, and the amount of carbon that had accumulated in
the bottom of the reservoir after a season of operation. It appeared to us that the accumulated
fines in the reservoir about equaled the loss in volume in the beds.
Ben Buchanan: There is a little problem with measuring volumes because I think the losses
that you have from burning the carbon do not necessarily reduce the size but many of those
losses are inside the particle and do not show up. You will get a weight loss but you do
not necessarily get the same volume loss.
M. A. Prosche: That could very well be. This was not a research project but strictly a
demonstration project to determine feasibility and costs.
Larry Echelberger: Has there been any attempt to reuse this water in your refinery after you
clean it up? Is it high in TDS? Do you attempt to reuse the water at all ?
M. A. Prosche: The discharge is primarily refinery waste water diluted with rain water.
Trie waste water consists of untreated service water, desalter waters, and water from tank
bottoms. So this water would require extensive pretreatment prior to any reuse. Our primary >
thrust now is to reuse stripped sour water which is essentially contaminated consensate.
Following the successful utilization of all of this water, we can then turn our attention to
the more difficult problems of reusing the waste water.
205
-------
BIOGRAPHY
Marvin A. Prosche is Manager of Refinery
Technology at Atlantic Richfield's Watson
Refinery. He holds a B.S. in ChE from the
University of Notre Dame and an M.S. in
ChE from the University of Illinois. His
experience includes 25 years of petroleum
refinery process design and process engineering.
206
-------
FIGURE I
WASTE WATER" ACTIVATED CARSON TREATMENT PLANT
TO RESERVOIR
"FROM RESERVOIR
u
1
V
BACKWASH -
TROUGH
t
CARBON
GRAVEL
LEOPOLD TILE
CARBON ADSORPTION CELL
12 CELLS
EACH 12' x 12' x 26* DEEP
Relative Efficiency Profile of Cell
After 208 Hours of Operation @ 250 GPM^and
650 to 600 PPM COD in Feed
CHLORINATOR
n
PUMpf~l
BACKWASH
EFFLUENT
SUMP
EFFLUENT BACKWASH
RETENTION SUMP
SWMP
•' » TO OOMINGUEZ
CHANNEL
Figure 2
Depth Down into Carbon Bed, Feet
207
10
11
-------
Typical Performance Data—Staggered Operation
Second Rainy Season
CL
CL
cf
g
V->
CD
"c
o
o
c
o
O
Q
O
O
400
300
200
100
802468
Millions of Gallons of Water Treated
10
Figure 3
12
14 16
c
o
.a
to
TO O
I 3
° R
c °
o O
•e a)
(3 3
0.4
0.3
02
0.1
0
802468
Millions of Gallons of Water Treated
10
12
14
16
-------
Performance Data—Bulk Processing
Second Rainy Season
Figure 4
Q_
Q_
cf
c
o
o
c
o
O
Q
O
O
o
•e
n)
a
c
o O
xj
is «>
<3 3
400
300
200
100
0
3 4 5 6 7 8 9
Millions of Gallons of Water Treated
10
4567 89
Millions of Gallons of Water Treated
10
11 12
11
12
-------
CARBON TRANSFER AND REGENERATION SYSTEM
FIGURE 5
I
CARBON
BED
CARBON REMOVAL
TROUGH
SPENT
CARBON
TANK
REGENERATED
CARBON
TANK
DEWATERING
SCREW
REGENERATION
FURNACE
SCRUBBER
SEALING
POT
EDUCTOR
-------
SESSION VI
"MISCELLANEOUS TOPICS"
Chairman
M.K. "Don"Hutton
Manager, Mechanical and Environmental Engineering
Kerr-McGee Refining Corporation
Oklahoma City, Oklahoma
Speakers
Lial F. Tischler
"Inherent Variability in Wastewater Treatment"
R.T. Milligan
"Reuse of Refinery Wastewater"
Sterling L. Burks
"Biological Monitoring of Petroleum Refinery Effluents"
Ronald G. Gantz
"API - Sour Water Stripper Studies"
211
-------
WATER POLLUTION ABATEMENT AT BP OIL CORPORATION'S
MARCUS HOOK REFINERY
B. A. McCrodden*
ABSTRACT
BP Oil Corporation's Marcus Hook Refinery has achieved first
stage ultimate oxygen demand standards prescribed by the Delaware
River Basin Commission by operating a wastewater treatment system'
consisting of rapid sand filtration, for removal of oil and sus-
pended solids, followed by granular activated carbon adsorption,
for removal of dissolved organic material. Associated equipment
includes backwash holding tanks, sludge thickeners, two-stage
centrifugation for oil-water-sludge separation, and a multiple
hearth furnace for carbon regeneration.
The 2.2 MGD Wastewater Treatment Plant has demonstrated
average removals by the rapid sand filters of 67, 67, and 20 per-
cent reduction for oil, suspended solids, and first stage ulti-
mate oxygen demand respectively. Average removals by the acti-
vated carbon adsorbers have been 74, 40, 27, 49, and 99 percent
reduction for oil, suspended solids, first stage ultimate oxygen
demand (FSUOD), chemical oxygen demand, and phenol respectively.
Capital cost of the Wastewater Treatment Plant, constructed
on a one-quarter plot, was $1,822,000, with a projected annual
operating cost of $183,000.
INTRODUCTION
A,2.2 MGD sand-filtration-activated carbon adsorption Waste-
water Treatment Plant, placed in operation at the Marcus Hook
Refinery of BP Oil Corporation, a subsiderary of the Standard Oil
Company (Ohio), in March, 1973, has demonstrated overall removals
of 80, 92, 42, 64, and 99 percent reduction for suspended solids,
oil, first stage ultimate oxygen demand (FSUOD), chemical oxygen
demand (COD),and phenol respectively.
The 105,000 BPD Class B Refinery, located in southeastern
Pennsylvania, discharges its effluent waters into the Delaware
River. In 1961, the Delaware Estuary became the subject of an
"Conservation Engineer, BP Oil Corporation
212
-------
intensive study by the then U.S. Public Health Service. The
Delaware Estuary Comprehensive Study (DECS) served as a basis
for a determination by the Delaware River Basin Commission (DRBC)
that the 1964 discharge loads to the Estuary exceeded its assim-
ilative capacity. Subsequently, stream quality objectives were
established and the Estuary divided into six zones. The total
first stage ulimate oxygen demand permitted to be discharged
without violating these objectives was determined for each zone
from base data developed by the DECS and its application of a
mathematical model. Under the doctrine of equitable apportion-
ment, each individual discharger to a zone received an alloca-
tion based upon the concept of equal percentage of raw waste
reduction in that zone. The Refinery's measured raw waste load
was 24,650 pounds per day FSUOD. Application of an 89.25 zone
percent reduction lead to an overall discharge allocation of
2650 pounds per day FSUOD. This allocation includes the net
contribution from the process wastewater and once through cooling
water streams.
An additional Delaware River Basin Commission requirement
is the secondary treatment of all process wastewater streams.
Secondary treatment for zone four of the Estuary is defined as
89.25 percent FSUOD reduction.
In 1969 a compliance schedule of 48 months was established.
The initial effort toward achieving compliance was the evalu-
ation of the existing API oil-water-solids Separator, through
which all process wastewater flow is directed. Monitoring of
API Separator influent and effluent FSUOD determined an average
68 percent removal, far below the DRBC's required 89.25 percent
for process wastewater streams. Accordingly, a project to de-
termine the treatability of the API Separator effluent, and a
project to reduce the API Separator's hydraulic loading were
undertaken. The latter project had as its basis an in-plant
water use survey which concluded that a reduction on process
wastewater flow to the API Separator could be accomplished by
installation of a brine cooler; replacement of barometric con-
densers with surface condensers; segregation of sanitary wastes
from the process wastewater stream; and further segregation of
this project are evidenced by a reduction in the hydraulic
loading from 3750 to 1450 GPM.
Treatability of the API Separator effluent was investigated
through the operation of a bench-scale sludge unit and an extend-
ed aeration pilot plant. With accumulated data as the basis,
a preliminary biological treatment system design was prepared.
The proposed full-scale design required intermediate facilities
for oil removal, two 369,000 gallon aeration basins, final
clarifiers, an anaerobic digester, and both biological and oily
sludge dewatering facilities.
213
-------
Review of an estimated biological treatment capital cost
of $2,500,000; an estimated annual operating cost of $220,500;
biological treatment-variability; land requirements; and excess
sludge generation, led to investigation of a sand filtration-
activated carbon adsorption treatment system.
Bench-scale single media filtration of API Separator efflu-
ent resulted in an outlet of acceptable quality for feed to an
activated carbon contactor. An adsorption isotherm, presented
as Figure 1, was prepared using filtered water and indicated a
theoretical loading of 0.3 gm of TOG adsorbed per gram of carbon.
This is equivalent to one pound of carbon exhausted per 1000
gallons of wastewater treated.
Based on the above results, operation of a sand filtration
pilot plant followed in line by a dynamic carbon column test was
continued for a six week period to establish design parameters.
Filtration rates.- ranging from 12-18 GPM/sq. ft resulted in
average removals of 79, 77, and 35 percent reduction for oil,
suspended solids and FSUOD respectively. Average removals by
activated carbon adsorption were 85, 62, 83, 65, and 99 percent
reduction for oil, suspended solids, FSUOD, TOG, and phenol
respectively as shown in Table 1. Projection of the carbon
column test data indicated an exhaustion rate of 0.86 pounds of
carbon per 1000 gallons of throughput.
With accumulated data as the basis, a preliminary sand
filtration-activated carbon adsorption design was prepared.
Comparison of an estimated capital cost of $2,000,000; an esti-
mated annual operating cost of $179,000; and reduced land area
requirements, with biological treatment preliminary design, led
to the decision to commence filtration-adsorption design. Figure
1A presents a schematic flow diagram of the treatment facilities
designed.
SAND FILTRATION DESIGN
Three parallel rapid sand filters were designed to remove
oil and suspended solids from the API Separator effluent. De-
sign removals were those achieved during pilot operation. An
intermediate basin was included in the design to control flow
surges and equalize influent overloads.
Each filter is 10 feet in diameter by eighteen feet six
inches overall height. Flow enters the bottom of the vessel and
rises vertically through a 10 inch pipe in the center of the
filter. A rated flow of 1000 GPM per filter corresponds to a
superficial hydraulic loading of 12.8 GPM/ft.
Flow to the filter system is controlled by a level control-
ler which maintains a constant level in the intermediate surge
214
-------
basin. The effluent flow from each filter is sensed by individ-
ual flow indicators. The flows are summed and equally divided
among the three filter.s by throttling each filter's effluent
control valve.
Under normal filtering conditions the vessel is full of
water to the vent connection on top. The water flow is down
through the filter media of 2.5 feet of anthracite and 4.5 feet
of sand; through the support gravel; and through the nozzles
which are inserted in a tube sheet. The water beneath the tube
sheet flows out through the outlet connection to a 30,000 gallon
filtered water holding tank.
Removal of suspended solids and oil trapped by the filters
is accomplished by backwashing with water stored in the filtered
water holding tank.
The initial step in backwash is to remove the water remain-
ing in the filter by applying air pressure to the top of the
vessle. This allows an up flow air and water scour to follow
and effectively remove adhering suspended solids and oil from _
the sand and anthracite particles. Scour rates are 7.1 GPM/ft.
and 7.1 SCFM/ft. As the scour water reaches the top of the
vessel, the air is shut off and the water rate increases to
25.1 GPM/ft. , thereby flushing the filter of trapped suspended
solids and oil. The backwash water overflows into the center
standpipe and is directed to a 30,000 gallon sludge blending
tank.
The backwash cycle is automatically operated by a Programmed
Timer which can be initiated by an interval timer, high differ-
ential pressure, or manually by pushbotton. The three filter
system is designed to allow only one filter to backwash at any
one time. The filters will automatically backwash in numerical
sequence. Although the filters were designed to operate with
one off-line, the mode of operation is to have an individual
filter off-line only during its backwash cycle. If the level
in the filtered water holding tank is low or the level in the
sludge blending tank is high, the backwash cycle cannot proceed
and an alarm is annunciated.
Table 11 summarizes the rapid sand filter design data.
ACTIVATED CARBON ADSORPTION DESIGN
Three parallel activated carbon adsorbers were designed
to remove soluble organic matter from the sand filter effluent
at a maximum flow rate of 2000 GPM. Design removals were those
obtained during pilot operation.
Each adsorber is a carbon steel vessel 10 feet in diameter
by 65 feet overall height and is lined with 12-15 mils of
215
-------
Plastite. The adsorbers each contain 92,000 pounds of granular
activated carbon in a bed depth of 45 feet. An additional 8000
pounds of carbon occupies the upper and lower end cone areas.
The upper and lower cone angles are 90 and 46 degrees respective-
ly, based on the angle of repose of granular activated carbon
immersed in water.
Flow to the three adsorbers is controlled by the level in
the filtered water holding tank, which acts as a feed surge
basin. The influent to each adsorber is distributed through a
circumferential manifold located just above the lower cone sec-
tion. The flow is directed downward under an internal cone,
then upward through a 3 foot diameter opening in the internal
cone. A design flow to each adsorber of 667 GPM corresponds to
an empty bed contact time of 40 minutes.
The upward flow through the2packed bed at a superficial
hydraulic loading of 8.5 GPM/ft. , is discharged through eight
internal septums which extend vertically from the upper cone.
The septums are stainless steel well screens which retain the
1.5 mm diameter activated carbon particles in the adsorber.
Filtered service water is provided at each septum for backflush-
ing, should plugging due to carbon fines occur.
Continuous adsorption is dependent upon the removal of
exhausted carbon from the adsorbers and the addition of regen-
erated carbon. One thousand pounds per day of spent carbon is
pulsed from each of the three adsorbers. During the pulse peri-
od, which occurs for each vessel every 24 hours, the adsorber
is taken out of service. The hydrostatic pressure available at
the lower cone apex is used to transport the carbon slurry to a
flooded collection tank. A pulse period of 1.4 seconds allows •
the desired 1000 pounds of carbon to be transferred under velo-
cities of 5 feet per second. Transfer lines are 4 inch schedule
40 carbon steel with schedule 80 long radius sweeps. Ball
valves are used in carbon slurry service. During this pulse
period, regenerated carbon is added to the top of the adsorber
from a carbon storage tank located above each vessel.
