ALKALINE SCRUBBING OF IN-SITU OIL
RETORT OFFGAS AT GEOKINETICS
HAL TABACK, P.E.
CONSULTING ENGINEER
KVB, INC.
IRVINE, CALIFORNIA
ROBERT GOLDSTICK
PRESIDENT
ENERGY DESIGN SERVICE
OJAI, CALIFORNIA
EDWARD BATES
OIL SHALE MANAGER
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO
P-352
SHALE
PRESENTED TO THE
18th ANNUAL OIL SHALE SYMPOSIUM
/ AIRPORT HILTON HOTEL
GRAND JUNCTION, COLORADO
APRIL 22-24, 1985
-------
NOTICE
This work was sponsored by the U.S. Environmental Protection Agency and
was performed under subcontract to MeteaIf fi Eddy, Inc., Boston, i
Massachusetts, under EPA Contract No. 68-03-3166. It has been subject to the
Agency's peer and administrative review, and it has been approved for
publication as an EPA document. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
-------
ALKALINE SCRUBBING OF IN-SXTU OIL SHALE
RETORT OFFGAS AT GEOKINETICS
By
Hal Taback, P.E.
Consulting Engineer
KVB, Inc.
Irvine, California 92714
Robert Goldstick
| President
Energy Design Service
;ojai, California
Edward Bates
Oil Shale Manager
U.S. Environmental Protection Agency
Cincinnati, OH
ABSTRACT
The EPA's «obile wet scrubber was need on a 200
ACFM slipstream of the Geokinetics retort offgas to
investigate the R^s removal efficiency and selectivity
(percent H^ removal/percent CO2 removal) as a function
of liquid/gas contact time, alkaline solution OH"
concentration, and the specific scrubbing chemical. A
venturi and a tray tower were used to produce contact
times of approximately 0.003 and 0.2 second,
respectively. Three alkaline solutions, HaOH, KOH, and
NH4OH were employed on each contactor at various
concentrations for a total of 22 runs. To analyze
these results and provide design criteria for future
alkaline scrubbers a sophisticated computer model
employing the penetration theory for liquid-phase mass
transfer was developed.
INTRODUCTION
Oil shale facilities proposed for Colorado and Utah
will produce substantial quantities of BjS and other
sulfur gases which could impact Class X airsheds such
as the Flattopn wilderness area. The Clean Air Act
requires stringent control of such •missions through
the use of best available control technology under PSD
permits. This report provides data characterizing in-
situ oil shale offgases from the Geokinetics (Seep
Ridge) plant in eautem Utah and assessing the
effectiveness at alkaline scrubbing processes in
-.
controlling the emission of BjB and other sulfur
compounds. This renults should assist developers and
permit writers in selecting appropriate controls for
the treatment of oil shale offgases. '
The offgas from the horizontal in-situ retort at
Geokinetics, in eastern Utah, contains approximately
0.15 percent (1500 ppnv) of H2s, 22 percent CO2, ».nd
0.10 percent NH3 in addition to N2 (60 percent), H2 (9
percent), CO (5 percent), CH4 (1.5 percent), and other
(2.25 percent). While these percentages are presented
on a dry basis, the offgas is actually saturated with
moisture. Also present, at levels of 0 to 10 ppmv
each, are organic sulfur species such as carbonyl
sulfide, mercaptans, thiophenes, and carbon
disulfide. Lovell, et al. (1962) and Desai, et al.
(1983) evaluated various sulfur-control processes and
concluded that caustic (NaOH) scrubbing could be a
candidate process if the selectivity of the scrubbing
process were sufficiently high. Selectivity here is
defined as the percent removal of H,S (and the organic
* \
sulfur) divided by the percent removal of the CO2.
Both HjS and CO2 are acid gases and CO2 i« present in a
concentration 150 times greater than that of H2s.
Therefore, any caustic or other alkaline scrubbing must
take advantage of relative solubilities and reaction
rates to achieve a high selectivity of HjS relative to
co2.
The full sulfur removal scheme for employing an
alkaline scrubber is shown in Figure 1. The retort
i
offgas enters the alkaline scrubber and gives up H2s
and CO2 to the scrubber liquid. The scrubber liquid is
cycled through a regenerator or stripper where the H2s
-------
flC
u_
CO
CO
o
§N
I f*. ^
CM
o
o
CO
CM
I
o
z
CO
CO
LU
O
O
DC
Q.
cc
o
CE
LJ
Z
111
O
LU
CC
O
•p
0)
cs
3
4J
•H
to
O
•H
4J
U
U)
10
in
0)
o
o
o
3
14
CO
<
O
I
as
|S
Si
f
HI
2o
_i in
cc
0)
•l-l
rt
10
A;
tSS
Sgo
DO"8
5WCO
UJU.
