United States
Environmental Protection
Agency
Atmospheric Research and ^
Exposure Assessment Laboratory - ;
Research Triangle Park NC 2771 1 '/
, t
Research and Development
EPA/600/S3-89/064 Aug. 1989
v>EPA Project Summary
Laboratory and Field
Evaluation of a
Methodology for
Determination of Hydrogen
Chloride Emissions from
Municipal and Hazardous
Waste Incinerators
S.C. Steinsberger and J.H. Margeson
Laboratory and field studies were
performed to develop and evaluate a
sampling and analytical technique for
measuring hydrogen chloride (HCI)
from stationary sources. Studies
were conducted in three phases: (1)
literature search and development of
a candidate sampling and analysis
protocol, (2) laboratory evaluation
and refinement of the protocol, and
(3) field evaluation. A modified
Method 6 sampling train was selected
for sample collection due to its ease
of operation, availability, and cost An
acidified water absorbing solution
was identified for collecting HCI in
the impingers. The acidified water
solution was selected to minimize the
potential for diatomic chlorine (C12)
to interfere with the HCI
determination. Ion chromatography
was selected as the most suitable
technique for the analysis of HCI. The
laboratory phase evaluated the HCI
collection efficiency of the sampling
protocol and the distribution of C12 in
the sampling train. A preliminary field
test was included in the laboratory
phase to indicate any further protocol
modifications. A ruggedness test was
designed to evaluate the effect of six
variables that may be encountered
when employing the sampling
protocol. A field evaluation was
conducted to determine the precision
and estimate the accuracy of the
sampling and analytical protocol. The
candidate method was also employed
to determine the bias and precision
of two HCI continuous emission
monitoring systems.
This Project Summary was
developed by EPA's Atmospheric
Research and Exposure Assessment
Laboratory, Research Triangle Park,
NC, to announce key findings of the
research project that is fully
documented in a separate report of
the same title (see Project Report
ordering information at back).
Introduction
The United States Environmental
Protection Agency (EPA) is currently
regulating emissions of HCI from
hazardous waste incinerators under 40
CFR 264.343 to 4 Ibs/hr or an HCI
removal efficiency of at least 99%. The
EPA is also currently considering regu-
lating HCI emissions from municipal
waste combustors (MWC's). Several state
and local agencies have already set HCI
emission limits for new MWC's and are
requiring installation of HCI continuous
emission monitoring systems (HCI
CEMS's) at certain facilities.
To support current and future
regulations on HCI emissions, a sampling
-------
and analysis method evaluation study
was conducted for the Quality Assurance
Division of EPA's Atmospheric Research
and Exposure Assessment Laboratory.
This method, designed to measure
hydrogen chloride emissions from
stationary sources, was developed and
then evaluated both in the laboratory and
in the field. Details of the evaluations are
presented including (1) laboratory
assessment of the sampling protocol
employing gas mixtures of HCI and C12,
the effect of variations in the sampling
train and technique, and the impact of
possible analytical interferents; and (2)
results of co-located duplicate and
quadruplicate-train sampling and con-
tinuous emissions monitoring at two
municipal waste incinerators.
The study was conducted in several
phases. The initial phase involved a
literature search which formed the basis
for development of the sampling and
analytical protocol. The procedures
chosen are outlined in the next section
and discussed in detail in the draft
method for measurement of HCI in
stationary source emissions, written in
Federal Register format, provided in
Appendix A of the report. The remaining
two phases consisted of (1) an initial
laboratory evaluation, including the
collection and analysis of preliminary
field samples from a stationary source
and a six-variable, one-blank rugged-
ness test; and (2) a final field evaluation
which included comparison of values
from the candidate protocol with
continuous emissions monitoring values.
Conclusions and recommendations are
made regarding the application, preci-
sion, and accuracy of the proposed
method.
