ALKALINE AND STRETFORD SCRUBBING TESTS FOR
AL FROM IN-SITU
RETORT OFFGAS
H2S REMOVAL FROM IN-SITU OIL SHALE
by:
. H. J. Taback, P.E.
G. C. Quartucy, P.E.
R. J. Goldstick, EDS
KVB, Inc.
Engineering and Research Division
Irvine, CA 92714
Under Subcontract to
Metcalf & Eddy, Inc.
Wakefield, MA 01880
EPA Contract No. 68-03-3166
Project Officer
Edward R. Bates
Hazardous Waste Engineering Research Laboratory
Cincinnati, OH 45268
Air and Energy Engineering Research Laboratory
Office of Research and Development
Research Triangle Park, NC 27711
•May, 1985
-------
NOTICE
The information in this document has been funded wholly or in part by
the United States Environmental Protection Agency under contract number EPA
68-03-B166 to Metcalf & Eddy, Wakefieldj MA. It has been subject to the
Agency's peer and administrative review, and it has been approved for publica-
tion as an EPA document. Mention of trade names or commercial products does
not constitute endorsement or recommendation for use.
-------
FORWARD
When energy and material resources are extracted, processed, converted
and used, the related pollution impacts on our environment and even on our
health often require that new and increasingly more efficient pollution
control methods be used. The Air and Energy Engineering Research'Laboratory-
Research Triange Park, N.C. assists in developing and demonstrating new and
improved methodologies that will meet these needs both efficiently and
economically.
This report provides data characterizing in-situ oil shale offgases
from the Geokinetics plant in eastern Utah and assessing the effectiveness of
Stretford and Alkaline scrubbing processes in controlling the emission of H,,S
and other sulfur compounds. The results should assist developers and permit
writers in selecting appropriate controls for the treatment of oil shale
offgases.
Frank Princiotta, Director
Air and Energy Engineering Research
Laboratory
Research Triangle Park, N.C.
ii
-------
ABSTRACT
These tests were conducted to evaluate the performance of two mobile
pilot-plant scrubbers for removing H-S and other reduced sulfur compounds from
the offgas of an in-situ shale oil retort in Utah.
A trailer-mounted scrubber system equipped with both a tray tower and
a venturi contactor was used to investigate each of three alkaline solutions;
sodium, potassium and ammonium hydroxide. The objective of this first test of
the alkaline scrubber was to shakedown the equipment, and investigate the
effects of scrubbing chemical, chemical concentration and liquid to gas
contact time on removal efficiency and H2S selectivity.
A skid-mounted Stretford scrubber system was also evaluated using a
scrubbing mixture of sodium carbonate, sodium vanadate, anthraquirione
disulfonic acid and water. A venturi scrubber was used through most of the
test as the sole contactor. Near the end of the test, a field-fabricated
packed tower was added in series downstream of the venturi in order to improve
the removal efficiency. Since this was the fourth test of the Stretford unit,
the test objectives were to obtain and maintain the highest removal efficiency
possible and to attempt to explain some lower removal efficiencies observed
during prior tests.
The retort offgas volumetric percent composition (dry) was approxi-
mately 59 N2, 23 CO2, 9 H2, 5 CO, 2 O2 2 CH4 plus 0.15 (1500 ppmv), H2S and
other reduced sulfur species. The gas was saturated with water and contained
a light mist of condensed water and oil particles.
The alkaline scrubber efficiencies varied directly with the OH~
concentration and gas/liquid contact time reaching 94 percent at the highest
OH~ concentration used in the tray tower and 50 percent at the lowest
concentration in the venturi. Conversely, it was found that the selectivity,
the percent removal of H-S divided by the percent removal of CO2, was highest
at the lowest OH~ concentrations and vice versa. It was found that
selectivity also varied inversely with gas/liquid contact time, the venturi
iii
-------
contactor'providing greater selectivity than the tray tower contactor. The
selectivity varied from a low of 9 to a high of 79. At the lowest OH~
concentration where the venturi produced a selectivity of 79, the tray tower
selectivity was only 22. The test results correlated well with a mathematical
scrubber model based on the penetration theory.
The H0S removal efficiency achieved for the Stretford plant was an
£•
average of 80 percent and a peak of 95 percent with the venturi contactor
alone and an average of 93 percent and a peak of 99.4 percent with the venturi
contactor followed by the packed-tower contactor. "
Neither the alkaline scrubber nor the Stretford removed significant
quantities of the organic sulfur compounds.
iv
-------
ACKNOWLEDGEMENT
The participation in the successful execution of this program by those
listed below is gratefully acknowledged.
EPA/OSEMB
E. R. Bates
J. O. Burckle
Project Planning On-Site Monitoring
Report Review
PEDCO Environmental, Inc. Stretford Plant Development
J. J. Carvitti Plant and S&A Operation
Monsanto Research Corp. Alkaline Scrubber Development and
T. Ctvrtnicek On-Site Operation
Metcalf & Eddy, Inc.
R. J. Swanson
Alkaline Scrubber Operation
Geokinetics, Inc.
J. M. Lekas
E. Costomiris
Pilot Plant Installation and General
Support
Dr. Richard C. Aiken
University of Utah
Penetration Theory Model Development
(as presented in Appendix B)
-------
TABLE OF CONTENTS
Section
FORWARD
ABSTRACT
ACKNOWLEDGEMENT , V
1.0 INTRODUCTION, SITE DESCRIPTION, FINDINGS, AND CONCLUSIONS 1
1.1 Introduction 1
1.2 Site and Process Description 4
1.3 Findings 14
1.4 Conclusions and Recommendations 19
2.0 SCRUBBER PILOT PLANT 22
2.1 Background (Review of H2S Removal Processes) 22
2.2 Description of Facilities 26
2.3 Operations 31
2.4 Theory of H2S/CO2 Selectivity 36
2.5 Data Analysis Techniques 59
2.6 Results 63
2.7 Findings 78
2.8 Two Stage System 79
2.9 Activated Carbon Process 82
3.0 STRETFORD PILOT PLANT 85
3.1 Process Description 85
3.2 Stretford Plant Operations 96
3.3 Analysis of Performance 111
4.0 QUALITY ASSURANCE 1 32
4.1 Gas Sampling 132
4.2 Water Data • 134
REFERENCES 140
APPENDICES
A. Sampling & Analysis Methodology A-1
B. Selective Scrubbing of Hydrogen Sulfide From Carbon B-1
Dioxide in Shale Oil Retort Offgas.
Part I Penetration Theory for Mass Transfer and B-1
Reaction of H2s, CO2 and NH3
Part II Penetration Theory Computer Listing B-29
Part III Computational Studies B-40
VI
-------
LIST OF FIGURES
Figure Page
1 Equipment at the GKI Facility 6
2 An example of a Geokinetics Horizontal In-Situ Retort 7
3 Schematic of Pilot Plant Installation in the Geokinetics Process 8
4 Photographs of the Scrubber Trailer Inlet and Outlet Plumbing 13
5 Removal Efficiency vs. Selectivity for Alkaline Scrubber 16
6 Alkaline Scrubbing Process Schematic In-Situ Retort Offgas 23
7 EPA's Mobile Wet Scrubber Trailer 27
8 EPA Scrubber Trailer Equipment and Flow Diagram 28
9 Sieve Tray Characteristic Curve 29
10 Characteristic Curves for Venturi with 3-cm-dia. Throat 30
11 Single-Reaction-Plane Concentration Profile 41
12 Chemical Enhancement Factor for H2S Single Reaction plane 42
13 Two Reaction Plane Concentration Profile 45
14 Chemical Enhancement of H^S Two Reaction Plane Model 46
15 Chemical Enhancement of CO- Two Reaction Plane Model 47
16 Effect of H2S Partial Pressure on Selectivity 49
17 Effect of COg Concentration on Selectivity 50
18 Variation of H-S and NH3 Removal Efficiency with Venturi Length 56
19 Selectivity as a Function of Residence Time or Venturi Length 57
20 Effect of Initial Hydroxide Concentration on Selectivity 58
21 Removal Efficiency and Selectivity Variation with Venturi Length 60
22 Effect of Temperature on Selectivity 61
23 Removal Efficiency for Tower and Venturi at [OH~] > 0.05 gmoles/ 66
liter
24 Removal Efficiency Results for Tower at [OH~] < 0.05 gmoles/ 67
liter
25 Removal Efficiency for Venturi at [OH~] < 0.05 gmoles/liter 68
26 Removal Efficiency for Tower and Venturi at [OH~] < 0.05 gmoles/ 72
liter
27 Selectivity at [OH~] < 0.05 gmoles/liter 74
28 H2S and CO- Removal Efficiency 76
29 Two-Stage Process 80
30 Tray Tower with Isolated Liquid Inlet 81
31 Overall View of Stretford Plant Installed at GKI 89
32 Simplified Flow Diagram of the Stretford Pilot Plant 91
33 Photographs of the Variable-Throat Venturi used on the 93
Stretford plant
34 Packed Tower Installed at Reaction Vessel Exit 95
35 Conventional Venturi Versus Jet Scrubber Venturi 108
36 Stretford Solution pH vs. Time 127
37 Primary Chemical Concentrations versus Time 128
vii
-------
LIST OF TABLES
Table Page
1 Geokinetics Estimated Retort Off-Gas Composition 10
2 Typical Changes in Off-Gas Ammonia and Hydrogen Sulfide Levels 11
During Burn of a Geokinetics Retort (ppm)
3 Summary of Alkali Scrubbing Results 15
4 Operating Conditions for Alkaline Scrubbing Test 32
5 Test Plan for Alkaline Scrubber 34
6 NH3 Concentration in Gas Stream 63
7 Water Analysis Data 64
8 Removal Efficiency for Scrubber Tests 65
9 (OH~) for Ammonia Tests • 69
10 Test Selectivity for the Alkaline Scrubber 73
11 Comparison of Theoretical and Experimental Select!vities , 77
12 Two-Stage Design Conditions 80
13 Key Design Parameters of the Stretford Pilot Plant 90
14 Geokinetics Project Schedule (1983/84) 98
15 Summary of GKI Accumulated Run Times 99
16 Summary of Primary Chemical Analyses 'I03
17 Electrical Requirement for Stretford Equipment (Amperes) 104
18 Determination of Elemental Sulfur Production 106
19 Stretford Operating Conditions Maintained during GKI Test Program 112
20 Summary of Retort Off-Gas Conditions 119
21 Reduced Sulfur Species Emitted (ppm) 120
22 Summary of Chemical Usage during Stretford Testing 124
23 Chemical Analyses Results 'l25
24 Reduced Sulfur Calibration Data 133
25 Water Analysis Data Quality Control 135
26 Comparison of Sample & Control Water Data Effect on Selectivity 138
viii
-------
SECTION 1.0
INTRODUCTION, SITE DESCRIPTION, FINDINGS, AND CONCLUSIONS
1.1 INTRODUCTION
1.1.1 Background
Removing hydrogen sulfide (H-S) and other reduced sulfur compounds
(carbonyl sulfide, carbon disulfide, mercaptans thiophenes, etc.) from shale
oil retort offgas with a wet scrubber requires a process that will selectively
react with the sulfur compounds and as little as possible with the carbon
dioxide (CO2) which is also present in much larger amounts than the I^S.
Typically the CO2 concentration in retort gas is 20 percent while the t^S
concentration will range from 0.1 percent (1000 ppm) to 4 percent, depending
on the particular retorting process used. This report covers the tests
performed on a direct-fired, in-situ (under the ground), retort for which the
lower H-S concentration applies.
Since both H0S and CO0 are acid gases, it is the objective of any
£l £*
scrubbing system to selectively remove as much H2S and other sulfur compounds
as possible while minimizing the reaction with the accompanying CO2. The
reasons for this selectivity are to conserve the scrubbing chemicals and to
concentrate the sulfur compounds so that they can be economically converted to
a solid recoverable or a safely-disposable form.
Two liquid scrubbing concepts were evaluated on this test, alkaline
and Stretford. The (Lovell et al, 1982 and Desai et al, 1983)* had identified
these as potential processes for removing reduced sulfur compounds from shale
oil retort offgas.
Field test data on retort offgas was limited especially for the
alkaline scrubber. Monsanto Research Corporation (MRC) reported achieving up
to 70 percent H2S removal with aqueous ammonium in tests performed on retorts
at DOE's Laramie Energy Technology Center.
*All references are listed on Pages 140 and 141
-------
The (Desai et al, 1983) primary concern with the alkaline process was
that the selectivity might not be'high enough to allow the removed sulfur to
be recovered with a Glaus process. Selectivity as used in this report is a
measure of the reactivity of the scrubbing solution with H2S compared to CO2
and is precisely defined as the percent removal of H2S from the gas stream
divided by the percent removal of CO2. The alkaline scrubber process
envisioned by the EPA involved: (1 ) removing H2S and the reduced sulfur
compounds with the scrubbing solution; (2) stripping the sulfur gases from the
scrubbing solution along with absorbed CO2; and (3) processing the
concentrated sulfur gases and CO2 stream in the Glaus unit to obtain elemental
sulfur. The Glaus process requires that the reduced sulfur (primarily H2S) be
at least eight percent of the feed gas with 15 percent or greater the
desirable concentration. To obtain this, a minimum selectivity of 10 and
preferably 30 or higher would be required. Because removal efficiency was
known to vary inversely with selectivity, there was a question as to whether
or not the selectivity could be achieved at an acceptable level of removal
efficiency.
Therefore, in 1983, the EPA modified one of their existing trailer-
mounted, wet-scrubber pilot plants to accommodate the potentially-combustible
retort off-gas. Extensive modifications were made to the wiring, controls and
power units in the EPA scrubber trailer and various safety devices were added
to explosion-proof the unit and protect its operators from H2S intrusion.
Pressure, temperature and pH sensors were installed to monitor the process.
Explosive gas and BUS detectors with alarms were installed in the control
room.
This was the first test of the modified pilot plant scrubber. A test
plan was prepared with regard to chemicals to be used for scrubbing, addition
rates, and solution pH levels to determine their effect on selectivity and
removal efficiency. Time was allowed for system-shakedown at the site since
there was no feasible way to completely check out the system before taking it
to the field. The field crew included personnel with technical skills to
rework the system as needed. The Qeokinetics, Inc. (GKI) facility at the test
site was well equipped to support any modification activity with welding,
crane and electrical services.
-------
For the Stratford process this was the fourth test of the EPA's pilot
plant which had been constructed as a transportable, skid-mounted unit. The
first field test of the Stretford pilot plant was made at Occidental Oil
Shale, Inc.'s (OXY) Logan Wash oil shale development mine near De Beque,
Colorado in June and July 1982. This mine site is where OXY has conducted all
of its oil shale research activities toward the development and commercial-
ization of the vertical modified in situ (VMIS) recovery process.
The second field test of the Stretford plant was made in September and
October 1982 at the GKI facility in Utah, the same site as this fourth test.
The third test in November 1982 represented a new application of the
Stretford pilot plant—coal gasification. The test site was the U.S. Bureau
of Mines (BOM)/Twin Cities Research Center (TCRC). The TCRC facility, which
is located in Minneapolis, Minnesota, contains a pilot-scale, low Btu, coal
gasifier.
For the Stretford process, the issue of selectivity per se is
unimportant because the process is inherently selective. Therefore, the
primary concern is removal efficiency. The performance of the Stretford pilot
plant with respect to H2S removal efficiency was improved significantly during
each of these three test programs. At each of the test sites, gas conditions
and composition were similar, which permitted comparisons and performance
trend analysis. Removal efficiencies of I^S improved from a low of 20 percent
at OXY, to 80 percent at GKI, and a maximum of 99+ percent at TCRC. These
incremental improvements in performance were obtained by various systematic
modifications to the pilot plant's process design and operating parameters.
It was desired to reproduce the 99+ percent removal efficiency on oil shale
offgas and to gain some insight as to the cause of the lower efficiencies
obtained in earler tests.
1 .1.2 Objectives
The objectives for the GKI tests were as follows:
1. For the Alkali Scrubber Pilot Plant
. Shakedown the equipment
-------
Investigate the influence of operating parameters pH,
gas/liquid contact time, scrubbing agent, etc. on the
selectivity and removal efficiency associated with
scrubbing reduced sulfur compounds in the presence of
high CC>2 concentration.
2. For the Stretford Pilot Plant
Duplicate on retort offgas the 99+ percent removal^
.efficiency attained in the TCRC coal gasifier tests.
(Upon achieving that), attempt to explain the low removal
efficiency on the 1982 test at GKI by deliberately
introducing upsetting changes to the plant chemistry and
then returning to the 99+ performance.
1.2 SITE AND PROCESS DESCRIPTION
The site of these sulfur scrubbing tests was the in-situ shale oil
pilot test facility of Geokinetics, Inc. (GKI) in eastern Utah, 70 miles south
of Vernal, Utah. This section describes the in-situ retorting process used by
GKI, the properties of the gas emitted by that process and the installation of
the two EPA pilot plants at the GKI site.
1.2.1 Shale Oil Production
Since early 1973, GKI has been developing a shale oil extraction
process designed for areas where oil shale beds are relatively thin and close
to the surface. Deposits with these characteristics have been found in areas
of Brazil, Morocco, Australia, the United States, and elsewhere throughout the
world.
In the southern Uintah Basin in the State of Utah, shallow oil shale
deposits in the Mahogany Zone exceed two billion barrels in place. Major
developers have generally ignored these deposits, and it was here that GKI was
able to acquire its lease holdings, which total 30,000 acres containing oil
shale seams averaging 30 feet in thickness and having an oil content of 22
gallons per ton.
In cooperation with the DOE, GKI is engaged in developing a true in-
situ extraction process for use on shallow oil shale deposits. Because the
process does not require the construction of a mine, surface retort, or
associated rock-moving equipment, the front-end capital cost of a commercial
-------
.operation is greatly reduced. GKI refers to their process as LOFRECO (low
front end cost) covered by U.S. Patent 4037657.
In the GKI horizontal in-situ retorting process, a specific pattern of
blast holes is drilled from the cleared surface through any overburden and
into the oil shale bed. Explosives are placed in these holes and detonated by
use of a carefully timed and planned blast system. The blast yields a well-
fragmented mass of shale with high permeability and also produces a slightly
sloping (approximately 4°) bottom surface that allows the produced, oil to
drain into a sump for collection. The fragmented zone constitutes the in-situ
retort. The void space in the fragmented zone comes from lifting the
overburden, producing a small uplift of the surface as shown in Figure 1(a)
and Figure 2. Submerged-type oil well pumps are used to lift the recovered
oil to surface storage tanks (see Figure 2).
Burning charcoal is introduced into drilled holes at the upper end of
the rubblized zone to ignite the retort. Air inlet piping is also installed
at this end of the retort. The burn front, consisting of a vertical wall
approximately 30-ft high, travels toward the deep or low end of the retort.
The objective is to retort the shale from one end to the other in a plug-flow
fashion by maintaining a burn front that occupies the entire cross section of
the bed. Typically the front travels at a speed of one 'foot per day. At
normal production with two retorts operating, the GKI plant produced
approximately 400 barrels/day of shale oil.
1.2.2 Retort Gas Properties and GKI Gas Processing
The GKI retort off-gas is brought to the surface for processing where
it is treated in four steps, shown schematically in Figure 3 and
photographically in Figure 1, before it is discharged to the atmosphere.
First, the gas passes through a condenser/demister located upstream of the two
blowers. The next treatment steps are the ammonia absorption, sulfur
recovery, and incineration. The latter three operations are performed in
series, with the treatment units arranged so that the desired treatment con-
figuration can be obtained by bypassing one or more process steps. Expected
operations during the scrubber test were to bypass the ammonia absorber and
treat the gas in the sulfur recovery unit and the incinerator. A maximum of
-------
H2S Removal
(Lo-Cat)
Tower
Incinerator
Ammonium
Tower
Ammonium
Tower
Incinerator
(b)
Figure 1. Equipment at the GKI Facility
-------
A-o T
4J
•H
w
I
H
i-l
id
-P
c
o
N
•H
i-l
O
S3
tn
u
•H
4-1
s
•H
8
m
o
0)
X
0}
CN
s
fa
w
in
0)
u
o
tr>
c
•H
5
3
4-1
•H
W
H
(0
•P
O
N
-H
Vl
fi
ID
U
•rl
•P
(1)
•H
,«
O
0)
JS
EH
W
(0
0)
u
£
o
-------
UJ
GC
LU
X
Q.
CO
O
2
DC
O
DC
UJ
z
Jo
—txh
i
5
30
CO UJ
DC
T
(P
N,
i
*l
gl
UJ
DC
r— •*
t
CO
<
0
u.
u.
0
1.
\
I
V OC
>
xo
^ i
I
r ^ !"
^ ^L
@UJ 1
5 1
o i
_J
| CQ
H-fCO
OC <
00
HfU_
UJ Ll-
CC O
Q
DC
O
u_
I- h-
111 Z
CC <
I- _J
CO Q»
Q.
Ill
•X*!
1
I
I
—«—A
1
I
— — J
UJ
1=
-------
10 Sm3/s cfm of gas at a maximum temperature of 82°C can be treated in the gas
processing operation. Typical retort gas composition (dry basis) as provided
by GKI is shown in Tables 1 and 2.
Sampling the retort gas was difficult because it was saturated with
water and contained some oil mist which condensed along with moisture on the
pipe walls. Over a week was spent by sampling and analysis technicians in
developing sampling trains that would not be fouled by the condensing shale
oil and moisture. The gas sampling apparatus is discussed in Appendix A.
1.2.3 Connecting the Scrubbers to the GKI Gas Processing Plant
As originally planned, the slip stream of retort gas for the two EPA
pilot plants was to be extracted from a six-inch sampling valve on the main
by-pass line of the gas processing plant. The outlet gases from the pilot
plants were to be returned to another six-inch valve just downstream of the
inlet valve. With this arrangement the inlet and outlet gas pressure would
be the same and the pressure for circulating the retort gas through the pilot
plants would be provided by their respective blowers.
Because GKI generates their own electricity, they have a limited
capacity. To save power, it was agreed to use the pressure differential
across the GKI blowers to drive the gas through the pilot plants as shown in
Figure 3. GKI reported their blower discharge pressure as +140 g/cm2 psig and
the suction pressure as -280 g/cm2 psig (i.e., 280 g/cm2 psi vacuum). There
was concern as to whether the pilot plants could operate under these
conditions, especially if the internal gas pressure were to drop below
atmospheric. However, by proper throttling at the respective discharge
valves, it was believed that a positive pressure could be maintained upstream
of that discharge valve.
The discharge pressure on the GKI plant varies depending on the
pressure drop in the plant, with the sulfur plant on stream the pressure is
rj
approximately 140 g/cm gage. When the sulfur plant was by-passed, the
discharge pressure dropped to 70 g/cm2 gage or lower. This pressure was still
sufficient to produce the required flow through the pilot plants. However,
the internal gas pressure in the pilot plants dropped below atmospheric. It
was found that the Stretford plant could operate with negative gas pressure
-------
TABLE 1. GEOKINETICS ESTIMATED RETORT OFF-GAS COMPOSITION*
Constituent Mean Gas Analysis, Volume %
Nitrogen 59
Carbon Dioxide 22
Hydrogen 9.3
Carbon Monoxide 5.3
Oxygen 2
Methane 1.44
Ethane 0.26
Ethene 0.16
Propane 0.20
Hydrogen Sulfide 0.15
Ammonia 0.10
Propene 0.10
1-Butene 0.038 .
Butane 0.037
Isobutane 0.014
2-Methylbutane 0.026
1-Petene , 0.015
Trans-Butene-2 0.007
Cis-Butene-2 0.004 ;
1,3-Butadiene 0.003
Iso-Hexane 0.004
Hexane 0.010
Carbonyl Sulfide 0.008
1-Hexene 0.001
Methyl Mercaptan 0.001
Carbon Disulfide <0.001
Thiophene <0.001
*Lekas, 1984
10
-------
TABLE 2. TYPICAL CHANGES IN OFF-GAS AMMONIA AND HYDROGEN SULFIDE LEVELS
DURING BURN OF-A GEOKINETICS RETORT (PPM)*
Mean
Standard Dev.
Ten-Day Means
1-10
11-20
21-30
31-40
41-50
51-60
61-70
71-80
81-90
91-100
101-110
111-120
121-130
131-140
141-150
151-160
161-170
171-180
181-190
191-200
201-210
211-220
221-230
231-240
H2s
1,382
599
61
125
220
479
947
1,506
1,431
1,586
2,048
1,754
1,990
1,734
1,186
1,493
1,801
1,960
1,606
1,852
1,569
1 ,589
1,674
1,181
1,353
2,012
NH3
958
530
16
13
69
444
862
824
1,013
1,061
1,207
595
1,142
914
1,053
7234
699
1,024
1,092
869
961
1,355
1,201
1,963
1,936
1,961
*Personal communication with James Lekas, Geokinetics.
11
-------
downstream of the venturi contactor. The alkali scrubber however could not.
The tray tower does not have a discharge pump and depends on gravity drain-
age. Therefore, a negative pressure in the tray tower defeats the automatic
drain which opens when the liquid reaches a high level point.
To allow the pilot plant to operate, the main blower from the scrubber
trailer was relocated from its position at the gas outlet inside the trailer
to a position outside the trailer where it served as a forced draft fan to
raise the inlet pressure. A four-inch diameter by-pass line and manual valve
was installed across the blower. To achieve a desired flow through the
scrubber it was necessary to manually trim the blower by-pass valve and adjust
the electric flow control valve in the trailer. The system was sensitive to
GKI's discharge pressure changes which at times caused mid-run adjustments and
even several aborted runs.
The scrubber blower was first installed with a four-inch line direct
from the GKI process. As mentioned earlier, the gas entering the scrubber and
the Stretford had significant water and oil mist. The blower soon became
flooded with this condensing liquid and a knockout tank (approximately 50
gallons) was installed, photographs of the final installation are presented
as Figure 4. After a day of operation, the knockout tank filled with oily
water,. A continuous drain was installed in the tank and the system functioned
well enough to complete the runs. The knockouts merely collected liquid
material running along the pipe walls. The suspended mist was carried into
the two processes. The scrubber trailer discharges had an oil slick on the
solution surface. The Stretford system had a foaming problem in their tanks
which may have been caused by the suspended oil as discussed in Section 3.0.
In future tests of the scrubber it would be prudent to send some sam-
pling crews to the site at least a few weeks before the equipment is shipped
to characterize the exhaust gases regarding condensed phases which can clog
sampling lines as well as the entire gas handling system. This will provide
time to fabricate and install the proper knockout devices before the field
test crew arrives on the site.
12
-------
Bypass
Valve
Inlet line from GKI
To Stratford
Stretford Knockout Tank
Scrubber Knockout Tank
Scrubber
Knockout
Tank
(b)
Return Line to
Figure 4. Photographs of the scrubber trailer
inlet and outlet plumbing.
-------
1.3 FINDINGS
1.3.1 Alkaline Scrubber
The alkaline scrubber system was operated using both the tray tower
and the venturi as the gas/liquid contactor. After relocating the main
blower, as described above, the equipment performed satisfactorily.
The alkaline scrubber was operated in a simple blowdown process where
the various alkali solutions were mixed to a specific concentration and fed
either into the tray tower or venturi contactors. In a real process unit, the
scrubber solution would be cycled through a stripper where the absorbed H-S
and CO0 would be removed. Then the solution would be returned to the original
/*
mixing tanks and recycled into the contactor. No significant alkali addition
would be required in that case. Since a stripper was not included as part of
the EPA pilot plant, the scrubbing solution was used on a once-through basis
then discharged to the GKI pond.
The experimental results for the alkaline scrubber are summarized in
Table 3 and Figure 5. The runs were conducted using alternately the tower
then the venturi at the same solution concentration. Three different solution
concentrations were used for each alkali except for the last four runs (No.
31-34) where only the tower was used to make two high concentration runs for
both NaOH and KOH.
It was generally found that the highest selectivity (percent removal
of H2S divided by percent removal of CO2) was obtained at the lowest solution
concentrations and at the shorter solution/gas contact times (i.e., with the
venturi contactor). Conversely, the highest H2S removal efficiencies were
obtained at the higher solution concentrations and the longer contact times
(i.e., with the tray tower contactor). A limit of 94 percent removal
efficiency was reached at an alkali concentration of approximately 0.9 gram
moles/liter where the selectivity is estimated at" approximately ten (analysis
of spent scrubber solution was not performed on that test as indicated in
Table 3). At the low concentration of 0.012 gram mole/liter the selectivity
reached as high as 79.
All three of the alkaline solutions performed similarly. The plot of
removal efficiency vs. selectivity in Figure 5 indicates the specific chemical
14
-------
TABLE 3. SUMMARY OF ALKALI SCRUBBING RESULTS
Contactor
Venturi
Venturi
Venturi
Venturi
Venturi
Venturi
Venturi
Venturi
Venturi
Tower
Tower
Tower
Tower
Tower
Tower
Tower
Tower
Tower
Tower
Tower
Tower
Tower
Alkali
NaOH
KOH
NaOH
KOH
NaOH
KOH
NH4OH
NH4OH
NH4OH
NaOH
KOH
NaOH
KOH
NaOH
KOH
NH4OH
NH4OH
NH.OH
KOH
NaOH
KOH
NaOH
OH~ Cone .
gmole/liter
0.012
0.012
0.023
0.023
0.045
0.046
0.049
0.29
2.0
0.012
0.012
0.023
0.023
0.045
0.046
0.049
0.29
2.0
0.89
1 .25
1.79
2.5
Removal
Efficiency %
52
53
48
48
70
71
60
62
67
52
54
54
59
83
88
64
91
93
94
93
92
94
Measured ,
Selectivity*'
79
71
60
51
(84)t
21
71
56
11
52
43
41
49
36
41
29
29
9
N/A
N/A
N/A
N/A
Run
No.
21
28
24
30
19
26
15
17
13
20
27
22
29
18
25
14
16
12
31
33
32
34
* Selectivity - A measure of the preferential removal of H2S over CO2 taking
into account the relative difference in concentration between the two gases.
In this report, selectivity is the ratio of percent removal of H2S to percent
removal of CO2«
t Data in brackets are suspected to be erroneous.
N/A - Selectivity values for these runs were not available because an analysis
of the spent scrubbing solution was not performed.
15
-------
100
u
-------
at each data point. All three solutions can produce removal efficiencies
above 90 percent at a selectivity 'to be considered a candidate for use with
the Glaus sulfur recovery process. All three show high selectivity at
recovery efficiencies high enough that with the use of multiple venturi
stages, a removal efficiency of over 95 percent should be obtainable for the
system. Since the system envisioned for using these chemicals involves
recycling the alkali, the relative cost of the individual chemicals is
insignificant. What may be significant are factors of corrosion, safety and
availability.
To analyze these data, a computer model of an alkaline scrubber was
developed employing the comprehensive penetration theory (See Appendix B).
Penetration theory (Danckwertz, 1970) treats the gas/liquid mass transfer to
allow contact time to be significant factor. Other models such as the two-
plane theory have implicit assumptions of equilibrium and cannot a.ccount for
the contact time difference between a tower and a venturi. The results
predicted by the penetration theory agree with the experimental results.
Based on the experimental results and the computer model, an alkaline
scrubbing system design concept is suggested which could achieve an H^S
removal efficiency of 95 percent with a selectivity approaching 40. This is a
two stage scrubber with the first stage being a venturi contactor and the
second stage a tray tower. The first stage removes 50 percent of the H2S in a
highly selective manner. The second stage removes 90 percent of the remaining
H-S at a lower selectivity. A summary of these performance values is as
follows:
TWO-STAGE ALKALINE SCRUBBER - CONCEPT I
Stage No
Contactor
Selectivity
Removal Efficiency
I
Venturi
110
50 percent
II
Tray Tower
40
90 percent
Combined
37
95 percent
17
-------
Another concept employing' a two-stage tray tower scrubber which
results in a higher removal efficiency but a lower selectivity is summarized
as follows:
TWO-STAGE ALKALINE TOWER SCRUBBER - CONCEPT II
Stage No
Contactor
Selectivity
Removal Efficiency
I
Tray Tower
40
90 percent
II
Tray Tower
40
90 percent
Combined
22
99 percent
This "two-stage" tray tower scrubber can be combined into a single
tower of double length.
The alkaline scrubber showed little removal of the organic sulfur
compounds. This is similar to previous results reported in the literature.
In reviewing the literature it was found that a commercial alkaline scrubber-
process exists, which has been successfully employed on the exhaust gases from
black liquor boilers in the pulp and paper industry, to remove organic sulfur
compounds as well as HpS. The primary difference is that the scrubbing
solution contains activated charcoal and a hypochlorite compound in addition
to the NaOH. The small amount of activated charcoal (less than 0.1 weight
percent) also aids in oxidizing HS~ to S-O^ and produces a saleable by-prodxict
of sodium thiosulfate. Prohocs, 1983 present the details of this system.
- , 0 stretford Plant
I . J . £. ————-———^-^—
The Stretford operated for over 200 hours. For 140 hours, the plant
operated with a venturi contactor. The venturi had been modified from that
used in previous tests in that the throat area could be adjusted to handle
variable gas flow rates. In this test the throat was adjusted to the smallest
throat area, 18 cm, and held constant during most of the testing.
18
-------
The maximum H-S removal efficiency measured while using the venturi
alone was 95 percent which was maintained only briefly. Over the period of
operation with this contactor alone the removal efficiency averaged 80
percent. The short increase to 95 percent was not explained.
A brief attempt was made to experiment with increasing the venturi
throat area. When no effect on removal efficiency was observed, the throat
area experiment was discontinued.
Because of the failure of the plant to achieve the 99+ percent removal
efficiency objective observed in the TCRC coal gasifier tests, the;plant was
equipped with a field-fabricated, packed-column contactor placed in series
with and downstream of the venturi. This device increased the removal
efficiency to the 99+ percent range during its period of operation. Because
of the make-shift nature of this field modification, there was no instrumenta-
tion to measure the flow rate of the scrubber liquid through the tower. Thus,
it was not possible to optimize the liquid distribution between the venturi
and the tower.
1.4 CONCLUSIONS AND RECOMMENDATIONS
1.4.1 Conclusions
Based on the findings reported herein, the following conclusions were
reached:
1. For shale oil retort offgas similar in composition to that
from the GKI process, the alkaline scrubber, in combination
with a stripper and a Claus plant, could be a viable means
of H-S removal. This overall conclusion is based on other
conclusions as enumerated below.