-------
•nd CO2 are distilled off and sent to a sulfur recovery
plant (such as Claus) where solid sulfur is produced
and the CO-, is released to the atmosphere. In this
system an important parameter is the H2S concentration
in the feed gas to the Claus or other sulfur recovery
plants. Dcsai, et al. (1983) indicated that the Claus
process night work with a feed of B percent H2s but at
least 15 percent was needed for confidence and 25
percent or higher was desirable. »e sulfur recovery
plant inlet: gas H2S concentration is equal to the
product of the H2s/CO2 ratio in the retort offgas times
the selectivity factor "for the alkaline scrubber. For
an oil shale offgas such as that from Geokinetics, the
•electivity required in the alkaline scrubbing process
can be calculated from the ratio of H2s to CO2 in the
retort offgas and the desired percentage of H2S in the
feed to the sulfur recovery plant. So»e values are:
Sulfur Plant Inlet Gas
H2s Percentage Desired
8>
15
25
Scrubber
Selectivity
Required
12
23
38
Thus, a scrubber selectivity of over 40 is desirable,
25 could be acceptable, and 10 is marginal for retort
offgas similar to Geokinetics.
Selectivity, while important, is only one
performance parameter determining H^ removal.
Unfortunately, many of the factors influencing scrubber
performance have conflicting effects. Typical examples
•re that increasing the hydroxyl ion concentration
(IOH~J) in the scrubbing solution and increasing the
gas-to-liquid contact time may increase removal
•fficiency of H^ while decreasing selectivity of H2S
relative to CO2.
To investigate these effects the EPA sponsored a
field testing program which is reported in detail by
Tabaclc, «t; al. (1985). This paper presents a summary
of those results and conclusions.
EXPERIMENTAL APPROACH
The EPA's nobile wet scrubber was installed at the
Geokinetics site to process a 220 ACFM slipstream of
retort offgas. The objectives of these tests were to
measure H<>S removal .efficiency and selectivity as a
function of (1) liquid/gas contact time, (2) scrubbing
solution [OB~], and (3) specific scrubbing chemical.
The mobile scrubber was equipped with both a venturi
and a tray tower contactor which produced liquid/gas
contact times of approximately 0.003 sec and 0.2 sec
respectively. Alkaline solutions of NaOB, KOH, and
KHfOH were employed alternately on each of the
contactors at various concentrations as summarized in
Table 1. A run consisted of discharging the contents
of the 1-n3 mix tank once through the\contactor for a
period of approximately 40 minutes. This simulates the
operation of the scrubber module in the system shown in
Figure 1.
The retort offgas was sampled upsjtreaa and
downstream of the scrubber and analyzed for specific
sulfur compounds and total reduced sulfur. The
sampling and the analytical procedure? that were used
for the specific reduced sulfur compounds are
essentially those specified in EPA Methods 15 and 16
(40 CFR 60, Appendix A, July 1, 1982);. This method
employs a gas chromatograph (GC) with a flame
photometric detector (FPD). In this procedure, a
continuous gas sample is extracted from the emission
source at a known rate, scrubbed to remove SO2 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. In addition, continuous real-time
analyses of total reduced sulfur (TRS) in the retort
offgas were made by first removing SO2« oxidizing the
•ample gas stream in a tube furnace, and then reading
the total sulfur as SO2 using a TECO!continuous SO2
monitor. This technique was derived from EPA Method
ISA (40 CFR 60, Appendix A, July, 19$2) and was used to
provide a real-time display of removal efficiency
during the test runs.
t
i
EXPERIMENTAL RESULTS
The test results are summarized in Table 2 and
Figure 2. Three different solution concentrations were
used for each alkali except for the last four runs (No.
19-22) where only the tower was used to make two high
concentration runs for both HaOH and KOH.
As expected, it was found that the highest
•electivity (percent removal of HjS divided by percent
removal of CO2> »as obtained*at the lowest solution
concentrations and at the shorter liquid/gas contact
times (i.e., with the venturi contactor). Conversely,
the highest HjS removal efficiencies were obtained at
the higher solution concentrations and the longer
contact times (i.e., with the tray tower contactor).