Procedures
The sampling and analytical protocol
evaluated in this laboratory and field
study was proposed based on a thorough
literature search. Candidate sampling
methods, absorbing solutions, and
analytical methods, as well as potential
interferents were reviewed. A modifica-
tion of the EPA Method 6 protocol was
chosen for sampling, and ion
chromatography was selected for analy-
sis based on (1) demonstrated speciation
of HCI and C12 and (2) accuracy of the
analytical technique, respectively, and
secondarily, the availability of equipment,
and universality of sampling and analy-
tical techniques (see Figure 1). Two
impingers containing a dilute sulfuric acid
solution (0.1 N H2S04) are followed by
one impinger containing a dilute caustic
solution (0.1 N NaOH) to provide high
HCI collection efficiency while minimizing
C12 interference.
In the first phase of the laboratory
evaluation, the sampling trains were
challenged with various concentrations of
HCI and C12 at different flow rates. The
ability of the absorbing solution to
efficiently collect and speciate one gas in
the presence of low (zero) to high levels
of the other was evaluated. The effect of
flow rate on the absorption capacity for
C12 in the acidic impinger solution was
also examined. All impinger samples
were analyzed separately by ion chroma-
tography. The concentration of the cylin-
der gases used were independently veri-
fied prior to the testing.
The preliminary field test was
conducted primarily to identify any
potential problems that might occur with
the sampling and/or analytical methods
when used at a typical HCI emission
source. The samples were taken down-
stream of acid gas and particulate control
equipment at a MWC where an HCI
continuous emission monitor was oper-
ating concurrently. Dual-train sampling
was utilized during the testing to identify
the effect, if any, of using stainless steel
versus glass probe tips. Comparison of
HCI tram values with the HCI GEMS
values provided information concerning
the proposed method's ability to follow
trends in HCI effluent levels.
After completion of the initial laboratory
and field studies, a ruggedness test was
developed to assess the effect on the
method of selected variables that may
affect actual sampling. The variables, or
deviations from standard procedure,
chosen for evaluation were low reagent
volume, increased impinger pH, longer
sampling time, elevated impinger tem-
peratures, higher sampling rate, and
elevated C12 levels. These six variables
plus control blank were combined in an
eight-run duplicate sample train test
matrix, which allowed the necessary
computations to identify which variable(s)
had a significant effect on the results.
The final phase of the method
evaluation consisted of a field test at a
MWC. The objectives of the test included
determination of the precision and
accuracy of the draft HCI protocol and
the bias and precision of HCI CEMS's. A
TECO HCI CEMS and a Bran and
Luebbe HCI CEMS were installed at the
MWC downstream of a lime-slurry
spray dryer and a three-field ESP. The
bias of the CEMS's and the precision of
the protocol were obtained concurrently
by conducting relative error test runs
using paired sampling trains. The
accuracy of the combined sampling and
analysis protocol was estimated emplo
ing 30-minute test runs consisting
dynamic spiking of the sampling trail
with HCI cylinder gas. The concentratk
of the HCI gas cylinders were determine
by independent analysis before and aft
the field test. Two additional relate
experiments were conducted
determine the amount of flue gas C(
absorbed by the alkaline imping
reagent and to compare the HCI resu
from the draft HCI protocol to tho:
obtained using a Method 5-tyf
sampling train employing an alkalii
impinger reagent.
Results and Discussion
The HCI collection efficiency in the fii
acidified midget impinger averaged 105
percent for a 442 parts per million (ppr
HCI gas mixture sampled at 2 liters p
minute (Ipm), with the second acidifii
impinger collecting only 0.4 percent. F
a gas mixture of 221 ppm HCI and 1!
ppm C12 sampled at 2 1pm, the H
collection efficiency for the first acidifii
impinger averaged 103.0 percent, wi
the second impinger collecting 3
percent. For a 393 ppm C12 gas mixtu
sampled at 2 1pm, the C12 collectii
efficiency of the first alkaline imping
averaged 88.2 percent, with each of tl
two acidified impingers collecting 0
percent. For the same gas mixtu
sampled at 0.5 1pm, the first tv
acidified impingers collected an avera
of 3.2 percent and 2.9 percer
respectively, with the first alkalii
impinger collecting 76.0 percent.
