EPA/600/A-97/057
Hydrogen Chloride Cylinder Gas Concentration
Analysis Using U.S. EPA Method 26
Richard C. Shores
Air Pollution Prevention and Control Division
National Risk Management Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
W. Gary Eaton, Eva D, Hardison, Constance V. Wall
Center for Environmental Measurements and Quality Assurance
Research Triangle Institute
Research Triangle Park, NC 27709
ABSTRACT
The EPA's Maximum Achievable Control Technology (MACT) standards will require the
measurement of hydrogen chloride (HC1) at sources. Continuous Emission Monitoring Systems (CEMS) are
expected to be used to ensure compliance with MACT standards. In the fall of 1995, the EPA Office of Air
Quality Planning and Standards (OAQPS) and the EPA National Risk Management Research Laboratory
conducted combustion emission measurement tests to evaluate proposed performance specifications for a
typical HC1 CEMS. To evaluate the relative accuracy of a CEMS versus the current compliance method
(EPA Method 26), the CEMS was first calibrated using HC1 from compressed gas cylinders and then used to
sample HCl-spiked emissions from a natural-gas-fired rotary kiln incinerator. Although at present there is no
National Institute of Science and Technology (NIST) certified reference material for HC1 gas, specialty gas
companies produce and certify compressed gas cylinders of HC1 for use in calibration and span checks of
CEMS.
As part of the quality assurance program for the CEMS study, Research Triangle Institute (RTI)
carried out an independent determination of the concentrations of HC1 in commercial compressed gas
cylinders used for calibration of the CEMS. A study of the potential for losses of HC1 in calibration or
sample delivery lines was also conducted. Results are reported in this paper.
INTRODUCTION
In the fall of 1995, the EPA Office of Air Quality Planning and Standards and the EPA National
Risk Management Research Laboratory conducted combustion emission measurement tests to evaluate
proposed performance specifications for a typical HC1 CEMS. Emissions were produced by operating the
EPA Rotary Kiln Incinerator Simulator (RKIS) and spiking its emissions with anhydrous HC1 to give target
HC1 levels from zero to approximately 100 ppm1. The CEMS sampled the RKIS continuously; several
Method 262 trains sampled emissions periodically. HC1 concentration values from the CEMS and the wet
chemistry Method 26 are to be compared to assess relative accuracy. A schematic diagram of the RKIS is
given in Figure 1. Sampling points for the CEMS and Method 26 sampling trains are identified. Acurex, Inc.
operated the RKIS and the CEMS. Energy and Environmental Research Corporation operated the Method 26
sampling trains and arranged for Method 26 impinger sample analysis for chloride by ton chromatography
(1C).
-------
RTI provided quality assurance services for these tests. Goals of the quality assurance program
were: (1) carry out a systems audit of the sample collection/handling and documentation activities for CEMS
and Method 26 analyses; (2) provide performance audit test solutions to the study's analytical laboratory for
1C analysis of chloride ions; (3) devise and apply a wet chemistry method to independently determine the
concentration of HC1 in the compressed gas cylinders used for calibration of the CEMS; (4) determine if line
losses occur during delivery of the calibration gases to the CEMS; and (5) compare the results of the cylinder
gas analyses to the manufacturer's stated concentrations.
This paper discusses the methodology used to determine the HCI concentration in the compressed
gas cylinders and to assess the potential for HCI losses in the gas calibration and sample delivery lines. The
differences between the cylinder gas manufacturer's stated values and the values found by analysis of
impinger samples for chloride ions are compared in terms of individual cylinder concentration levels and in
terms of regression relationships for the entire set of cylinders.
METHODOLOGY
Sampling from Cylinders
An abbreviated version of the Method 26 sampling train was selected to capture gases from the
calibration cylinders for an assay. The sampling train arrangement is shown in Figure 2. The first and
second midget impingers contained 15 mL of 0.1 N sulfuric acid (H2SO4) absorbing solution; the third and
fourth impingers contained 0.1 N sodium hydroxide (NaOH) in water, and the fifth impinger contained silica
gel to dry the air prior to volume measurement by the dry gas meter. The bases of the impingers were
immersed in an ice bath during sampling. A stainless steel needle valve, in line between the cylinder regulator
and the impingers, was used to adjust the flow rate. The flow rate of cylinder gas was determined with a
NIST-traceable soap film flowmeter and set to a known flow rate between 1 and 2 liters per minute. The total
volume sampled was indicated by the displacement indicated by the calibrated dry gas meter.