2. For GKI-type process offgas and based on these tests, the
performance of an alkaline scrubber with a tray tower
contactor similar to that in the EPA pilot plant can achieve
an H-S removal efficiency of at least 90 percent with a
selectivity of approximately 30. Under the same conditions
a single venturi contactor in place of the tray tower would
remove only 50 to 60 percent fi^S but with a selectivity of
70 to 80.
19
-------
Based on the computer model developed to analyze these test
results, the removal-efficiencies and selectivity above are
applicable to offgas with lower H_s concentrations than
found at GKI. This suggests a concept of multiple scrubbing
actions to increase the H_s removal. Because this increased
removal efficiency is accompanied by a reduced selectivity
which could present a problem for the Glaus plant, the cost
effectiveness of this concept requires a design study.
Based on a three gas component (H2S, NH3 and CO2) analysis
by the computer program, the principal reactant for the H2S
in the retort offgas is the NH., in that same offgas. In
that NHo is present in the GKI offgas in similar molar
quantities to that of the H2s, the scrubber performance
observed on these tests may not be applicable to retort
offgas with little or no NH3. This also suggests that the
water and the NH3 in the offgas would be an effective
scrubbing agent without any alkali addition to the water.
Scrubbing in this manner would certainly improve the
selectivity but the H2s removal efficiency obtainable is
uncertain.
The alkaline scrubber removal efficiency and selectivity
seemed to have little dependency on the alkali used. This
is consistent with the above concept that it is the NH3 in
the offgas itself that is reacting the H2S. Since the NH3
and H2S concentrations are variable, it is likely that some
of the H2S is reacted by the alkali. Therefore, it is
likely some alkali will always be needed. However, the
choice of scrubbing alkali may be made on such factors as
cost, maintenance, safety, availability, crew comfort, etc.
rather than performance.
The absorption of H2s and CO2 in the alkaline solution
appears to be fully reversible by distillation. The 'sulfur
in the scrubber solution is primarily in the form of
sulfide. The sulfate or sulfite level determined in the
scrubbing solution was equal to that in the water supply.
The sulfide will distill off as H2S (along with CO^) while
the sulfate will not. It had been suggested (Desai et al,
1983) that the H2S would not be recoverable from the
alkaline solution (presumably because it would be oxidized
by the Oo in t*16 offgas). This does not seem to be the case
based on this test.
With an adequate contactor, the Stretford process can obtain
removal efficiencies of 99 percent. These tests suggest
that if adequate H2s removal cannot be achieved with a
venturi, then a packed tower is a workable option for
improving performance.
20
-------
8. To insure continued satisfactory performance of a Stretford
plant in processing retort offgas, it is important to
provide effective removal of hydrocarbon mist and other
particulate matter from the gas before it enters the plant.
1.4.2 Recommendations
The following recommendations are made regarding continued investiga-
tion of reduced sulfur compound removal from shale oil retort offgas:
A. It is recommended that a preliminary design study be conducted
to determine the effect of removal efficiency and selectivity on the
design of a sulfur removal system based on an alkaline scrubber and a
Glaus plant. The objective of this study would be to provide cost
tradeoff data necessary to optimize a sulfur removal plant for any
future installations.
B. To continue the research and development of the alkaline
scrubber, it is recommended that the EPA mobile scrubber pilot plant be
deployed for a further series of tests. The objective of these tests
would be as follows:
1. Explore the effect of OH~ concentration on removal
efficiency and selectivity.These tests would cover the
concentration range from 0.0 to 0.012 gram moles/liter and
from 0.05 to 1.0 gram moles/liter using both the venturi and
tray tower.
2. Investigate the combined venturi and tray tower concept
postulated in this report to see if the 95 percent removal
efficiency and 37 selectivity is achievable. The operating
parameters for this test would be selected after the field
results from the concentration tests (above) are known.
3. Investigate the use of hypochlorite solution and charcoal in
the NaOH scrubbing solution to improve the organic sulfur
removal. Since neither the alkaline scrubber nor the
Stretford plant was effective in removing organic sulfur
compounds from the offgas, this test will determine whether
or not this process will be as effective on shale oil retort
offgas as it has been in the paper industry.
21
-------
SECTION 2.0
SCRUBBER PILOT PLANT
This section describes the facilities, theory of operation and results
of the scrubber pilot plant tests. It concludes with a concept design for a
potentially viable alkaline scrubber for H2S in a high-CO2-concentration gas,,
2.1 BACKGROUND (REVIEW OF H2S REMOVAL PROCESSES)
Under EPA sponsorship, two studies of various H2S removal processes as
were conducted (Lovell, et al 1982 and Desai, et al 1983). These processes
were evaluated with regard to removal efficiency, waste disposal requirements,
safety requirements, overall treatment costs, state of development, licensing
requirements and compatibility with EPA's concept of mobile pilot plant
scrubbers. The six processes that were deemed to have applications to shale
oil retort offgas were:
Lo-Cat (TM)
NaOH Scrubbing
Amine Scrubbing
Aqueous Ammonia Scrubbing
Stretford
Unisulf
The Stretford process will be discussed in Section 3.0. This
discussion is concerned only with the caustic and aqueous ammonia scrubbing.
In this report, caustic has been extended to include KOH as well as NaOH.
NH.OH, NaOH and KOH are referred to collectively in this report as alkali or
4«
alkaline material.
The caustic or ammonia scrubbing process (i.e., alkaline scrubbing
process) consists of: (1) a scrubber to selectively remove H2S from the
retort gas, (2) a regenerator (distillation unit) to release the absorbed H2S
gas as well as the co-absorbed CO2 gas and permit the recycling of the
scrubbing liquid, and (3) a Claus plant to recover sulfur from the H2S rich
gas. The process schematic is shown in Figure 6.
22
-------
M-l
rH
CO
co
'.10
Cn
w
CO
(d
en
tn
C
•H
co
co
d)
U
O
U!
id
O
U) -P
D C
rH 1-1
U ft
CN
8
CO
I
0) Vl
c u
id D
t^j ti
t-H U
< CO
I
S-l
0)
c
Q) H
CP O
0) JJ
K fd
In-Situ Retort
Offgas
0)
•H U
3 >-
tJ1 U
•H 0)
^J (N
•H O
S 0
CO
O
CO
-P
!H
O
-P
Q)
'
-P
-H
CO
I
G
H
O
•H
4->
1
a;
u
to
: co
co
0)
: o
o
M
U
CO
a)
c
•H
23
-------
The alkaline scrubbing process removes H2S from the retort gas by
absorption with chemical reaction. The H2S is removed from the gas by
reaction with OH~ to form bisulfide, HS~. When the HS~ solution is heated,
the H0S is regenerated in concentrations high enough for sulfur recovery in a
£
Glaus process. Any absorbed ammonia will be removed from the scrubber liquid
and can be recovered in the regeneration process. The Glaus process is most
effective with greater than 15 percent H2S concentration of the feed gas
although the minimum acceptable concentration is 8 percent. As the retort
off gas contains CO., in great excess of H2S (150:1 for the in-situ retort), the
CO- absorption rate can be high and can be the primary limitation to the
^
process.
There are two primary performance considerations in the scrubber
design, removal efficiency and selectivity. The removal efficiency desired is
at least 95 percent based on the expected allowable sulfur emission rates for
regulating future oil shale processing. The selectivity is critical because
of the need to obtain an acceptable concentration of H2S in the Glaus feed
gas. The removal efficiency, is the overall percent reduction in reduced
sulfur while the selectivity is the relative preference given to absorption of
H2S over that of CO2 considering the great difference in their concentration
level., For this report selectivity is defined as:
% H S absorbed
S = 2
% CO absorbed
The required selectivity for the scrubber is determined by the equation:
C x R
where: S = selectivity = % H2S absorbed/% CO2 absorbed
C = Glaus feed gas ratio of H2S/CO2
R = Retort offgas ratio, CO2/H2S
For a Glaus feed gas ratio of (H2S/CO2) 0.08 and retort offgas CO2/H2S
ratio of 150, the selectivity required is
24
-------
S = 0.08 x 150 = 12
The minimum criteria is to achieve an H2S/CO2 ratio of 0.08. However,the
higher the selectivity, the less CO- absorbed and the lower the steam
requirement to re-vaporize in the regeneration stage. To achieve the moderate
H2S/CO2 ratio of 0.25 requires a selectivity of
S = 0.25 x 150 = 38
Lovell, et al, 1982 has reported a selectivity of 29 as a maximum for
the Japanese Diamox process (essentially an ammonia scrubbing process) and did
not select alkaline scrubbing on the basis that the maximum selectivity of
approximately 30 is insufficient for a cost effective system.
One of the primary objectives of this project was to examine the
potential for achieving higher selectivities by maximizing the effect of the
different reaction rates for H2S and CO2 absorption. Essentially, the H2S
absorption reaction is instantaneous while the CO2 absorption rate is finite
(6000 liters/gmole-sec). This suggests that limiting the reaction time and
controlling the relative gas/liquid mass transfer coefficient should result in
higher selectivities. Consequently, these tests were run with a tray tower at
a residence time of 0.2 sec and a venturi with a residence time of 0.003 sec.
While it is desirable to have both a high selectivity and a high
removal efficiency, the literature shows that these two parameters usually
change in opposite directions. Low solution concentration and short
gas/liquid contact time increase selectivity but lower removal efficiency.
The optimum scrubber design requires a tradeoff of these parameters.
To provide a means to analyze the experimental data obtained in this
program and to assist designers in optimizing alkaline scrubber performance an
analytical computer program was developed. The penetration theory
(Danckwertz, 1970) was used for modeling mass transfer in this program and the
gas/liquid contactors assumed were Venturis since these are discrete short
interval contactors and have the most controllable operating parameters.
25
-------
2.2 DESCRIPTION OF FACILITIES
The equipment used for the alkaline scrubbing test was contained in
the EPA's Mobile Wet Scrubber Pilot Plant shown in Figure 7. A schematic
diagram of the equipment inside the scrubber trailer is shown in Figure 8.
The gas treatment equipment consists of a spray tower, venturi/eyelone com-
bination, sieve tray tower, and a demister. The system can be operated in
series with none, one or more treatment units excluded from operation.
Peripheral equipment consists of a Roots blower, a sump tank, fabric filter/
holding tank combination, a pump/mix tank, an air cooler, feed and recycle
pumps, liquid control valves, gas temperature and flow/monitoring devices, gas
pressure monitoring devices, and liquid flow and pH and monitoring
instruments.
In these tests only the sieve tray tower or venturi/eyelone unit was
used for gas treatment. The spray tower was not included in the gas train.
As discussed in Section 3^3 the Roots blower was relocated upstream of the
trailer inlet for the GKI tests to boost the inlet pressure.
2.2.1 Sieve Tray Tower
The sieve tray tower consists of four trays within an 46-cm-dia. pyrex
glass column. Three sets of trays are available for this tower with varying
hole diameter and spacing. The open area is the same for all trays. The
sieve tray perforation size used for these tests was 0.32 cm. The sieve tray
characteristic curves are shown in Figure 9.
2.2.2 Venturi/Cyclone
The Venturi scrubber consists of three interchangeable venturi throat
sections (3.5, 6.0, 8.5 cm dia) which allow operation over a wide range of
pressure drops and liquid-to-gas (1/g) ratios. Each venturi throat has a
length of 30.5 cm and two radial inlet water nozzles 5.1 cm below the throat
entrance. After leaving the venturi the scrubbed"gas enters the cylone
separator. The venturi throat used for these tests was the 3.5 centimeter
diameter and the characteristic curve for the venturi scrubber is shown in
Figure 10. The reader is referred to the "EPA Scrubber Trailer Operation
Procedure," (Ctvrtnicek, 1984) for additional information regarding the
details of the scrubber trailer and specific equipment contained within.
26 ;
-------
Figure 7. EPA's mobile wet scrubber trailer
27
-------
co
CTl
r-l
Q)
u
•H
c
-p
VI
I;
u
M
a>
it!
1
M
Cn
to
•H
B
•a
4-)
c
a)
I
M
a)
t-i
•H
1
H
U
M
03
0)
M
g.
•H
28
-------
100
u
c 10
i i i i i i i i i i i I i i i i I r
500 scfm'
450 scrm*
400 scfm— -
300 scfm— —
200 scfm——
I I I I i 1 I l I I I I
I i l l l
10
100
JL/G, gpm/103 scfra
Figure 9. Sieve tray characteristic curve (Ctvrtnicek, 1984)
29
-------
100 r
u
10
1 L-
250 scfm
200 scfm
I till
10
100
L/G, gpm/103 scfm
Figure 10. Characteristic curves for venturi with 3-cm-dia. throat.
(Ctvrtnicek, 1984)
30
-------
2.2.3 Retort Gas
The scrubber trailer was operated on a 0.094 Sm /S ACFM slipstream of
retort off-gas. The slipstream was removed from the discharge side and
returned to the suction side of the GKI blower as discussed in Section 1.2.3.
2.2.4 Alkaline Chemicals
The alkaline chemicals used for these tests were sodium hydroxide
(NaOH), potassium hydroxide (KOH), and ammonium hydroxide (NH4OH). Concen-
trated solutions in 0.21m -gal drums were used and supplied to the chemical
mix tank by means of a drum pump.
The duration of the tests was limited by the capacity of the chemical
mix tctnk. The addition of a precision metering pump and flow controls to
provide for continuous concentrated alkali feed would allow for continuous
operation for future tests.
2.3 OPERATIONS
2.3.1 Schedule of Activities
Testing operations were initiated on May 5, 1984. These early
operations involved equipment shakedown, sampling system development and
interface problems with the GKI retort process. Consequently, tests from
May 5 to May 8 did not yield quantitative data. The test runs reported were
performed over a three day period, May 9 through 11.
2.3.2 Operating Conditions
The operating conditions for the scrubber tests are shown in
Table 4. The inlet pressure to the scrubber trailer averaged 84 x 10 Pascal
(12.2 psia) and the inlet gas temperature averaged 54°C. For reference the
atmospheric pressure at Kamp Kerogen during the tests averaged 79 x 10 Pa
(11.4 psia). The H-S concentration of the inlet gas was fairly constant for
~ • >
most of the runs at 1,280 ppm. However, for runs 12-17 the H2S was
considerably higher with a level of 1780 ppm. The gas flow to the trailer
31
-------
EH
H
EH
Oi
H
CQ
ai
U
CO
H
Z
&,
g
CO
z
o
H
EH
M
Q
z
O
CJ
EH
^
W
0
^
K
%
EH
O
r— 1
u
0)
)H U
o> a
4J -P
(C 1C
0)
CXi
&
&
fa
o
oT
1 3
•P
a
CD
1
ro
O
*~
X
to
o
to to
n) a
O ft
«T
•s
to
to
CO
tH
1
Ifi
4->
g
H
CO
o
w
1
CO
S3
•P
C
M
Qj
<
4-
C
C
H
c
J.
ontac-
o
z
en o en ro en o
ro ro ro ro ro CN
o o o o o o
VD vD vD VD VD VD
^* ro *sr ro ^* ro
en oo vo ^* en oo
CM CN r- <- CN CN
M fN
(^ ro ^* LO vO f^
en m
^— m
ro ro
O O
0 0
ro ro
0 0
ro CN
CN CN
en O
o «-
0 O
0 0
00 CO
CN CN
0 0
m m
CN ro
ro CN
t- CN
in in
r- -a-
CM in
vo m
en oo
f- r-
ro en
CN ro
oo oo
•H
M M
IS
SI
oo en
en co
t~- tN
ro ro
O O
VD 0
•a1 m
ro ro
0 0
CN t-
CN CN
0 O
0 0
0 0
CO CO
CN CM
0 0
VD in
CM ro
CM t-
CN in
in in
o in
co in
•** in
r- oo
•d1 o
C3 ^5*
00 00
•H
ll tj
Towe
Ventu
o »-
-<
a n> s
V & -P
c o c
$*$
00 Ol O
CN CN ro
01 Ol
r^ r—
ro ro
0 0
ro en
CN CN
CN CN
^ en
0 0
o o
0 0
00 00
CN CN
in r-
vo in
ro ro
ro *-
r— CN
in in
00 O
*r ^
oo CN
O 0
co co
VO CM
in •*
00 00
H M
0) 0)
g g
EH EH
T- CN
ro ro
01 01
ro ro
0 0
r~ en
CN i-
CN CN
en 01
en o
O t-
0 0
0 0
CO 00
CN CM
<- o
co o
ro ^*
^r CN
in t-
m in
en ro
ro ro
<* ^ •
01 r-
r-- oo
ro t^
ro *tf
co CO
U ^
II,
,
ro '•S4
ro ro
"t
CM
CN
00
01
o
0
o
in
ro
co
I
32
-------
averaged 1.00 Sm^/s and was fairly constant during the runs. The inlet makeup
water temperature averaged 22°C and showed little variation.
The liquid flows were maintained at 0.38 1/s for the sieve tray tower
tests and approximately 0.32 1/s for the venturi tests. There was, difficulty
in controlling the liquid flow to the venturi and consequently the flow rates
show considerable variation ranging from a low of 0.29 1/s to a maximum of
0.35 1/s.
2.3.3 Scrubber Operating Problems
A. Inlet Gas Pressure—
As discussed in Section 1.2.3, the inlet gas pressure from the GKI
facility was insufficient to operate the scrubber train. The system requires
a positive internal pressure. Therefore, it was necessary to relocate the
Roots blower to the gas inlet to boost the pressure. A knock-out drum was
also constructed and installed to prevent solids (stones in pipe line) and
excessive slugs of water from entering the blower.
B. Liquid Flows—
Control of the liquid flow to the venturi at flow rates greater than
0.32 1/s was erratic and, therefore, it was decided to maintain a maximum flow
rate of 0.32 1/s.
C. Gas Leak—
A gas leak developed at the flange on the inlet valve requiring
shutdown and repair.
D. Liquid Level Control—
The liquid discharge from the tower sump was controlled by a high/low
liquid level.controller activating the drain valve. Initially the range was
too small resulting in continuous on-off operation. Increasing the high/low
range resolved the problem.
2.3.4 Test Plan
The test plan is shown in Table 5. The primary objectives of the test
plan were to:
33
-------
TABLE 5. TEST PLAN FOR ALKALINE SCRUBBER
Run No.
12
13
14
15
16
17
18
19
20
21
22
24
25
26
27
28
29
30
31
32
33
34
Alkali
Ammonia
ti
"
it
ii
ii
NaOH
II
11
II
* II
II
KOH
11
11
II
II
11
"
II
NaOH
II
Cone
gmoles/liter
2.0
II
0.05
II
0.3
II
0.05
It
0.012
II
0.023
II
0.05
II
0.012
II
0.023
II
0.9
1.8
1.25
2.5
Contactor
tower
venturi
tower
venturi
tower
venturi
tower
venturi
tower
venturi
tower
venturi
tower
venturi
tower
venturi
tower
venturi
tower
tower
tower
tower
34
-------
1. Evaluate the comparative scrubbing efficiency of ammonium
hydroxide, sodium hydroxide, and potassium hydroxide.
2. Evaluate the effect of concentration on H2S/CO2 selectivity for
these alkaline scrubbing chemicals. :
3. Evaluate the performance of the tower and venturi with regard to
both removal efficiency and H2S/CO2 selectivity.
The test plan was not designed to evaluate variations in the liquid to
gas ratio or other variations in the tower or venturi characteristics (i.e.,
variation in venturi throat diameter, variation in sieve tray perforations).
2.3.5 Operating.Procedure
Due to the problems with the GKI plant and the gas analysis, it was
necessary to compress the individual test periods into a relatively short
time. Eighteen tests were required in a 60 hour test period allowing only two
hours per run. The following procedure proved to be effective in meeting this
brisk schedule.
The mix tank was prepared with the proper solution strength in accor-
dance with the test plan shown in Table 5. Gas flow was maintained during the
down periods. Therefore, it was only necessary to initiate liquid flow to the
contactor to begin the run.
There were two essential timing factors to consider. The first factor
concerned the capacity of the mix tank. As the plant was running on a once-
through basis with dilute alkaline make-up at a rate of 0.38 1/s gpm, there
was only a 40 minute maximum run time.
The second aspect of the timing requirements was due to the gas
sampling procedure. The gas analysis system was evaluating total reduced
sulfur (TRS) alternating every ten minutes from the scrubber outlet to the
Stretford outlet. Therefore, it was necessary to start the test run precisely
at the beginning of the Stretford measurement cycle. This would give TRS
readings for the scrubber outlet at 10-20 minutes and 30-40 minutes into the
35
-------
cycle. The 10-20 minute readings typically did not indicate the steady state
condition which meant that the 30-40 minute readings were critical.
As it was necessary to refill the make-up tank after each run, the
timing to prepare for the next run was also essential. This procedure proved
to be effective once the timing sequence was fine-tuned.
2.4 THEORY OF H2S/CO2 SELECTIVITY
The following discussion is a review of the past research an the use
of alkaline chemicals to scrub H2S in the presence of large CO2 concentra-
tions,, This discussion is presented at this time in order to provide the
reader with a background to evaluate the test results. A reader familiar with
the theory of selective absorption may go directly to Section 2.5 for a
discussion of the test results. He may also care to refer to Appendix B for
the mathematical derivation of the scrubber model.
To briefly summarize, the absorption of H2S by alkaline scrubbing must
occur with adequate removal efficiency while limiting the amount of CO2
absorption. The key factors that affect the relative absorption rate of H2S
over CO0 are liquid alkaline concentration, contact time and presence of NH^
£
in the gas. The theoretical analysis for the H2S selective absortion provides
both a basis for data correlation and a predictive model for evaluation of
this scrubbing process at varying conditions.
2.4.1 Mass Transfer Rate
Absorption of a species from a gas to a liquid occurs by mass transfer
first through a gas film to an interface and then through the liquid film to
the bulk liquid (Danckwertz, 1970).
The absorption rate is determined by the equation:
g ^ g g 1 *• 1 1
where N = mass transfer rate, gmoles/hr
K = gas side coefficient, gmoles/hr - m
" -y
K, = liquid film coefficient, moles/hr - irr
C = concentration, gmoles/liter
36
-------
subscript (1) = liquid
subscript (g) = gas
superscript (i) = interface
superscript (o) = bulk fluid
g = chemical reaction enhancement factor, unitless
2
A = surface area for mass transfer, m ;
The gas film coefficient is determined by the physical characteristics
of the system (type of contactor, flow rates, physical properties, etc.) and
is not affected by the chemical type or concentration of the scrubbing
solution.
The rate of mass transfer of any chemical species in the liquid film
is a product of the concentration difference, the mass transfer coefficient
and the chemical enhancement factor for that species (which is usually
expressed as a multiplier of the liquid film coefficient).
A key element in optimizing selectivity is the relative importance of
the gas and liquid film coefficients. The liquid phase reaction of H2S is
instantaneous while the CO2 absorption reaction is finite. Therefore, the
absorption of H2S is limited by the gas film resistance while the CO2
absorption is liquid film controlled. Consequently, maximizing the gas film
coefficient while minimizing the liquid film coefficient can significantly
increase selectivity. This discussion is continued in Section 2.4.3.
The gas bulk concentration of the species is determined by the process
conditions. The H2S concentration is approximately 0.15 percent and the CCX,
is approximately 23 percent. The liquid interface concentration is determined
from the gas concentration by the solubility and volatility of the species.
These properties are both affected by temperature, ionic strength •
(concentration of ionic species) and other dissolved components. The chemical
enhancement is determined by the chemical type and composition of the scrubber
liquid.
A. Gas Concentration—
As one of the purposes of this program is to evaluate the'selectivity
of the alkali scrubbing solution for H2S over CO2 it is useful to look at the
37 ;
-------
initial parameters . The rate of absorption of a species is directly related
to its concentration. Therefore, the ratio of concentrations for H2S and CO2
indicates the nature of the problem. For the typical gas at OKI with 1500 ppm
H2S (0.15%) and 23% CC>2, the relative absorption rate or concentration ratio
CO2/H2O is 23/0.15 = 150. This indicates that disregarding selectivity the
absorption rate of H2S will be less than one hundred and fiftieth that of CO2.
B. Solubility—
The solubility of the species in the liquid determines the interface
concentration.
There is a natural selectivity of this system for H2S based on the
relative solubility of CO2 and H2S. Essentially, the higher solubility of H2S
makes it easier to absorb than CO2 and, therefore, increases the
selectivity. This physical selectivity, 6, is defined as
solubility CO«x gaseous concentration CO
e _
~
solubility H S x gaseous concentration H S
£. £
At 25°C the solubility in water of H2S = 1.8 x 10~3 mole fraction and
CO2 = 0.6 x 10~3 mole fraction. Therefore the physical selectivity is
6 = (0.6 x 23%) CO /(1.8 x 0.15%) H S = 50
£t £
The physical selectivity (i.e., due to solubility) predicts a three-
fold increase in absorption of H2S over that of CO2 based only on the
concentration conditions and results in a decrease from 1 50: 1 to 50: 1 for the
CO2/H,,S absorption ratio.
C. Chemical Enhancement —
When the gas species being absorbed reacts with the scrubbing solu-
tion, the absortion rate is increased due to the elimination of the species .
The chemical enhancement factor is determined by: (1 ) the rate of reaction,
*If it were not for the selectivity of H2S over CO2 it would be impossible to
use alkali solution to remove H2S from the GKI retort gas.
38
-------
(2) the concentration of the species, and (3) the diffusivity (ease at which
the species dissolved in the gas moves through liquid) of the species. The
chemical enhancement factor 3, is defined as:
0 = Kxa/ Kxa ;
where K = actual absorption coefficient
K, = absorption coefficient without reaction
xa
The relative chemical enhancement, o, is defined as the ratio of
chemical enhancement for each species, i.e.
0=3 (H2s)/3 (C02)
Note: This term is sometimes referred to in the literature as selectivity.
The terminology, relative enhancement, is used in this report to distinguish
this item from the selectivity used in this report as defined in Section 2.1
(ratio of removal efficiencies H-S to CO-).
It is the relative liquid phase reaction rates and reaction mechanisms
that account for the highly selective H2S absorption required for the alkaline
scrubbing process to be economically feasible. This chemistry is presented
next.
2.4.2 Chemistry
A. General Kinetics—
For the absorption reaction of H_S in alkali solution with no CO-
present the initial reaction equation is:
H2S + OH~ -»• HS~ + H2O (5)
and the reaction rate defined as:
rH S ~ ~kH S [OH~] tH2S] (6)
39
-------
where r = reaction rate
k = rate constant
[OH~] = hydroxyl ion concentration - bulk
[H2S] = EUS concentration - interface
The system is also characterized by the chemical equilibrium constant
^ = - [HSJ - (?)
[H2S] [OH ]
Note: The second dissociation to S~2 (HS~ + OH~ -»• S~2 + H2O) is relatively
small and can be neglected.
The concentration profile for this system is presented in Figure 11
which shows the reactant concentration variation in the liquid film.
This model assumes a single reaction plane where the reaction of H2S
with OH~ takes place.
The above rate and equilibrium equations can be combined to evaluate
the relative rate of chemical absorption to that of physical absorption, i.e.
chemical enhancement. The reader is referred to Astarita, 1964 and
Danckwertz, 1970 for a complete description of this derivation. The resulting
equation for chemical enhancement is:
wheret [OH~] = hydroxyl ion concentraton in the bulk liquid, gmoles/liter
[H-S] = H~S concentration at the interface, gmoles/liter
This approach is based on the single-reaction-plane concentration
profile as shown in Figure 1 1 . Experimental results reported in the
literature for the alkaline scrubbing of CO- and H-S are shown in Figure 12.
The one-reaction-plane model represents the lower boundary for the data, i.e.
predicting chemical enhancement lower than realized in the experimental
investigation (Astarita, 1967).
40
-------
•H
M-l
O
C
.3
c
-------
(U
rtJ
rH
ft
c
o
•H
-P
u
ft)
(U
c
•H
to
co
f
s
i-l 1^5
o en
C LJ
a> Id
15 4J
0) w
c5
(3
c
m
u
•H
U
(N
r-i
0)
^
cn
•H
42
-------
B. Two Reaction Plane — '
The above analysis assumed no interrelationship between the H2S and
other gases present. In a scrubber involving the simultaneous absorption of
H0S and CO, in aqueous hydroxide the prediction of chemical behavior is con-
£ £
sidered complicated by the interaction of the various reactants.
The complete chemistry for the H2S-C02 absorption in alkaline solution
is as follows:
CO2 + OH~ -»• HCO3~
H2S + OH~ -»• HS~ + H20 (10)
HCO3~ + OH~ •»• CO3~2 + H2O (11)
HS~ + OH" -»• S~2 + H2O (12)
H2S + CO3~2 -»• HS~ + HCO3~ (13)
HS~ + CO3~2 -»• S~2 + HCO3~ (14)
To simplify the analysis Reactions 12 and 14 can be neglected as the
equilibrium values for S~2 are very small.
Reactions 9, 10, 11 and 12 can be considered instantaneous, regardless
of reactant concentrations when compared with the diffusional process.
Reaction 9 is only instantaneous at OH~ concentrations greater than 0.01
gmole/liter (where Reaction 9 is followed immediately by Reaction 1 1 ). When
the OH~ concentration is low, i.e., when HCO3~ and CO3~2 coexist, Reaction 9
is too slow to affect the absorption rate. Therefore, Reactions 9 and 11 can
be combined and the remaining reactions to be considered are:
C02 + 20H" ->• C03~2 + H20 (15)
H2S + OH~ •*• HS~ + H20 (10)
43
-------
H2S + C03~2 ->- HS~ + H03~ (13)
HC03~ + OH~ -»• CO3~2 + H2O (11)
These reactions can all be considered as instantaneous and irrevers-
ible. Therefore, none of the couples of reactants involved may coexist in
appreciable concentration levels at any point of the liquid.
The concentration profiles resulting from these reactions are shown
in Figure 13. The primary reaction plane is the reaction interface for
Reaction 9, the reaction of CO2 with OH~. Between the primary reaction plane
and the interface the concentration of OH~ must approach zero.
The CO3~2 ions are formed at the primary reaction plane. But since
CO,~2 can not coexist with H2S, there can be no H2S in the vicinity of the
primary reaction plane. Therefore, a secondary reaction plane located between
the interface and the primary reaction plane exists where the reaction of H2S
—2 -
with CO3 takes place.
The CO2 is physically absorbed and diffuses from the surface to the
_ o
primary reaction plane where it reacts with OH . The CO3 ions formed
diffuse toward the bulk of the liquid and towards the interface. The H2S,
physically adsorbed, diffuses to the secondary reaction plane where it reacts
with the CO3~2 to form HS~.
The CO3~2 and HCO3~ ions loop back and forth in Zone II and Reaction
10 never actually takes place. However, the net results of Reaction 13 (which
takes place at the secondary reaction plane) and Reaction 9 (which takes place
at the primary reaction zone) is Reaction 10.
The reader is referred to Astarita, 1965 for a detailed description of
the equations developed to calculate the plane depth and relative chemical
enchancement factor.
Application of the penetration theory to the two reaction plane model
has been investigated (Onda, 1972). Figures 14 and 15 show the chemical
enhancement factors for H2S and CO2 as a function of OH~ concentration. The
H2S delta shows good correlation with all three models, but a significantly
better data fit with the unsteady-state penetration theory than the two film-
44
-------
c
o
Q)
D
5
o-
x2^-
Distance
Figure 13. Two reaction plane concentration profile
45
-------
40
PS
O
fa
CO.
EH
U
U
U
30
20
10
Penetration Theory
Two Reaction Plane
Single Reaction Plane
Model
Two Film Model used;for
Data Analysis
Experimental Results
Reported in Reference
0.4 0.8
OH~ CONCENTRATION, gmoles/liter
1.2
Figure 14. Chemical Enhancement of H2S Two Reaction Plane Model
(Onda, 1972)
46
-------
40
30
CM
8 20
10
Figure 15.
experimental results
reported in reference
single reaction
plane model ~
•
penetration theory
two reaction planes
0.4 0.8
OH , gmoles/liter
1.2
Chemical enhancement of CO2 two reaction plane model
(ONDA, 1972)
47
-------
theory models. The data for the CO- enhancement factor shows good correlation
with the penetration theory but poor correlation with the more simple models.
Figure 16 shows the effect,of reducing the partial pressure of H2S.
The relative enhancement factor (o) increases dramatically below a partial
pressure of 0.4 atm., H2S. This feature of high relative chemical enhancement
at low H~S concentrations can be effective for maximizing selective absorption
in multi-stage scrubbing systems.
Figure 17 shows the results of experimental runs by Astarita, 1965 to
evaluate the effect of CO3~2 ion on H2S scrubbing. Solutions of NaOH and
Na-CO, were prepared with varying OH~ concentrations but maintaining a total
2 3
OH~ + CO3~2 concentration of 1 gmole/liter. The absorption for H2S over CO2
increases dramatically with decreasing OH~ concentrations. The ratio of
chemical enhancement increases from 10:1 at 1 molar NaOH to 50:1 at 0.1 M NaOH
and 0.9 M Na2CO3. These data confirm the scrubbing effect of the CO3~2 ion
for H0S. The data also indicate that the presence of the OH~ ion is more
important than the CO3~2 for chemical enhancement. This figure also indicates
_2
the competing nature of the selective absorption process as the high CO3
concentrations result in high chemical enhancement ratios but lower removal
efficiency. :
The above discussion and experimental results were evaluated for
gas/liquid contact times of the order of magnitude of 0.1 seconds. It has
been reported that reducing this contact time can improve relative chemical
enhancement due to the higher reaction rate for H2S over CO2.
In an analysis of the scrubbing efficiency of CO3~2 in a spray tower,
Aiken,, et al, 1983 used a series of gas sample ports to follow the
concentration of H-S and CO2 as a function of distance from the spray nozzle,
which is equivalent to residence time for reaction. The results indicate that
the H..,S concentration in the gas was reduced to its minimum value at the first
gas sample port while the CO2 concentration in the liquid continued to
increase. This confirms that limiting the contact time for reaction should
favor H0S selectivity.
48
-------
20
-P
c
0)
I
u
JS
C
rfl
U
•H
g
Q)
U
CO.
