EPA-600/4-77-008b
January 1977
Environmental Monitoring Series
STANDARDIZATION OF METHOD 11 AT A
PETROLEUM REFINERY: Volume II
Environmental Monitoring and Support Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park,. North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL MONITORING series.
This series describes research conducted to develop new or improved methods
and instrumentation for the identification and quantification of environmental
pollutants at the lowest conceivably significant concentrations. It also includes
studies to determine the ambient concentrations of pollutants in the environment
and/or the variance of pollutants as a function of time or meteorological factors.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
STANDARDIZATION OF METHOD 11
AT A PETROLEUM REFINERY
VOLUME II
by
George W. Scheil
Michael C. Sharp
Midwest Research Institute
Kansas City, Missouri 64110
EPA Contract No. 68-02-1098, Task 8
EPA Project Officer
M. Rodney Midgett
Quality Assurance Branch
Environmental Monitoring and Support Laboratory
Research Triangle Park
North Carolina 27711
Environmental Monitoring and Support Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park
North Carolina 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Monitoring and
Support Laboratory, U.S. Environmental Protection Agency, and approved for
publication. Approval does not signify that the contents necessarily re-
flect the views and policies of the U.S. Environmental Protection Agency,
nor does mention of trade names or commercial products constitute endorse-
ment of recommendation for use.
-------
FOREWORD
Midwest Research Institute (MRl) under Task 6 of EPA Contract No. 68-
02-1908, conducted in-house work toward the standardization of Method 11
(Federal Register. Vol. 39, pp. 9321-9323, March 8, 1974) as applied to the
analysis of H S in petroleum refinery fuel gases. The results of this in-
house work are given in Volume I of this report.
Further work on the method was done under Task 8 of EPA Contract No.
68-02-1908. This continuation of efforts was a collaborative test to evalu-
ate the accuracy and precision of the modified Method 11. The results of
the collaborative test are given in this report. The collaborators' data
sheets are in Volume III.
Approved for:
MIDWEST RESEARCH INSTITUTE
<\
L. J. Snannon, Director
Environmental and Materials
Sciences Division
April 4, 1977
111
-------
ABSTRACT
A collaborative test was run of the revised Method 11 procedure that
was developed and is reported in Volume I of this report. Ten collaborators
were selected from a total of 24 interested organizations. Part of the
screening process was to require each potential collaborator to analyze a
set of liquid samples in accordance with the procedure.
A test manifold was constructed which could generate controlled, simu-
lated refinery fuel gas streams using spiked natural gas. The major constit-
uent of the stream was methane
-------
CONTENTS
Foreword iii
Abstract . iv
Figures vi
Tables viii
Acknowledgments ix
1. Introduction 1
2. Conclusions 3
3. Recommendations 4
4. Selection of Collaborators 5
5. Hydrogen Sulfide Sampling System 9
6. Experimental Design „ 18
7. Collaborative Test 25
8. Results of Analyses ..... 34
9. Statistical Analysis of Collaborators' Results 39
Appendices
A. Collaborator Selection Documents 53
B. Instructions to the Collaborators. 85
v
-------
FIGURES
Number
1 Block Diagram of H~S Sampling System 10
2 Close-Up View of the Spike Gas Cylinders and
Metering Equipment 12
3 View of the Spike Gas Assembly 12
4 Schematic Drawing and Photographs of the Diffuser ... 13
5 Schematic Drawing and Photographs of the Sampling
Manifold 14
6 Burner Assembly for Combustion of Excess Gas 16
7 Sampling Table and Exhaust Hood 17
8 Collaborators Sampling Trains 28
9 Collaborators Analyzing Samples 28
10 Collaborator Analyzing Test Samples 29
11 Participating Collaborators and Midwest Research
Institute Staff, Hydrogen Sulfide Collaborative Test,
February 23 through 27, 1976, MRI Field Station,
Grandview, Missouri 30
12 Sampling and Analysis Data Sheet 32
13 Data Sheet for Standardizations 33
VI
-------
FIGURES (concluded)
Number Page
14 Bias Results by Collaborator, ,H S Standards Data .... 43
15 Bias Results by Collaborator, H S Field Data 47
16 Theoretical Calibration Curves, H S Liquid Samples ... 50
A-l Sampling System 56
A-2 H S Sampling Train 71
VII
-------
TABLES
Number
1 Collaborator Mean Square Errors 7
2 Analysis of Variance Model for Liquid Samples 22
3 Analysis of Variance for Calibration Gas Samples. ... 23
4 Analysis of Variance for Field Samples 24
5 Collaborative Test Log 26
6 H«S Concentration (mg/dscm) Determined by Collaborators
for Standard Liquid and Cylinder Samples 35
7 H2S Concentration (mg/dscm) Determined by Collaborators
at Test Levels A, B, C 36
8 Data Summary of MRI Measurements at the H2S Measure
Concentration During the Manifold Test 37
9 Analysis of Variance of H S Standards Data 40
10 Components of Variance, H2S Standards Data 41
11 Bias and MSB, H2S Standards Data 42
12 Analysis of Variance H2S Field Data 45
13 Components of Variance, H2S Field Data 45
14 Bias and MSE, H2S Field Data 46
15 Liquid Standard H2S Results by Collaborator 49
16 Summary of Analysis Results 51
viii
-------
ACKNOWLEDGMENTS
Task 8 was conducted under the technical management of Mr. Paul C.
Constant, Jr., Head, Environmental Measurements Section of MRI's Environ-
mental and Materials Sciences Division, who is the program manager.
Dr. George Scheil was task leader. He was assisted by Messrs. John LaShelle,
Bruce DaRos, Thomas Merrifield, and Michael Sharp of MRI. Texas Analytical
Control, Inc., of Houston, Texas, provided the solid-state H^S monitor used
during the collaborative test.
IX
-------
SECTION 1
INTRODUCTION
This report covers the tests and evaluations conducted by Midwest
Research Institute (MRI) of proposed methods to replace the present Method
11 procedure (Federal Register, Vol. 39, pp. 9321-9323, March 8, 1974). The
testing is being done for the Environmental Protection Agency (EPA) under
Contract No. 68-02-1098, "Standardization of Stationary Source Emission Mea-
surement Methods." This report covers Tasks 6 and 8 of the contract. The
first task covered the search for a method which is free of the mercaptan
interference, its evaluation, and field testing of the final method. The
work done on Task 6 is in Volume I of this report. Task 8 included a col-
laborative test of the proposed, new Method 11 and is covered in this vol-
ume (II) of the report.
A collaborative test is a procedure in which a group of persons from
different laboratories conduct sampling and analysis under identical condi-
tions using the same method. It provides information on the variability of
method results between laboratories as well as the reproducibility of a sin-
gle analyst's results. A properly designed collaborative test should demon-
strate the reliability of the method being tested under typical, realistic
sampling and analysis conditions.
Task 8 began in December 1975 with the selection of collaborators.
A total of 24 organizations received a Request for Proposal (RFP) which out-
lined the duties of each collaborator. As a part of the selection process
each potential collaborator was required to analyze a set of liquid samples
of pure cadmium sulfide suspended in absorbing solution. After eliminating
those who made serious analytical errors or who failed to meet the require-
ments of the RFP, the 10 lowest bidders were accepted as collaborators. Con-
struction and testing of the sampling manifold for the test proceeded con-
currently with the collaborator selection. The collaborative test took
place during February 22 to 27, 1976, at MRl's Deramus Field Station in
Grandview, Missouri. As part of the test each collaborator was required
to analyze a series of synthetic liquid samples, a set of three reference
cylinder gases and to sample simulated fuel gases from a sampling manifold.
After allowing the collaborators to check their results and report changes,
the results were statistically analyzed by MRI.
-------
The following sections of this report cover the selection of collabora-
tors, a description of the sampling system used in the collaborative test,
the experimental design of the test, a discussion of the collaborative test,
the collaborators' analysis results, a statistical analysis of the test re-
sults and finally the general conclusions and recommendations from the study,
Volume III of this report contains only copies of the original collabo-
rator data sheets.
-------
SECTION 2
CONCLUSIONS
The major conclusions that can be drawn from the results of this col-
laborative test are:
1. The "Tentative Method for the Determination of Hydrogen Sulfide
Emissions from Stationary Sources," as shown in Appendix A of this report,
is adequately written for those knowledgeable of sampling and analysis
techniques.
2« If the tentative procedure is followed by people knowledgeable of
the sampling and analysis techniques given therein, then in the absence of
mercaptans the results will be on the average 4% low, + 47°,* over the range
100 to 400 mg/dscm. If a set of such people, each sampling independently,
follow the method, then results will be on the average 4% low, + 870.*
3. When methyl and ethyl mercaptans are present in addition to H S
at a total level of 150 ppm, the results vary from 2% low at an H S concen-
tration of 400 mg/dscm to 14% high at an H S concentration of 100 mg/dscm.
The presence of mercaptans has no significant effect on precision.
* Precision is for the 95% confidence limit (2 a).
-------
SECTION 3
RECOMMENDATIONS
Based upon the conclusions that have been drawn from the results of
this collaborative test, it is recommended that the tentative method be
used as written as a replacement for the original Method 11 with the addi-
tion of the information given in Conclusion 3.
-------
SECTION 4
SELECTION OF COLLABORATORS
To select participants for the collaborative test, MRI sent letters
to more than 60 organizations. Those organizations that expressed interest
in the test then received an RFP and a set of synthetic liquid samples.
The samples had to be titrated according to the proposed method and the re-
sults were used in selecting competent collaborators.
Twenty-one organizations responded in the affirmative to the letter
seeking interested organizations. Three additional organizations were added
to the list, giving a total of 24. A copy of the proposed RFP was submitted
to the project officer for approval. After making the suggested changes, the
RFP was sent to the 24 organizations on December 22, 1975. A copy of the
RFP is included in Appendix A.
The samples shipped to the potential collaborators contained 1 to 28
mg of cadmium sulfide suspended in 45 ml of absorbing solution (equivalent
to 25 to 700 mg/dscm assuming a 10 liter sample volume).
The synthetic samples were tested before shipment. When the samples
were analyzed by the normal procedure, low recovery efficiencies were found--
approximately 60%. However, by making two simple changes in the analysis
procedure, 100% recovery could be obtained. First, the samples had to be
uniformly suspended since otherwise numerous unreacted particles could be
seen in the flask after titration. Vigorous shaking by hand prior to sample
recovery was only partially successful, but placing the sample bottles in
an ultrasonic bath for 10 min broke up all the particles. The second change
was that 30 sec of vigorous shaking was needed after the acidified iodine
had been added to the sample in the iodine flask. Since 30 sec of shaking
was one of the noncritical variables checked in the ruggedness test, making
these changes for the liquid samples should not affect the experimental de-
sign.
-------
To check the stability of the liquid samples, one set was prepared
and analyzed after standing at room temperature for 1- and 2-week intervals.
A second set was placed outdoors and analyzed in the same way. After 1 week,
no change in value could be measured. After 2 weeks, recoveries of 103 to
105% were obtained. The higher values were found in the outdoor samples which
were frozen solid when brought indoors for analysis. None of the frozen sam-
ples were broken or otherwise damaged. Therefore, it was assumed that the
samples would survive shipment and would give predictable values if analyzed
within 2 weeks of preparation.
Most of the liquid samples were shipped on December 31, except for
three sets sent after New Year's Day since three companies requested delayed
shipments due to vacations. The shipments contained instructions for the
modified analysis (see Appendix A) and each organization was required to
report the results of any four of the six samples shipped. This allowed for
losses in handling or for practice purposes.
A total of 22 organizations responded to MRI's RFP for the collabora-
tive test.
One collaborator withdrew before evaluation and one submitted incom-
plete data. Of the 20 remaining collaborators, five failed to execute the
method without gross errors.
After these five were eliminated, the remaining 15 collaborators were
rank-ordered via their mean square errors* (see Table 1). It is not the pur-
pose of collaborator selection to construct an optimum group, but rather to
construct a representative group of collaborators for the field test. For
this reason, the only collaborators eliminated from Table 1 were Cll and
C13, because these two are outliers (Dixon-Massey test, r = 0.517, or <
0.10). 22
Ten of the remaining 13 collaborators were selected purely on the basis
of cost. These are identified in Table 1 by a line under their number. The
10 collaborators are:
The Research Corporation of New England
125 Silas Deane Highway
Wethersfield, Connecticut 06109
Monsanto Research Corporation
Dayton Laboratory
1515 Nicholas Road
Dayton, Ohio 45418
* The mean square error (MSB) is a standard statistical descriptor of
"total uncertainty," namely, MSB = (bias) 2 + (variance).
-------
TABLE 1. COLLABORATOR MEAN SQUARE ERRORS
Collaborator MSE (% of true value)
2 7.4
^5
6
jj
12
18
19
11
j.
3
16
.17
I4-
11
13
8.1
9.2
9.4
9.6
10.6
11.2
11.2
12.5
12.6
14.1
16.0
19.0
26.0
27.1
Clayton Environmental Consultants
25711 Southfield Road
Southfield, Michigan 48075
York Research Corporation
One Research Drive
Stanford, Connecticut 06906
Washington State University
College of Engineering
Pullman, Washington 99163
California Department of Health
Health and Welfare Agency
Air and Industrial Hygiene Laboratory
2151 Berkeley Way
Berkeley, California 94704
Interpoll, Inc.
1996 West County Road C
St. Paul, Minnesota 55113
Roy F. Weston, Inc.
Weston Way
Westchester, Pennsylvania 19380
-------
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, North Carolina 27709
Ecology Audits
11061 Shady Trail
Dallas, Texas 75229
-------
SECTION 5
HYDROGEN SULFIDE SAMPLING SYSTEM
The system shown in Figure 1 was used to generate known levels of HLS
in natural gas while also controlling the levels of the interferences ana
other compounds normally found in refinery fuel gas streams. The test assem-
bly was constructed at MRI's Deramus Field Station in Grandview, Missouri.
