74-SRY-2
(REPORT NUMBER)
AIR POLLUTION EMISSION TEST
Union Oil Corporation
(PLANT NAME)
Wilmington, California
(PLANT ADDRESS)
U. S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Water Programs
Office of Air Quality Planning and Standards
Emission Standards and Engineering Division
Emission Measurement Branch
Research Triangle Park, N. C. 27711
-------
P.O. Box 13454• UNIVERSITY STATION • GAINESVILLE, FLORIDA 32604-904/372-3318
environmental science and engineering, inc.
AN EQUAL OPPORTUNITY EMPLOYER
73-011-034
fNVlRONMENTAL SCIENCES KlFf*f*CF ENVIRONMENTAL ENGINEERING
METEOROLOGY Am QUALITY MANAGEMENT
AQUATIC BIOLOGY DISPfRSION MODELING
FISHERIES BIOLOGY OCEANOGRAPHY
fOLLUTION BIOASSAYS COASTAL ENGINEERING
VflLDUfe ECOLOGY . M_ w»—U 1 O7/1 HYDROLOGY
BOTANY rlCl I LII 1 3 I *f GCOLOGY
CHEMISTRY GEOLOGICAL EttGINEffH*G
fAOtOCHfMtSJRY WASJEWAJEft MANAGEMENT
BIOSTATISTICS SOLID WASTE MANAGEMENT
CNVIRONMENTAL ECONOMICS OCCUPATIONAL SAf£TY/H£ALTH
SOURCE TEST REPORT
EMISSIONS FROM SULFUR RECOVERY PLANT
Union Oil Corporation
Wilmington, California
for
The Environmental Protection Agency
United States Government
Report No. 74-SRY-2
E. R. Hendrickson, Ph.D.
Senior Advisor
John D. Bonds, Ph.D.
Project Manager
OFFICE/LABORATORY LOCATION: FIVE MILES WEST OF INTERSTATE 75 ON STATE ROAD 26 (NEWBERRY ROAD)
-------
SOURCE TEST REPORT
REPORT NO: 74-SRY-2
PLANT TESTED: Union Oil Corporation
Wilmington, California
EMISSIONS FROM: Sulfur Recovery Plant
TESTOR: Environmental Science and Engineering, Inc.
Post Office Box 13454
University Station
Gainesville, Florida 32604
CONTRACT NO: 68-02-0232, Task Order No. 34
-------
TABLE OF CONTENTS
Page No.
1.0 INTRODUCTION 1
2.0 SUMMARY AND DISCUSSION OF RESULTS 3
3.0 PROCESS DESCRIPTION 8
4.0 LOCATION OF SAMPLING POINTS . . 19
5.0 ANALYTICAL PROCEDURES 21
APPENDICES .
A. COMPLETE EMISSION TEST RESULTS WITH SAMPLE
CALCULATIONS
A-l. CALCULATIONS . . A-l
A-2. SULFUR COMPOUNDS BY GAS CHROMATOGRAPHY ... A-4
A-3. C02, CO AND 02 BY CONTINUOUS MONITORING . . A-7
A-4. S02 BY EPA METHOD 6 A-10
A-5. .OUTLET FLOW CALCULATIONS A-l3
A-6. MOISTURE AND FLOW RATE (EPA METHODS 1, 2
AND 4) A-14
A-7. ORSAT BY EPA METHOD 3 A-20
A-8. N0¥ BY EPA METHOD 7 A-26
A
A-9. ODOR DETERMINATION BY EPA DRAFT METHOD
(DILUTION) A-30
A-10. VISIBLE EMISSIONS BY EPA METHOD 9 ...... A-34
B. FIELD DATA
B-l. SULFUR COMPOUNDS BY GAS CHROMATOGRAPHY ... B-l
B-2. S02 BY EPA METHOD 6 . . . B-34
B-3. MOISTURE ...... B-37
B-4. TRAVERSE POINTS B-40
environmental science and engineering, inc.
-------
TABLE OF CONTENTS (continued)
Page No.
B-5. VISIBLE EMISSIONS B-41
B-6. ODOR EMISSIONS B-56
C. LABORATORY REPORT
C-l. S02 BY EPA METHOD 6 C-l
C-2. NOY BY EPA METHOD 7 . . . • C-4
A
D. CALIBRATION STANDARDS D-l
E. PROJECT PARTICIPANTS E-l
environmental science and engineering, inc.
-------
1.0 INTRODUCTION
Under Section III of the Clear Air Act of 1970, as ammended, the Environ-
mental Protection Agency is charged with the establishment of standards of
performance for new stationary sources which may contribute significantly
to air pollution. A performance standard is based on the best emission
reduction systems which have been shown to be technically and economically
feasible.
