United States
               Environmental Protection
               Agency
 Atmospheric Research and
 Exposure Assessment Laboratory
 Research Triangle Park, NC 27711
               Research and Development
 EPA/600/SR-92/208  March 1993
EPA       Project  Summary

               Volatile Organic Sampling
               Train— Soot  Interference  Report
               Gary D. Hinshaw, Frank W. Wilshire, and Larry D. Johnson
                 Boiler  performance studies spon-
               sored by the U.S. Environmental Pro-
               tection Agency (EPA) have shown that
               industrial  boilers co-firing hazardous
               waste are capable of achieving destruc-
               tion and removal efficiencies (DREs) of
               99.99 % for Resource Conservation and
               Recovery  Act  (RCRA) Appendix VIII
               compounds designated as principal or-
               ganic hazardous constituents.  Evalua-
               tion of boiler DREs for volatile  organic
               compounds (VOCs) is accomplished in
               part by sampling those compounds as
               described in EPA Method 0030, the
               Volatile Organic Sampling Train  (VOST).
               A hysteresis effect, resulting from soot
               deposits on boiler surfaces, has been
               observed in previous studies and is
               the subject of several related reports.
               Sorption behavior of organic com-
               pounds on deposits of soot in  boilers,
               as demonstrated  by these hysteresis
               studies, has led to concern that similar
               sorption effects from soot captured by
               the paniculate filter of the VOST might
               cause a low bias in measurement re-
               sults. It seemed prudent,  then, to de-
               termine if such an effect is measurable,
               to investigate the magnitude of  any ob-
               served effect on VOST sampling, and,
               as appropriate,  to also determine
               whether the effect is greater for higher
               boiling point VOCs as expected.
                This study examined the effects of
               soot deposits within the sampling ap-
               paratus upon  levels of VOCs recov-
               ered. Specifically, collection efficiency
               was evaluated by comparing VOC lev-
               els collected in a control VOST,  when a
               soot-free  paniculate filter (i.e., glass
wool plug) was used In the sampling
probe, to VOC levels recovered while
using a soot-laden paniculate filter in
the VOST probe. Samples were col-
lected at routine VOST sampling rates
and analyzed by gas chromatography
and mass spectrometry (GC/MS) as de-
scribed in EPA Methods 0030, 5040,
and 8240. Because of time and  cost
considerations, the soot was generated
only from No. 2 fuel oil,  as opposed to
that produced from hazardous waste.
Soot deposition on the filters varied
from 4 to 70 mg.
  Volatile organic compounds evalu-
ated were selected to span the range
of boiling points covered by VOST
methodology. Emphasis was  directed
to substances  in the upper range  of
VOC boiling points (i.e., 120 to 130 °C),
because sorption phenomena often cor-
relate with boiling points, and thus the
higher  boiling point compounds were
deemed more likely to be seriously af-
fected.
  The data collected were subjected to
descriptive  and pattern  recognition
techniques (i.e., percent difference and
rotated principal component analysis).
Results of these statistical analyses in-
dicated that at least two of the higher
boiling point VOCs appear to be nega-
tively biased  by the presence of soot
on the glass wool filter.  However, it is
not likely that these findings would have
a major impact on previously collected
hazardous waste incinerator data, since
the VOST-Soot  effect was determined
at moderately high soot loadings, which
are atypical of properly operating haz-
                                                              Printed on Recycled Paper

-------
ardous waste Incinerators. Also, a level
of sooting where the effect ceases to
cause significant  deterioration of  the
data was not established In this study.
Consequently, the data gathered In this
study was not totally conclusive In de-
termining  whether  soot may Impact
ORE determinations from boilers using
the VOST methodology.
   This Project Summary was devel-
oped by EPA's Atmospheric Research
and Exposure Assessment Laboratory,
Research  Triangle Park, NC, to  an-
nounce 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
   EPA-sponsored studies of boiler perfor-
mance under both normal and transient
conditions  have  shown that the boilers
can  achieve a destruction and removal
efficiency  (ORE) of at  least 99.99% for
Resource Conservation and Recovery Act
(RCRA)  Appendix VIII  compounds
(40CFR261) that are principal organic
hazardous constituents  (POHCs). How-
ever, the ORE measurements made  dur-
ing some of the earlier EPA studies could
have been affected  by the presence of
deposits of unburned, carbonaceous ma-
terial (i.e.,  soot)  either (1) on cold boiler
surfaces or (2) within the organic sam-
pling system. The first concern  deals with
a phenomenon referred to as hysteresis.
The second concern, however, is the pri-
mary focus of this project. Specifically, the
quality of data on volatile  organic concen-
trations in  boiler stack gases may be im-
pacted by the presence of soot if the soot
Is trapped on the glass  wool  plug  (i.e.,
particulate filter) typically used in the probe
of a volatile organic sampling train (VOST)
and if the soot functions as  an active
sorbent. If sufficient effect is present, the
current VOST sampling technique may not
be appropriate for certain boiler applica-
tions.
 Experimental Procedure
   The primary objective of this project was
to determine if combustion-generated soot
 will sorb  a sufficient quantity  of  volatile
 organic compounds  (VOCs) to cause an
 erroneously low measurement of VOCs
 by the  VOST method. The type of  soot,
 that is, its source and conditions of forma-
 tion, may affect its propensity to  sorb
 VOCs.  Thus, soot generated by actual
 mixtures of hazardous waste and fuel may
 differ from soot generated by hydrocarbon
 fuels alone in its effect upon VOST  mea-
surements. However, for the initial "range-
finding" experiments described herein, soot
was generated in only one way, by No. 2
fuel oil combustion only, without any haz-
ardous waste  constituents present  (i.e.,
no RCRA Appendix VIII compounds).
  These  experiments followed a two-step
format. In the first step, the soot deposi-
tion step, the soot was collected upon the
glass wool plug of a VOST system probe
by withdrawing soot-laden gas through the
probe of  the experimental  VOST system
at a flow rate of 0.5 L/min. (The mass of
soot deposited was controlled by adjust-
ing the deposition time, ranging from 6 to
9 min for Experiments 14-16 and from 10
to 18 min for Experiments  17-19.) During
the second  step, which  was sampling,
gjass wool doped with soot_was used in
the experimental probes, but a clean glass
wool filter was used in the control probe.
(The control probe was positioned in the
VOST adapter after  the soot deposition
step.)
   A mixture of target VOCs was  intro-
duced downstream of the flame zone fol-
lowing soot deposition, and the test VOST
and a control VOST were operated con-
currently. The  target  VOCs are shown in
Table 1.  The VOCs in Table 1 were se-
lected by EPA personnel to span the range
of boiling points covered  by the VOST
method and also to concentrate on the
upper range of boiling  points  (approxi-
mately 120 to 130 °C) associated with the
VOST method. All compounds except oc-
tane  are Appendix VIII compounds that
could be used as POHCs for DRE deter-
mination.
   Sampling procedures   followed  EPA
Method 0030, sample preparation proce-
dures followed Method 5040 (thermal de-
sorption),  and  analysis procedures
incorporating gas  chromatography/mass
spectrometry  (GC/MS) followed Method
 8240. In 'the  next phase' of the' experi-
 ment, a  well-mixed, diluted VOC mixture
was injected directly into the VOST probes,
 instead of being injected into the middle
of the quartz flow tube of the soot genera-
tor, as was the case during Phase 2.

