MIDWEST RESEARCH INSTITUTE
I
REPORT
SOURCE TESTING REPORT
EMB PROJECT NO. 72-PC-12
UNION OIL COMPANY
Lemont, Illinois
by
E. P. Shea
Midwest Research Institute
Kansas City, Missouri 64110
EPA Contract No. 68-02-0228
(MRI Project No. 3585-C, Task No. 12)
MIDWEST RESEARCH INSTITUTE 425 VOLKER BOULEVARD, KANSAS CITY, MISSOURI 64110 • AREA 816 561-0202
-------
SOURCE TESTING REPORT
EMB PROJECT NO. 72-PC-12
UNION OIL COMPANY
Lemont, Illinois
by
E. P. Shea
Midwest Research Institute
Kansas City, Missouri 64110
EPA Contract No. 68-02-0228
(MRI Project No. 3585-C, Task No. 12)
MIDWEST RESEARCH INSTITUTE 425 VOLKER BOULEVARD, KANSAS CITY, MISSOURI 64110 ° 816561-0202
-------
PREFACE
The work reported herein was conducted by Midwest Research
Institute (MRI) pursuant to a task order issued by the Environmental
Protection Agency (EPA) under the terms of EPA Contract No. 68-02-0228.
Mr. E. P. Shea served as the Project Chief and directed the MRI field
team consisting of: Messrs., Gary Kelso, gas sampler; Fred Bergman, gas
chromatographic operator; Douglas Weatherman, methylene blue analyst; and
Mike Serrone, iodometric titration analyst. Mr. Winton Kelly, EPA was
Project Officer and Mr. Richard Burr, EPA was the Field Engineer responsible
for collecting process data. Mr. E. P. Shea prepared this final report.
Approved for:
MIDWEST RESEARCH INSTITUTE
Paul C. Constant, Jr.
Program Manager
26 January 1973
ii
-------
I. TABLE OF CONTENTS
Page
II. Introduction 2
III. Summary of Results 8
A. Observations 8
B. Suggested Method 15
IV. Complete Gaseous Results 21
A. Gas Chromatograph 22
B. lodometric Titration Procedure 36
C. Spectrophotometric Method 41
V. Sampling and Analytical Procedures 46
VI. Process Operating Conditions 56
Appendix A - Determination of H2S in Stack Gases 57
Appendix B - Determination of l^S in Refinery Fuel Gases 62
Appendix C - Field Data Sheet 69
Appendix D - Laboratory Report 81
Appendix E - Field Log 98
Appendix F - Project Participants and Titles 100
-------
II. INTRODUCTION
Under the Clean Air Act of 1970, as amended, the Environmental
Protection Agency is charged with the establishment of performance standards
for stationary source categories which may contribute significantly to air
pollution. A performance standard is a standard for emissions of air pollu-
tants which reflects emission limitations attainable through the best emis-
sion reduction systems that have been adequately demonstrated (taking into
account economic considerations).
The development of realistic performance standards requires
accurate data on pollution emissions within the various source categories.
Sampling and analytical techniques have to be developed to acquire the data.
A method for sampling and analysis of H2S is needed for the fuel gas systems
in the petroleum refining industry. An analytical system for analysis of
sulfur compounds using a gas chromatograph equipped with a 36-foot by 0.085
inch polyphenyl ether and orthophosphoric acid Teflon column and a flame
photometric detector has been used by EPA.— The objective of this project
was to compare the chromatographic technique with a methylene blue-spectro-
photometric method and an iodometric titration method at the Union Oil
Company Refinery in Lemont, Illinois. This report presents the results of
the tests run at the refinery for EPA to determine: (1) the applicability
_!/ J. D. Mulik, R. K. Stevens and R. Baumgardner, "An Analysis System
Designed to Measure Multiple Maladrous Compounds Related to Kraft
Mill Activities," TAPPI WATER and Air Conference, 4-7 April 1971.
-------
of the methylene blue-spectrophotometric method for analysis of t^S in the
ranges of concentration present in the effluent from the amine scrubber,
(2) the applicability of the iodometric titration method, (3) the compari-
son of the results from the wet methods with the results from the gas
chromatograph and (4) changes in either or both procedures to make them
applicable to analysis of H2S in the concentrations present in the refinery
gas stream.
Appendix A presents the EPA draft method for sampling and analysis
of H2S by the methylene blue-spectrophotometric technique. Appendix B
presents the EPA draft method for sampling and analysis of H£S using the
iodometric tiration. These methods were tried, but had to be modified in
the field before any results could be obtained. The mpdifications to the
methods are described in .this report.
On Sundary, 25 June 1972, the equipment was shipped by truck to
Illinois. Messrs. Kelso and Serrone drove the truck. Messrs. Kelly and Burr
arrived at the site Sunday afternoon. On Monday, 26 June 1972, Messrs.
Shea, Bergman and Weatherman arrived at Lemont, Illinois, and the crew then
drove to the refinery arriving there about 10:15 a.m. A meeting was
scheduled with the refinery personnel for 10:15, but had to be delayed
until 11:30, because of a truck fire across the street from the refinery.
The process engineers who were members of Union Oil's fire brigade, were at
the fire.
The equipment was installed Monday afternoon. Figure 1 shows the
refinery fuel gas producing section with the MEA scrubbers and the fuel gas
mix tank.
-------
UNION OIL COMPANY
CHICAGO REFINERY
FUEL GAS H2S REMOVAL
H2S ABSORBERS (MEA) .
FUEL GAS
PRODUCERS
5-50 gr H2S
per 100 scf
Crude
District
Unit
2 MM scfd
10 MM scfd
Sour Gas
22.0 MM scfd
100
psig
Lean
Amine
Control
Valve
&
L«
L»-to
Rich Amine
to
Regenerator
3-30 gr H2S
per 100 scf
2.0 MM scfd
Sour Gas
2.5 MM scfd
100
psig
Lean
Amine
1.5 MM scfd
Rich Amine
To
Regenerator
Fuel
Gas
Mix
Drum
Fuel Gas
-^ to
Firing Line
Natural Gas
0.25grH2S
Sample Point
Figure 1
-------
The sampling point was located after the control valve between
the amine scrubber and the fuel gas mix drum. This control valve regulates
the pressure to the mix drum at 50-60 psig from the 120-130 psig amine
scrubbing unit. This is point A in Figure 1. A Teflon lined pressure
regulator was installed after the 3/4 inch valve in the low pressure line
as shown in Figure 2. A 1/4-inch Teflon tube was attached to the regulator
to serve as the sample line. The tubing was enclosed in 3/8-inch copper
tubing for protection from accidental damage and steam tracing to keep the
temperature of the gas stream above ambient to prevent condensation. The
sampling point for the impinger was 40 feet from the regulating valve. A
Teflon tee was installed in the sample line with one branch for impinger
sampling and the other for GC sampling. The GC was connected to the
impinger sampling tee with 10 feet of Teflon tubing. In the morning, gas
flow was started in the sample line at least 20 minutes before a sample was
withdrawn to ensure that the line was purged. The steam tracing was started
at the same time. A continuous flow was maintained in the sample line until
sampling was finished for the day. After the last sample was withdrawn
the gas and steam valves were closed.
Ten gas samples were taken on Tuesday, 27 June 1972, using the
midget impingers with Cd(OH)2 solution. Two of the samples were ruined
because of excess pressure buildup in the sampling train, and one because
the 30 ml of Cd(OH)2 was exhausted. Of the seven good samples four were
analyzed by the iodometric titration and three by methylene blue-spectro-
photometric method.
-------
UNION OIL COMPANY
CHICAGO REFINERY
FUEL GAS H2S REMOVAL
Open
12 In. Pi
3/4 In. Valve closed
High Pressure
Control Valve Open
3 Ft -4 Ft
Y
3/4 In. Valve
X
Ground Level
Low Pressure
6 Ft
Sample Location A (Cat Cracker - Coker Stream)
Figure 2
-------
On Wednesday there were 14 samples taken in the Cd(OH)2 midget
impinger sampling train. One sample was lost, four were analysed using the
methylene blue-spectrophotometric method and nine by the iodometric titra-
tion. Problems were encountered in the gas chromatograph analysis all day
due to the increasing concentration of H2S saturating the detector.
There were 19 samples taken with the impinger train on 29 June
1972, two samples were lost, four were analyzed with the methylene blue-
spectrophotometric, and 13 were analyzed using the iodometric titration
method. The final changes in the sample handling and analytical methods
were made this day.
In all 24 analyses were made on the refinery gas stream on Friday
30 June, using the chromatographic setup. The standard gas (46 ppm H2S)
was injected into the chromatograph four times to check standardization
before shutting down. Since the standardization checked previous results
very closely, the instrument was shut down and the equipment packed for
return to Kansas City.
There were 24 samples collected in the Cd(OH)2 impinger train on
Friday. Six samples were analyzed using the modified methylene blue-
spectrophotometric method, and the other 18 using the modified iodometric
tritration method. The first sample was taken at 8:53 a.m. and the last
sample was finished at 5:52 p.m. The equipment was packed and the truck
loaded for return to Kansas City. We were finished at 8:00 p.m., and left
the refinery at this time. On Saturday, two men drove the truck to Kansas
City, while the other two returned by air to Kansas City.