Figures 3 and 4 Bhow the variation of Bf removal
efficiency with lOH"). A limit of 94 percent removal
efficiency was reached at an loir) 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
2). At the low IOH-) of O.012 grami«ole/liter the
nelectivity with the venturi reached as high as 79 with
a removal efficiency of just over 50 percent.
The effect of IOH"] on selectivity for the venturi
and the tower is plotted in Figure 5. Mote that the
venturi selectivity is more sensitive to tOH~] than
that of the tower.
-------
TABLE 1. SCRUBBING TESTS CONDUCTED
Run No.
Alkali
Cone.
gmoles/liter
Contactor
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Ammonia
it
H
n
M
•
NaOH
ft
M
*
•
H
KOH
M
tl
M
M
M
H
•
NaOH
•
2.0
If
0.05
H
0.3
• H
0.05
H
0.012
«
0.023
M
0.05
•
0.012
•
0.023
•
0.9
1.8
1.25
2.5
tower
venturi
tower
Iventuri
i tower
venturi
tower
Iventuri
tower
iventuri
tower
j venturi
tower
venturi
tower
'venturi
tower
[Venturi
tower
tower
tower
tower
i
Gas flow approximately 0.1 sm3/s (220 ACFM)
Liquid flow approximately 0.38 £ /s (6 gpm) i
Pressure differential across contactor: Venturi: 34 nmHg (18" H20)
Tray tower: 23 mmHg (12" H2O)
-------
TABLE 2. 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
NH4OH
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 i
i
52 ;
43 i
41
49
36
41 ;
29
29
9
N/A'
N/A|
N/A
N/A
Run
No.
10
16
12
18
8
14
4
6
2
9
15
11
17
7
13
*
5
I
1 »
21
20
22
* Selectivity - A measure of the preferential removal of H2S over CO2 talcing
into account the relative difference in concentration between the two gases.
In this paper, selectivity is the ratio of percent removal of H_s to percent
removal of CO^* i
t Data in brackets are suspected to be erroneous. ;
N/A - Not available. Selectivity values for these runs were not available
because the spent scrubbing solution was not analyzed. ;
-------
100
TRAY TOWER
VENTURI
N= NaOH
K= KOH
= NH4OH
x
g OPEN SYMBOLS: VENTURI
Q. SOLID SYMBOLS: TRAYi TOWER
Q.
SELECTIVITY
40 60
% REMOVAL H2S
% REMOVAL CO2
Figure 2. Removal Efficiency vs. Selectivity for Alkaline Scrubber
-------
oc
UJ
o
O co
X Z
DO
oc
s
O CO
* Z
in
q
d
q
d
eo o
9 Ss
o -^
CO
1®
o
CM
q
d
g
cr>
o
o
v
o
o
o
o>
o
00
o
N.
o
(D
O
in
•H
u
3
4J
c
Q)
•0
o
u
c
o
•H
O
•H
14J
U-J
(0
O
1
0)
1-1
IT>
•H
&-,
!N30d3d 'AON3IOIdd3 IVAOINSd '
-------
i 1—i—r
o
t
CO
-6
D
O)
I
X
x _ o
S S 5
2 ^ Z
OD<
O
O
O
O
o
oo
o
f-.
o
tt>
o
in
o
O
TO
in
•
CM
O
•
CVJ
0)
CO
£
in O
^ e
O)
i—K
I
±
C>
in
•
o
w
(U
in
O
I
O
I
•D
10
o
(U
U
•H
UJ
(0
s
s.
T
-------
Discussion of Results
All three of the alkaline solutions performed
•inilarly. The plot of removal efficiency vs.
selectivity in Figure 2 indicates the specific chemical
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 Claus or
other sulfur recovery processes. All three show high
selectivity at recovery efficiencies high enough that,
with the use of multiple venturi stages, a satisfactory
H2S removal efficiency (>95 percent) could be obtained
for the system. Since the system envisioned for using
these chemicals (Figure 1) involves recycling the
alkali, the relative cost of the individual chemicals
is insignificant. Hhat may be significant are factors
of corrosion, safety and availability as well as the
cost of stripping the gases from the alkaline solution.
The test results indicate that the selectivity for
the venturi is highly sensitive to the [OH~] with a
rate of change, ds/d[OH~], of -1700 liter/gmole in the
[OH~J range of 0.01 to 0.04 gmole/liter. The tower
results! show a rate of change of only -300 liter/gmole
in the sane [OH~] range. This effect is believed due
to the presence of NH3 in the retort gas. The short
residence tine in the venturi (0.003 seconds) results
in a high dependence of selectivity on [OH~] due to the
direct dependence of CO2 enhancement. In other words,
the short residence time Beans that the CO2 has a
limited time to react. However, as the H-,S absorption
is controlled by the gas film, its absorption rate is
independent of [OH~1 at values less than 0.03
gooles/liter.