There does not appear to be ,
interaction between HCI and C
affecting either the HCI collectii
efficiency or the retention of C12 by tl
acidified impingers. The sample flow n
appears to affect the distribution of C
throughout the train with a higher fl<
rate reducing the amount of C12 retain
in the acidified impingers. A higher fl<
rate does not appear to reduce the H
collection efficiency at the levels teste
Based on these observations, tl
acidified midget impinger sampling tra
operated at a sampling rate of 2 1p
appears to minimize the high H
measurement bias caused by C12 to le
than 5% for the conditions tested.
The preliminary field test indicated tt
both stainless steel and glass probe ti
could be used for HCI sampling. The I-
emission trends indicated by an install
HCI CEMS were reflected by the resu
of the manual sampling. The relativt
high moisture level at the sour
combined with extended sampling tim
resulted in the first impinger becomi
-------
Mr TEFLON TUBE
MTTEFLON TUBE
(OPTIONAL)
TAMLEM
STEEL
FTTTMa
(OPTIONAL)
GLASS UNER WRAPPED
WITH HEAT TAPE
TEFLON FLTER LOCATION
(ten «M* Mnpmtura > 300- F)
(H0un>1C)
VtotfriB (figura IB)
Purging (FjgtnlA)
NEEDLE
VALVE
Figure 1.
HCI Sampling Train
full of condensed flue gas moisture. A
water knockout impinger was incorpora-
ted into the sampling train for the field
evaluation test.
The ruggedness test was used to
assess the sensitivity of the method to
selected variables which may affect sam-
pling. The results showed percent
differences for the six variables of less
than ± 2.5%, indicating that the method
was insensitive to the selected variables:
low reagent volume, increased impinger
pH, longer sampling time, elevated
impinger temperature, higher sampling
rate, and elevated C\2 levels. These
results, in conjunction with the earlier
laboratory evaluation, indicated that at
Gig levels up to 50 ppm, the measure-
ment of HCI is not biased significantly.
The field test involved paired midget
impinger train sampling using the
sampling train shown in Figure 1. As
indicated in Figures 2 and 3, flue gas HCI
levels determined by the manual method
were in good agreement with the levels
indicated by the TECO HCI CEMS. The
Bran and Luebbe HCI CEMS was able to
follow the changes in the HCI flue gas
levels, but was biased low by
ipproximately 60 percent (4 ppm). The
SUFK3E
TANK
specific results of the field test are as
follows:
- The average precision (expressed as
the relative standard deviation) of the
HCI sampling and analysis protocol
was 6.2% at an average flue gas HCI
concentration of 3.9 ppm and 3.2%
at an average concentration of 15.3
ppm. The average relative standard
deviation for the moisture determina-
tion employing the midget impinger
train was 4.5% and 3.2%, respec-
tively, at the same concentrations.
- The average relative error of the HCI
sampling and analysis protocol,
established by dynamic spiking, was
5.5% and 7.1% for HCI gas mixtures
of 9.7 and 34.3 ppm, respectively
- The relative errors and biases
relative to the manual HCI method
for the TECC-R HCI CEMS were
1.6% and 6.8%, and 0.07 ±079
ppm and 0.68 ± 1.58 ppm, at
average flue gas HCI levels of 3.9
and 9.9 ppm, respectively.
- The relative errors and biases
relative to the manual HCI method
for the Bran and LuebbeR CEMS
were 69% and 58%, and -2.66 ±
0.90 ppm and -5.7 ± 2.35 ppm, at
average flue gas HCI levels of 3.9
and 9.9, respectively.
- The precisions (standard deviations)
for the TECO CEMS were 0.75 ppm
and 1.50 ppm at average flue gas
HCI levels of 3.9 and 9.9 ppm,
respectively. The precisions (stan-
dard deviations) for the Bran and
Luebbe CEMS were 0.87 ppm and
2.30 ppm at the same flue gas HCI
levels.