As the ball valve at the adsorber apex closes to stop spent
carbon flow, filtered service water is. introduced to flush the
line, thereby preventing carbon bridging. Freezing problems
are avoided by draining the transfer line following completion
of the water flush.
A cone bottom carbon collection tank receives the spent
carbon and acts as the regeneration furnace feed tank. A ball
valve at the apex of the collection tank pulses carbon for 8
seconds into a dewatering screw at two minute intervals. Fil-
tered water is added at the apex to prevent carbon bridging, and
is added to the dewatering screw to further wash the carbon of
free oil which was "filtered out" in the adsorber. Overflows
216
-------
from the collection tank and dewatering screw are directed to a
carbon settler from which the carbon is ejected into the de-
watering screw and the water overflows to be reprocessed.
Table III summarizes the activated carbon adsorption design
data.
THERMAL REGENERATION DESIGN
A five foot diameter multiple hearth furnace was designed
to thermally regenerate the spent carbon. The dewatered carbon
enters the six hearth furnace through an 8 inch inlet for re-
generation at a design rate of 125 pounds per hour. The carbon
is moved downward through the fire brick lined hearths by stain-
less steel rabble arms. In the first hearth, which is unfired
but maintains a temperature of 1100:F., any remaining moisture
is vaporized. Hearths four and six, numbered from the top, are
tangentially fired by two burners using refinery fuel gas at
rates of 188 and 68 CFH respectively, to maintain respective
temperatures at 1725°F. and 1750°F.
In an atmosphere controlled by addition of steam at a design
rate of 125 pounds per hour, the adsorbed organics are volati-
lized and oxidized. To assure complete oxidation, all flue
gases pass through an integral afterburner fired by refinery
fuel gas and maintained at a temperature of 1250 F. Recircula-
tion of shaft .cooling air provides sufficient oxygen for com-
bustion. Prior to emission to the atmosphere, the flue gases
pass through a two foot diameter, four plate, Impinjet wet scru-
bber using filtered service water for gas cooling and particu-
late removal to 0.04 grains per standard cubic foot (dry).
Temperature indicator controllers maintain the desired
temperature in the fired hearths. Furnace safety features in-
clude ultra-violet flame scanners which cause an alarm to annun-
ciate should the combustion air blower, induced draft fan, or
the shaft cooling air fan fail. Abnormally high or low fuel gas
pressure will cause the main gas safety valve to close, result-
ing in a flame-out at all burners.
Regenerated carbon is discharged from the furnace into a
12 cubic foot cone bottom quench tank flooded with filtered
service water. Temperature reduction, the addition of make-up
carbon, and the formation of a carbon slurry occur in the quench
tank. As the carbon level in the quench tank increases a rotat-
ing bindicator is stopped and a timed sequence is initiated to
transfer the regenerated carbon to one of three 96 cubic foot
carbon storage tanks located above each adsorber.
During the time controlled sequence, the carbon slurry flows
by gravity into a 5 cubic foot blow case. Filtered service
water is then introduced into the blow case to pressure the
217
-------
carbon at velocities of 5 feet per second through 2 inch trans-
fer lines of schedule 40 qarbon steel with schedule 80 long
radius sweeps. The slurry transfer is followed by a water flush
and an air drain to clear the line. In the event a high level
is indicated by a storage tank bindicator, the carbon is auto-
matically transferred to the next storage tank.
Carbon addition to the adsorbers from the storage tanks,
which occurs during the pulsing of spent carbon from the bottom
cone, is judged complete by a bindicator located in the upper
cone. Should the bindicator indicate a low level, the adsorber
may not be brought back into service.
An additional safety feature is an atmospheric vent from
the top of the adsorber to its carbon collection tank. In the
event a number of septums plug simultaneously, excess flow will
be vented, and overflow the collection tank to the carbon sett-
ler. A pressure gage is located on the vent line to indicate
such an occurance.
Table IV summarizes the regeneration furnace design data.
SOLIDS DEWATERING SYSTEM DESIGN
A solids handling system was designed to separate the sludge
removed at the Sand Filters, API Separator, and Emulsion Treater
into an oil, water, and solid phase. Upon separation, the oil
is recovered, the water is returned for reprocessing, and the
solids are disposed on a sanitary landfill.
The three intermitent sludge streams noted above are mixed
in a 30,000 gallon sludge blending tank and transferred to a
26,000 gallon circular thickener at a rate of 60 GPM. The design
loading of 30 pounds per square foot per day results in an
underflow concentration of 2.5 percent solids. The thickener
underflow of 20 gallons per minute and overflow of 40 gallons
per minute are directed to the sludge holding tank and API
Separator respectively. Should an emulsion layer accumulate on
the thickener, it is skimmed directly to the sludge holding
tank. The sludge holding tank acts as a feed surge basin for
the scro 1 centrifuge.
Feed the scroll centrifuge, flowing at 20 GPM, passes
through a double pipe heat exchanger which maintains an'outlet
temperature of 150 F. Operating at 2600 RPM, the scroll centri-
fuge discharges a stream of 50 percent solids, and an oil-water
stream. The solids are carried by conveyor belt to a holding
container to await disposal. The liquid centrate is directed to
the disc centrifuge feed sump.
The disc centrifuge feed, at 20 GPM, passes through a J
double pipe heat exchanger which maintains an outlet temperature
218
-------
of 180 F. An additional 25 GPM of filtered service water also
enters the disc machine at 180 F. to establish a nozzle seal.
Operating at 6350 RPM, the disc centrifuge discharges an oil
stream for recovery, a water stream for reprocessing, and a
solids-water stream, also for reprocessing through the API
Separator.
SAND FILTER PERFORMANCE
The three parallel rapid sand filters have demonstrated
average removals of 67, 67, and 20 percent for oil, suspended
solids, and FSUOD respectively. The average influent flow of
1795 GPM corresponds to an average superficial hydraulic loading
of 7.8 GPM/ft/ .
The original design mode of operation has been revised
to permit all filters to equally divide the flow, rather than
have two filters in-service and one off-line, thus the lower
average superficial hydraulic loading noted above. With the
current mode of operation, an individual filter is only off-line
during a backwash cycle.
Analysis of the data presented in Table V indicates average
removals of 6.6, 4.6, and 2.1 Ibs./D/ft. . for oil, suspended
solids, and FSUOD respectively. The units of pounds per day per
square foot, represents a simultaneous evaluation of concentra-
tion and hydraulic loadings.
Table VI presents an analyses of sand filter backwash water.
Based on this data the backwash high rate flush duration was
increased to 7 minutes and the flow rate was increased to
greater than 2000 GPM. The total backwash duration averages 20
minutes and is dependent upon the time required for the pressur-
ized removal of water remaining in the filter. The time requir-
ed is a function of the differential pressure across the sand
filter when a backwash is initiated. The backwash interval is
currently set at 3.5 hours, ie,. each individual filter is
backwashed every 10.5 hours.
Although the design included a differential pressure
override to initiate backwash, this option is not currently
utilized since the maximum differential pressure reached during
the above backwash interval has been 3 PSI.
To date the factor limiting the backwash interval has
been the capacity of the sludge blending tank which receives
backwash water.
An operating problem encountered has been a decrease in
effluent quality due to a backwash cycle. The two filters
remaining on line experience a "shock" as the individual flow
rates increase to include the portion of flow previously handled
219
-------
by the third filter. The decrease in effluent quality was due
to the expelling of suspended solids trapped in the filter, and
the decrease in suspended solids removal at the increased
superficial hydraulic' loading. An increase in effluent quality
was obtained by setting the controls to maintain the established
flow rate to each on-line filter during a backwash cycle.
A decrease in effluent quality was also observed during
the removal of water remaining in the filter by applying air
pressure print to backwash. Table VII presents an analyses of
this water and indicates that suspended solids are removed from
the filter and contaminate the effluent water stored in the
filtered water holding tank.
The installation of drain lines to direct this water
removed from the sand filter to the API Separator is expected
to result in an increase in filtered water quality.
ACTIVATED CARBON ADSORPTION PERFORMANCE
The three parallel activated carbon adsorbers have demon-
strated average removals of 74, 40, 27, 49, and 99 percent for
oil, suspended solids, FSUOD, COD, and phenols respectively.
The average influent flow of 1450 GPM corresponds to an empty
bed contact time of 55 minutes. The average-superficial hydrau-
lic loading experienced has been 6.2 GPM/ft. .
The 345 GPM difference between filter and adsorber flow
is due to the utilization of filtered water as unit service
water and as backwash water.
Table V prese ts adsorber influent and influent data.
Review of this data must consider the operating period duration,
especially with respect to the carbon detention period of 92
days in the adsorber bed since the data was obtained during a
period when a portion of the bed was esentially virgin carbon.
The carbon adsorbers utilize the removal merchanisms of
adsorption, filtration, and biological degradation.
During initial operation, the adsorbers were receiving an
influent containing 1 ppm of dissolved oxygen, and discharging
an influent containing zero dissolved oxygen. The oceuranee •>
of anerobic digestion in the adsorbers was indicated by an
increase in sulfide concentration from an influent of 0.6 ppm
to an effluent of 13.4 ppm. The influent concentration of or-
ganic sulfur compounds was determined to be 0.01 ppm. Further
investigation of the effluent odor problem due to the sulfide
produced in the adsorbers indicated the presence of butyl mer-
captan, tiophene, and dimenthal sulfide in the adsorber effluent.
220
-------
In order to correct the effluent odor problem, air was
injected into the sand filter effluent manifold and an increase
was o tained in the adsorber influent dissolved oxygen concen-
tration to 3.0 ppm. 'The average decrease in dissolved oxygen
though the adsorfeers has been 93 percent, leaving an effluent
dissolved oxygen residual of 0.2 ppm. Although activated carbon
has an affinity for oxygen, a 15 percent increase in ammonia
concentration through the adsorbers indicates that biological
activity is occurring. The detection of dissolved oxygen in the
effluent and a significant reduction in the odor problem have
occurred, however, an increase in sulfide concentration to an
average of 7 ppm is indicative of continuing anaerobic digestion.
In order to maintain aerobic conditions and eliminate the sul-
fide production, injection of hydrogen peroxide will be evaluated
with respect to its effect on sulfide production, dissolved
oxygen concentration, flow rate, and phenol, B0D, and COD re-
movals.
The average TOG loading on the spent carbon removed from
the adsorbers has been 0.22 pounds TOC/pound carbon. The average
COD loading has been 0.42 pounds COD/pound carbon while the
phenol loadings experienced have been 0.09 pounds phenol/pound
carbon.
The low phenol loading: lead to an increase in effluent
phenol concentrations following introduction of 100 GPM of
Foul Condensate Stripper bottons containing an average 300 ppm
phenol. The increased effluent phenol concentrations noted in
Table V are a result of the spent carbon wave front moving
toward through the carbon bed. In order to achieve design re-
movals, the carbon regeneration rate was increased to 250 pounds
per hour, thereby providing additional adsorptive capacity.
It was observed that during the period in which the spent
carbon wave front had moved upward in the adsorber, the increase
in phenol effluent concentration occurred at low adsorber in-
fluent values. This was a result of the adsorbed phenol achiev-
ing equilibrium with the phenol in solution in the wastewater.
Again, lowering the spent carbon wave front will provide addi-
tional adsorptive capacity and eliminate the occurrence of this
phenomenon.
An operating problem encountered has been the decrease in
effluent quality due to the pulsing of spent carbon from an
off-line adsorber. The two adsorbers remaining on-line experi-
ence a "shock" as the individual established flow rates increase
to include that portion of flow previously handled by the third
adsorber. The decrease in effluent quality was due to the ex-
pelling of oil and carbon fines from the adsorber. A similar
expelling occurs from the third adsorber as flow is restored.
In order to reduce the "shock" to the carbon bed, the time
required to open and close the adsorber influent valve will be
221
-------
set to maintain the established flow rate to each on-stream
adsrober during the carbon pulse cycle.
A second problem encountered has been the inability to
achieve the design maximum flow rate of 2000 GPM through the
adsorbers. Entrapped suspended solids, oil, and carbon fines
contribute to an excessive pressure drop through the bed. Fol-
lowing an individual bed pulse, an increase in the maximum
flow rate to the adsorber is observed. Currently, the pulsing
and regeneration rate has been increased to remove the entrapped
matter and virgin fines in an attempt to achieve the design flow
rate. Flushing the adsorbers has also been successful in obtain-
ing an increased flow rate. The flush is accomplished by clos-
ing the adsorber effluent valves and introducing 150 GPM into
the base of the vessel and allowing the bed to expand into the
carbon storage tank above the adsorber. The flush water con-
taining carbon fines overflows the carbon storage tank to the
carbon settler, from which the water is directed to the API
Separator and the settled carbon fines are transferred to the
regeneration furnace where they are combusted.
THERMAL REGENERATION^ PERFORMANCE
Spent carbon regeneration has been achieved using a six-
hearth furnace fired> by refinery fuel gas.
Regeneration at the 3000 pound/day rate is controlled by
measurement of the regenerated carbon apparent density and
comparing the value with virgin carbon. To date," the average
density of regenerated carbon has been 49.1 g/lOOcc, which com-
pared with a virgin carbon value of 45.1 g/lOOcc. Changing the
atmosphere in the regeneration furnace by varying the amount of
steam injected has been the method used to control the carbon
density. The current steam addition rate is 30 LB./HR. A
decrease in steam addition was found to allow additional com-
bustion of adsorbed materials and thereby decrease the carbon
density.
Adsorption isotherms prepared for both regenerated and
spent carbon were equivalent indicating regeneration to the
carbon's adsorptive capacity was achieved.
Carbon losses to date have been 10 percent per regenera-
tion cycle. This high loss value may be attributed to mechani-
cal and start-up problems rather than to excessive carbon attri-
tion, and is expected to approach the design 5 percent loss as
equipment modifications are completed. These modifications in-
clude the installation of a surge tank associated with the spent
carbon collection tank and revisions to the carbon settler to
increase its capacity.
SOLIDS DEWATERING SYSTEM PERFORMANCE
222
-------
An analytical evaluation of the solids dewatering system
has not been performed to date. Operating problems encountered
resulted in an extended shutdown of the disc centrifuge and the
inability to accept API Separator and Emulsion Treater sludges.