The calibration cylinder gases were sampled at two locations. The first sampling point was within 2
feet (0.6 meter) of the cylinder regulator. The second sampling point was at the end of the approximately 75-
foot (23-meter) long Teflon supply line which connected the cylinder to the CEMS inlet. Stainless steel
regulators (high purity gas type) were attached to the cylinders. These regulators were dedicated and not
removed from the cylinders. The regulators were first purged with cylinder gas to flush out any air or
moisture. Then, while a low flow of gas was maintained, a stainless steel needle valve was connected to the
outlet of the regulator. Teflon lines and connections were used throughout; open lines were capped to prevent
entrance of moisture. The needle valve was purged for several minutes, and the flow through the valve was
adjusted to approximately 2 liters per minute using a soap film flowmeter to confirm the actual flow. The
outlet from the needle valve was connected, again via Teflon tubing and fittings, to the inlet of the first
impinger of the abbreviated Method 26 sampling train.
The contents of each impinger were transferred quantitatively to a separate 100-mL volumetric flask,
diluted to the mark with deionized water, and the flask's contents were poured into separate pre-washed, 125-
mL capacity, high-density polyethylene bottles. The labeled bottles were refrigerated at approximately 4
degrees Celsius until 1C analysis for chloride began.
Analysis by Ion Chromatography
Impinger solutions were analyzed for chloride ion by 1C by both RTI (as the QA laboratory) and
Energy and Environmental Research Corporation (project's analytical laboratory). The laboratories both
-------
used a Dioncx DX300 ion chromatograph with a conductivity detector, sodium carbonate/sodium bicarbonate
eluent solution, and a 5-point calibration curve. However, the analysis ranges of the two laboratories differed
significantly. The QA audit laboratory operated on a fixed zero to 2 ^g/mL range and diluted all of the
samples with eluent solution to adjust the chloride ion concentration to within this range. Eluent was used for
dilution to minimize the "water dip" and other matrix effects which affect peak shape and thus quantitation.
On the other hand, the project's analytical laboratory calibrated over a much broader range, zero to 100
^g/mL, and analyzed samples without dilution. Quality control and quality assurance samples were analyzed
by both laboratories.
RESULTS AND DISCUSSION
HC1 Cylinder Gas Concentrations
Concentrations of HCl in six compressed gas cylinders were determined by 1C analysis of Method 26
impinger solutions for chloride ions by the audit and contract laboratories. Results are given in Table 1. For
individual cylinders, the best agreement between the manufacturer's concentrations and those determined by
either laboratory is for cylinders containing stated concentrations of 8.66,20.5, 38.2, and 79.0 ppm. Seven of
the eight analytical results were within 10% of the stated value and four of the eight results were within the
manufacturer's accuracy range of +/- 5%, The largest differences were found for the two lower concentration
cylinders (stated concentrations of 2.49 and 5.47 ppm); differences ranged from -9.6% to 28.5%. Due to the
sampling protocol, impingers which collected HCl from these cylinders gave a less concentrated solution,
with the result that the lower extreme of the 1C calibration range (less than 3% of full scale in three of four
cases) was used for analysis. Since matrix effects from the water and 0.1N to 0.03N sulfuric acid were
present, the chloride ion area or peak height was more difficult to determine accurately at the lower
concentrations. Figure 3 illustrates the 1C chloride ion peak and the effect of the solution matrix on the
baseline; Figure 4 shows the peak for higher concentrations.
The results of the cylinder analyses were also considered as a group by examining the fit of the linear
regression of the manufacturer's concentration values versus those found by the laboratories. At the 95%
confidence level, and using Student's t criteria, the slope of the regression equation for either set of laboratory
data was found to be significantly different from a slope of 1.00. However, the slopes of the results were not
significantly different from a slope of 1.05, a value representing the vendor's stated accuracy of •+•/- 5%.