The test facility provided ample room and electrical power for the test and
included a large exhaust hood over the sampling manifold to avoid buildup
of noxious fumes.
The equipment specifications for the sampling system are as follows:
1. Gas regulator - A standard natural gas regulator was used to drop
the 10 psig main gas supply pressure down to a stable 5 psig pressure for
the system.
2. Dry gas meter - A Singer Model AL-425 dry gas meter measured the
natural gas flow rate. The meter had a large observation index for greater
reading accuracy. The meter was calibrated against a Precision Scientific
1 ft /rev. wet test meter.
3. Thermometer - The gas stream thermometer was a 0 to 50°C bimetallic
thermometer mounted in the gas meter outlet with the entire stem in the gas
flow.
4. Precision pressure gauge - Stainless steel 0 to 15 psig, Matheson
Gas Products No. 63-5615, 0.25% accuracy, readable to 0.05 psig (2.5 ram Hg).
Checked against a mercury manometer at working pressure.
5. Gases, regulators - All gases and regulators were obtained from
Matheson Gas Products except for nitrogen which was prepurified grade, Union
Carbide. Shutoff valves were those supplied with the regulators and lecture
bottle controls.
6. Filters - Stainless steel, 7 \un pores, Type 4FR, Nupro Company.
The filters prevented particulates from reaching the rotameters.
-------
r
~i
r
HjS Cylinder
*
Corrosion Res
Single Stage
M 1 Shut Off
from Nj 7 Mi'ron F;l
Micrometer xjx K?i— 4 Doubl«
Valve !&/ K/ ' Micron»,
^"' Rotate,
/ . . ((MOOcc/rr
cc/ ntin f
Regulator Safety Vent _ . .
toAtmosphere Pr.c,»on
. Pressure Gauge
i Gas Cock s-**.
" . Rj . . i . A *A -
Gos Regulator Dry Gas /"K
(5 pig ) Meter W
(Appro* . Thermometer
53-100
L./min) i
Methyl Mercopton
(Lecture Bottle)
COS (Lecture Bottle) NJilos^n'cJiind.r SO2 C/linoer Propylene Cylinder
t t r— *— -i
istant 1 1
Re Ma tor -a i :
« I
Two Stage High Corrosion Resistant „ ! ' Single Stage
Pressure Regulator Single Stage Regulator Regulator
1, JL ,0HS $i 1 Shut Off Volve Jf (9) (-Shut Off Volv.
Volve JX) H Shut Off Volve (V) 1 Shut Off Volve Un«2""TL !$J Shut Off Volve Jf j
jf ! JT 60 1 Micrometer Vo ve J[ (V) L Micrometer Valve
! t ' T ' t i T
ef 7 Micron Filter
_ ... ... Rotometer Glass Capillary
7M,cron F.lter (0-700cc/min) -I0»ce/m.n
. , KO 1 1 Micrometer Valve 6O 1 Micrometer Volve 1 '
er Valve "V" V* j
I * Bubbler ' j
Rotometer
in) j(0-IOOcc/min)
Rotometer Contai ni ng j Rotometer
(0-10X)cc/min) Ethyl Mercapton i (Q-4 l/min)
1 , 1 ( r Ice Bath | J
'
^.'-^^ Teflon TfZZ— "
t"^\ Sampling t^. — ,
1
J
r\ .
L }
Ftow Restrictor Flame Out Burner and
Control Valve Safety Valve Flue Stack
Figure 1 - Block Diagram of H S Sampling System
-------
7. Rotameters - Supplied by Matheson Gas Products with stainless steel
housing. Nitrogen rotameters use 602 tubes, propylene uses a 603 tube, H S
uses a 601 tube, and methyl mercaptan and carbonyl sulfide each use a 610
tube with a micrometer valve built into the housing.
8. Micrometer valves - All are standard Whitey valves, except that
for H S which is a Nupro SS-2MGD double micrometer valve for very low flow
rates.
9. Glass capillary - The SC>2 for flow control was a short section of
dropping mercury electrode capillary.
10. Ethyl mercaptan bubbler - The bubbler used to saturate the N
carrier with ethyl mercaptan was a 200 mm glass vacuum trap with side arm
outlet. In operation the trap was filled with mercaptan outdoors, capped
and brought inside and connected to the system. The rate of mercaptan addi-
tion was calculated from the partial pressure assuming a temperature of 2 C
in an ice bath:
1 ("1Q/, CT —
log 10 P (mm Hg) = 6.95206 - ^j^'(°C)
Figures 2 and 3 show the gas spiking assembly mounted in the fume
hood.
11. Diffuser - The diffuser was the same as that used at the labora-
tory test facility (Figure 4) except that the end plates were replaced with
stainless steel plates with 0-ring seals and 1 in. O.D. stainless steel
tubes welded to the end plates. This change was necessary to eliminate leaks
in the assembly. The inlet tube to the diffuser was drilled and tapped for
1/4 in. NPT (National Pipe Thread) to accept the lines from the gas measur-
ing system.
12. Sampling manifold - The sampling manifold is shown in Figure 5.
To obtain gastight seals in the manifold, the original end plates were re-
moved and 0-ring sealed stainless steel plates with tubing similar to those
used for the diffuser were installed. An additional 0-ring was used between
the two Teflon subassemblies, and the ports were drilled out to accept 1/8
in. Teflon tubing which was expoxied in place for all ports. Stainless steel
Swaglok caps were used to seal off the ends of all unused ports. The diffuser
and sampling manifold were connected by a 2 m equilibration section of 3 cm
O.D. Teflon pipe.
If Lange, Handbook of Chemistry. 10th ed., p. 1429 (1961).
11
-------
Figure 2 - Close-up View of the Spike Gas Cylinders
and Metering Equipment
Figure 3 - View of the Spike Gas Assembly
L2
-------
Photo 1 - Top View Diffuser Components: Housing, End
Sections, Spiraler Tube, Teflon Screens, Retaining
Rings.
Photo 2 - External View of Diffuser.
T/
7L.
7ZL
Exploded cross section of all-Teflon diffuser with inlet (A),
end section (B), spiraler tube with angled holes (C),
prescreen block with holes (D), five sets of fine mesh
Teflon screen and retaining blocks (E), end section (F),
exit (G) and diffuser housing (H). Double cross-hatched
end plates are stainless steel.
Jffl!)
V/////////////////////////////S
Figure 4 - Schematic Drawing and Photographs of the Diffuser
-------
Photo 1 - Sampling Manifold External View. The bottom
section was not used for this test and the end plates were
modified as described In the text.
:"
Photo 2 - Internal View (Right Component is Inverted
in this Photo).
rrn
rrn
rm
Cross section of all-TFE Teflon manifold with inlet (A), mixing impeller (B), main chamber flow
spreader (C), exits to collaborator ports (D), channel to exhaust manifold (E), double cross-
hatched assembly plates at top and bottom are stainless steel.
Figure 5 - Schematic Drawing and Photographs of the Sampling Manifold
-------
13. Control valve - A 1/2 in. NPT multiturn valve was used to control
the gas flow rate. Since the burner had a regulator to drop pressure to
about 10 cm water pressure and with little flow resistance between the mani-
fold and the main supply regulator, this valve provided the primary flow re-
striction in the system. Thus, the sampling manifold remained very near the
5 psig supply pressure.
14. A Cox 250,000 Btu natural gas burner was installed outside the
building with a 30 cm diameter stack approximately 4 m tall (see Figure 6).
The burner orifice was increased to the maximum and a 1/4 in. tube was run
to the center at the flame to augment the gas flow. A pressure sensor with
alarm was connected to the burner such that if a flame-out occurred shutting
down the burner, then the alarm would sound.
All gas cylinders except for nitrogen and propylene were installed in
an exhaust hood together with the controls and rotameters. Two short sec-
tions of stainless steel tubing connected the metered gases to the diffuser
inlet tubing.
The sampling manifold was mounted on a 3 m diameter circular table with
the sampling ports fastened at the edge of a raised, center platform. An ex-
haust hood was then suspended directly above the manifold. The lower edge
of the hood had a fringed rubber curtain which extended below the sampling
ports yet allowed access to the ports when necessary. The sampling table
and hood are shown in Figure 7.
The test manifold can supply H S levels of 50 to 500 mg/dscm for up
to 30 separate devices. A total of 20 liters/min can be withdrawn from the
manifold at a pressure of 5 psig without upsetting the system.
The entire system was leak tested above normal working pressure. Sev-
eral of the ports in the sampling manifold leaked and were plugged from the
inside of the manifold. Each remaining port was then tested for flow restric-
tions by measuring the maximum flow through each line. About 30 of the 45
ports passed both checks and these 30 ports were used for the collaborative
test.
A solid-state H S monitor was then used to check each port for consis-
tent FLS levels. The spike assembly was set for approximately 300 mg/dscm of
H S. All ports showed identical H S levels + 5%. With the sampling system
set to 151 mg/dscm of H S, duplicate sample trains gave 135 and 136 mg/dscm
of H S. At 288 mg/dscm, duplicate trains gave 328 and 292 mg/dscm; at 387
mg/dscm duplicate trains gave 405 and 414 mg/dscm.
15
-------
Figure 6 - Burner Assembly for Combustion of Excess Gas
L6
-------
Figure 7 - Sampling Table and Exhaust Hood
I /
-------
SECTION 6
EXPERIMENTAL DESIGN
GENERAL CONSIDERATIONS
The major considerations that entered into the experimental design of
the collaborative test were:
Analysis Method
All collaborators shall follow the revised cadmium sulfate procedure,
copies of which were provided with the RFP's for the test. MRI would check
equipment specifications, make calibration checks on all dry gas meters,
and any other checks necessary to ensure that the procedure was being fol-
lowed.
Location of Test
The test will take place at MRI's field station which is located approx-
imately 15 miles south of MRI in a suburban area of Kansas City.
Sampling Time
All sampling times shall be 15 min. For the reference gases, no special
order of analyses is required. For the 3 days of sampling from the test mani-
fold, all collaborators shall sample simultaneously. Duplicate trains will
be used for all sampling.
Number of. Collaborators
In making all preparations for the test, it was assumed that 10 quali-
fied collaborators would participate in the test.
Test Levels
There would be three challenge levels. Each test manifold level would
comprise H S, natural gas and possible interferences. The three levels of
challenge would be 100, 200, and 400 mg/m of H S.
18
-------
Standard Samples
There would be two types of standard samples given to each collaborator
at the field site: eight liquid samples (two each at Levels 1, 2 and 3, one
sample near zero, and one blank) and six gas samples (two each at Levels 1,
2 and 3). One set of standard samples would be given to each collaborator
before the 3-day test starts to be analyzed on-site. A similar set would
be given to each collaborator after completion of the 3-day test to be ana-
lyzed on-site.
Test Schedule
The field work would be executed as follows:
Day Activity
Sunday Collaborators travel to Kansas City. Site
will be available to collaborators Sunday
afternoon for unpacking equipment, set up,
etc.
Monday Orientation. Equipment set up. Calibration
of collaborator's equipment. Analysis of
standard samples.
Tuesday Sampling at Level 1. Analysis of samples.
Wednesday Sampling at Level 2. Analysis of samples.
Thursday Sampling at Level 3. Analysis of samples.
Friday Analysis of second set of liquid and gas
samples. Packing of equipment. Turn in of
field test data and results of analysis.
Saturday Complete packing and return home.
Every effort would be made to have the collaborators ship their equip-
ment and materials so that the equipment and material would arrive at MRI
a week prior to Day 1 of the field test. If the collaborator is driving and
is bringing his equipment with him, this advance delivery would not be nec-
essary.
MRI would assist the collaborators with the shipment of their equipment
back to their home organization to expedite site-clearing operations.
19
-------
True Value of H S
A set of three cylinders of H S in methane would provide the standard
gas samples. Each cylinder would be carefully analyzed by the supplier be-
fore shipment and again upon return of the cylinders after the test. MRI
would also check the reported levels prior to the test using the proposed
test method.
The liquid standard samples were prepared by suspending accurately
weighed amounts of pure cadmium sulfide in the absorbing reagent. A random
selection of the samples were analyzed by MRI prior to the test.
The primary calibration of the H S levels in the manifold was the cal-
culated dilution factor of pure H S added to the natural gas stream. The
H9S rotameter was calibrated with H S at different pressures using a soap
film bubblemeter. The natural gas flow was measured with a dry gas meter
calibrated against a wet test meter. The pressure and temperature were also
measured during operation which allowed the H^S spike level to be calculated
using the gas laws and empirically measured rotameter temperature and pres-
sure corrections. Further checks were made using a potentiometric titration
with standard silver nitrate and flame photometric gas chromatography anal-
yses. A commercial solid-state H S sensor was also used to monitor the in-
tegrity of the spike but was not used as a primary calibration method.
Simulation of a Typical Refinery Fuel Gas Stream in the Test Manifold
Refinery fuel gas streams can vary widely in composition. However, they
normally consist of hydrogen, methane, and/or natural gas with smaller amounts
of H S, SO , mercaptans, C -C unsaturates, COS, amines used in scrubbers
and water present in varying proportions. For this test commercial natural
gas was used as the primary constituent. The natural gas in Kansas City nor-
mally is methane with < 5% ethane, and small amounts of nitrogen, argon, and
higher hydrocarbons. Constant amounts of propylene (~ 4%), SO (*• 50 ppm),
methyl mercaptan (•* 100 ppm) and ethyl mercaptan (** 50 ppm) were added during
all tests to generate a stream which would be more difficult to analyze than
the majority of real samples.
FORMAL DESIGN
The H S collaborative test can be considered as furnishing three data
sets:
1. Analysis of known liquid samples to estimate purely analytical
errors.
2. Analysis of known calibration gas samples (without interferences)
to estimate sampling and analysis errors.