10
\
experimental results reported.in
reference :
10
Legend:
r\ = Chemical enhancement due
to solubility only
6 = Chemical enhancement
total '
a = Selectivity
(all are unitless)
0.5
1.0
Ambient Pressures, Atmospheres
Figure 16- Effect of H S partial pressure on selectivity.
(Onda, 1972)
49
-------
u
tt)
CO
^v
to
0)
in
o
Q)
•4-1
(0
to
m
s-i
to
to
id
S
60
— —2
total (OH + CO3 ) cone. = 1 gmoles/liter
\
\
\
\
40
20
Legend:
See Figure 5-12
60
40
tr
Ui
D
20
Figure 17.
0.5 1.0
OH , gmoles/liter
-2
Effect of CO concentration on selectivity
(Onda, 1972)
50
-------
2.4.3 Analysis of Three Component System (H^S-NH^-CC^) Using Penetration
Theory
The literature reviewed above has the following limitations with
respect to oil shale applications:
1. It does not account for the presence of ammonia in the gas
acting as a scrubbing agent. The ammonia in the retort gas
reacts with H-S increasing selectivity and removal
efficiency.
2. It makes ho provision for estimating performance of a
venturi scrubber with short residence times to maximize
selectivity. The characteristics of a venturi scrubber of
short residence times 0.003 seconds and relatively high gas
phase coefficients favor selective H2S absorption. ;
Therefore, to evaluate the test data and to be able to extrapolate
these test results into a realistic design concept, a computer program was
developed incorporating a venturi scrubber model with three component
absorption, with reaction mass transfer model, all based on the penetration
theory. This computer model was developed from fundamental principles. The
reader is referred to Appendix B for a complete description of the
mathematical derivation of the model. The essential features of the system
are presented below. ;
The model calculates the selectivity and removal efficiency for the
H0S NHo-CO0 gas in contact with an alkali solution in a venturi scrubber. The
2, J £• '
calculation technique:
1. Determines the chemical enhancement factor from the
penetration theory model based on the concentration of the
gas and liquid.
2. Calculates the mass transfer rate based on the physical
characteristics of the venturi and chemical enhancement.
3. Updates the concentration profile based on the mass transfer
rate.
4. Repeats the above routine for small intervals along the
length of the venturi.
51
-------
A. Penetration theory for mass transfer and reaction of H2S-CO2-NH3 —
The presence of NH3 in the retort gas significantly affects removal
efficiency and selectivity f or H2S .
The gas reaction is:
H2S + NH3 •»• HS~ + NH4+ (1
Three cases must be considered depending on the initial concentrations of N
and H2S
H2S
II.
III. NH3 < H2S
The species which is in lesser amount (NH3 or H2S) will be> consumed at
the interface (Reaction 16) and will not exist inside the liquid film. Its
absorption rate will be entirely controlled by the gas film; liquid film
resistance to mass transfer will be effectively zero. Its interfacial
concentration can be set to zero for computing the rate of transfer across the
gas film. The species in excess will diffuse into the liquid phase and
react. Carbon dioxide, which is unaffected by the presence of NH3, diffuses
into the liquid and reacts according to the equation:
CO + 20H~ * C0~2 + HO (13)
£ - J ^
This reaction is also instantaneous and irreversible. There will be a
reaction plane at which CO2 and OH~ are consumed instantaneously.
Case I. [NH.^ > [H2S]i
H2S reacts at the interface; the excess NH3 (dissolved) is consumed by
the instantaneous and irreversible reaction,
NH + H+ -»• NH* (17)
52
-------
The species to be considered are NH , HS , CO , CO , and OH . The two
species which react instantaneously and irreversibly at a plane are CO2 and
OH~ according to Reaction 13) above.
Reaction between CO,, and ammonia (or NH ) can be neglected because
4
of unfavorable equilibrium constants (K ~ 10 ). All the other species
undergo physical diffusion only. The enhancement factor for H2S and NH3 in
the liquid film is infinite, i.e., absorption of H2S and NH, is entirely
controlled by gas film resistance. The interfacial concentration of both H2S
and NH3 can be set equal to zero to calculate the rate of absorption across
the gas film.
Solving the partial differential equations for diffusion to determine
the chemical enhancement for CO2 results
E = instantaneous enhancement factor = „ ..
erf* {«__} 1/2.
where D is the diffusion coefficient (cm/sec2), a is determined by the
equation
r_a_i 1/2 _ Co ,Dc r_a i 1/2 r_a _a .1
DC Ai °A °A AC
and subscripts A = CO2 and C = OH~
These equations are shown here merely to illustrate the form of the
solution. In this form, it is not possible to obtain a physical sense of the
process. Only by a parametric study using these relationships can the process
be understood. By comparing results predicted by this model with experimental
measurements, such as performed in this test program, the validity of these
abstract relationships can be evaluated.
* erf = error function, a standard mathematical function
53
-------
Case II [NH3]i = [H2S]i
This case is very similar to Case I. Both H2S and NH3 are consumed at
the interface by the Reaction 12.
H S + NH -»- HS~ + NH* <1 6>
Reaction 14 does not occur since there is no excess NH3» Reaction 13 does
occur, however. The concentration profile and enhancement factor for CG>2
remain the same as in Case I except that
. .
L 4Ji L Jx
Case III. [ H2S]i > [NH^
This is the most complex and challenging case mathematically. NH3 is
converted to NH at the interface by the reaction
H S + NH ->• HS~ + NH* (16)
The excess H2S along with CO2 diffuses into the liquid and reacts with OH~.
The mathematical modeling expressions for this case are presented In Appendix
B and as stated earlier are abstract and difficult to relate directly to
physical phenomena. Basically, the approach taken is to use the two reaction
plane theory discussed in Section 2.4.2.B and add penetration theory
expressions which provide for a time variation of the concentration of each of
the chemical species. This model can account for a process where the
gas/liquid contact time is of the order of milliseconds such as in a
venturi. It can also treat dimensional aspects of the system such as venturi
geometry and liquid droplet size which can assist a designer in optimizing the
venturi contactor.
B. Venturi Scrubber for Multicomponent Mass Transfer with Reaction —
Once the chemical enhancement factors have been determined for a
specific concentration profile, the mass transfer rate must be determined by a
mass balance.
54 ;
-------
The mass balance equations provide for the relationship between liquid
and gas phase concentrations as material is transferred from the gas to the
liquid phase. The rate of mass transfer is determined from the
characteristics of the contactor. The gas-side mass transfer coefficient, kG,
is computed with consideration to the droplet size and varying relative
velocity. The liquid-side mass transfer coefficient for physical absorption
is taken from the penetration theory.
C. General Results—
The following discussion presents the results of a typical analysis.
The application of the model to the test data is presented in Section 2-5.
Figure 18 shows the removal efficiency of H2S and NH3 versus distance
down venturi throat; Figure 19 shows selectivity, S, defined as
% removal H S
S =
% removal CO
versus length of venturi. Figure 18 indicates that 60 percent of the H2S and
70 percent of the NH, is removed in a single pass through the venturi. Only
two percent of the CO2 is removed (not shown on the figure). Most H2S removal
occurs early in the throat. The corresponding selectivity shown in Figure 19
indicates that a maximum in the selectivity is likely at some intermediate
venturi length. This agrees with the results of Hsieh and Aiken (1984) and
can be explained by the notion that up to and including the region of the
peak, H2S is gas film controlled while CO2 is liquid film controlled. The gas
film coefficient is high for short contact time but decreases as the contact
time increases. This is because the liquid droplets accelerate and the
relative velocity between the gas and the droplets decreases while the liquid
film coefficient does not decrease as rapidly.
Figure 20 shows the dependency of the selectivity on reactant
concentration. The selectivity is seen to decrease substantially with
increase in OH~ concentration. The CO2 reaction is aided more by increased
OH~ concentration than the H2S reaction.
55
-------
+J
•H
n I i
o
CM
-6
O
-p
0)
,-P
I
o
OJ
•H
U
•H
in
4-1
W
(S
O
m
s c
o c
0)
C i *]
o
-H -H
^ ^
•H -P
J"4 G
Hi Q)
OD
rH
CP
•H
O
IT)
O
CO
56
-------
O
O
rH
o o
CO VD
I
0)
00
CM
o
rsi
e
U
tn
0)
CO
•H
>H
3
4J
(1)
a)
•H
U
G
•H
en
0
to
O
•H
-P
U
w
(0
T3
•p
U
0)
i-H
CD
W
0)
M
3
57
-------
o
o
o
o
0)
JJ
•H
Q)
iH
O
as
o
(N
o
o
in
O O
<* n
o
o
•H
-H
+J
o
0)
i-H
(U
w
c
o
c
0
•H
-P
0)
U
o
U
•H
X
o
m
•H
-H CQ
M-l X
-P fi
O 0)
a) a
4-1 Oi
o
CM
58
-------
As discussed above, the required design for the venturi scrubber is
based on the trade off between removal efficiency and selectivity. Figure 21
shows the model results for both removal efficiency and selectivity. For a
given venturi length (i.e., residence time) the removal efficiency and
selectivity can be easily determined. The figure shows that a venturi
designed for the peak selectivity of 110 at 12 cm length can provide a 50
percent removal efficiency.
The computer model was used to investigate the effect of liquid
droplet size on selectivity. The base case assessed a droplet diameter of
30 pm. It was found that increasing the droplet size to 60 pm can improve
selectivity by as much as 20 percent. This is again due to the effect on the
gas film coefficient. Larger liquid droplets accelerate more slowly to
maximum velocity during which time the differential velocity between the
droplets and the gas is high. High differential velocities result in high gas
film coefficients and therefore favor E^S removal. Conversely, small liquid
droplets accelerate faster and favor CO2 absorption because of both the lower
gas film coefficient and the greater liquid surface area.
The effect of temperature on selectivity was also evaluated. The
model only considers temperature effects with respect to vapor-liquid
equilibrium. There is a slight increase in selectivity with temperature as
shown in Figure 22.
2.5 DATA ANALYSIS TECHNIQUES
2.5.1 Removal Efficiency
The removal efficiency was calculated from the H2S inlet and outlet
concentrations as
n (H2S) = % removal H2S = (H2S ppm in - H2S ppm out)/H2S ppm in
59
-------
o
o
o
CO
o
•vf
o
CM
4J
t7>
0)
IT)
CM
O
ON
O
a;
i-H
C
0)
-P
C
Q)
-P
-H
O
•H
-P
m
Hi
•H
•P
O
0)
U
C
•r-l
U
i
§
p;
O
M
3
o
LO
o
(N
60
-------
35
30
•H
4-)
0
0)
r-l
(1)
w
25
"
20
I I
300 325
Temperature, °K
350
Figure 22. Effect of temperature on selectivity. (Appendix B)
61
-------
2.5.2 Selectivity
Selectivity is defined as S = n (H2S)/n (CO2>
and n (CO2) = % removal CO2 = moles of CO2 absorbed/moles of CO2 in the retort
gas.
The moles of CO2 input is calculated from total moles in = flow,
liter/sec x 3600 sec/hr. x molar density (gmoles/liter).
The molar density is determined by the ideal gas law,
n/V, gmoles/liter = P/RT = (P, atm) / (0.082 x (T, °C +273).
Moles of CO2 in = moles of gas in x % CO2 (0.23)
Moles of CO2 absorbed was determined by analysis of the scrubber water
discharge for each run.
!
Moles of CO2 absorbed = liquid flow, liter/sec x 3600 sec/hr x
[(HCO~ + CO~ out - CO in) gmoles/liter]
j -j 3
2.5.3 Data
A. Gas Analysis Data—
1. H0S—The inlet and outlet gas analysis technique is described in
^.—
Appendix A. The inlet and outlet H2S concentrations were used for the data
analysis.
2. NH3—NH3 was determined for the retort gas and scrubber gas
effluent. The results are shown in Table 6. The tests with high NH4OH
concentration showed considerable removal of the NH3 resulting in large
increases in the exit gas composition. At the low NH4OH concentration, the
NH3 stripping was significantly less with the exit gas concentration
increasing by 30-120 percent.
62
-------
TABLE 6. NH3 CONCENTRATION IN GAS STREAM
Run
12
13
14
15
16
17
NH4OH
gmoles/liter
2
2
0.049
0.049
0.29
0.29
Inlet NH3
ppm
1190
414
611
442
464
461
Outlet NH3
ppm
24256
6787
1332
575
3007
947
Increase,
; % NH,
1938
1539
118
30
548
105
B.
Water Analysis Data—
The water analysis techniques are described in Appendix A. The
pertinent results used in the data analysis are shown in Table 7. The molar
concentrations were calculated from the equation.
gmoles/liter =
mg/liter
1000 mg/g
x
gmoles
MW, g
Table 7 also shows the sulfate values from the water analysis. No appreciable
sulfate was found and, in fact, the scrubber effluent had less sulfate than
the make-up water. The samples were also measured for sulfite concen-
tration. However, the sulfite values were too low to offset the interference
from the sulfide ion in solution.
2.6 RESULTS
2.6.1 Removal Efficiency
The removal efficiency results from the test program are presented in
Table 8 and Figures 23, 24, and 25. There is some question as to the correct
[OH~] concentration to use when evaluating the data for ammonia. Ammonia is a
weak base with the following reaction
NH4 + OH
63
-------
TABLE 7. WATER ANALYSIS DATA
Run No.
12
13
14
15
16
17
18
19
20
21
22
24
25
26
27
28
29
30
Water
Sulfide
Mg/1
670
800
400
400
490
520
400
420
280
310
290
310
660
290
250
190
300
310
«
Ammonia
Mg/1
20,000
1,200
1,800
1,200
7,300
4,300
250
180
340
260
260
230
280
1,700
370
250
290
220
1
Carbonate
mg/1*
19,000
14,000
4,100
2,200
6,200
2,900
1 , 1 00
1,800
600
840
600
840
1,300
3,000
960
840
960
1,600
26
Bicarbonate
mg/1*
<1
<1
<1
<1
<1
°
3,200
370
1,600
850
2,000
980
2,600
4,000
1 , 500
980
1,600
490
250
Hydroxide Sulfate
mg/1 mg/1
11,000 160
10,000 33
34 61
200 53
1,700; 66
1 , 600 38
<1 63
<1 75
<1 ! 100
<1 110
<1 110
<1 120
-------
TABLE 8.
H2S REMOVAL EFFICIENCY (E) FOR SCRUBBER TESTS
Run No.
12
13
14
15
16
17
18
19
20
21
22
24
25
26
27
28
29
30
31
32
33
34
H2S,
in
1780
tl
11
It
II
n
1280
n
n
ii
ii
ti
1280
II
II
It
tt
If
998
1048
1065
1003
ppm
out
128
595
645
704
167
683
218
384
614
609
588
666
154
372
583
596
519
660
60
83
75
62
AH2S,
ppm
1657
1185
1135
1076
1613
1097
1062
896
666
671
612
614
1126
908
697
697
761
620
938
965
990
941
E, %
93
67
64
60
91
62
83
70
53
52
54
48
88
71
54
53
59
48
94
92
93
94
Alkali Contactor
NH4OH tower
venturi
tower
venturi
tower
venturi
NaOH tower
venturi
tower
venturi
tower
venturi
KOH tower
venturi
tower
venturi
tower
venturi
KOH tower
KOH tower
NaOH tower
NaOH tower
OH~
Concentration
gmoles /liter
2.0
2.0
0.049
0.049
0.29
: 0.29
0.045
0.045
0.012
0.012
; 0.023
0.023
0.046
0.046
0.012
0.012
0.023
0.023
0.89
1.79
1.25
2.49
65
-------
0)
-p
•H
U)
01
in
CN
QJ
-P
•H
tH
[0
(U
S3
O
S
1
^
a
g
w
o
"*
in
o
OD
-------
0)
-p
U)
0)
o
ffi
o
<*
as
-p
0)
tH
c
•H
4J
g
as
o
as o
o as ^
m o 33
sws
OD<]
O
O
,-H
O
(Tv
O
CO
O
r-
O
in
o
in
o
o
o
n
O
CN
o
M
0)
-P
•f->
en
cu
rH
as
o
o
o
LD
o
o
V
as
o
4J
id
a)
I
tn
-P
en
0)
K
u
c
0)
-H
u
•H
rd
1
0
S-l
'AousTDTrpga
67
-------
in
o
o
o
o
cn
o
oo
o
r-
o
10
o
in
33
O . 03
iti O
a M
33
°^
w
3
C '
•H
0
C •
S3
O
I
S3
OD<1
o
o
ro
O
r-t
O
dJ
-P
•H
cn
0)
CM g
O &i
O p-.
rn
o
'ADUBTDTJja
Q)
4J
•H
0)
1-1
I
in
o
o
v
o
-P
n3
•H
d
0)
S-J
0
U-l
u
c
0)
(0
i
LO
CN
Q)
S-l
68
-------
[NH + ] [OH ]
and K = - - = 1.75 x 10 at 25°C
However, since the [OH~] is a product that is consumed, the dissociation
reaction is driven to the right. Thus, continuous renewal of [OH~] is
provided rather than an equilibrium condition. The exit scrubbing liquid
[OH~] can be considered the minimum [OH~], while the inlet NH4OH concentration
can be considered as a maximum [OH~] . Consequently, an arithmetic average of
the inlet and outlet [OH~~] was used for ammonia data analysis. These values
are summarized in Table 9.
TABLE 9. [OH~] FOR AMMONIA TESTS
Inlet NH4OH
Cone. , gmoles/liter
2.0
0.29
0.049
Exit
[OH~], gmoles/liter
0.1
0.002
0.016
Average
[OH~], gmoles/liter
1.0
0.025
0.154 ;
A. [OH~] Greater Than 0.05 gmoles/liter—
The removal efficiency varied from a low range of ~55-70 percent
(venturi) to a maximum of 80-93 percent (tower) (Figure 23). There was a
consistent trend showing higher removal efficiencies with the tower than the
venturi at equal OH~ concentrations. This is expected as the longer residence
time in the tower provides for longer time for absorption.
B. [OH~] Less Than 0.05 gmoles/liter—
1. Tower—The removal efficiency for the tower runs at [QH~] <.05
gmoles/liter is shown in Figure 24. The removal efficiency varied from 55
percent at the lower [OH~] of ~0.012 gmoles/liter to 88 percent at ~0.05
gmoles/liter of OH~. Two distinct trends are apparent. At [OH~] greater than
69
-------
0.025 .gmoles/liter, there is a significant improvement in removal efficiency
with higher [OH~]. This is to be expected on the [OH~] determines' the
relative chemical enhancement. However, at [OH~] less than 0.025
gmoles/liter, and greater than 0.01 gmoles/liter, the removal efficiency is
relatively independent of [OH~], This result is due to the NH3 present in the
retort gas. As the H2S is removed by its reaction with NH3 (Reaction 12), the
minimal (if any) dependence of H2S removal with [OH~] is consistent with the
theoretical model discussed in Section 2.5.
It should be noted that this NH^-H-jS reaction will occur at [OH~]
approaching zero which indicates removal of H-S with a water scrubber without
alkaline feed. This will affect the process and plant design for the in-situ
plant analysis.
In a concept design for an in-situ shale oil retort offgas processing
plant (Denver Research, 1983), the retort gas is first treated in .an absorber-
cooler, which "condenses light oils and ammonia containing water." The
material balances given in this report indicate that with a 3:1 ratio of NHg
to H2S in the retort gas, only 1.4 percent of the H2S is absorbed while 92.3
percent of the NHo is absorbed. This material balance is not consistent with
either the theoretical or experimental results. Both the H2S and NH3 will be
I
removed in the absorber and this fact will affect the process and .plant design
downstream. Qualitative observation of actual H2S removal during the plant
Startup tests with water recirculation through the venturi showed a 10 to 15
percent H2S removal efficiency. Similar conditions for the tower ;were not run
but it would be expected that the longer residence times in the tower would
result in greater H2S removal.
There was no significant difference in the performance with any of the
scrubbing chemicals at equivalent OH~ concentrations.
2. Venturi—The effect of [OH~] on removal efficiency for the venturi
runs at [OH~] less than 0.05 gmoles/liter is shown in Figure 25. .The removcil
efficiency ranged from 48 percent to 70 percent and showed a minimal
dependence on OH~ concentration. These results also indicate the same
"leveling off" of the dependence of removal efficiency with [OH~] below OH~
concentrations of 0.025 gmoles/liter. At OH~ greater than 0.025 gmoles/liter,
70
-------
there is a slight increase in removal efficiency with [OH~] from 55 percent to
65 percent. These results are consistent with the discussion in the previous
section on the scrubbing effect of Nf^ in the retort gas.
There was no significant difference in performance for any; of the
scrubbing chemicals at equivalent [OH~],
3. Summary-Figure 26 shows the combined results for the tower and
venturi runs.
The two contactors show similar performance at the low [OH~] with the
tower performance improving more rapidly than the venturi at [OH~] between
0.02 and 0.05 gmoles/liter. For both contactors, the performance at 0.05
gmoles/liter [OH~] approaches their maximum values of 93 percent and 66
percent for the tower and venturi respectively. It appears that the effect of
the NHg in the retort gas which results in the lack of dependence at low [OH~]
becomes less important at [OH~] greater than 0.25 gmoles/liter for both
contactors.
2.6.2 Selectivity
The results for the selectivity analysis are presented in Table 10 and
Figure 27. :
A. Tower— :
The selectivity for the tower runs ranged from a low of 9 for the high
OH~ concentrations to a high of 52 at the low OH~ concentrations. This trend,
increasing selectivity with decreasing OH~ concentration is consistent with
the previous theoretical development. The ammonia test results using- the
average NHLOH concentration do not correlate well showing lower selectivities
than NaOH and KOH at equal [OH~1. This is inconsistent with theoretical
analysis and is most likely due to the emperical approach using th
-------
•P
•H
in
o
o
o
ro
O
M
0)
4-)
M
0)
o 51
o r
M
O
O
o
0)
0)
1-1
o
IT)
O
o
V
•H
B
0)
u
o
•H
m
u-i
w
(C
1
V
o
o
o
CTi
o
r--
o
in
'AouBTDTjga
72
-------
TABLE 10. TEST SELECTIVITY FOR THE ALKALINE SCRUBBER
Removal Efficiency ;
Run
Tower
20
27
22
29
18
25
14
16
12
Venturi
21
28
24
30
19
26
15
17
13
Chemical
NaOH
KOH
NaOH
KOH
NaOH
KOH
NH4OH
NH4OH
NH4OH
NaOH
KOH
NaOH
KOH
NaOH
KOH
NH4OH
NH4OH
NH4OH
OH gmole/liter
0.012
0.012
. 0.023
0.023
0.045
0.046
0.049
0.29
2.0
0.012
0.012
0.023
0.023
0.045
0.046
0.049
0.29
2.0
H^S, %
52.0
54.5
54.1
59.5
83.0
88.0
63.8
90.6
92.8
52.4
53.4
48.0
48.4
70.0
70.9
60.5
61.6
66.6
CO-y, %
1.0
1.27
1.32
1.22
2.32
2.14
2.18
3.15
10.38
0.66
0.75
0.79
0.94
0.84
3.42
0.85
1.10
6.11
Selectivity
52
43
41
49
36
41
29
I 29
9
79
71
60
51
84*
21
• 72
56
11
*suspect water data
73
-------
CO
o
O
o
O
O
O
l-i
a)
4-1
•H
tn
0)
rH
O
JS m
•H O
S
8
SB
O
-P
itJ
-P
•H
•H
*J
O
0)
i-H
.0)
fcj
o
co
o
in
o
ro
O
CN
U =
74
-------
0.003 seconds compared to approximately 0.2 sec. for the tray tower the high
gas film coefficient and presence of NHj in the retort gas.
C. Comparison between Tower and Venturi— :
The test results indicate that the selectivity for the venturi is
highly sensitive to the OH~ concentration with a rate of change, dS/dOH~, of
-1700 liter/gmole in the OH~ concentration range of 0.01 to 0.04
gmole/liter. The tower results show a rate of change of only -300 liter/gmole
in the same [OH~] range. This effect is due to the presence of NH3 in the
retort gas. The short residence time in the venturi (0.003 seconds) results
in a high dependence of selectivity on [OH~] due to the direct dependence of
CO- enhancement. In other words, the short residence time means that the CO2
has a limited time to react. However, as the H2S absorption is controlled by
the gas/film, its absorption site is independent of the [OH~] at [OH~] values
less than 0.03 gmole/liter.
Figure 27 also shows that, at [OH~] greater than 0.03 gmoles/liter the
tower provides higher selectivity than the venturi.
This is due to the combined effect of the gas film coefficient and the
presence of NHo. The higher gas coefficient in the venturi essentially
increases the availability of the CO2 at the scrubbing liquid interface.
Consequently, the liquid phase chemical enhancement factor, which '-is a direct
function of [OH~], has a substantial effect on the CO2 absorption rate. In
the tower, the gas film coefficient is lower which decreases the relative
importance of the liquid film and, therefore, decreases the dependency of the
CO2 absorption on OH~ concentration. Since the H2S removal is determined
solely by the gas film coefficient - due to the presence of NH3 in the gas,
the sensitivity of H2S absorption to OH~ concentration in both tower and
venturi is decreased. ;
Figure 28 shows the H2S and CO2 removal efficiencies. At [OH~] below
0.035 gmoles/liter, the tower and venturi show similar C02 removal
efficiencies. However, at [OH~] greater than 0.035 the venturi rempves CO2
more readily than the tower and, therefore, has lower selectivity.
These results indicate a clear choice of alternatives in deciding
between a tower or venturi scrubber based on process requirements. If
75
-------
un
o
•sf
o
o
o
CN
O
Vl
0)
•P
•H
g
tn
ffi
O
O
O
u
c
0)
•rH
u
M-l
o
(3
CM
o
u
12
en
CM
ffi
CO
M
0)
n
3
ijl
I I
o
co
o
r--
o
in
76
-------
select!vities greater than 50 are needed, the venturi is required to take
advantage of the high selectivity at the short contact time. However, the
venturi scrubber will only provide 50 to 60 percent removal efficiency per
stage. If a selectivity less than 50 is acceptable, the tower is more
effective in that both removal efficiency and selectivity is greater than with
the venturi.
D. Comparison with Theoretical Model— :
The theoretical model described in Appendix B was used to evaluate the
correlation with the test results. The model results were evaluated at
distances of 20.5 and 24 cm from the point of liquid injection. This range
was used due to the fact that the effective scrubbing in the first 5 cm is
questionable since it takes approximately 5 cm to achieve complete
atomization. The resulting selectivities are shown in Table 11.
TABLE 11. COMPARISON THEORETICAL AND EXPERIMENTAL SELECTIVITIES
OH~, Theoretical Experi-
gmoles/liter 24 cm* 20.5 cm* Average mental
0.045 25 31 28 21
0.023 43 53 48 55
0.012 66 82 74 75
Avg.
Difference, %
25
14.6
: 1.4
13.7
*Distance along venturi throat that fluid is injected.
The theoretical selectivities are in good agreement with the test
results with respect to both trend and absolute values. There is excellent
agreement at the low concentration range (1.4 percent) while the higher
concentration (0.045 gmoles/liter) has a 25 percent deviation. This
information is also shown in Figure 27 as a range of predicted values for each
concentration. The agreement of the theoretical model with the test results,
particularly at the lower concentrations which are of primary interest when
77
-------
evaluctting a venturi scrubber, indicate the model can be used for predictive
studies of multi-stage performance.
2.7. FINDINGS
The following summarizes the pertinent alkaline scrubber results
obtained from the test program:
1. All three alkaline solutions provided similar results at
equal [OH~3 for both scrubbing efficiency and selectivity.
Correlation of the ammonia results were complicated by
incomplete dissociation of the weak base.
2. Removal efficiencies of 85-90 percent with a selectivity of
30 can be achieved in a tray tower with as low as 0.045
gmoles/liter OH~ concentration.
3. Selectivity in the tower was only slightly dependent on
[OH~] ranging from 45-50 at 0.012 gmoles/liter [OH~] to 25-
30 at 0.045, gmoles/liter [OH~].
4. Removal efficiencies of 55-65 percent can be achieved in the
venturi with nominal dependence on OH~ concentration.
5. Selectivity of 70-80 can be realized in a venturi at low
(0.012 gmoles/liter) OH~ concentrations.
6. At [OH~] less than 0.03 gmoles/liter, the venturi had higher
selectivity than the tower, while at [OH~] greater than 0.03
gmoles/liter, the tower exhibited higher selectivity than
the venturi. This occurs because of the effect of the NH3
in the retort gas which provides for H2S removal based
solely on the gas film coefficient. Therefore, at low [OH~]
the venturi scrubber is effective for EUS removal with
nominal CO2 absorption. However, at high [OH~], the CO2
absorption increases more rapidly in the venturi than the
tower due to the higher gas film coefficient in the venturi.
The following summarizes the alkaline scrubber results from the
theoretical model analysis for the venturi:
7. The NH3 in the retort gas reacts with the H2S at the gas-
liquid interface. Therefore, the removal efficiency is only
marginally dependent on the [OH~],
8. Removal efficiencies for NH3 and H2S are similar.
78
-------
9. The selectivity is significantly affected by contract' time
with a maximum selectivity of 110 occuring at approximately
0.0015 seconds contact time.
10. The agreement between the theoretical model and field test
results is excellent at low OH" concentrations (0.012 to
0.025 gmoles/liter) which is the primary range of interest
for the venturi.
11. Variations in temperature and liquid droplet size can have a
significant effect on selectivity.
2.8 TWO STAGE SYSTEM
The above findings suggest two alternative alkaline scrubber design
concepts for further consideration and evaluation. One system combines the
high selectivity of the venturi with the high removal efficiency of the
tower. The other design concept uses a tower for maximum H2S removal and
isolated liquid input to maximize selectivity for use with a Glaus plant.
2.8.1 Venturi-Tower
The design objective for this concept is to obtain a minimum removal
efficiency of 95 percent with selectivity greater than 30 which cannot be
obtained with either the venturi or the tower in a single stage. , The venturi
can have the high selectivity but the low removal efficiency requires too many
stages for the I^S removal requirements. The tower can approach the 95
percent removal efficiency but selectivity drops below 30 at removal
efficiencies greater than approximately 85 percent.
A two stage system that will exploit the specific design features for
each unit can provide a system (shown in Figure 29) that will meet the above
process requirements.
The first stage is a venturi designed for peak selectivity based on
contaict time and OH~ concentration. (See Figure 19). The theoretical model
indicates a maximum selectivity of 110 will result in a 50 percent removal
efficiency. The CO2 removal efficiency is 0.4 percent.
79
-------
H S = 1500 ppm
CO0 = 22%
= 2500 gmoles
hr
0.012 gmc
I
Venturi
le/liter OH
H2S = 750 ppm
75 ppm H2S
1 Tower
0.045 gmoles/liter
OH~
(Basis)
10 gmoles/hr CO,
55 gmoles/hr CO,
Figure 29. Two-Stage Process
The second stage, a tray tower, is designed for approximately 85-90
percent H2S removal efficiency with a stage selectivity of 40 using an OH~
concentration of 0.045 gmoles/liter. The CO2 removal efficiency is 2.2
percent.
The net result from this design is a 95 percent removal efficiency
with a selectivity of 37. These results are shown in Table 12.
TABLE 12. TWO STAGE DESIGN CONDITIONS
Position
C02'
gmoles CO
OH~, gmoies/liter
Gas Liquid
Inlet 1500
Stage 1-Venturi 750
Stage2-Tower 75
22
21.9
21.4
2500
2490
2435
65
H2S removal Efficiency =
CO~ Removal Efficiency =
10
55
1500 - 75
1500 .
2500 - 2435
2500
0.012
0.045
= 95%
= 2.6%
Selectivity = 37
80
-------
2.8.2
Tower - Tower
If a lower selectivity can be tolerated and a higher removal
efficiency is required, a multi-stage tower can be used. Selectivity can be
increased by providing isolated stages with respect to scrubbing liquid to
take advantage of the higher selectivity at low H2S concentrations illustrated
in Figure 16. Figure 30 shows a schematic for this tower design.
Scrubber Liquid
Inlet
Retort Gas
t
Gas to Process
Scrubber Liquid
Outlet"
Figure 30. Tray Tower with Isolated Liquid Inlet ;
The H-S concentration to each stage is reduced by approximately 50 percent per
stage.
By using fresh scrubber solution for each tray, the removal
efficiency/tray is maintained but selectivity should increase because of the
lower inlet H2S concentration to each stage. The performance of this concept
is summarized as follows:
H2S Cone., ppm
CO2 Cone., %
H2 removal eff. =
1500-15
1500
Inlet
(Assumed)
1500
22
= 99%
Stage 1
Exit
150
21.5
Scrubber
, Exit
15
21.0
22—21
CO2 removal eff. = —r^ = 4.5%
Selectivity =
99
4.5
22
22
81
-------
2.9 ACTIVATED CARBON PROCESS
A variation of the caustic scrubbing process using activated carbon as
a catalyst has been developed by the Pulp and Paper Research Institute of
Canada (Prohocs, 1983) for the purpose of controlling H2S emissions from black
liquor recovery furnaces. The flue gas concentrations from these furnaces are
similar to the retort gas concentration from in-situ retorts.
This process appears to have three process advantages over the basic
alkaline scrubbing process: (1) higher H2S removal efficiencies, (2) removal
of organic sulfur compounds, and (3) a more salable byproduct (sodium
thiosulfate).
The GKI tests and the analysis reported above indicate that a
scrubbing efficiency of 93-95 percent can be achieved but only on the H^S.
The organic sulfur compounds are not removed. Since the organic sulfur
compound can account for one to four percent of the total sulfur in a typical
retort gas, the net result is that a scrubbing efficiency of 96 to 99 percent
on the H2S is required to obtain a net sulfur removal efficiency of 95
percent.
Scrubbing the offgas from a black liquor recovery furnace containing
H2S in the presence of large amounts of CO2 was accomplished using as little
as 0.03 weight percent of activated carbon in suspension. The principal
reaction steps are:
1 . H2S + CO3~2 -»• HCO3~ + HS~
2. Partial adsorption of HS~ on the surface of the activated
carbon
3. Oxidation of HS~ to S2O3~2
The reactions are:
Absorption of CO2:
2NaOH + CO2 •»• Na2CO3 + H2O (17)
Na2C03 + C02 + H20 -»• 2NaHCO3 (18)
Absorption of H2S:
Na2CO3 + H2S -»• NaHCO3 + NaHS (19)
82
-------
NaHCO3 + H2S •»- NaHS + C02 + H2O (20)
Reactions 17 and 18 govern the (initial) chemical composition and, more
importantly, the pH of the scrubbing solution.