Sampling Line Study
To assess the possibility for sampling or calibration line losses, calibration gas samples from
cylinders were taken by Method 26 immediately downstream of the cylinder regulator and at the exit end of
the approximately 75-foot (23-meter) long delivery tube, just prior to the CEMS inlet. The study of the
lowest concentration cylinder (designated value 2.49 ppm) showed an apparent increase in concentration;
however, the analytical results were so variable as to be uninterpretable. Results from studying the cylinder
with a stated concentration of 38.2 ppm showed a diminishment of only 2.7%. Since this is well within the
precision of the collection/measurement method at this concentration, it appears that no detectable line losses
occurred.
Quality Assurance Samples
Each laboratory ensured that its analytical system was in control by analyzing control samples,
blanks, and spikes at approximately 20-sample intervals. Two quality assurance solutions were also analyzed
twice as blind samples. The first was a water-diluted acid precipitation standard with a chloride
concentration of 2.31 ^g/mL. This solution was suitable for the QA laboratory's range of zero to 2
-------
(agreement within 1%) but was unsuitable as an audit solution for the project's analytical laboratory which
operated over a zero to 100 Mg/mL range (high by 0.9 yug/mL or 40%.)- The second solution was 12.0
/ug/mL of sodium chloride in 0.03N sulfurie acid. For analysis of this sample, the QA laboratory was high by
15%; the analytical laboratory was high by 12%.
CONCLUSIONS
Sampling and analysis difficulties were encountered in applying Method 26 to the analysis of gases
in cylinders containing 5 ppm HC1 or less. However, for concentrations in the range of 8 to 100 ppm,
independent analysis gave results that were within 10% (and often within 5%) of the manufacturer's
designated concentration. The results from this short quality assurance study point out the good potential for
reliable production and use of compressed gas standards of HC1 in a nitrogen matrix. To achieve higher
confidence in compressed gas standards for HC1 would require a program of development, multi-instrument
analyses, and round-robin testing such as is used to certify NIST or EPA standard reference or research grade
materials. In addition, explicit instructions would be required for the proper care and use of the HC1 cylinder
standards.
REFERENCES
1. Quality Assurance Test Plan. "Support of OAQPS HC1 Monitoring Standards." Prepared by Acurex, Inc.
for the U.S. Environmental Protection Agency; National Risk Management Research Laboratory, Research
Triangle Park, NC. Document No, QTRAKCR06,10/n. August 1995.
2, Code of Federal Regulations, Part 60, Appendix A. "Method 26-Determination of Hydrogen Chloride
Emissions from Stationary Sources." Promulgated April 22,1994.
ACKNOWLEDGEMENTS
The cooperation of employees of Energy and Environmental Research Corporation and Acurex, Inc.
during the conduct of the quality assurance studies is appreciated and acknowledged. Joe McSorley, now
retired from the Agency, was the OAQPS Project Officer for the project. The research described in this paper
was funded wholly or in part by the U.S. Environmental Protection Agency through contract No. 68-D3-0045
with Research Triangle Institute,
-------
Table 1. HC1 Cylinder Gas Concentrations as Determined by EPA Method 26
Cylinder Number
A020112
A0200100
A020130
XA-1658
A020015
A01980
Manufacturer's Stated
Concentration,
ppm (a)
2.49
5.47
8.66
20.5
38.2
79.0
QA Audit Laboratory
Results, ppm
(% difference) (b)
3.20 (28.5)
6.32 (15.5)
8.96 (3.5)
21.3 (3.9)
40.7 (6.5)
84.2 (6.4)
Project Laboratory
Results, ppm
(% difference)
2.25 (-9.6)
6.68 (22.1)
9.42 (8.7)
20.0 (-2.4)
39.0(2.1)
88.9 (12.5)
(a) Certified working standard. Manufacturer's stated analytical accuracy +/- 5%.