20
-------
3. Analysis of known field samples with realistic interferences pres-
ent in the gas stream.
In all cases, three H S levels of approximately 100, 200, and 400 mg/
dscm will be used. In addition, blanks and very low level samples will be
included in the liquid samples. At all times all 10 collaborators will be
observing the same stream so that direct estimates of accuracy and preci-
sion are possible.
The three analyses of variance models are presented in Tables 2 through
4.
21
-------
TABLE 2. ANALYSIS OF VARIANCE MODEL FOR LIQUID SAMPLES
X. - P+ C. + L.. + CL.. + e , where
ijk i ij ij (ij)K
Vi = overall mean response,
Ci = i collaborator (i = 1, ..., 10),
LJ = jth H2S level (j = 1, ..., 5 including a blank level),
CL. . = collaborator-level interaction,
error in ijth cell (k = 1,2 replicates for each cell)*,
X. = ijk response, i.e., ijk difference between observed
ijk
and true value.
Source
Collaborator (C)
Level (L)
C x L
Error (e)
d.f. EMS
9 o-e2 + 8 oc2
4 ffe2 + 2 a(,L2 + 2Q ^2
27 CT 2 -f 2 arT 2
40 cre2
* The first replicate before and the second replicate after the field
test.
22
-------
TABLE 3. ANALYSIS OF VARIANCE FOR CALIBRATION GAS SAMPLES
Xijkl= " + Ci + Lj + CLij
LTjk(k)
u = overall mean response,
th
CJL = i collaborator (i = 1, ..., 10),
LJ = jth H2S level (j = 1, ..., 3),
CL.. = collaborator-level interaction,
T, /-\= k train within i collaborator
(k = 1, 2 trains per collaborator),
LT
i\ ~ level-train interaction,
:(ijk)l
= t
error in
*1
ith
cell (1 = 1, 2 replicates per cell),
th
X = ijklcn response, i.e., ijklcn difference between observed H0S
, , 2
and true value.
Source
Collaborator (C)
Level (L)
C x L
Train (T)
L x T
Error (e)
d.f.
EMS
9
2
18
10
20
60
o-
ae
CTe
CT
e
CTe
CTe
2 .
2 .
2
2 -
2 .
2
f- 6 OVr
f 4 CTC
f- 2 aL
•• 6 am:
T
1- 2 aL,
+12 a
+ 4 CTCL
40
23
-------
TABLE 4. ANALYSIS OF VARIANCE FOR FIELD SAMPLES
Xijkl ' * + Ct + Lj + CL-. + Tk(i) + LTjk(1) + e(..k)1 , where
u = overall mean response,
Ci = 1th collaborator (i = 1, ..., 10),
L - jth H2S level (j = 1, ..., 3),
CL... = collaborator-level interaction,
\(i) = k train within i collaborator (k = 1, 2 trains per
collaborator) ,
LTjk(i) = level-train interaction,
ijk)l = 1 error in ijk cell (1=1, ..,5 replicates per cell)
e
= iJk1 response, i.e., ijkl difference between observed
H S and true value.
Source d.f. EMS
Collaborator (C) 9 o-e2 + 15 or2+ 30 ac2
Level 2 ae2 + 10 o-CL2+ 5 aLT2+ 100
C x L 18 ae2 + 5 o-LT2+ 10 aCL2
Train (T) 10 ae2 + 15 c^2
L x T 20 cre2 +
Error (e) 240
-------
SECTION 7
COLLABORATIVE TEST
The collaborative test was conducted at MRI's Deramus Field Station
on February 22 through 27, 1976. Instructions for the test were sent to the
collaborators prior to the test (see Appendix B). A log of the test is given
in Table 5. Photographs taken during the tests are shown in Figures 8, 9,
and 10. Figure 11 is a group photo of the test participants.
Most of the collaborators arrived at the site on February 22 to set
up their equipment. All collaborators were present by 0800 on February 23
and following the orientation meeting proceeded to analyze the three gas
cylinders and a set of five liquid samples. The liquid samples included one
,blank and one very low level CdS sample in addition to the three levels of
~ 100, 200, and 400 mg/dscm assuming a 10-liter sample volume. Except for
the two collaborators who did not come on Sunday to set up, everyone finished
by 5 p.m. The dry gas meters of half the collaborators were checked on
February 23 and the remaining meters were checked by MRI against a 3 liter/
revolution wet test meter on February 24. Five runs at a level of ~ 200 mg/
dscm H S were made on February 24. Five runs at ~ 400 mg/dscm were made on
February 25 and six runs at a level of 100 mg/dscm were made on February 26.
On all 3 days a constant interference level was maintained of ** 3% propy-
lene, 100 ppm methyl mercaptan, 50 ppm ethyl mercaptan and 50 ppm SO . For
the sixth run on February 26, 100 ppm carbonyl sulfide was also added. All
collaborators finished testing by 5 p.m. on all 3 days of manifold testing.
On February 27, a second set of liquid samples was analyzed and the three
cylinders retested. All results were submitted and equipment packed by 3
p.m. on February 27.
No significant deviations from the proposed method were observed during
the test, although several collaborators did have varying degrees of diffi-
culty in getting the sampling trains to operate properly. It was discovered
during troubleshooting that some trains can require more than 25 cm of water
pressure to simply overcome the water column resistance of the impingers.
Increasing the test pressure should eliminate this problem. Several procedu-
ral errors were noted in the test log. Most of the errors were obvious and
probably caused by lack of familiarity with the method, which was unavoid-
able since even the original Method 11 procedure has been used infrequently.
25
-------
TABLE 5. COLLABORATIVE TEST LOG
Date
2-23
2-23
2-24
2-24
2-24
2-24
2-24
2-24
2-25
Dry test meter
calibrations check
Standard samples
Level A Run 1
Level A Run 2
Level A Run 3
Level A Run 4
Level A Run 5
Dry test meter
calibration check
Absorbing solution
pH test
Comments
Meters for six of the collaborators were tested
Some collaborators had trouble making connec-
tions to the reference gas cylinders. A few
bottlenecks occurred when more collaborators
wanted to sample a cylinder than there were
ports available. The liquid samples were
issued with no problems. A few samples were
spoiled during analysis and replacements
were issued. Several problems passing leak
checks. One train backed up during sampling.
Start at 0905, stop at 0920, System stable.
Two collaborators (3,7) discovered leaks
during the test-only one (3) was serious.
Collaborator 8 pulled cadmium solution
into the peroxide on both trains.
Start at 1005, Stop at 1020, System stable.
Collaborator 8 again pulled over the cad-
mium solution. Collaborator 7 had a brief
leak.
Start at 1105, stop at 1120, System stable.
Collaborator 6 forgot the charcoal filters
during purge.
Start at 1345, stop at 1400, System stable.
Collaborator 6 had a leak in one train for
about 1 min. Collaborator 3 had a flow loss
in one train.
Start at 1445, stop at 1500, System stable.
Meters for the remaining four collaborators
were checked.
pH's were between 2.80 and 3.10.
26
-------
TABLE 5 (Concluded)
Date
2-25
2-25
2-25
2-25
2-25
2-26
2-26
2-26
2-26
2-26
2-26
Test
Level B Run 1
Level B Run 2
Level B Run 3
Level B Run 4
Level B Run 5
Level C Run 1
Level C Run 2
Level C Run 3
Level C Run 4
Level C Run 5
Level C Run 6
2-27
Standard samples
Comments
Start at 0905, stop at 0920, System stable.
Collaborator 8 backed up both trains.
Start at 1015, stop at 1030, System stable.
Start at 1115, stop at 1130, System stable.
Start at 1345', stop at 1400, System stable.
Start at 1440, stop at 1455, System stable.
Start at 0855, stop at 0910, propylene and
methyl mercaptan showed some drift.
Start at 0950, stop at 1005, System stable
Collaborator 9 lost one sample--wrong
impinger order.
Start at 1040, stop at 1055, System stable.
Start at 1315, stop at 1330, System stable.
Start at 1410, stop at 1425, methyl mercaptan
varying, H~S reading rose briefly.
Start at 1505, stop at 1520, carbonyl sulfide
on at *« 80 ppm. Methyl mercaptan, carbonyl
sulfide, 1^S, and ethyl mercaptan drifted
10 to 20% during test.
No problems reported.
27
-------
Figure 8 - Collaborators Sampling Trains
Figure 9 - Collaborators Analyzing Samples
28
-------
Figure 10 - Collaborator Analyzing Test Samples
29
-------
Figure 11 - Participating Collaborators and Midwest Research Institute Staff,
Hydrogen Sulfide Collaborative Test, February 23 through 27, 1976,
MRI Field Station, Grandview, Missouri
Left to right, standing: David Henry, Ecology Audits; Kent Shoemaker, Clay-
ton Environmental Consultants; Mark Childers, Ecology Audits; Paul C. Constant,
Midwest Research Institute; Larry Hackworth, Research Triangle Institute; Dave
Harsch, Washington State University; Denny Wagoner, Research Triangle Institute;
Lee Bamesberger, Washington State University; Robert Epstein, York Research
Corporation; Windle McDonald, Monsanto Research Corporation; Harold Hubbard,
Midwest Research Institute; Lewis Cash, Monsanto Research Corporation; Bruce
DaRos, Midwest Research Institute; Barry Jackson, Roy F. Weston; George Scheil,
Midwest Research Institute; Rodney Midgett, EPA, Quality Assurance Branch;
and John LaShelle, Midwest Research Institute.
Left to right, kneeling: Harilal Patel, Interpoll; Perry Lonnes, Interpoll;
James Fries, California Department of Health; Emil DeVera, California De-
partment of Health; Victor Hansen, Clayton Environmental Consultants;
Michael Carey, Roy F. Weston; Nicola Grappone, The Research Corporation of
New England.
Not pictured: Sam Cha, The Research Corporation of New England; and
Michael Horowitz, York Research Corporation.
30
-------
All collaborator data for the test was recorded in duplicate on the
data forms shown in Figures 12 and 13. One copy of the data was submitted
before leaving the test site and the second copy was retained by the col-
laborators. After returning to their home laboratories> the collaborators
checked their results and reported any errors or changes to MRI.
31
-------
MIDWEST RESEARCH INSTITUTE
DATA SHEET
HYDROGEN SULFIDE - CADMIUM SULFATE METHOD
SAMPLING
Collaborator^
Run No.
Sampling Train No.
INITIAL READINGS
Dry Gas Meter
Meter Temp. °C_
Time
Rotameter
Type Ball Used: Glass
Sampling Flow Rate
Barometric Pressure mm Hg
Date
Sampled by
Port No.
FINAL READINGS
Dry Gas Meter
Meter Temp. °C
Time
Rotameter
Steel (Circle)
Measured Gas Meter Vol. (Vs)
REMARKS:
Meter Correction Factor*
Si At Standard Cond. (V .___)
. STD -
ANALYSIS
Analyzed by_
Time at I Addition
2
Time at Titration
Volume of Iodine Added to Sample
Normalities: .01 N Iodine
Blank: .01 N Thinsulfate
Volume ml
Volume of Iodine Added to Blank
H S Concentration_
2 mg/DSCM
REMARKS:
Burette: (.01 N Thiosulfate)
Final RDG.
Initial RDG.
Volume of Thio (V )
.01 N Thiosnlfate
jnl
ml
*Vs
— where Vs is the true gas volume and Vm is the gas volume indicated by the
Vm
gas meter.
Figure 12 - Sampling and Analysis Data Sheet
32
-------
MIDWEST RESEARCH INSTITUTE
DATA SHEET
HYDROGEN SULFIDE - CADMIUM SULFATE METHOD
STANDARDIZATIONS
Collaborator Analyzed by Date_
0.1 N Sodium Thiosulfate
Weight of K Cr 0 g
Volume of K Cr 0 Solution ml
Final Buret Reading
Initial Buret Reading
Volume of 0.1 N Thio
Sodium Thiosulfate Volume (V ) ml
Normality (N )
0.01 Iodine Volume of Iodine
Final Buret Reading
Initial Buret Reading
Volume of .01 N Thio
Thiosulfate Volume (V ) ml
Normality (NT)
Blanks
Time at !„ Addition
Time at Titration
Final Buret Reading
Initial Buret Reading
Volume of .01 N Thio
Volume of Thiosulfate ml (VT)
REMARKS:
Figure 13 - Data Sheet for Standardizations
33
-------
SECTION VIII
RESULTS OF ANALYSES
The collaborators' results are summarized in Tables 6 and 7. Copies
of the original data sheets are in Volume III of this report. The results
of MRI's measurements of the H S levels during the manifold tests are sum-
marized in Table 8,
The liquid samples were prepared by weighing portions of 99% pure cad-
mium sulfide to the nearest 10 \ig. Each sample was then stored dry until
the collaborative test. Shortly before issuing each set of samples, MRI
added 45 ml of absorbing solution to each sample and the samples were mixed
for at least 10 min in an ultrasonic bath. The collaborators were instructed
to assume a sample volume of 10 liters in the calculations. The true value
for each sample was calculated thus:
u c ,. «-• r IA \ - 34.1 • (sample wt) • 100
hV S concentration (mg/dscmj = •,,,—^—
2 ° 144.5
where sample weight is expressed as milligrams of CdS
The H S standard gases were prepared using research grade methane. Each
cylinder was analyzed by the supplier using a flame photometric procedure
before shipment and again after the collaborative test. MRI also checked
the cylinders before the collaborative test using the cadmium sulfide train
and using the potentiometric method. Both methods agreed within + 5% of the
reported values. The reported ILS concentrations measured in parts per mil-
lion were converted to milligrams per dry standard cubic meter by multiply-
ing by 1.429. Cylinder 1 was analyzed as 75 ppm initially and 69 ppm follow-
ing the test. Cylinder 2 was reported as 250 and 240 ppm and Cylinder 3 was
reported as 130 and 127 ppm. The average of the two analyses was taken as
being the true value.