- Flue gas CC>2 absorbed by alkaline
impinger reagents was not found to
be significant in either the midget
impinger train and the Method 5-
type tram.
- The midget impinger tram and the
Method 5-type tram produced sim-
ilar HCI results at a flue gas HCI
concentration of 21 2 ppm. However.
the Method 5-type train produced
significantly lower HCI results than
the midget impinger train at a flue
gas concentration of 4.8 ppm. The
low bias may have been a result of
unreacted lime collected on the filter
or the glass-fiber filter itself absorb-
ing gaseous HCI from the sample
-------
HC1 OUTLET CONCENTRATIONS - 9/15/88
Wheelabrator Millbury - Unit 2
TJ
O
X
E
0.
OL
iu -
9 -
8 -
7 -
6-
5 -
4 -
3-
2 -
1 -
LEGEND
TECO
Bran it Luebbe
Impinger Results
r\
/"\ r^\ 1 \
/ \ / NN-. ' \
\ . / X"
.... , \»' ~^ /
S/V i-x
N^'"~^v / V / V
10:15 11:15 12:15 13:15 14:15 15:15 16:15 17:15 18:
Clock Time
Figure 2 Flue Gas HCt trends indicated by HCI CEMS's under normal acid gas conditions
Conclusions and
Recommendations
A midget impinger train employing an
acidified impinger reagent and operated
at a sampling rate of 2 1pm provides
acceptable HCI collection efficiency at
HCI levels up to 500 ppm and is not
susceptible to significant C\2 interfer-
ence at C12 levels less than 50 ppm.
The method, as described, may also be
suitable for determining C12 emissions.
The method is insensitive to slight
changes in reagent volume, impinger pH,
sampling time, impinger temperatures,
and sampling rate that may occur during
actual use.
The precision and bias demonstrated
for the HCI method are acceptable, and
the method can also be used for
moisture determination. The agreement
between the manual method and the
TECO HCI CEMS, calibrated with HCI
cylinder gases, was acceptable at rela-
tively low flue gas HCI levels.
A nozzle oriented opposite the gas
flow and a Teflon filter can be used with
the manual method probe assembly to
avoid collection of particulate matter and
loss of gaseous HCt through reaction
with glass surfaces and alkaline particu-
late matter. A glass wool plug or a glass
fiber filter should not be used to prevent
particulate matter from entering the tram,
since this will increase loss of HCI due to
reaction with alkaline particulate matter.
A 1-hour sampling time is recom-
mended to decrease any bias introduced
by the reaction of HCI with glass
surfaces and alkaline particulate matter.
-------
HC1 OUTLET CONCENTRATIONS - 9/16/88
Wheelabrator Millbury - Unit 2
*?
TJ
O
e
a.
a.
45 -
40 -
« _
30 -
25 -
20 -
15-
10 -
5 -
o -
LEGEND
TCCO
\ Bran & Luebbe
; « Impinger Results
#A
.' I '
' ':
/' \\
/' \\
/.' \\
'i \\ r\
/' \\ / \
.' \\
* i * « / \
' i * ** »
/' \\ .-. / , *"-'""\
," \ vv^. --:---' V ' v \
j: \ ""' / \
^ ^'^-' V..r x--
--'' ""--_
11:15 12:15 13:15 14:15 15:15 16:15 17:15 18:15
Clock Time
Figure 3. Flue Gas trends indicated by HCI CSMS's under elevated acid gas conditions
-------
-------
S. C. Steinsberger is with Entropy Environmentalists, Inc., Research Triangle
Park, NC 27709. J. H. Margeson (also the EPA Project Officer, see below)
is with the U.S. Environmental Protection Agency, Research Triangle Park,
NC 27711.
The complete report, entitled "Laboratory and Field Evaluation of a Methodology
for Determination of Hydrogen Chloride Emissions from Municipal and
Hazardous Waste Incinerators," (Order No. PB 89-220 5861 AS; Cost:
$15.95, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Atmospheric Research and Exposure Assessment Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
EPA/600/S3-89/064
0000853
------- |