(b) % difference = 100 x (laboratory result - manufacturer's value)/(manufacturer's value)
Table 2. Comparison of HC1 Cylinder Gas Concentrations as Determined at the Cylinder and at the Entrance
to the CEMS (Line Loss Study)
Cylinder Number
(Manufacturer's
Value, ppm)
A020112
(2.49)
A020015
(38.2)
Concentration, ppm, as
Found at the Cylinder
3.20
40.7
Concentration, ppm, as
Found at the CEMS
Inlet
4.35
39.6
Percent Change from
Cylinder to CEMS Wet
35.9
-2.7
-------
HCI Injection
Secondary Combustion Chamber / /
Reference
5 Plug Flow Section
Stack Transition Section
Afterburner
Mixing Chamber
Rotary Leaf
Spring Seal
View Port
Main Burner Tt
Kiln Section Transition Section
Numbers = Sampling Points
Letters = Injection Points
Figure 1. Schematic of Rotary Kiln Incinerator
HCI
Cylinder
Needle
Valve
0.1 N H2SO4 0.1 N NaOH Silica Gel
Figure 2. EPA Method 26 Sampling Train Modified to Collect Cylinder Gases
-------
8.0
7.0
6.0
5.0
% 4.0
| 3.0
8 2.0
cc
1.0
0.0
-1.0
-2.0
Chloride
0
2 3
Time (minutes)
Figure 3. Ion Chromatograph Trace for Chloride Ion Showing Matrix Effects
40
30
Response
-*• ro
0 0
0
C
Chloride
A
1
V
\ 1 11 1
) 12345
Time (minutes)
Figure 4. Ion Chromatograph Trace for Chloride Ion at Higher Concentrations
-------
NRMRL-RTP-P-214
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before eompletin
1. RE
3. R
4. TITLE AND SUBTITLE
. REPORT DATE
Hydrogen Chloride Cylinder Gas Concentration
Analysis Using U.S. EPA Method 26
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
R. C. Shores (EPA, NRMRL-RTP), and W. C. Eaton,
E. D. Hardison, and C. V. Wall (RTI)
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
10, PROGRAM ELEMENT NO.
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, North Carolina 27709
11. CONTRACT/GRANT NO.
68-D3-0045
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Air Pollution Prevention and Control Division
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Published paper; 9/95-5/97
14. SPONSORING AGENCY CODE
EPA/6QO/13
*SUW-EMENTABY NOTES APPcD project officer is Richard C. Shores, Msdl Drop 91, 919/
541-4983. For presentation at Measurement of Toxic and Related Air Pollutants,
RTP.NC, 4/29-5/1/97.
16-ABSTRACT The paper discusses hydrogen chloride (HC1) cylinder gas concentration
analysis using U.S. EPA Method 26. (NOTE: EPA's maximum achievable control
technology (MACT) standards will require the measurement of HC1 at sources. Con-
tinuous emission monitoring systems (CEMS) are expected to be used to ensure corn-
compliance with MACT standards. In the fall of 1995, the EPA Office of Air Quality
Planning and Standards and the EPA National Risk Management Research Laboratory
conducted combustion emission measurement tests to evaluate proposed performance
specifications for a typical CEMS.) To evaluate the relative accuracy of a CEMS
versus the current compliance method (EPA Method 26), the CEMS was first calibra-
ted using HC1 from compressed gas cylinders and then used to sample HC1-spiked
emissions from a natural-gas-fired rotary kiln incinerator. Although at present
there is no National Institute of Science and Technology (NIST) certified reference
material for HC1 gas, specialty gas companies produce and certify compressed gas
cylinders of HC1 for use in calibration and span checks of CEMS. As part of the
quality assurance program for the CEMS study, RTI carried out an independent de-
termination of the concentrations of HC1 in commercial compressed gas cylinders
used for calibration of the CEMS. Results are reported in the paper.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution Monitors
Hydrogen Chloride
Gas Cylinders
Concentration (Composition)
Measurement
Emission
Pollution Control
Stationary Sources
EPA Method 26
Continuous Emission
Monitors (CEMs)
13 B
07B
13 D
07D
14G
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report/
Unclassified
21. NO. OF PAGES
20. SECURITY CLASS (This page}
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
------- |