Reactions 19 and 20 have unfavorable equilibrium coefficient
absorption of H2s . Removing this limitation of H2S absorption due to the
unfavorable equilibrium is the purpose of the oxidation reaction —
2NaHS + 20.. ^ carbon Na s o + H O (21)
^ ^ £• O £
Reactions 19 and 20 have a very unfavorable equilibrium with respect
to the absorption of H2S, particularly in the presence of the more acidic C02r
present in concentrations of 10-16 percent by volume.
With NaOH or Na2CO3, in the initial alkaline solution, the absorbed
CO2 will also depress the pH. At 70°C (the typical scrubbing temperature
range) the pH is depressed to values of 8.5 to 9.0 which significantly
decreases the H2S absorption rate and therefore limits the removal
efficiency. But Reaction 21 under normal conditions of scrubbing, proceeds
very rapidly to the right, thus allowing more H2s to be absorbed by Reactions
19 and 20. A significant amount of the Na-jS-jOg is further oxidized to
Na2so4. H2S removal efficiencies of 99 to 99.9 percent were readily achieved
at H2S inlet concentrations of <1100 ppm.
Reaction 21 requires 2-3 percent oxygen in the flue gas. At H2s
concentrations over 100-120 ppm, a separate aeration step was required. The
aeration step occurs prior to the recycle of the scrubbing fluid. The
aeration residence time and rate requirements are a function of H,2S load and
the required exit gas concentration.
Removal of organic sulfur compounds can be enhanced by adding chlorine
gas to the alkaline scrubbing solution or using hypochlorite solution. The
absorption/adsorption and possible oxidation mechanisms of the organic sulfur
compunds were not determined. However, early laboratory tests indicated that
removal efficiency of the organic sulfur compunds may be improved by addition
of activated carbon. In addition, laboratory tests with dilute sodium
83
-------
hypochlorite (500-700 ppm) resulted in an exit concentration of less than one
ppmv of organic sulfur compounds regardless of the inlet concentration. This
process is more expensive than the basic alkaline scrubbing process and could
result in trace emissions of chlorinated organics.
84
-------
SECTION 3.0 '
STRETFORD PILOT PLANT
3.1 PROCESS DESCRIPTION
The Stretford is a regenerative process that converts H2S in the
retort off-gas to elemental sulfur. It uses air oxidation to regenerate the
chemicals reduced during the offgas treatment. The original Stretford process
was developed in the early 1950's by the North Western Gas Board and the
Clayton Aniline Company. The original pilot plant was operated on town gas at
the Stretford Road Gas Works in the village of Stretford, England. The North
Western Gas Board later became part of the British Gas Corporation (BGC),
which currently licenses the process worldwide to engineering and construction
companies.
The Stretford process has been in use for more than 25 years, and more
than 90 commercial Stretford plants are currently in service worldwide for the
following specific gas-treating applications:
Coal gasification . Glaus tail gas
. Coke oven gas . Geothermal power generation
Refinery fuel gas . Carbon disulfide manufacture
SNG (petroleum) plant gas . Ore roasting '
Natural and associated gases . Sewage sludge digester gas
The most common application of this technology is for sulfur recovery (as part
of the Beavon process for treating Claus plant tail gas). The commercial use
of Stretford technology directly on synthetic fuel process gas streams has not
been practiced. In the United States, a number of demonstration plants have
been installed on coal gasification process gas streams; however, a variety of
operating problems have limited the performance of these plants.
85
-------
3.1.1 Process Chemistry
The process chemistry of the Stretford technology is based on the
absorption of H2S and subsequent liquid-phase oxidation of the captured H2S to
elemental sulfur in an alkaline solution of sodium, vanadium, and anthraqui-
none disulfonic acid salts.
The Stretford liquor is a dilute solution of sodium carbonate
(Na2CO3), sodium metavanadate (NaVO3), and sodium salts of the 2:6 and 2:7
isomers of anthraquinone disulfonic acid (ADA). The solution is maintained
at a pH of 8.5 to 9.5 and a temperature of approximately 43°C.
Removing the H2S from the gas stream and converting it to elemental
sulfur is basically a five-step chemical process, as shown in the following
simplified chemical reactions: ;
1. The H2S reacts with the sodium carbonate to form sodium hydrosul-
fide and sodium bicarbonate:
H2S + Na2CO3 •»• NaHS + NaHCO3 (3-1)
2. The hydrosulfide then reacts with sodium metavanadate to form
elemental sulfur, a quadravalent vanadium salt, and sodium
hydroxide:
2NaHS + 4NaVO3 + H20 •> Na2V4Og + 2S + 4NaOH (3-2)
3. The quadravalent vanadium salt then reacts with ADA* to regenerate
the sodium metavanadate:
Na2V4Og + 2NaOH + H2O + 2ADA -»• 4NaVO3 + 2ADA«2H (3-3)
4. The sodium hydroxide and sodium bicarbonate reaction products
further react to form sodium carbonate: '.
NaOH + NaHC03 •*• Na2CO3 + H2O (3-4)
*The chemical formula for 2:7 ADA is: NaSO3 Vx'vA^' NaSO3
86
-------
5. The reduced ADA** reacts with oxygen to regenerate the ADA:
2ADA'2H + 02 •»• 2ADA + 2H2O (3-5)
The overall process reaction can be written as the oxidation of H2S to
f
elemental sulfur:
2H2S + O2 -»• 2S + 2H2O (3-6)
Several side reactions that form nonregenerable compounds :are possible
in the Stretford process. If sodium hydrosulfide contacts absorbed oxygen in
either the absorber or the oxidation tank (which can occur if the system lacks
adequate vanadium _ levels or is removing H2S at levels above design), sodium
thiosulfate forms according to the following reaction:
2NaHS + O2 -»• Na2S2O3 + H2O ', (3-7)
The amount of dissolved oxygen in the process liquor is pH-dependent.
The rate of Reaction 3-7 is also dependent upon pH and will decrease as pH
increases. The rate of EUS absorption (Reaction 3-1) is also pH-dependent,
which in turn is strongly influenced by the carbon dioxide content of the
gas. High carbon dioxide concentrations, such as found in the gases from a
shale oil retort, can cause the process to operate at lower pH levels, which
reduces the overall removal efficiency. ]
Any SO2 present in the feed gas is also absorbed and ultimately oxi-
dized to form sodium sulfate according to the following reaction: i
2SO2 + O2 + 2Na2CO3 •> 2Na2SO4 + 2CO2 (3 --8)
Any hydrogen cyanide present in the feed gas forms sodium ; thiocyanate
according to the following reaction:
2HCN + 2NaHS + O -»• 2NaCNS + 2H
-------
The nonregenerable compounds will build up in the system and eventu-
ally impede the performance of the Stretford process by interfering with the
principal chemical reactions. These compounds must be removed from the pro-
cess either by purging them from the system or by recovering them in a regen-
eration system.
3.1.2 Plant Design Description
The transportable pilot plant is mounted on three skids. It requires
gas inlet/outlet connections, a condensate collection connection,> and an
electrical service connection with the host site facility. The plant's design
configuration reflects the simplicity required to achieve the necessary
mobility for assorted host sites and yet contains all the necessary elements
to provide a workable, commercially representative Stretford process. An
overall view of the Stretford plant is shown in Figure 31.
Table 13 provides a summary of the plant's design conditions with res-
pect to gas characteristics and composition. Additional operating flexibility
was incorporated into the pilot plant to withstand any anticipated gas
conditions for processing oil shale retort offgas. This includes a maximum
j^S concentration of 3,000 ppmv and operating pressure from atmospheric to
five psig.
Figure 32 presents a simplified process flow diagram of the pilot
plant. This diagram depicts the basic design configuration of the plant,
including a variable throat venturi scrubber gas/liquor contactor, reaction
tank, oxidizer, pump tank, and slurry tank.
Before the retort offgas stream entered the Stretford, it was pre-
conditioned in upstream equipment (vacuum blowers and mist eliminators)
operated by GKI to remove any residual product oil. Thus, the gas stream
was pre-cleaned and saturated prior to entering the Stretford. Normally,
the gas stream would then enter the Stretford at- the blower skid, which con-
tains a compressor suction drum and booster fan assembly. (The compressor
suction drum, "knock-out drum," served to remove any slugs of condensate that
might have been carried over from the upstream product collection[ equipment..
The booster fan assembly includes the booster fan, two silencers, and an
emergency bypass line.) However, during this test program, the saturated gas
88
-------
Packed
Tower
Venturi
Contactor
Reaction
Vessel
S&A
Taller
Slurry
Tank
Reaction
Vessel
Pump
Tank
Figure 31. Overall View of Stretford Plant Installed at GKI.
89
-------
TABLE 13. KEY DESIGN PARAMETERS OF THE
STRETFORD PILOT PLANT
Inlet gas flow rate
Blower outlet flow rate
Gas inlet temperature
Blower outlet gas temperature
Blower discharge pressure
Pressure drop across venturi
Inlet gas composition,
concentration (volume)
H2°
CO-
CH,
H-
NH,
H2S
C2H6
C2H4
C3H8
C3H6
COS
cs.
RSH
0.77 am3/s (1640 acfm)
0.74 am3/s (1560 acfm)
38°C (100°F)
46°C (115°F)
70 g/cm2 (1.0 psig)
35 g/cm2 (0.5 psig)
6%
53% '
30.7%
1 .51%
1 .51%
5.85% ;
0.22%
0.13%
0.33%
0.22%
0.17%
0.17%
0.22%
<110 ppm
<100 ppm
<100 ppm ;
90
-------
II
'"Vlt
1 ) jw
II
11
I. t
5-
5§
s
a
-P
o
-p
(L
J-l
4J
-------
stream entered the stretford process at the venturi inlet by completely by-
passing the blower skid. The blower was not necessary for this test program,
because an adequate differential pressure existed across the GKI blower, thus
allowing the retort gas to be easily introduced to the stretford system from
the discharge side of the GKI blower and to be exhausted to the inlet side of
the GKI blower. Any initial concern regarding the stretford discharge being
upstream from the Stretford inlet was dispelled by considering the small vol-
umetric flow of the Stretford plant compared to that of the GKI plant which is
0.33 Sm3/S scfm compared to 10.4 Sm3/S. Because the gas passes through the
GKI compressor between the outlet and inlet the uniformity of composition at
the Stretford inlet was considered to be good. The small amount of dilution
caused by this plumbing arrangement had no effect on the program results. The
experiment is to determine the removal efficiency from inlet to outlet. As
long as the inlet value was measured after the dilution, the test results are
valid.
The pressurized gas stream first enters the variable throat venturi
scrubber, where the gas stream comes in contact with the Stretford solution.
The solution is delivered to the top of the venturi through a single feed
line with a spray nozzle. The stretford solution injected into the venturi
scrubber consists of a dilute solution of sodium carbonate, sodium ammonium
vanadate, and the 2:7 isomer of anthraquinone disulfonic acid (ADA) for the
removal of hydrogen sulfide.
A variable throat venturi designed by PEI was used during this test
progr=im. The venturi is shown in Figure 33. The top photograph shows an
overall view of the entire venturi with the elbow joint connector.' The lower
photograph shows the inside of the variable-area throat. The throat of the
venturi was six inches long, and had a diameter of six inches when fully
open. Fully closing the venturi gave an area equivalent to that of a three
inch diameter throat.
At design conditions of 0.71 m3/s retort off gas flow (twice that for
this test), the superficial gas velocity through the throat varied .from 40 m/s
to 1515 m/s, depending upon the venturi position. This was equivalent to gas
residence times from 0.004 to 0.001 seconds.
92
-------
W '"" _ * IJ-'
&e "*«•* * - ** « " <
Figure 33. Photographs of the variable-throat
venturi used on the Stretford plant.
-------
An elbow joint connects the venturi contactor to the reactor. The
reactor serves as a gas-liquid separator, collects the spent process liquor
from the venturi, and provides holdup for completion of chemical .reactions.
The reactor has a liquid inventory of approximately 1060 liters.
The retort off-gas exited from the top of the reactor into the GKI
blower suction line. The reactor discharge gas line was initially fitted with
a baffle-type mist eliminator to minimize solution carryover. Near the end of
the test program, a packed tower was fitted to the reactor outlet. The pur-
pose of the packed tower was to increase the gas-liquid contact time, thereby
increasing the H2S removal efficiency. The packed tower was constructed of a
three-meter long, 30-cm dia., steel pipe packed with 2.5-cm dia. Raschig
rings. The process solution was injected countercurrent to the gas flow
through the packed tower.
The packed tower is shown in Figure 34, and is the tall column located
.on top of the reaction vessel (on the left side of the photograph). The tube
entering the packed tower near the top is the solution injection ;line.
The variable throat venturi. in the inlet line can be seen at the ^bottom of the
photograph at the left hand edge.
The reduced process liquor flows from the reactor to the .oxidizer.
The function of the oxidizer is to reoxidize the Stretford liquor (replenish
the reduced ADA), separate the sulfur product from the liquor by air flota-
tion,, strip bicarbonate formed in the process from the liquor (as, carbon
dioxide),•and strip any ammonia absorbed from the gas stream. The stripped
carbon dioxide and ammonia are removed from the process via an atmospheric
vent stack in the oxidizer. Oxidation air is introduced into the base of the
oxidation tank through a dispersion ring. The air is further dispersed into
the liquid by a mixer. The oxidizer tank has a liquor inventory of 5,000
liters/hour (to the weir overflow).
The sulfur product is generated as a froth at the top of the oxidizer.
This froth contains approximately seven percent (by weight) sulfur. The froth
overflows a slurry weir into the slurry tank at a rate of approximately 38
liters (at design conditions). The slurry tank functions both as a slurry
94
-------
Figure 34. Packed tower installed at reaction vessel exit.
95
-------
receiver and a decanting tank. The slurry tank is agitated and has a maximum
liquid inventory of approximately 3400 liters.
The plant is also equipped with a pump tank that provides liquid surge
capacity within the system while accommodating the recovered process liquor
and the addition of makeup chemicals and makeup water. The pump tank is
agitated and has a maximum liquid inventory of approximately 4,500 liters.
The pump tank is served by the solution feed pump, which delivers
process liquor to the gas contactor. The liquor is routed through an electric
coil solution heater before it enters the gas contactor. The solution heater
allows the process liquor to contact the incoming gas at approximately the
same temperature. This feature benefits the performance of the process with
respect to chemical consumption, nonregenerative byproduct formation, and
removal efficiency.
A modification was made to inject additional heat into the system,
because of anticipated cold weather conditions. The compressed oxidizer air
was originally cooled to near ambient temperatures in an aftercooler. This
aftercooler was bypassed, which allowed the heat of compression to be added
to the oxidizer, where the heat is essential.
3.2 STRETFORD PLANT OPERATIONS
3.2.1 Introduction
This section describes the operation of the Stretford pilot plant at
the GKI site in Kamp Kerogen. In order to gain a complete understanding of
the mechanics of H2S removal by the Stretford process, a knowledge of the
operating parameters and their effect on plant performance is required. This
knowledge is also necessary for the development of full-scale designs based
on the pilot plant experience. This section presents both the proposed and
actual schedules of events, and summarizes the parameters maintained during
the test program. These process parameter values and their influence on
operations are discussed in detail in Section 3.3. Finally, operating
problems are outlined, along with field-implemented corrective actions and
some suggested alternatives.
96
-------
3.2.2 Activities Schedule
Table 14 presents the comparison of projected and actual schedules.
From the table, it is apparent that all activities did not proceed as planned.
The most significant deviation was that the testing did not occur as a con-
tinuous series. GKI equipment failures, power outages, and sampling and
analysis equipment malfunctions were the most frequent causes of delay during
this test program.
Table 14 further shows that the planned unit upset was never per-
formed. This step was eliminated due to the problems encountered in obtaining
the desired H,,S removal efficiences. These equipment problems caused the
expenditure of additional sampling crew man-hours that were not originally
plannesd.
3.2.3 Summary of Operations
Startup of the Stretford pilot plant was achieved with a minimum of
difficulty. From a mechanical standpoint, the unit worked well except for a
few minor problems. These problems included the following:
. Failure of the solution heater due to corrosion of two of
the heater elements.
. Failure of the slurry tank mixer motor.
Actual operations with oil shale off-gas being processed through the
Stretford unit amounted to 204 hours. The test series time period totalled
255 hours. The total system operating time divided by total time available to
operate was 80 percent. Of the total down time of 50 hours, 14 hours were due
to system operating problems. The remaining 36 down time hours were due to
GKI shut-downs. Excluding these 36 hours, the plant availability was 94
percent. The Stretford pilot plant run time for this program is summarized in
Table 15 on a daily basis.
The test equipment malfunctions hindered the program's original goals,
as discussed in Section 3.2.2. In spite of this, deliberate changes were made
to the plant operating conditions, so that the effect of these changes could
be documented. The aim of the work was to gather information on how different
97
-------
TABLE 14. GEOKINETICS PROJECT SCHEDULE
(1983/4)
Activity
Delivery of pilot plant
Assembly of pilot plant
Connection of plumbing and
electrical lines by GKI
Delivery of sampling
equipment
Plant startup/stabilization
Continuous testing
Deliberate system upset
Install and test packed tower
Complete testing - evaluate
recovery from system upset
Disassembly of pilot plant
Projected
11/7/83
11/8-11/83
11/8-11/83
4/30/84
5/1-2/84
5/3-7/84
5/8/84
—
5/9-13/84
5/14-16/84
Actual
1 1 /7/83
11/8-11/83, .
4/24-27/84
4/24-27/84
4/30/84
5/3-4/84
5/5-1 1 /84
-f
5/11-14/84
"§
5/14-16/84
No. of
Personnel
4
4
4
7*
7*
7*
7*
7*
7*
5
* Includes three sampling and analysis technicians.
System upset not conducted due to low initial H2S removal efficiences
* Tests not run
98
-------
H
S
Ei
2
OS
Q
£
H Q)
m V! -P
P5 C5 CH
-H
en • o CM
in j^ 3 -i-1
ft to
C cj M (0
(C S (C ,£
r-l 8 -P 13
ft "O CO CO
? *t *>
0 U o?
H *~
CO ^
3 O
o ^
j rt. in
0) '* '
eg "•>
•H t—
§•
o i ° °
U
o e
CD tn
3 >s >i
Q W iH
•HO O
a
* i N "
"-1 § 5 °
in "^
« M
2 co >i
K S1 .M CN •"
5 00
a
• o s "* °
1-1 -H 3 m
Is °
•* CO
§^ 5 » •*
So* N
^ 14-1 D
00 CO '*
(0 CM \ \
Q ^- m m
o
•H
+J
(0
O
ft
0
rH
(0
e
o
s
in
CM
0
o
o
in
•
m
o
x?
CM
in
CM
in
V^
m
o
•H
-P
(0
M
0)
Cn
o
,—1
(C
g
o
in
•
»-
o
o
o
m
•
oo
in
0
•<*
CM
S
CN
^t
in
_c
o
•a
4J
^
rC
CO
H
O
in
*
*£>
CN
in
•
r«-
0
0
in
CM
•
in
r--
in
r-
VD
CM
xj
CM
*x^
m
e
CO
4J
tn co c
>i H 0
in 3 -H
tn -p
O CO U
•P 0) (0
S-l 0)
CO CM M
3
*c ,c c
&^ -H
CO -H
^ffi §
.-1 -H
(C CO 3
M-l M i-H
3 0
>-l CO CO
co tn
!5 CO 'O •
O S-l 0) r-4.
ft ft C CO
M -H tn
H (1) (0 CO
« > M CO
0 0 T3 >
O
m
CM
M
in
r-
in
o
o
in
•
oo
m
CM
CO
vo
CN
00
^^
m
H
O
•
^j
•H
(8
ft
J_J
M
0)
(C
CO
^3
C C
O S
•H O
•P T3
3 -P
•35
co tn
m
CM
CO
ro
in
^
m
Is-
Is-
m
r-
^*
o
0
o
in
in
*~
o
in
CN
CM
""N^
in
P
3
o
'O
4J
3
J3
CO
H
O
0
VD
ro
in
^
CM
m
*
t*-
o
in
CN
C3
fO
in
CN
,_
CN
r-
CN
0
^X^
in
0 4J
•P 0
1C
C -P
e c
3 0
iH O
O
o -o
•H
•a 3
CO &
U r-(
1C *>»
ft tn
*O CT1
CO
i-H CD
i-H ^
(0 O
•P J-l
tn ft
C E
M -H
0
VD
n
o
in
CN
2
in
VD
in
^
t—
0
m
t-~
'"
00
CM
CN
t-
^s^.
in
(
.§
-H
4J
1C
V4
CO
O-i
• o
iH
(XJ
g
0
0
VD
ro
0
in
CN
2
' 0
in
•^
;P
0
• ^
CN
i
CM
CM
CN
xj-
i
-------
operating parameters affected H2S removal by,the Stretford plant. The opera-
ting conditions maintained during this test program are discussed in detail in
Section 3.3.2.
3.2.4 Unit Operating Conditions
This section gives a brief overview of the Stretford pilot plant oper-
ating conditions. A more complete discussion of these operating conditions is
presented in Section 3.3.2.
A. Inlet Gas Conditions—
The inlet gas flow to the Stretford was initially measured by the use
of a U-tube manometer in conjunction with an Accutube probe placed vertically
in the inlet gas duct upstream of the gas/liquid contactor. The Accutube
probe has two sets of four openings at various distances from the pipe center-
line; one set faces upstream and the other faces downstream. The openings
measure a velocity profile by comparing the high and low pressures, observed
by thes upstream and downstream openings. The differential pressure, in inches
of water, is displayed on the Capsuhelic gauge. Given this differential pres-
sure, the gas flow is calculated by using the pressure reading, barometric
pressure, gas temperature, internal pipe diameter, gas specific gravity, and
an orifice constant.
During the final portion of the program, an S-type pitot tube was
installed in place of the Accutube. The pitot tube was used in conjunction
with both a U-tube manometer and a Capsuhelic differential pressure gauge.
The switch from the Accutube to the pitot tube was required because of the
failure of the Accutube.
The gas conditions encountered at GKI were similar to what was orig-
inally expected with the exception of lower H2S levels. The originally
planned H2S concentrations of approximately 2000 ppm were not encountered
during this, program. The EUS concentrations varied between 718 ppm and
2175 ppm, and averaged 1233 ppm during this program. The inlet gas
temperatures remained within the expected range.
100
-------
B. Plant Process Stream Conditions—
The two process streams of concern in the Stretford process are the
liquor stream and the oxidizer air stream. The flow rates of both streams
were measured by making use of in-line orifice plates to establish a differ-
ential pressure between the upstream and downstream faces. The differential
pressure measured was sensed by a Meriam bellows-type indicator calibrated
with the orifice constant of the plate used. Temperature and pressure are
measured with in-line thermometers and pressure gauges.
The liquor stream characteristics are very important in operating
and diagnosing the performance levels of the system. The solution flow rate
varied from a minimum of 0.5 liter/sec to a maximum of 2.9 liter/sec. The
solution temperature set point was approximately 43°C. This solution
temperature provides a good balance between reaction kinetics and unwanted
byproduct formation (with a corresponding chemical loss).
C. Chemical Concentrations and Additive Rates—
The three primary chemical ingredients required in a Stretford solu-
tion are a carbonate source, a vanadium source, and the 2:7 isomer of ADA.
Sodium carbonate was used as the carbonate source, a 6.3 volume percent sodium
metavanadate solution (ELVAN K) was used as the vanadium source, and ELVADA
was used as the source of the 2:7 isomer of ADA. In addition to these ingre-
dients, optional chemicals can be added to improve some aspect of plant
performance. These optional chemicals included an antifoaming agent to con-
trol the foaming tendency of Stretford solution and a combination flotation
aid/biocide (ELVAFORM). The purpose of the ELAFORM is to control the micro-
biological activity in the Stretford solution and to assist in sulfur froth
formation. All of these chemicals were included in the Stretford solution
used at GKI.
PEI's proposed concentrations of the primary chemicals in the Stret-
ford solution for this test series were Na2CO3 - 25.0 g/liter, ADA - 9.6 g/
liter, and NaVO, - 3.12 g/liter (as vanadium). These concentrations varied
daily as make-up water and chemicals were added. In order to maintain the
desired concentrations of the primary chemicals in the Stretford solution, a
series of chemical analyses were performed on a regular basis. A summary of
101
-------
the chemical analysis test results is presented in Table 16. These results
are discussed in detail in Section 3.3.2.
The pH of the Stretford solution was kept high to increase the van-
adium solubility and H2S removal. The pH level fluctuated during this test
series;, due to the daily addition of sodium carbonate. The pH value ranged
between 12.3 and 7.4 during the GKI test program. This compares to a system
start-up pH of 9.0.
The oxidizer air stream in the oxidizer tank is used to regenerate the
reduced Stretford solution and to float the elemental sulfur into the slurry
tank as a froth. The air also strips the bicarbonate and ammonia from the
solution. The compressor was originally set up to route the air though an
aftercooler to remove the heat of compression. The aftercooler was bypassed
during this test series in order to help maintain the Stretford solution
temperature at the desired level of 43°C. This was thought to be necessary
due to the expected low ambient temperatures.
The oxidizer air flow was varied between 0.017 and 0.042 Sm3/s during
the course of the testing. Previous tests showed that flows over 0.038 Sm /s
scfm were excessive, while flows in the range 0.021-0.038 Sm /s were
acceptable for oxidation purposes. [
When foaming occurred in the oxidizer task (attributed to condensed
oil in the scrubbing solution), the air flow was reduced to correct the
problem. Excessive foaming ultimately caused the recirculation of a solution
that was not totally reoxidized, which also reduced the H2S removal
efficiency. ;
D. Consumption of Utilities—
1 „ Electric—The motor control center of the pilot plant was ;equipped with
a 480-V, 300-amp main breaker indicating a power demand of 144 kW. Actual
consumption, even at full loads, was less than this. Table 17 presents a
breakdown of the current drawn by the various electrical components of the
system during operation. Actual demand was equal to 85.4 kW. Base_d on the
hours available for operating (205 hours), the total electrical consumption
was 17,500 kWh.
102
-------
TABLE 16. SUMMARY OF PRIMARY CHEMICAL ANALYSES
Test High Value Low Value Number of Tests
Specific carbonate, g/liter 49.3 10.6 10
Vanadium^ g/liter 3.5 2.1 ' 5
ADA, g/liter 11.0 4.6 7
pH 12.3 7.4 49
!
Oxidation, rel. mv +90 -87 49
Thiosulfate, g/liter 1.05 1.05 1
103
-------
TABLE 17. ELECTRICAL REQUIREMENT FOR STRETFORD EQUIPMENT
(Amperes)
Equipment
Gas blower
Solution heater
Solution pump
Condensate pump
Pump tank agitator
Slurry pump
Oxidizer agitator
Slurry tank agitator
Air compressor
Control transformer
Heat trace
Totals
Current
Overload
Rating*
60.1
§
9.9
0.7
2.1
1 .4
9.9
1 .7
60.1
15. 0#
15. 0#
Current
Demand When
Operating'''
85.0
51 .0
7.1
1.7
3.1
2.4
13.6
2.4
85.0
14.0
10.0
275.3
Current
Demand-This
Program ;
0
43.0 **
7.1
0
3.1
2.4 ;
13.6
0
85.0
14.0
0
163.3
Current
Breaker
Rating
100
60
20
15
15
15
20
15
100
15
15
390
* Per line — multiply by 1.7 for total three-phase current demand
f 480 V
§ Variable with manually adjusted limit switch '
# Ratings given for circuit breaker
** Time weighted average
104
-------
Electricity was produced on-site by diesel engine powered generators.
Several power failures were experienced during the test. ,
2. Water—Water was added to the Stretford solution to replenish liquid
levels lost because of evaporation and carryover in the outlet gas. The
sulfur slurry (25 percent solids) purge which is usually a source: of water
loss was required only once during these tests. After the initial charge of
9,500 liters, nine water additions were necessary. Total water usage was
about 17,000 liters.
E. Sulfur Production—
The pilot plant was designed to recover an average of 2.6 kg. of
elemental sulfur per hour, or 65 kg per day. The sulfur was collected in the
slurry tank, where it settled to the bottom. The clear liquid layer was
recycled back into the process, and a portion of the slurry was drained into
210 liter drums once during the course of the test program. Precise
measurements of the solids contents were not made. Slurry samples were taken
in a graduated cylinder and the solids were allowed to settle. The solids
content was then estimated by the ratio of the volume of solids in the
cylinder to the total sample volume.
The exact amount of elemental sulfur produced is unknown;' however, it
is estimated that approximately 300 kg. of sulfur were recovered during the
test program. The average values for gas flow rates (in standard: cubic feet
per minute) and the amount of H2S removed (in parts per million by volume)
were determined for each day for which H2S concentration data were available.
The data used to calculate the elemental sulfur production are contained in
Table 18. Based on these numbers and their corresponding durations, the total
amount of H2S removed was calculated. The weight of H^S removed was
calculated and multiplied by the ratio of the molecular weight of'• sulfur to
the molecular weight of H^S to find the amount of elemental sulfur produced.
This equated to a production rate of approximately 1.72 kg/hr or 41 kg/day
(based on a 24-hour day). Thus the estimated sulfur production was only
66 percent of the design value. If the program average gas flow rate had been
at the design level of 0.70 Sm3/s instead of an actual rate of 0.28 Sm3/s, the
design sulfur removal rates may have been met. Lower than expected incoming
105
-------
TABLE 18. DETERMINATION OF ELEMENTAL SULFUR PRODUCTION
Date
1984
5/4
5/5
5/6
5/7
5/8
5/9
5/10
5/11
5/12
5/13
5/14
Total
Average H2S
Concentration,
Inlet
N/A*
1584
1719
N/A
1377
1638
1314
1144
1141
981
1121
— —
Outlet
N/A
447
261
N/A
244
278
248
228
10
131
92
— —
ppmV
A
—
1137
1458
—
1133
1390
1066
916
1131
850
1029
— —
Gas
Flow
gm3/s
0.344
0.253
0.235
0.221
0.476
0.320
0.246
0.281
0.266t
0.251
0.209
— —
AH2S
Flow
Sm3/s
—
0.0003
0.0003
—
0.0005
0.0004
0.0003
0.0003
0.0003
0.0002
0.0002
"- —
AH2S
Flow
kg/hr
—
1.49
1 .77
—
2.76
2.31
1.36
1.31
1.54
1.09
1.13
"^ —
Hours
On Line ;
8.5
24
24
16.75
18.25
15.5
21.25
17.5
24
24
8.75
177. 25§
AH2S
kg
—
35.6
42.4
—
50.8
35.5
28.8
23.2
37.2
26.4
9.7
~~
Sulfur
Produced
kg
— :
33.5
39.9
—
47.8
33.5
27.0
21.9
35.0
24.9
9.1
273.0
* N/A - data not available
t Average of 5/11 and 5/13 data - 5/1 2 data not available
§ Total of hours when H-,S concentration data were available.
106
-------
H-S concentrations also contributed to the lower than expected elemental
sulfur production rates. ;
During this test program, the slurry tank was drained once during the
test 417 liters removed) and again after the test. The sulfur content of the
drained solution was not determined as part of this test program.
3.2.5 Operating Problems
A. Performace of Gas-Liquid Contactor—
On the GKI test a variable throat area venturi was used as a con-
tactor. This replaced the venturi originally supplied with the plant. A
brief description of the two venturi design variations follows in order to
enhance the understanding of the design and intended use of the contactor that
was supplied with the pilot plant. Figure 35 shows the difference between the
more conventional venturi (in which the liquid is dispersed into the gas
stream) and the configuration of a jet venturi scrubber. In the conventional
venturi, the gas enters from the top and supplies most of the powe.r input to
the scrubber. This power is generated by a fan. The liquid is pumped into
the venturi throat through nozzles or is cascaded down the inside of the
contactor. In the throat section, energy is transferred from the gas to the -
liquid to atomize it and create intimate contact. As the combination of gas.
and liquid leaves the throat, some of the energy is regained by the gas
stream.
Although much of this description also fits-the jet venturi scrubber,
the jet venturi acts as an ejector, in that the gas is aspirated into the
venturi by the high-pressure, high-flow liquid stream. The liquid pressures
in jet venturi applications are usually 7000 to 17,500 g/cm .
The design of the contactor that was supplied with this pilot plant
more closely resembles the jet scrubber design. The liquid enters through a.
nozzle at the top, and the gas enters from the side. The gas was designed to
supply the motivating force, with an inlet pressure of one to five psi from a
booster blower. The liquid is supplied from a pump with a maximum -delivery
o
pressure of 3500 g/cm • This pressure distribution is not a representative
one for H2S removal by a jet venturi scrubber.
107
-------
GAS IN
LIQUID IN
THROAT SECTION
GAS AND LIQUID
OUT.
GAS IN
LIQUID IN
SPRAY
NOZZLE
GAS AND LIQUID
OUT
CONVENTIONAL VENTURI
JET SCRUBBER
Figure 35. Conventional venturi versus jet scrubber venturi.
108
-------
As mentioned previously, the gas-liquid contactor used during this
program was a variable throat venturi. When the throat plug was lowered
completely, the venturi throat was free of obstruction. The venturi throat
diameter was 15 cm at this condition. When the throat plug was fully raised,
the throat area was reduced so that the equivalent diameter was 7.5 cm.
The venturi was designed to operate at a gas flow of up to
approximately 0.42 m /s and to receive a gas pressurized to between 105 and
140 g/cm2. Lower than expected pressures may have inhibited gas/liquid
contact causing lower removal efficiencies than previously achieved.
As mentioned previously, a packed tower was added to the reactor
vessel gas outlet near the end of the program. This tower was installed to
enhance the system H2S removal efficiency, which had peaked at about 95 per-
cent while using only the venturi scrubber. This modification gave H2S
removcil rates in excess of 99 percent.
B. Loss of Solution Heat—
The design of this pilot plant included a trim heater for system
startup and temperature maintenance. A continuous heater was deemed
unnecessary because of the high inlet gas temperature (60°C) and saturated
conditions. Even though these conditions were approached at GKI, the evening
ambient temperatures and moisture conditions were such that significant
quantities of moisture evaporated from the system solution tanks. > A cover was
fitted to the pump tank, while the slurry tank remained uncovered. During the
initial portion of the testing, the solution loss averaged about 30 liters per
hour.