In order to set realistic performance standards, accurate data on pollutant
emissions is normally gathered from the stationary source category under
consideration.
The sulfur recovery system at Union Oil Company of California's Los Angeles
Refinery at Wilmington, California, was designated as a well controlled
stationary source and was thereby selected by the Office of Air Quality
Planning and Standards (OAQPS) for an emission testing program. Tests
were conducted on Sulfur Recovery Unit. No. 2 during March 5-7, 1974. The
tests were performed by personnel from Environmental Science and Engineering,
Inc., Gainesville, Florida, and the Environmental Protection Agency, Emission
Measurement Branch, OAQPS, Research Trinagle Park, North Carolina.
The sulfur recovery system consists of a 100 LT/D Claus Sulfur Recovery
Unit followed by a Beavon Sulfur Removal Process (BSRP) Unit. The BSRP unit
treats the tail gas from the Claus Unit to remove additional sulfur before
the gas is released to the atmosphere.
environmental science and engineering, inc.
-------
Tests were conducted at a point downstream from the BSRP unit. The tests
were designed to determine the average emission rates during four-hour
sampling periods on each of three days. The emissions measured were:
sulfur compounds (hydrogen sulfide, carbonyl sulfide, carbon disulfide,
sulfur dioxide), hydrocarbons, carbon monoxide, nitrogen oxides, visible
emissions and odors.
-2-
environmental science and engineering, inc.
-------
2.0 SUMMARY AND DISCUSSION OF RESULTS
2.1 A complete summary of all data collected for the three four-hour
tests completed at Union Oil Company, Wilmington, California, is
included as Table 1. To facilitate a comparison of results all
concentrations are presented as ppmv dry (except % dry for C02
and 02), and all emission rates are standardized as grams per
hour (gm/hr). Sample calculations and conversions are presented
in Appendix A.
2.2 Sulfur compound concentrations were determined by gas chromatography
(COS, CS2, H2S, S02, and total sulfur), and EPA Method 6 (S02).
Additional total sulfur data which was scheduled to be collected
from the Meloy Sulfur Analyzer was not obtained due to instrument
malfunction.
Average sulfur dioxide concentrations obtained by EPA Method 6 and
gas chromatography were 4.0 ppmv and 1.0 ppmv, respectively. These
values are within reasonable agreement, especially since EPA Method
6 cannot be expected to yield accurate results at the S02 concen-
trations present in the stream.
The total sulfur results were obtained by summation of the individual
sulfur compounds as sulfur dioxide. The values obtained for total
sulfur ranged from 16 ppmv to 19 ppmv. The average value for the
three tests was 17 ppmv. The Meloy Total Sulfur Analyzer was
-3-
environmental science and engineering, inc.
-------
Table. 1
DATA SUMMARY - INLET AND OUTLET OF STRETFORD ABSORBER
UNION OIL COMPANY
Wilmington, California
Ojte
5 Kirch 1974
5 Kirch 1974
6 Kirch 1974
6 Kirch 1974
7 Kirch 1974
7 Kirch 1974
CUtc
5 Kirch 1974
6 Kirch 1074
7 Kirch 1974
location
Outlet
Inlet
Outlet
Inlet
Outlet
Inlet
AVCRAGC
location
Outlet
Outlet
Outlet
mwc
Flnw R.itr
WriJ/:i...f))
65.5
68.6
71.6
74.3
68.8
71.3
(OUTUr) 68.6
C0(l) Mj(3)
2.55 > 103 37.2
J.78 x 10' 42.1
2.C6 > 10J 48.8
2.73 « 10} 42.7
Orttthtt
*f6lttur» X COj XOj XCO
4.2 5.4 0.6 0
H.6 9.0 0.4 0
5.0 5.5 0.5 0
4.7 9.0 0.4 0
3.3 6.0 0.3 0
4.4 9.0 0.3 0
4.J 5.6 0.5 0
MASS [MISSION MItS, o»/hr
50^(4) COS(4) HjS(4) CSj(4)
IS. 4 1.64 > 13*
7.9 1.79 « I02
8.2 1.52 » 102
10.5 1.6S i 102
»Dia/ rjrj
*"*"' *°'UI
5.8 0.02
5.7 . 0.09
5.9 0.02
5.8 0.04
15(5) W,(6)
1.9S > I02 8.4
1.91 i 102 0
1.62 i I02 0
1.83 i 1C7 2.8
Eflilttlon Concf ntrjtlont , piwfv, dry
ryqnctSC Cjt Cftronut^'jr^iin Odor Concf ntrailo
CO, ppnv (1) SOj, ppmv(3) S02(4) COS! 4) N;S() CjJ Chrotu'.o^riph (7) tfA Dilution
(2) Pjrj-ojrctlc (i) lotul Sulfur it tOj (Sumitlon of 6.C. Sulfur Co^poun^l) 16) !?* >Vf>o
-------
inoperative during the testing period and no comparative data
were obtained.