-------
III. SUMMARY OF RESULTS
A. Observations
The calibration curve for the methylene blue analysis prepared
in the laboratory before the field work was started was based on the draft
method. Because several changes were made in the methylene blue method
for analysis of H2S in the field, an attempt was made to establish a stan-
dardization curve for the E^S levels obtained in the refinery gas stream.
It was not possible to obtain a standard curve for the methylene blue-
spectrophotometric determination of H2S at concentrations above 25 ppm.
The stream at the refinery contained from 200 to 500 ppm H2S. Therefore,
the results using the methylene blue-spectrophotometric determination for
H2S were valueless for the refinery gas stream that was tested at Lemont.
Comparisons will not be made to the GC data because the concentration of
H2S cannot be calculated for the methylene blue analysis.
Table I presents a comparison of the analytical results from the
flame photometric gas chromatograph analysis and the iodometric titration.
During the 4 days of sampling, there were 34 direct comparison analyses
between the iodometric titration and the GC. Fourteen of these comparisons
were made while the iodometric titration was being modified and before the
final method for sampling and analysis was established. On the 27th there
was one comparison and the results are so far apart that it proved nothing.
The flame photometric detector was saturated and therefore the gas chromato-
graph data is not reliable. The lower readings for H2S on 27 June are
8
-------
TABLE I
COMPARISON OF GC TO IODOMETRIC TITRATION
GC
1040
1050
1100
-
-
-
1126
1208
1335
1356
1417
-
0906
0918
0955
1003
1008
1014
1018
1023
1030
1035
1040
1045
1050
1055
1100
1119
1124
1129
1134
1139
1144
1149
1154
Time
lodometric
Titration
1111-1121
1337-1347
1501-1511
1525-1535
1156-1208
1402-1417
1424-1434
1439-1449
1523-1537
1549-1559
1601-1611
1710-1720
0925-0935
1042-1052
1057-1107
1117-1127
1130-1140
FOR H2S
DETERMINATION
Concentr at ion
Method Titration
EPA
a/
a/
a/
a/
a/
a/
a/
b/
b/
b/
y
Q /
D /
y
•
y
y
y
(ppm)
469
452
439
445
-
325
195
216
260
340
370
350
405
388
284
380
334
285
GC
(ppm)
262
273
261
-
-
-
362
405
423
449
375
-
430
546
498
488
437
470
281
415
454
462
515
491
479
450
485
475
491
498
525
516
499
495
495
Date
6/27/72
6/27/72
6/27/72
6/27/72
6/27/72
6/27/72
6/28/72
6/28/72
6/28/72
6/28/72
6/28/72
6/29/72
-------
TABLE I (Concluded)
Time
GC
1241
1246
1251
1339
1334
1351
1413
1437
1448
1459
1000
1005
1030
1050
1115
1130
1155
1247
1300
1315
1330
1345
1415
1430
1447
1500
1517
1524
1617
1630
1645
1700
1715
1730
1745
1800
a/
b/
c/
d/
lodometric
Titration
1250-1300
1311-1321
1330-1340
958-1003
1027-1030
1050-1059
1119-1126
1252-1302
1322-1332
1412-1422
1427-1437
1521-1529
1613-1624
1630-1640
1643-1653
1658-1706
1715-1725
-
1739-1752
-
First modificatii
instead of Whai
Second modificat;
see Section V.
Final modificatii
See Section IV-A
Concentration
Method
EPA
£/
c/
£/
£/
c/
c/
£/
£/
£/
c/
£/
c/
c/
c/
c/
c/
£/
£/
Titration
(ppm)
544
560
519
398
355
345
312
411
437
442
402
335
394
293
396
360
373
406
GC
(ppm)
550
584
766
600
648
780
460
449
557
489
253d-/
354
321
347
368
378
392
440
472
448
476
400
412
403
478
500
640
378
366
293
335
330
448
372
446
431
Date
6/29/72
6/29/72
6/29/72
6/29/72
6/29/72
6/29/72
6/29/72
6/29/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
used fritted disc glass crucibles for filtering CdS
No. 40, see Section V.
used 10% HCl to wash glassware instead of water,
see New Method, Section III-B.
23.
10
-------
exactly opposite to the findings on the other dates. On the 28th there
were three comparisons. There was such a discrepancy between results that
validity of the iodometric titration procedure was in serious doubt. On
the 29th the first 10 comparisons showed the same large spread between the
iodometric and the GC analyses. Therefore, further changes were made in
the procedures for analyzing by the iodometric method. The changes in the
iodometric titration procedure and the final method used are detailed in
Section IV.
The last two comparison runs on the 29th and all of the comparison
runs on the 30th (with the possible exception of the first chromatograph
analysis) appear to be reliable comparison runs.
Figures 3 and 4 are plots showing the results of analysis in
comparison of the values of the GC and iodometric titrations taken on the
29th and 30th of June. Figure 3 is the comparison of the data for the
29th of June and shows the concentration of H2S in ppm on the ordinate and
the time at which the sample was taken on the abscissa. The dots are results
of the GC analysis and the dashes are the iodometric analysis of the impinger
samples. The dash corresponds to the sample time when the refinery stream
was being bubbled through the impinger train to collect the sulfide. Upon
examination, Figure 3 shows that the concentration of H2S as determined by
the gas chromatograph was constantly changing. The only iodometric titra-
tion samples that are shown on this figure are the ones using the final
modified method. Examination of Figure 4 shows that the variation in
11
-------
800
700
600
E
Q.
Q.
00
x 500
u
400
300
250
0906
• G. C. Sample
_ Impinger Sample
0956
1046
Figure 3
_
1116
1226
1316
1406
1456
TIME
Comparison of GC with lodometric Titration
(29 June 1972 Start 0906)
-------
E
Q.
Q.
CO
CN
X
u
500
450
400
350
300
225
— G. C. Sample
• Impinger Sample
I
_L
I
0850 0940
1030 1120 1210 1300 1350
TIME - MlNUTES
1440
1530
1620
1710 1800
Figure 4 - Comparison of GC with lodometric Titration
(30 June 1972)
-------
concentration during the sampling period was again extreme and the concentra-
tion of H2S in the refinery gas stream varied from about 200 to 500 ppm during
the sampling period. The concentration data obtained by the iodometric
titration procedure do not vary quite as greatly as the GC because the GC
is an instantaneous sample collected over a 1- to 3-second interval while
the impinger sample is collected during a 5- to 15-minute period. This
would tend to average out the fluctuations that occur in this stream and
would tend to yield a flatter curve than the instantaneous sampling of the
GC. During our sampling period not only did the concentration of H2S vary
from time to time in the refinery gas but the pressure to our sample line
varied quite widely. Instead of 50 to 65 psig supply,the pressure varied
from about 40 to 75 pounds at the point where we connected our pressure
regulator and sample delivery line. This wide fluctuation in pressure put
a strain on the pressure reducing valve and required periodic attention and
a change of valve settings for proper delivery of gas to the sampling trains.
The sampling location was downstream of the control valve, thus changing
pressures could indicate changing flow rates of the gas stream.
The variation in results is well within "t 107o in all cases and
there are a number of comparison readings that show essentially no difference
in concentration at all. The conclusions that can be drawn from examination
of the data in Table I and Figures 3 and 4 are: (1) that the iodometric
titration method as modified is a reliable method for on-the-spot analysis
of refinery gas streams for hydrogen sulfide and, (2) the results are well
within an average experimental error of 1 10%, using the gas chromatograph
data as reference.
14
-------
B. Suggested Method
Soon after starting the sampling and analytical work at the re-
finery, it became apparent that the procedures, as written, were not adequate
for the sampling and analysis of I^S at the concentrations present in the
refinery stream. Therefore, several modifications in the analytical and
sampling procedure were made. For the determination of I^S in refinery fuel
gases by the iodometric titration, the recommended procedure based on our
work is:
. . . 1. Recommended Procedure of Determination of H2S in Refinery
Fuel Gases
1. Principal and Applicability
1.1 Principal
H2S is collected in a series of midget impingers and
reacted with alkaline cadmium hydroxide to form cadmium sulfide.
The cadmium sulfide is dissolved in the impingers that it was
collected in by adding an acidified iodine to the midget im-
pingers. A known volume of iodine solution is used with a
known normality. The excess iodine is titrated with thio-
sulfate solution with starch as the indicator.
1.2 Applicability
This method should be applied only when specified
by the test procedure for determining compliance with the
new source performance standards.
15
-------
2. Interferences
3. Apparatus
3.1 Sampling
The apparatus used shall be the same as that used in
Method 4, Federal Register. Volume 36, p. 24887, 23 December
1971, except that two dry impingers should be put into the
sampling train to make certain that no moisture is carried
into the dry gas meter.
3.2 Analysis
Glass stoppered 500 ml iodine no. flask for reaction
vessel.
4. Reagents
4.1 Sampling
4.1.1 Absorbing solution - Mix 4.3 g cadmium sulfate
(3CdS04 8H20) and 0.3 g of NaOH in 1 liter of distilled water.
Mix well before using.