34
-------
TABLE 6. H2S CONCENTRATION (mg/dscra) DETERMINED BY COLIABORATORS FOR STANDARD LIQUID AND CYLINDER SAMPLES
LO
Ul
Collaborators
Standard 1
samples MRI Coll
Test No. 1 0 6.5
133 132.7
175 174.2
421 416.7 .
4.7 25.6-
Cylinders
Test No. 1
Cyl. 1 157.7
(103 mg/DSCM) 118.9
Cyl. 2 364.1
(350 rag/DSCM) 377.4
Cyl. 3 204.5
(184 mg/DSCM) 199.5
Standard
Samples MRI Coll
Test No. 2 0 -11.6
110 110.4
340 328.9
251 245.8
2.36 1.7
Cylinders
Test No. 2
Cyl. 1 117.3
(103 mg/DSCM) 107.2
Cyl. 2 329.8
(350 mg/DSCM) 346.8
Cyl. 3 192.4
(184 mg/DSCM) 190.1
2
MRI Coll
0 < 8(0)
4 10.0
204 198.6
100 100.6
377 363.8
116.0
117.1 -
366.3
354.7
193.3
197.3
MRI Coll
470 442.9
242 230.0
106 111.0
4.01 8.8
0 < 8(0)
101.3
97.4
309.4
305.6
186.8
184.3
3
MR! Coll
357 333
113 98.4
2 -4.8
0 -4.8
182 174
88.9
82.8
331
333
171.0
168.0
MRI Coll
0 0
2.83 6.8
139 132
187 175
367 339
94.9
94.3
328
338
175
183
4
MRI Coll
2.83 1.2
0 0.9
341 316.6
127 110.3
224 208.8
93.07
90.39
334.9
335.7
181.7
179.5
MRI Coll
450 427.6
98.1 91.9
192 175.6
0 0.3
3.06 0.5
93.4
91.0
333.2
333.9
179.7
178.1
5
MRI Coll
179 179.1
4.5 9.5
379 370.7
0 6.1
131 134.3
90.8
92.3
330.2
323.6
172.1
173.6
MRI Coll
232 220
121 122.1
334 320.4
0 2.71
4.25 3.22
88.8
89.4
317.4
318.6
170.1
169.7
6 7
MRI Coll MRI Coll
030 4.62
138 43 2.8 9.74
231 224 98 96.92
4.5 14 256 251.62
460 433 358 343.92
93 92.94
93 89.92
342 343.93
339 332.37
177 228.92
179 227.16
MRI Coll MRI Coll
174 178 2.59 0.43
448 415 0 -2
131 126 137 126.06
0 7 218 199.57
3.06 5 408 381.61
95 100.85
95 102.06
338 355.67
334 335.06
180 189.34
174 182.83
8
MRI Coll
3.7 11.9
171 149.6
123 110.5
381 331.5
0 5.1
100. 1
100.6
232.5-'
257.2-
176.9
172.4
MRI Coll
136.0 113.1
396 333.3
2.1 4.19
0 2.69
238 208.5
83.8
90.30
284.3
286.4
155.8
153.2
9
MRI Coll
99 103.61
243 237.85
348 339.95
4 8.54
0 9.37
"87.31
87.87
297.08
314.77
166.5
168.41
MRI Coll
370 368.72
97 121.74
189 201.27
3.5 17.08
0 13.73
111.13
106.78
312.50
312.84
173.92
174.70
10
MRI Coll
0 5.78
4.5 11.73
104 105.4
257 248.2
387 374.3
100.9
90.6
329.0
332.5
177.7
170.1
MRI Coll
196 191.9
397 3/0.4
133 132.8
2.4 10.2
0 8.5
95.47
94.14
330.9
333.9
184.5
180.1
1 — • — — — — .
57Error during sampling or analysis. Sample deleted
-------
TABLE 7. H7S CONCEOTRATION (•g/dsca) DETERMINED BY COLLABORATORS AT TEST LEVELS A, B, C
Level RunS/
la
Ib
2a
2b
A 3a
3b
4a
4b
5a
5b
la
Ib
2a
2b
B 3a
3b
4a
4b
5a
5b
la
Ib
2a
2b
C 3a
3b
4a
4b
5a
5b
6a
6b
H2S
present
180
180
188
191
192
342
380
378
426
426
100
100
99
97
102
100
Collaborators
1
187.9
199.2
191.9
203.9
215.1
208.2
713.8
216.0
223.2
218.5
375.3
333.7
406.7
407.7
384.2
395.1
447.7
432.8
434.2
415.5
109.3
108.5
107.4
105.0
115.3
110.9
127.0
123.9
128.4
127.6
123.1
122.6
2
271.0
277.9
210.0
214.2
221.9
204.1
212.2
203.8
211.1
214.1
356.3
351.2
399.0
340.3
392.2
388.4
420.4
446.6
436.4
427.2
114.0
116.7
118.2
117.7
121.7
122.2
123.3
114.4
120.4
118.6
117.9
121.3
3
217h/
246-^
174
177
194
194
195
236-^
197
338
346
376
364
369
373
399
393
397
403
111
112
114
113
113
112
110
114
113
116
117
117
4
184.2
181.0
182.8
176.0
196.6
193.2
198.4
197.6
200.7
202.4
334.8
333.3
375.0
374.0
379.7
373.5
395.1
402.2
404.1
399.4
115.99
113.7
114.3
111.6
lll.l
110.7
112.0
111.9
116.9
115.4
111.5
112.8
5
181.7
181.7
177.3
173.8
190.3
189.3
193.7
195.9
194.2
194.7
328.4
327.7
364.8
362.1
365.3
361.0
386.1
383.7
382.4
390.0
108.6
109.4
111.1
111.8
109.5
109.4
110.6
110.4
115.5
115.6
113.5
112.5
6
180
182
174
177
195
196
197
199
200
204
337
339
375
386
377
376
402
405
400
405
113
113
113
112
111
119
114
115
114
113
120
116
7
191.83
179.72
180.34
173.29
175.55
187.19
230.466
190.750
202.66
196.13
347.99
337.70
439.55
417.51
439.61
426.05
416.55
408.72
424.32
422.01
118.61
116.34
119.86
112.98
117.73
113.48
118.95
115.22
128.71
116.74
120.76
112.99
8
122.2^
144.5^
114.2^
150. 1-^
144.4
157.4
164.7
204.8
167.0
220.8
312.6^
331.6
314.5
366.3
356.3
416.9
393.2
378.5
371.7
110.8
110.6
102.4
103.0
99.4
100.4
109.7
109.9
113.7
107.9
106.6
105.8
9
170.79
176.43
164.02
172.82
183.88
188.94
183.91
186.81
184.25
191.05
315.15
321.98
290.51
354.18
342.31
354.03
372.88
384.65
384.43
375.27
109.61
115.31
71.35 ,
50.19^
102 . 20
115.13
114.48
118.31
123.08
118.03
121.52
125.24
10
219.1
221.5
205.8
205.6
224.6
216.3
193.6
217.7
224.3
227.1
364.3
369.9
392.6
401.1
388.6
394.1
411.2
409.4
420.8
417.3
182.66
144.4
144.9
143.05
142.9
143.8
138.9
143.9
150.3
146.1
151.1
129.0
a,/ Each collaborator ran two trains simultaneously each run.
_b/ Error (hiring sampling or analysis. Sample deleted.
-------
TABLE 8. DATA SUMMARY OF MRI MEASUREMENTS AT TOE HjS MEASURE CONCENTRATION DURING THE MANIFOLD TEST
Potentloraetric tltratlons
Level Run
A la
Ib
2a
2b
3a
3b
4a
4b
5a
5b
B la
Ib
2a
2b
3a
3b
4a
4b
5a
5b
C la
Ib
2a
2b
.3 a
3b
4a
4b
5,i
5b
6a
6b
1
Vol. 2
tttrant Gas vol. sampled
(ml) (1) at 760 ran. 21°C
11.7
11.3
12.0
11.3
14.0
14.5
17.8
14.7
14.5
14.4
18.7
18.9
20.8
20.9
17.5
18.4
20.3
20.1
21.5
21.8
12.3
12.8
12.8
12.7
12.7
13.3
12.5
11.5
12.6
13.2
12.3
13.0
16.19
16.80
16.19
16.80
18.89
19.60
18.86
19.57
18.86
19.57
18.94
19.65
18.75
19.45
16.12
16.73
17.84
18.06
18.74
19.45
19.17
19.96
19.12
19.84
19.07
19.79
19.04
19.76
18.89
19.60
18.89
19.60
3
l^S measured
(mg/dscm)
188
175
193
175
193
193
246
196
200
192
348
339
391
379
383
388
401
392
405
395
98
98
102
98
102
101
too
89
102
103
99
101
4
II 2S
Flow rate
(ml/ratn)
20.25
19.98
21.06
21.18
21.37
38.67
42.82
42.20
50.85
50.61
11.30
11.34
11.38
11.38
11.34
11.05
Dilution factor calculi
5 6
Gas Pres. H2S flnw rate at
(mm Hg) manifold (ml/mln)
994
992
991
988
990
989
988
986
991
989
998
996
996
997
996
995
18.73
18.49
19.53
19.71
19.82
35.89
39.82
39.26
47.40
47.11
10.43
10.49
1.0.53
10.63
10.57
10.29
at tons
7
Total flow rate/ 8
mln at manl- HjS cone.
fold conditions (ppm)
150.0
150.9
153.4
151.3
150.6
151.4
151.2
151.3
148.6
149 . 3
148.2
148.6
156.0
151.5
149.6
148.1
124.9
122.5
127.3
130.3
131.6
237.1
263.4
259.5
319.0
.315.5
70.4
70.6
67.5
70.2.
70.7
69.5
9 10 11
H2S cone. Avg. potentlometrlc Final avg.
(mg/m3) result (mR/dscra) (mj?/dscm)
178
175
182
186
188
339
376
371
456
451
101
101
96
100
101
99
182
184
193
196
196
344
385
386
396
400
98
100
102
94
102
100
180
180
188
191
192
342
380
378
426
426
100
100
99
97
102
100
-------
The methods used to determine the H S content of the manifold stream
showed a consistent 15% spread in the H S measured. The gas chromatography
results are the lowest of the three. H S is apparently irreversibly absorbed
by reactive surfaces with the chromatograph since rapid, repeated injections
show increasing peak heights until.a plateau is reached. If the injections
are then made with longer delay times, the peaks decrease in height. There-
fore, the gas chromatography results were not used in determining the true
H S levels.
The potentiometric titrations of H S (and mercaptans) gave a change
in potential much lower than expected. However, duplicate samples agree
within + 5% and runs made with the standard gas cylinders showed 100% col-
lection efficiencies.
The gas dilution measurements showed excellent stability, and initially
agreed with the GC analyses. After measuring the flow of H S at ambient and
at 5 Ib pressure MRI showed that the published equation for rotameter pres-
sure corrections is in error for the rotameter used in the test. Using the
empirical corrections, the dilution factors were in reasonable agreement
with the potentiometric titrations, and both were in agreement with the pre-
dicted mercaptan interferences. The average of the potentiometric and dilu-
tion results was used as the true value for each run.
38
-------
SECTION IX
STATISTICAL ANALYSIS OF COLLABORATORS1 RESULTS
The H S collaborative test data consist of three sets of results:
(a) standards data, (b) the actual field data, and (c) liquid standards
data. These sets were analyzed individually and results for each block
will be presented. A summary description of all results will be presented
last.
(DRY) STANDARDS DATA
Each collaborator produced two replicate determinations for three test
cylinders (103, 350, and 184 mg/dscm), and this procedure was repeated at
the end of the week. Thus, the data set is described by the 10 x 3 x 2 x
2 factorial model:
X. .. = H + C. + L. + CL. . + T. + CT.. + LT., + CLT. + e .,
ijkl x j ij k ik jk ijk (ijk)l
where U- = overall mean response
C = i collaborator (i = 1, ..., 10) (random factor)
i
L. = jth test level (j = 1,2,3) (fixed factor)
T, = kth test (k = 1,2) (fixed factor)
k
CL = collaborator-level interaction, etc.
ij
e . = e residual in ijk cell (1 = 1,2 everywhere)
\iJ k)L
X. = ijkl response where a response is a bias, i.e., collaborator
reading - true value (mg/dscm).
Five observations were deleted as outliers from the data set; two due
to mechanical failure (C7-L3-T1); and three as statistical outliers (Dixon-
Massey outlier test, C8-L2-T1 and the first value of Cl-Ll-Tl).*
* \[22 = 0.744 (p < 0.005), and \fTI = 0.421 (p < 0.10).
39
-------
Before performing an analysis of variance of the data set, the cell
ranges were compared to the H S levels as a check on the homeoscedastic
assumption of analyses of variance. It was found that the average cell
range (R) was proportional to the H S level (R = 2.9, 3.5, 7.7 versus L =
103, 184, 350); i.e., ae is proportional to the H S level (rather than
constant).
Therefore a log transformation was applied to the data; in particular
Z. = In (100 + X. .). The analysis of variance was then performed and
the results are shown in Table 9.
TABLE 9. ANALYSIS OF VARIANCE OF H S STANDARDS DATA
Source
Collaborator (C)
Level
Test
CL
LT
CT
CLT
Error
(L)
(T)
(e)
dF
9
2
1
18
2
9
16
57
1.
0.
0.
0.
0.
0.
0.
0.
SS
2059
7699
0886
5772
2009
8869
2995
1879
0.
0.
0.
0.
.0.
0.
0.
0.
MS
1340
3849
0886
0321
1005
0985
0187
0033
40.
12.
0.
9.
5.
29.
5.
64,
00,
90,
73,
49,
89,
68,
F
P <
P <
NS
P <
P <
P <
P <
0.
0.
0.
: o.
0*
0.