The failure of one of the heater elements during the test program
required that the element be bypassed. This reduced the heating capacity by
one-third and resulted in a corresponding increase in the time necessary to
bring the system temperature back to the desired level following the addition
of water to the system. • • '
This shortcoming became apparent following the installation of the
packed tower. When installing the packed tower, the mist eliminator was
inadvertently left out of the system. The lack of the mist eliminator, com-
bined with the installation of the packed tower, resulted in a system solution
109
-------
loss of about 155 liters per hour. The required amounts of cold water makeup
resulted in low solution temperatures during much of the testing with the
i
spray tower. ;
To maintain the solution temperature, on future tests the solution
heater should be replaced with a larger capacity heater to reduce the system
recovery time.
C. Sulfur Flotation—
One of the major problems encountered at GKI was the lack of sulfur
flotation in the oxidizer tank. The system showed good sulfur flotation from
its startup on May 4, 1984 through May 10, 1984. At that time, a major system
upset occurred when the oxidizer began foaming out of control. The foaming
was brought under control by the addition of an antifoaming agent. Once the
foaming was under control, a flotation aid was added to the system. This did
not, however, solve the problem. Various additions of the flotation aid were
tried throughout the remainder of the program, along with varying the oxidizer
air flow rates. None of these changes resulted in improved sulfur flotation..
D. Solution Foaming—
As mentioned above, one of the major problems encountered at GKI was
. - . _ *
excessive foaming in the oxidizer. During the May 10 upset, the antifoam
agent was added at frequent intervals until the foaming was brought under
control. During this time, the oxidizer air flow was reduced in ah effort to
reduce the foaming. In order to keep up with the flow of foam, a larger
pulley was fitted to the slurry pump. This increased the slurry flow rate to
the oxidizer tank from 8 liters/min to 40 liters/min and was necessary to
prevent the slurry tank from overflowing.
The procedures outlined above are not solutions to the problem of
excessive foaming, but rather are only a temporary treatment.
The cause of excessive foaming is not known; however, it is possible
that hydrocarbon (oil mist) carryover in the retort off-gas may have been
responsible. It is believed that oil mist carryover may have occured when the
knock-out drum on the blower skid was completely filled. During that period,
all of the oil normally collected in the knock-out drum would have been
110
-------
carried over to the Stretford plant. (Referring to Figure 4(a), the
horizontal inlet line continues to the Stretford process while a vertical pipe
is 'T'ed off the inlet line and runs vertically down to the knockout tank.
There is no gas flow in the vertical line or through the knockout tank. If
the knockout tank and vertical drain line fill up with condensed liquid, the
remaining condensed liquid will drain into the Stretford process.) This would
explain the sudden nature of the upset. The rate of liquid accumulation in
the knock-out drum was much faster than anticipated. The use of a prequench
or saturation chamber upstream of the gas/liquid contactor(s) might control
this problem by scrubbing out the condensed hydrocarbon before the; gas reaches
the venturi contactor.
3.3 ANALYSIS OF PERFORMANCE
The process streams of the Stretford pilot plant were tested for
selected species to evaluate removal efficiencies and interferences with
process chemistry. Additionally, the influent and effluent emissions were
characterized. This section details the characterization of the gaseous and
liquid components and summarizes the program test results.
3.3.1 Program Test Results
The Stretford pilot plant process streams were tested for selected
species in order to measure removal efficiences, evaluate interference with
process chemistry, compute material balances, and characterize pollutants in
the influent and effluent. The two process streams-of primary concern for
this Stretford test program were the retort off-gas and the Stretford solu-
tion. This section discusses the results of the analyses performed on both of
the process streams. The results of the gas analyses are discussed in the
first part of this section, while the second part contains the results of the
Stretford solution analyses.
Table 19 summarizes the operating conditions maintained during this
test program. This table contains the results of both the gas analyses and
the solution analyses. These gas and solution test results are discussed in
detail in the remainder of this section. Some of the test results presented
111
-------
Jg
§
P5
1
1
o
H
EH
H
H
^H
O
S
g
H
<
i-i
a
§
tn
§
H
Q
1
H
E-l
1
O
§
IT*
S
CO
•
I-I
1
a
**
B
g
C i-*
o n
Tj Tj
(C C g
2 £
o£
S.
o • ""a
.-4^5^
gS8£
• (*1
•£ c v.
flj t*i O*
U w J*
en
C C -S
s&s
S8>
§Ss
** 3 i1
a! £ 0 «
N Z
1 i !£
^ ^ w
4J
S S ° °
H
" " 5 o
S S t. "
,-, * « w
'o 5 B J'
»».
.*>>
S«
is *
V
S
32-ii
1 2-1 !
wo01
3 c 2 g
0.
ill"
U ^ ***B
w b. o: w
« «8
3 *J -H ,
o § S
o a
i
0> «
*J CO
•0 at
a ^~
»
1
o
^H
ffl m
o *J
-H 1- -H
e ti i-
4) •»-> 3
^"S
_ 0 '$
10 -H
•H vOU3CD<->O>I
^ ^ ^ji ^ ^> ^j ^ i ^ r~r*-Nm£-*r.«>2i
.* ,. ...
iitii iiii.««.»»
11111 ' ' ' ' S?SS5*
SSSSiSSSSSoSSSSS
oooo lvooom^'^'<»in*fi^o
wwS2SoS»«gffi-*SPc!2
oooo§2°§oo^2i£i2"i£
ooO'-o«-or»mmf-oooopM
«^ooo(so-g»osssgs
oooooooooooooooo
'
S2s55s2sce-5.ssaa
u
«}
!
•o
V
1
"a
^4 -H
kt U
3 3
4J 4J
C C
S S
u «
>|
5
Q) ^H
U -H
^4 fl
> > C
M «J O
ai -H
a AJ AJ
0 «
jj *J ^
§
-------
-P
o
o
rH
H
•H >H
*J ** *
2S1
SS
IB |
l
§s:&
IQ *J .* S £
* W « ^ «
§ I S 1.2
~H *J -H ~H **
ja 3 a xi 3
O fH V O i-l
w«
Z TJ
4J m
BJ m
u v
^ u
^H o
« b
o a
IIIIIIIIII
i« u>r«- ooiinr-^ *• m*>
OC3OOOOOOOOOOOO
Doddddd ddddddcSddd ooooooooooo
ICOCOODtOCOCOODCOCOaoeOCOCOCOCOGOOOCDoS (DGO COCOCOaiCD CD COCOP-P-P-P-P-P-P-P**
O 4J O S-
w x e
f« -F! _ *
is*
r- r-r^r-p-r- P-
2lSjiUj-j(vi(^(5S(5<%CM{N(M(NCM(*lf^f^nl(^ O»(N rNO)04fN(N CN ^^ °* " " *^ " i.! i-I « X X
3 S S S O S O O O S S O S S O S O O O O O O OOOOO O «5O«3«JOOOOOOtS
oodo'oooddoooo'oooocsoo do ooooo o oooooo'ooooo
55*S5?SS55S*5?-55?333 33
S5555S5SSSPS«SSS?gSaSR S
-f^iP- P-P* P--P* p-r- p^co
p» in m V (n t^
t^ r«- m a* o» m
!!!!!!!!! i ! ! ! ! ! a < !!!!!!
" ^H fl)
5 ££
.3 333335388 -8
W W tfl '
O O O '
oooooo
« § ggig§gg§§§,
3S3SS55iSS535;5553S3i3 55 35553 3 33333S3353S
d d d d d d d d d d d d d d d d d d d d do ddddd d d d d d d d d_d o o
>>>>>>>>>>>>>>>>>>>> >> > > > > > > »>»»»>>
-
o.o oooooooo*-
ss
Q »-
113
-------
C
-H
•P
I
(Ti
g£
e c .
ss
y ^H -, j
01 i, S w
IB
u W = o
s="II
a Si
°> s 3
sis'
tn ^ ^5
3 g| *
o o
5 2 •
18
, . . . . ^
*J*JiliJ4J4JjJ4J*J.H
?««.. '^'l OOOOOOOOOO(8«(0(0(8(a<0 0«
^O^^OjJj} -U*J CCCCCCCCCCbbhhuMbB-Oki
^ o. a a Q, oo «:™«!«!nm«m««55555-'2.'2Suu
•-JOgJJ ^^ 4J*J4J*J4J4J*i4JjJ*J-IM^^^H^I^-IM-rt-H
I « ft)
I ^4 M4 i
.a .o
g
iJ li 41 4*
OOOO
e c e c
4J *J 4J *J
SIQ IB IT)
S o a
S3
(Qc5?iTfOfn*rN ^r^'rii*pmmr*}*r
t*'t**t*'t*fp*r~t*'F+ r*t**r*r*r~t*'t^t*'r*r*'r*t**r*^r*t*'r*t^^ ' ' i*- p- p» p*p*t*-f^r*r^r*r^r^
p^r^r^r-t^Mvv^Vfnrj.
OOOOOOOO O O O O O O O O O O O O O O O O O O O i ! O O O O O O O O O O O O
OOOOOOOO OOOOOOOOOOOOOOOOOOO OOO OOOOOOOOO
r*\Dr-f*p-vD\ovO\
<
o)fMr-M(^v*oiv ^•v^oimf^-eocowo) OOOOCNCNCDO ^i-ci (MOOOMV-V^^
in^rmmu)^M)i£> •vO'£)'£ii>^r^fl'^T^1 ivDH)y3i£>mtnr*co i ir*r-f* r^r^aasoacosco
i f i *n o
*HW'»<*J OOOOOOOOO
114
-------
13
0
4-J
i
,
:*
cidizer
0
o
V)
o
w
o
S
4J
c
m
cn
°
3
m
w
S
Oxidation
Potential
H mv
381
3 ° S
• ""e
•*
0^
g 0 *
•S 1"*
* fcg
1* 3 ,,
^ ° "
» W W
-2 IB J
Eh K W
•H
g _
to u >
= o a
M ° >
i! v x
M a. |
Sg-
° SC-T"
Contct P
Device R
B * Sn
S
3S
« en
Q *-
C
•H
4J
M
" SI
m "4
t i
m m
O 0
•- m
— o
fl fO
fM CM
P- P-
o o
. 0 0
£ m
CO CO
1 t
I 1
« B
L) O
ffl «
r- P-
o o
p* p*
S S
0 0
p* in
0 0
1
o cn
i i
mm
O 0
o o
o o
0 0
c- so
CO
35
P* CM
PJ fM
in CD
CM fM
0 0
cn O
in \c
CD
7
tn
CO
O
o
CO
o
0
fM
9
(S
CO
m
p*
CM
PJ
0
O
~
ap
10
CO
0
m
m
CM
CM
CO
o
0
fM
CO
CO
in
fM
CO
CO
m
CM
CM
in
CM
0
CM
CO
— in
?7
m m
0 0
cn m
fM CM
CM CM
m co
o o
0 0
9 m
CM O
CO
oJS
o *n
m *
S J£
CM in
fM PJ
m CM
fM CM
0 0
fM O
P" CO O
10 xn in
i t i
mmm
000
in in in
o O o
000
m.0 *
i i t
i t t
\o cn m
CM PJ fM
in *r T
•-mo
S m S
fM •- CM
000
Pi CO O
•^ CM rn
fM fM fM
0> 0) O 9) 0) 4> 0)
11 Till 1
cccceceessfisseoe
a)SwSw»»'OO
*j AJ 3 *J *j 4J u a B « mm 01 o M <
itiTTTTiiiiiitiiiiiiiti
ooooooooooooooooooooooo
mm«m«M(nm«fnfMrMCMcnm cncMfMCMCMMfM
"
ooooooooooooooo Jooooooo
ooooooooooooooo ooooooo
«o. «.«»-,--*»«-- «««««,«
C4| 11111 till II II IOOO — *-O--
a»"»88a8aaafla8"saaaa
'^ooaQQaQSoSoouoo^'^^™''''^^
,,o,..o
j °mm^CMCMfMfMfM«^»-^'^»-fMPJfV
o
1
m
CD
o
N
m'
o
0
CO
5:
(N.
PJ
p-
P4
—
o
CM
® 41 « «
H J3 A J3
« IB (8 fl
•H -H -H -H
ii il
32S2
75 M fM
1 1 1
CO O i- m
m o o cn
CO CO CO CO
o o o o
co r- P-* P-
fM CM CM CM
o o o o
0000
«u> * a,
•-moo
fM fM CM P>i
CO ^ O CM
CM •- PJ CM
r- r- P- P-
PJ m m m
Q 0 Q Q
o o o o
§** ^! (N m
o o o
m
0) 0) 0)
« io us
•H >H -H
« fQ <0
g gi
222
2 Q Q
o *r o
7
mmm
o o o
P" \O 10
fM P4 PJ
CM P4 Pi
p* *- ~~
o o o
o o o
fM •- •-
CM PJ (*>
r- M fM
v0 ui ^
CO CO X
CM n m
fM CM *•
CM O \0
CM fM fS!
r- — *-
o e S
1 1
in o m
*9* in vo
o o o
115
-------
•o
I
•H
+J
I
O1
.-I
1
«= 1
23 a
32
i.
.2-
g B 1
o '•> £
-.
8s
si::
'° fc « CO
3=15
ISI*1
853-.*
"£3 =
« S2
H = '> '
° o «
U Q
3
:
,a.a.aa£.a.a£.a
o o o o e
OOOOOOOOOCCCCCC
cccccccccoooooo
2 555
.* -H -H -H -H 5 TJ
*J4J*JAJ*JO)-««
I *» O -I O
i 3 *J •fHbuUkiw *g
-------
I
o
u
m
O it)
4J *J
, o « - u
I"-*-
slit
S3-I1
85,
• ?s^
U£S6
»»s
2 ** •*« .
s §>•
u a
4> ft) SJ 4) «J « ft)
fl «J fl fl flj
oooooooccccccc
ccccceeooooooo
^4 -^ -H ^4 -H -4 -rt
4u)wB(«(«(«t«(ata«(0«««
WWWU1MWWUOOUOUO
I
r* r- ^ *- m r-
mr*»r»\£'£if)inin«9^'mmr>i(N«-'*cor<*\0QOOaio>i cocor-r-»-ooff*
oooooooo
v.*.^..-^.,-,.,. i i
inir>*rvi»f^r^(^-coCT>»-f*O'-f<>^fTr(NeD<-NcNr-ma3ff>coxr-coa)a3a)xo3 socor-oooofyfNfN
!itfiCrt'^^t^H^^^-(^i»HOinirtr*'OOff»ff»oCT> cor^p"C>) _ .- —,.— --
;zzz* ra« «« R3« lefntNCNmrtnfnvn '^'*i'^'*»SJS^;x£v
oocaOuOOOOoO'**"'-'-'-'-'-^1*- ^-».».^.U'inw>e'*'*-
cooDcoiDcocomomxfi
<:«:<<£ irt in ui m in*-^>r»^v-«-oicoaD
'oooooo oooooooooo
inoinOinooOininmooooof4r»p.(*1Hi^i/i(n'"^irtf*l^»-^(N OOOOOOOO'vt<
oooooooooo«~*-«-*-»-»-«-»-'-'-cNr«if«if«) ooooooSoot
I
117
-------
in Table 19 represent averages or estimated data. This was necessary because
the chemical analysis schedule used and the time required to take, a complete
set of gas data.
A. Program Test Results - Gaseous—
The inlet gas samples were collected in a horizontal section of duct
work upstream of the venturi contactor. The outlet samples were taken in a
vertical section of duct downstream of the Stretford's outlet butterfly valve.
The gaseous results are presented in Table 19. The gas flow was
initially measured using an Accutube flow indication device and a, U-tube
manometer. When the Accutube failed, an "S" type pitot tube, used in conjunc-
tion with a Magnehelic gauge, was substituted. Table 20 shows that the daily
average flow rates, expressed on a wet basis at standard conditions (20°C and
760 mm Hg), varied between 0.209 and 0.476 Nm^/s. The program average gas
flow rate was 0.284 Nm3/s. '
The gas exit temperature varied from 37 to 48°C and averaged 41 °C.
The system inlet static pressure varied between 588 and 876 mm Hg, and aver-
aged 648 mm Hg. The static outlet pressure averaged 628 mm Hg; the pressure
readings varied between 506 and 841 mm Hg.
The daily reduced sulfur emission data are summarized in Table 21.
These data were collected according to the general procedures set forth in EPA
Methods 15 and 16. These methods call for the use of an on-line, semicontin-
uous sample extraction system, a dynamic dilution system and a gas chromatog-
raph with flame photometric detector (GC/FPD). The details of these test
methods are presented in the appendix.
The major indicator of performance during this test series is the
H2S removal efficiency. The large increase in H2S removal efficiency between
May 11 and May 12 (Table 19) was due to the installation of the packed
tower. Prior to the installation of the tower, the average recorded H2S
removal efficiency was 80 percent, while the maximum recorded H2s removal
efficiency was 95 percent. In comparison the recorded H2S removal-efficiency
averaged 93 percent following the tower installation. The maximum H2S removal
efficiency measured during this time was greater than 99 percent.
118
-------
TABLE 20. SUMMARY OF RETORT OFF-GAS CONDITIONS
Date
1984
5/4
5/5
5/6
5/7
5/8
5/9
5/10
5/11
5/12
5/13
5/14
Overall Average
Volumetric
Flow Rate
Nm3/S*
0.344
0.253
0.235
0.221
0.476
0.320
0.246
0.281
DOS
0.251
0.209
0.284
Temperature °C
42
45
42
42
42
39
41
37
38
39
36
41
Absolute
Pressure
mm Hg^
, 630
623
626
645
689
701
717
614
623
626
625
I 648
Measured at Stretford inlet, reported at standard temperature (20°C) and
presisure (760 mm Hg).
t Outlet temperature reading given, inlet temperature indicator OOS
§ Inlet pressure
119
-------
TABLE 21. REDUCED SULFUR SPECIES EMITTED** (ppm)
H2S Concentration
Date
1984
5/5
5/6
5/7t
5/8
5/9
5/10
5/11
5/12
5/13
5/14
Inlet
Avg
1584
1719
—
1377
1638
1314
1144
1141
981
1121
Range
1322-1730
776-2165
—
1367-1898
1398-1935
1245-1761
1015-1253
953-1249
718-1125
1091-1137
Outlet
Avg
447
261
—
244
278
248
228
10
131
92
Range
385-693
16-559
—
75-395
188-343
235-301
190-240
6-15
7-140
14-138
COS MeSH
Concentration Concentration
Inlet
45
190
—
N.D.
35
N.D.
26
99
88
94
Outlet Inlet
59 N.D.*
36 N.D.
—
54 N.D.
82 N.D.
53 N.D.
52 N.D.
72 N.D.
76 N.D.
79 N.D.
Outlet
N.D.
7
—
5
4
N.D.
N.D.
14
18
18
* N..D. = none detected, minimum detectable level
t No data available for 5/7 due to sample system problems
** GC/FPD measurements
120
-------
Table 19 shows two cases (May 6 and May 8) where H_S removal effici-
encies exceeded 90 percent before the installation of the packed tower. In
both cases, the high H2s removal efficiencies were maintained for only a short
time. A review of both the unit log and operating data provided no explana-
tion for the high H2S removal efficiencies. Following the installation of the
packed tower on May 11, Table 19 shows two sets of H2S removal efficiency
data. The first set of data, taken on May 12, shows about ten hours of
operation with H2S removal efficiencies in excess of 98 percent. These data,
taken when operating with maximum solution flow to the packed tower and the
venturi throat plug in the fully closed position (i.e., 46 cm^ throat area),
are believed representative of the H2S removal efficiencies that can be
sustained by the Stretford plant when operated with the venturi contactor and
the packed tower. The second set of H2s removal data, taken during May 13 an
14 using the same venturi throat area, show H2S removal efficiences ranging
between 83 percent and 98 percent. The first 13 data points were taken while
operatting with no solution flow to the packed tower. These data show that the
H2s removal efficiency reached an equilibrium value of about 88 percent for
operation without the packed tower. The sudden increase in H2S removal
efficiency from 88 percent to 95 percent was caused when the solution flow to
the packed tower was restarted. The gradual increase in characteristic
removal efficiencies after this point were due to gradual increases in the
solution flow rate to the packed tower.
The lower than expected H2s removal efficiences measured prior to
the installation of the packed tower were probably due to several factors.
The primary factor was probably the low residence time that the solution was
allowed in the reaction vessel. Previous data have shown that a residence
time of approximately 15 minutes is necessary during the treatment of lean
(300-500 ppm) H2S streams. Attempts were made during the program to maintain
a minimum solution residence time of 15 minutes by controlling the solution
flow rate and reaction vessel level. Unfortunately, this was not! always
achieved.
i
The lack of solution residence time in the reaction vessel causes
problems to occur with the reactions involving the formation of sulfur
particles from the HS~ radical and the reduction of the vanadium to its
121
-------
valence of four. If the NaHS is not completely reacted before it is trans-
ferred to the oxidizer, the formation of thiosulfate (a stable unwanted
byproduct) increases greatly and lessens the opportunity for the yanadate/ADA
reaction to be completed. This incomplete reoxidation of the main chemicals
before; recycling back to the pump tank causes numerous problems. One of these
is an attempt to oxidize the NaHS molecule with an already reduced vanadium
molecule. This lowers the efficiency and increases the chance of thiosulfate
formation. The unwanted cycle is self-perpetuating.
Another problem is that the final conversion to elemental;sulfur can
form elsewhere in the system (in the piping and pump tank), where it will
become a suspended solid. This is known to have been a problem during this
test program. While draining the system at the end of the program, consider-
able sulfur deposits were found at the bottom of the pump tanks in addition to
deposits at the bottom of both the reaction vessel and the oxidizer tank.
Another limiting factor was in the contact between the retort off-gas
and the Stretford liquid. This became obvious following the installation of
the packed tower. As noted previously, a large gain in H2S removal efficiency
was noted following the installation of the tower. Since the tower serves
only to increase gas/liquid contact time and area, it follows that the contact
time and area were the limiting factors in H2S removal efficiences.
The removal efficiences for both the carbonyl sulfide (COS) and methyl
i
mercaptans (MeSH) were negligible during this program. The variation of a few
parts per million at the measured levels of inlet concentrations can be
explained by the limitations of the sampling and analysis procedures.
B. Program Test Results - Liquid—
The Stretford solution is a dilute solution of sodium carbonate
(Na2CO3), sodium metavanadate (NaVO3), and sodium salts of the 2:6 and 2:7
isomers of anthraquinone disulfonic acid (ADA). These chemicals are referred
to as the primary chemicals. The Stretford solution is intended to be main-
tained at a temperature of 110°F and a pH of 8.5 to 9.5. ;
Due to daily solution loss resulting from evaporation and carryover
with the retort gas, it was necessary to add water and primary chemicals on a
daily basis. The daily makeup rate for the primary chemicals is summarized in
122
-------
Table 22. The daily makeup rates were based on measured solution concentra-
tions, the liquid inventory, and the design feed gas conditions. Table 22 also
indiccites the consumption of ELVAFORM (a combination biocide flotation aid)
that was added to control aerobic microbial growth and to assist in sulfur
flotation.
In order to maximize I^S removal efficiencies, it was important to
maintain the proper concentrations of the primary chemicals. A daily routine
of chemical analyses was established in order to accomplish this. Table 23
presents the complete results of the chemical analyses performed during this
test program. The analyses performed included the following:
pH
. oxidation level
. sodium carbonate
ADA
. vanadium :
. thiosulfate
The results of the pH tests are plotted versus time in Figure 36.
This figure shows both the individual pH data and the daily average pH. The
individual pH data show large variations between consecutive tests in many
instances. The daily averages show that the pH was only in the desired range
of 8.5 to 9.5 during four of the nine test days. During the remaining five
days, the average pH was above 9.5. As mentioned in Section 3.1.1, the rate
of H2S absorption is pH-dependent; as the pH level decreases below the design
levels;, the H-S removal efficiency decreases. Thus, it appears that operation
with the pH in excess of the desired range would have had no adverse effect on
the H0S removal efficiency.
In Figure 37, the primary chemical concentrations as determined by
laborcitory analysis are plotted versus time.
The purpose of the sodium carbonate {Na-CO,) in the Stretford solution
is to react with the incoming H^S to form sodium hydrosulfate (NaHS) and
sodium bicarbonate (NaHCOg). Thus, low levels of Na-CO^ would result in
reduced H-S removal efficiency. The carbonate concentration during this test
123
-------
TABLE 22. SUMMARY OF CHEMICAL USAGE
DURING STRETFORD TESTING
Date
1984
5/3*
5/4
5/5
5/6
5/7
5/8
5/9
5/10
5/11
5/12
5/13
5/14
Total1"
Na2C03
kg
226
45
0
23
33
136
23
23
54
0
130
59_
752
ADA
kg
111
22
16
0
45
10
11
61
25
0
39
0_
340
NaVOj
(EL VAN K)
kg
90
18
9
23
13
0
9
0
12
0
39
12
225
H2°
liter
9,304
1,804
1,137
0
1,308
868
902
803 '.
1,270
0
5,264
3,676
26,336
ELVAFORM
liter
7.6
0.99
0.99
0.99
0.99
1.2
1.5
0.99
1.5
2.0
0.99
0.00
19.7
* Indicates initial start-up charge.
"*" Includes chemicals remaining in system at completion of programs
Note: antifoaming agent of less than one gallon was added during the test.
124
-------
TABLE 23. CHEMICAL ANALYSES RESULTS
Date
1984
5/5
5/6
5/7
5/8
5/9
pH
9.93
10.00
9.61
9.32
10.84
9.86
9.78
7.00
9.45
9.38
9.67
10.70
9.30
9.30
9.49
12.28
9.60
9.95
9.50
9.45
9.51
7.35
7.40
9.66
9.60
9.44
8.60
9.94
Oxidation Sodium
Level Carbonate ADA Vanadium . Thiosulfate
rel MV g/liter g/liter g/liter g/liter
-81 22.8 7.6
-6 to -10
+75 to +90
-10 to -11
+20 to +55
-21
+5 11.6 — 2.1
0
-31 :
-25
+46 to +50
-16 ;
+1 15.2 7.76
0 to -10
-15.5 to -17.5
-51
-32 ;
-50
+27 14.4 — 3.3 '
-32 ;
-14
-45 I
-21
-21 31.3 4.60
+4
-60
-87
-75
125
-------
TABLE 23. (continued)
Date
1984
5/10
5/11
5/12
5/13
PH
9.63
9.53
9.30
9.22
9.53
9.43
10.12
9.44
9.31
9.93
12.07
9.52
9.43
9.34
9.32
12.30
9.87
9.34
9.42
9.33
9.31
10.43
10.88
Oxidation Sodium
Level Carbonate
rel MV g/liter
-25 30.9
-10
-52
-66
-42
-55
-37 26.7
+4
-19
-25
-50
-26 32.6
-4
+55
-15
-17
-32
+5 31.7
-1
-5
-no
-28
-13
ADA Vanadium Thiosulf ate
g/liter g/liter g/liter
5.4 3.5
s.s — :
3.3
7.6 — 1.1
5/14
11.0
2.9
126
-------
Daily average pH
5/5 5/6
5/8
5/9 5/10
Date 1984
5/11 5/12 5/13 5/14
Figure 36. Stretford Solution pH vs. time
127
-------
35
30
25
20
15
0)
4J
•H
cr>
.3
4J
(C
c
0)
u
I 10
Target carbonate
concentration
25.0 g/1
Carbonate
I I Vanadium
ADA
0
Target vanadium
concentration
3.12 g/1
I I
I
I I
5/5 5/6 5/7 5/8 5/9 5/10 5/11 5/12 5/13 5/14
Date 1984
Figure 37., Primary Chemical Concentrations versus Time.
128
-------
series varied from 11.6 to 32.6 g/liter and averaged 24.1 g/liter. In compar-
ison, the design carbonate concentration was 25.0 g/liter based on an inlet
H2S concentration of 2,000 ppmV. The carbonate concentration was well below
the desired level of 25.0 g/liter during the first four days of testing,
averaging only 16.0 g/liter. During the remainder of the program, however,
the carbonate concentration averaged 30.6 g/liter well above the design
level,. The large increase in carbonate concentration seen in Figure 37
between May 8th and 9th corresponds to a large carbonate addition ,on May
8th. However, the increase in carbonate concentration was not reflected in
the solution pH, as would have been expected nor did it affect the removal
efficiency. :
The function of the sodium metavanadate (NaVO,) in the Stretford
solution is to react with the NaHS formed by the .reaction between H2S and
Na2CO.j, producing elemental sulfur. Thus, low NaVOg levels would inhibit the
formation of elemental sulfur in the solution during this test program, the
vanadium concentrations varied between 2.1 and 3.5 g/liter and averaged 3.02
g/liter. The desired vanadium concentration was 3.12 g/liter, based on
expected inlet gas H2S concentrations. Figure 37 shows that the vanadium
concentration was the most stable of the primary chemical concentrations
during this testing.
The ADA's function in the Stretford solution is to regenerate the
NaVOg,. The ADA is regenerated by oxygen in the oxidizer tank. The ADA con-
centreitions varied from 4.6 to 11.0 g/liter during this program as shown in
Figure 37. The average ADA concentration was 7.5 g/liter, compared to a
desired concentration of 9.6 g/liter. As was the case with the carbonate data,
large increases in ADA concentration could be traced to large ADA additions on
the previous day.
3.3.2 Summary of Findings
During its second field test at the Geokinetics Kamp Kerogen shale oil
retort site, the Stretford test program was run in two distinct parts. The
two program parts were as follows:
Testing with only one gas-liquid contacting device (the
variable-throat venturi scrubber)
129
-------
. Testing with two gas-liquid contacting devices in serijes
(the venturi scrubber and a packed tower)
The Stretford was operated for a total of 205 hours during a twelve day period,
between May 3, 1984 and May 14, 1984. Operation with the venturi scrubber
accounted for 142 hours of operation. The remaining 63 hours of operation
were with the venturi scrubber and the packed tower.
The H2S removal efficiency averaged 80 percent during the initial
portion of the testing. During this time, a maximum H-S removal efficiency of
95 percent was achieved on two separate occasions, but documented H-S removal
efficiencies in excess of 90 percent were maintained for only 5 hours. For the
remainder of this portion of the program, the H2S removal efficiencies
remained in the 80-90 percent range. Operating changes designed to increase
the I^S removal efficiency (i.e. decreasing the venturi area and increasing
the solution residence time in the reaction vessel) did not seem to cause a
significant increase in H2S removal efficiency. The available data give no
clue as to why the H2S removal efficiencies peaked and dropped on two separate
occasions; the lack of continuous H2S removal data is the limiting factor in
the ability to interpret the data.
During the second part of the program, when operating with, the venturi
scrubber and the packed tower, the H2S removal efficiency averaged;93 percent.
This number would have undoubtedly been considerably higher if the system had
been operated continuously with the venturi area set at its minimum and
maximum solution flow to the tower. When this was done (the latter half of
May 12), the H2S removal efficiency averaged 99 percent over a period of
10 hours. During the period between 4 pm on May 13 and the end of the
program, the solution flow to the spray tower was deliberately shut down then
restarted. This was done to allow the system to reach a steady-state
condition while operating in its original configuration. Once this condition
was achieved, and a steady-state H2S removal efficiency of about 88 percent
was achieved, the solution flow to the spray tower was restarted, increasing
H2S removal efficiency as the solution flow rate to the spray tower was
increased.
130
-------
The concentrations of the primary chemicals varied significantly
during the course of the program. The carbonate and ADA concentrations varied
the most, while the vanadium concentration remained relatively constant.
Four major problems were encountered during this test program, as
listed below:
. inadequate venturi contactor performance
. lack of sulfur flotation
. excessive solution foaming (for a few hours)
. contamination of sampling and analysis system
The performance of the original gas-liquid contactor (the variable-
throat venturi scrubber) was never up to the expected levels of 96-99 percent
H2S removal. It is believed that the reason for this was the low liquid
pressures to the scrubber, which resulted in poor atomization of the Stretford
solution. The installation of the packed tower dramatically increased H2S
removal efficiencies.
The system showed good sulfur flotation from startup on May 4, 1984
until May 10, 1984. At that time, a major system upset occured when the
oxidiaer began foaming out of control. Following the upset, the system showed
poor sulfur flotation for the remainder of the program. While the cause of
the foaming was not determined with any degree of confidence, it was most
likely due to contamination of the Stretford solution by oil carried over in
the retort off-gas. i
Clogging of the sampling and analysis system resulted in large gaps in
available H,? removal efficiency data, as shown in Table 19. The clogging was
caused by liquid and solid particulate matter in the offgas finding its way
into the sample system. The most frequent points of clogging were the preci-
sion valves used to control the dilution air flow. These had to be
disassembled and cleaned frequently until improved mist knockouts were
developed.
Excessive solution foaming was a problem only during the above-
mentioned upset. The problem was brought under control within a few hours
using an antifoaming agent.
131
-------
SECTION 4.0
QUALITY ASSURANCE
A formal quality assurance (QA) program was conducted for this test.
Separate quality assurance project plans were prepared and approved by EPA and
KVB project directors, project directors from PEI and MRC, and their
respective QA officers. These QA plans defined the test objectives, sampling
and analysis procedures, calibration procedures and frequency, sample custody
procedures and management responsibility. This section presents certain data
that will indicate the degree of error associated with the reported data.
4.1 GAS SAMPLING
Samples of retort offgas were taken at three locations, upstream from
both the Stretford and the alkaline scrubber and downstream from each of those
units. Daily calibration checks were performed. Each instrument was
calibrated using certified gases of known concentration. Often three
concentrations were used to establish a calibration curve. After the
instrument calibration, recovery checks were made on the sampling lines by
drawing calibration gas through the full system.
An example of one day's calibration of the gas chromatograph with a
flame photmetric detector (GC/FPD) is presented as Table 24.
Each day after calibrating the GC/FPD and the continuous total reduced
sulfur (TRS) monitor, recovery checks were made. First, the three trains were
checked for leaks. Following the leak check, H2S calibration gas was
introduced to the sampling probe with excess flow; 1000 ppm H2S was used for
the inlet and 100 ppm H2S was used for the outlets. The calibration gas was
pumped through the entire sampling system and diluted. The dilution rate was
measured with a bubble tube so a dilution factor could be established.