4.2 Analysis
4.2.1 Iodine Solution, 0.01 N
Standardize daily as follows: Pipette 100 ml of the 0.01 N
iodine solution into a 500-ml conical flask. Titrate with the standard
0.01 N thiosulfate solution until the solution is a light yellow. Add
a few drops of the starch solution and continue titrating until the
blue color just disappears.
16
-------
4.2.2 Sodium Thiosulfate Solution.
Sodium thiosulfate solution, standard 0.1 N. For each
liter of solution, dissolve 24.8 g of sodium thiosulfate
(Na2S203.5H20) in distilled water and add 0.01 g of anhydrous
sodium carbonate (Na2 CO^) and 0.4 ml of chloroform (CHClg) to
stabilize. Mix thoroughly by shaking or by aerating with nitro-
gen for approximately 15 minutes, and store in a glass-stoppered
glass bottle.
Standardize frequently as follows: Weigh into a 500-ml
volumetric flask about 2 g of potassium dichromate (K2Cr207)
weighed to the nearest milligram and dilute to the 500-ml mark.
Use dichromate which has been crystallized from distilled water
and oven-dried at 360°F to 390°F. Dissolve approximately 3 g
of potassium iodide (KI) in 50 ml of distilled water in a glass-
stoppered, 500-ml conical flask, then add 5 ml of 20% hydro-
chloric acid solution. Pipette 50 ml of the dichromate solu-
tion into this mixture. Gently swirl the solution once and
allow it to stand in the dark for 5 minutes. Dilute the solu-
tion with 100 ml to 200 ml of distilled water, washing down the
sides of the flask with part of the water. Swirl the solution
slowly and titrate with the thiosulfate solution until the solu-
tion is light yellow. Add 4 ml of starch solution and continue
with a slow titration with the thiosulfate until the bright blue
17
-------
color has disappeared and only the pale green color of the
chromic ion remains.
4.2.3 Sodium Thiosulfate Solution.
Sodium thiosulfate solution, standard 0.01 N. Dilute 100
- 0.01 ml of the standard 0.1 N thiosulfate solution in a volu-
metric flask to 1 liter with distilled water.
4.2.4 Starch Indicator Solution.
Suspend 10 g of soluble starch in 100 ml of distilled
water and add 15 g of potassium hydroxide pellets. Stir until
dissolved, dilute with 900 ml of distilled water, and let stand
1 hour. Neutralize the alkali with concentrated hydrochloric
acid, using an indicator paper similar to "alkacid" test ribbon,
then add 2 ml of glacial acetic acid as a preservative.
Test for decomposition, by titrating 4 ml of starch
solution in 200 ml of distilled water, with the 0.01 N iodine
solution. If more than 4 drops of the 0.01 N iodine solution
are required, make up a fresh starch solution.
5. Procedure
'5.1 Sampling
5.1.1 Assemble the sampling train as shown in Figure 4-1.
Federal Register. Volume 36, p. 24887 connecting two midget
impingers with cadmium hydroxide, two dry midget impingers and
one midget impinger filled with silica gel in series. Place
18
-------
15 ml of the collecting solution in each of the first two
irapingers.
5.1.2 Purge the connecting line between the fuel
gas sampling valve and the first impinger. Connect the sample
line to the train. Read and record the initial reading on the
dry gas meter.
5.1.3 Open the flow control valve and adjust the
sampling rate to read between 1.0 and 1.5 liters/minute. Read
and record the meter temperature.
5.1.4 Continue sampling for 10 minutes. At the end
of this time, close the flow control valve and read the final
meter volume and temperature.
5.1.5 Disconnect the impinger train from the sampling
line and cap the open ends. Remove to the sample analytical
area.
5.2 Sample Recovery and Analysis
5.2.1 Combine the 50 ml of the 0.01 N iodine solu-
tion with 50 ml of 4 N hydrochloric acid in a 250-ml beaker.
Shake and carefully transfer the contents of the impingers
to a 500-ml iodine flask. Rinse the impingers and connecting
glassware with 50 ml of the acidified iodine solution and care-
fully transfer to the iodine flask. Replace stopper after each
additon of solution and shake. Rinse the impingers and connect-
ing glassware with 25 ml of the acidified iodine solution and
19
-------
transfer to the iodine flask. Rinse the impingers and connecting
glassware with the last 25-ml portion of the acidified iodine
and transfer to the iodine flask. Then rinse the impingers,
glassware and 250-ml beaker with distilled water and quantatively
transfer to the iodine flask. Stopper and shake well. Analyze
within 30 minutes.
5.3 Analysis
5.3.1 Titrate the solution in the flask with 0.01
normal sodium thiosulfate solution, until the solution is light
yellow. Add 4 ml of the starch solution and continue titrating
until the blue color just disappears.
5.3.2 Run a blank determination with 30 ml of collect-
ing solution handled according to Section 5.2 to adjust the
value of the iodine consumption.
6. Calculations
Sample Calculations of t^S by lodometric Titration
_ 0.2618 x [ANi - (BN2-C)] g/scf
H0S ' -
VMS
C = ANB - BNB
17.7 x Pm
VMg = "TV x Vm cu ft
a (Tin + 460)
0.2618 = (34 gr/mole H2S) (0.0154 gr/mg) (1,000 mg/g)
1,000 x 2 H2S equil
20
-------
Calculations for Sample No. 24, 30 June 1972.
VMS = 17.7 x 29,92 x 0.273 = Q>264 cu ffc
88 + 460
C = (50 x 0.01)-(47 x 0.01) = 0.03
C ° 0.2618 x [50 x 0.01-(18.8 x 0.01 - 0.03)] =
H2S 0.264
0.254 gscf = 25.4 g/100 cf = 406 ppm
Cu 0 = Concentration of hydrogen sulfide in refinery gas, gr/scf
tlnO
Pm = Meter pressure in. of mercury absolute
Tm = Meter temperature °F
Vm = Meter volume cu ft
A = Volume of standard iodine used ml
Ni = Normality of the standard iodine solution
B = Volume of standard sodium thiosulfate used ml
N2 = Normality of the standard thiosulfate solutions
VMS = Volume of gas sampled at standard conditions
AN = Volume of standard iodine used x the normality standard iodine
"D
used in the blank
BNR = Volume of standard thiosulfate used x the normality of the
standard thiosulfate used in the blank
IV. COMPLETE GASEOUS RESULTS
This section contains the results from the gas chromatograph
analysis, the iodometric titration and the methylene blue-spectrophotometric
analytical methods.
21
-------
A. Gas Chromatograph
The results of the flame photometric gas chromatographic analysis
of the refinery gas stream are contained in Table II. All of the data
recorded are presented in this table.
There are a total of 60 analyses and all but two of them are
reliable readings based on our experience with the gas chromatograph and
the knowledge gained of refinery operations during this program. We have
reported the time, the date, the gas chromatograph reading, the calibration
factor, and the concentration of H2S in ppm and grains/ 100 cubic feet.
The GC was standardized using a 46 ppm standard H2S gas. Because
of the high hydrocarbon content of the sample gas, the flame went out and
had to be relighted before the refinery gas could be analyzed by the
chromatograph. Three samples of refinery gas were analyzed. The chromato-
graph began malfunctioning and it was not possible to use the chromatograph
for analysis. The nature of the trouble pointed to the electrometer.
On June 28th, Mr. Bergman found and repaired a break in the power
cable from the electrometer to the detector head. The chromatograph was
standardized with the 46 ppm H2S and five analyses were made. The sample
loop for the preceding tests was 20 inches of 1/8-inch Teflon tubing. The
200 to 500 ppm concentration of H2S in the refinery stream began to satur-
ate the detector so the sample loop was shortened by cutting 2- inch pieces
out of it starting at 1515. The detector was still being saturated at 1720.
At 1730 the sample loop was shortened to 8 inches and standard l^S gas was
22
-------
TABLE II
CONCENTRATION OF H2S BY FLAME PHOTOMETRIC
Gr/100 cf
16.4
17.1
16.3
6/28/72 1126 452 0.80 362 22.6
25.4
26.4
28.1
23.4
6/29/72 0906 506 0.85 430 26.9
33.2
31.2
30.5
27.3
29.4
17.6
26.0
28.4
28.8
32.2
30.7
29.9
28.2
30.3
29.5
30.7
31.2
32.8
32.2
31.2
31.0
31.0
34.4
36.5
47.9
37.5
40.5
48.7
29.7
28.0
34.8
30.6
GAS CHROMATOGRAPHY
Time
1040
1050
1100
1126
1208
1335
1356
1417
0906
0918
0955
1003
1008
1014
1018
1023
1030
1035
1040
1045
1050
1055
1100
1119
1124
1129
1134
1139
1144
1149
1154
1241
1246
1251
1258
1339
1344
1351 ^
1413
1437
1448
1457
Reading
420
438
419
452
506
528
560
468
506
643
585
575
513
551
330
488
533
544
605
578
562
527
568
558
578
586
618
608
588
583
583
649
686
900
Saturated
705
763
918
540
528
655
575
23
Factor
0.625
0.625
0.625
0.80
0.80
0.80
0.80
0.80
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
Detector
0.85
0.85
0.85
0.85
0.85
0.85
0.85
ppm
262
273
261
362
405
423
449
375
430
546
498
488
437
470
281
415
454
462
515
491
479
450
485
475
491
498
525
516
499
495
495
550
584
766*
-
600
648
780
460
449
557
489
-------
TABLE II (Concluded)
Time Reading Factor ppm Gr/100 cf
1000
1005
1030
1050
1115
1130
1142
1155
1247
1300
1315
1330
1345
1415
1430
1447
1500
1517
1524
1617
1630
1645
1700
1715
1730
1745
1800
177
250
227
245
260
280
290
326
618
700
664
727
591
609
595
708
740
730
432
418
335
383
377
511
425
510
492
1.43
1.43
1.43
1.43
1.43
1.43
1.43
1.43
0.675
0.675
0.675
0.675
0.675
0.675
0.675
0.675
0.675
0.675
0.875
0.875
0.875
0.875
0.875
0.875
0.875
0.875
0.875
253*
358
324
350
372
400
415
465
440
472
448
476
400
412
403
478
500
492
378
366
293
335
330
448
372
446
431
15.8
22.4
20.2
21.9
23.3
25.0
25.9
29.0
27.5
29.5
28.0
29.8
25.0
25.8
25.2
29.8
31.2
30.8
23.6
22.9
18.3
20.9
20.6
28.0
23.2
27.9
26.9
These results are questionable.