Hydrogen sulfide, as determined by the gas chromatograph, was below
detectable during the testing period.
Carbonyl sulfide concentrations, as determined by gas chromatography,
ranged from 15 ppmv to 17 ppmv with a mean value of 16 ppmv for the
three tests.
The carbon disulfide concentration was below detection on the first
day of testing. On the succeeding tests an unknown compound with a
retention time slightly less than carbon disulfide was obtained which
prevented the determination of low level concentrations of carbon
disulfide. No attempt was made to identify this compound since
reference materials were not readily available and the time frame
for the project was not flexible enough to allow any field research.
2.3 Carbon dioxide, carbon monoxide, and oxygen were determined by con-
tinuous methods (NDIR and paramagnetic) and by the Orsat method. The
following paragraphs compare the results obtained by the various methods.
Carbon dioxide concentrations were determined by NDIR and by Orsat.
The results obtained varied from 5.4% to 6.0% (Orsat) and 5.6% to
5.8% (NDIR). The average values for the three tests were 5.6% and
5.7% (NDIR). This agreement between the two methods is within
experimental error.
-5-
environmental science and engineering, inc.
-------
Oxygen concentrations were determined by Orsat and by paramagnetic
analyzer. Average values obtained for the Orsat method was 0.5%
and for the paramagnetic method was 0.2%. Although the values
are not in good agreement, the Orsat method has limitations at
low concentrations and the paramagnetic analyzer values should
be considered to more closely approximate the true concentration
of oxygen in the sample gas stream.
Carbon monoxide concentrations were measured by NDIR..and Orsat.
The concentrations determined by NDIR ranged from 556 ppmy to 604
ppmv. These concentrations are below the range applicable to Orsat
measurements as is shown by the fact that no carbon monoxide was
measured by the Orsat method.
2.4 Nitrous oxides were determined by EPA Method 7. The concentration
of NOX varied from less than detectable to 1.1 ppmv as NOo (daily
average) for the test period.
2.5 Visible emission were determined by qualified observers in accordance
with EPA Method 9. The average visible emissions obtained for the
three tests was zero.
2.6 Hydrocarbons were not determined during the source tests performed
due to a malfunction of the instrumentation.
environmental science and engineering, inc.
-------
2.7 Odor concentrations were determined according to an EPA Draft
Method (Dilution Method). The odor concentration ranged from
2.0 to 3.5 odor units per standard cubic foot (ou/SCF), with an
average of 2.6 ou/SCF. This means that when the stack gas was
diluted by an average factor of 2.6, the odors were non-detectible
by 50% of the evaluation panel.
2.8 Moisture and flow rates were determined on the inlet and outlet of
the Stretford Absorber according to EPA Methods 1, 2, and 4. The
moisture content of the inlet varied from 4.4% to 11.6% with an
average value of 6.9%. The moisture content of the outlet ranged
from 3.3% to 5.0%, with an average value of 4.2%. Flow rates for
the inlet and outlet averaged 71.6 DNM3/min and 68.6 DNM3/min,
respectively.
2.9 The overall results from the various tests S02, CO, C02,
COS, C$2 » etc.) seem to be in agreement with the results expected
from the Beavon Sulfur Removal Process. Both the magnitude and
concentration range of the individual pollutants agree with theorized
results.
-7-
environmental science and engineering, inc.
-------
3.0 PROCESS DESCRIPTION
3.1 Claus Sulfur Recovery
In petroleum refining, various processes generate "sour" gas streams
which contain not only sufficient amounts of hydrocarbons to be used
as a fuel gas, but also contain contaminants such as carbon
dioxide and hydrogen sulfide. These fuel gases are treated to remove
C02 and H^S, and in regenerating the treating solutions by steam strip-
ping "acid" gases are evolved which contain concentrated H_S and s&me
co2.
Most refineries recover the H2S as elemental sulfur by the Claus process
shown below:
Reheat
Reheat
Reheat
, t
' ' Reaction
'|~~) Furnace j
/A
/,
C' r—
Cond.
Boiler Feed Water
*
<-
/ i
wcrto'
**
/*
CO!
^
v
id.
/
)
C
_
'
Coiv/nrtcr y
'
c
r
_
Cond.
'----- -
C
i k
j'rd Stayed
Convertery
V >
k
Cond.
x "
•
1
Removal Unit
h ->
Steam »
f
• i
r Liquid
Sulfur
L , u..,i, , ,
Sulfur Pit
Figure 1. Schematic diagram of Claus Process.
For the high concentrations of H2S usually found in'refinery acid gases,
the "straight through" variation of the Claus process is used.