Figure 5 shows that, at concentrations 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 NH^. The higher gas film diffusion
coefficient in the venturi essentially increases the
availability of CO2 at the scrubbing liquid
interface. Consequently, the liquid-phase chemical
enhancement factor, which is a direct function of ;
(OB~), 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 dependence of the
CO2 absorption on (OH~], Since, the HjS removal is
determined solely by the gas film coefficient (due to
the presence of NH3 in the gas), the sensitivity of H2?
absorption to IOH~] in both tower and venturi is .
decreased. ,
i ' ;
These results indicate a clfcar choice of alterna-
tives in deciding between a tower or venturi scrubber
based on process requirements. j If selectivities
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 remova'l efficiency per
stage. If a selectivity less than 50 is acceptable,
the tower is more effective in that combined removal
efficiency and selectivity is greater than with the '
venturi. i
A theoretical computer model was developed by Aiken
(1985) using the penetration theory to correlate the
venturi test results obtained in this program at
TABLE 3. COMPARISON OF THEORETICAL AND EXPERIMENTAL SELECTIVITIES;
Selectivity (Unitless)* '
IOH-)
gaoles/liter
• 0.045
0.023
0.012
24 cat
25
43
66
Theoretical
20.5 cut
31
53
82
Average
28
48
74
Experi-
mental
21
55
75
Avg.
Difference,
percent
25
14.6
1«4
13.7
•Selectivity is the percent H^ removal/percent CO2 removal
tDistance along venturi where liquid is injected
-------
1-1
U
0)
0!
W
in
OJ
•H
U/SSHU= A1IAI10313S
-------
%.i.ca witli tdie -tes-t. conditions And provide a
basis for detailed design of future alkaline
scrubbers. One of the sensitive parameters in the
analysis was the point of injection of the scrubber
fluid into the gas stream. The actual injection point
is 20.5 cm from the beginning .of the venturi
contactor. However, the injected liquid requires a few
• centimeters to atomize. Therefore, computer runs were
•ade for distances of 20.5 and 24 cm. The computed
•electivities are shown in Table 3 compared to the test
results, test agreement is between the test results
and the numerical average of the computed results at
20.5 and 24 cm.
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 at the
higher concentration (0.045 gmoles/liter) there is a 25
percent deviation. This information is also shown in
Figure 5 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
evaluating a venturi scrubber) indicates .that the model
can be used for predictive studies of multi-stage
performance.
Some other interesting results derived from the
modeling study, which included the chemical reaction of
C02, H2S, and NH3 in the retort offgas with the
alkaline solution in a venturi contactor are:
1. The KH3 in the retort offgas reacts with the
H2S at the gas/liquid interface. Since the
HHj present in the offgas is at
approximately the same concentration as the
H2s , the removal efficiency is only
marginally dependent upon -*he IOH~].
2. The removal efficiencies for HH3 and H2S are
similar.
3. The HjS selectivity is significantly
affected by contact time with a maximum
•electivity of 110 occurring at
approximately 0.0015 sec contact time.
4. Variations in temperature and liquid droplet
size can have a significant effect on
•electivity.
Scrubber Concept Designs
These experimental and theoretical results suggest
two alternative alkaline scrubber design concepts for
futuro consideration. One system combines the high
•electivity of''the venturi with the high removal
efficiency of the tower. The other uses a tray tower
for miiximum H2S removal and isolated liquid scrubbing
•Cages to maximize the selectivity for use with a
•ulfur recovery plant. |
For the first concept, the 'first stage is a venturi
designed for peak selectivity based on contact time and
[OR~]. The theoretical model indicates that a maximum
•electivity of 110 will result in a 50 percent removal;
efficiency. The CO2 removal efficiency is 0.40
percent.
The second stage is a tray tower designed! for 90
percent H2S removal efficiency Iwith a stage aielectivity
of 40 using an [OH~] of 0.045 gmoles/liter. These
\
values are based on the experimental results obtained
in this program as shown in Figure 2. The CC>2 removal
efficiency is 2.2 percent. ;
The overall performance for this design is a 95
percent H^ removal efficiency ;with a selectivity of
37. These values are calculated as follows:
"2s
CO2 cone.
ppm
*
Inlet
(Assumed)
1500
22
; Stage 1
Venturi Exit
! 750
! 21.9
Stage 2
flower Exit
75
21.4
H2S removal eff. - L^
22-21.4
CO2 removal eff
Selectivity
22
95*
2.7%
#-»
The second concept is essentially two towers in
series; but, in fact, the two tower stages could be
combined into one partitioned vessel. Each stage has
90 percent B2s removal efficiency and a selectivity of
40. The overall performance of the scrubber is 99 :
percent removal efficiency with a selectivty of 22. ;
These values are calculated as,follows:
!