132
-------
CN
1
10
M
c
•H
tt)
23
DI
1
^
CO
X
15;
i
CM
CO
r_>
a
a
it
c
r-
ir
r-
r-
CT
a
c
c
JQ
£
Cu
y
K
rH
CO
§ §
M W
•ft CJ
M H
a Ct<
I§
CO
rH
cn
en
cn
o
CO
cn
cn
o
rH
VO
r-
• cn
cn
o
in
in
CO
cn
cn
o
™
H
EH
5
M
8
i
5
'"'N*
W
p
Q
2
ja_
§ § *
uo-
*
1
CO
t
o
CM
rH
(tf
1 u "i
E 8 •— '
a
s
^
1
*iH
rH
^
ui
•o
c
j:
4J
0
o
c
H
*
o
CO
i
£
CO
cc
T*
^*
cc
cn
CO
co
in
CO
CO
CO
CO
I
o
CO
s
"i
c
H
a
i
i
c
c
a
if
CO
1
5
cc
S
CM
CO
VD
r-
VD
O
VD
CM
VD
CO
r-
CO
8
ac
CO
0
1
VD
in
a
•s
c-
If
f
0
£
'-'
CO
in
M1
en
in
CO
CO
CO
(*
in
^
n
O
4.
(t
V,
ae
in
o
I
o
co
3
,
t-
r^
rH
en
in
rH
cn
CO
o
CO
CM
S
0
1.
CM
f
rH
CSI
C\
rH
CO
o
CM
in
in
CM
CO
CM
CO
£
s
a
o
•H
^t
co
cr
U"
rH
It
rj
•O
C
K
*
CO
CO
+
r-
CM
•^
VD
CO
I-
p^
VD
CO
^
CM
VD
CO
•*
VO
VD
CO
^t
o
[^
co
CM
X
e
T
cn
VD
1-
cc
r
vO
cn
a
•-;
t*
VO
f
c
c
r
Cl
so
o
in
CM
^
O
CO
en
r*-
o
CO
cn
o
o
CO
rH
cn
o
CO
CM
CM
r^
•s.
c
o
in
a
VD
o
CO
rH
O
CO
o
CM
CO
CO
in
m
cn
VD
CN
CO
u
cn
CO
CM
r-
cn
CO
CM
CO
in
CO
CM
cn
rH
cn
CM
CM
cn
VD
g
0)
.c
a
H
dp
d
CO
CM
in
CO
CM
in
CO
••»
CO
CO
in
CO
VD
CO
CM
CO
CM
X
dP
O
§
M
cn
c
rH
O
r-
VD
VD
O
*""*
en
CO
o
rH
o
o
CO
8
e
O
O
c
CM
C
co
c
cn
cn
CO
VD
C
^
cn
CM
CQ
"
dP
O
C
VD
cs
CN
rH
r-
co
rH
VD
rH
O
o
rH
VD
CM
U
CO
in
rH
U*
in
^
in
CO
rH
in
rH
CO
CM
(U
C
£
a
|
o
its.
ID
M
I
V
a
co
en
133
-------
After calculating the dilution factor the recovery gas was sampled by
the GC/FCD and the continuous TRS monitor. The analytical response was
applied to the respective calibration curve and translated to ppm at the
instrument. The instrument ppm was multiplied by the dilution factor to
obtain ppm at the stack. This calculated stack value should equal the
calibration gas cylinder concentration for a 100 percent efficient recovery.
Whenever the recovery check yielded a lower value than expected, the sampling
train was checked for gas concentration at various points in the system. Due
to the reactive nature of H2S, degradation was a problem. Once the sampling
system had been corrected and the gas flow had equilibrated, recovery checks
ususally improved. ,
A recovery factor was calculated from the check and applied to the
data for that day. ;
ppm analyzed x dilution factor x recovery factor = ppm stack
4.2 WATER DATA
A key element in the scrubber analysis was the carbonate, bicarbonate
and hydroxide concentrations. To provide an indication of the accuracy of
these laboratory values as well as the other water analysis, a series of
control samples not identified to specific runs were submitted for, analysis
along with the primary samples. The results of these control samples are
compared to those for the primary samples in Table 25.
The average differences between the sample and control values for
carbonate, bicarbonate, and hydroxide are summarized below.
Average Difference, %
carbonate 16.3
bicarbonate 27.6
hydroxide 21.1
Total Alkalinity 9.0 ;
The carbonate and bicarbonate values have average differences of 16.3
percent and 27.6 percent respectively, based on the absolute value of the
134
-------
TABLE 25. WATER ANALYSIS DATA QUALITY CONTROL
Run No.
14
30
18
17
29
19
15
27
25
21
22
15
27
25
21
22
24
16
12
28
26
20
13
Quality
Control
J
O
D
M
QQ
F
A
DD
AA
G
C
A
DD
AA
G
C
S
H
P
TT
RR
U
SS
Deviation
Component Sample
Sulf ide, mg/1 400
310
400
Sodium, mg/1 1 70
180
1,000
Ammonia, mg/1 1,200
370
400
260
300
Total Organic Carbon 280
230
260
250
160
Alkalinity, mg/1 as CaCO3 2,200
20,300
96,000
2,200
8,300
2,300
84,000
Control
460
330
400
170
170
970
1,700
540
398
380
344
230
290
280
330
• 160
2,100
19,000
120,000
2,500
7,500
2,200
84,000
S-C
+60
20
0
! o
10
30
500
170
2
120
44
50
60
20
80
0
100
1 , 300
24,000
300
800
100
0
-
%
15.0
6.5
0.0
7.1
0.0
5.5
3.0
2.8
41.7
45.9
0.5
46.2
14.7
29.8
18.0
26.0
7.7
32.2
0.0
16.8
4.5
6.4
25.0
13.6
9.6
4.3
0.0
9.0
(continued)
135
-------
TABLE 25. (Continued)
Run No.
24
16
12
28
26
20
13
24
16
12
28
26
20
13
24
16
12
28
26
20
13
24
16
12
28
26
20
13
Quality
Control Component
S Carbonate
H
P
TT
RR
U
SS
S Bicarbonate
H
P
TT
RR
U
SS
S Hydroxide
H
P
TT
RR
U
SS
S Sulfate
H
P
TT
RR
U
SS
Deviation
Sample
840
6,200
19,000
840
3,000
600
14,000
980
<1
<1
980
4,000
1,600
<1
<1
1,700
11,000
<1
<1
<1
10,000
120
66
160
110
100
100
33
Control
960
6,700
13,000
800
3,200
720
11,000
610
<1
<1
1,300
3,400
1 , 200
<1
<1
2,300
13,000
<1
<1
<1
11,000
110
98
120
57
98
81
51
S-C
-120
6,000
40
-200
120
4,000
370
-320
600
1
-600
-2000
1,000
f
10
32
40
53
2
19
-V8
%
-14.3
-8.1
31.6
4.8
-6.7
-20.0
28.6
16.31
37.8
-32.7
15.0
25.0
27. V
35.3
-18.2
10.0
21 .1
8.3
48.5
25.0
48.2
2.0
19.0
54.5
29.4
(continued)
136
-------
TABLE 25. (Continued)
Run No.
24
16
12
28
26
20
13
24
16
12
28
26
20
13
24
16
12
28
26
20
13
Quality
Control Component
S Total Dissolved Solids
H
P
TT
RR
U
SS
S Total Suspended Solids
H
P
TT
RR
U
SS
S Total Solids (mg/1)
H
P
TT
RR
U
SS
Sample
37,000
80,000
130,000
45,000
32,000
22,000
130,000
120
150
150
83
74
57
180
37,000
80,000
130,000
45,000
32,000
22,000
130,000
Deviation
Control S-C
1 , 700
690
610
1 , 700 Error
3,500
1,200 ;
580
70 50
66 84
35 115
25 58
18 56
8 49
51 129
1 , 800
760
650
1 , 700 Error
. 3,500
1,200
630
%
41.7
56.0
76.7
69.9
75.7
86.0
71.7
68.2
,
137
-------
deviation. However, the total alkalinity average difference is only 9.0
percent. This is because the carbonate and bicarbonate are determined on the
sample titration. A shift in the endpoint determination results in a higher
carbonate value and a lower bicarbonate value or vice versa.
Evaluating the deviation for individual runs results in the following
Run
24
28
26
20
Quality
Control
S
TT
RR
U
Carbonate
% Deviation
-14.3
4.8
-6.7
-20.0
Bi carbonate
% Deviation
37.8
-32.7
15.0
25.0
Total
% Deviation
23.5
-27.9
8.3
5.0
16.2
The average error for the total carbonate-bicarbonate is 16.2 percent.
To evaluate the effect on selectivity of these variations in
carbonate-bicarbonate values, the selectivity for these runs was recalculated
based on the control data concentrations. These results are shown in
Table 26.
TABLE 26. COMPARISON OF SAMPLE & CONTROL WATER DATA
EFFECT ON SELECTIVITY
Run #
20
18
16
24
26
Contactor
Tower
Tower
Tower
Venturi
Venturi
Chemical
NaOH
NaOH
NH4OH
NaOH
KOH
Sample
% CO9
1 .0
3.2
10.4
.79
3.42
Selectivity
51 .8
28.7
8.9
60.4
20.8
Control
% CO?
.93
4.57
7.06
.58
3.43
Selectivity
55.7
19.8
13.1
83.0
20.7
Error
-3.9
8.9
-4.2
22.6
_ .1
7.9
138
-------
With the exception of run No. 24, all of the error values are less
than ten percent indicating reasonable agreement and accuracy for the
calculated selectivty values.
The control samples for total Dissolved Solids is obviously in error.
As these values do not directly affect the results, this discrepancy
was not investigated.
The control samples for the remaining components with the exception of
Total Suspended Solids showed reasonable agreement with the original samples.
139
-------
REFERENCES
1. Aiken, R., unpublished data and correlation (1976). ;
2. Aiken R., et al., "Selective Absorption of H2S From Larger Quantities
by Absorption and Reaction in Fine Sprays," AICHJ; Vol. 29, No. 1,
January 1983, p. 66.
3. Astarita, Gianni and Gioia, Franco, "Industrial & Engineering
Chemistry Fundamentals," Vol. 4, No. 3 August 1965, p. 317.
4. Astarita, Gianni and Gioia, Franco, "Industrial & Engineering
Chemistry Fundamentals," Vol. 6, No. 3, August, 1967.
5. Astarita, Gianni and Gioia, Franco, "Chemical Engineering Science",
Vol. 19, 1964, pp. 963.
6. Cooney, D.O., "Modeling Venturi Scrubber Performance for H^S Removal
from Oil-Shale Retort Gas," presented at AIChE Annual Meeting,
Washington, D.C. (1983).
7. Ctvrtnicek, T.E., "EPA Scrubber Trailer Operating Procedure", EPA
Contract No. 68-03-2784, April, 1984.
8. Danckwertz, P.V., "Gas-Liquid Reactions," McGraw, Hill, New York, 1970
9. Denver Research Institute, April, 1983, Pollution Control Technical
Manual: "Modified In-Situ Oil Shale Retorting Combined with Lurgi
Surface Retorting", EPA Report # EPA-600/8-83-004.
10. Denver Research Institute, April, 1983, Pollution Control Technical
Manual: "TOSCO II Oil Shale Retorting with Underground Mining", EPA
Report # EPA-600/8-83-003.
11. Desai, B.O., Day, D.R., and Peters, J.A., "Air Pollution .
Investigations of Oil Shale Retorting: In-Situ and Surface, Task 1:
Evaluation of Sulfur Removal Technologies", EPA Contract No.
68-03-2784, February, 1983.
I
12. Edwards, T.J., Mauer, G., Newman, J., and J.M. Pransmitz, "Vapor-
Liquid Equilibria in Multicomponent Aqueous Solutions of Volatile Weak
Electrolytes", AIChE J 24, 966 (1978).
13. Gupalo, Y.P., and Ryazantser, Y.S., Chem. Eng. Sci. 27, 61 (1972).
14. Ingebo, R., "Drag Coefficients for Droplets and Solid Spheres in
Clouds Accelerating in Air Streams, NACA Tech Note 3762 (1956).
140
-------
15. Lekas, James, 1984, Personal Communication with Geokinetics, Inc.,
Salt Lake City, UT.
16. Lovell, R.J., Dylewaki, S.W., and C.A. Peterson, "Control of Sulfur
Emissions from Oil Shale Retorts", EPA Report 600/7-82-016, NTIS PB82-
231945, (April 1982).
17. Mason, D.M., and R. Kao, "Correlation of Vapor-Liquid Equilibria of
Aqueous Condensates From Coal Processing," presented at symposium on
Thermodynamics of Aqueous Systems, Washington, D.C., October, (1979).
18. Nomhebiel, G., "Gas Purification Processes for Air Pollution Control,"
Newnes-Butterworth, London, 1972.
19. Onda, K., Takeuchi, H., et al., "Journal of Chemical Engineering of
Japan," Vol. 5, No. 1, 1972. !
20. Perry, J.H., ed., "Chemical Engineers' Handbook," Fourth edition,
McGraw-Hill, New York (1963). |
21. Perry, R.H., "Chemical Engineers" Handbook, McGraw Hill, NY, 5th
Edition, 1973.
22. Prohocs, S, "Control of Kraft Recovery Furnace Emissions with
Simultaneous Heat Recovery using the PPRIC/BCRC Scrubbing Process",
Pulp and Paper Research Institute of Canada, Miscellaneous Report No.
MR37, July, 1983. i
23. Stern, A.C., "Air Pollution", Vol. IV, Academic Press, New York, 1977.
24. Uchida, S. and Wen, C.Y., "Gas Absorption by Alkaline Solutions in a
Venturi Scrubber," Ind. Eng. Chem. Process Des. Dev. J_2^ 437 (1973).
25. Van Krevelen and D. W. Hoftijzer, "Composition and Vapor Pressure of
Aqueous Solutions of Ammonia, Carbon Dioxide, and Hydrogen Sulfide,™
Recueill 68 191 (1949).
141
-------
APPENDIX A
SAMPLING AND ANALYSIS METHODOLOGY ;
A.1 GAS STREAMS '
Two sampling and analysis (S&A) methods were used to determine the
retort offgas composition during the test program run at GKI, instrumental and
wet chemical. The majority of the tests were run using the electronic
instruments described in Section A.1.1. The ammonia concentrations were
determined using the wet chemical analysis method described in Section A.1.2.
A.1.1 Instrumental S&A
There were essentially three separate sampling and analysis systems
used during this test program. One system was used to measure specific
reduced sulfur compounds; this system is described in Subsection A. Another
system provided a continuous, real time measurement of the total organic
sulfur in the gas stream; this is described in Subsection B. The third
system, used to measure the non-sulfur gas components, is described below in
Subsection C.
Ao Sampling and Analysis for Specific Reduced Sulfur Compounds—
The sampling and analytical procedures that were used for the reduced
sulfur compounds are essentially those specified in EPA Methods 15 and 16 of
the Federal Register.* The method employs a gas chromatograph (GC) with a
flames photometric detector (FPD). In this procedure, a continuous gas sample
is extracted from the emission source, scrubbed in a cold SO2 scrubbing
solution, and diluted with clean dry air. An aliquot of the diluted sample is
then analyzed for the following sulfur compounds: hydrogen sulfide (H2S),
carbonyl sulfide (COS), carbon disulfide (CS2), methyl mercaptan |(MeSH), and
thiophene.
* 40 CFR 60, Appendix A, Reference Methods 15 and 16, July 1, 1982.
A-1
-------
The sampling system, shown in Figure A-1, consisted of stainless steel
probes, Telfon SO2 scrubbing systems, utilizing a citrate buffer solution,*
Telfon sample transfer lines, a dilution unit, GC-FPD, an integrator, and a
calibration gas source. The samples were collected semicontinuously from three
points; the combined inlet to both the Stretford and the alkaline scrubber and
the respective outlet from each process.
The GC used was a Perkin-Elmer Model 990 with an FPD. This GC is
equipped with a 10-port valve for automatic injection of the sample from the
sample loop and for backflushing a precolumn that traps high-molecular-weight
sulfur and hydrocarbon compounds. The sample loop for the GC is a 1/8-in OD
Teflon tube, the length of which was adjusted to vary the amount of sample
injected. The columns and conditions used in this analysis were as follows:
Precolumn 33 cm x 0.32-cm OD Teflon tubing with
Carbopack BHT 100 40/60 mesh.
Analytical column 2.7 m x 0.32-cm OD Telfon tubing with 60/80
Carbopack B/1.5% x E60/1.0% H3PO4
Carrier gas Helium at 50 cc/min
Oxidant Air at 101 cc/min
Fuel Hydrogen at 76 cc/min
Column temperature 60°C for 2 minutes; 25°C/min to 135° and hold
for 7 minutes
The analysis procedure was as follows. The diluted sample was purged
through the loop and injected into the precolumn. The ^S, COS, MeSH, CS2»
and thiophene passed through the precolumn and were separated by the analyti-
cal column according to the above-mentioned temperature program. The DMOS and
other high-molecular-weight sulfur and hydrocarbon compounds were removed by
the precolumn, which was backflushed after each injection. The H2S, COS,
MeSH, and CS2 concentrations were determined by comparison with calibration
*Citrate buffer comprises 284 gram sodium citrate +41 grams anhydrous citric
acid in one liter of deionized water.
A-2
-------
INLET
PROBE
HEATING
SYSTEM
GAS FLOW
2?
FILTER
TEFLON
SAMPLE
LINE
TEFLON IMPINGERS-
WITH CITRATE f
BUFFER SOLUTION >
EXCESS VENT J
OUTLET
GAS FLOW
HEATING
SYSTEM
-A-
•FILTER
HIGH LEVEL
STANDARD
GAS
LOW LEVEL
STANDARD
GAS
NaOH
INLET
I SAMPLE
SOLENOID
SWITCHING
I VALVE
VENT
I
GAS CHROMAT06RAPH
WITH FPO
Calibrated
Standard
INTEGRATOR
STRIP CHART
RECORDER
L
MOBILE
LABORATORY |
Figure A-1.
Sampling and analytical system for
reduced sulfur compounds.
A-3
-------
gases generated with the permeation system. Thiophene was quantitated from a
standard tank. ;
Prior to each day's analysis, the GC-FPD was calibrated by generat-
ing at least three concentrations of each component of interest and
determining the FPD response curve by using a linear least-squares program.
At the beginning and end of each test day the sample system was checked for
sample recovery by injecting calibration gas through each sample probe, SO,,
scrubber, Teflon transfer line, and dilution system to the GC. The recovery
percentage was then calculated and used to adjust analytical results.
Samples were typically collected alternately from the inlet and outlet
of the Stretford and alkaline scrubber units. Each sample analysis required
approximately 12 minutes. ;
B. Sampling and Analysis for Total Reduced Sulfur Gases—
A continuous real-time analyses of total reduced sulfur (TRS) in the
retort offgas was made by oxidizing the sample gas stream in a tube furnace
and reading the total sulfur as SO2 using a Thermal Electron Corp. (TECO)
continuous SG>2 monitor. The system is shown on Figure A-2. This technique,
which was derived from EPA Method 15A, was used to provide alternate TRS
measurements from three gas sampling locations: the inlet to both the
Stretford and alkaline scrubbers and the respective outlets from those two
units. :
Referring to Figure A-2, combustion air is added to the oxygen-
deficient flue gas by dynamic dilution. A portion of the diluted sample
(2 liter/min) is drawn off a venting manifold and pumped through |a quartz
combustion tube. The combustion tube is heated to 1000°C in a tube furnace.
The sample stream flows from the combustion tube to a second venting mani-
fold,. The TRS monitor takes its sample from this second manifold. The
instrument response to the sample is recorded by a strip chart recorder.
The TRS monitor is calibrated at the zero to 100 ppm range with
hydrogen sulfide (H2s). The H2S used for calibration is diluted and mixed in
a porous plug dilution system. Each flow is measured with a bubble tube. The
A-4
-------
4J
tfl
W
•H
W
>1
i-H
(C
3
CO
0)
U
3
•8
(0
s
EH
I
-p
3
rH
•H
Q
o o
o o
oo
tJl
•H
A-5
-------
calibration relationship between ppm H2s and percent chart is expressed by the
slope and intercept of its linear regression.
The oxidation efficiency of the tube furnace is checked by comparing
the response of the gas chromatograph (GC/FPD) to the TRS monitor. The
comparison is not direct since the analytical range of the two instruments is
different. The TRS monitor calibration was checked daily with a mid-scale
precision point. After the point was stable, the range was changed to the 500
ppm scale to verify the output was 1/5 of the 100 ppm scale. i
C. Sampling and Analysis for Non-Sulfur Gas Components—
Sampling and analysis for (CO, CO2, O2, H2, CH4, and N2) gases were
conducted using a Baseline Industries, Inc. Model 1030-A GC with a thermal
conductivity detector (TCD). Samples were collected on a semi-continuous
basis from three sites, the inlet gas line to the Stretford unit and the
alkaline scrubber and the outlets from the respective scrubbers.
As shown in Figure A-3, samples were drawn from the source through a
coalescing filter and a condenser unit (to remove particulates and moisture)
and conveyed to the laboratory trailer by means of a diaphragm pump. Both
source lines were continuously purged. At the trailer, a valve manifold
system and diaphragm pump were used to draw a sample from either purge line
to fill the GC sample loop. Injections to the GC were made automatically
from a 1.0-ml sample loop with a 10-port pneumatic valve. Samples were taken
alternately from the inlet and outlet sites, with a new injection 'approxi-
mately every 15 minutes.
The GC conditions for this analysis were as follows:
Column 1 1.2 m x 0.32-cm stainless steel with 50/80
mesh Porapak N
Column 2 1.8 m x 0.32-cm stainless steel with
Molesieve 5A, 40/60 mesh
Carrier gas Helium at 25 ml/min !
Column temperature Isothermal at 75°C
Injection temperature 100°C
Detector temperature 100°C
A-6
-------
OUTLET
INLET
t
3-WAY BALL VALVE
SS PROBE
/
3-«AY BALL VALVE
SS MtOBE
COALESCING
FILTER
AA
Q— CAL GAS
CONDENSER —*-r]
316 S.S. UJ
PLE LOOP
WKP
Figure A~3. Non-sulfur gas components sampling train schematic.
A-7
-------
The analytical procedure was as follows. After its injection, the
hydrogen elutes through both columns in less than one minute. At this point
the O2, N2, CH4, and CO have eluted out of Column 1 (Porapak N) and into
Column 2 (molesieve). Carrier flow was then switched to bypass Column 2
trapping these components. The CO2 was then eluted from Column 1 to the
detector, and Column 1 was backflushed to elute organics in the C2 to C4
range. After the backflush step is completed, Column 2 was opened to carrier
flow and the remaining components elute in the following order: O2, N2, CH4>
an CO.
The GC-TCD was calibrated daily with a range of gas standards. Two
calibration mixtures containing each of the following components at concen-
trations of one and five percent are used to establish the low calibration
scale: H2, O2, N2, CO, CO2, and CH4.
A standard containing 30 percent CO2 and 40 percent N2 in helium was
used for higher calibration standards.
This standard was injected at the sample probe to verify sampling
systesm integrity. A standard of 0.5 percent propane was used to calibrate for
the backflush organic peak.
A<>1.2 Ammonia S&A Procedure
A. Sampling— ,
The retort gas was sampled for ammonia content simultaneously upstream
and downstream of the alkaline scrubber trailer during the Nf^OH scrubbing
tests. Two Andersen portable Method 5 type sampling consoles were used. The
sampling train consisted of a 1-cm OD stainless steel probe, four standard
impingers in an ice bath, a vacuum pump and dry gas meter. The first two
impingers contained initially 1000 ml of a 0.02 N H2SO4 solution (0.55
ml/liter or 12 drops concentrated [37N] H2so4/liter).
The third impinger was empty. The fourth impinger was filled with
desicant. A glass wool plug was used inside the probe to capture
particulates. :
A-8
-------
The sampling rate was maintained at between 0.00014 and 0.00019
am^/s. 0.14 Sm3 of gas was sampled. After sampling, the probe and connecting
glassware were washed with the 0.2N H2SO4 absorbing solution and the washing
liquid was collected in the first impinger. Finally, all impinger liquids
were combined.
B. Analyses—
The collected solution was analyzed with an Orion Model 907 Micro-
processor lonanalyzer equipped with an ammonia electrode. The system was
calibrated prior to the analysis and again after the analysis.
The samples were allowed to equilibrate to laboratory temperature (The
same temperature used for the calibration solutions).
The sample was first divided into 90 ml aliquots in 250 ml plastic
beakers. Each aliquot was analyzed as follows. The clean electrode was
immersed in the sample. A teflon stirring bar was placed in the beaker and
one ml of 10N_NaOH was added. After stirring sample for two minutes the NH,
concentration displayed on the instrument was recorded. The display read ppm
NH3 by weight„ This was repeated for each aliquot and the results averaged.
To avoid contamination the electrode was rinsed with distilled water and
blotted with clean tissue before each immersion. ;
The retort gas NH3 concentration - ppm (vol) was computed using the
following equation:
(ppm wt NH liq) (liq., vol., ml)
(dry ppm V) = 0.049 DSCF (16UC) of sampled flue gas
A.2 LIQUID STREAMS
Ao2.1 Scrubber '••
Liquid samples were taken from the scrubber effluent stream after
approximately 20 minutes of test operation to assure steady state
conditions. The samples were separated into seven containers and preserved in
accordance with Table A-1.
A-9
-------
TABLE A-1. ALKALI SCRUBBER WATER SAMPLES (Each Run)
(a) Preservation Methods
Pollutant
Container
Preservation
Sulfide
Ammonia/Ammonium
Alkalinity
Dissolved and Sus-
pended Solids
Total Solids/
Sulfate/Sulfite
Total organic and
inorganic carbon
Sodium
pH
500-mL amber glass
500-mL amber glass
500-mL amber glass
500-mL amber glass
Add zinc acetate (several crystals),
Cool to 4°C
Adjust to pH <2 w/H2SO4, Cool to 4°C
Cool to 4°C ;
Cool to 4°C
500-mL amber glass Cool to 4°C
500-mL amber glass Adjust to pH <2 w/H2SO4, Cool to 4°C
500-mL plastic
Adjust to pH <2 w/HNO3, Cool to 4°C
Analyze on site
(b) Analytical Methods
Method No.*
Alkalinity (CaCO3)
Bicarbonate (HCO3)
Carbonate (CO3)
Hydroxide (OH)
Residue, Filterable (TDS)
Residue, Non-filterable (TSS)
Residue, Total (TS)
Sulfate (SO4)
Sulfide (S)
Sulfite (SO3)
310.1
i
310.1
310.1
310.1
160.1
160.2
160.3
375.2
376.1
377.1
*EPA-600/4-79-020 "Methods for Chemical Analysis of Water and Wastes"
A-10
-------
A. 2.2 Stratford
A number of chemical analyses were conducted on the Stretford solution
during this test program. These chemical analyses were performed in order to
determine the following solution properties:
pH
. oxidation level :
. carbonate concentration
. anthraquinone disulfonic acid (ADA) concentration
. vanadium concentration
. thiosulfate concentration
The proposed test plan* specified that the chemical analyses be
performed at fixed intervals. These intervals are shown in Table A—2, along
with the method of analysis used and the desired levels. The samples for each
of the required chemical analyses were taken from a sample line located at the
bottom of the solution heater.
Each of the six chemical analysis methods are described below.
A. pH and Oxidation Level Analytical Procedure—
The pH and oxidation measurements were performed with an Orion
Model 907 Microprocessor lonanalyzer, pH probe and oxygen sensing probe
respectively.
pH Procedure
1. Keep the lonanalyzer plugged in at all times; switch to Standby
when not in use. Suspend the pH electrode in deionized water or
pH 7 buffer when not in use. Keep fill arm cap on when not in
use.
*Proposed Test Plan, Pilot Plant Testing of Stretford Technology on Oil Shale
Retort Off-Gas at Geokinetic's Kamp Kerogen Facility. Second field test;
Pedco Environmental, Inc., September 30, 1983, Appendix C.
A-11
-------
TABLE A-2. PROPOSED CHEMICAL ANALYSIS SCHEDULE
FOR STRETFORD PILOT PLANT
Analysis
pH
oxidcition level
sodium carbonate
Method
electro-chemical
electro-chemical
distillation/
Schedule
6 per day
6 per day
1 per day
Desired
Level
8.5-9.5
;
' 25.0 g/liter
ADA
Vanadium
Thiosjulfate
titration
spectr ophotometry
titration
titration
1 every 2 days
1 every 2 days
1 every 2 days
9.6 g/liter
3.12 g/liter
<20%
A-12
-------
2. Make sure calibration buffers and samples are at room'temperature.
Calibrate the system daily as described here. Set the slope dial
to 100 percent and the temperature dial to room temperature.
Remove the rubber cap from the electrode fill arm. Immerse the
electrode in pH 7 buffer, turn the mode switch to pH/0.01, and set
the display to 7.00 by using the calibration knob. Switch beaker,
and immerse it in pH 10 buffer. Switch to pH/0.01 and set the
display to 10.00 with the "% slope" dial. Do not change the tem-
perature setting. Switch to Standby and rinse as before.
3. Immerse the electrode in the sample, switch to pH/0.01, and record
the sample pH. Switch and rinse between samples, as above.
[
4. Add electrode filling solution through fill arm as needed to keep
the level within one inch of the arm. ;
Oxidation Level Procedure
The oxygen content of the Stretford solution was determined by
substituting the oxygen sensing electrode into the Orio lonanalyzer in place
of the pH electrode. A calibration solution was not used.
B. Specific Carbonate Analytical Procedure—
Reagents; ;
1. Absorbing solution:* Dissolve 22 g NaOH and 1.0 g Na2CO, in
deionized water and dilute to one liter. Add a few crystals of
thymolphthalein indicator.
2. Standard HC1, 0.5N_: Dilute 42 ml concentrated HC1 to one liter.
Standardize against Na-CO., solution carried through the entire
procedure.
3. Barium chloride crystals.
4. Hydrogen peroxide, three percent: Dilute 100 ml of 30 percent
H2°2 to one Uter. Prepare every three days. Refrigerate.
!
5. Hydrochloric acid, 6N; Dilute 500 ml of concentrated; HC1 to one
liter. ;
6. Sodium carbonate standard: Dissolve 15.8948 g of Na2CO3 in
deionized water and dilute to one liter. (9.0 mg CO, per ml.)
*EPA-600/4-79-020, "Methods for Chemical Analyses of Water and Waste Waters.
A-13
-------
7. HC1, 2.4N_: Dilute 200 ml concentrated HC1 to one liter.
Procedure;
1. Assemble the apparatus as shown in Figure A-4. Pipette 15 ml of
absorbing solution into the absorber. Put glass beads in the
distilling flask.
2. Measure an aliquot of sample calculated to contain 100 to 150 mg
of CO3 into the distilling flask. Add water to cover;the bottom
of the thistle tube. Add ten ml three percent H2O2»
3. Turn on the vacuum so a gentle stream of bubbles is generated in
the absorber. Add 20 ml of 6N_ H2SO4 through the thistle tube.
Before the acid is completely drawn into the flask, attach the
LiOH tube to the tube inlet. Adjust the vacuum as necessary
during the distillation. Turn on the cooling water flow to the
condenser.
4. Bring the solution to a gentle boil and hold for about: two to
three minutes. Turn off the flame and continue to draw air
through the system for 15 minutes. If the absorbing solution does
not remain blue, too much sample was used. Start again with less
sample.
5. Remove the absorber and transfer the absorbing solution to a
beaker. Rinse the absorber into beaker. Moisten a strip of lead
acetate test paper with 2.4N_ HC1. With a stirring rod, transfer a
drop of absorbing solution to the paper. If the paper turns
black, H^S has distilled over. Repeat the test, increasing the
amount or strength of the H2O2 added, until no H2S distills over.
6. When a sulfide-free distillate is obtained, add two to three g
Bad2 while stirring. Lower the pH electrodes into the
solution. Add rapidly but dropwise, enough 2.4N^ HC1 to bring the
pH to about 10. From that point, titrate the solution stepwise
with 0.5N_ HC1, recording the number of milliters used and the pH
after each addition. Titrate at least to pH 3.5. Make small
additions near pH 8.3 and 4.5; larger ones can be used between
these values.
7. Titrate 15 ml of absorbing solution in the same way each day.
S tandardi z ati on;
1. Prepare the apparatus as described above. Pipette 15 ml of
standard sodium carbonate into the distilling flask. Carry this
solution through the entire distillation and titration procedure.
N HC1 = 4'4" :
ml HC1
A-14
-------
Cooling Water
LI OH
TUBE
TO LOW
VACUUM SOURCE
ABSORBER
DISTILLING
FLASK
BURNER
Figure A-4 Carbonate distillation apparatus.
A-15
-------
Calculation;
1. Graph the results of the titration, milliliters versus pH»
Connect the points with a smooth line. Determine the milliliters
used between the two inflection points.
.. . • _ „ __ (T-B) N (10.6 x 105)
g/liter of Na.CO_ = :
2.3 ml sample •
where T = milliliters of acid for sample
B = milliliters of acid for blank
N = normality of acid
Anthraquinone Disulphonic Acid (ADA) Analytical Procedure\
Reagents;
1. Sodium hydroxide solution, 30 percent W/W (NaOH) •
2. Sodium dithionite powder
3. Anthraquinone disulphonic acid (ADA) standard solution, 250 mg.
Calibration (Perform with each set of samples);
1. Prepare a series of 100-ml volumetric flasks containing 0, 3, 5,
and 10 ml of the 0.250 mg/ml ADA standard solution. The flasks
contain 0, 0.75, 1.25, and 2.5 mg ADA.
!
2. Add dithionite and NaOH, dilute, mix, and measure absorbance as
described in "Procedure" below. |
3. Plot absorbance (Y axis) versus mg ADA (X axis).
Procedure;
1. Familiarize yourself with the operation of the Spectronic 70.
Allow the instrument to warm up one hour before each use.
2. Pipette five ml of filtered Stretford solution into a 100 ml
volumetric flask and dilute to volume with deionized water.
3. Pipette four aliquots of this solution (five ml each) into 100 ml
volumetric flasks. To each flask add approximately 0.1g sodium
dithionite powder and ten ml of 30 percent NaOH solution.