-3-
science and
enc.
-------
In this process, HoS is partially oxidized in the reaction furnace:
H2S + 3/2 02 , - > S02 + H20 + 131 Kcal
The S02 then reacts with the remaining hUS:
2H2S + S02 ; - * 3S + 2H20 + 28 Kcal
The overall reaction, commonly called the Claus Reaction, is
3H2S + 3/2 02 7 - » 3S + 3H20 + 150 Kcal
Stagewise condensation, reheating, and catalytic conversion steps push
the Claus reaction to the right and remove most of the sulfur gases as
elemental sulfur. The efficiency of the Claus Process increases with
increasing H2S inlet concentration and the number of conversion stages.
Some side reactions occur in the reaction furnace which lower Claus effi-
ciencies. Carbonyl sulfide (COS) and carbon disulfide (CS2) are formed
at high temperatures if the acid gas contains C02 or hydrocarbons or both:
CH4 + 4S , - > CS2 + 2H2S
C02 + H2S 7 - > COS + H20
COS + H2S ; - -> H20 + CS2
Though present in relatively small quantities, COS and CS2 are not
-9-
environmental science and engineering, inc.
-------
recovered in the Claus process and do become significant in the Claus
tail gas after h^S and S02 levels have been reduced by 90-95% or more.
Also, the Claus reaction, being exothermic, is favored by lower tempera-
tures, hence the reaction continues in the tail gas to some extent. A
typical tail gas analysis from a Claus plant at 94% sulfur removal is:
Component % Volume
H2S 0.85
S02 0.42
S8 0.05
COS 0.05
CS2 0.04
H20 33.10
CO 0.22
C02 2.37
N2 61.30
H2 1.60
Thus any tail gas treatment to remove sulfur levels below 500 ppm must
address all five sulfur constituents. One of these processes to remove
sulfur from tail gases is the Beavon Sulfur Removal Process, described
in the following section.
-10-
environmental science and engineering, inc.
-------
3.2 Beavon Sulfur Removal Process - Commercial Status
The Beavon Sulfur removal process was developed by the Ralph M. Parsons
Company and the Union Oil Company and is licensed by the Union Oil Com-
pany of California. The following table summarizes the status of
Beavon Sulfur Removal Process units applied to Claus tail gases at
the time the tests were made. Additional units have been committed
since that time.
Company/location
Union Oil Company/
Wilmington, California
Mobil Oil Company/
Tbrrance, California
Atlantic-Richfield/
Philadelphia, Pennsylvania
Getty Oil Company/
Delaware City, Delaware
Kobe Steel Co./Japan
Texaco, Inc./
Long Beach, California
Unknown/
Carribbean
Union Oil Co./
Rodeo, California
Onstream Date
July 1973
July 1973
September 1973
November 1973
October 1973
March 1974
April 1974
November 1974
Number/Capacity of
Claus Plant. LT/D
2/100*
3/100
1/140
1/300
1/220
* During EPA tests, sulfur recovery averaged 35 LT/D.
3.3 Beavon Sulfur Removal Process - Process Description
A schematic diagram of the Beavon Sulfur Removal Process is shown as Figure
2. The Claus plant tail gas is reheated in a combination reheater/reducing
gas (CO and Ho) generator and then fed to a hydrogenation reactor where
-11-
environmental science and engineering, inc.
-------
all sulfur species are converted to H2S and CO is converted to C02. The
reactions are:
SJ. U
T n^ — —
S02 + 2H2
CS2 + 2H20 —
COS + H20
CO + H?0
•> H2S
-» H2S +
— » 2H2S
-» H2S +
-> co2 +
2H20
+ co2
co2
Ho
tJater vapor is then removed from the tail gas by a quench tower and re-
turned to sour water strippers.
After quenching, the gases are fed to a Stretford solution absorber. The
Stretford process is well established, with the initial plant having begun
operation about 15 years ago. There are approximately 60 operating
plants world wide.
»
In the Stretford absorber, the H9S is absorbed in an aqueous solution of
sodium carbonate, sodium metavanadate, and anthraquinone disulfonic acid
(ADA). The sodium carbonate provides solution alkalinity for initial
absorption of the H2S per the following equation:
H2S + Na2C03 - * Na+ + HS~ + NaHC03
The HS~ ion then reacts with metavanadate to precipitate colloidal sulfur
as follows:
HS" + 2V5+ - » S + 2V4+ + H+
The vanadium, in turn, is reoxidized by the ADA:
V4+ + ADA - » V5+ + Reduced ADA
-12-
environmental science and engineering, inc.
-------
From the reactor, the solution flows through a series of oxidizers where
the ADA is oxidized with blown air. The sulfur is recovered as a froth,
filtered, washed, melted and transferred to the Claus plant sulfur pit.