24
-------
passed through the chromatograph several times to help clean out the detec-
tor. The chromatograph was shut down and all sampling discontinued for
the day. There were seven samples passed through the GC from 1500 to 1720,
but since the detector was saturated on all samples no attempt was made to
calculate the concentration of H2S.
Thursday morning, 29 June 1972, the sample loop on the chromato-
graph was shortened to a volume of 0.1 ml before analysis began. Standard
HoS (46 ppm) was passed through the chromatograph for standardization.
Sampling began at 9:06 a.m. Three analyses were made to scan for components
other than I^S, CH.,SH (methyl mercaptan) was the only other component de-
tected in more than trace quantities. At 9:55 it was decided to inject gas
from the sample line into the instrument every 5 minutes, 13 analyses were
made in this manner. At 11:00 a.m. a 19-minute run was made on one injec-
tion. The only components found were H2S and CH3SH. Nineteen more analyses
were made with 5 minutes between injections. At 3:00 p.m. the chromatograph
again started malfunctioning. The indications were that there was no H2S
in the effluent. The standard (46 ppm) H2S was injected into the chromato-
graph. The trace showed a blip instead of a peak. Several more injections
were made at 10-second intervals with the standard gas, and finally, a peak
about one-fourth of the proper size, did show up. It was theorized that the
column had become saturated with the amine (MEA) from the scrubber system.
A 1-foot section of the column was cut off and the column reversed. Then
several more injections of standard gas were made at 10-second intervals.
A peak about one-half of the normal size was obtained.
25
-------
At 6:00 p.m., it was decided to bake the column over night.
The temperature was raised from 60°C to 110°C, the nitrogen carrier gas
was left on and the rest of the instrumentation turned off.
Friday morning the temperature of the column was reduced to 60°F,
the instrumentation turned on and the flame lit. After the chromatograph
stabilized, standard (46 ppm) HoS was injected into the column several times
for standardization. Analysis of the refinery stream started at 10:50 a.m.
and there were five analyses before lunch. Between 12:47 p.m. and 4:00 p.m.,
11 analyses were made. At 4:00 p.m., four injections, at 4-minute intervals,
were made with the standard (46 ppm) I^S to check standardization. Then
eight more injections were made and analyses recorded.
Table III shows the sample calculations for determining the
hydrogen sulfide concentration from the GC analysis.
The following week after the field test, the chromatograph was
reassembled in Kansas City and standard gases run through it to obtain
calibration for the various sulfur compounds. Forty-six parts per million
hydrogen sulfide gas was injected and the response was the same as obtained
during our work at Lemont, Illinois, the previous week. In addition, a
mixture of the gases, carbonyl sulfide, methyl mercaptan, and sulfur dioxide
was injected to obtain calibrations to determine if any of them were present
in the refinery gas tested at Lemont. Separate calibrations were also ob-
tained for carbon disulfide and ethyl mercaptan. Then a mixture of nitro-
gen, COo and S07 was injected to determine if C02 and S0? were present in
26
-------
TABLE III
SAMPLE CALCULATIONS FOR HYDROGEN SULFIDE CONCENTRATION
1. Hydrogen Sulfide Concentration from Gas Chromatograph Analysis
R x 64 x 46 ppm
Rs x 4
Calculation for 1,800 hr on 30 June 1972
C = 492 x 64 x 46 = 4n
H2S 840 x 4 PPm
CH s = Concentration of H2S in the refinery gas, ppm.
R = The readings on the recorder chart from the chromatograph when
the refinery gas is analyzed.
Rs = The reading on the recording chart when standard I^S is analyzed
the refinery gas. After examining the calibration curves and the chromato-
graph recordings from the work at Lemont, Illinois, it was determined that
the only materials found by chromatography were H2S, methyl mercaptan and
a trace of carbon dioxide. Table IV shows the relative retention (a) to
the internal standard of methane for the components which were expected to
be in the refinery gas that was analyzed. Table V shows the components that
were resolved when the large 20-inch sampling loop was installed in the
chromatograph. These data were taken on Tuesday, June 27th, and show that
there was a trace of methane, ethane, a concentration of H2S of about 273
ppm, five unknown hydrocarbons with the relative retention of 2.06, 2.41,
2.64, 2.92, and 3.58. Methyl mercaptan and carbon disulfide at concentra-
tions of about 9.4 ppm and 0.9 ppm, respectively, were also found.
27
-------
TABLE IV
RELATIVE RETENTION n, TO INTERNAL STANDARD OF METHANE
Component
CH4
C2H6
C02
C3Hg
H2S
COS
S02
CH3SH
CS2
C2H5SH
Retention
Time-/
14.8
16
» 17
18
23
25
45.5
64
120
167
Relative
Retention
.0
,08
.15
,22
,55
,69
,07
4.32
8.11
11.28
a/ Measured in millimeters on recorder chart L&N No. 492000, with recorder
chart speed of 1 inch per minute.
TABLE V
COMPONENTS RESOLVED USING LARGE
SAMPLING LOOP ON 27 JUNE 1972
(Samples Taken at 10:40 and 12:08")
a
1.0
08
55
06
2.41
2.64
2.92
3.58
4.32
8.11
Compound
CH4
C2H6
H2S
Unknown Hydrocarbon
Unknown Hydrocarbon
Unknown Hydrocarbon
Unknown Hydrocarbon
Unknown Hydrocarbon
CH3SH
2
Approximate Concentration (ppm)
273
9.4
0.09
Note: See Figure 5 page 32 for peaks attributed to these compounds.
28
-------
Table VI contains the report on the concentration of methyl mercaptan found
in the refinery stream on June 28th, 29th and 30th. The data are reported
in parts per million with the time and date at which it was found.
TABLE VI
CONCENTRATION OF METHYL MERCAPTAN IN REFINERY STREAM
Time
11:26
9:06
9:18
9:55
10:03
11:54
10:00
10:05
10:50
11:15
11:30
11:42
11:55
12:47
13:00
13:15
13:30
13:45
14:15
14:30
14:47
15:00
15:17
15:24
16:17
16:30
16:45
17:00
17:15
17:30
17:45
18:00
(28-30 June 1972)
Date
6/28/72
6/29/72
6/29/72
6/29/72
6/29/72
6/29/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
6/30/72
Methyl Mercaptan
(ppm)
7.5
6.0
13.3
17.5
10.2
13.6
3.4
6.8
2.6
3.4
8.5
8.3
10.9
11.1
11.2
12.3
11.7
11.2
10.7
12.8
12.5
13.3
14.2
15.6
9.9
9.6
11.1
12.8
15.4
14.3
16.7
12.8
29
-------
Figures 5 through 9 are reproductions of the recorder chart from
the gas chromatographic analysis for H2S and other components in refinery
gas. Figure 5 shows the first determinations on the 27th of June and also
shows the line voltage fluctuation that gave us trouble that morning and
blew several fuses in the chromatograph as well as the spectrophotometer.
Examination of the figure shows the components that can be identified in
the refinery gas stream. These are hydrogen sulfide, methane, ethane,
methyl mercaptan, five unknown hydrocarbons and carbon disulfide.
Figure 6 is the recorder trace for the chromatograph and shows
the standardization of the instrumentation as well as a trace for E^S and
methyl mercaptan. Figure 7 is a reproduction of the recorder trace for the
afternoon of the 28th. This was the day that the detector was becoming
saturated because of the high concentration of hydrogen sulfide in the
refinery gas stream. During this period the sample loop was being shortened
so that an analysis could be made using the gas chromatograph. The first
trace shows a rounded point rather than a sharp, well-defined point
indicating that the photo detector was still being saturated. Both hydrogen
sulfide and methyl mercaptan can be seen in the trace on this day. Figure
8 shows a trace from the 29th of June and shows the clean spectrum that
should be obtained using the detector. Also, it shows the period when a
sample was injected approximately every 5 minutes. The only two components
being resolved were t^S and CH3SH, with the possible exception of a slight
trace of CC^. Carbon dioxide would give a negative peak in respect to
and methyl mercaptan. Figure 9 is a trace of the 30th of June and shows
the type of recording and detection experienced at the time of shutting
down at 1800 hours on 30 June.