Revisions to the disc centrifuge systems will be completed
to provide flow control of the feed at 20 GPM, replacing the
existing level controlled mode of operation. In addition, the
sealing water will be flow controlled at 25 GPM to prevent the
flushing of bearing grease which occurred previously, causing
a bearing burnout and resulted in the extended shutdown.
Sludges from the API Separator and Emulsion Treater caused
severe plugging problems in the solids dewatering system. The
inlet screens at the sludge blending tank, three inch transfer
lines, pump impellers, and the double pipe heat exchangers ex-
perienced plugging due to the debris contained in these sludges.
The problem has been corrected by installing a comminutor prior
ro the in-line "centrifugal pump that transfers sludge from the
blending tank to the thickener.
Extended operation at design conditions will permit eval-
uation of the centrifuge system.
ECONOMIC EVALUATIONS
Tables VIII and IX present Wastewater Treatment Plant capi-
tal and annual operating costs totaling $1,822,400 and $183,000
respectively.
The estimated annual operating cost is presented due to
the current operating period during which labor, maintenance,
and carbon addition were not representative of normal operating
conditions.
Considering equipment life to be 10 years, and the cost
of capital 35%, Wastewater Treatment Plant unit costs, including
filtration adsorption and solids dewatering, are 24
-------
CREDITS
This project was supported in part by Demonstration Grant
No. 12050GXF from the Research and Monitoring Division of the
U.S. Environmental Protection Agency.
224
-------
B P OIL CORPORATION
MARCUS HOOK, PENNSYLVANIA
CARBON ADSORPTION ISOTHERM OF FILTERED WASTEWATER
TYPE OF CARBON:
FILTRASORB 300
INITIAL TOC-36MG/L
4l =300MG/GM OF
C0 CARBON
10 20
TOC REMAINING, MG/L (..C0)
225
-------
isa
ro
B P OIL CORPORATION
MARCUS HOOK, PENNSYLVANIA
SCHEMATIC FtOW DIAGRAM Of PROPOSED SAND FILTER - ACTIVATED CARBON TREATMENT FACILITIES
CONUHIMTEP.
stomMici
. 1
EJISTIM »WSI
SEPMITM
...
.
D
|_L_
L
^
£_T~
>r
£
Vr
[KULSIOK TIEtlEC eOTTOI
>PI eoiioii SIIIDCC
_«!EI
0 --I
ue I—i
~*
*==^
c
CE»T!i=l*
FEES TIM
1
Mb r«l ' fill
1^=
" M >»
£
II IKUIN TIE'Tt«
EFFtfdl nSMl FILTEt
BOY F.
Figure IA
-------
TABLE 1
Sand Filtration Pilot Plant Performance
Oil Suspended Solids FSUOD
mg/1 Percent mg/1 Percant mg/1 Percent
Inf Eff Removal Inf Eff Removal Inf Eff Removal
60 11 79 35 8 77 76 49 35
Activated Carbon Adsorption Pilot Plant Performance
Oil Suspended Solids FSUOD
mg/1 Percent mg/1 Percant mg/1 Percent
Inf Eff Removal Inf Eff Removal Inf Eff Removal
12.3 1.8 85 83 62 57 9 83
TOG Phenol
mg/1 Percent mg/1 Percent
Inf Eff Removal Inf Eff Removal
37 13 65 2.7 0.02 99
227
-------
TABLE 2
Rapid Sand Filter Design Data
Filter Media
Anthracite
Depth
Volume
Particle Diameter
2.5 FT,
195 FT:
0.25 FT*
Sand
Depth
Volume
Particle Diameter
Gravel Support
Depth
Rated Flow (Each of Three Filters)
Filter Diameter
Center Standpipe
Filter Area
Hydraulic Loading
Liquid Capacity With Media Installed
Maximum Allowable Pressure Drop
Thru Media
Design Pressure
Backwash Interval
Backwash Water Flow
Low Rate
High Rate
Backwash Air Flow
4.5 FT,
350 FT*
1 mm
1.25 FT
1000 GPM
10 FT
10 IN,
78 FT^
12.8 GPM/FT'
- 4800 6AL
6.5 PSI
47.5 PSI
12 HOURS
As Percent of Filtrate 1.3%
550 GPM 7.1% GPM/FT/
1960 GPM 25.1 GPM/FT ,
550 SCFM 7.1 SCFM/FT'
228
-------
TABLE III
Activated Carbon Adsorption Design Data
Rated Flow (Each of Three Adsorbers) 667 GPM
Adsorber Diameter 10 FT
Adsorber Bed Depth 45 FT
Contact Time (Empty Bed) 40 MIN 7
Hydraulic Loading 8.5 GPM/FT
Design Inlet Pressure 60 PSI
Pressure Drop Thru Carbon 35 LB
Carbon Inventory
Carbon Bed 92,000 LB
Adsorber Total 100,000 LB
Theoretical Carbon Capacity 0.3 LB TOC/LB Carbon
Carbon Dosage 0.86 LB Carbon/1000 GAL Throughput
Activated Carbon Properties
Filtrasorb 300
Total Surface Area 950-1050 M2/g
(N7 BET Method) -
Bulk DSnsity 26 LB/FT
Particle Density Wetted in Water 1.3-1.4 g/cc
Mean Particle Diameter 1.5-1.7 mm
Iodine Number, minimum 950
Ash Max 8%
Moisture Max 2%
229
-------
TABLE IV
Thermal Regeneration Design Data
Furnace
Regeneration Rate
Steam Addition Rate
Fuel
Fuel Rate
Hearth 4
Hearth 6
Afterburner
Combustion Air Rate
Hearth 4
Hearth 6
Afterburner
Design Temperatures
Hearth 4
Hearth 6
Afterburner
60"x6 Hearth with
Integral Afterburner
125 LB/HR
125 LB/HR
Refinery Fuel Gas
188 CFH
68 CFH
310 CFH
5000 CFH
1800 CFH
8120 CFH
1725-°F.
1750°F.
1250°F.
230
-------
TABLE V
r-o
CO
Wastewater Analyses
Parameter
API Separator
Influent
Sand Filter
Influent
Carbon Adsorber
Influent
Carbon Adsorber
Effluent
Average
Flow, GPM
Suspended Solids, ppm
Oil, ppm
COD, ppm
Phenol, ppm
Phenol, ppm
BOD5, ppm
FSUOD, ppm
A
B*
1795
165
--
590
900
Range Average Range
1480-1895
18-118
8-144
131-522
0.48-20.9
9.1-20.9
59-144
75-174
1795
75
107
341
5.6
13.5
95
115
735-
5-
3-
110-
0.77-
9.6-
26-
35-
1585
74
66
461
21.8
21.6
109
155
Average
1450
25
35
239
5.7
14.0
68
92
=N° Re duct ion
67
67
30
28
20
a
o
•H
u
13
-------
TABLE VI
Analyses of Sand Filter Backwash Water
Time After
Initiating
High Rte
(1960 GPM)
Backwash Suspended Solids
Mins. ppiri
0 87
1 19,370
2 9,010
3 4,160
4 3,010
5 850
6 300
232
-------
TABLE VII
Analyses of Water Removed from Sand Filters
by Pressure Prior to Backwash
Time,After
Air Pressurization Suspended Solids
Minutes ppm
0 13
1 42
2 296
3 196
4 156
5 77
6 91
7 2,010
8 5,140
9 4,900
233
-------
TABLE VIII
Wastewater Treatment Plant
Capital Cost
Item Cost
Engineering $ 125,000
Surge Basin 30,400
Sand Filters 193,300
Carbon System 358,100
Carbon Charge 120,000
Solids Dewatering 152,300
Tanks 93,200
Pumps . 13,100
Building 143,800
Piping 178,400
Electrical 199,400
Instrumentation 133,700
Structural 22,100
Foundations 45,900
Concrete 12,800
Total: $1,822,400
234
-------
TABLE IX
Estimated Wastewater Treatment Plant
Annual Operating Costs
(Filtration, Adsorption, Regeneration,
and Sludge Dewatering)
Labor $ 44,000
Maintenance 74,000
Power 10,000
Carbon Regeneration 55,000
($0.05/LB., Carbon at
3000 LBS/Day)
Total $183,000
Includes Steam, Electricity, Labor Fuel and
5% Carbon Make-up.
235
-------
SECTION 6
Full-Scale Powdered Activated Carbon Treatment
x
CASE HISTORY: USE OF POWDERED ACTIVATED CARBON
IN AN ACTIVATED SLUDGE SYSTEM
Joyce A. Rizzo
Sun Oil Company, Marcus Hook, Pennsylvania
INTRODUCTION
In an effort to expand the performance of existing biological treatment facilities, full
scale trials utilizing powdered activated carbon were conducted at Sun Oil Company's
Corpus Christi, Texas Refinery. The main objective of the trials was to reduce the effluent
suspended solids loading for compliance with the 1977 NPDES and State permit limitations.
Three separate trials were initiated over a one-year period comprising five months of
operation with powdered activated carbon addition to the aeration system of the 2.16
MGD biological treatment plant. Although the powdered carbon addition could not
guarantee compliance with the 1977 suspended solids' criteria, improvement in the existing
system's performance was significant for COD and BOD removal as well as suspended solids.
The treatment costs ranged from 1.7 to 4.3$/10 gallons depending on the influent flow
and quality. Data have been compiled showing reductions in'the refinery's final effluent
loadings of up to 56% for suspended solids, 36% for COD and 76% for BOD.
FACILITIES
A simplified flow diagram of the treatment scheme at the Corpus Christi Refinery is
shown in Figure 1. The system's operating conditions are outlined on Table 1 . The daily
average charge to the plant is only half the design capacity; although the peak design
rate is often operated during periods of rain and high ballast loading. The refinery waste-
water flow (averaging about 600 GPM) is equalized with ballast and contaminated storm
water in a 10 MM gallon pond. The combined flow from the pond (averaging about 750
GPM) is pumped through a dissolved air flotation unit for removal of any excess oil and
suspended solids. The effluent from the DAF is split between two 760,000 gallon
rectangular aeration basins. Air is introduced into each of the basins by two 75 hp
floating mechanical aerators. When operating both basins in parallel at average flow
conditions, the retention time is about 22 hours; at peak flow, this is reduced to 11 hours.
The mixed liquor volatile suspended solids is maintained at about 2200 ppm. With both
systems operating, the food to mass ratio averages only about 0.13 Ib COD/lb MLVSS or
0.07 Ib BOD/lb MLVSS in each basin; this, of course, can be doubled by operating only
one aeration system.
Each aeration system is coupled with a 55" diameter x 12' deep clarifier. The over-
flow from the clarifiers combines for discharge. The settled sludge is collected from each
clarifier, combined and returned to the aeration basins. The return rate usually averages
about 50% of the charge rate.
236
-------
OBJECTIVE
Basically, the performance of the biological treatment system is satisfactory and the
effluent readily complies with the refinery's NPDES and State permits. With the exception
of suspended solids, the limits established for compliance in 1977 are also currently being
met. The consistently effective removal of suspended solids in the final clarifier seems to
be the only area of concern. Even with substantial organic removal in the biological
system, there appear to be two major factors causing the extra suspended solids carryover
from the final clarifiers:
Variability of influent organic and hydraulic loadings. The biological population
is thus unstable with erratic growth and variable efficiency.
Excessive aerator foaming in the aeration basins. This foam, laden with biological
solids, tends to carryover into the clarifier and out to the effluent. The problem
becomes most severe at high flow periods.
Interim limitations had been set by the State and NPDES permits allowing time to
investigate in-plant improvements which could possibly permit compliance with the 1977
TSS criteria without construction of additional facilities. Full scale trials appeared
necessary due to the nature of the problems noted — bench scale studies tended to show
excellent settling characteristics which do not reflect the hydraulics of the operating system.
Initial testing was performed with the addition of polymers to the aeration basin effluent.
The polymers did improve the solids settling somewhat but their benefit was overshadowed by
the problems still being caused by aeration foaming and shock loads. Conversely, defoamers
were effective in controlling the foam, nothing more.
In the refinery, every precaution is taken to reduce the possibility of shock loads and
spills. However, no matter how efficient, there is no way to completely eliminate the
variability: turnarounds, emergency shutdowns and the like. This, coupled with the influx
of highly variable loads of ballast, dictated that the treatment plant operation would have
to compensate for unstable influent conditions to the plant.
APPLICATION OF POWDERED ACTIVATED CARBON
The addition of powdered activated carbon was initiated in an effort to resolve all the
operating problems causing the suspended solids carryover ~ the objective being compliance
with the 1977 suspended solids' limitations. Although this objective was not achieved, some
interesting data resulted. A summary of the three separate carbon trials conducted is shown
on Table 2. Unfortunately, it was not possible to operate a "blank" system simultaneously
with the carbon system. Therefore, the data obtained 30 days prior to the start of each
trial has been utilized as the base period for that trial. As can be noted on Tables 3, 4 and
5, for the most part, the influent loadings of each base period and respective trial are
comparable; this is especially true of Trial 3. During each trial, operation was maintined
as closely as possible to the conditions observed during the base period. No special problems
were encountered that would lead one to believe that any of the trials and base periods'
operation were different, especially on the average. Although, the comparison is not ideal,
it is relevant, especially in respect to the "realworld" effect on the plant's effluent. The
237
-------
percent reductions shown between the base period and each trial are meant as a guide to
point up the improvement in effluent quality; the actual values, of course, cannot be
taken as absolute. From an operation's standpoint, there appears to be no question that
the addition of the powdered activated carbon to the aeration system realizes a significant
improvement in the effluent quality and reduces its variability.
The first trial was considerably longer than the last two — mainly, because the
carbon dose was gradually built up in the system over a four-week period; whereas in the
other trials, the carbon was batch loaded in a couple of days up to the operating level.
The gradual addition of the carbon allowed the daily observation of its effect in order to
determine the appropriate loading to be maintained. In order to minimize the amount of
carbon needed to load and maintain the system, only one aeration basin was operated
during the trials; the same was true of the respective base period. The carbon~was
manually batch added daily to the dissolved air flotation effluent as it entered the
aeration basin. During the first trial it was found that the optimum operating carbon level
needed in the aeration system was 450 ppm with approximately 1000 ppm resulting in the
recycle sludge. All three trials were conducted at these levels. This requires about
6100 pounds of carbon to charge one aeration system up to operating levels. An easy
indication of the carbon level in the system was found to be the foam level in the basin —
the appropriate amount of carbon would eliminate the foam completely; as the carbon
level dropped, the foam would build up. Of course, laboratory analysis utilizing a
standardized comparison procedure confirmed the operating level of carbon on a concen-
tration basis.