The oxidized solution is continually recycled to the absorber for con-
tacting additional gases.
Tail gas from the Stretford unit is not incinerated on units in operation
thus far. Overall sulfur recovery is 99.8+%.
BEAVON SULFUR REMOVAL PROCESS
Tail Gas
S02 COS CS2
Reducing Gas Generator
Stretford
Absorber
Cobalt-Molybdate Catalyst
Oxidation
and Froth
Tanks
Sulfur
Melter
Molten
Sulfur
Figure 2. Schematic diagram of Beavon Sulfur Removal Process
environntental science and engineering, inc.
-------
3.4 Plant Operation During Emission Measurements
Process variables were monitored during the test periods to assure
that measured emissions were representative of normal plant operation.
These variables included:
(1) acid gas flows from amine units
(2) acid gas flow from sour water strippers
(3) tail gas H2S/S02 ratio
(4) excess hydrogen from Beavon reactor
(5) Stretford absorber liquid flow rates
During the test periods, the. process, was very steady as indicated
by the monitored method variables.
Since the refinery was running at 80,000 barrel/day crude versus a
normal 108,000 barrel/day rate due to restricted crude supply, the
Claus plant was operated well below the designed 100 LT/D rate. From
acid gas feed rates, sour water stripper gas rates, and known h^S
concentrations in these streams, sulfur feed was calculated:
Date Sulfur Produced LT/D
3/5 33.7
3/6 35.8
3/7 33.5
Corresponding process data taken during these emission tests are
summarized in detail in Tables 2, 3 and 4. Data pending confidential
-14-
environmental science and engineering, inc.
-------
determination are stored in the confidential files of the Emission
Standards and Engineering Division, OAQPS, Research Triangle Park,
North Carolina 27711.
-15-
environmental science and engineering, inc.
-------
Table 2. PROCESS DATA SHEET - UNION OIL TEST
Emissions From Sulfur Plant Measured March 5-7, 1974
OBSERVER: C. Sedman
liOCATION: Wilmington, California
DATE: March 5, 1974
Time
Amine acid gas, SCFH
SWS acid gas, SCFH
Tail gas H2S/SC2
Excess Hp from reactor, %
Absorber liquid rate, GPM
1330
36,000
22,000
3.0
1400
36,500
22,000
3.3
Confidential
. Detei
-mined Cc
1430
37,000
22,000
2.9
1500
38,000
22,500
3.1
1530
38,500
22,500
3.5
1600
38,000
22,500
3.4
Status Being Determined
)nfidenti
al
.1630
38,000
23,000
3.2
1700
39,000
22,500
3.1
1730
39,000
22,000
3.0
1800
39,000
22,500
3.0
1830
39,000
22,500
2.9
cr>
i
Sulfur Feed
Amine: [38>000][.82l[32][24] c
[380][?;:40]
SWS : [22>400][.29&][32][24]c
[380][2240]
33.7 LT/D
-------
Table 3. PROCESS DATA SHEET - UNION OIL TEST
Emissions From Sulfur Plant Measured March 5-7, 1974
OBSERVER: C. Sedman
LOCATION: Wilmington, California
DATE: March 6, 1974
Time
Amine acid gas, SCFH
SWS acid gas, SCFH
Tail gas H2S/S02
Excess H2 from reactor, %
Absorber liquid rate, GPM
1130
36,000
31 ,000
3.6
1200
36,000
.33,000
3.5
Confidential
Determined Cc
1
1230
36,000
34,000
3.2
1300
36,000
35,000
3;0
1330
36,000
36,000
2.8
1400
36,000
35,500'
2.9
Status Being Determined
>nfident
al
1430
36,000
34,000
3.4
1500
36,000
34,000
2.9
1530
36,000
34,000
2.6
1600
36,000
34,500
2.8
Sulfur Feed Amine: [36,000][.821][32][24] = 26J LT/p
[380][2240]
SWS :
[380] [2240]
= 9.1 LT/D
35.8 LT/D
-------
Table 4. PROCESS DATA SHEET - UNION OIL TEST
Emissions From Sulfur Plant Measured March 5-7, 1974
OBSERVER: C. Sedman
LOCATION: Wilmington, California
DATE: March 7, 1974
Time
Amine acid gas, SCFH
SU'S acid gas, SCFH
Tail oas H7S/S02
Excess Hp from reactor, %
Absorber liquid rate, GPM
1100
36,000
26,000
3.2
1130
36,000
25,500
3.0
Confidential
Detei
'mined Cc
1200
36,000
24,500
2.8
1230
36,000
24,500
2.8
1300
36,500
24,500
3.0
1330
37,000
24,500
3.5
Status Being Determined
mfidenti
al
1400
37,500
24,500
3.2
1430
37,000
24,500
2.7
1500
36,000
24,500
2.8
1530
36,000
24,000
3.1
Sulfur Feed
Amine: [36,400][.821][32][24]
[380][2240]
SWS : [24,700][.296][32][243
[380][2240]
26.9 LT/D
6.6 LT/D
33.5 LT/D
-------
4.0 LOCATION OF SAMPLING POINTS
The sampling points selected for emission tests at Union Oil Company,
Wilmington, California, are shown in Figure 3.