30
-------
Figure 5 - GC Trace
31
-------
70 0
71
It**
Figure 6 - GC Trace
32
-------
Figure 7 - GC Trace
33
-------
$00
?*
/CO
Figure 8 - GC Trace
34
-------
Figure 9 - GC Trace
•" A
35
-------
Figure 10 is a calibration chart for hydrogen sulfide. A standard
gas (46 ppm hydrogen sulfide) was introduced into the gas chromatograph four
different times and the readings obtained were used to determine the multi-
plying factor for converting from scale reading to concentration of hydrogen
sulfide in the refinery gas. Figure 11 is a copy of the calibration curve
for methyl mercaptan, carbon dioxide, carbonyl sulfide, and sulfur dioxide.
One thing that becomes apparent in the examination of this calibration curve,
is that carbonyl sulfide and hydrogen sulfide are eluted from the column at
nearly the same time and therefore, it would be difficult to separate and
quantify carbonyl sulfide and hydrogen sulfide with the 36-foot Teflon
column used at Lemont, Illinois, Union Oil Refinery.* It was from calibra-r
tion curves of this type that the value of methyl mercaptan and the other
constituents were determined so that the readings on the chart could be
converted into ppm.
B. lodometric Titration Procedure
Table VII contains the results of the iodometric titration for
determination of hydrogen sulfide in the refinery gas stream from the amine
scrubber.
* The carbonyl sulfide, if present, would show a double peak with hydrogen
sulfide or a peak on the back side of the hydrogen sulfide peak.
Examination of the chromatograph trace on June 27th and 28th (Figures
5 and 6), when the large sample loop was used, does not show any
indication of the carbonyl sulfide peak. Therefore, the concentra-
tion of carbonyl sulfide is less than 0.5% of the concentration of
hydrogen sulfide in this refinery gas stream.
36
-------
Figure 10 - Standardization Using 46 ppm H0S
37
-------
cos
Figure 11 - Standardization Using Gas Mixture
38
-------
TABLE VII
CONCENTRATION OF H2S DETERMINED BY
IODOMETRIC TITRATION
Sample No.
and Time
2
4
6
7
1
3
4
5
8
9
10
13
1
4
5
6
7
8
9
10
14
15
16
17
18
1
2
3
4
5
6
7
9
11
13
14
17
19
20
21
22
23
24
1042-1052
1337-1347
1501-1511
1525-1535
1156-1208
1402-1417
1424-1434
1439-1449
1523-1537
1549-1559
1601-1611
1710-1720
0925-0935
1042-1052
1057-1107
1117-1127
1130-1140
1250-1300
1311-1321
1330-1340
1523-1533
1539-1550
1554-1607
1612-1626
1632-1642
0853-0908
0929-0939
0958-1003
1027-1030
1034-1044
1050-1059
1119-1126
1252-1302
1322-1332
1412-1422
1427-1437
1521-1529
1613-1624
1630-1640
1643-1653
1658-1706
1715-1725
1739-1752
VM
(cu ft)
0.318
0.323
0.323
0.318
0.323
0.517
0.565
0.393
0.301
0.325
0.369
0.322
0.335
0.358
0.333
0.394
0.413
0.335
0.307
0.388
0.381
0.300
0.303
0.302
0.521
0.304
0.364
0.262
0.251
0.578
0.349
0.368
0.423
0.316
0.309
0.298
0.354
0.034
0.365
0.359
0.377
0.347
0.273
TM
(°F)
80
84.5
84.5
84.5
74
72
71.5
71.5
72
73
72.5
71
68.5
70
71
71
70
74.5
75
74.5
74.5
75.5
75.5
76
75.5
75
79
84.5
91
88
87
85.5
87
87
87.5
89
88
88
88
88
88
88
88 ,c
VMS
(cu ft)
0.312
0.315
0.315
0.310
0.322
0.516
0.563
0.391
0.299
0.323
0.367
0.321
0.336
0.358
0.332
0.393
0.413
0.333
0.304
0.386
0.379
0.296
0.299
0.298
0.515
0.300
0.358
0.255
0.241
0.559
0.338
0.357
0.410
0.306
0.299
0.288
0.342
0.294
0.353
0.338
0.364
0.336
j 0.264
Gr/100 cf
29.3
28.2
27.4
27.8
20.3
12.2
13.5
16.2
21.2
23.1
21.8
25.3
24.2
17.2
23.7
20.9
17.8
34.0
35.0
32.4
24.1
23.8
20.3
19.4
17.0
18.4
21.4
24.9
22.2
15.4
21.6
19.5
25.7
27.3
27.6
25.1
20.9
24.6
18.3
24.8
22.5
23.3
25.4
ppm
469
452
439
445
325
195
216
260
340
370
350
405
388
284
380
334
285
544
560
519
386
380
324
310
272
294
342
398
355
246
345
312
411
437
442
402
335
394
293
398
360
373
406
Date
6/27/72
6/28/72
6/29/72
6/30/72
-------
The data are presented as sample number, time of sampling, date sampled,
the meter volume, meter temperature of the dry gas meter in the sampling
train, the calculated volume of gas samples at standard conditions, and
the concentration of H2S in grains/100 cubic feet and ppm. In all there
were 43 iodometric titrations. The titrations on Tuesday the 27th and
Wednesday the 28th are of very little value as it is certain that the method
of sampling, handling and analysis was allowing the escape of hydrogen sul-
fide when the cadmium sulfide was dissolved in hydrochloric acid. Samples
1 through 7 on Thursday the 29th were also analyzed before the final ana-
lytical and sampling procedures were established and these values are of
questionable value. Samples Nos. 8 through 18 on Thursday the 29th and all
of the samples on Friday the 30th were collected and analyzed by the modified
procedure which is reported in Section III-B of this report.
Concentration of Hydrogen Sulfide by Iodometric Titration
= 0.2618 x [ANj_ - (BN2-C)] g/scf
V" v^
C = ANB - BNB
17.7 x Pm „ -
VMq = ; x Vm cu ft
S (Tm + 460)
0.2618 = (34 gr/mole H2S> (0-0154 gr/mg) (1,000 mg/g)
1,000 x 2 H2S equil
Calculations for Sample No. 24, 30 June 1972
40
-------
VM 17.7 x 29.92 x 0.273 . „,.
VMS - x 0.264 cu ft
88 + 460
C = (50 x 0.01)-(47 x 0.01) = 0.03
_ 0.2618 x [50 x Oo01-(18.8 x 0.01 - 0.03)]
C
V 0.264
H0S
0.254 gscf = 25.4 g/100 cf = 406 ppm
CH Q = Concentration of hydrogen sulfide in refinery gas, gr/scf
tlrt O
Pm = Meter pressure in. of mercury absolute
Tm = Meter temperature °F
Vm = Meter volume cu ft
A = Volume of standard iodine used ml
N, = Normality of the standard iodine solution
B = Volume of standard sodium thiosulfate used ml
N2 = Normality of the standard thiosulfate solutions
VMg = Volume of gas sampled at standard conditions
ANg = Volume of standard iodine used x the normality standard iodine
used in the blank
BN = Volume of standard thiosulfate used x the normality of the
standard thiosulfate used in the blank.
C. Spectrophotometric Method
The modified method for methylene blue-spectrophotometric deter-
mination of hydrogen sulfide in the refinery gas stream was used in trying
to draw a standard curve so that the absorbance data obtained in the sampling
at the Union Oil Refinery at Lemont, Illinois, could be related to the con-
centration of H2S in ppm. An attempt was made to prepare a calibration
41
-------
curve for 5, 10, 25, and 50 ppm H2S. Figure 12 shows that at the values
below 10 ppm, the curve has the right slope, but is straightened out above
10 ppm and went backwards at all values above 25 ppm. It was not possible
to get an absorbance reading above 35 ppm H^S.
Because several changes were made in the methylene blue-spectro-
photometric method for analysis for H2S in the field, a new standardization
curve for the methylene blue method was needed.
The curve, Figure 13, shows the results of making up a 200 ppm,
350 ppm, and a 500 ppm standard using analyzed reagent grade cadmium sul-
fide. The curve has a negative slope, or in other words, higher parts per
million H2S gave a lower absorbance reading. This same type of response
was found in the calibration work for a standard curve before going to the
refinery. It was then decided to see if mixing the hydrochloric acid and
the ferric chloride before dissolving the cadmium sulfide would allow prep-
aration of a calibration curve in the range of 200 to 500 ppm hydrogen sulfide.
As shown in Figure 14, the curve does have a less steep slope; however, it
is still negative and is not useful for determining the concentration of
H2S in the refinery stream at the concentrations found in the Union Oil
Refinery at Lemont. While dissolving the cadmium sulfide in the acid
solution, the smell of H2S escaping from the flask was noticeable. The
higher the concentration the more t^S escaped while dissolving the cadmium
sulfide. The ability to analyze the refinery stream for l^S using the
methylene blue method is entirely dependent on the solubility of H2S in
water. In all cases, where concentrations above 100 ppm H2S existed, H2S
did evolve from the solution.