While gradually adding the carbon daily, it was noted that a distinct improvement
in the plant performance did not come about until the 450 ppm level was attained —
however, by letting the level drop off gradually, the peak performance could be main-
tained down to 300 ppm. It was also noted that more frequent additions of smaller
quantities appeared more effective than larger batch additions less frequently.
Referring to Table 2, you will note that the carbon addition rate was much higher
during the first two trials; this came about for a couple of reasons:
More frequent batch additions of larger quantities ( 1000 pounds at a time) were
necessary to compensate for shock loadings. Although the presence of the carbon
in the system greatly reduced the chances of a shock that could inhibit the
biological activity, it could not guarantee that a reduction in activity would not
occur. However, in general, it was found that if caught in time with the addition
of a large batch of carbon, a lapse in system efficiency could be effectively
brought back up to normal almost immediately without adverse effects on the
effluent quality.
At the time of the third and most recent trial, the biological system was more
stable and it took less carbon to obtain the same results. Better methods of
equalization had been put into operation and ballast was being successfully
treated at a rateable basis.
The flow rate during the third trial averaged 800 GPM versus 625 - 660 GPM for the
first two trials (this was due primarily to increased ballast water treatment). The higher
238
-------
flow with the reduced carbon addition rate accounts for the lower 1.7<:10 gallon treatment
cost compared to Trial 1's 4.3$. It is important to note that these costs only reflect the
carbon usage, the manpower utilized in adding the carbon is not included. More efficient
means of carbon addition could, of course, be employed with minimal capital investment
but was not deemed necessary for a trial basis.
On Table 6, the treatment costs obtained are summarized along with the reduction in
effluent loading observed during each trial.
From this data, one is erroneously led to believe that the efficiency seemed to improve
because of a lower carbon dose (and subsequently lower cost). As explained, the higher
carbon dosages observed during Trials 1 and 2 are directly attributable to a higher frequency
of shock loadings to the system. The only valid conclusions to be drawn from the costs
veijsus efficiency data presented is that the carbon treatment costs range from 1.7$ to 4.3$/
10 gallons depending on the variability of the influent and the frequency of batch additions
to solve overload problems. Again, it is important to remember that the percent reductions
are determined from the trial versus its pre-trial base case and should only be thought of as
a guide for comparison purposes.
Total Suspended Solids Reduction. The average effluent suspended solids data obtained
are outlined on Table 3. The reduction in the effluent suspended solids' loading was
dramatic, between 49.3% and 55.7%. The existing permit limit is easily met; however,
the effluent, at the very low level of 405 Ibs/day daily average for Trial 3, still would not
comply with the 327 Ibs/day limit of the 1977 permit. In addition, although the daily
maximum values while using carbon were significantly lower and under better control,
several peaks still would occur. Referring to Table 7, in the case of Trial 3, four data
points out of 20 were above the 561 Ib/day 1977 allowable limit, for 80% compliance;
100% compliance is easily obtainable with the existing permit. Of course, the odds of
being able to comply with the 1977 limits are better, but'still not acceptable.
Referring to the frequency distributions for Trial 3 plotted on Figures 2 and 3, the
reduction in effluent suspended solids with carbon addition becomes very obvious. The
chance of complying with the 561 Ibs/day 1977 maximum was increased from 40% during
the base period to 82% with carbon; of course, these percentages are only relative not
absolute. Although the permit is written on a mass basis, the effluent concentration of
suspended solids is a basis more easily compared between various facilities and is there-
fore, shown in Figure 3. On each figure, the flat slope of the carbon trial versus the
base period also indicates the more consistent effluent obtainable with carbon treatment,
the variability having been significantly reduced.
Organic Removal Efficiency - as mentioned, the addition of powdered carbon also
improved the organic removal efficiency of the system. A summary of the influent and
effluent COD loadings of the system are shown on Table 4. This table also points up the
uniformity of influents between the base period and trial period for Trials 1 and 3. As
mentioned, the operation was maintained as consistent as possible for each case. During
Trial 2 the influent COD loading was about 25% higher than during the base period ~
yet, the effluent loading was 30% lower showing a significant reduction on a relative
base. The removal efficiency during Trials 1 and 3 was increased significantly up to 70%
239
-------
from 62.9% and 54.8% respectively for each trial. The ultimate reduction in COD
loading to the effluent being 35.7% for Trial 3 as compared to base period 3.
The daily COD removal efficiencies for Trial 3 are plotted in Figure 4. It can be
seen that not only was the removal efficiency significantly increased, the variability
of the efficiency was reduced. Again, resulting in an effluent of consistently better
quality.
Figure 5 is a frequency distribution plot of the effect of the carbon addition on the
effluent COD concentration for Trial 3. Not only does the slope again show more
consistent results but, for example, the 95% point has been reduced from 280'ppm down
to 150 ppm, a 46% reduction assuming a comparable base.
The effect of powdered carbon on BOD is summarized in Table 5. It is obvious from
the influent loadings that the data obtained during Trial 3 are most relevant since the
base period and trial had the most comparable influent conditions. Again, as in the case
of COD, during Trial 2, the influent loading was 38% higher than the base period; yet,
the effluent was 51% lower. The reduction in BOD was most sharply noted during Trial
3 when the effluent BOD never exceeded 6 ppm — the reduction from the base period
was 76%.
The frequency distribution comparing the base period and Trial 3 is shown in Figure
6. Again the results are significant. The curve for the effluent BOD during the carbon
addition period is virtually flat, showing definite uniformity and consistency of effluent.
Figure 7 denotes the distribution of the percent BOD removals obtained — again, the
curve is relatively flat, the variability being virtually eliminated.
Oil Removal. No data were compiled on oil and grease removal since the system is
not required to handle any large amounts of oil. The effluent oil and grease usually
averages less than 5 ppm with an occasional peak around 10 ppm; the influent is normally
20-25 ppm. Any improvement would be difficult to observe at these low levels. On
occasion, however, the influent oil level can be higher, as did happen during the post
period of Trial 2 when there was an upset in the Refinery. The carbon still in the system
readily adsorbed the excess oil, keeping it in the sludge. Of course, the sludge had to
be removed from the system, thus spending the carbon. However, a possible effluent oil
and grease violation was avoided at only the expense of a carbon charge.
SPECIAL NOTE
Sludge wasting had always been erratic and unpredictable in this system due to the
variable influent and unstable bug growth. When as stable as possible, about 15,000
gallons per day of recycle sludge would be wasted from one operating aeration system.
During the carbon trials the wasting was reduced to 10,000 gallons per day under similar
circumstances. The basis for this reduction being that the sludge was considerably
thicker so less volume needed to be handled to dispose of the same mass.
240
-------
SUMMARY
As evidenced by the data presented, the addition of powdered activated carbon to the
aeration system of the refinery's existing wastewater treatment facilities did not reduce the
effluent suspended solids enough to comply with the 1977 permit limitations. However,
significant improvements in the system's performance were observed as follows:
Improved organic and suspended solids removals and reduced effluent loadings
% REMOVAL
Trial 3 Without* Carbon With Carbon % Reduction in Effluent Loading
COD 54.8 70.5 3577
BOd 93.5 98.5 75.8
TSS - - 49.3
*Pretrial base period
More uniform effluent quality
Clearer effluent
Elimination of foam in the aeration system
More consistent sludge wasting at 2/3 the volume
Reduced chances of biological upsets — the use of powdered carbon does not
eliminate biological upsets but it does appear to reduce the opportunity of their
occurance; and when they do occur, it appears to maintain the effluent quality
under better control. When caught in time, it was found that a reduction in
biological activity can effectively be brought back up to normal by massive batch
addition of carbon to the existing operating level.
These improvements were obtained by building up the aeration system to a 450 ppm
operating level of carbon with about 1000 ppm in the recycle sludge. The system was
maintained at this level by the, batch addition of about 100 Ibs/day of carbon (or 10 ppm)
at an average cost of 1.7<£/10 gallons of water treated. The amount of carbon necessary
to operate the system for peak performance could be readily determined by observing:
' The aeration basin foam level - the proper operating level of carbon eliminates
foam in the system.
' Clarity of the effluent - the presence of carbon in the system removes the tint
usually characteristic of biological effluent.
It is important to note that the powdered activated carbon utilized in these trials has
a high bulk density of 44 Ibs/cu. ft. This becomes an important factor when improved
settleability is the primary objective of the carbon addition to the system. In addition,
the higher bulk density reduces the opportunity for carbon loss to the effluent.
DISCUSSION
James F. Dehnert: We have done considerable work with powdered carbon on a pilot basis
and we~~have done some of the work that you did. A couple of problems we ran into and I
wondered if you had solved them, that is the analysis of the mixed liquor solids. Did you
determine actual carbon percentage or was this a calculated value?
Joyce Rizzo: We did it by a standard comparison. What we did was make up standard
241
-------
filters with carbon which we compared. We took samples of our aeration system before
we added carbon to it. We added known dosages of carbon and set up a set of standards.
So all the numbers we have obtained are 50 ppm, but we could keep track of it pretty
easily.
James F. Dehnert: We had not devised a system to determine how much carbon was
actually in our system at any one time. The second question: you said you didn't test
very many parameters. We have a problem in California with meeting fish toxicity, did
you ever run any tests on that?
Joyce Rizzo: Everyone knows that refinery effluent is not toxic. Maybe California
is different - they may have a higher breed of fish down there, I don't know. Obviously,
we didn't evaluate toxicity.
Les Lash: You mentioned holding about 450 milligrams per liter of activated carbon,
Joyce, in your aeration basin, I guess. What is the retention time in that aeration basin?
Joyce Rizzo: About 12 hours at average operating conditions.
Les Lash: When we ran the pilot plant on municipal waste for the EPA on powdered
activated carbon in Salt Lake, we used reactive clarifiers and if you didn't allow enough
settling, of course the effluent went over black. Now that is a little different than the
clear you were talking about.
Joyce Rizzo: The only time we ever had a carryover problem with carbon was during a
time when the bug population was upset and they would tend to take the carbon with
them. If they were going out, they weren't going to leave the carbon behind. On a
normal basis we had no problem with carbon carryover at all. It stayed right with the
sludge. In fact, it improved the settleability tremendously. If you look at the clarifier,
it has the appearance of being black; of course, the aeration system would turn black
versus the dark rich brown color and it appeared like it was all carbon. The carbon
color would take over completely, and if you looked at the clarifier surface, it appeared
black but the effluent water was virtually clear.
Milton Beychok; I am just a bit curious, I don't know if there is anyone here from
duPont or not, but something like four years ago, duPont first used this and called it
PACT, powdered activated carbon treatment, and they were licensing it. Does the use
of this involve any patent problems or royalty on the part of duPont?
Joyce Rizzo: No, this is quite a different system than what duPont is utilizing in their
PACT project, they are also utilizing a regeneration system which I be>":ve is also a
part of their patent on that. We used ICI carbon, and the system that v/e used was out-
lined to us from ICI as to how to operate. I don't think we did anything really different.
I have talked to many people who are also using powdered carbon in their aeration
system in a similar manner.
Ed Sebesta; I note you have a centrifuge for your waste activated sludge there. Did
242
-------
you continue to use a centrifuge both before and after carbon addition and what was the
effect?
Joyce Rizzo; We don't use a centrifuge all the time by the way, only when we have to.
We have plenty of land out there for land farming and a lot of times put the sludge out wet
from the aerobic digester. We did not see any real difference in our sludge handling
characteristics on the solid waste handling system - none at all. Actually, the amount of
carbon that we used - our aerobic digester is three-quarter million gallons like our aeration
system - and the amount of carbon that we used and the amount that was wasted was
virtually lost in the system.
Ed Sebesta: I have another question and you indicated that 'the effluent was very clear. I
think the data indicated about an average of 40 parts per million suspended solids. This
is pretty high for a very clear effluent, do you have any explanation for the concentration
numbers?
Joyce Rizzo: Very fine solids. Color really is what I am talking about. 40 ppm does not
represent optimal operation, 10-20 ppm would be more representative of the best effluent
obtainable with the carbon. The solids that would settle out by the way would be ones
that would settle out after hours. You could take a sample of the effluent and you wouldn't
think there were suspended solids in it, but if you let it sit for a while, the solids that
would settle would be very, very fine. I would say about half the solids that went out in
the effluent were carbon solids. Very, veryjine solids.
Pat DeJohn: I'd like to make a couple of comments. One that Mr. Lash made, the carbon
that you were using has a density of about 45 pounds per cubic foot, and consequently it
settles very rapidly. I believe what they were using out at Salt Lake was a carbon that
had a density of about 15 pounds per cubic foot. The higher density is the reason the
carbon doesn't go out over the weir. The other thing about the patent situation, duPont
just got a patent issued about two months ago on the PACT process and right now there are
some negotiations going on with respect to that.
Jan is Butler: You mentioned sometime just spreading the wet waste activated sludge. Do
you experience any odor problems?
Joyce Rizzo: No - none at all. If we put it out wet, the solids settle rapidly, and there
is no odor problem at all. We are using (I have been gone now from Corpus Christi for
about six months), but at the time I left we were using the centrifuge all the time for the
waste.
Anonymous; Do you add carbon continuously?
Joyce Rizzo: Corpus Christi has just purchased another load of carbon for usage. As I
understand it, their main goal is to use it mostly for batch addition for shock loadings and
things like that.
243
-------
Fred Goudy: Does the use of powdered carbon have any long range implications with
regard to the operation of the activated sludge plant at Corpus Christ! ?
Joyce Rizzo; They are using it.
Fred Goudy: Continuously?
Joyce Rizzo: I am not sure, I can't answer that for fact right now. I know they have
just purchased some more carbon,\and I know they have it in stock and I know that their
intentions are to use it. Now whether they will use it all the time, I don't know at this
time. 1 might mention what I think about carbon application in the aeration system.
Like I said, although we really were oriented to suspended solids removal, it is obvious
that some of the other things that I have mentioned about the system are very true. That
is, if you have a plant that is sitting there hydraulically loaded and you need more
capacity, it seems to help and gives you that extra capacity that you need. As I
mentioned none of the reductions that I presented, of course, should be taken as absolute
numbers because it was pretrial versus trial period. But there is no question that the
improvement in the plant performance is there and I think the improvement is really
dependent on the variability of your influent.