The outlet sample location, downstream from the Stretford absorber, was
used to obtain samples for the EPA van (gas chromatographic systems),
sulfur dioxide, moisture, velocity, integrated bag, fixed gases (carbon
monoxide, carbon dioxide and oxygen), Orsat and the odor analysis panel.
The sampling port at this point consisted of a gate valve with a 2"
opening.
The inlet sample location, upstream from the Stretford absorber, was used
to determine moisture and velocity. The sampling port at this location
consisted of a gate valve with a 2" opening.
No major problems were encountered in using these locations for sampling.
-19-
environmental science and engineering, inc.
-------
oooo
Incinerator - not normally
used
For Incinerator Air
Outlet Sample Point
7
0
v
T,
7
0^
V
*• •
,
f
\
\
X:
=•
•v.
J
/
N
s
/
/
/
X
-
_
-
/
\
\
\
^
\
V
\:
*?
J
— S tret "ford Absorber
2'-11 I D Piprt v/i th Shut Off
Valves, 24" In Length
(Check Velocity Here)
c ^/ Prorp*^ fi^s
n Inlet
—^
/^
24" Duct
/ i" Sflnpl c Line
Figure 3. Schematic diagram of sampling locations, Union Oil
of California, Wilmington, California.
-20-
environmental science and engineering, inc.
-------
5.0 ANALYTICAL PROCEDURES
A brief summary of each procedure is presented in this section.
5.1 Sampling Procedures
Sample gases were extracted from the emission source for the deter-
mination of sulfur-containing compounds using a 3/16" FEP Teflon
tubing sample line, heated and maintained at 100°C. This sample
line terminated in the dilution manifold in the EPA mobile laboratory.
Either direct or diluted sample could be withdrawn from the system as
dictated by the analytical range of the instruments. The samples
were simultaneously analyzed for total sulfur, carbon disulfide,
sulfur dioxide, hydrogen sulfide and carbonyl sulfide. FEP Teflon
parts or Teflon coated parts (including the sample pump heads) were
used throughout the system to take advantage of the minimum reactivity
of the Teflon to low level concentrations of sulfur compounds. Figures
4. 5, and 6 show the sample dilution system in the EPA mobile laboratory
the sample handling system for CO, C02 and 02> and a flow system for
the sample from the source to the collection and analysis locations.
Sample gases for the determination of CO, C02 and Oo were obtained
from the EPA sample line and the concentrations of the gases were
determined by instruments located in a truck adjacent to the EPA
mobile laboratory. An integrated bag sample was obtained at the same
source point as the EPA sample and this sample was used for NOX analysis
and Orsat determination of CO, C0? and 02- NOX and S02 were determined
in situ in the mobile laboratories assembled by ESE.
-21-
environmental science and engineering, inc.
-------
Probe —h
Filter
(glass wool)
TO
ro
I
Heating Sample Line-
Stack Wall
Filter
Pump
(Heated)
TO INSTRUMENTS
10:1 102:1 103:1
Permeation
Tube
Calibration
Gas
s/
k
1
i
agm
d)
l
/
/
K
j-(
• v^
^
Ven
/
JVent
)
)
'-v
k
^
t
c
V
H /
^Vent
)
)
J \,
/
S
^
K
r
\.
/
*
%
^Vent Diluent Air
Posi
1 .Disp
L*. Pump
^ (150
P Val
/
\
\
Dilution Box Heated
r To 100°C
t
V
ive
acement
cc/min.)
/ay
e
). Vent
/
v
^
k
/
|
:
6
c
31
: ' i
- •*
is
) C
/
Cl
Dry
1 F
I !
-
> c
[|
.
•
c
a
3
25
\
ean
Air
1350 cc/min
£— Flowmeter
Figure 4.0. Sample Dilution System in
EPA Mobile Laboratory
-------
To Atmosphere (—.O-r- —*>
--— Dilution System
Calibration;: jl j
Gao (07) ; ! !
* V ' I I
I jl
H?0
i
°2
Stan
djrd
.1
i
|
°?
^naly2cr
riH
Standard
Gas (CO)
Silica
Gel
Avcarite
T
tj"Ji"
CO
Standard
NOIR
CO
—- Sample Inlet
J
Atciosphcre <-~-—y
Gas
•Cor.t'31
Figure 5.0. . Schematic of sampling system for C0?, CO, and 02..