01
01
01
05
01
01
As can be seen from Table 9, all effects except Test are statistically
significant. Such a high number of significant effects is the usual result
when the replicate is a duplicate measurement only rather than a genuine
recreation of the same experimental conditions.
The primary results of interest from the analysis are the estimates
of^components of variance (see Table 10) and, in particular, the components
cr2 and cr2 (or their square roots). Note that, due to the log transforma-
G \j
tion of the data, Oe, etc., are theoretically constant percentages of the
H S level; i.e., a constant coefficient of variation (cr/|J.) exists in the
data. The observed CV1s in the data set for replicate readings were 2.50,
1.69, and 1.95% (versus theoretically constant 2.16%), so this idealization
does not seem unrealistic.
In other words, from Table 10 we have:
A collaborator produces readings with a standard deviation = 2.16%
of the H S level being measured.
The standard deviation of (mean) collaborator determinations of an
H S level is 3.93% of that H2S level.
40
-------
TABLE 10. COMPONENTS OF VARIANCE, H S STANDARDS DATA
Source
C
L
T
CL
LT
CT
CLT
e
02
e
02
e
02
e
02
e
02
e
02
e
02
e
02
e
EMS
+ 12 02
c
•f 4 02 + 40 02
CL L
+ 6 02 + 60 02
CT T
+ 4 a2
CL
+ 2 02 + 20 Q2
CLT LT
4- 6 o2
CT
+ 2 cr2
CLT
02 (Z- Scale) CV (%)
0.010891 3.93
0.008822
"0"
0.007192
0.004087
0.015875
0.007712
0.003297 2.16
The "total" standard deviation of a set of collaborators reading
an H S level is Jcr2 + 02 = 4.43% of that H S level.
Of course, the uniform CV over the experimental H S range (100 to 350
mg/dscm) might not hold for other H S values.
The other result of primary interest is the bias exhibited by the method.
These results,* along with the average mean square error (MSB) are shown in
Table 11.
In general, the bias is ••*-57oJ i.e., the H S method has a collection
efficiency of about 0.952 on the standard cylinders. The individual bias
curves are shown in Figure 14.
FIELD DATA
The field data are structurally similar to the standards data with
three H S levels and 10 collaborators, but with five runs per level. Formally,
the experimental design model is:
X. = \L + C. + L. + CL. . + R 4- CR + LR + CLR. ., + e.. ., .t
ijkl i j ij k ik jk ijk (ijk)l
The bias was estimated from actual results (not transformed but from the
Z-scale).
41
-------
TABLE 11. BIAS AND MSE, H S STANDARDS DATA
2
H S level
103 mg/dscm
184 mg/dscm
350 mg/dscm
Avg (212 mg/dscm)
Bias (B)
mg/dscm (%)
-5.68 (-5.5)
-6.20 (-3.4)
-18.78 (-5.4)
-10.22 (-4.8)
a/
NJMSE i-
5.91%
4. 03%
5.82%
5.26%
\JMSE 2^
7.09%
5.62%
7.02%
6.57%
al Within a collaborator.
b/ Between collaborators.
42
-------
CO
350
TRUE H2$ (/yg/DSCM)
Figure 14 - Bias Results by Collaborator, H S Standards Data
-------
The terras in (2) are the same as in (1) except that run (R) has replaced
test (T), and there are five runs per H S level. As with the standards data,
the cell ranges are proportional to the H S level, so that a log transforma-
tion was applied.*
In this data set all of C8's results at Level A and all of ClO's results
at Level C were discarded as outliers. Thirteen other values were also dis-
carded. After these values were replaced by artificial readings, the analysis
of variance was performed (Table 12) and components of variance estimated
(Table 13). The individually observed a 's were 2.33, 2.41, and 2.09% of the
Q
H S level, respectively.
The CT , cr values from Table 13 are very similar to the corresponding
fo
rom Table 10; i.e., no deterioration in precision was observed
between the standards and field results.
' / '
The biases in the field test are shown in Table 14 (individual bias
curves are shown in Figure 15). The biases are different in the field than
in the standards; namely, the field biases are about +2.6% on the average
(although level dependent), whereas the standards biases were negative
(-4.8% on the average).
LIQUID STANDARDS
Each collaborator also observed five standard liquid samples each dur-
ing two tests. Unlike the other two data sets, however, in this block the
H S level was not tightly controlled, so that each collaborator "saw" 10
unique H S values (of course the general H S span was similar for all col-
laborators).
Thus, instead of an analysis of variance, an analysis of covariance
(with covariate = H S level) was performed.
No difference between tests was found (f(1,96) < 1), but in most other
respects statistically significant influences were found:
. There was a significant collaborator effect; i.e., not all collabo-
rators read the same H S value even after adjustment for the vari-
ability in levels (F(9,89) = 7.22, p < 0.01).
. One regression line does not describe the data as well as 10 indi-
vidual regressions (F(18,80) = 11.47, p < 0.01).
* This time Z = log X was used where X = H S reading (not bias). This was
done for convenience only.
44
-------
TABLE 12. ANALYSIS OF VARIANCE H S FIELD DATA
Source df SS MS
Collaborator (C) 9
Level
(L)
Run (R)
CL
CR
LR
CLR
Error
(e)
2
4
16
36
8
61
.129
0.
72.
0.
0.
0.
0.
0.
0.
5511
2612
3589
1399
1037
2449
1445
0627
0.
36.
0.
0.
0.
0.
0.
0.
0612
1306
0897
0087
0029
0306
0024
000486
125.
4,152.
30.
17.
5.
12.
4.
93,
94,
93,
90,
97,
75,
94,
P *
P <
P *
P <
P *
P <
P f
C 0.
C 0.
C 0.
: 0.
C 0.
C 0.
C 0.
01
01
01
01
01
01
01
TABLE 13. COMPONENTS OF VARIANCE, H S FIELD DATA
Source
C
L
CL
CR
CLR
R
LR
e
EMS
02 + 30 02
e C
02 + 100 02 + 10 02
e L CL
02 + 10 02
e CL
02 + 6 o2
e CR
02 + 2 02
e CLR
02 + 60 02 + 6 02
e R CR
02 + 20 02 +2 02T
e LR CLR
02
02 (Z- Scale)
0.002025
0.3612
0.0008258
0.0003992
0.0009415
0.001447
0.001410
0.000486
CV (%)
4.50
2.20
45
-------
TABLE 14. BIAS AND MSE, H S FIELD DATA
H2S Level
100 mg/dscra
186 mg/dscm
390 mg/dscm
Avg (225 mg/dscm)
Bias (B)
mg/dscm (70)
+14 (+14.0)
+10 (+5.4)
-7 (-1.8)
+6 (+2.6)
a/
\JMSE 1-
14.2%
5.8%
2.87,
NfMsF^7
14. 97o
7.47,
5.370
_a/ Within a collaborator.
b/ Between collaborators.
46
-------
20i-
-20
100
C8 Missing at 100 /ug/DSCM
CIO Missing at 390/ug/DSCM
• True Value
380
TRUE H2S (/ug/DSCM)
Figure 15 - Bias Results by Collaborator, H S Field Data
-------
. Not all collaborator slopes are equal (F(9,80) = 9.51, p < 0.01).
The regression on the means is not completely linear (F(9,89) =
8.00, p < 0.01).
The individual regression curves are described in Table 15, and illus-
trated in Figure 16. Note that the inequality of regression curve slopes is
mostly due to one collaborator (C8). The average bias in the liquid standards
data set is -3.75%, a value not very different from the dry standard result.
The standard deviation within a collaborator cannot be directly computed
for this data set because a collaborator did not repeatedly observe a fixed
H S value.
The standard error of the estimate from the AOCV is 3.83%, but this
includes lack of fit. The collaborator standard deviation is
-------
TABLE 15. LIQUID STANDARD H S RESULTS BY COLLABORATOR
Collaborator
Cl
C2
C3
C4
C5
C6
C7
C8
C9
CIO
All
a (intercept)
-0.46
-1.16
-1.63
-2.46
4.12
7.96
1.69
6.51
12.97
8.14
3.91
b (slope)
0.985
0.961
0.937
0.945
0.958
0.924.
0.944
0.844
0.955
0.930
0.936
r (correlation)
99.97o +
99.97» +
99.97, +
99.970 4
99.97o +
99.97o +
99.97» +
99.97= 4
99.97o
99.97= +
99.67=
Note: Y (collaborator reading) = a + b X ("true value").
49
-------
Ln
O
24
22
20
18
16
14
12
10
8
6'
41
2
(0.0)
-2
-4
-6
J L
J L
_L
J_
_L
J I
8 9 10 11 12 13 14 15 16 17 18 19 20
TRUE H2S
Figure 16 - Theoretical Calibration Curves, H S Liquid Samples
-------
TABLE 16. SUMMARY OF ANALYSIS RESULTS
a
e
CTc
aT\|CTe + CTC
B
Standards
2.16
3.93
4.48
-4.81
Field
2.20
4.50
5.01
«.tf
Liquid
Standards
3.83s7
3.53
5.21
-3.75
U)
_a/ But this includes lack of fit, and is not directly com-
parable to the other two values.
jb/ But this bias is level dependent; the individual values
are.+14% (L = 100 mg/dscm), +5.4% (L = 186 mg/dscm),
and -1.8% (L = 390 mg/dscm).
51
-------
APPENDIX A
COLLABORATOR SELECTION DOCUMENTS
53
-------
REQUEST FOR PROPOSAL
1. Background
Midwest Research Institute (MRI) is under contract with the
Environmental Protection Agency (EPA) to evaluate a modification of
Method 11 - Determination of Hydrogen Sulfide Emissions From Stationary
Sources in EPA's program for the Standardization of Method 11, which is
a tentative method for determining the concentration of l^S in fuel gas
streams at petroleum refineries. This evaluation comprises three major
phases: in-house laboratory evaluation, field collaborative test, and
statistical analysis of the results of the collaborative tests.
MRI has completed the first phase and is now embarking on the
second phase—a field collaborative test. This test will be held in the
Kansas City, Missouri, area and will comprise ten collaborators sampling
and analyzing strictly according to the modified Method 11. The collabora-
tors' results will be analyzed by MRI.
2. Purpose
The purpose of this task is to provide EPA with data on the ac-
curacy and precision of the modified Method 11. (A copy of the modified
Method 11 forms Appendix A of this request for proposal.)
3. Statement of Work
A collaborative test program will be undertaken to achieve the
purpose given in Section 2 of this RFP. MRI, as prime contractor, will
54
-------
coordinate the testing. Each of the 10 collaborators participating in the
test will participate according to the Collaborative Test Plan given in
Section 4 of this RFP.
4. Collaborative Test Plan
4.1 Location of Test
The test will take place at MRI's field station which is located
approximately 15 miles south of MRI in a suburban area of Kansas City.
(See Appendix B.)
4.2 Date of Test
The test is planned for the week of February 22, 1976. See
Section 4.6 of this RFP for test schedule.
4.3 Sampling System
Sampling by the collaborators will be done indoors using the
sampling system shown diagrammatically in Figure A-l. Each collaborator
will attach his sampling trains to the sampling ports of the manifold
which will be operated at 5 psi (see Figure A-l). The ports are 1/8 in.
Swagelok with nut and ferrule. The collaborator will need to furnish
the appropriate matching connections for his trains to connect to his
1/4-in. O.D. Teflon sampling lines.
4.4 Number of Collaborators
There will be 10 different organizations, each of which will
serve as an independent, unbiased qualified collaborator.
55
-------
Mixer
Dry-Test
Meter
. Constant-
' « -1 Pressure
T Regulator
Commercial Source
of Natural Gas
Sampling
Ports
Sampling Lines
Collaborators'
Trains
Cylinder
of H2S
Rotameter-—T T
Critical—-i
Orifice
mi
Cylinders of Possible Interferences
1 - Methyl Mercaptan
2 - Ethyl Mercaptan
3 - SO2
4 - Carbonyl Suifide
5 - Isopropylene
6 - Monoethanol Amine *
* Conditional
Note: Manifold operated at 5 psi.
Exhaust
Figure A-l - Sampling System
-------
4.5 Experimental Design
The collaborative test can be considered as furnishing three
data sets:
(1) Analysis of known liquid samples to estimate purely
analytical errors,
(2) Analysis of known calibration gas samples (without inter-
ferences) to estimate sampling and analysis errors,
(3) Analysis of known field samples with realistic inter-
ferences present in the gas stream.
In all cases, three l^S levels will be used. In addition, blanks
will be included in the liquid samples. At all times all 10 collaborators
will be observing the same stream so that direct estimates of accuracy
and precision are possible.
Each of the collaborators will sample from the sampling system
with two trains simultaneously. Each collaborator's trains will be identi-
cal to the extent possible. This will provide each collaborator with
replicate samples per run. Thus, a run will produce 20 samples.
There will be three challenge levels. Each level will comprise
H2S, natural gas and possible interferences.
There will be three sampling test days. Each day there will be
a different challenge level. There will be at least 5 runs/test day.
57
-------
In addition, there will be two types of standard samples given
to each collaborator at the field site: eight liquid samples (two each at
levels 1, 2 and 3 and one blank) and six gas samples (two each at levels 1,
2 and 3). One sample of each type and level, plus one liquid blank, will
be given to each collaborator before the 3-day test starts—to be analyzed
on-site. A similar set will be given to each collaborator after completion
of the 3-day test, to be analyzed on-site.
4.6 Test Schedule
The field work will be executed as follows:
Day
Sunday
Monday
Friday
Date
February 22
February 23
Tuesday February 24
Wednesday February 25
Thursday February 26
February 27
Saturday February 28
Activity
Collaborators travel to Kansas City. Site
will be available to collaborator Sunday
afternoon for unpackaging equipment, set
up, etc.
Orientation. Equipment set up. Check of
calibration of collaborator's equipment.
Analysis of standard samples.
Sampling at Level 1. Analysis of samples.