Inlet i
(Assumed) Stage 1 Exit Scrubber Exit
H2S cone.
°2
ppm
1500 ,
H2s removal eff. -
CO2 removal eff. » ,
Selectivity - i
150
21.5
1500-15
1500
~22
15
21.0
99%
4.5%
CONCLUSIONS
The findings reported above lead to the following;
conclusions: | ;
-------
t. For shale ell retort offgas similar In
composition to that fro* the Geokinetics
process, tho alkaline scrubber, in
combination with • stripper and • Cl»u»
plant, could be a technically viable Beans
of Bf removal. A. design/cost study for •
such a system would determine its economic
viability.
2. Per eeokinetics-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 H2S removal efficiency
of at least 90 percent with a selectivity of
approximately 30. Under the sane conditions
a single venturi contactor in place of the
tray tower would remove only 50 to 60
percent HjS but with a selectivity of 70 to
80.
3. Bused on the computer model developed to
analyze these test results, the removal
•ffficiencies and selectivity above are
applicable to offgas with lower H2S
concentrations than found at Geokinetics.
This suggests a concept of multiple
scrubbing stages to increase the H2s
nimoval. ••cause this increased removal
efficiency is accompanied by a reduced
selectivity (which could present a problem
by lowering the H2S concentration in the
f*ed gas to the sulfur recovery plant), the
cost effectiveness of this concept requires
• design study.
4. Based on a theoretical analysis of the
three-gas component system (B2S, HH3, and
CO2), the principal reactant for the H2S in
the retort offgas is the KH3 in that same
offgas. In that HH3 is present in the
Geokinetiea offgas in similar mole
qinantities to that of the HjS , the scrubber
performance observed on these tests may not
tie applicable to. retort offgas with little
<>r no HH3« this suggests that the water and
1die HH3 in the'off gas would be an effective
Dcmbbing ageat without any alkali additive
to the watery Scrubbing in this manner
irould certainly improve the selectivity, but
1th* effect oa removal efficiency obtainable
:ls uncertain.
offgas itself that is reacting the H2s .
Since the HH3 and H^ concentration* are
variable, it is likely that some of the H2s
is reacted by the alkali; Bierefore, !lt is
likely some alkali will always be needed.
However, the choice of scrubbing alkali may
be made on such factors as co»t,maintenance,
safety, availability, and crew comfort.
6. The absorption of HjS and CO2 in the
alkaline solution occurs;in reactions which
should be fully reversible by
distillation. The sulfur in the scrubber
solution is primarily in the fora of
sulfide. The sulfate or sulfite level
determined in the scrubbing solution was
equal to that in the water supply. The
•ulfide will distill off as H2s (along with
CO2) while the sulfate and sulfite will not.
RETEREHCgS
Aiken, R. C., "Selective Scrubbing of H2s
from C02 in Shale Oil Retort Offgas Bused on
the Penetration Theory,"! Appendix B to EPA
Report cited below, Taback, et al., 1985.
!
Oesai, B. O., Day, D. R., and Peters, 3. A.,
"Air Pollution Investigations of Oil lihale
Retorting: In-Situ and Surface, Task li
Evaluation of Sulfur Removal Technologies,"
(Draft Report) EPA Contract Ho. 68-03-2784,
February 1983.
fcovell, R. 3., Dylewaki, S. ¥., and
fetarson, C. &., •Control of Sulfur
Emissions from Oil Shale Retorts," EPA
Report 600/7-82-016, HTIS PB82-231945, April
19S2.
Taback, B. J., puartucy, C. C., and
Coldstick, R. 3., "Alkaline and Stratford
Scrubbing Tests for HjS Removal from In-Situ
Oil Shale Retort Offgas,- KVB Report Ho.
•07430-1982, EPA Contract 68-03-3166, EPA
Air mad Energy Engineering Research
laboratory, Research Triangle Bark, SC,
February 1985. ,
5. 'Die alkaline •ember removal «fficiency and
aeleetivity aeemed to have little dependence
en the alkali used. This is consistent with
the above concept that it is the HH3 in the
------- |