Dave Story: You mentioned adding polymers; could you elaborate a little bit more on
that?
Joyce Rizzo: Yes, we didn't do too much work with polymers because it was the very
first thing we tried to reduce suspended solids. We were adding about 5 or 6 parts per
million of some polyelectrolytes to our aeration basin overflow and we saw a reduction
in effluent suspended solids and a little bit better settling characteristics, but nothing
very significant. I think this is mostly due to our foaming problem because we still had
a lot of foam carrying over to the clarifier which is a significant problem.
Anonymous: How do you account for the reduction in sludge wasting?
Joyce Rizzo: We didn't have to maintain as many bugs in the system and they didn't have
to work as hard. But the real reduction was in volume not mass. The recycle sludge
was thicker and more uniform with carbon addition.
BIOGRAPHY
i
Joyce A. Rizzo is a staff engineer in the
Advanced Management and Methods Department of
Suntech, Inc., a subsidiary of Sun Oil Company.
She is based in Marcus Hook, Pennsylvania. Ms.
Rizzo joined Sun Oil in 1971 as a process engineer at
their Marcus Hook Refinery. Prior to her move into
Suntech six months ago, she spent three years in
Process Engineering at Sun's Corpus Christi Refinery.
She holds a B.S. in Chemical Engineering from
Northeastern University in Boston.
244
-------
TABLE I
"OPERATING CONDITIONS OF WASTEWATER TREATMENT PLANT"
FLOW
Daily average 1.08 MOD
Daily maximum 2.16 MOD
Design 2.16MGD
BIOLOGICAL SYSTEM
Recycle
MLSS
MLVSS
Retention
(Per Basin)
Sludge Age
SETTLING
2 - 760,000 Gal. Rectangular Aeration Basins
2-75 HP. Floating Aerators/Basin
50% of Charge Rate
3000 PPM
2200 PPM
11 Hours at Maximum Design Rate
22 Hours af Average Operating Rate
12 Days
2 -55Ft. Diameter Circular Clarifiers 12 Ft. Deep
,2
Average Rise Rate, GPD/FT
Peak Rise Rate, GPD/Ft
227
455
SOLIDS HANDLING
1 - 760,000 Gallon Aerobic Digester
1 - Solid Bowl Centrifuge
4 Acres of Land Farm
Wasting Rate 30,000 GPD
VSS 7,500 PPM
TABLE 2
"SUMMARY OF POWDERED ACTIVATED CARBON ADDIT|ON TRIALS"
Aeration System: 450 PPM PAC
Recycle System: 1000 PPM PAC
Trial 1
Bose Case 1
Trial 2
Base Case 2
Trial 3
Base Case 3
*At 22
-------
TABLE 3
"EFFECT OF POWDERED CARBON ON DAILY AVERAGE EFFLUENT SUSPENDED SOLIDS"
1977 Permit: 327 Lbs/Day Daily Average
Base Case 1
Trial 1
Base Case 2
Trial 2
Base Case 3
Trial 3
Flow
GPM
626
629
645
657
839
799
Permit
PPM
43
42
41
32
34
Effluent
PPM Lbs/Day
TT5~ 861
50 381
163
72
79
42
1262
565
799
405
Reduction
55.7
55.2
49.3
TABLE 4
"EFFECT OF POWDERED CARBON ON COD"
Base Case 1
Trial 1
Base Case 2
Trial 2
Base Case 3
Trial 3
Flow
GPM
626
629
645
657
839
799
Influent
PPM
459
457
343
444
367
379
Lbs/D
3445
3446
2658
3500
3698
3632
Effluent % %
PPM
170
135
266
183
166
112
Lbs/D
1277
1020
2059
1445
1670
1073
Removal
62.9
70.4
22.5
58.7
54.8
70.5
Reduction
20.1
29.8
35.7
Base
Trial
Base
Trial
Base
Trial
1
1
2
2
3
3
TABLE 5
"EFFECT OF POWDERED ACTIVATED CARBON ON BOD"
Flow
GPM
626
629
645
657
839
799
Influent
PPM
152
213
152
227
188
207
Lbs/D
1144
1607
1173
1898
1895
1981
PPM
15
15
30
13.
12
3
Effluent
Lbs/D
116
114
232
5 113
124
30
Removal
%
89.9
92.9
80.2
94.0
93.5
98.5
Reduction
%
1.7
51.3
75.8
246
-------
TABLE 6
"COMPARISON OF CARBON TREATMENT COSTS
AND SYSTEM PERFORMANCE"
% Reduction*
Trial
Trial
Trial
1
2
3
Dose
PPM
24
19
9
-------
FIGURE I
TREATMENT PLANT FLOW DIAGRAM
BALLAST
a
TANK FARM
REFINERY
API
SEPARATOR
EFFLUENT
EQUALIZATION
a
IMPOUNDMENT
I POWDERED
j CARBON
j ADDITION
I
I
H AERATION
FINAL
EFFLUENT
I
AEROBIC
DIGESTER
-H CENTRIFUGE
SLUDGE
TO
LAND
FARM
-------
EFFLUENT
TSS
LBS/DAY
2800
2400
2000
1600
1200
800
400
0
240
200
160
EFFLUENT
TSS 120
PPM
80
40
0
I
FIGURE 2
EFFECT OF POWDERED CARBON
ON EFFLUENT TSS LOADING
CARBON
i i i i i i
I i
> 10 30 50 70 90 95
% OF VALUES LESS THAN
FIGURE 3
EFFECT OF POWDERED CARBON
ON EFFLUENT TSS
99
BASE CASE
WITH CARBON
J I
I
10 30 50 70 90 95
% OF VALUES LESS THAN
99
TRIAL 3
TRIAL 3
249
-------
COD -
REMOVAL
EFFLUENT
COD
PPM
FIGURE 4
COMPARISON OF COD REMOVAL EFFICIENCY
100 TRIAL 3
90
80
70
60
50
40
30
20
10
0
300
280
240
200
160
120
80
40
0
I
AVG.»54.8 %
\
AVG.= 70.5%
CONTROL
PERIOD
JULY, 1975
TRIAL'
AUGUST
1975
FIGURE 5
EFFECT OF POWDERED. CARBON
ON EFFLUENT COD
BASE CASE
•"""^ ^ •*.«•*"'* WITH CARBON
TRUL 3
I i
10 30 50 70 90 95
% OF VALUES LESS THAN
250
99
-------
EFFLUENT
BOD
PPM
BOO
REMOVAL
FIGURE 6
EFFECT OF POWDERED CARBON
ON EFFLUENT BOD
60
50
40
30
20
10
TRIAL 3
^- WITH
CARBON
J 1—''ill
J L
5 10
30 50 70
90 95
99
% OF VALUES LESS THAN
FIGURE 7
EFFECT OF POWDERED CARBON
ON BOD REMOVAL
IOO
98
94
90
86
82
78
74
TRIAL 3
-WITH ^^
- CARBON
90 95
% OF VALUES LESS THAN
251
99
-------
CASE HISTORY
THE USE OF POWDERED ACTIVATED CARBON
WITH A BIODISC-FILTRATION PROCESS
FOR TREATMENT OF REFINERY WASTES
J.F. Dehnert
Environmental Director, Avon Refinery, Lion Oil Co.
ABSTRACT
A description of the development of a supplemental petroleum waste
water treating plant utilizing a Rotating Biological Surface Unit and
Powdered Activated Carbon followed by clarification and filtration
from laboratory and pilot plant studies through construction, start up
and operation to meet July 1977 NPDES discharge requirements.
Starting in 1972 Pilot Plant studies were conducted to compare the
performance of activated sludge, trickling filter, RBS and activated
carbon absorption processes in treating the Avon Refinery Waste Water.
The primary objection was to meet the EPA guideline discharge limits
plus the California State limits on fish toxicity. After several months
of study the treatment scheme of a RBS Unit plus solids removal facili-
ties was selected to meet the Federal standards and powdered activated
carbon was selected to meet the toxicity limits.
INTRODUCTION
With the adoption of the 1972 Amendments to the Clean Water Act,
the Staff at the Avon Refinery near San Francisco, then operated by
Phillips Petroleum Co. embarked on an investigative program to determine
the waste water treatment necessary to meet the limitations which would
eventually be placed on the refinery discharge through the NPDES Program.
At that time and until January 1975, the refinery discharge was already
subject to limitations imposed by the California Regional Water Quality
Control Board on 5-day BOD, oil and grease, settleable solids, suspended
solids, coliform and fish toxicity. In addition, limitations were in
effect on receiving water quality with respect to pH, dissolved oxygen,
undissociated NH^H, chromium, lead, H2S, Fish Toxicity, floating oil,
discoloration or turbidity and odor. The existing waste water treatment
included sour water stripping, API gravity separation, dissolved air
flotation and pH equalizing surge ponds followed by a 108-acre bio-
oxidation pond. The company had also segregated the refinery sewers so
that as much as possible of uncontaminated storm run off could bypass
the process water treatment. This storm run off was combined with the
bio-oxidation pond effluent and discharged in an underwater diffuser in
252
-------
the main channel of an arm of San Francisco Bay which receives the
Central California Valley drainage. This treating process is illustrated
in Figure 1.
INVESTIGATION
At the s£art of the investigation, no specific Federal limitations
had been determined and, therefore, the studies were mainly concerned
with a comparison of generally accepted methods of waste water treatment
to determine which of these would be most effective in removing or
reducing the known pollutants in the Refinery effluent. A primary
objective was to determine a treatment scheme that would result in a
waste water discharge that would meet the more restrictive California
State fish toxicity limitations that were being proposed.
In early 1972, working with an engineering contractor, three pilot
plants were installed at the Refinery with a slip stream of the waste
water going to the bio-oxidation pond serving as the raw feed. These
units were an air agitated activated sludge unit, a 21-foot trickle
filter and a four-column granular activated carbon unit preceded by a
mixed media filter. These plants were operated from June 1972 through
1973 in parallel or in series under wide variations of operating
conditions such as hydraulic loading, recycle rates and suspended solids
concentrations.
During this period of operation, different sets of tentative EPA
guidelines were issued for the Petroleum Industry. In each case, the
effluent from the biological treating pilot plants failed to meet these
guidelines and the proposed California fish toxicity standards. From
the experimental data it appeared the only way these limitations could
be met was by the use of activated carbon at high regeneration rates as
a final treating step. It appeared that the waste water contained some
non-biodegradable or at least "refractory" organic material as indicated
by COD and TOC tests.
In the Spring of 1973, information was received describing the
rotating biological surface units which were being proposed for treating
industrial waste as an improved alternate to the other biological treat-
ment systems. There were several advantages claimed for this process
such as high biomass concentration; low volume, high density sludge
production and low power requirement.
Subsequently, arrangements were made to install a four stage pilot
rotating biological^ surface unit at the refinery and to compare its
performance with the other pilot plants.
The RBS test program consisted basically of three periods determined
partly by a difference in the quality of raw waste water and partly by the
type of operation of the unit. During the first period, a direct comparison
was made between the trickle filter, the activated sludge and the RBS units
with the same feed going to all three units. For the second test period,
253
-------
the feed rate to the RBS unit was reduced to a very low figure to
establish nitrifying bacteria in the bio-mass. During the third period,
a series of hydraulic loading tests were performed where the rate was
varied from 1/2 gpm to 18 gpm representing hydraulic loading of 0.2 to
6.8 gal/day/sq.ft. of surface area.
As a result of the pilot plant study, it was concluded that the
removal of organic pollutants by the RBS unit compared favorably with
the trickling filter and the activated sludge processes and, for some
waste water parameters, the RBS unit appeared to be superior. The pilot
plant operation verified most of the claims made by the manufacturer,
particularly with respect to energy requirements and ease of operation.
The removal of organics by the RBS unit was very similar to the
trickle filter and the activated sludge with each unit able to achieve
about the same percentage removal and final concentrations in the final
effluent at their optimum operation.
A portion of the test program was devoted to establishing nitri-
fication, and determining the relationship between hydraulic loading and
the degree of conversion of ammonia to nitrate. This was accomplished
by operating the unit at a very low feed rate and adding sodium
bicarbonate to increase the alkalinity- At the low rate, it was possible
to lower the ammonia concentration from 15 to 20 mg/1 to less than
1 mg/1; however, as the feed rate was increased, nitrification decreased
and eventually stopped altogether. Contrary to what was expected,over
50% of the conversion of ammonia to nitrate took place in the first stage.
From the data obtained, it was concluded that, if nitrification is
desired in a.commerical unit, it would have to be designed for about
one-half the hydraulic loading that would be required for organics
remo-, al.
One of the most noticeable differences between the RBS effluent and
the effluent from the other bio systems was the suspended solids content.
Although the suspended solids did increase with feed rate, even at
relatively high hydraulic loading, the RBS effluent had lower suspended
solids than the best operation of the other processes. At low feed
rates, the R^S effluent after 30 minutes of settling, exhibited a
sparkling appearance that was achieved on the other processes only by
filtration or activated carbon treatment of the effluent.
Static bioassays were conducted weekly on samples of the various
pilot plant effluents using the APHA standard methods to determine the
96-hr, median toxicity (TLm). Although the RBS unit was not the answer
to the toxicity problem at the Avon Refinery, in general, this effluent
was less toxic than the effluents from either the trickling filter or the
activated sludge. Activated carbon absorption remained as the only waste
water treatment that would produce a completely non-toxic water (100%
survival) from the waste water stream.
Activated carbon treatment data are presented in Tables I through
VIII.
254
-------
Table I presents early data showing the effect on fish toxicity
of treating various RBS effluents with powdered activated carbon.
s-
Tables II and III present data on the effect of powdered carbon on
other parameters and indicate that toxicity is improved although other
parameters are not greatly affected.
Tables IV and V illustrate the effect of pH changes on toxicity and
possibly the ability of carbon to absorb the toxicants.
Tables of VI and VII show comparisons of two different powdered
activated carbons and indicate that selection of the proper carbon source
can make a very great difference in the ultimate success of carbon
treatment.