Sampling at Level 2. Analysis of samples.
Sampling at Level 3. Analysis of samples.
Check of calibration of collaborator's
equipment. Analysis of second set of
standard samples. Packaging of equipment.
Turn in of field test data and results of
analysis.
Complete packaging and return home.
58
-------
A test day is an 8-hr work day (Monday through Friday) from
8:00 AM to 5:00 PM, with 1 hr off for lunch.
4.7 Collaborator Sampling and Analysis
Each collaborator will follow the modified Method 11 (see
Appendix A) exactly. He will do his work entirely independent of any of
the other collaborators. He will operate two trains simultaneously during
a run. The analysis of all samples (see test schedule, Section 4.6 of
this RFP) will be done at MRI's field station.
Each collaborator will need two people in the field. One will
be the person who performs the sampling. The other will be the person who
analyzes the samples. These jobs cannot be interchanged during any por-
tion of the test.
MRI will supply each collaborator deionized water, a 15-a, 115-
v, AC electrical service, analysis-bench space, and data forms. Also, a
barometer, an analytical balance (1-mg sensitivity) and a pH meter will
be part of MRI's test facilities available to collaborators.
Each collaborator shall calibrate each of his dry gas meters in
strict accordance with the procedure given in the modified Method 11. The
results of these calibrations shall be submitted to MRI at least 10 days
before the field test commences. The calibrations of the collaborators
dry gas test meters will be checked with MRI's wet test meter at the test
site.
59
-------
Each collaborator shall prepare the potassium dichromate stan-
dard at his home laboratory and standardize sodium thiosulfate at the
field site.
Each collaborator will need to furnish a 1/8-in. O.D. Teflon
tube to fit a 1/8-in. tube fitting (compression type). This is his con-
nection to gas-cylinder set-ups where he will get his standard gas samples.
5. Collaborators
5.1 Qualifications
Each collaborator is required to provide MRI with its qualifica-
tions to do the required testing and analysis. These qualifications shall
include:
(1) Capabilities and experience of personnel who will do the
work;
(2) Equipment available for use on the project;
(3) Analysis of the standard liquid H~S samples; and
(4) Related past and current programs of your company concerning
sampling of stationary sources. Experience in the measure-
ment of t^S would be desirable.
5.2 Project Personnel
Each project person must be identified and his direct and unique
qualifications given. This should include past sampling and analysis
experience by Method ll--when and where this work was done.
60
-------
Each person's function should be identified. The person who
analyzes the standard samples that are included as a part of this request
for proposal must also be the field analyst who will analyze test and
standard samples at the test site. The person identified as the one who
will be the operator of the H2S sampling trains must be the one who performs
this function at the collaborative test.
The project personnel will be considered to be "key personnel"
and shall not be substituted by the collaborator after the proposal is
received by MRI. Failure to provide the key personnel committed in the
proposal shall make the collaborator nonresponsive and will be cause to
terminate his subcontract.
5.3 Equipment
Each collaborator shall furnish all equipment and material as
specified in Section 5 and 6 of the modified Method 11 procedure (with
the exception of that specified in Section 4.7 of this RFP) necessary to
(a) perform calibration of his equipment at his home laboratory before
going to the field, and (b) perform the field sampling and analysis in the
field according to the plan given in Section 4 of this RFP.
Each collaborator shall have in the field three sampling trains
that meet the specifications given in modified Method 11. Two of these
are for operating two trains simultaneously per run. The other is to
serve as a backup. Also adequate chemicals, hardware, spare parts, etc.,
should be taken to the field. A sampling valve (needle valve or equivalent)
61
-------
will be needed with each train for its connection to a port of the manifold
in the sampling system (see Figure A-l).
The proposal shall identify the equipment and material that will
be taken to the field for the collaborative test.
5.3 Standard Samples
Each request-for-proposal recipient will be sent standard liquid
samples to be analyzed at his home laboratory by the analyst who will serve
as the collaborative field-team analyst in Kansas City. The results of their
analyses need to be part of the proposal submitted to MRI in response to
this request for proposal.
These standard samples will be shipped to the collaborator one week
after the request for proposal is mailed. This delay is to provide the
respondee with time to acquire the necessary material to perform the analysis,
since the samples must be analyzed immediately upon their arrival.
The cadmium absorbing solution must contain Antifoam B. All col-
laborators shall use the primary sample recovery method in the analysis of
all samples. The acidified-iodine extraction of the impingers will not be
used.
MRI will include a small amount of Antifoam B with the standard
samples shipped to the collaborators. The collaborators must obtain their
own Antifoam B for use in the field sampling.
62.
-------
5.4 Shipment of Equipment
Each collaborator should ship its field equipment and material to
Midwest Research Institute, 425 Volker Boulevard, Kansas City, Missouri,
Attention: George Scheil, prepaid and in sufficient time so that it will
arrive at MRI a week prior to Day 1 of the field test. If the collaborator
is driving and is bringing his equipment with him, this advance delivery is
not necessary. MRI should be apprised of the shipment method.
5.5 Lodging
The Ramada Inn shown on the figures of Appendix B is a recommended
motel for collaborators. This motel, as shown in the figures, is 35 miles
from the Kansas City International Airport and a few miles from MRl's field
station where the test will take place.
6. Contractual Document
It is MRl's intent to issue each collaborator a cost-plus-fixed-fee
type purchase order to authorize the work.
7. Insurance Coverage
Each collaborator assumes the risk of all injuries, including
death resulting therefrom, to all persons including collaborators,
employees, or any member of the public, and damage to and destruction of
property by whomsoever owned resulting from the prosecution of any work
or obligation undertaken or required by this project unless caused solely
by the negligent acts or omissions of MRI, its employees, servants, and
employees, from and against any and all liability arising therefrom,
63
-------
including all expenses legal or otherwise, incurred by them in the investi-
gation, defense, and settlement of any claim or suit.
Each collaborator shall secure before commencing and shall main-
tain during the performance of the work (a) Comprehensive General Liability
Insurance with the minimum Bodily Injury Limits of $300,000 each occurrence
and Property Damage Limits of $100,000 each occurrence; (b) Automobile
Liability Insurance with minimum Bodily Injury Limits of $300,000 each
occurrence and Property Damage Limits of $100,000 each occurrence; (c)
Statutory Workmen's Compensation and Occupational Disease Disability
Insurance; and (d) Employers' Liability Insurance with Limits at $25,000
each occurrence. Collaborators shall also furnish to MRI evidence of such
insurance coverage in the form of Certificates of Insurance, together with
evidence that the insurance carrier has assumed the Liability of the col-
laborator hereunder either by furnishing to MRI a properly executed "Assump-
tion of Contractual Liability" endorsement or by Certificate of Contractual
Liability Insurance. All Certificates of Insurance shall stipulate that
MRI will be given 10 days' written notice prior to any change, substitution
or cancellation prior to the normal expiration date.
64
-------
TENTATIVE METHOD FOR THE DETERMINATION OF HYDROGEN SULFIDE EMISSIONS
FROM STATIONARY SOURCESi/
a_/ A tentative method is one which has been carefully drafted from avail-
able experimental information, review editorially within the Quality
Assurance Branch, EMSL, and has undergone extensive laboratory evalu-
ation. The method is still under investigation and, therefore, is
subject to revision.
65
-------
1.0 Principle and Applicability
1.1 Principle
Hydrogen sulfide (l^S) is collected from a source in a series of
midget impingers and absorbed in pH 3.0 cadmium sulfate solution to form
cadmium sulfide (CdS). The latter compound is then measured iodometrically.
An impinger containing hydrogen peroxide is included to remove SC>2 as an
interfering species.
1.2 Applicability
This method is applicable for the determination of hydrogen sul-
fide emissions from stationary sources only when specified by the test pro-
cedures for determining compliance with the new source performance stan-
dards.
2.0 Range and Sensitivity
Q
The limit of detection is approximately 8 mg/m (6 ppm). The
n
maximum of the range is 740 mg/mj (520 ppm).
3.0 Interferences
Any compound that reduces iodine or oxidizes iodide ion will in-
terfere in this procedure, provided it is collected in the cadmium-sulfate
impingers. Sulfur dioxide in concentrations of up to 0.1 mole percent is
eliminated by the peroxide solution. Mercaptans coprecipitate with hydrogen
sulfide. In the absence of l^S, only traces of mercaptan are collected. A
mercaptan concentration equal to that of the h^S present yields results
that are approximately 57, high. Carbon oxysulfide of 20% does not interfere.
Certain carbonyl-containing compounds react with iodine and produce recurr-
ing endpoints. However, acetaldehyde and acetone at concentrations of 1 and
370, respectively, do not interfere.
Entrained hydrogen peroxide produces a negative interference
equivalent to 1007o of that of an equimolar quantity of hydrogen sulfide.
Avoid the ejection of hydrogen peroxide into the cadmium sulfate impingers.
4.0 Precision and Accuracy
Replicate analyses should not deviate by more than 570 relative
standard deviation. The collection efficiency of hydrogen sulfide measure-
ments has been established at 96 + 27» of the absolute value based on known
standards.
66
-------
5.0 Apparatus
5.1 Sampling Train
5.1.1 Sampling line - Six to 7 mm (1/4 in.) Teflon* tubing to
• connect sampling train to sampling valve. Depending on sampling stream
pressure, a pressure-reduction regulator may be required just prior to the
' Teflon sampling line.
If significant amounts of water or amine are present in the
sample stream, a corrosion-resistant cold trap should be used immediately
after the sample tap. The trap should not be operated below 0°C to avoid
condensation of 63 or C^ hydrocarbons.
5.1.2 Impingers - Five midget impingers, each with 30 ml capacity.
1 The internal diameter of the impinger tip must be 1.00 mm + 0.05 mm. The
impinger tip must be positioned 4 to 6 mm from the bottom of the impinger.
5.1.3 Glass or Teflon connecting tubing for the impingers.
5.1.4 Ice bath container - To maintain absorbing solution at a
low temperature.
! 5.1.5 Silica gel drying tube - To protect pump and dry gas meter.
5.L.("> Sampling valve - Needle valve or equivalent to adjust gas
flow rate. Stainless steel or other corrosion-resistant material.
5.1.7 Dry gas meter - Sufficiently accurate to measure sample vol-
ume, to within L7» and calibrated with a wet test meter over the range of flow
rates used in sampling. Gas volume for one dial revolution must not be more
than 10 liters. The gas meter should have a gas petcock or equivalent on the
outlet connector which can be closed during the leak test.
5.1.8 Flow meter - Rotameter or equivalent, to measure a 0.5 to
2.0 liters/rain (1 to 4 CFH) flow rate.
5.1.9 Graduated cylinder - 25 ml size.
5.1.10 Barometer - To measure atmospheric pressure to within
'+ 2.5 mm (0.1 in.) Hg.
5.1.11 U-Tube manometer - 0-30 cm. water column. For leak check
procedure.
* Mention of trade names of specific products does not constitute
endorsement by the Environmental Protection Agency.
67
-------
5.1.12 Rubber squeeze bulb - To pressurize train for leak check.
5.1.13 Tee, pinchclamp, and connecting tubing - For leak check.
5.1.14 Vacuum pump - Required for air purge.
5.1.15 Needle valve or orifice - To set air purge flow to 1 liter/min.
5.1.16 Tube packed with activated carbon - To filter air during
purge.
5.1.17 Volumetric flask - One 1,000 ml.
5.1.18 Volumetric pipette - One 15 ml.
5.2 Sample Recovery
5.2.1 Sample container - Iodine flask, glass-stoppered; 500 ml
size.
5.2.2 Pipette - 50 ml volumetric type.
5.2.3 Graduated cylinders - One each 25 and 250 ml.
5.2.4 Flasks - 125 ml, Erlenmeyer.
5.2.5 Wash bottle.
5.2.6 Volumetric flasks - Three 1,000 ml.
5.3 Analysis
5.3.1 Flask - 500 ml glass-stoppered iodine flask.
5.3.2 Burette - 50 ml.
5.3.3 Flask - 125 ml. Erlenmeyer.
5.3.4 Pipettes, volumetric - One 25 ml; two each 50 and 100 ml
5.3.5 Volumetric flasks - One 1,000 ml; two 500 ml.
5.3.6 Graduated cylinders - One each 10 and 100 ml.
68
-------
6.0 Reagents
Unless otherwise indicated, it is intended that all reagents con-
form to the specifications established by the Committee on Analytical Re-
agents of the American Chemical Society, where such specifications are
available. Otherwise, use best available grade.
6.1 Sampling
6.1.1 Cadmium sulfate absorbing solution - Dissolve 41.0 g of
3CdSO^-8 H20 and 15.0 ml of 0.1 M sulfuric acid in a 1-liter volumetric
flask that contains approximately 3/4 liter of deionized distilled water.
Dilute to volume with deionized water. Mix thoroughly. pH should be
3.0 + 0.1. Add 10 drops of Dow-Corning Antifoam B.* Shake well before.
use. If Antifoam B is not used, the alternate acidified iodine extraction
procedure must be used.
6.1.2 Hydrogen peroxide, 37° - Dilute 30% hydrogen peroxide to
3% as needed. Prepare fresh daily.
6.1.3 Water - Deionized, distilled, to conform to ASTM specifi-
cations D1193-72, Type 3.
6.2 Sample Recovery
6.2.1 Hydrochloric acid solution (HC1), 3 M - Add 240 ml of
concentrated HCl (specific gravity 1.19) to 500 ml of deionized, distilled
water in a 1-liter volumetric flask. Dilute to 1 liter with deionized
water. Mix thoroughly.
6.2.2 Iodine solution, 0.1 N - Dissolve 24 g of potassium
iodide (Kl) in 30 ml of deionized, distilled water. Add 12.7 g of re-
sublimed iodine (12) to the potassium iodide solution. Shake the mixture
until the iodine is completely dissolved. If possible, let the solution
stand overnight in the dark. Slowly dilute the solution to 1 liter with
deionized, distilled water, with swirling. Filter the solution if it is
cloudy. Store solution in a brown-glass reagent bottle.