4. To three of the flasks pipette about 0.5 x, x and 1.5 x mg of ADA
from the standard solution respectively, where x = the_mg of ADA
in the aliquot. Assuming the Stretford solution contains 6.2 g/1
of ADA the volumes of standard to be added are 3, 5 and 10 ml or
0.75, 1.25, and 2.5 mg. Dilute each aliquot to volume.
A-16
-------
5. Measure the absorbance of each aliquot using the 0 rag ADA standard
as zero, at 414 ny. Determine the measured mg from the calibra-
tion curve.
7.
Plot the added mg on the X-axis versus the measured mg on the Y--
axis. The x intercept is the actual amount of ADA (mg) in the
sample aliquot.
If the absorbance is less than 0.1, use a larger aliquot of the
dilute sample solution. If the absorbance is greater than that of
the highest standard, use a smaller aliquot. (Note: adjust the
volume of standard added accordingly.)
Calculation:
Grams ADA =
mg of ADA from graph (step 6) x 20
ml of diluted sample used
A
b
s
o
r
b
a
n
c
e
Calibration Curve
mg ADA
mg ADA
measured
in sample
(Plot from procedure
Step 6, for use in
calculation step)
mg of ADA in
Stretford solution
sample
mg ADA added to sample
Sample Absorbance Curve
A-17
-------
D. Vanadium Analytical Procedure
This test must be performed in an exhaust hood. Highly toxic and
corrosive sulfur trioxide gas is evolved during the procedure.
Reagents;
1. Sulfuric acid, 50 percent (H2SO4)
2. Concentrated nitric acid, 70 percent (HMO,) ;
3. Potassium permanganate solution, 0.5 percent KMnO4)
4. Sodium nitrite solution, 0.5 percent (NaNO-) ;
5. Sulfamic acid solution, 10 percent
. . r
6. Concentrated phosphoric acid, 85 percent (H3PO4)
7. Sodium diphenylamine sulfonate powder, (SDS) .
8. Ferrous ammonium sulfate solution, M3.025N_ (FAS)
9. Vanadium standard solution, 0.015jb^
1
Procedure;
1. Pipette 25 ml of filtered Stretford solution into a 250-ml
Erlenmeyer flask. Cautiously add 25 ml H2SO4.
2. Add 25 ml HNCU. Bring solution to boil on hotplate in hood. Boil
until greenish color develops and copious white fumes of SO, are
evolved.
3. Remove from hotplate and allow to cool.. Dilute to approximately
100 ml with deionized H2
-------
Calculations;
0.375
N of FAS =
ml FAS
ml FAS x N x 50.9
Vanadium, g/liter = . _ .
ml of sample
where N = normality of FAS
E. THIOSULFATE ANALYTICAL PROCEDURE I
Reagents; i
1. Calcium chloride crystals (CaCl2)
2. Sodium hydroxide solution, 10 percent (NaOH)
3. Sulfuric acid solution, 25 percent (f^SO^) ;
4. Iodine/iodide solution, 0.1 N_ in I2, 0.24 N_in KI (I2)
5. Phenylarsine oxide solution, 0.113 _N_ (PAO)
6. Starch indicator solution
Procedure;
1. Pipette 25 ml of Stretford solution into a 250-ml beaker. Add
30 g CaCl-. Add enough distilled water to allow the use of a pH
probe, and adjust the pH to 10 to 11 with NaOH.
2. Heat the solution, while stirring, to 85°C.
3. Cool the solution to room temperature, and filter through Whatman
42 filter paper into a 250 ml Fleaker. Wash the filter cake with
water, using a minimum of three rinses.
4. Add 25 ml H-SO*, mix, pipette in 50 ml I2, and mix.
5. Titrate with PAO to a pale yellow color, add 1 to 2 ml starch
indicator, and titrate to the disappearance of the blue color.
Adjust sample volume so 10 ml < ml PAO < 50 ml. i
6. Standardize the I2 with each set of samples by pipetting 20 ml of
it into a solution of"" ----- • - . -- - - — • .
with PAO as in step 5.
it into a solution of 15 ml H2SO4 in about 50 ml water. Titrate
A-19
-------
Calculations;
' 2.26
_
N of
_
2 ml
(ml I x£I ) - (ml PAD x .113)
Sodium thiosulfate, g/1 = 158 x -
ml of sample
A-20
-------
APPENDIX B
SELECTIVE SCRUBBING OF H2S FROM C02
IN SHALE OIL RETORT OFFGAS BASED
ON THE PENETRATION THEORY
Prepared by
Dr. Richard C. Aiken, Consultant
c/o University of Utah
Salt Lake City, UT 84112
-------
PART 1
Penetration Theory for Mass Transfer and Reaction of H?S-CO?-NH-3
The simultaneous chemical absorption of C02> H2S and NHg from
exhaust gases by an alkaline scrubbing liquid is, considered
here. The following reactions are assumed to occur.
i ) H2S + NH3 * HS" + NH*
This reaction is instantaneous and irreversible. Since both
components enter the liquid film from the gas phase, three cases
have to be considered depending on the relative amounts of HoS
and NH3 at the interface: '
I. [NH3]i > [H2S]1 ;
II. [NH3]i = [H2S]i
III. [NH3]i < [H2S].
The species which is in lesser amount will be consumed at the
interface and will not exist inside the liquid film. Its
absorption will be entirely controlled by the gas film; liquid
film resistance to mass transfer will be effectively zero. Its
interfacial concentration can be set to zero for computing the
rate of transfer across the gas film. The species in excess will
diffuse in and react. Carbon dioxide diffuses in -and reacts
according to
11) • C02 + 20H" = CO'2 + H20 .
B-l
-------
This reaction' is also instantaneous and irreversible. t There will
be a reaction plane at which C02 and OH" are consumed instantan-
eously.
Case I.
H2S is annihilated at the interface; the excess NH3 ( dl ssol ved )
is consumed by the instantaneous and irreversible reaction
NH3 + H+ -»• NH* . . la
_2
The species to be considered are NH4, HS~, C02, C03 and OH
The two species which react instantaneously and irreversibly at a
plane are C02 and OH" according to ii) above.
C02 + 20H" = CO'* + H20. Ib
Reaction between CO? and ammonia (or NH.) can be neglected
— 4
because of unfavorable equilibrium constants (K ~ 10 ). All
the other species undergo physical diffusion only. The
enhancement factor for H2S and NH3 in the liquid film is
infinite, i.e. absorption of H2S and NH3 is entirely controlled
by gas film resistance. the interfacial concentration of both
H2S and NH3 can be set equal to zero to calculate the rate of
absorption across the gas film. At any time t, the concentration
profile in the liquid film is shown in Figure B-la.
i
Let
A = C00 ; C = OH
x1 = location of reaction plane
B-2
-------
Gas/Liquid
Interface
-Reaction Plane
x1(t) = DISTANCE INTO LIQUID
a. Case I - [NH3]i > [H2s]i
*• Gas /Liquid
Interface
0)
'IS
'rH
jo.
Jg
1
Ig
1 (0
sir
CO
OH~
OH~
b. Case III - [n,s]i > [NH,]i
Figure B-l. Concentration Profiles
B-3
-------
_aA
at
A ax2
0 < x < x
3C
IT
= D
r Z '
c ax
X ' < X< X .
A(0,x) = 0;
C(0,x) = C0);
A(t,0) = A-
C(t.x') = 0;
A(t.x') = 0;
C(t,») = C0.
Following the treatment by Bird, Stewart and Lightfoot (1960), we
can write ;
x1 = /4ot ,
and a is calculated from the flux relation
Dn
A ax
X1
c ax
X '
The analytical solution is
= A. [1 -
erf
erf {-j=p*;
A
0 < x < x
B-4
-------
c = cn [ i -
*
erfc c
er
fc <-^>V2 '
X ' < X <
a is obtained by solving the nonlinear equation:
i UA
A
From the concentration profiles we can calculate the rate of mass
transfer at the interface:
= - D
_3A
A 3x
Ai
x=0
The average rate of absorption up to time t is
NA ' 2NA
z = 0
The enhancement factor is:
= instantaneous enhancement factor =
B-5
-------
and
E. = average enhancement factor up to time t = 2E.
Case II. [NH3]i = [HgS]1 ;
This case is very similar to Case I. Both H2S and NH3 are
consumed at the interface by the reaction
H2S + NH3 •* HS" + NHj
Reaction la does not occur since there is no excess NH3.
Reaction Ib does occur, however. The concentration profiles and
enhancement factory for C02 remain the same as in Case I except
that
[NH+]. = [HS']..
Case III . [H2S].j > [N H 3 ] i
This is the most complex, and interesting case mathematically.
NH3 is converted to NH. at the interface by the reaction
H2S + NH3 = HS~ + NH*. .
The excess H2S along with C02 diffuses into the liquid and reacts
with OH~. This process can be modelled using the tw.o-reaction
plane approach of Astarita -(1965). Alternatively, the penetra-
tion theory equations may be used and the enhancement factors for
H2S and C02 in the liquid film calculated, as by Onda et al .
(1972). :
B-6
-------
Fixing the H2S concentration at the interface is trickey. We can
just set it equal to the difference in solubi11 ties ' between H2S
and NH3. The profile at any time t is shown in Figure B-lb.
The expressions for the concentration profiles and enhancement
factors given by Onda et al. (1972) follow: ;
Let
A = CO2, B = H2S
A = AI [ 1 -
erf {7?DAT~}
~
B - B1 [ 1 - -^
erf {•
xl
/4DBt
EA erf
•B erf
where
B-7
-------
/4DAt
A - DB
and
/4DAt
4^ and 2 are obtained by solving equations (31) and (32) in
Onda et al . !
B-8
-------
Design of Venturl Scrubber for Multicomponent
Mas;; Transfer with Reaction
Consider a differential segment of a venturi scrubber of length
dz, in which several gaseous species (S species in total) are
absorbing into a reactive liquid. For the nth component, a mass
balance based on the liquid side gives:
Nnadz ' d(W - LmdXn + XndLm
dLm = N.adz, Nt = J N.
v "*•
so
Nnadz '
LmdXn dXn Nna N.aX
dz = M m., nv or T-H = -J3 5 I
N a-N.aX dz L L
n t n mm
Define the overall coefficient KoL
Y - Y .
where M. - n ni
n *:
B-9
-------
Mn is derived as follows:
The overall mass transfer coefficient, based on liquid side for
multicomponent absorption with reaction, is defined from
Nn ' kg R(Yn ' Yni> = kL~'
-------
Substitute for Nn:
dXn KoLnap (Xn*
„ .
2- [ z K . - (X.
= °Lap J-
Take
n=l,2,...S
Now derive the similar relation for mass transfer of component n
based on the gas side: i
Nnadz « -d(6mYn) - -
- - Ntadz
Nnadz ' - GmdYn + YnNt'adz
dz " G
V
G
m m
where
P<¥
B-ll
-------
<
_
lea - Mr~-' » X 5?
Kg p kL Xn n
N = I N
To relate the gas and liquid-side balances, consider
'GmdYn - YndGm = LmdXn + XndLm'
- -Ntadz,
dLm = N.adz,
so that
GmdYn + Yn(Ntadz) = LmdXn
dYn
- Gm dir
Rearrange to:
dz "• Gm " Gm
B-12
-------
dGm
m _ N
HI Nta
' V
Assuming that the liquid atomizes instantly into the droplet form
with a constant mean diameter at the point of entry to the gas
stream, the force balance yields (Uchida and Wen, 1973):
where t is time (s) after gas contact in the throat section. The
gas velocity is taken to be a constant in this region. For the
drag coefficient, the relation of Ingebo (1956) is used:
Cd - 27/Re0'84 ,
I
i
where
Re = d |Vr-V. | pr/y
Distance along the throat is obtained from
li = V
dt V
The gas-side mass transfer coefficient, kg, is computed with
consideration to the droplet size and varying relative
velocity. We use the correlation of Gupalo and Ryazantser
(1972):
B-13
-------
Sh = 0.991 Pe1/3(l+Re/4) °'27
krRTdn
G P
'
where
^
The liquid-side mass transfer coefficient for physical, absorption
is taken from the penetration theory as ;
D
k. = 2 / -£-
L irt
where D|_ is diffusion rate into infinitely dilute solution.
The parameter a, surface area/volume of unit, is calculated from
a = 6(l-e)/dp, ;
where
e = 1-LQ/(VLA) .
Vapor-liquid equilibria
Let n = 1 for H2S, n = 2 for C02, and n = 3 for NH3. We
recognize for inlet partial pressures of H2S and NH3 of the same
B-14
-------
order of magnitude that the NH3, being much more soluble than
H2S, will be present at the gas-liquid interface in larger
concentrations than H2S. Furthermore, as NH3 and H2S participate
in a very fast reaction (usually, as here, considered
instantaneous),
NH3 + H2S + NHj + HS" ,
to a first approximation we will assume that the interfacial
concentration of H2S is zero and the concentration at NH3 is
equal to its value for physical absorption, with gas film
controlling, minus that for H2S under physical absorption (no INH3
present), .also gas film controlling. This will be approximately
true until NH3 has been depleted in the gas phase enough that its
single-solute physical absorption leads to a concentration equal
to that of H2S. After that point, further loss of NH3 reverses
the role of NH3 and H2S: the interfacial concentration of NH3 is
zero and that of H2S is computed by subtracting the single-solute
physical absorption concentration of NH3 from H2S (Van Krevelen
and Hoftijzer 1949). !
The use of physical absorption data for calculation of
interfacial concentration is appropriate here since the liquid is
assumed stagnant and reaction fronts are set up for the fast
_ 2
reaction involving OH" and C03 that move away: from the
interface. ;
For equilibrium of H2S take
where log H: = 102.325 - 4423.11 T'1 - 36.6296 log T +"0.013870T
(Mason and Kao, 1979). Here the ionic strength was taken to be
zero for this infinite dilution Henry's constant; T is in degrees
Kelvin.
B-1.5
-------
For NH3? the equilibrium expression '••
X3 = PY3/H3p ,
where lnH3 = - 157.552/T + 28.1 luT - 0.049227T - 149.006 as
given by Edwards et al . (1978).
For physical absorption of C02
where ;
log H2 = 3.822 - 7.8665 x 10~4 exp(T/100)
- 0.04145 (T/100)2-17.457(T/100)-2
as given by Mason and Kao (1979). ;
Parameters used in this study ' : •
Physical and chemical parameters used in this study appear in
Table B-l; operating parameters and their ranges in Table B-2.
A standard case representing a most probable actual operating
condition appears in Table B-3 (only shown are parameters given
ranges in Tab! e B-2) .
B-16
-------
Table B-l. Physical and Chemical Parameters
_ 3
= 8x10 g/cm.s
_ "* 3
P = 7.3 xlO g/cm
PL = Ig/cm
p = 0.0562 gmol/cm
g = 980 cm/s2
DH s = 2.21 x 10"5 cm2/S
Dco2 = 10"5 CI"2/S
D _ ~ (1.7 - 2.7) DCQ
OH LU2
DH s (gas phase) = 0.424 cm2/s
C02 + OH"'* HOD' kf = 6000 £/gmol s K = 3.4 x 107
H2S + OH" * HE" + H20 Keq = 8.9 xlO8
B-17
-------
Table B-2. Operating Parameters
VL (inlet) 200-5000 cm/s
dp = 10-70 ym
venturi throat diameter = 3.5 cm
venturi throat length = 25-40 cm
temperature: 30 degrees C
Lm = 1.7 gmol/s cm2
Gm = 0.4 gmol/s cm2
P = 1 atm ;
inlet OH~ concentration = 0.01 - 2 gmol/£
inlet gas H2S concentration = 50 - 2000 ppm!
inlet gas NH3 concentration = 10 - 2000 ppm
inlet C02 concentration 10 - 30% (vol)
B-18
-------
Table B-3. Standard Case Operating Conditions
VL = 200 cm/s
d = 30 ym
venturi throat length =30.5 cm
inlet liquid OH" = 0.025 gmol/A
inlet gas H2S = 1400 ppm
inlet gas NH3 = 950 ppm
inlet gas C02 = 22% (vol)
B-19
-------
RESULTS AND DISCUSSION
The venturi design equations and the penetration theory
equations described in this report were coded in a computer
program and run on a UNIVAC 1100/61 at the University of Utah
Computer Center. A partial listing of this program appears in
Part II, Page B-36 with output for the base case presented on
Pages B-43 and B-44.
In addition to the listed program, subroutines on the UNIVAC
library for stiff differential equation solution and simultaneous
nonlinear algebraic equations were utilized.
Figure B-2 shows the percent removal of ^S and C02 versus
distance down venturi throat; Figure B-3 shows selectivity, S
defined as
% removal H,S
S =
_, _ ,
% removal CO,,
versus distance down the venturi throat. Figure B-2 indicates a
substantial portion of H2S (59%) is removed in a single pass
through the venturi , whi le only a small fraction of C0£ (1.8%) is
removed (NHg is reduced 69%). Most F^S removal occurs early in
the throat. The corresponding selectivity shown in Figure B-3
indicates that a maximum in the selectivity is likely some
intermediate distance down the throat. This agrees with the
results of Hsieh and Aiken (1984) and is a result of the fact
that up to and including the region of the peak h^S is gas film
controlled while C0£ is liquid film controlled; the gas film
coefficient is quite high for small contact time but decreases as
the gas-liquid relative velocity decreases, while the liquid film
coefficient does not decrease as fast.
Figure B-4 shows the dependency of the selectivity on
reactant concentration. The selectivity is seen to decrease
substantially with increase in OH" concentration. C0£ is
apparently aided relatively more than HgS by the reactant.
B-20
-------
( "TOA)
o
Cl
in
CN
o
CN
m
in
in
CN
o
CN
in
a
o
•P
C
O
Q
O
fl
.(tf
t!
-H
Q
o
o
H
O
O
CN
i-l
O
O
O
o
o
CO
o
o
o
o
o
o
CN
01
-a
x
o
c
o
CO C
CU
C >
a> '
01 c
o 3
s- o
-a -a
>>•
J= O)
o
i— i/i
(O •rj-
> -a
O '< i
E 4-
cu o
CU -P
to o
(O C
O 3
4-
O)
CO
-------
O
O
O
oo
o
CN
-p
Cn
0)
CD
CM
O
IN
CN
O
tn
c
0)
0)
CO
0)
(D
O
0)
T)
•H
w
14-1
O
O
-H
4-J
0
3
in
•H
+J
U
0)
f-l
CO
CO
ca
0)
Cn
•H
fe
B-22
-------
o
o
o
o
0)
+J
•H
0)
!-f
o
4J
•rH
>
•H
4J
O
0)
rM
0)
to
c
o
c
o
-rH
4J
(ti
M
4J
c
Q)
O
8
0)
•o
•H
O
(N
O
•H
4J
•H
•rH
M-l
O
•P
O
0)
H
O
O
in
I
CQ
0)
M
Cn
•H
B-23
-------
Comparison with the alkaline scrubber tests on gases from an
oil shale in-situ retort of Geokinetics (Uintah County, Utah) is
shown in Table B-4 (data taken from the main body of this report,
Table 10 on Page 73). The model results are the same as tabu-
lated in Table 10. The agreement between model and experiment, is
excellent, we think, for this complex system. We claim that Run
No. 19 has an erroneous value for the selectivity as indicated in
Table 10, page 73 of the main text. It is not possible for the
selectivity to increase with a further OH" concentration
increase.
Finally, we consider the result of three scrubbers in series,
with fresh scrubbing liquid (OH" = 0.025 gmole/liter) in each
pass; Figures B-5a to B-5c contains the results. Panels a and b
contain the removal percentages of H2S or NH3 and C02,
respectively: panel c the overall selectivity. It may be seen
that the selectivity suffers quite significantly from the
multiple passes through the venturi chain. While the percent H2S
removed increases from 70 to 82 to 93 in the three passes, the
selectivity drops from 32 to 23 to 15.
Further discussion ;
The venturi model with penetration theory does not now
include any adjustable parameters whatever, although rather
!
approximate values for physical constants have been chosen. The
atomization zone is not included in the model as its physics are
quite complex; only an empirical-based approach could be taken if
experimental data were available for this region, which it is
not. Note, however, that Hsieh and Aiken (1984) and Bendall and
Aiken (1982) found experimentally quite unfavorable selectivity
in the atomization region of a pressure nozzle and this is likely
to be so for the venturi. The overall selectivity would thus
decrease somewhat for the unit. ;
Our results would indicate that a low value reactant
concentration, less than 0.01 gmole/liter, be used in a several-
B-24
-------
TABLE B-4 COMPARISON OF MODEL TO GEOKINETICS SCRUBBER DATA
Model Predicted Selectivity
t°H 3 at Venturi Lengths Measured Selectivity
gmole/liter 20.5 cm 24 cm Average Average Run Numbers
°-045 31 25 28 21 26
0.023 53 43 48 55 24 and 30
°'012 82 66 74 75 21 and 28
B-25
-------
w
Q)
W
a
(D
0)
O
-a-
o
CO
o
CM
O
rH
O
H
n3
O
-p
c
0)
o
03
Q)
M
(Q
W -H
0) O
•H Id
S^
0) 0)
•H 14-1
n °
5-51
C -H
OJ 4-1
> O
(U
i-l iH
-d
-OD %
M-l
o
-P
u
a)
CO
Q)
tn
•rH
fa
HO S
B-26
-------
pass venturi serial chain. The residence time of each venturi
should be optimized and quite likely significantly shorter than
the 30.5 cm throat of the present configuration.
The smaller the droplets the better would be the selectivity
and extent of H2S removal since the gas film can limit transfer
of H2S but not C02 (not documented herein). Similarly, the
higher the gas flow rate the better will be the selectivity (not
documented herein); this is so because smaller droplets are
produced as well as a large relative velocity, .which also
increases the gas film coefficient. Again, however, we must
qualify these observations with the fact that no consideration is
given in the model to the atomization process. ]
Extent of Removal vs. Selectivity ;
We seek a single measure of the performance of a mass
transfer unit for the selective and extensive removal of hydrogen
sulfide from carbon dioxide. The only parameter used up to this
point has been defined as the selectivity on page B-19.
This does not consider at all how much hydrogen sulfide is
removed. Thus a process may have excellent selectivity while
transferring a negligible quantity of hydrogen sulfide! An
alternative may be to weight more heavily the percent removal of
hydrogen sulfide:
a
(% removal of hydrogen sulfide) ;
% removal of carbon dioxide
where a is greater than unity. The result of doing this is
indicated in Figure B-6 as a function of distance down the
venturi. Note on this figure a= i corresponds to the curve of
Figure B-2 but has a different character here. It was seen on
the original that a maximum in selectivity occurred at about 12
cm down the throat; this resulted from a decreasing gas film mass
transfer coefficient as the gas-liquid velocity decreased and in
increasing liquid film coefficient for carbon dioxide as the
atomization zone is neared due to liquid mixing. Because we
cannot at this level of model development specify well such
behavior, we have chosen to omit it here.
B-27
-------
•H ^
00
CM
O
CM
VD
i-l
w
u
EH
s
O
oo
4J
•rH
>
•H
4-1
O
0)
O
c
•H
14-1
c
•H
0)
4J
a)
nj
ft-
c
O
•H
•H
4J
c
•H
I
4J
VD
A
O
CN
rH
O
O
O
00
O
CN
fe
B-28
-------
As a increases in Figure B-6, an optimum is reached in
selectivity that occurs later in the throat as a becomes
larger. This is quite reasonable, as more emphasis on extent of
removal would favor higher residence times. Clearly a good
choice of a should be greater than unity; we choose here a = 3.
With this value of a , we consider a train of four Venturis
in series. We design each venturi so as to maximize the new
selectivity ( a= 3). Figure B-6 shows the result of doing this
with a = 1; the curve is rather steeply decreasing in
selectivity with distance, i.e., as traveling from venturi to
venturi - but is still quite superior to the selectivity given in
and reproduced in Figure B-7 here as the dashed curve, in which
the entire length of three Venturis (30.5 cm) was used (optimum
lengths here were 9.5, 9.5, 10.5, 11.5 cm, respectively). The
curve with a = 3 is much less steep, indicating our choice of
venturi length is consistent with good selectivity and extent of
removal. Total percent hydrogen sulfide removal after the three
Venturis was 93 percent for a = 1, compared with ;90 percent
for a = 3. I
(
An indication of temperature effects is shown in Figure B-
8. There is seen to be a weak selectivity advantage to elevated
temperatures. Note, however, our model includes temperature
effects only in the vapor-liquid equilibrium and no effects on
reaction rates. The actual temperature dependence of selectivity
has been shown to be in the opposite direction for hydroxide
solutions (Garner, et al . , J. Appl , Chem. _8, 325, 1958).
B-29
-------
130
100
X
EH
H
>
H
EH
a
50
a = 2
a = 3
a defined by:
Selectivity,=
(% H2S Removal)
(% CO2 Removal)
a
01234
VENTDRI NUMBER ;
r
Figure B-7. Selectivity vs. venturi pass; solid curves based on
optimum selection of venturi length for two different
definitions of selectivity. Also shown selectivity
for full 30 cm venturi (dashed curve). >
B-30
-------
1 I
35
EH
H
>
H
EH
O
30
25
a = 1
20
I I I
300 325 350 -
TEMPERATURE, °K
Figure B—8. Effect of Temperature on Selectivity
B-31
-------
Nomenclature '
Q
A = local venturi cross sectional area, cm
Cj =drag coefficient >
dp = droplet diameter, cm
En = enhancement factor for component n in liquid.
e = local void fraction in venturi
p
g = acceleration due to gravity, cm/s
6Q = volumetric gas flow rate, cm^/s
k_ = gas phase mass-transfer coefficient, gmol/s cm2 atm
kL = liquid phase physical mass-transfer coefficient, cm/s
LJVJ = molar liquid mass velocity gmol/s cm ;
LQ = volumetric liquid flow rate, cm^/s |
N^ = mass transfer flux of component n, gmol/s cm2
S 2
N1t = total mass flux = ( I Nn), gmol/s cm
n = l
P = total pressure, atm
P = average molal liquid density, gmol/cm3
PG = density of gas, g/cm3 • ;
PL = density of liquid, g/cm
S = number of species transferred from gas to liquid
t = contact time, s
V(. = gas velocity, cm/s
«*
Vt = liquid velocity (droplet velocity), cm/s
Xn = mole fraction of n in the liquid (mixing cup)
Xni- = mole fraction of n in the liquid at the interface
*
X = liquid phase mole fraction of A that would be in
equilibrium with gas of mole fraction Yn. ;
Y
ri = mole fraction of n'in gas i
B-32
-------
Yni- = mole fraction of n at the interface in the gas \
*
Y = gas phase mole fraction of n in equi1ibriurn with bu1k
concentrationinliquidn
z = absorber length, cm • .
B-33
-------
PART II
PENETRATION THEORY COMPUTER PROGRAM LISTING
-L IIII SSSSS I III IIII IIII N NN 66666
-L ii sssssss mm n n NN NN 6666666
-L II SSS SS TT II NN NN 666 66
-L ; II SS TT II NNN NN 66
-1- II SSS n II NNNN NN 66
-t- II SSSSSSSS H II NNNN NN 66
LL II SSSSSSSS H II NNNNNNN 66 666
->- II SSS H II NNNNNN6666
-I- II SS TT II NN NNNN 66 66
-L II SS SSS TT II NN NNN 666 666
IIHIIII II SSSSSSS TT II NN NN 6666666
llllll.ll IIII SSSSS TT IIII NN N 66666
* « * UNIVERSITY OF UTflH COMPUTER CENTER - UNIVflC 1180 BflTCH/TINE-SHftRIN6 EXEC LEV. 38RMP/PROD5 SITE * U OF U * * * *
WER-. 352788 INPUT DEVICE: CRTC9 PflRT-MHBER: 8 flCCOUNT-MMER: 352788 USER-ID: 352788 CREflTtH: 352788
FILE-NWC ON«TEPRINT$ OUTPUT DEVICE: CC2 CREflTED flT: 88:22:13 NOV 28,1984 PRINTED AT: 88:23:87 NDV 28,1%4
B-34
Ol OTACCTOQat O7ACgTae^«
-------
3PRT HOPE
FURPUR FH8R1 UC1E E38 S74T11 11/28/84 68:22:25
B-35
-------
J»II)
FIX THE INITIAL CONDITIONS FOR ALL DDEs
1
Y10LD = 0.214E-3
Y20LD = 0.212
Y30LD = 0.915E-4
C40LD = 0.023E-3 :
6MOLD = 3.831
LMDLD * 1000.0/60.0
WRITE HEADING FOR OUTPUT !
WRITE (6,41)
FORMAT C1',5X, 'SECTION f',5X,'6flS FLOW,5X,'LIO FLOW, 5X,' TICS'
»,5X,'YC02',5Xf'COH-'l5Xl«YNH3'//)
BEBIN STEPPING ALONG THE THROAT OF THE VEHTURI \
DO 31 I = 2.NEON
B-36
-------
57 J = 1-1
58 H = STEP(J)
59 C
68 C CflLCULflTE INTERFflCIH. CONCENTRflTIONS
61 C
62 H2SI = P»YiOLD/HH2S
63 C02I = P»Y20LD/HC08
64 NH3I = P«Y30LD/HNH3
65 C
66 C CHOOSE THE SET OF DDEs TO INTEGRATE
67 C
68 IF (NH3I.GE.H2SI) THEN
69 NH3I = NH3I-H2SI
78 STOR = NH3HHNH3/P
71 IFLflS = 1
72 IF (J.EQ.1) THEN
73 C40LD = C40LD - H2SI
: « 1 "!i 74 END IF
' 75 CflLLSOLVEKC40LD,C02I,E2IIERROR )
76 IF (IERROR.EQ.1) THEN
77 PRINT *, 'ROOT FINDER DOES NOT CONVERSE'
78 STOP
79 END IF
M ELSE
81 IFLflS = 2
82 H2SI = H2SHH3I
83 CflLLSOLVE2(C40LD,C02I,H2SIfEl!E2lIERROR)
84 IF aERROR.6T.129) THEN
- - | 85 PRINT t, 'ROOT FINDER DOES NOT CONVERSE1
• 86 PRINT f, IER
87 STOP
88 END IF
89 END IF
9e c
91 C BEGIN INTESRflTINB flPPROPRIflTE ODE SET
92 C
'33 IF (IFLflB.EQ.1) THEN
94 KL1 = 2.»SQRT(Di/TME3 TERJC = KL2l£EiflREfl(J)tY2aj)fP/HC02
»* FLUX = TERMl+TERWtTERK
US C
Kfc C EULER'S METHOD
117 C
«« Y1NEH = Y11JL1HSTEP(J)*(Y10LIWLUX-TEW1)/6MLD
i 1W 97 Y3O = Y2OJHSTEP(J)f(Y20LI>fFLUX-TER«2)/6MLD
: • f ^ "« Y3NEM«Y30UHSTEP(J)*(Y30LDiFLUX-TERW)/aaj)
HI 910 = 6NOLD - STEP(J)iFUK
H2 LNNEU * LKLO + STEP(J)«FUIX
il3 C4NEH = C40JH5TEP(J)«2.liFUJX-TER»e)/{R*TEI«»»
' B-37 .
-------
11* * WOLD)
115 ELSE
»6 W-l = 2.0»SQRT(Di/(TINE(J)*PI))
117 KL£ = 2.«*SQRTCD2/ariHE(J)*PI))
H8 TERM = KLl*El*P*fiREft(J)«Y10LD/HH2S
119 . TERB2 = KL2«E2*P«flREfl-
Ul * TERM1))/(R*TEMP«6WLD)
i;32 END IF
133 C
134 C WRITE RESULTS flT THIS STEP
135 C
136 99 WRITE (6,42) STEP(J),9ICW,LMr&rYlNEW,Y2®l,C4«W,Y3NEH
137 42 FORHflT (' ',5X,F7.317X,E9.4I4XtE9.4,4(2XfEB,3))
iJrO C
139 C UPDflTE FUNCTION VflLUEES FOR THE NEXT ITERflTION
1^ C
141 Y10LD = Y1NEM
1*3 Y30LD = Y3NEH
144 C40LD = C4NEW
1*5 GMOLD = GNNEW
1*6 LKLD = LNNEU
147 C
14B 31 CONTINUE
141 C
159 STOP
151 END
152 C
155 C
15* SUMOUTINE PARW (N, NEON, BK, TINE, AREA, STEP)
1515 INTEGER N.NEDN
1» REflL VLBflR,eK(N)fTWe(N),H,flREft{N),VL(l),HK{12)tSTEP(N)
I?/ EXTEffW. FUNXN,FDU
15(1 C
1KI COMMON /BftSPRP/ V6,DENS6,VISCSIDIFFB
160 CONNON /DROP/ DP.DENSL.DENNOL
161 COMNON /CONDS/ TEMP,P,R
IK! COMMON /LOCflL/ 6,RflT
163; C
164 DATfl B,VULD/988.I,5B.W/
165 C
1«, RflTIOl = DENS6tDP/VISC6
167 RBTI02 = DENS67DENSL^)P
16fi RflT ^ I.T5»
169 MTI03 = DP/DIFF6
178 RATI04 = DIFFB/
-------
171
172
173
174
175
176
177
178
179
1188
181
1.82
IIB3
184
185
186
187
1.88
189
198
191
192
193
194
195
1%
197
198
C
C
C
C
C
C
C
C
21
C
SET PflRflMETERS FOR BEflR'S METHOD
M = 1
METH = 2
MITER = 2
INDEX = 1
IHK = M
H = 1.8E-7
TOL = 1.8E-5
Z = 8.8
ZEND = 8.81
VLU) = VLOLD
CflLCULflTE FOR EflCH SEGMENT DF THE THROflT THE flVERflBE
LIQUID VELOCITY, RESIDENCE TIME flND THE GflS PHflSE MftSS
TRflNSFER COEFFICIENT
WRITE (6,21)
FORMflT ('l',5X,'SECTIONf',5XI'flV6. LID. VEL',5X,'CONTCT. TIME'
* 5X,'6ftSM.TCOEFF.',5X,'flREfl'//)
DO 18 I = 2,NEQN
J = 1-1
STEP(J) = ZEND .