Table VIII is a part of a very large table of data obtained on a
granular carbon test conducted over a six month period. It is presented
to further illustrate that long after the carbon was "exhausted" with
respect to COD removal it would continue to produce a non toxic effluent
and that it could be rejuvenated by a hot water backwash.
By the time these pilot plant studies were complete, the final
guidelines had been issued by EPA and the N.P.D.E.S. permit for the
refinery had been issued by the Regional Water Quality Control Board.
This permit outlined not only the discharge limits but a compliance
schedule for submitting a conceptual plan, completion of Engineering,
start of construction and completion of construction. At this time a
thorough review was made of all the accumulated pilot plant data and the
conceptual plan developed. It appeared that of the many parameters of
water quality, COD,' suspended solids and fish toxicity would control the
design of the treatment system. Included in the consideration was the
volume of water which varied considerably with the seasonal storm water
entering the process or oily sewers, since practically all of the annual
rainfall in this location occurs between the first of November and the
first of April. Several alternate plans were considered but all were
basically a supplemental biotreatment, solids removal and activated
carbon treatment.
During the long period of monitoring the raw waste quality it became
evident that treating requirements were also cyclic in that both COD and
toxicity increased during the winter months but during part of the year
the discharge would probably meet the 1977 limits without much, if any,
additional treatment. In our studies with granular activated carbon we
noted that in several instances long after the carbon was "exhausted"
with respect to removal of COD it would still produce a non toxic water.
From this information it appeared the biotreating system should be capable
of handling wide variation in waste loading and that a carbon system should
be designed to be used only when necessary. From Capital cost considera-
tions, possible ease of handling and the indication that relatively small
quantities would be required, the decision was made to use powdered carbon
on a periodic and throw away basis rather than use a granular bed system.
255
-------
In the summer of 1975 the conceptual plan illustrated in Exhibit
II was put out for bids to Engineering-Construction Firms as a "turn key"
project. As a result Engineering was completed by December 1975, field
construction started in February 1976, and completed by January 1, 1977,
all well within the compliance schedule.
At this writing the treating facilities are still in the process of
starting up primarily because of an extraordinary length of time required
for biomass to develop on the RBS units and then delays in correcting
minor difficulties with certain mechanical equipment, instruments and
electrical control systems.
Our principle concern was the difficulty in establishing the biomass.
During the pilot plant phase we had started up three different pilot
plants charging similar waste water and in all cases a good growth was
established within 3 to 4 weeks. However, in the case of our commerical
unit after four weeks there was only a very slight indication of biogrowth
on the first stage. It was determined that low water temperature and
relatively low soluble BOD were responsible for the apparent lack of
bioactivity.
With increased temperature and the addition of higher strength
waste water, supplied with a portable pump for 10 days, we observed an
increase in the growth of the biomass extending through all three stages.
Up to this point only the RBS and the clarifiers were in operation with
the plant effluent returning to the feed surge ponds. However, with the
establishment of biomass, the filter, polymer injection system and sludge
digester were all put into operation. The carbon system was operated for
a short period primarily to test it mechanically.
Only limited data has been obtained at this time however, they
indicate the plant will perform satisfactorily and the waste water
discharge will be in compliance with the July 1, 1977 limitation.
BIOGRAPHY James F. Dehnert
James F. Dehnert is the Environmental
Director for the Avon Refinery of Lion Oi' Com-
pany at Martinez, California. He has a B.S.
degree in Chemical Engineering and a B.S. de-
gree in Chemistry from Washington State University.
He has been employed at this Refinery for thirty
years with various assignments in Techanical
Service, Economic Planning and Unit Operations.
He served as an Area Operation Supervisor be-
fore becoming involved in Environmental assign-
ments.
256
-------
EXHIBIT I
QPE-R.ATIOM
01
Acm PLAUT
API ]DAF
FOUL WATER
STR.IPPER.
CA.WA.t-
DISCHARGE
-------
EXHIBIT II
-------
TABLE 1
TREATMENT OF RBS EFFLUENT WITH
.POWDERED ACTIVATED CARBON
Test
No. 1
Sample
RBS Feed
RBS Eff
RBS Eff +20 ppm PAC
TLm or Survival*
74
90
90%
No. 2
RBS Feed
RBS Eff
RBS Eff + 10, ppm PAC
80
90%
100%
No. 3
RBS Feed 80
2nd Stage RBS 92
2nd Stage RBS 4- 10 ppm PAC
90%
No. 4
RBS Feed
RBS Eff
RBS Eff -I- 10 ppm PAC
< 35
< 75
60%
No. 5
RBS Feed
RBS Eff
RBS Eff + 10 ppm PAC
33
64
60%
No. 6
RBS Feed
RBS Eff
RBS Eff + 20 ppm PAC
" " + 35 ppm PAC
" " + 50 ppm PAC
33
>69
90%
90%
100%
*Survival in undiluted waste
259
-------
TABLE II
f
ACTIVATED SLUDGE EFFLUENT TREATED WITH PAC
Parameter
Toxicity (% Survival)
COD mg/1
NH3(N) mg/1
Oil mg/1
Naphthenic Acids mg/1
Cr(T) mg/1
Cu mg/1
Zn mg/1
Act Sludge Eff
0 (24 hr)
108
35
0.2
1.5
0.02
0.20
0.03
Act Sludge Eff
+100 ppm PAC
100 (96 hr)
84
28
0.1
0.6
0.02
0.25
0.02
TABLE III
ACTIVATED SLUDGE EFFLUENT TREATED WITH PAC
Carbon Dosage (ppm)
Parameter
Toxicity (% Survival)
COD mg/1
Phenol mg/1
Oil mg/1
Naphthenic Acid mg/1
50
100
150
0
150
4.8
0.1
3.1
10
130
4.9
0.1 "
4.3
100
120
4.7
0.1
3.5
100
120
5.2
-v
0.1
3.1
260
-------
TABLE IV
EFFECT OF pH ON TOXICITY
RBS Effluent @ 7.2 pH
RBS Eff Lowered to 6.5 pH
RBS Eff Raised to 8.5 pH
40% Survival
0% Survival
90% Survival
TABLE V
EFFECT OF pH ON CARBON TREATMENT
RBS Feed 7.1 pH
36 TLm
RBS Feed @ 6.5 pH + 30 ppm PAC 65 TLm
RBS Feed @ 7.0 pH + 30 ppm PAC 61 TLm
RBS Feed @ 7.5 pH + 30 ppm PAC 80 TLm
TABLE VI
COMPARISON OF TWO CARBONS
Sample
RBS Effluent (as is)
Toxicity
58 TLm
TOG (mg/1)
68
RBS Eff +60 ppm PAC-A
+90 ppm PAC-A
85 TLm
93 TLm
57
57
RBS Eff + 15 ppm PAC-B
+30 ppm PAC-B
+60 ppm PAC-B
93 TLm
90% Survival
100% Survival
57
54
44
-------
TABLE VII
COMPARISON OF TWO CARBONS
PAC-A Added Continuously to RBS Pilot Plant and
Additional PAC-A or PAC-B Added to RBS Effluent
RBS Eff 4- 45 ppm PAC-A 40 TLm
RBS Eff 4- 195 ppm PAC-A 86 TLm
RBS Eff + 45 ppm PAC-A 4-50 ppm PAC-B 93 TLm
RBS Eff 4-45 ppm PAC-A 4-75 ppm PAC-B 100% Survival
262
-------
20
TABLE VIII
GRANULAR CARBON TREATMENT OF TRICKLING FILTER EFFLUENT
COD mg/1
Toxicity (Survival)
Column
Day
1
2
3
4
5
6
7
8
9
10
.Feed
180
180
180
200
260
300
-
-
270
320
No.l
70
80
100
120
210
260
240
250
270
300
No. 2
60
30
50
50
140
160
170
200
220
250
No. 3
30
30
30
' 30
80
110
100
140
170
210
No. 4
10
30
40
50
90
100
70
100
140
170
Column
Feed No . 1 No . 2
100
100
90
' 70
0
100
61TLm 100
No.3 No.4
220
210
170
160
100
30 180 130 130 120 110
34TLm
71TLm
100
100
45
50
55
60
70
75
100
Hot
190
150
150
Hot
170
160
170
Water
160
140
150
Water
170
160
170
Wash
120
140
140
Wash
160
130
150
120
130
140
130
130
130
120
100
140
130
45TLm
53TLm
\
59TLm
45TLm
80
0
100
0
80
0
100
100
100
100
263
-------
SESSION VIM
ADD-ON GRANULAR ACTIVATED CARBON
Chairman
Nicholas D. Sylvester
Professor of Chemical Engineering
University of Tulsa, Tulsa, Oklahoma
Speakers
Fred M. Pfeffer
W. Harrison and L. Raphael ian
"Pilot-Scale Effect on Specific Organics Reduction
and Common Wastewater Parameters"
R. H. Zanitsch
R. T. Lynch
"Granular Carbon Reactivation-State of the Art"
L. W. Crame
"Activated Sludge Enhancement: A Viable
Alternative to Tertiary Carbon Adsorption?"
264
-------
"GRANULAR CARBON REACTIVATION: STATE-OF-THE-ART"
R. H. Zanitsch
Engineering Director, Calgon Environmental Systems Division
R. T. Lynch
Process Engineer, Calgon Corporation
Use of granular activated carbon for treatment of industrial waste-
water is receiving widespread acceptance. In the past several years, 100
adsorption systems have been installed in industrial plants. Applications
range from dye plant wastewater reuse to removal of toxic materials. Gran-
ular carbon is being used to treat flows as low as 1,000 gallons per day
to as high as 20,000,000 gallons per day in industrial waste applications.
It is being employed as a pretreatment step to remove toxic materials prior
to biological treatment, as the main treatment process and for tertiary
treatment of biological plant effluents.
In most industrial wastewater applications, cost of virgin carbon pro-
hibits using it on a throw-away basis. Chemical regeneration is feasible
in only a limited number of applications and regenerant disposal remains a
problem. Thermal reactivation is in most cases complete, efficient, and
economical whether it is performed on-site or on a contract basis at a
central reactivation facility.
The technology of reactivation with industrial waste carbons has de-
veloped in only the last ten years. There are now approximately twenty re-
..activation systems installed in the United States which are reactivating
industrial wastewater carbons.
New thermal reactivation processes (such as fluidized beds and elect-
ric furnaces) are now being developed but no commercial experience with
industrial wastewater carbons has been developed in the United States.
For the purposes of this presentation, we will discuss our experience with
the design and operation of multiple hearth furnaces and rotary kilns as
they relate to industrial wastewater applications.
THE THERMAL REACTIVATION PROCESS
Granular carbon is usually wet when fed to the reactivation furnace.
Water concentration is a function of carbon size, water temperature dur-
ing the dewatering step, and the amount of adsorbate on the carbon. In
practice, moisture content varies from 40 to 50 percent on a wet spent
basis.
The reactivation process can be divided into three steps:
1. Evaporation of moisture on the carbon (Drying).
r
265
-------
2. Destructive distillation of organics resulting in pyrolysis of
a portion of the carbon from the organic materials (Baking).
3. Activation of the carbon by selectively burning carbon deposited
during the organic removal step (Activation).
During the drying step, carbon temperature is increased to approxi-
mately 212°F (100°C) and moisture evaporates into the gas phase. As moist-
ure evaporates it is also possible for highly volatile organics to be steam
distilled.
The second step is termed baking or pyrolysis of the adsorbate. Dur-
ing this step, carbon temperature increases to approximately 1200-1400°F
(649-760°C). A portion of the organic molecules are thermally cracked to
produce gaseous hydrocarbons which are driven off. The remaining lower
molecular weight organics are distilled. During this process, a carbon char
is deposited in the pore structure of the original activated carbon.
The final step is activation of the carbon - a chemical reactivation
whereby carbon char deposited during the baking step is combusted along
with a small amount of the original carbon. By this time, temperatures
are in the range of 1600-1800°F (871-982°C).
Since the fixed carbon and the granules are both carbon, the^process
requires that fixed carbon be selectively gasified with minimum gasifi-
cation of the granular carbon. Steam is added to the furnace and oxygen
concentration is controlled to promote gasification of the fixed carbon
while minimizing burning of the original* granular carbon.
REACTIVATION SYSTEM DESCRIPTION
The basic sequence for thermal reactivation is as follows: (See Ex-
hibit 1).
Spent carbon is removed from the adsorbers and transferred as a slurry
to a spent carbon storage tank. Spent carbon is then transferred to an
elevated furnace feed tank from which it is metered, at a controlled rate,
to a dewatering screw. The dewatering screw is an inclined screw conveyor
which serves the dual purpose of gravity draining slurry water from the
granular carbon and providing a water seal for the top of the furnace. A
timer operated valve is used to meter carbon to the dewatering screw.
Drained, but wet, spent carbon then gravity flows into the furnace where
it is dried, baked, and reactivated as discussed earlier. Reactivated
carbon exits the furnace by gravity and enters a quench tank. The quench
tank serves the dual purpose of wetting the reactivated carbon and provid-
ing a bottom seal for the furnace. The carbon is then transferred to a
reactivated carbon storage tank from which it is then returned to the ad-
sorbers as needed. In most industrial waste applications, an afterburner
and scrubber are provided for destroying organics and removing residual
particulates from the furnace off-gases.
266
-------
The spent carbon storage tank should be designed for five to ten days
storage of carbon in order to allow for routine furnace maintenance and un-
scheduled shut-downs. This is usually a lined carbon steel tank with a
cone bottom to facilitate carbon flow. The reactivated carbon storage tank
is usually sized in the same manner with the same materials of construction.
In the case of the reactivated carbon storage tank, facilities should be
provided for adding virgin makeup carbon to the system as required.
The furnace feed tank is usually sized for at least one shift of op-
eration. The feed tank insures a constant carbon feed to the furnace in-
dependent of the large storage system. This tank is usually a cone-bottom,
lined carbon steel tank.