6.2.3 Standard iodine solution, 0.01 N - Pipette 100.0 ml of the
0.1 N iodine solution into 1-liter volumetric flask and dilute to volume
with dionized, distilled water. Standardize daily as in Section 8.1.
This solution must be protected from light. Reagent bottles and flasks
must be kept tightly stoppered.
* Mention of trade names of specific products does not constitute
endorsement by the Environmental Protection Agency.
69
-------
6.3 Analysis
6.3.1 Sodium thiosulfate solution, standard 0.1 N - Dissolve
24.8 g of sodium thiosulfate pentahydrate (^28203-5H20) or 15.8 g of an-
hydrous sodium thiosulfate (^28203), in 1 liter of deionized, distilled
water and add 0.01 g of anhydrous sodium carbonate (NaoCOo) and 0.4 ml of
chloroform (CHC^) to stabilize. Mix thoroughly by shaking or by aerating
with nitrogen for approximately 15 min and store in a glass-stoppered, re-
agent bottle. Standardize as in Section 8.2 below.
6.3.2 Sodium thiosulfate solution, standard 0.01 N - Pipette
50.0 ml of the standard 0.1 N thiosulfate solution into a volumetric flask
and dilute to 500 ml with distilled water.
6.3.3 Starch indicator solution - Suspend 10 g of soluble starch
in 100 ml of deionized, distilled water and add 15 g of potassium hydroxide
(KOH) pellets. Stir until dissolved, dilute with 900 ml of deionized dis-
tilled water and let stand for 1 hour. Neutralize the alkali with concen-
trated hydrochloric acid, using an indicator paper similar to Alkacid test
ribbon, then add 2 ml of glacial acetic acid as a preservative.
(Note 1: Test starch indicator solution for decomposition by titrating
with 0.01 N iodine solution 4 ml of starch solution in 200 ml
of distilled water that contains 1-g potassium iodide. If
more than 4 drops of the 0.01 N iodine solution are required
to obtain the blue color, a fresh solution must be prepared.)
7.0 Procedure
7.1 Sampling
7.1.1 Assemble the sampling train as shown in Figure A-2, connect-
ing the five midget impingers in series. Place 15 ml of 3% hydrogen per-
oxide solution in the first impinger. Leave the second impinger empty.
Place 15 ml of the cadmium sulfate absorbing solution in the third, fourth,
and fifth impingers. Place the impinger assembly in an ice bath container
and place crushed ice around the impingers. Add more ice during the run,
if needed.
7.1.2 Connect the rubber bulb and manometer to first impinger,
as shown in Figure A-l. Close the petcock on the dry gas meter outlet. Pres-
surize the train to 30-cm. water pressure with the bulb and close off tubing
connected to rubber bulb. Train must hold a 30-cm. water pressure with not
more than a 1 cm. drop in pressure in a 1-min interval. Stopcock grease is
acceptable for sealing ground glass joints.
70
-------
Used for
Air Purge
SAMPLING VALVf!
Used for
Leak Check
TEFLO^J SAMPLING LINE
•
/ MIDGET IMPINGERS
FUEL GAS
LINE
SILICA GEL TUBE
DRY GAS METER
Used for
Air Purge
Figure A-2 - ii2s Sampling Train
71
-------
7.1.3 Purge the connecting line between the sampling valve and
the first impinger. Close valve and connect the sample line to the train.
Open the petcock on the dry gas meter outlet. Record the initial reading
on the dry gas meter.
7.1.4 Open the sampling valve and then adjust the valve to obtain
a rate of approximately 1 liter/min. Maintain a constant flow rate during
the test. Record the meter temperature.
7.1.5 Sample for at least 10 rain. At the end of the sampling
time, close the sampling valve and record the final volume and temper-
ature readings.
7.1.6 Disconnect the impinger train from the sampling line. Con-
nect the charcoal tube and the pump, as shown in Figure A-2. Purge the train
with clean ambient air for 15 min to ensure that all l^S is removed from the
hydrogen peroxide. For sample recovery, cap the open ends and remove to a
clean area that is away from sources of heat. The area should be well
lighted, but not exposed to direct sunlight.
7.2 Sample Recovery
7.2.1 Discard the contents of the hydrogen peroxide impinger.
Carefully rinse the contents of the third, fourth, and fifth impingers
into a 500 ml iodine flask.
(Note 2: The impingers normally have only a thin film of cadmium sulfide
remaining after a water rinse. If Antifoam B was not used or
if significant quantities of yellow cadmium sulfide remain in
the impingers, the alternate recovery procedure must be used.)
7.2.2 Pipette exactly 50 ml of 0.01 N iodine solution into a
125-ml Erlenmeyer flask. Add 10 ml of 3 M HCl to the solution. Quanti-
tately rinse the acidified iodine into the iodine flask. Stopper the
flask immediately and shake briefly.
7.2.2 (Alternate) Extract the remaining cadmium sulfide from
the third, fourth, and fifth impingers using the acidified iodine solution.
Immediately after pouring the acidified iodine into an impinger, stopper it
and shake for a few moments, then transfer the liquid to the iodine flask.
Do not transfer any rinse portion from one impinger to another; transfer it
directly to the iodine flask. Once the acidified iodine solution has been
poured into any glassware containing cadmium sulfide, the container must be
tightly stoppered at all times except when adding more solution, and this
must be done as quickly and carefully as possible. After adding any acidi-
fied iodine solution to the iodine flask, allow a few minutes for absorp-
tion of the H2S before adding any further rinses. Repeat the iodine
72
-------
extraction until all cadmium sulfide is removed from the impingers. Ex-
tract that part of the connecting glassware that contains visible cadmium
sulfide.
Quantitatively rinse all of the iodine from the impingers,
connectors, and the beaker into the iodine flask using deionized, distilled
water. Stopper the flask and shake briefly.
7.2.3 Allow to stand about 30 min in the dark for absorption of
the HoS into the iodine, then complete the titration analysis as in Sec-
tion 7.3.
(Note 3:CAUTIONl Iodine evaporates from acidified iodine solutions.
Samples to which acidified iodine have been added may not be
stored, but must be analyzed in the time schedule stated above
in 7.2.3.)
7.2.4 Prepare a blank by adding 45 ml of cadmium sulfate absorb-
ing solution to an iodine flask . Pipette exactly 50 ml of 0.01 iodine so-
lution into a 125-ml Erlenmeyer flask. Add 10 ml of 3 M HCl. Follow the
same impinger extracting and quantitative rinsing procedure carried out in
sample analysis. Stopper the flask, shake briefly, let stand 30 min in
the dark, and titrate with the samples.
(Note 4: The blank must be handled by exactly the same procedure as that
used for the samples.)
7.3 Analysis
(Note 5: Titration analyses should be conducted at the sample-cleanup area
in order to prevent loss of iodine from the sample. Titration
should never be made in direct sunlight.)
7.3.1 Using 0.01 N sodium thiosulfate solution, rapidly titrate
samples in iodine flasks using gentle mixing, until solution is light yel-
low. Add 4 ml of starch indicator solution and continue titrating slowly
until the blue color just disappears. Record Vt, the volume of sodium thio-
sulfate solution used (ml).
7.3.2 Titrate the blanks in the same manner as the samples. Run
blanks each day until replicate values agree within 0.05 ml. Average the
replicate titration values which agree within 0.05 ml.
8.0 Calibration and Standards
8.1 Standardize the 0.01 N iodine solution daily as follows: Pipette
25 ml of the iodine solution into a 125-ml Erlenmeyer flask. Add 2 ml of
3 M HCl. Titrate rapidly with standard 0.01 N thiosulfate solution until
the solution is light yellow, using gentle mixing. Add four drops of starch
73
-------
indicator solution and continue titrating slowly until the blue color just
disappears. Record Vt, the volume of thiosulfate solution used (ml). Re-
peat until replicate values agree within 0.05 ml. Average the replicate
titration values which agree within 0.05 ml and calculate the exact normal-
ity of the iodine solution using equation 9.1. Repeat the standardization
daily.
8.2 Standardize the 0.1 N thiosulfate solution as follows: Oven-dry
potassium dichromate (I^C^Oy) at 180 to 200°C. Weigh to the nearest milli-
gram, two grams of potassium dichromate into a 500 ml volumetric flask, dis-
solve in deionized, distilled water and dilute to exactly 500 ml. In a
500-ml iodine flask, dissolve approximately 3 g of potassium iodide (KI) in
45 ml of deionized, distilled water, then add 10 ml of 3 M hydrochloric acid
solution. Pipette 50 ml of the dichromate solution into this mixture. Gent-
ly swirl the solution once and allow it to stand in the dark for 5 min. Di-
lute the solution with 100 to 200 ml of deionized distilled water, washing
down .the sides of the flask with part of the water. Titrate with 0.1 N
thiosulfate until the solution is light yellow. Add 4 ml of starch indi-
cator and continue titrating slowly to a green end point. Record Vt, the
volume of thiosulfate solution used (ml). Repeat until replicate analyses
agree within 0.05 ml. Calculate the normality using eq. 9.2. Repeat the
standardization each week.
9.0 Calculations
Carry out calculations retaining at least one extra decimal figure
beyond that of the acquired data. Round off results only after the final
calculation.
9.1 Normality of the Standard Iodine Solution.
where:
Nj = normality of iodine, g-cq/liter;
Vj = volume of iodine used, ml;
NT = normality of sodium thiosul fate, K-O<] / I i t rr ; ,md
VT = volume of sodium thiosulfate used, ml.
9.2 Normality of the Standard Thiosulfato Solution.
NT = 2.04
74
-------
whore:
W = weip.ht of K2Cr2°7 uscd' 8'
VT "= volume of Na^S^O- used, ml;
NT - normality of standard thiosul f'ale solmimi, >•.-<••!/ •' '
and
2.04 a conversion factor.
(6 eq I2/tnole K2CrM (1,000 ml/liter)
(294.2 g K2Cr207/mole) (10 aliquot factor)
9.3 Dry Gas Volume - Correct the sample volume nira.Mi r.'d l>, ti
p.as meter to standard conditions (21nC) and /(>() mm !!}•..
where:
V * volume at standard conditions of j'.as sampli- ilin-uj'li
the dry f.as meter, standard liters;
V = volume of j',a.-: sample llironc.li I lie dry |-,a:. m« i . i <\<*< ii i
conditions), liters;
T , = absolute temperature at standard cond i I i cur;, .' ' . K,
T = average dry gas meter temperature, "K;
P, = barometric pressure at the orifice meter, mm llj;; and
P = absolute pressure at standard conditions, 760 rran lit-,.
std
9.4 Concentration of H2S - Calculate the concentration of M0S in
the gas stream at standard conditions using equation:
KRVjNj - VTNT) sample - (VjNj - VTNT) blank]
Crt c ™ iii'
"2 V
where (metric units):
CH q = concentration of H~S at standard conditions, myj
2K = conversion factor = 17.0 x 103
(34.07 g/mole H2S) (1,000 liters/m3) (1,000 mr./|-.)
(1,000 ml/liter) (2H2S eq/mole)
75
-------
V-r = volume of standard iodine solution, ml ;
NT = normality of standard iodine solution, p-cq/lit.or:
VT = volume of standard sodium thiosulfatc solution, ml
N-r *" normality of standard sodium thiosulfatt- solution,
g-eq/liter; and
dry gas volume at standard conditions, lit IT.-.
mstd
10.0 Stability
The absorbing solution is stable for at least 1 month. Sample
recovery and analysis should begin within 1 hour of sampling to minimize
oxidation of the acidified cadmium sulfide. Once iodine has been added
to the sample, the remainder of the analysis procedure must be completed
according to Sections 7.2.2, 7.2.3, and 7.3.
76
-------
MAPS
77
-------
lull-mat ionol
Airport
)
-^T3
N
KANSAS CITY. MISSOURI
871 h St.
Driving disfonco
from KCI lo
Romao'a Inn:
Approxiirntol)'
35 rni.
HA/MAD A INN
Map -- Kansas City In Vernal M-M--! Airport to n.-inmda Inn
78
-------
^ RAMADA INN
TRUMAN
CORNERS
FIELD STATION
llr\p: Il.-ii.n'la Inn to Field Station
79
-------
INSTRUCTIONS FOR SAMPLE ANALYSIS
80
-------
These test samples are similar to the liquid standard samples
which will be used during the collaborative test. Temperature changes
from -20 to +30°C do not adversely affect the samples. Over a 2-week
period, the samples show only slight changes in value. Please complete
the sample analyses within 2 weeks of the shipping date (December 29).
The sample analysis results and the standardization data must
be recorded on the data sheets provided and the data sheets returned to
MRI with your proposal. Report the results of any four of the six sam-
ples. Be sure to record the sample numbers on the proper data sheets in
the remarks section. Report the analysis results as mg/dscm. Assume
that the sampling data for each sample was as indicated on the data sheets.
A vial containing Antifoam B is included with the samples since
this is not a commonly available item.
Please complete the standardizations data before analyzing the
samples. Each sample should be treated as if it was the contents of an
impinger train. The samples are suspended in the normal absorbing solu-
tion and the sample volume is identical to the total amount used in the
sampling train (3 x 15 ml = 45 ml).
The synthetic samples are less reactive than the freshly pre-
cipitated cadmium sulfide obtained by Method 11. To obtain accurate re-
sults it is necessary to maintain the precipitate as a finely divided
suspension.
81
-------
To completely suspend the sample for analysis, the following
changes must be made in the analysis. First, shake each sample thoroughly.