-------
Emission Source
ro
-p.
i
Swagelok
Tee
SO,
Heated Teflon Lines
EPA
Van
rs°s
Hr
po
-\!_
ESE
Truck
rco2
) CO
COS ) 02
-CS2 I Lab
1
H2°
Odor
Bag
I
Orsat
NOX
Rnq
Non-Heated Lines
Pump and Dry
Gas Meter
Pump and Flowmeter
Figure G.O. Schematic of sampling systems.
-------
5.2 Analytical Procedures for'Sulfur Compounds
Sulfur compounds were measured by qas chromatography and by wet
•chemical methods. The analytical methods for the various sulfur
compounds are described in the following paragraphs.
5.2.1 Sulfur compounds by Gas Chromatography
Sulfur compounds, when introduced into a hydrogen-rich flame, produce
strong luminescent emissions between 300 and 423 nm. Through the
use of a narrow band optical filter that permits transmission at
394 nrc, a flame photometric detector (FPD) can measure the chemilum-
inescent emissions produced by the $2 species and can differentiate
between sulfur containing and non-sulfur containing compounds.
Through the use of a gas chromatograph (GC) equipped with the
appropriate analytical columns, it is possible to separate and
quantify the various sulfur compounds.
. Applicability of Method
The compounds of interest in emissions from sulfur recovery systems
are hydrogen sulfide (H2S), carbon disulfide (CS2), sulfur dioxide
(S02) and carbonyl sulfide (COS).
The two GC/FPD systems available in the EPA mobile laboratory are
capable of the separation and quantitation of all of the compounds
-25-
environmental science and engineering, inc.
-------
of interest with the exception that COS and F^S could not be determined
simultaneously on any one system due to the relatively small dif-
ference in the retention times and the relatively large differences
in concentration which results in the overlap of peaks. The
difficulty presented by the h^S - COS separation and determination
was overcome through the use of a scrubbing system which effectively
removes one component (I^S) from the sample. Silver wool, which
reacts readily with H?S, was installed in one of the GC systems
between the sample loop and the analytical column. Removal of the
makes possible the determination of COS while the other system
determines FLS + COS. The difference between the two systems
gives the ^S concentration.
Instrumentation and Standards
GC/FPD System - The system provided in the EPA mobile laboratory
was assembled from components available from various commercial
sources.
Sulfur Compound Permeation Tubes - Provided by EPA and gravi-
metrically calibrated by EPA personnel.
Analysis of Samples
The sample gas was extracted from the test source and diluted with
clean, dry, sulfur-free air in the dilution system.. Diluted sample
was continuously flowed through the sample loop and injected at
fifteen minute intervals throughout the test. The fifteen minute
interval was selected due to the retention time of CS2.
-26--
environmcntal science and engineering^ inc.
-------
Responses obtained from each compound were compared to the
standard curve for that component and the concentrations were
determined. In the series of tests conducted, two GC/FPD
systems were utilized. GC#1 was equipped with a scrubbing sys-
tem and was used for a direct determination of COS concentration.
GC#2 was used to determine SOoj CSo, and H^S was determined on
GS#2 by subtracting the COS concentration found with GC#1 from
the ^S + COS concentration. The analytical methods used for
the analysis of sulfur compounds in this series of tests were
the same as described in an EPA preliminary draft method entitled
"Semicontinuous Determination of Malodorus Reduced Sulfur Emis-
• sions From Stationary Sources". Example calculations for the
determination of H^S by difference are given in Appendix A.
5.2.2 Titrimetric Method for the Determination of Sulfur Dioxide
Sulfur dioxide was oxidized to sulfate in the presence of hydrogen
peroxide according to EPA Method 6 as outlined in the Federal
Register, 36, No. 59, Part II, Aguust 17, 1971. The sulfate
which was formed and collected was subsequently titrated with a
standardized solution of barium perchlorate in the presence of
thorin indicator, and the sulfur dioxide concentration was
calculated.
-27-
environmental science and engineering, inc.
-------
5.3 Analytical Procedures for Carbon Monoxide. Carbon Dioxide and Oxygen.
Carbon monoxide, carbon dioxide and oxygen were monitored continuously
from the source during the three four-hour tests. The sample was
obtained as described in section 5.1.
Instrumentation and Standards
Carbon Dioxide - Beckman Model 315A NDIR configured for 0-5% carbon
dioxide.
Carbon Monoxide - Beckman Model 315B NDIR configured for 0-1000 ppm
carbon monoxide.
Oxygen - Beckman Model F-3, paramagnetic oxygen analyzer capable of
measuring 0-25% oxygen.