Spent and reactivated carbon are transferred in slurry form using
either eductors, blowcases or slurry pumps. In the case of eductors and
pumps, dilution water must be provided in order to reduce slurry concen-
tration to less than one pound per gallon to minimize carbon abrasion and
line erosion. Eductors are generally applicable in non-corrosive services
where static head is not great. Pumps can be used satisfactorily in high-
head applications, but are subject to erosion and plugging. Eductors and
pumps require use of dilution water which necessitates installation of
water recycle systems. The blowcase is an efficient method of, transferring
carbon. Both air and water have been used to pressurize blowcases. In the
case of a blowcase, carbon is transferred in a much denser slurry (three
pounds per gallon) and, therefore, care must be taken to maintain control
over line velocities to minimize abrasion and wear. Material for carbon
slurry lines should be compatible with the wastewater. As long as slurry
lines are flushed free of carbon after each transfer, galvanic corrosion
of carbon steel lines will not be a problem; however, if the wastewater is
corrosive, more exotic materials of construction should be used. All car-
bon slurry lines should be equipped with flush connections to facilitate
flushing and unplugging.
CAPITAL AND OPERATING COST ESTIMATES
r<
Based on our experience with the design, installation, and operation
of multiple hearth furnaces and rotary kilns for reactivating industrial
waste carbons, we have estimated the installed cost of reactivation systems
to reactivate 5,000, 10,000, 30,000, and 60,000 pounds per day. (See Ex-
hibit 2). The capital cost curve shown in Exhibit 3 represents a total
installed cost including all equipment, site preparation, foundations, in-
stallation, startup, and indirects. We have assumed that necessary util-
ities and off-site facilities are available at the battery limits. As you
can see, we estimate the total installed cost of a 10,000 pound per day
reactivation system to be approximately $1.25 million plus or minus 20 per-
cent. The time required to design, procure, install, and startup a re-
activation system is usually estimated to be two years assuming a twelve-
month delivery time on the furnace and associated equipment.
We have also estimated direct operating costs for reactivating 5,000,
267
-------
10,000, 30,000, and 60,000 pounds per day of industrial wastewater carbons as
shown in Exhibit 4.
In order to develop these costs, the following elements were consider-
ed:
1. Labor was estimated to be one operator per shift at a rate of
$10/hour. An allowance of 25 percent of the labor cost for
supervision was also included.
2. Fuel was estimated at 8,000 BTU's per pound at a cost of $3/million
BTU's. This estimate includes afterburner operation and an allow-
ance for inefficiencies due to interruptions and reduced feed
rates. Approximately half of the fuel consumption is required for
the afterburner and idling.
3. Power costs for the reactivation system are minimal and were as-
sumed to cost $0.03/KWH.
4. Steam costs were based on an average demand of one pound of- steam
per pound of carbon for reactivation at a cost of $4/1,000 pounds.
5. Maintenance costs for an industrial wastewater application can
range from 8 to 15 percent of the reactivation system cost per
year. For this estimate, we assumed a maintenance cost of 8 per-
cent per year.
6. Makeup carbon costs were based on an average carbon loss rate of
7 percent and a virgin carbon cost of $0.57/pound delivered.
Carbon losses can range from as low as 3 percent to greater
than 10 percent depending on design and operation of the system. •
Most industrial waste systems operate in the 5 to 7 percent loss
range. Makeup carbon costs represent the highest individual cost
element in the direct operating cost estimate and, therefore, all
efforts should be made to minimize carbon losses through good de-
sign and operation.
7. A general plant overhead of 10 percent of the above cost was al-
lowed to cover such items as insurance, taxes, monitoring, ac-
counting, and administration.
As can be seen from Exhibit 4, the direct operating cost for a re-
activation system handling industrial wastewater carbons ranges from
$0.11 to $0.19/pound over the range investigated. This does not include
depreciation or amortization of investment. The economies of scale are
obvious. We feel these costs can range plus or minus 20 percent, but in
general, reflect the cost to operate a reactivation system on industrial
waste applications.
268
-------
MULTIPLE HEARTH FURNACE
Exhibit 5 is a cross-sectional view of a multiple hearth furnace. The
furnace consists of a cylindrical refractory-lined steel shell containing
several refractory hearths and a central rotating shaft to which rabble arms
are attached. From four to eight hearths are used in carbon reactivation
furnaces. The center shaft and rabble arms are cooled by air supplied by
a centrifugal blower discharging air through a housing into the bottom of
the shaft. A sand seal at the top of the furnace and a sand or water seal
at the bottom are used to seal the furnace against introduction of extran-
eous air.
In operation, wet spent carbon is introduced through a chute into the
outside of the top hearth of the furnace. The rabble arms are equipped
with solid alloy rabble teeth which rake the carbon towards the center where
it drops to the hearth below. The teeth on the rabble arms are arranged to
move the carbon in a spiral path. The action is gentle to minimize at-
trition. The top hearth is termed an "in" hearth since carbon flow is in-
ward.
The second hearth is consequently an "out" hearth where the carbon is
moved outward by the rabble teeth. Out hearths have a series of holes
around the periphery of the hearth through which the carbon drops to the
next lower hearth.
In this manner, carbon passes through the furnace until it is finally
discharged through a chute in the bottom hearth into the water filled
quench tank. The chute extends under the water level in the quench tank
to provide a seal.
Drying is accomplished in the upper one-third of the furnace. Dis-
tillation and pyrolysis of the adsorbate occurs in the next one-third.
Activation of the carbon is completed in the bottom one-third of the furn-
ace.
Burners are mounted tangentially on the furnace shell in burner boxes.
Usually burners are placed on the bottom two or three hearths and on one
upper hearth below the lowest drying hearth. However, if desired, burners
can be mounted on any hearth including the drying hearths.
On small furnaces, two burners per fired hearth are used. On larger
furnaces, three burners are installed. The burners are of the nozzle-mix
type burning fuel oil or natural gas. Dual fuel burners are commonly-
employed to burn gas when it is available and fuel oil at other times.
Steam addition ports are provided on the bottom two or three hearths
to add steam for control of the reactivation process.
The center shaft is driven through a variable speed drive at 0.5 to
2.5 rpm.
269
-------
A number of furnaces have been installed with integral or "0" hearth
afterburners. This is less costly from a capital cost standpoint than the
separate afterburner.
DIRECT FIRED ROTARY KILN
Exhibit 6 is a simplified sketch showing a direct-fired rotary kiln.
The kiln is a refractory lined steel shell enclosed on each end with re-
fractory lined stationary hoods. This sketch depicts a counter-current
operation where gas flow is opposite the carbon flow. Co-current operation
is also possible and one carbon reactivation kiln is currently operating
in this manner.
The kiln is mounted on two or three sets of trunions depending on the
length of the unit. The kiln is sloped from the feed to the discharge end
and one set of thrust rolls are used to maintain the kiln in position on
the trunions. Proper training and alignment of trunions is important to
minimize excessive wear of the trunions and tires.
The kiln is driven through a variable speed drive coupled to a speed
reducer and pinion gear which meshes with a bull or girt gear mounted on
the kiln shell. The kiln is equipped at each end with hoods. Rotary seals
are used to seal between the rotating kiln shell and the stationary hoods.
The hoods are refractory lined.
\
A feed screw or chute is used to feed wet carbon into the kiln. Flights
are usually employed to advance the damp carbon and to shower the carbon in
the feed end to obtain high heat transfer rates during the evaporation step.
Flights are also used in the first portion of the baking step up to a point
where the temperature reaches approximately 1200-1400°F (649-760°C). Ma-
terial of construction for the flights is a function of carbon corrosive-
ness and reactivation conditions in the kiln.
The hot reactivated carbon, at a temperature of 1600-1800°F (871-982°C),
discharges from the kiln and falls down the discharge chute into a water-
filled quench tank. The discharge chute extends under the water level in
the quench tank to form a seal to eliminate air leakage into the kiln.
A burner is mounted in the discharge hood to provide heat for the re-
activation process. Either fuel oil or gas may be burned. The burner
air-to-gas ratio is adjusted to minimize oxygen concentration in the kiln.
A steam addition port also is provided in the discharge hood to admit steam
into the kiln for control of the reactivation process.
The exhaust gases, at a temperature of 500-800°F (260-427°C), leave
the kiln through a duct connected to the feed hood. Iii most installations,
gases are passed through an afterburner for complete combustion of organics
and burning of carbon fines swept out of the kiln. New installations, as
is the case with the multiple hearth furnace, will probably require in-
stallation of a wet scrubber to meet air pollution codes in most areas of
the country.
270
-------
kiln .I/?7 -in' factivation Pro"ss is controlled by varying the
rate and h r° PT *?"*" retentio» **•*'«* by adjusting the steam
bon to b b"rne* tempf atfe at the ^ttings required for the particular car-
bon to be reactivated. A steam rate of 0.6-1.2 pounds per pound of carbon
and a temperature of 1600-2000'F (871-1093'C) are ranges encountered in
practice.
OPERATING AND MAINTENANCE PROBLEMS
, A number of unique operating and maintenance problems have been ex-
perienced in reactivating industrial wastewater carbons. These problems
include:
Corrosion
Slagging
Poor Reactivated Carbon Quality
High Carbon Losses
Feed Interruptions
Hearth Failures
Slurry Line Erosion and Corrosion
Corrosion
Selection of proper construction materials for carbon storage and
handling systems is very important. We have found lined carbon steel
tanks to be satisfactory, but proper selection and application of lining
material is extremely important. Lining material should be corrosion
and abrasion resistant. We recommend thorough corrosion coupon testing
prior to making a final selection. Erosion of lining material at carbon
outlet nozzles, followed by corrosion of the metal, has been a problem.
We have installed sacrificial wear plates or stainless steel cones on tanks
in order to minimize this problem. Dewatering screws and quench tanks are
generally constructed of 304 or 316L stainless steel. In general, these
materials are satisfactory for most applications. However, the dewatering
screw is exposed to the spent carbon slurry and, therefore, its material
must be compatible with the wastewater. Corrosion of rabble arms and teeth
in multiple hearth furnaces and lifting and drying flights in rotary kilns
can be a problem when handling chlorinated hydrocarbons and organic sulfur
compounds. Special attention must be given to material selection to mini-
mize this problem.
Slagging
Formation of clinkers and slag in the furnace is generally a function
of sodium and/or organic phosphate content of the spent carbon. Slag
formation can be minimized by pretreatment of carbon and maintenance of
proper furnace conditions. Formation of slag can generally be attributed
to constituents in the water contained in the pores of the carbon as it
271
-------
enters the furnace. These chemicals react with alumina and silica in the
furnace refractory resulting in slag formation.
Hearth Failures
Hearth failures in multiple hearth furnaces can generally be attributed
to cyclic operation. Frequent feed interruptions, resulting in temperature
excursions on the upper hearths, will weaken hearths and ultimately lead to
failure. By minimizing the number of feed interruptions and maintaining
continuous furnace operation, upper hearth life can be maintained for three
to five years. Another problem leading to hearth failure is brick attack
by sodium compounds; which leads to the slagging problem discussed earlier.
Also, improper dewatering or improper operation of the dewatering screw,
which would result in excessive amounts of water entering the top hearth,
can result in thermal shock which leads to failure. In general, hearth life
is a function of the operating philosophy of the furnace. If frequent feed
interruptions due to improper carbon feed system design or cyclic operation
are encountered, poor hearth life can be expected.
Carbon Losses
As mentioned earlier, makeup carbon cost is the single most important
cost element for a reactivation facility. By properly designing the ad-
sorbers, the carbon transfer and handling systems and the carbon storage
and reactivation systems, losses can be controlled at the 5-7 percent level.
Within the reactivation furnace itself, carbon losses should not exceed
1-3 percent. Carbon which is lost is due to oxidation during the activation
step. This can be controlled by maintaining oxygen levels in the activation
zone at 0-2 percent, or roughly that required for destruction of organics
without sacrificing carbon. Most carbon losses in a granular carbon system
occur due to backwashing of carbon in adsorbers, abrasion in slurry lines,
spillage, and carryover in overflow lines. These losses can all be mini-
mized by proper design of the basic system. Care must be given to overflow
rates, backwash rates, and slurry line velocities, and frequent checks must
be made to see that good housekeeping and operating techniques are being
followed.
Slurry Line Erosion and Corrosion
As mentioned earlier, construction material of spent carbon slurry
lines should be compatible with the wastewater in order to minimize cor-
rosion. If spent carbon or reactivated carbon is allowed to accumulate in
a carbon steel slurry line, galvanic corrosion can be expected. Therefore,
flushing of all slurry lines after each transfer is recommended. If the
wastewater is extremely corrosive, we recommend lined steel or stainless
steel slurry piping be considered.
Erosion of slurry lines can be attributed to excessive transfer
27?
-------
velocities. We Recommend a slurry line velocity of 3-5 feet per second
which is sufficient to prevent settling and minimize abrasion. Also,
slurry lines should be as direct as possible with a minimum number of
bends. We recommend that long-radius bends be used to minimize abrasion.
Also, we recommend that all bends be accessible for periodic inspection
and replacement. Flush connections should be provided at frequent inter-
vals on all slurry lines in case a line becomes plugged.
FURNACE SELECTION CRITERIA
Choice of a reactivation furnace depends on many factors. A thorough
analysis of each type of equipment plus reactivation characteristics of
the carbon are necessary to make a final decision on which piece of equip-
ment to use.
Both multiple hearth furnaces and kilns are being employed to react-
ivate granular activated carbon used in industrial wastewater treatment.
The quality of reactivated carbon that can be achieved is the same for both
units.
Major parameters influencing the selection of a reactivation furnace
are as follows:
, 1. Capital Cost - Total installed costs for either a multiple hearth
furnace or a rotary kiln are approximately the same. The purchase
price is generally higher for a multiple hearth furnace. However,
the installation costs are lower which tends to make installed
costs the same. Site preparation, foundations, and structural
costs are higher for a rotary kiln because of the greater area re-
quired to install a kiln.
2. Area Requirements - Area requirements are much greater for a rotary
kiln than for a multiple hearth furnace. The kiln also requires
more foundations and structural steel for walkways than a multiple
hearth furnace. The multiple hearth furnace is higher than a kiln
which means more structural steel is required to support the car-
bon feed equipment.
3. Fuel Consumption - Fuel consumption is higher in a rotary kiln be-
cause of higher heat losses. In a multiple hearth furnace, in-
sulation is used behind the wall brick to minimize heat loss.
This is not possible in a rotary kiln. Surface area is also higher
in a rotary kiln than a multiple hearth furnace of equivalent ca-
pacity. Fuel consumption for each will be in the following ranges
depending on capacity and operating rate as a fraction of rated
capacity.
273
-------
BTU/LB Carbon*
Multiple Hearth Furnace 2500-4500
Rotary Kiln 3500-8000
*Does not include afterburner fuel requirements.