42
-------
50
40
10
•
j
0 0.1 0.2
Absorbance
Figure 12- Calibration Curve for t^S Using Methylene Blue-Spec 20
43
-------
500
400
300
a.
a
10
CN
U
200
100
0.1
SPECTROPHOTOMETER - 20
B&L. FILTER 670 A
_L
0.2 0.3
ABSORBANCE
0.4
0.5
Figure 13 - Calibration Curve for lUS Using Methylene Blue-Spec 20
44
-------
500 _
400
300
I
CN
X.
(J
200
100
Jl
SPECTROPHOTOMETER - 20
B& L. FILTER 670 A
I
I
0.1 0.2 0.3
ABSORBANCE
0.4
Figure 14- Calibration Curve for t^S Using Methylene Blue-Spec 20
45
-------
Based upon the work at the refinery and at MRI in trying to
establish calibration curves for the methylene blue-spectrophotometric
method for analyses of H2S, the conclusion is that the method is not appli-
cable for concentrations above 100 ppm of H2S and is of questionable value
above 50 ppm H2S.
V. SAMPLING AND ANALYTICAL PROCEDURES
Figure 15 is a schematic of the equipment used to obtain the gas
samples. A pressure reducing valve with all Teflon contact parts was in-
stalled in the outlet of a bleed valve on the low pressure line from the
amines scrubbers. A 1/8-inch Teflon tube about 40 feet long was attached
to the outlet of the pressure reducing valve. The Teflon tubing was jacketed
with a 3/8-inch copper tubing to provide steam tracing and also prevent
damage to the line. A Teflon tee was put into the sample line 40 feet from
the pressure regulator. One branch of the tee went to the gas chromatograph
sample valve. The other end of the tee was connected to the sampling train.
The sampling train consisted of five impingers in series, 15 ml of cadmium
hydroxide solution in the first two, two dry impingers and a silica gel
impinger. The impingers were submerged in an ice water bath to be certain
that all of the moisture in the line was condensed in the dry impingers and
the silica gel impinger before the gas was discharged to the dry gas meter.
46
-------
Refinery
Gas line
From Scrubber
Pressure
Reducing
Valve
l/8"Teflon Tubing
Copper Jacket
Pressure Gauge
Therometer -_ ^_^^
"99
2 - Cd (OH)2 Impinger
Dry Impinger
Silica Gel
( r^r^TMTKT^
Drygas
Meter
To G.C. Sample Valve
Ice Water Bath
Impinger
Train
Figure 15 - Schematic of Sampling Equipment
-------
The EPA suggested sampling procedure is contained in Appendix A
for methylene blue and Appendix B for iodometric titration. These sampling
procedures were modified by the addition of three impingers, two dry and one
filled with silica gel following the two impingers containing 15 ml each of
cadmium hydroxide solution.
When the absorbing solution is made up, the cadmium hydroxide
solubility is so low that it precipitates out and forms a white crystalline
precipitate, this should be mentioned in the procedure for preparation of
the absorbing solution. It also should be stressed, that before the absorb-
ing solution is poured into the impingers the bottle of absorbing solution
must be thoroughly shaken to make sure that the precipitated hydroxide does
actually get into the impingers.
The procedure specifies a sampling rate of Z-rl/2 liters/minute
for 10 minutes, for a total of 25 liters of gas sampled. However, the first
day, it was found that 25 liters of gas, at the concentration of H2S in
this refinery stream, was exhausting the 30 ml of cadmium hydroxide before
the end of the sampling period. At the concentration of l^S (200 to 500 ppm)
in the refinery gas stream, an adequate sample volume was 10 to 15 liters
of gas. Using this sample volume, the cadmium hydroxide was not exhausted
during sampling.
At the start the procedure as presented in Appendix A, for the
Recovery of the Collected Sample, was followed. But a problem arose in
transferring the contents of the impingers quantitatively to a plastic
storage bottle. The washing of each impinger and the connecting
48
-------
glassware with the distilled water did not remove all of the precipitated
cadmium sulfide present in the impingers and the connecting glassware.
Attempts were made to rinse the glassware with 10% by weight HC1,
but loss of H2S when the CdS dissolved was observed.
According to the proposed method in Appendix B, the CdS precipitate
was filtered through a Whatman No. 40 filter paper, but this did not work.
The filter paper did not retain the cadmium sulfide. Filtering was then
tried through glass crucibles with the porous plates (filter crucible) and
rinsing the sample container and the precipitate with 10% ammonium hydroxide
solution. This was not successful either, due to the inability to recover
all the CdS from the glassware and the inability of the crucible to retain
the precipitate.
Additionally, the retained filter cake could not be dissolved
and recovered because of the loss of H^S.
Because of these problems, the intermediate filtering step was
discarded and sample recovery was accomplished according to the procedures
given in the suggested method outlined in Section III-B.
The method using the iodometric thiosulfate titration is a usable.
method as long as the analysis is performed within 1 hour of the sampling
time. Transportation of the samples back to a central lab for analysis is
impractical unless it is transported back in the same container that the
sample was obtained in. In other words, it would have to be transported
back in the midget impingers and analyzed in the midget impingers at the Iab0
49
-------
The preferred method is to take the sample and dissolve the cadmium hydroxide
in the impingers and then transfer qualitatively the dissolved materials to
the iodine flask. Once the cadmium sulfide has been dissolved in the acidi-
fied iodine solution, the analysis should be completed within 30-45 minutes
before the loss of iodine begins to be troublesome.
Figure 16, is a schematic of the analytical setup for the gas
chromatographic flame photometric analysis of H2S.
The sampling for the GC consisted merely of injecting gas from the
refinery gas stream into the sample valve and from the sample valve directly
to the gas chromatograph column into the flame photometer and detector.
The GC is a Perkin Elmer Mark II* with a stainless steel gas sampling valve,
Melpar Model FTD100* flame photometric detector, a Microtech Model 8176*
solid state electrometer, and a Leeds and Northrup Speedomax XL* recorder.
Nitrogen was used as the carrier gas and a mixture of hydrogen, oxygen,
and air was burned in the flame photometer.
Due to the mutual solubility and reaction of methylethylamine
and H2S problems were experienced in the operations of the gas chromatograph.
The Teflon column became saturated with methylethylamine and would not elute
H2S to the detector. In the future for GC work, it would be wise to include
a column to scrub out the methylethylamine ahead of the 36-foot Teflon
column used for hydrogen sulfide resolution.
* Mention of a company name does not indicate endorsement, by EPA.
50
-------
STANDARD
H2S 46ppm IN N2
Sample
.Line
rtf
N2
AIR
25 cc/mi
-33-
15 cc/mln
mm
Flame Photometer and Detector
Sample Gas Chromatograph
Valve Column
Electrometer
70 cc/min
Power Regulator
Flowmeter
Assembly
J110 Volt
Recorder
Figure 16 - Gas Chromatograph Analytical Train
-------
Figures 17 through 26 are photographs taken at the refinery
during the sampling program. Figure 17 shows the sampling train with the
impingers in the ice water bath, the Teflon sample line and the dry gas
meter for recording the temperature and amount of gas sample.
Figure 18 shows the Teflon sampling line to the gas sampling
valve on the chromatograph as well as the general layout of the chromato-
graphic analytical equipment.
Figure 19 is a side view of the gas chromatograph showing the
set up, and Figure 20 is the view of the gas chromatograph showing the
cylinder of the l^S standard gas, the sample valve, the gas chromatograph,
the flame photometer, and detector, the electrometer, the flowmeters and
the recorder integrator and the power supply regulator.
Figures 21 through 25 show the sequence of analysis using the
iodometric titration. Figure 21 shows the impingers with the yellow cad-
mium sulfide precipitate, Figure 22 shows the flask with the dissolved
sulfide in the acidified iodine before titration. Figure 23 shows the
flask during titration and before the addition of the starch indicator.
Figure 24 shows the flask after addition of the starch indicator, and
Figure 25 shows the flask at the end point of the iodometric titration.
Figure 26 shows the Bausch and Lomb Spectral Photometer Model No. 20* with
a blank and the H2S methylene blue solutions ready for analysis.
* Mention of a company's name does not indicate endorsement by EPA.
52
-------
Figure 17 - Impingers and Dry Gas Meter
for Hydrogen Sulfide Sampling
Figure 18 - Teflon Sample Line to Gas
Sampling Valve on GC
Figure 19 - Side View of Gas Chromatograph
53
-------
Figure 20 - Gas Chromatograph Setup
Showing - H2S Standard, Sample Valve -
GC-Flame Photometer-Electrometer-Flow
Meters-Recorder Integrator
Figure 21 - Impingers with CdS
Figure 22 - Flask with CdS - Acidified
Iodine Before Titration
54
-------
Figure 23 - Flask During Titration Before
Addition of Starch Indicator
Figure 24 - Flask After Addition of
Starch Indicator
Figure 25 - Flask Showing End Point of
lodometric Titration
Figure 26 - B and L Spec 20 with Blank
and HoS-Methylene Blue Solutions
55
-------
VI. PROCESS OPERATING CONDITIONS
This section to be furnished by EPA.