Then place the samples in an ultrasonic bath for at least 10 min before
sample recovery. After removing each sample from the ultrasonic, shake
briefly, and proceed with the sample recovery steps without delay. Rinse
the contents of the bottle into an iodine flask in the normal manner and,
after adding the acidified iodine to the flask, stopper the flask and
shake vigorously for ^30 sec. If the ultrasonic bath is not used, if the
sample recovery is not done immediately after removal from the ultrasonic,
or if the sample/acidified iodine mixture is not shaken vigorously for 30
sec, recoveries will be low and erratic. The remainder of the sample
analysis is the same as for a normal sample.
82
-------
Collaborator_
Run No.
MIDWEST RESEARCH INSTITUTE
DATA SHEET
HYDROGEN SULFIDE - CADMIUM SULFATE METHOD
SAMPLING
Date
Sampling Train No.
Sampled by_
Port No.
INITIAL READINGS
Dry Gas Meter J?/ g. 6 3
Meter Temp. °C / g"
Time i3l 5
Rotameter
FINAL READINGS
. 8"?
Dry Gas Meter
133&
Meter Temp. °C
Time
87
Rotameter
llir- "«^
Type Ball Used:
-------
MIDWEST RESEARCH INSTITUTE
DATA SHEET
HYDROGEN SULFIDE - CADMIUM SULFATE METHOD
STANDARDIZATIONS
Collaborator Analyzed by Date
0.1 N Sodium Thiosulfate
Weight of K Cr 0
Volume of K Cr 0 Solution ml
Final Buret Reading
Initial Buret Reading
Volume of 0.1 N Thio
Sodium Thiosulfate Volume (V ) ml
T
Normality (N )
0.01 Iodine Volume of Iodine
Final Buret Reading
Initial Buret Reading
Volume of .01 N Thio
Thiosulfate Volume (V_) ml
Normality (NT)
Blanks
Time at I~ Addition
Time at Titration
Final Buret Reading
Initial Buret Reading
Volume of .01 N Thio
Volume of Thiosulfate ml (V_)
REMARKS:
84
-------
APPENDIX B
INSTRUCTIONS TO THE COLLABORATORS
85
-------
INSTRUCTION TO COLLABORATORS
A number of details with regard to the collaborative test have
not been mentioned or have surfaced in conversations with various prospec-
tive collaborators. The following comments should cover most questions
that you may have with regard to the test.
1. Calibration data for the 3 dry gas meters (including spare
unit) must be submitted to MRI 10 days prior to the test date, i.e., by
February 13. Calibration must be done by wet test meter or spirometer (bell
prover). Data must include specifications on the wet test meter (or spiro-
meter) and a simple block diagram showing the gas flow path in the com-
ponents of the test assembly. For example, the normal configuration using
a wet test meter is:
Exhaust
Compressed
Ga s
Source
-
Need Le
Vcilvv
| Dry Gas
~* JMeU-r ,
->
W.T.M.
Calibration can be done with air, nitrogen or natural gas. The data must
include all pertinent pressures, temperatures, etc. necessary to calculate
the meter correction factor, a sample calculation, and the final correction
factors reported to 1 part per thousand.
2. The test manifold is mounted on a round table approximately
12 ft in diameter. The port assembly is mounted on a small elevated plat-
form in the center of the table. An exhaust hood is suspended above the
platform with a flexible rubber skirt which extends below the ports. Each
86
-------
collaborator will be assigned a sector on the table approximately 3 ft along
the circumference of the table. Both sampling trains must fit within this
area (an equivalent space is available under the table). Exhaust gases
from the trains should be vented to the exhaust hood. The 110 VAC power is
available at the table and at a number of other locations in the building.
Please note that the pumps are to be used only during the air purge. During
the actual sampling the pressure head of the manifold (~ 4 to 5 psig) forces
the gas through the trains. Upon arrival at the site you will be assigned
two sample ports which will be used throughout the test. Each manifold
port terminates at 1/8 in. swagelock nut w/ferrules. Cap plugs seal off
any unused ports.
3. Space for sample recovery, titrations, etc., will be provided
at a number of benches and tables in the room. Each collaborator will have
at least 6 ft of bench/table space in addition to the sampling table area.
The building has only one sink. All solution wastes must be dumped into
waste buckets which MRI will provide. Deionized water is provided at a
large reservoir with spigot. Each collaborator should bring a poly-pac or
similar water bottle to avoid the creation of a traffic jam at the deionized
water tap.
4. Since up to 25 people will be working in an area approximately
30 x 60 ft, large, cumbersome pieces of equipment should not be used. For
example, anyone bringing the Rockwell meters used in RAG consoles should
dismount the meter and bring only the essential components. Every effort
87
-------
must be made during the test to plan ahead and work as rapidly and effi-
ciently as possible. Monday will be the critical day because it is the
first and due to the large number of tasks to be completed.
5. The specifications on irapinger tip diameter and spacing must
be adhered to for the 3 CdSO^ impingers. If you can keep these impingers
properly identified, peroxide and dry impingers need not meet the precise
specifications. The specification on impinger tip diameter is not a mis-
print. The ruggedness test showed that this is a critical factor in col-
lection efficiency. Most commercially available impingers will meet the
specifications. The entire test will go much faster if you can bring
enough spare impingers so that two sets of impingers can be prepared while
another two sets are sampling and being purged.
6. The liquid samples which each collaborator will analyze on
Monday and Friday are similar to those sent by mail. The samples used
during the test will be freshly prepared and agitated in an ultrasonic by
MRI before issuing them so that you may proceed immediately with the anal-
ysis after receiving them. All the liquid samples must be analyzed against
a blank using your own absorbing solution.
7. Safety glasses must be worn at all times in the test building.
No food or beverages are allowed in the building and smoking is also pro-
hibited. A guest house is immediately adjacent to the test building where
none of the above regulations apply. A drinking fountain, coffee, etc.,
will be available in the guest house.
88
-------
8. Ice is available from a number of sources in the vicinity.
1 9. Anyone who wishes may come to the test facility on Sunday
afternoon to unpack, assemble equipment, etc. However, no sample analyses
can be done until Monday. The grounds of the MRI field station are normally
kept locked except 8 AM through 5 PM weekdays. However someone will be on
I
duty at the site from 12 noon until 6 PM on Sunday and from 7 AM until the
i completion of all work each weekday. Someone will also be available on
Saturday, the 28th if needed. The test facility phone number is 761-1272.
10. All collaborators are to be present at the test site at 8 AM
on Monday for a brief orientation meeting. The MRI field station location
is shown on the map attached to the RFP. The test facility and guest house
are the buildings nearest the main gate.
11. Data forms for standardizations and sample results will be
provided by MRI. Each data form consists of two data sheets stapled to-
gether with carbon paper between them. Since no report is required of the
sample results, the data sheets are to be a complete and original record of
all data. One copy of all data for the day must be submitted to the MRI
test supervisor before leaving the site each evening. Each collaborator
will complete all necessary calculations for computing the final results.
The second copy is to be retained by the collaborator. MRI requests that
each collaborator check their data sheets and calculations after returning
home and report any errors or amendments within 3 weeks of the completion
of the test. We invite any comments on the test or the method from the
participants.
89
-------
12. On Monday, the 23rd, each collaborator must analyze a set of
liquid samples, run at least one simultaneous pair of samples from each of
the three reference gas cylinders and have the dry gas meters checked against
MRI's wet test meter. MRI personnel will assist in the meter checks. On
each of the next throe days each collaborator will make at least five simul-
taneous , paired runs from the test manifold. On Friday another set of
liquid samples will be analyzed and the reference gas cylinder sampling also
repeated. After completing all required tests the collaborators may pack
their equipment either Friday or Saturday. No testing is scheduled for
Saturday.
13. Each reference gas cylinder will have a pressure regulator
and enough sampling connections to allow up to 3 collaborators to sample
each gas simultaneously. Each collaborator must provide a 1/8 in. O.D.
Teflon tube to fit the compression fittings of the sampling connections.
14. All solution standardizations must be done on-site. How-
ever, the potassium dichromate standard should be dried and weighed at the
home laboratory. The weighed portions may be shipped dry in sealed vials
or in solution form. The other solutions may be shipped preweighed and dry
or as stock solutions. Each collaborator is responsible for bringing enough
reagents for the entire test schedule. In the event of loss or damage of
the reagents, MRI will provide enough reagents to complete the test. MRI
will also have a limited number of spare components for emergency use only.
Each collaborator is expected to bring sufficient spare parts to cover
breakage or malfunctions. This must include 3 complete sampling trai.ns.
90
-------
15. Any equipment which is shipped prepaid to MRI in advance of
the test will be delivered to the test site. Please notify George Scheil
or Paul Constant if you intend to ship equipment to the site and give ship-
ping date, number of boxes and shipping firm so that MRI can check in the
shipments and begin tracing any lost shipments. Any equipment shipped
should arrive at MRI 1 week prior to the test. The MRI phone number is
(816) 561-0202.
16. Each sampling period will be 15 min long. All sampling and
analysis must conform to the method. All cadmium sulfate absorbing solu-
tion must contain Antifoam B and the impingers will be rinsed with water
only-no HCl-12 extraction will be used during sample recovery.
91
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA 600/4-77-008b
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
STANDARDIZATION OF METHOD 11 AT A PETROLEUM REFINERY -
VOLUME II
5. REPORT DATE
January 1977
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
George W. Scheil and Michael C. Sharp
Midwest Research Institute
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
10. PROGRAM ELEMENT NO.
1HD621
11. CONTRACT/GRANT NO.
68-02-1098
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Monitoring and Support Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA-ORD
15. SUPPLEMENTARY NOTES
16. ABSTRACT
A collaborative test was run of the revised Method 11 procedures that was
developed in Volume I. Ten collaborators were selected from a total of 24 interested
organizations. Part of the screening process was to require each potential colla-
borator to analyze a set of liquid samples in accordance with the procedure. A test
manifold was constructed which could simulate typical refinery fuel gas streams.
During February 22 to 27, 1976, a total of 16 runs at three different H?S levels were
made, as well as a series of standard sample analyses. MRI then made a statistical
analysis of the collaborators' results. Without mercaptans present, results averaged
4 percent low with a standard deviation of 2 percent for any single collaborator and a
standard deviation between collaborators of 4 percent. With mercaptans present at
150 ppm, precision was unaffected but the results varied from 2 percent low to 14 per-
cent high at He concentrations of 400 and 100 mg/dscm. The tentative procedure was
found' to be adequately written for those knowledgeable of sampling and analysis
techniques.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COS AT I Field/Group
Air pollution
Gases
Hydrogen sulfide
Measuring
Petroleum refineries
Tests
Collaborative testing
Methods standardization
Methods evaluation
Stationary sources
13B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. OF PAGES
91
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
-------
INSTRUCTIONS
1. REPORT NUMBER
Insert the EPA report number as it appears on the cover of the publication.
2. LEAVE BLANK
3. RECIPIENTS ACCESSION NUMBER
Reserved for use by each report recipient.
4. TITLE AND SUBTITLE
Title should indicate clearly and briefly the subject coverage of the report, and be displayed prominently. Set subtitle, if used, in smaller
type or otherwise subordinate it to main title. When a report is prepared in more than one volume, repeat the primary title, add volume
number and include subtitle for the specific title.
5. REPORT DATE
Each report shall carry a date indicating at least month and year. Indicate the basis on which it was selected (e.g., date of issue, date of
approval, date of preparation, etc.).
6. PERFORMING ORGANIZATION CODE
Leave blank.
7. AUTHOR(S)
Give name(s) in conventional order (John R. Doe, J. Robert Doe, etc.). List author's affiliation if it differs from the performing organi-
zation.
8. PERFORMING ORGANIZATION REPORT NUMBER
Insert if performing organization wishes to assign this number.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Give name, street, city, state, and ZIP code. List no more than two levels of an organizational hirearchy.
10. PROGRAM ELEMENT NUMBER
Use the program element number under which the report was prepared. Subordinate numbers may be included in parentheses.
11. CONTRACT/GRANT NUMBER
Insert contract or grant number under which report was prepared.
12. SPONSORING AGENCY NAME AND ADDRESS
Include ZIP code.
13. TYPE OF REPORT AND PERIOD COVERED
Indicate interim final, etc., and if applicable, dates covered.
14. SPONSORING AGENCY CODE
Leave blank.
15. SUPPLEMENTARY NOTES
Enter information not included elsewhere but useful, such as: Prepared in cooperation with, Translation of, Presented at conference of,
To be published in, Supersedes, Supplements, etc.
16. ABSTRACT
Include a brief (200 words or less) factual summary of the most significant information contained in the report. If the report contains a
significant bibliography or literature survey, mention it here.
17. KEY WORDS AND DOCUMENT ANALYSIS
(a) DESCRIPTORS - Select from the Thesaurus of Engineering and Scientific Terms the proper authorized terms that identify the major
concept of the research and are sufficiently specific and precise to be used as index entries for cataloging.
(b) IDENTIFIERS AND OPEN-ENDED TERMS • Use identifiers for project names, code names, equipment designators, etc. Use open-
ended terms written in descriptor form for those subjects for which no descriptor exists.
(c) COSATI FIELD GROUP - Field and group assignments are to be taken from the 1965 COSATI Subject Category List. Since the ma-
jority of documents are multidisciplinary in nature, the Primary Field/Group assignment(s) will be specific discipline, area of human
endeavor, or type of physical object. The application(s) will be cross-referenced with secondary Field/Group assignments that will follow
the primary posting(s).
18. DISTRIBUTION STATEMENT
Denote releasability to the public or limitation for reasons other than security for example "Release Unlimited." Cite any availability to
the public, with address and price. -
19.8.20. SECURITY CLASSIFICATION
DO NOT submit classified reports to the National Technical Information service.
21. NUMBER OF PAGES
Insert the total number of pages, including this one and unnumbered pages, but exclude distribution list, if any.
22. PRICE
Insert the price set by the National Technical Information Service or the Government Printing Office, if known.
EPA Form 2220-1 (9-73) (Reverie)
------- |