4. Capacity Turndown - Capacity turndown ratio is defined as the per-
cent of rated capacity at which the furnace can be operated while
producing good reactivated carbon with reasonable carbon loss.
Capacity turndown for the equipment being evaluated in this paper
are as follows:
Multiple Hearth Furnace - 33 Percent
Rotary Kiln - 50 Percent
The multiple hearth furnace can be operated at a lower fractional
capacity because of the greater degree of control that can be ob-
tained in various zones of the furnace. In a rotary kiln, with
only one burner and one steam addition point, kiln speed is the
major parameter that can be varied to operate at lower capacities.
5. Degree of Control - Better reactivation process control can be
achieved in a multiple hearth furnace because the furnace is di-
vided into distinct zones according to the number of hearths in
the furnace. Each hearth can be equipped with burners, steam
addition, and air addition which can be controlled independently.
Thus, it is possible to control temperature and vary the atmosphere
in each hearth to optimize carbon reactivation.
f'
In a rotary kiln, the steam port, and burner can only be mounted in
the firing end of the kiln. With this arrangement, the degree of
control that can be achieved is less than in a multiple hearth
furnace. In a properly sized kiln, this is not a distinct dis-
advantage and good carbon reactivation can be achieved. However,
as discussed previously, capacity turndown is not as great in a
kiln.
6. Corrosion and Slag - Many industrial waste streams contain in-
organic impurities which can cause corrosion and slag formation in
the reactivation furnace. These impurities are mostly chloride
and sulfur salts of calcium and sodium. The multiple hearth furn-
ace has more exposed alloy parts than a kiln and is, therefore,
more susceptible to corrosion. Rabble teeth and arms are expen-
sive, long delivery castings as opposed to the alloy flights in a
kiln which are fabricated from readily available plates. Also,
considerable corrosion of flights can occur in a kiln before re-
placement is required.
Slag buildup in a multiple hearth furnace will require periodic
274
-------
shutdowns to remove accumulated material. Slag in a rotary
kiln will be discharged, with the reactivated carbon, into the
quench tank where it can be removed without shutting down the
process.
7. Maintenance - Experience with reactivating industrial wastewater
carbons indicates higher maintenance costs in a multiple hearth
furnace. The factors responsible are:
a. Corrosion and slag formation resulting in shutdowns for
repairs.
b. Rabble teeth and arms are more expensive to replace than
alloy flights used in a rotary kiln.
c. Multiple hearth furnaces are more difficult to work on.
It takes more man-hours to rebuild a hearth than to re-
place brick in a kiln. Because of these factors, down-
time to affect repairs is longer in the multiple hearth
furnace.
d. More instrument components are required with a multiple
hearth furnace.
8. Effect of Feed Outages - The upper hearths in a multiple hearth
furnace can be damaged from temperature cycling caused by inter-
ruptions in furnace feed. Periodic planned shutdowns can be con-
ducted without hearth damage.
Feed outages are usually not a major problem in a rotary kiln. The
refractory is much less effected by temperature cycling in a kiln
than the hearth refractory in a multiple hearth furnace.
9. Operating Factors - Operating factors for kilns and multiple hearth
furnaces are as follows:
r1
Rotary Kiln 85-95 Percent
Multiple Hearth Furnace 75-90 Percent
Multiple hearth furnaces must be shutdown more often to clean slag
and replace rabble teeth. Tftien a furnace is down for repairs, the
work requires more man-hours to complete than similar work on a
rotary kiln.
Based on the above parameters, the multiple hearth furnace offers
the following advantages over a rotary kiln:
• Better control of temperature and atmosphere
• Lower fuel consumption
• Greater capacity turndown
275
-------
• Less area required
• Lower carbon losses from carryover and attrition
The rotary kiln advantages are:
• Less corrosion and slag formation
• Less downtime
• Lower maintenance costs and easier maintenance
• Less effect from feed outages
• Easier to operate
SUMMARY AND CONCLUSIONS
Granular activated carbon has been demonstrated to be effective in
treatment of a wide variety of industrial wastewaters. Both multiple hearth
furnace and rotary kilns can satisfactorily reactivate spent carbons used in
industrial wastewaters provided adequate consideration is given to selection
of materials, sizing of equipment, and operating philosophy. Experience
gained over the last ten years indicates that corrosion, slagging, poor re-
activation quality, carbon losses and line erosion can all be minimized
through good design. Although the same types of problems exist in in-
dustrial purification and municipal water and waste treatment applications
using granular activated carbon, they are magnified in industrial waste-
water applications where wastewater quality, and thus carbon exhaustion
rates, are more variable and substantially more corrosive. However, our
experience with reactivating over 100,000,000 pounds of spent carbon for
more than 75 different industrial wastewater applications, indicates that a
high quality product can be produced on a reliable, economical basis.
276
-------
BIOGRAPHIES
Roger H. Zanitsch is Engineering Director
of the Calgon Environmental Systems Division
for Calgon Corporation. He joined Calgon as a
Project Engineer in 1969 and was later named
Project Manager of the Environmental Engineer-
ing Department. Zanitsch received his BS degree
in Civil Engineering from the University of Cin-
cinnati and an MS degree in Environmental Engin-
eering from the same school. Zanitsch is a mem-
ber of the Water Pollution Control Federation.
Richard T. Lynch is a senior engineer in the
Process Engineering Group of Calgon Corporafion's
Engineering Department. Me has a B.S. degree in
Chemical Engineering from the Universify of Florida.
He has been a project manager for-the design of
several carbon adsorption reaction systems treating
industrial waste streams. He is a member of the
American Institute of Chemical Engineers and a
registered professional engineer in Florida.
277
-------
EXHIBIT 1
TYPICAL REACTIVATION SYSTEM
FLOW DIAGRAM
Exhaust
Gas
Spent Carbon
From Adsorbers
Scrubber
Spent
Carbon
Storage
5-10
Days
Blowcase
\
M. H. Furnace
Or
Rotary Kiln
React
Carbon
Storage
5-10
Days
Reactivated Carbon
To Adsorbers
Blowcase
278
-------
EXHIBIT 2
INSTALLED COSTS OF REACTIVATION SYSTEMS
CAPITAL COSTS
Capital Cost Estimate
($ Million)
Purchased Equipment
Installation*
Reactivation Rate - 1,000 Lbs/Day
5,000
10.000
30.000
60.000
0.24
0.61
0.85
0.36
0.91
1.27
0.77
1.93
2.70
1.20
3.00
4.20
*Ihstallation costs include foundations, structural equipment setting,
electrical, instrumentation, site preparation, engineering contractor,
overhead and profit, and indirects.
Fuel - 8,000 BTU/Lb @ $3/106
BTU
Power @ 3
-------
EXHIBIT - 3
10,000
9 -
8 -
7 —
6 —
5 -
4 —
O
O
O
rsj
0 1.000
9
_l 7
< 6
W 5
z
<
O
100
CAPITAL COST ESTIMATE
FOR REACTIVATION SYSTEMS
I I I
I
I
I I I I I
I I I I I
I I I I I
1,000
CARBON
•56769 2
10,000
REACTIVATION RATE -
6789
100,000
5 6789
LBS/DAY
-------
EXHIBIT - 4
10 Or
9 —
8 -
7 -
6 -
S
Z
o
m
oc
<
o
to
i-
v>
o
o
o
a:
UJ
a.
o
o
UJ
10
9
8
DIRECT OPERATING COST
FOR REACTIVATION SYSTEMS
1,000
J_
I I I I I I
3 496789 2 3 456789
10,000 100,000'
CARBON REACTIVATION RATE - LBS/DAY
8 6 7 8 9
-------
EXHIBIT 5
Carbon In
Gas Out
Hearth
6'
3' Diameter
Rabble Arm
Rabble Teeth
Two Burners and
Steam Inlets at
Hearth 4, 5 and 6,
not shown
Carbon Out
CROSS SECTIONAL VIEW OF MULTIPLE HEARTH FURNACE
USED AT POMONA WATER RECLAMATION PLANT
282
-------
EXHIBIT 6
Dewatered
Carbon
Carbon
Feed
Screw
.Gas Discharge
Duct
Discharge
End Seal
>^Firing Hood
=#-*- Steam
Fuel
Hood
Ring
X^ Discharge
Chute
283
-------
DISCUSSION
L. L. Krohn, Union Oil Co.; Have you made an analysis of the off-gases
from that furnace?
Roger Zanitsch; Yes, we have. We've looked at in excess of 100 different
industrial waste carbons. We've analyzed the off-gases before and after
afterburning on many of these, to determine what temperature and residence
time is needed for organic destruction.
L. L. Krohn. Union Oil Co.; Do you have any feel for the particulate matter
coming out?
Roger Zanitsch; We have systems operating that are designed with a 2 second
residence time which is primarily there to consume the particulates. If
you have a half second residence time which is certainly sufficient to des-
troy most of the organics present, you'll still have some carbon fines that
will need to be scrubbed. We feel that the 1-2 second residence time at
1800°F can destroy the carbon fines as well as the organics.
L. L. Krohn, Union Oil Co.; Considering the new source review - can we hope
to build this system?
Roger Zanitsch; Reactivation furnaces are now in operation and many are
being designed for industrial waste applications. Technology exists to
handle essentially all air pollution control requirements at a reasonable
cost.
Mac McGinnis. Shirco. Inc.; We have made some of the economics that you're
talking about for our electric regeneration furnace and compared them with
similar economics as you have presented here from multiple hearth and other
approaches and just a couple of comments - a couple of factors that we have
included that you haven't mentioned are in the area of utilities, scrubber
water which on small capacity units may be a fairly significant contribution
of operating costs; and the other factor you mentioned quality of the pro-
duct, laboratory labor, lab time to confirm that the product is indeed of
the desired quality, can be a fairly significant contribution.
Roger Zanitsch; I'm glad you brought this up. In the analysis that I show-
ed, the operating cost included a 10% general plant service allowance on
the total operating cost to cover overhead items such as accounting, qual-
ity control, etc. As far as scrubber water cost and disposal, it can be a
factor. In those installations where we have scrubbers, we've recycled
water through the pretreatment system to remove the carbon fines. Frankly,
we haven't found this to be a significant cost factor.
284
-------
DISCUSSION
Mac McGinnis, Shirr.o. Inc^ Well, a half cent here and half cent there, it
oegins to add up. The other general comment is you've indicated that there
is considerable data on regeneration costs in multiple hearth furnaces in
particular and you've showed us some trend lines in terms of direct operat-
ing costs. Can you comment on any specific data, you know, accumulated over
a period of time that indicates an actual cost figure for some specific ap-
plication?
Roger Zanitsch; The numbers which I presented are based on our experience
in operating both small and large furnaces.
Mac McGinnis. Shirco. Inc.; One last comment - would you say then that the
actual data would fall within that plus or minus 20% about your nominal
curve?
Roger Zanitsch; On industrial waste applications, yes. In process applic-
ations, such as the decolorization of sugar solutions, operating costs are
substantially lower since they have a constant feed and a very predictable
product.
Colin Grieves, Amoco Oil Co.; First, would you care to comment on some of
the new technology which you eluded to? And second, would you like to say
anything about regeneration of powdered activated carbon?
Roger Zanitsch; As far as the new technologies are concerned, I was person-
ally thinking of the electric furnace and the fluidized bed furnaces. The
Japanese have several different types of furnaces. Most of the experience
with the newer furnaces has been in either pilot-scale or on the commercial
scale, but in considerably less corrosive application than you have in in-
dustrial wastes. In industrial waste applications, the big awakening has
been in the areas of corrosion, maintenance costs, and feed interruptions.
The new technologies have not been demonstrated in this type of service.
As the new technology develops, it's going to take some time to gain the
experience necessary to apply these new furnaces in the industrial waste
effort. As far as powdered carbon activation, I don't really feel qualified
to discuss it on the basis that I would only be expressing my opinions
since no commercial experience has been developed.
285
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-79-177
3. RECIPIENT'S ACCESSION-NO.
4. Tl.TLE AND SUBTITLE
/Activated Carbon Treatment of Industrial Wastewaters-
Selected Technical Papers
5. REPORT DATE
August 1979 issuing date
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO
ndustrial Sources Section, Source
.Obert S. Kerr F.nvirrmmpntfl1
tent Branch
oratory
9. PERFORMING ORGANIZATION NAME AND ADDRESS /
Industrial Sources Section - Source Management Branch
Robert S. Kerr Environmental Research Laboratory
P.O. Box 1198
Ada, Oklahoma 74820
10. PROGRAM ELEMENT NO.
1BB610
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
Robert S. Kerr Environmental Research Laboratory
U.S. Environmental Protection Agency
P.O. Box'1198
Ada, OK 74820
13. TYPE OF REPORT AND.PERIOD COVERED
In-House
14. SPONSORING AGENCY CODE
EPA 600/15
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Because of the tremendous interest in the organic constituent removal by
activated carbon, the two industrial categories displaying the most interest are
the petroleum refining and petrochemical industries. EPA's Office of Research and
Development has co-sponsored two technical symposia for the petroleum refining/
petrochemical industries, and activated carbon treatment as an important section
of both agendas. The technical papers presented research activities conducted by
consultants, industries, and EPA.
\
The presentations made at these symposia have been arranged into the following
sequence: (1) State-of-the-Art, (2) Organic Compound Removal, (3) Granular Pilot-
Scale Studies, (4) Powdered Activated Carbon Pilot-Scale Studies, (5) Full-Scale J
Granular Activated Carbon Treatment, (6) Full-Scale Powdered Activated Carbon Treat-
ment, and (7) Activated Carbon Regeneration.
/
Economics of Activated Carbon Treatment are presented in the applicable in-
dividual technical papers and is not a separate topic for this report.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Activated Carbon Industrial Waste Treat
Mixed Media Filter ment
Environmental Engineering
Chemical Engineering
Organic Chemistry
Refining
Refining Wastewaters
Processes & Effects
Characterization
Measurement & Monitorinc
Bench-Scale Plants
Petrochemical Industry
68-D
71 -C
89-B
94-E
97-R
3. DISTRIBUTION STATEMENT
Release to public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
308
20,
Y. CJ-ASS (This page)
SI
22. PRICE
EPA Form 2220-1 (9-73)
286
•U.S. GOVERNMENT PRINTING OFFICE: 1979 0-620-007/6322
------- |