Standard Gases - Obtained from Matheson Gas Products, Inc., Cucamonga,
California. All standards were either primary or certified and were
analyzed by Matheson Gas Products, Inc. All standards consisted of
the component of interest with the balance of the mixture as nitrogen.
Analysis of Samples
The sample gas was extracted from the test source and flowed through
the instruments. In the case of the carbon dioxide the sample was
diluted with clean, dry, carbon-dioxide free nitrogen in order to
maintain the concentration within the operating range of the instrument.
A schematic of the instrument flow system is presented as Figure 5.
Responses obtained from the instrument were displayed on a strip chart
recorder and these responses were compared to the appropriate standard
curve to obtain the concentrations of the different constitutents.
. -28-
environmental science and engineering, inc.
-------
EPA Method 10 (Federal Register,39, No. 47, March 8, 1974) was used
as a guideline in the determination of carbon monoxide concentrations.
5.4 Analytical Procedure for Nitrogen Oxides
Nitrogen oxides were measured according to EPA Method 7 (Federal Register.
39, No. 47, March 8, 1974). A portion'of the contents of the integrated
bag sample was collected in an evacuated flask which contained sulfuric
acid and hydrogen peroxide. After the oxides of nitrogen had been
oxidized to nitrate, the nitrate was reacted with phenoldisulfonic
. acid and a spectrophotometric method was used to determine concentration.
5.5 Procedure for the Determination of Moisture
The moisture content was measured in accordance with EPA Method 4
which appeared in Federal Register. 36_, No. 59, Part II, August 17,
1974.
5.6 Procedure for the Determination of Stack Gas Velocity
The stack gas velocity was determined after sample and velocity
traverse points were located. EPA Method 1 and 2 as they appeared
in Federal Register, 36_, No. 59, Part II, Aguust 17, 1974 were used
for these determinations.
5.7 Procedure for the Determination of Visible Emissions
EPA Method 9, as outlined in the Federal Register. 36, No. 247, Part II,
-29-
environmental science and engineering, inc.
-------
December 23, 1971 was used as the guideline in the determination of
visible emissions. A certified observer was used to observe emissions
from the sample source.
5.8 Carbon Dioxide, Oxygen and Carbon Monoxide by Orsat
Orsat determinations were made on the integrated bag sample in accor-
dance with EPA Method 3 which appeared in Federal Register, 36, No.
247, Part II, December 23, 1971.
5.9 Analytical Method for Odor Emissions
The Determination of Odor Potential from Stationary'Sources (Dilution
Method), an EPA draft method, was used as a guideline for the odor
emissions portion of the report. This method is based on the fact
that the human olfactory sense is very perceptive to trace quantities
of odorous compounds and that the selection of the odor panel,
methods of dilution of sample gas, etc., odor potentials may be
determined.
5.9.1 Selection of the Odor Panel
Students from a local high school were selected by conducting a
screening test as outlined in Figure 7.0. In the screening test the
potential panel members were required to identify the odd sample in
each set of 3. Samole concentrations ranged from 1% to .001% of
vanilla extract and methyl sal icy!ate in benyl benzoate.
-30-
environmental science and engineering, inc.
-------
Collection and Analysis of Samples
The samples for odor analysis were collected in the apparatus shown
in Figure 8.0.
Analysis of Samples
The samples were collected at each site as prescribed by the project
officer. The teflon bags containing the samples were transported to
the high school by the personnel responsible for the analysis. The
samples were diluted with clean, dry, odor-free air and analyzed by
the panel on a detectable, non-detectable basis. Blank sample (clean,
dry, odor-free air) were given to the panel periodically to insure the
integrity of the odor panel procedure.
The data obtained were plotted on log probability paper and the best
line through the plotted data was determined by the method of least
squares. The odor concentration for each sample was used to calculate
the odor emission rate according to the equation E = CVA, where:
E = odor emission rate, odor units/minute
C = odor concentration, odor units/SCF
V = velocity of source, feet/minute
and A = cross-sectional area of the stack, square feet.
-31-
environmental science and engineering, inc.
-------
FiGu:-.;;_7.o_
Layout of Scrocvn'm Test
V=--Vanill;i Extr.ict. /•;=-- Kr.-lhyl Solkyalc
I3li= Denzyl Hcnzontc
l.OS
1.0%
0.5%
M ) (V ; (M
0.12
o.cns
0.001S
7
l
o.ooi;>
-32-
-------
CO
CO
I
Vacuum
Pressure Pump
Rubber Gasket
Teflon Bag
Rigid Container
DI il«i% r»r» T« i*^ i ^ ~
t\ u w J «. I 1 M ^/ 11. ^j
Stainless Stsc-1
Swags Lck
Heated Tcr.cn :
Line to Probe
On-Off Valve
Sampling Bag Assembly
Figure 8.0
------- |