56
-------
APPENDIX A
DETERMINATION OF H2S IN STACK GASES
The following represents the draft method of sampling for hydrogen
sulfide and analyzing by the methylene blue-spectrophotometric method. This
method was furnished by EPA. Observations on this method and the revised
method are tabulated in Section IV.
57
-------
DETERMINATION OF H2S IN STACK GASES
1. Principle and Applicability
1.1 Principle
H2S is collected in a series of midget impingers and reacted with
alkaline Cd(CH)2 to form CdS. The CdS is filtered and rinsed with cold water.
It is then dissolved in HC1 and reacted with N,N dimethyl - p-phenylenediamine
in sulfuric acid to produce methylene blue, which is determined spectrophoto-
metrically at 670 mm.
1.2 Applicability
This method should be applied only when specified by the test
procedures for determining compliance with the New Source Performance Standards.
2. Range and Sensitivity
This method is reported as usuable down to 5 ppb, and to beyond
500 ppm.
3. Interferences
SOo in concentrations 100 times greater than HoS is reported to
cause some interference.
4. Apparatus
4.1 The apparatus used shall be the same as that used in Method 4,
F.R. 36, 24887, 23 December 1971.
4.2 Spectrophotometer shall be capable of operation at 670 nm.
58
-------
5. Reagents
5.1 Sampling
5.1.1 Absorbing solution - Mix 4.3 g cadmium sulfate (30dSO^
8H20) and 0.3 g of NaOH in 1 liter of distilled water. Mix well before
using.
5.2 Analysis
5.2.1 Amine-acid stock solution - Add 50 ml of concentrated
sulfuric acid to 30 ml of water and cool. CAUTION: Add H2S04 slowly or
spattering may result. Add 2 g of N,N-dimethyl - p-phenylenediamine. Stir
until dissolved.
5.2.3 Ferric chloride solution - Dissolve 100 g of ferric chloride
hexahydrate (Fe2Cl3 - 6H20) in water and dilute to 100 ml.
5.2.4 4 N HC1 - Dilute 340 ml of concentrated HC1 to 1 liter.
*
6. Procedure
6.1 Sampling
6.1.1 Assemble the sampling apparatus as shown in Figure 4-1,
F.R. 36, 24887, connecting two midget impingers in series. Transfer 15 ml
of absorbed solution to each impinger.
6.1.2 Sample at a rate proportional to the stack velocity until
a 25-liter sample has been obtained.
6.1.3 Transfer the contents of the impingers quantatively to a
plastic bottle, washing each impinger several times with distilled HoO.
59
-------
6. 2 Analysis
6.2.1 Filter the contents of the sample container 'through each
of two sintered glass crucibles. Rinse several times with 2 ml portions of
cold water. Discard the effluent. Dissolve the CdS in 4 H HC1 and immediately
add 3 ml of amine test solution and one drop of ferric chloride solution.
Agitate after each addition. Transfer to 50 ml volumetric flasks, make up
to volume, and allow to stand for 30 minutes. Run a blank in the same
manner using unaspirated absorbing solution. Determine the absorbance of
the sample at 670 nm vs the blank. Determine the l^S concentration from a
previously prepared standard curve.
7. Standardization
7.1 Using analyzed reagent grade CdS, prepare a working calibration
curve starting at Section 6.2.1 of the procedure. The calibration curve
should have a minimum of five points.
8. Calculations
CH2S = Ccds x VSOLN x 135756 X °-0154 gr/mg X Vj
CT, g = Concentration of H^S in gas analyzed
Cr,c = Concentration of CdS from calibration curve in mg/ml
Lido
Vg = Volume of gas sample in SCF
= Fi-nal dilution volume of solution read in spectrophotometer
= Ratio of mol. wt H0S/mol wt CdS.
144.46
60
-------
9. Bibliography
Jacobs, M. B., M. M. Braverman, and S. Hochheiser. "Ultramicrodetermina tion
of Sulfide in Air," Anal. Chem.. 29, 1349 (1957).
61
-------
APPENDIX B
DETERMINATION OF H2S IN REFINERY FUEL GASES
The following represents the draft method of sampling for hydrogen
sulfide and analyzing by the iodometric titration method. This method was
furnished by EPA. Observations on the method and the revised method are
tabulated in Section IV.
62
-------
DETERMINATION OF H2S IN REFINERY FUEL GASES
1. Principle and Applicability
1.1 Principle
H2S is collected in a series of midget impingers and reacted with
alkaline Cd(OH)2 to form CdS. Then CdS is filtered and rinsed with 10% NH.OH.
It is then dissolved in HC1 and absorbed in a known volume of iodine solution.
The excess iodine is titrated with thiosulfate solution, with starch as the
indicator.
1.2 Applicability
This method should be applied only when specified by the test
procedures for determining compliance with the New Source Performance
Standards.
2. Range and Sensitivity
3. Interferences
4. Apparatus
4.1 Sampling
The apparatus used shall be the same as that used in Method 4,
F.R., 36, 24887, 23 December 1971.
4.2 Analysis
Glass stoppered 500 ml iodine number flask for reaction vessel.
63
-------
5. Reagents
5.1 Sampling
5.1.1 Absorbing solution.
Mix 4.3 g cadmium sulfate hydrate 3 CdS04'8H20 and 0.3 g of NaOH
in 1 liter of distilled H-O. Mix well.
5.2 Analysis
5.2.1 10% Ammonium hydroxide solution
Add 380 ml of concentrated ammonium hydroxide (s.g. 0.90) to
620 ml of distilled H20.
5.2.2 10% Hydrochloric acid solution.
Mix 230 ml of concentrated HC1 (s.g. 1.19) and 770 ml of distilled
H20.
5.2.3 Iodine solution, O.lN.
Dissolve 25 g of potassium iodide (KI) in 30 ml of distilled H20
in a 1 liter graduated cylinder. Weigh 12.7 g of resublimed iodine (I2)
into a weighing bottle and add to the potassium iodide solution. Shake the
mixture until the iodine is completely dissolved. Slowly dilute the solu-
tion to 1 liter with distilled 1^0, with swirling. Filter the solution, if
cloudy, and store in a brown glass stoppered bottle.
5.2.4 Iodine solution, 0.01 N.
Dilute to 1 liter with distilled water 100 ml * 0.01 ml of the
0.1 N iodine solution in a volumetric flask.
64
-------
5.2.4 Iodine solution, 0.01 N.(Continued)
Standardize daily as follows: Pipette 100 ml of the 0.01 N iodine
solution into a 500-ml conical flask. Titrate with the standard 0.01 N thio-
sulfate solution until the solution is a light yellow. Add a few drops of
the starch solution and continue titrating until the blue color just
disappears.
5.2.5 Sodium thiosulfate solution.
Sodium thiosulfate solution, standard 0.1 N. For each liter of
solution, dissolve 24.8 g of sodium thiosulfate (^28203'51^0) in distilled
water and add 0.01 g of anhydrous sodium carbonate (Na£ CO-j) and 0.4 ml of
chloroform (CHC13) to stabilize. Mix thoroughly by shaking or by aerating
with nitrogen for approximately 15 minutes, and store in a glass-stoppered
glass bottle.
Standardize frequently as follows: Weigh into a 500-ml volumetric
flask about 2 g of potassium dichromate (lOjC^Oy) weighed to the nearest
milligram and dilute to the 500-ml mark. Use dichromate which has been
crystallized from distilled water and oven-dried at 360°F to 390°F. Dissolve
approximately 3 g of potassium iodide (KI) in 50 ml of distilled water in a
glass-stoppered, 500-ml conical flask, then add 5 ml of 20% hydrochloric
acid solution. Pipette 50 ml of the dichromate solution into this mixture.
Gently swirl the solution once and allow it to stand in the dark for 5 minutes.
Dilute the solution with 100 ml to 200 ml of distilled water, washing down
the sides of the flask with part of the water. Swirl the solution slowly
and titrate with the thiosulfate solution until the solution is light yellow.
65
-------
Add 4 ml of starch solution and continue with a slow titration with the
thiosulfate until the bright blue color has disappeared and only the pale
green color of the chromic ion remains.
5.2.6 Sodium thiosulfate solution.
Sodium thiosulfate solution, standard 0.01 N. Dilute 100 * 0.01 ml
of the standard 0.1 N thiosulfate solution in a volumetric flask to 1 liter
with distilled water.
5.2.7 Starch indicator solution.
Suspend 10 g of soluble starch in 100 ml of distilled water and
add 15 g of potassium hydroxide pellets. Stir until dissolved, dilute with
900 ml of distilled water, and let stand 1 hour. Neutralize the alkali with
concentrated hydrochloric acid, using an indicator paper similar to "alkacid"
test ribbon, then add 2 ml of glacial acetic acid as a preservative.
Test for decomposition, by titrating 4 ml of starch solution in
200 ml of distilled water, with the 0.01 N iodine solution. If more than
four drops of the 0.01 N iodine solution are required, make up a fresh
starch solution.
6. Procedure
6.1 Sampling
6.1.1 Assemble the sampling train as shown in Figure 4-1, F.R. 36
24887 connecting two midget impingers in series. Place 15 ml of the collec-
tion solution in each impinger.