DIF = ZEND-Z
VL(1) = VLOLD
CALL DBEARIM.FUNXN.FDU 7 H.VL TBflJ TH irru UTTTD runn
m * IHK,HK,IER)
2B8 IF (IER.GE. 132) THEN
3W PRINT •, »6EflRS METHOD FfllLS - IER= '.IER
882 STOP
2«3 END IF
S84 C
2(fi Z = ZEND
2% IF (I.LT.7) THEN
2«7 ZEND = ZEND+«.81
208 60 TO 69
819 END IF
2118 IF (I.ED.7) THEN
211 ZEND = 8.1
212 60 TO 69
213 END IF
214 IF U.LT.ia> THEN
815 ZEND = ZEND+&1
216 ELSE IF (I.LT.22) THEN
217 ZEND = ZBBH«.2
218 ELSE IF (I.LT.26) THEN
219 ZEND = ZEND+8.25
220 ELSE
221 ZEND = ZEN&fl.8
222 END IF
223 69 VUCM = VL(1)
224 VLBflR« (VLME»*VLOLD)««.5
2?5 VLOLD * VLNEM
226 TIME(J) * DIFF/VLBflR
227 PECLET = RflTI03*flBS(VLBAR-VG)
B-39
-------
228
1229
230
231
1232
i>33
234
235
236
c!37
236
e!39
240
£41
242
£43
244
245
246
247
248
249
250
251
£52
253
254
255
2)6
257
SS&
259
260
261
3£
2£3
264
265
266
267
268
269
278
271
272
273
274
27,5
276
277
278
271)
280
281
282
281}
284
22
C
10
C
C
C
.C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
PECLET = 8. 991*PECLET«8. 3333
REYNLD = RATIOHABS(VG-VLBflR)
6KU) = «MTI04»PECLET*(1.8+0.25*REYNLD)«0.27
VOIDA6 = 1.0-315.8/(9.6«VLBflR)
AREfl(J) = 6.0*(1.0 - VOIDfl6)/DP
WRITE<6,22) STEP(J),VLBAR,TINE(J),6K(J),AREfl(J>
FORNflT C I,5X,F7.3,10X,E9.4,18X,E9.4,8X,E9.4,5X,E9.4)
CONTINUE
RETURN
END
SUBROUTINE HENRYS (T)
COWON /HENRY/ HC02,*2S,H«3
HENRY'S CONSTANT FOR H2S
DUMP = 182.325-4423. 11/T-36.6296«AL0610(T)+8.81387*T
HH2S = 18MDUKP
HH2S = HH2S*10ee.0
HENRY'S CONSTANT FOR C02
DUMP = 3.822-7.8665E-4«EXP(T/188.l3)-e.84145i
-------
285 CONST = (Rfl-1.8)/Rfl
286 Rfl = SORT(Rfl)
287 Q = C40LD/(2.8*C02I)
288 END IF
289 _
^ F2^HRftfE(F(6lESS2}»EXP(a3NST*fiUESS2w2)+£RF(6UESS2/Rfl)-r(
291 HF = (6UESS2-«£SS1)*8.5D8
ra ftV = 6UESS1+HF
1293 F=0*Rft*ERF(AV)»EXP(CONST«flV»flV)+ERF(flV/Rft)-1.8
S94 IF (F. EO. 8.8D8. OR . HF. LT. TOLL) THEN
295 ROOT = flV
296 £2 = 1.8/ERF(ROOT)
297 IERROR = 8
-298 60 TO 75
,PM(2)
33J> C
331' TERM = EXP<-X(1)«2)/ERF(X(D)
33H TER«2 * EXP(-X(2)«e)*PflR(i)/ERF(X<2))
33SI TERK3 = EXP(1.67t(X(l)«2-X(E)«2);
3411 TERH4 = ERF(1.67fX(D)
341 TER« = ERF(1.67*X(2))
-------
342
343
344
345
346
347
348
349
350
.351
352
353
.354
355
•^ff
C
C
C
c
c
TERM6 = EXPHU2)«2)*PflR(2)/(l.e-ERF(X(2»)
FU) = TERM1*TERH3+2.8»TERM2-TER«6
F(2) = {TERW*TERM3/(TERK5-1.8))f(TER«4-TERM5)-TERK2
RETURN
END
SUBROUTINE FUNXN (H.Z.VL.VLPRIN)
REflL VL(H),VLPRIK(M)
COMMON /LOCflL/ 6,WT
COMMON /BflSPRP/ VB,DENS6,VISCB,DIFB
DO 1 I = 1,N
VLPRW(I) = (6+RflT*(VB-VL(I))*(flBS(V6-VL(I)))««.16)/VL(I)
333 1 CONTINUE
m c
^1 RETURN
:
-------
SECTION t 6flS FLOW LIQ FLOW YH2S YCQ2 COH-
YNH3
m
.816
.828
.638
.848
.658
.668
.188
.288
.368
.488
.588
.768
.968
1.188
1.388
1.588
1.788
1.988
2.188
2.368
2.588
2.758
3.888
3.258
w* uvW
4.588
5.586
6.568
7.586
8.588
9.598
18.588
11.588
12.588
13.568
14.568
15.586
16.586
17.566
18.568
19.586
24.568
21.568
22.568
23.568
24.568
25.566
26.588
27.568
28.586
29.568
.3883+01
.3883+881
.3883+681
.3883+681
.3883+881
.3883+881
.3883+681
.3882+881
.3882+681
.3882+681
.3882+681
.3881+881
.3881+881
.3881+881
01
.3879+881
.3879+681
.3879+681
.3878+881
.3878+881
.3877+681
.3877+681
.3876+661
.3874+681
.3874+881
.3873+681
.3872+881
.3872+681
.3871+681
7ACJUM1
• wWUU'fWi
.38674481
• 36634001
.3861+661
• 38ovT001
.3859+681
.3857+881
.1667+02 .128-002 224+00 .230-04 .960-03 - initial conditions
.1667+882 .127-462 .228+688 .229-984 .947-683
.1667+682 .127-982 .228+888 .229-884 .938-883
.1667+882 .-27-882 .228+888 .229-684 .987-883
.1667+882 ..26-882 .228+888 .229-884 .888-883
.1667+882 .125-882 .228+888 .229-884 .858-883
.1667+882 .125-882 .228+888 .229-884 .844-883
.1667+882 .124-882 .228+888 .229-684 .837-883
.1667+882 .123-882 .228+888 .229-884 .827-683
.1667+882 .121-682 .226+888 .229-684 .816-683
.1667+882 .126-682 .228+888 .229-884 .884-683
.1667+682 .118-682 .228+668 .229-684 .798-883
.1667+682 .116-682 .228+888 .229-684 .775-883
.1667+882 .114-682 .228+868 .229-684 .759-683
.1667+882 .112-682 .228+888 .229-684 .741-683
.1667+882 .118-682 .226+888 .229-684 .724-663
.1667+682 .187-682 .228+888 .229-884 .786-683
.1667+882 .185-882 .228+888 .229-664 .687-883
.1667+882 .163-882 .228+688 .229-684 .669-683
.1667+882 .188-682 .228+688 .229-884 .658-683
.1667+882 .981-683 .228+888 .229-884 .632-683
.1667+682 .958-683 .226+688 .229-684 .614-683
.1667+682 .934-683 .226+888 .229-684 .596-683 '•
.1667+682 .911-683 .228+888 .229-664 .578-683
.1667+682 .887-683 .226+888 .228-684 .566-683
.1667+882 .665-683 .228+888 .228-684 .543-863
.1667+682 .839-683 .226+888 .228-684 .524-683
.1667+882 .815-683 .219+668 .228-684 .566-683
.1667+662 .792-883 .219+868 .228-664 .469-683
.1667+662 .771-663 .219+668 .228-664 .473-683
.1668+682 .751-463 .219+688 .228-664 .458-683
.1668+882 .731-683 .219+888 .228-684 .444-883
.1668+882 .714-883 .219+688 .228-684 .431-683
.1668+862 .697-883 .219+888 .228-684 .419-863
.1668+682 .681-683 .219+688 .228-664 .467-683
.1668+682 .666-663 .219+686 .228-684 .397-663 ;
.1668+882 .652-683 .219+666 .228-684 .387-683 :
.1668+682 ,639-683 .819+668 .228-664 .377-663
.1668+662 .627-663 .218+666 .228-664 .369-663
.1668+662 .616-663 .218+666 .227-664 .361-663
.1668+662 .665-663 .210+488 .227-664 .353-683
.1668+M2 .595-683 .£16+869 .227-464 .346-663
.1668+662 .586-463 .gl8+8N .227-684 .346-663
.1669+662 .578-663 .218+666 .227-684 .334-663
.1669+662 .572-663 .217+680 .227-464 .336-663
.1669+662 .567-663 .217+666 .227-664 .326-683
.1669+662 .563-683 .217+484 .226-664 .324-663
.1669+662 .556-663 .217+668 .226-664 .319-683
.1669+662 .549-663 .216+686 .226-664 .314-663
.1669+682 .542-663 .216+668 .226-464 .369-663
.1669+662 .533-683 .216+446 .226-464 .363-663
.1669+682 .525-663 .216+446 .225-464 .298-463
6PRTH3PE
B-43
-------
SECTION t WG. LIQ. VEL CONTCT. TIKE BflS H.T COEFF.
AREA
.000
UARNIN6 WITH
.010
.020
.030
.040
.050
.060
.100
.200
.300
.400
.500
.700
.900
1.100
1.300
1.500
1.700
1.900
2.100
2.300
2.500
2.750
3.000
3.250
3.500
4.500
5.500
6.500
7.500
8.500
9.500
10.500
11.500
12.500
13.500
14.500
15.500
16.500
17.500
18.500
19.500
20.500
21.500
22.500
23.500
24.500
25.500
26.500
27.500
28.500
29.500
. 50+02
FIX ERROR (IER =
.1429+004
.3339+004
.4260+004
.4964+004
.5548+004
.6052+004
.7027+004
.8942+004
.1088+005
.1219+005
.1317+005
.1422+005
. 1528+005
.1604+005
.1660+005
.1703+005
. 1737+005
.1763+005
.1785+005
.1802+005
.1817+005
.1830+005
.1842+005
.1851+005
.1859+005
.18714005
.1884+005
.1891+005
.1895+005
. 1897+085
. 1898+005
.1899+085
.1899+005
.1899+005
• 1 7004*005
.19004605
* l«nM»rf^53
* 1 «nW' wv3
.1900+005
• liMWrfMt)
• l^nwr^VD
• IVAvMMD
.1900+005
« l^W WflD
• i ^HI 00D
• 17004005
.19004005
• 19001005
• 1
-------
PART III
•COMPUTATIONAL STUDIES
OF THE SIMULTANEOUS CHEMICAL ABSORPTION
OF THREE GASEOUS COMPONENTS
INTO A REACTIVE LIQUID
B-45
-------
The situation considered here is the simultaneous absorption of
three gases into a liquid containing a nonvolatile solute with
which two of the gases react; the third gas reacts with both the
absorbed gases in the liquid phase, but not in the gas phase.
Such a situation occurs in the absorption of a gas containing
C02, H2S and NH3 in an alkaline solution. C02 and H2S react with
the alkali and the dissolved NH3 reacts with the dissolved C02
and H2S.
Let A = C02
B = H2S '
C = OH-
E = NHo - .
The reaction scheme considered is:
A + nxC -»• Pls rate = k1AC
B + n2C -»• P2, rate = k2BC
A + n3E -»• P3, rate = k3AE
B + n4E -»- P4, rate -: k4BE
A material balance over a differential element of liquid results
in the unsteady state diffusion equations with reaction terms.
If =>A 7 - klAC - k3AE
B-46
-------
= D
- k2BC -
~
3E
-9-
oX
9f = DE —2 - n3k3AE - n4k4BE
oX
The boundary conditions are:
a(x,0)
B(x,0)
C(x,0)
E(x.,0)
= 0 ;
= 0 ;
• co ;
= 0 ;
A(0,t) = A1 ;
B(0,t) = B1 ;
-§ (0,t) = 0 ;
A(-,t) = 0 ;
B(-,t) = 0 ;
Q
E(0,t) = Ei ; E(-,t) = 0 .
The following dimensionless variables are introduced:
1/2
n -
x ,
e = k2C0t,
a - A
-
e =
B-47
-------
Pi = ki/k2 ,
P3 = k3/k2» P4 = k4/k2
rA = DA/DB
rc = DC/DB,
rE = DE/DB
m.
niAi
m
B = n2Bi/C0 >
m,
ni '
The terms in equation 1 become:
JA _ . aa
at Ai Te
kjAC =
k3AE = k3AiEiae ;
=---
at iae at " 2oi
B-48
-------
92B
v 32b . 3n
"
k2C0Bi 32b
k2BC =
be
3C =
"aT
3c 36
Te
2 3c
= c
2 0 2
^ u ^
k2C0 32C
n ?"
UB 3xZ
ac
***
Substituting these relations into Equation 1, we get the
dimensionless form of the model equations:
***
3E
3e
36
Te
32E
_ r e ,71x
~ ~
k2C0Ei 32e
n3k3AE = n3k3A.Ei ae ;
= n4k4B(JE1 be ;
B-49
-------
aa _ 3 a
"ae ' rA TT
ac -
ab
. . .
' bc * b4mE be
ac a c .
Te = rc T~? " bimAac - mBbc
OT\
ae
" r
ae 'E
' b3q3mAae -
a(n,0) = 0 ;
b(0,e) = 1 ;
a(»,e) = 0 ;
b(n,0) =
b(0,e) = 1 ;
c(n,0) = 1;
c(-,e) = 1 .
e(n,0) = 0 ;
e(0,e) = 1 ;
:e(»,e) = 0
B--50.
-------
These nonlinear coupled PDEs must be solved numerically to give
concentration profiles as function of 0 and other parameters;
however, before they can be solved numerically, the boundary
condition at n •»• » has to be eliminated by an additional
variable transformation. Let
15 „ -1
"
2 _ ,.3
" *
and
n)3
2 97
a _ /a^2 3^ a^c a
2" ~ ^TV' --- 7 -- 7 T
ay 3y ar ay
with this transformation, equation 2 now becomes:
Te = rA5 7~l + 2rA5 If - Plac - P3mEae
-|| = 54 l!| + 2c3 Jb _ bc
B-51
-------
2
1-| + 2YE?3 || - P3q3mAae -
a(n,0) = 0; a(0,e) = 0 ; a(l,e) = 1;
c(n,0) = 1 ; c(0,e) = 1 ; -|| (1,0) = 0 ;
e(n,0) = 0 ; e(0,e) = 0 ; e(l,e) = 1 .
Parameter values for which I is to be solved:
= rc = rE = X " DA = DB = DC = DE = 2>0 x
ii) TII = 1 ; n2 = 2 ; n3 = 1 ; n4 = 2 ;
111) Ai = .02 m/lt, Bi = .002 m/U, E1 = .0016 m/lt; GQ= .2 m/lt;
iv) kx = 1.0 x 103 It/m s ; i
B-52
-------
v)vary t, pj , p3, p4 suitably
MASS TRANSPORT PARAMETERS:
The instantaneous rate of absorption of A and B is
and
N = - D -
A UA 3x
x=0
= - n
DB T7
x=0
9a
195
5 = 1
ab
—
Nc = - Dc --
E E 3x
x=0
1/2
je
a?
5 =
-
_ (VBi)
5=1
i) For a given parameter set, look at S vs t; is there an
optimum contact time?
ii) For a given contact time, other parameters remaining
B-53
-------
constant, look at S vs. p; does the effect of one reaction
being much faster 'level off?
B-54
-------
0)
u
W
•H
•O
* 8
tt) 0)
Tl 03
•H
XI H.
O O'
•H •
T3 O
C >i-
O 1-1'
£! 0)
^ 4-)
(d (0
u e
I
W 0)
CU 4-J
2^
S-) rt! ;
>i Q)
SI'
O M
dJ
0) 4J
H C
•H -H
8
cu
IT)
I
m
Q)
^
B-55
-------
LSODE, RUN NO. 1 - TEST CASE FOR NEW APPROACH
INITIAL T = .000
FINAL T = .200 + 003
INTERVAL T = .200 + 002
NUMBER OF ODES = 44
INTEGRATION ALGORITHM = 2
1 - NONSTIFF (MF = 10)
2 - STIFF (MF = 22)
ERROR CRITERION - REL
MAXIMUM ERROR = .100-006
B-56
-------
INPUT PARAMETERS ARE
INTERFACIAL CONG. OF COMPONENT A (AI) = .200000-001
INTERFACIAL CONC. OF COMPONENT B (BI) = .200000-002
INTERFACIAL CONG. OF COMPONENT E (El) = .160000-002
INITIAL CONCNTRATION OF LIQUID C (CO) = .200000+000
REACTION RATE CONSTANT FOR A+C—P (KA) = .100000+004
DIFFUSIVITY (D) = .200000-004
RATIO OF KA/KB (PI) = .100000-001
RATIO OF K3/K2 (P3) = .100000-001
RATIO OF K4/K2 (P4) = .100000+001
MOLES/LT C02
MOLES/LT H2S
MOLES/LT NH3
MOLES/LT OH~
LT/MOLE*S
CM**2/S
DIMENSIONLESS
DIMENSIONLESS
DIMENSIONLESS
AT DIMENSIONLESS TIME = .000000
POSITION
.000000
.100000+000
.200000+000
.300000+000
.400000+000
.500000+000
.600000+000
.700000+000
.800000+000
.900000+000
.100000+001
CONG. OF A
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
CONC. OF B
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
CONC. OF C
.100000+001
.100000+001
.100000+001
.100000+001
.100000+001
.100000+001
.100000+001
.100000+001
.100000+001
.100000+001
.100000+001
CONC. OF D
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
AT DIMENSIONLESS TIME = .200000+002
POSITION
.000000
.100000+000
.200000+000
.300000+000
.400000+000
.500000+000
.600000+000
.700000+000
.800000+000
.900000+000
.100000+001
CONC. OF A
.000000
.294016-003
.573658-002
.229578-001
.598230-001
.119076+000
.211053+000
.336968+000
.508210+000
.725232+000
.100000+001'
CONC. OF B
.000000
.878867-005
.507091-003
.508065-002
.238339-001
.663402-001
.145651+000
.267508+000
.446546+000
.685229+000
.100000+001
CONC. OF C
.100000+001
.999993+000
.999837+000
.999637+000
.999254+000
.998972+000
.998557+000
.998324+000
.998005+000
.997918+000
.997784+000
CONC. OF D
.000000
.320827-003
.609640-002
.236652-001
.608238-001
.120241+000
.212272+000
.338080+000
.509087+000
.725723+000
.100000+001
AT DIMENSIONLESS TIME = .400000+002
B-57
-------
POSITION
CONG. OF A
CONG. OF B
CONG. OF G
CONG. OF D
.000000
.100000+000
.200000+000
.300000+000
.400000+000
.500000+000
.600000+000
.700000+000
.800000+000
.900000+000
.100000+001
.000000
.381004-003
.631849-002
.233125-001
.603422-001
.119459+000
.211469+000
.337276+000
.508463+000
.725367+000
.100000+001
.000000
.878720-005
.507203-003
.508038-002
.238355-001
.663400-001
.145654+000
.267509+000
.446549+000
.635230+000
.100000+001
.100000+001
.999991+000
.999766+000
.999626+000
.999178+000
.998959+000
.998480+000
.998310+000
.997928+000
.997904+000
.997707+000
.000000
.435383-003
.686030-002
.241300-001
.615044-001
.120743+000
.212818+000
.338483+000
.509418+000
.725899+000
.100000+001
AT DIMENSIONLES TIME = .600000+002
POSITION
CONG. OF A
CONG. OF B
CONG. OF C
AT DIMENSIONLESS TIME = .800000+002
CONG. OF D
000000
100000+000
200000+000
300000+000
,400000+000
500000+000
,600000+000
,700000+000
,800000+000
,900000+000
,100000+001
.000000
.411941-003
.642041-002
.233804-001
.604340-001
.119529+000
.211544+000
.337332+000
.508508+000
.725391+000
.100000+001
.000000
.878584-005
.507283-003
.508009-002
.238366-001
.663395-001
.145656+000
.267509+000
.446551+000
.685230+000
.100000+001
.100000+001
.999995+000
.999726+000
.999631+000
.999136+000
.998965+000
.998437+000
.998316+000
.997885+000
.997909+000
.997664+000
.000000
.485275-003
.702366-002
.242392-001
.616516-001
.120856+000
.212937+000
.338573+000
.509490+000
.725938+000
.100000+001
POSITION
.000000
.100000+000
.200000+000
.300000+000
.400000+000
.500000+000
.600000+000
.700000+000
.800000+000
.900000+000
.100000+001
CONG. OF A
.000000
.427272-003
.644276-002
.234000-001
.604555-001
.119548+000
.211561+000
.337346+000
.508519+000
.725397+000
.100000+001
CONG. OF B
.000000
.878485-005
.507335-003
.507987-002
.238372-001
.663391-001
.145657+000
.267509+000
.446552+000
.685230+000
.100000+001
CONG. OF C
.100000+001
.100000+001
.999704+000
.999640+000
.999112+000
.998973+000
.998413+000
.998325+000
.997861+000
.997918+000
.997640+000
CONG. OF D
.000000
.515536-003
.706740-002
.242778-001
.616935-001
.120893+000
.212971+000
.338601+000
.509512+000
.725950+000
.100000+001
AT DIMENSIONLESS TIME = .1000000+003
B-58
-------
POSITION
CONG. OF A
CONG. OF B
.€ONC. OF C
CONG. OF D
.000000
.100000+000
.200000+000
.300000+000
.400000+000
.500000+000
.600000+000
.700000+000
.800000+000
.900000+000
.100000+001
.000000
.436048-003
.645007-002
.234085-001
.604631-001
.119556+000
.211568+000
.337352+000
.508523+000
.725399+000
.100000+001
.000000
.678412-005
.507368-003
.507971-002
.238376-001
.663366-001
.145658+000
.267508+000
.446552+000
.685230+000
.100000+001
.100000+001
.100001+001
.999691+000
.999648+000
.999098+000
.998982+000
.998399+000
.998334+000
.997847+000
.997927+000
.997626+000
.000000
.536724-003
.708486-002
.242982-001
.617116-001
.120911+000
.212987+000
.338615+000
.509522+000
.725956+000
.100000+001
AT DIMENSIONLESS TIME = .120000+003
POSITION
CONG. OF A
CONG. OF B
CONG. OF C
AT DIMENSIONLESS TIME = .140000+003
CONG. OF D
000000
100000+000
200000+000
300000+000
400000+000
500000+000
600000+000
700000+000
800000+000
900000+000
100000+001
.000000
.441300-003
.645351-002
.234131-001
.604668-001
.119560+000
.211571+000
.337355+000
.508525+000
.725401+000
.100000+001
.000000
.878357-005
.507391-003
.507958-002
.238379-001
.663383-001
.145658+000
.267508+000
.446553+000
.685230+000
.100000+001
.100000+001
.100001+001
.999684+000
.999657+000
.999090+000
.998990+000
.998391+000
.998342+000
.997839+000
.997935+000
.997618+000
.000000
.552230-003
.709493-002
.243119-001
.617226-001
.120923+000
.212997+000
.338623+000
.509528+000
.725959+000
.100000+001
POSITION
.000000
.100000+000
.200000+000
.300000+000
.400000+000
.500000+000
.600000+000
.700000+000
.800000+000
.900000+000
.100000+001
AT DIMENSIONLESS
CONG. OF A
.000000
.444482-003
.645546-002
.234159-001
.604689-001
.119562+000
.211573+000
.337356+000
.508527+000
.725401+000
.100000+001
CONG. OF B
.000000
.878314-005
.507407-003
.507948-002
.238381-001
.663380-001
.145659+000
.267508+000
.446553+000
.685230+000
.100000+001
CONG. OF C
.100000+001
.100002+001
.999679+000
.999664+000
.999086+000
.998997+000
.998386+000
.998349+000
.997834+000
.997942+000
.997613+000
CONG. OF D
.000000
.563719-003
.710186-002
.243217-001
.617302-001
.120931+000
.213003+000
.338629+000
.509532+000
.725962+000
.100000+001
TIME = .160000+003
B-59
-------
POSITION
CONG. OF A
CONG. OF B
CONG. OF C
CONG. OF D
.000000
.100000+000
.200000+000
.300000+000
.400000+000
.500000+000
.60.0000+000
.700000+000
.800000+000
.900000+000
.100000+001
.000000
.446416-003
.645661-002
.234175-001
.604702-001
.119563+000
.211574+000
.337357+000
.508527+000
.725402+000
.100000+001
.000000
.878280-005
.507420-003
.507940-002
.238382-001
.663378-001
.145659+000
.267508+000
.446553+000
.685230+000
.100000+001
.100000+001
.100003+001
.999676+000
.999670+000
.999083+000
.999004+000
.998383+000
.998355+000
.997831+000
.997949+000
.997610+000
.000000
.572258-003
.710690-002
.243290-001
.617358-001
.120937+000
.213008+000
.338634+000
.509535+000
.725963+000
.100000+001
AT DIMENSIONLESS TIME = .180000+003
POSITION
CONG. OF A
CONG. OF B
CONG. OF C
AT DIMENSIONLESS TIME = .200000+003
POSITION
CONG. OF A
CONG. OF B
CONG. OF C
CONG. OF D
000000
100000+000
200000+000
300000+000
400000+000
,500000+000
,600000+000
,700000+000
,800000+000
,900000+000
,100000+001
.000000
.447592-003
.645732-002
.234185-001
.604710-001
.119564+000
.211575+000
.337358+000
.508528+000
.725402+000
.100000+001
.000000
.878252-005
.507429-003
.507934-002
.238383-001
.663376-001
.145659+000
.267507+000
.446553+000
.685230+000
.100000+001
.100000+001
.100003+001
.999674+000
.999676+000
.999081+000
.999009+000
.998381+000
.998361+000
.997829+000
.997954+000
.997608+000
.000000
.578610-003
.711063-002
.243343-001
.617399-001
.120942+000
.213012+000
.338637+000
.509538+000
.725965+000
.100000+001
i
CONG. OF D
000000
100000+000
200000+000
,300000+000
,400000+000
,500000+000
,600000+000
,700000+000
.800000+000
,900000+000
,100000+001
.000000
.448307-003
.645775-002
.234191-001
.604715-001
.119565+000
.211575+000
.337358+000
.508528+000
.725402+000
.100000+001
.000000
.878229-005
.507436-003
.507928-002
.238384-001
.663374-0.01
.145659+000
.267507+000
.446553+000
.685230+000
.100000+001
.100000+001
.100004+001
.999673+000
.999680+000
.999079+000
.999014+000
.998380+000
.998366+000
.997828+000
.997959+000
.997606+000
.000000
.583336-003
.711341-002
.243384-001
.617430-001
.120945+000
.213015+000
.338640+000
.509539+000
.725966+000
.100000+001
B-60
-------
UNIVERSITY OF UTAH COMPUTER CENTER 1100/61-H2 1100/OS:38R5MP/PROD514-74 BUILD-1
RUNID: SRINI ACCOUNT 581000 EXEC-ID: 38R5MP/PROD5
PROJECT: SELECTIVITY OPTS: CPTN USER-ID: 352700 P/S: S/CR1
EST. TIME: 02:00:00 EST. PAGES: 50 EST. CARDS: 0
*PROGRAM: SUBMIT
INITIATION TIME: 02:29:02 NOV 20, 1984
*I/0 TIME:. 00:00:00.086 ER TIME:
*CPU TIME: 00:00:00.012 AVE-MEM-SIZE:
*TOT UNITS: 00:00:00.849 CONDITION WORD:
TERMINATION TIME: 02:29:03 NOV 20, 1984
*PROGRAM: SUBMIT/CURRENT
*INITIATION TIME: 02:29:03 NOV 20, 1984
*I/0 TIME: 00:00:00.305 ER TIME:
*CPU TIME: 00:00:00.011 AVE-MEM-SIZE:
*TOT UNITS: 00:00:00.876 CONDITION WORD:
TERMINATION TIME: 02:29:05 NOV 20, 1984
*PROGRAM: CONNECT
*INITIATION TIME: 02:29:05 NOV 20, 1984
*I/0 TIME: 00:00:00.040 ER TIME:
*CPU TIME: 00:00:00.003 AVE-MEM-SIZE:
*TOT UNITS: 00:00:00.078 CONDITION WORD:
*TERMINATION TIME: 02:29:06 NOV 20, 1984
*PROGRAM: CONNECT
*INITIATION TIME: 02:29:07 NOV 20, 1984
*I/0 TIME: 00:00:00.071 ER TIME:
*CPU TIME: 00:00:00.003 AVE-MEM-SIZE:
*TOT UNITS: 00:00:00.284 CONDITION WORD:
TERMINATION TIME: 02:29:08 NOV 20, 1984
*PROGRAM: ILSODE/DOUBLE
*INITIATION TIME: 02:29:08
*I/0 TIME: 00:00:00.360
*CPU TIME: 00:01:13.633
*TOT UNITS: 00:01:15.763
NOV 20, 1984
ER TIME:
AVE-MEM-SIZE:
CONDITION WORD:
TERMINATION TIME: 02:34:48 NOV 20, 1984
#FILE: 352700*SUBMIT$PF(1)
SPINI FIN
;I/0 TIME:
;CPU TIME:
;TOT UNITS:
00:00:00.751
16K
0000,0000,0000
00:00:02.560
45K
0000,0000,0000
00:00:00.035
9K
0000,0000,0000
00:00:00.209
9K
0000,0000,0000
00:00:01.770
44K
0000,0000,0000
DELETED: 02:34:48 NOV 28, 1984
00:00:01.113
00:01:13.662
00:01:24.153
ER TIME:
AVE-MEM-SIZE:
KILO-MEM-SECS:
00:00:09.378
41K
3493.212
BILLING INFORMATION 1100/61 UNITS: 83.017
TOTAL-RUNS B/D 114/019 AS OF 02:11: NOV 20, 1984 [
CARDS-IN: 22, OUT: 0 PAGES: 16 TAPEES - LIB/OC:00/00
*EST. COST: $2.08 USED BY USR-ID: $1673.80 USR-ID HAS LEFT: $2026.20
ACCT-BAL $1697.14 AT 09:06 NOV 19, 1984 EXPIRATION*23:59 DEC 31, 1986
B-61
-------
Nomenclature '
A = local venturi cross sectional area, cm
Cc| = drag coefficient
dp = droplet diameter, cm
En = enhancement factor for component n in liquid.
e = local void fraction in venturi ',
p ; -
g = acceleration due to gravity, cm/s :
i
GQ = volumetric gas flow rate, cm^/s
kg = gas phase mass-transfer coefficient, gmol/s cm2 atm
kL = liquid phase physical mass-transfer coefficient, cm/s
LM = molar liquid mass velocity gmol/s cm
LQ = volumetric liquid flow rate, cm3/s
Np = mass transfer flux of component n, gmol/s cm^
S 2
Nt = total mass flux = ( E Nn), gmol/s cm
n = 1 ;
P = total pressure, atm
p = average molal liquid density, gmol/cm3 ;
PG = density of gas, g/cm3
PL = density of liquid, g/cm
S = number of species transferred from gas to liquid
t = contact time, s
E
Vq = gas velocity, cm/s . ;
Vj_ = liquid velocity (droplet velocity), cm/s
Xn = mole fraction of n in the liquid (mixing cup)
Xni = mole fraction of n in the liquid at the interface
*
Xn = liquid phase mole fraction of A that would be in
equilibrium with gas of mole fraction Yn.
y
'n = mole fraction of n in gas
B-62
-------
Yn.j = mole fraction of n at the interface in the g;as
*
Yn = gas phase mole fraction of n in equilibrium with bulk
concentration in liquid n
z = absorber length, cm
B-63
-------
TECHNICAL REPORT DATA
(Pleate nag Imurucn'ons on tht rtvene before completing)
1. REPORT NO.
EPA-
. TITLE AND SUBTITLE
Alkaline and Stretford Scrubbing Tests for H2S
Removal from In-Situ Oil Shale Retort Offgas
7. AUTHOR^)— ~
H. J. Taback, G. C. Quartucy, R. J. Goldstick
9. PERFORMING ORGANIZATION NAME AND ADDRESS
KVB, Inc.
18006 Skypark Boulevard
Irvine, California 92714
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Air and Energy Engineering Research Laboratory
Research Triangle Park, NC 27711
15. SUPPLEMENTARY NOTES
3. RECIPIENT'S ACCESSION NO.
6 REPORT DATE
B. PERFORMING ORGANIZATION CODE
I. PERFORMING ORGANIZATION REPORT NO
KVB72 807430-1982
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
EPA 68-03-3166
13. TYPE OF REPORT AND PERIOD COVERED
Final 1/84 - 12/84
14. SPONSORING AGENCY CODE
EPA/
16. ABSTRACT
Two mobile pilot-plant scrubbers were evaluated for removing reduced
sulfur compounds from the offgas of an in-situ retort at Geokinetics. The alkaline
scrubber had a tray tower arid a venturi contactor used alternately with NaOH, KOH
and NH OH to investigate the effects of scrubbing chemical, chemical concentration
and residence time on removal efficiency and H S selectivity. The Stretford plant
employed a venturi contactor. Near the end of the test, a packed-tower contactor
was added downstream of the venturi. The Stretford test objectives:were to repeat
a 99+ percent removal efficiency observed on the previous test and to attempt to
explain some lower removal efficiencies observed prior to that. The alkaline scrub-
ber efficiency varied inversely with selectivity. At high solution concentration
in the tower, 94 percent removal was achieved at a selectivity of 9. At low concen-
tration in the venturi the removal was 50 percent and the selectivity was 79. The
Stretford achieved 99+ percent removal with the packed tower and 95 percent with
the venturi. A computer model of the alkaline scrubber based on the penetration
theory was developed and agrees well with the observed performance. Based on this
model, it appears possible to design an alkaline scrubber system including a Glaus
plant which can achieve 95 percent H2S removal at a selectivity of 37.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Oil Shale Emissions
Control Technology
Caustic Scrubber
Alkaline Scrubber
Aqueous Ammonia Scrubber
ps. DISTRIBUTION STATEMENT
Release to Public
EPA Form 2220-1 (»-73)
b.lDENTDFIERS/OPEN ENDED TERMS
Hydrogen Sulfide Control
Control Technology
Performance
Penetration Theory
Analysis
19. SECURITY CLASS (THitRtpori)
Unclassified
30. SECURITY CLASS (ThLt fegej
Unclassified
c. COSATI Field/Croup
21. NO. OF PAGES
22. PRICE
B-64
------- |