66
-------
6.1.2 Purge the connecting line between the fuel gas sampling
valve and the first impinger. Connect the sample line to the train. Read
and record the initial reading on the dry gas meter.
6.1.3 Open the flow control valve and adjust the sampling rate
to 2.5 liters/min. Read and record the meter temperature.
6.1.4 Continue sampling for 10 minutes. At the end of this time,
close the flow control valve and read the final meter volume and temperature.
6.1.5 Disconnect the impinger train from the sampling line and
cap the open ends. Remove to the sample clean-up area.
6.2 Sample Recovery
6.2.1 Transfer the contents of the impingers quantitatively to a
plastic storage bottle. Wash each impinger and the connecting glassware
several times with distilled water, and add these washings to the storage
container. Cap and seal the storage container.
6.3 Analysis
6.3.1 Filter the contents of the sample container through each of
two retentive filter papers, (eq. Whatman 40). Rinse the sample container
and the precipitate twice with 10% ammonium hydroxide solution.
6.3.2 Place the filter papers with the precipitate in the iodine
number flask. Add a measured excess of Standard 0.01 N iodine solution and
immediately add 50 ml of 1070 HC1 solution. Stopper and shake. (Enough
standard iodine solution should be added to give a back titration equivalent
to approximately 50% of the added iodine solution.)
67
-------
6.3.3 Titrate the solution in the flask with 0.1 N sodium thio-
sulfate solution until the solution is light yellow. Add 4 ml of the starch
solution and continue titrating until the blue color just disappears.
6.3.4 Run a blank determination beginning at 6.3.1 to adjust the
value of iodine consumption.
7. Calculation
7.1 Gas volume sampled at standard conditions (70°F, 29.92 in Hg)
v = 17.7 5n V ft3
(Tm + 460) m
Pm = meter pressure "Hg abs
Tm = meter temperature, °F
•3
V_ = meter volume ft
0.2618(AN, - BN,)
CH2S = =—-1 gr/SCf
ms
where
A = volume of standard iodine used, ml
N- = normality of the standard iodine solution
B = volume of standard sodium thiosulfate used, ml
N£ = normality of the standard thiosulfate solution
Vms = volume of gas sampled at standard conditions
0.2618 = (34 g/mole H2S) (0.0154 gr/mg) (1000 mg/g)
(1000) (2 H2S equivalent)
68
-------
APPENDIX C
FIELD DATA SHEET
The Field Data Sheets are contained in this Appendix.
69
-------
GAS SAMPLING FIELD DATA
Material Sampled for_
Date
Plant
Ambient Temp
Run No.
/(£>
/i
Power Stat Setting
Filter Used: Yes
Operator
No
Location
F Stack Temperature
Stack Dimensions
Irapingers with
Ini)ingers with
Total number of Impingers
Sample Bottle No.
Impinger Bucket No.
Meter Box No. U'*2.
ml of
°F
.£ y/4-L \/
\/ ml of /
-
70
-------
Clock
Time
Meter (ft3)
Flow Meter
Setting (cfh)
Meter Temperature
in
o.ooo
/ooo
- A
/ p
'/<*> ra*.
77
0.321
71
1/3- 1
0
nsv
0.3*7
n.ooo
\
O.&&0
134-7
0.323
2.0
1411
0 .BOO
2.0
3.0
-------
GAS SAMPLING FIELD DATA
Material Sampled for
Date <£ $ J ifflf 72-
Vs<- —
Plant UW/DAI Vs<- —*vL/&T_£ c Location
Ambient Temp. "7J °F Stack Temperature_
Run No.,^?'^ 'T&S1' F}/)f Stack Dimensions
Power Stat Setting
Filter Used: Yes No
Operator
Impingers with _ _ ml of
Impingers with _ ml of / c
Total number of Impingers
Sample Bottle No. —
Impinger Bucket No.
Meter Box No.
-------
Clock
Time
Meter (ft3)
Flow Meter
Setting (cfh)
Meter Temperature
in
//ft
O.ooo
lot
73*
0. QOO
n
O.ooo
O . & b O
/7
O.f/7
• o
7
3.o
7/
0.313
O. 6x30
1
Q.232-
72*
2.0
n.ooo
3.0
/**>/
A//
^L
23.
£.
^_0
O.M'
73
-------
7-2-
'e^f
f/*
&<
M// '
j
Clock
Time
/Ct2f
' U&
J7/0
n^o
•
Meter (ft3)
O BOO
O 3 2^=*
0.00O
0. 3^P-
Flow Meter
Setting (cfh)
3-.o
3.o
a~*
<3.
-------
GAS SAMPLING FIELD DATA
Material Sampled for_
Date
Plant
[}
to i /
Location
Ambient Temp . G? f? f~ °F Stack Temperature
Run No.
Stack Dimensions
Power Stat Setting
Filter Used: Yes
Operator_
No
I
Impingers with
J-
Impingers with
Total number of Impingers_
Sample Bottle No.
Impinger Bucket No.
Meter Box No. Cs
ml of
ml of
(/ '
75
-------
92-
Clock
Time
Meter (ft3)
0.000
0.33^
Flow Meter
Setting (cfh)
Jo
_£0_
Meter Temperature
in
oQfo
0.3V/
0.000
fi.OOO
-2. ?
/of?
1//7
O.3S3
0.31^
*J_jtW
o.ooo
o,
s.i
3.0
//a./
0.
/,
\33o
Mo*
D.&ev
0.301
-------
Clock
T±me
Meter (ft3)
Flow Meter
Setting (cfh)
Meter Temperature
in
.00°
73 '
fa
/fro
o.
0.000
oo
0.3
2.3
T&Vf
I.I
0*3 80
1L
O.QQO
1.4-
ll /
0.302.
2.1
3.1
77
-------
GAS SAMPLING FIELD DATA
Material Sampled for
Date 30
JL
Ifr/y
Plant L/AJ/osJ ([//(-- ^oL^^'f Location
Ambient Temp. X"«J °F Stack Temperature
Run No. _ . _ Stack Dimensions
Power Stat Setting — ~" _
Filter Used: Yes No fX^
Operator
Impingers with / S"~ _ ml of
CJ **H )-,
/ _ Impingers with
Total number of Impingers_
Sample Bottle No. —
Impinger Bucket No.
Meter Box No. Cf "~
78
-------
•30
•"£*/
Clock
Time
Meter (ft3)
Flow Meter
(cfh)
Meter Temperature
in
0-000
owe
030+
A3.
09.29
Q.OOO
2.3-
79
o.o&o
1.1
O
f.e
3.S
2.3.
3.7
f?
^'
0.3 t,l
1?
Jtt?-
O.Zf?
4). OOo
?
130(0
^_
^
fl-Q^o
^i£
V.ZJt,
79
-------
72-
Clock
Time
Meter (ft3)
Flow Meter
Setting (cfh)
Meter Temperature
in
w
4)2-
0.309
'f
0.000
14-31
a
/.f
. wo
0.349
3.1
3.1
0.
2.
Q.QOO
f
o.
O'.ooo
O.ooo
I.-J
0.3o4
xrf
5.
SZ'f
tl'f
/Iff
sn
f
0,
sn
t&f
0.273
80
-------
APPENDIX D
LABORATORY REPORT
(Titration Data)
This section presents the iodometric titration and the methylene
blue-spectrophotometric field data sheets.
81
-------
//
-?
00
-------
/
n,l
/
"r .
&< O/
2-3,0
83
-------
/I
O
84
-------
85
-------
&^
7:3
86
-------
72^
IAJ$
^o s ^ //
2
/?'•
IhtO
f /:3 5
87
-------
-*
J^^y?A
sr.B
Scv^/e^JL/J
^
3O.T
4-7,5
fl?
88
-------
1S '3
7 d : S'O
•//;
: 3
89
-------
-*5**°fl&&$£
/
-------
91
-------
7
92
-------
(C:
93
-------
94
-------
i«j£ "<3+~^t, ~&^s <£&?//
^
95
-------
4 2.
-30(0 -
/ :
I «/ 6" **> - / jT^ 3
, /
96
-------
/ <= O
y
&./&> .
97
-------
APPENDIX E
FIELD LOG
Table E-l presents the actual time during which sampling was
conducted.
98
-------
TABLE E-l
SAMPLING LOG
Time
Location
Amine scrubber effluent
low pressure line
Amine scrubber effluent
low pressure line
Amine scrubber effluent
low pressure line
Pollutant
H2S
CH3-SH
H2S
CH3'SH
H2S
Date
6/27/72
6/28/72
6/29/72
Began
10:30
11:25
9:06
Ended
15:35
17:20
16:42
CH3-SH
Amine scrubber effluent
low pressure line
H2S
CH3-SH
6/30/72
8:53 18:00
99
-------
APPENDIX F
PROJECT PARTICIPANTS AND TITLES
Name
Paul Constant
Evan P. Shea
Gary Kelso
Fred Bergman
Doug Weatherman
Mike Serrone
Title
Program Manager
Project Chief
Testing Engineer (gas)
Analytical Chemist (GC)
Analytical Chemist
Analytical Chemist
100
------- |