 Table 1.  Volatile Organic Compounds Injected
 During Phase 2 Testing
Compound Name
Chloroform
Carbon tetrachloride
Toluene
Tetrachloroethene
Octane
Chlorobenzene
1, 1, 1,2-Tetrachloroethane
Boiling Point fC)
61
77
111
121
126
131
131
  A mixing chamber was designed and
installed to improve the mixing of the VOC
mixture (from the diluent bag) with air. A
VOC mixture was introduced at the inlet
of the mixing chamber, and air was intro-
duced at the outlet of the chamber  at
near-sonic velocity.  The entire chamber
and the  inlet and outlet lines were electri-
cally heated. The combined  mixture, di-
luted with air and well mixed, then entered
the VOST adapter through a small tube
that passed through a port. The gases
then passed into a  large stationary sec-
tion of glass tubing (electrically heated),
into which the VOST probes were inserted.
A sliding section of slightly larger tubing
(also electrically heated) was used to help
contain the gases and prevent contamina-
tion by  residual organic  material in the
combustion system.' A nominal flow of 6 L/
min (dry,  standard  basis) was used,  of
which only 1 L/min was withdrawn by the
two VOST systems during the sampling
step of each experiment.
  As  in Phase 2, only one type of soot
was tested during  Phase 3; that is, only
soot generated from No. 2 fuel oil was
used. To  deposit soot on the filter, the
soot-laden gas was withdrawn through the
probe of the experimental VOST system
at a flow rate of 0.5 L/min; the mass of
soot deposited  was controlled by adjust-
ing the deposition time (ranging from 6 to
9 min for  Experiments 14 to 16 [filter
type B]  and  from 10 to 18 min for Experi-
ments 17 to 19 [filter type A]). The deposi-
tion time was limited by the ability of the
sampling pump to accommodate the pres-
sure drop across the filter. Baseline ex-
periments, in which no soot was generated,
were run for comparison to experiments
involving soot generation. The soot mass
deposited on the glass wool plugs ranged
from 3.5 to 70 mg.

 Results andJ)is.cusslon^_,,..

 Analytical and Quality Control
 Results
   All VOST samples were analyzed within
 2 weeks after collection. Percent differ-
 ences between the control and experi-
 mental  VOST samples were calculated to
 determine if any bias between the two
 VOST systems existed.
   Quality assurance/quality  control (QA/
 QC)  results for Phase 3 were generally
 very  good. A  performance audit (PA)
 sample was prepared by the Midwest Re-
 search  Institute (MRI) Quality Assurance
 Department and  submitted  for analysis.
 The PA sample was analyzed by GC/MS
 with  the  same instrumental parameters
 planned for use with the VOST samples.

-------
Four of the eight target compounds were
identified and quantified in the PA sample,
and the  remaining four were not found.
Results were then reported to the project
QA Coordinator prior to analysis of actual
project samples.  All target compounds
present in the PA sample were correctly
identified and were quantified with recov-
eries of 100% to 108%, based  upon the
actual concentrations.
  The percent differences for the injected
VOCs were  less than 10% (with one ex-
ception) in the baseline experiments. This
suggests that the VOC mixture was more
homogeneous in the vicinity of the VOST
probes. The high reproducibility shown by
the two VOSTs allows more confidence in
assessing any soot effects observed.
  The percent differences for both octane
and chlorobenzerie  Indicated a clear  re-
duction with soot on the filter. In addition,
a positive bias was  observed in the case
of benzene, which  was monitored as a
residual  product of incomplete combus-
tion (PIC).

Statistical Analysis
  The concentration data generated in this
phase  of the program were subjected to
rotated  principal  component  analysis
(RPCA).  In brief,  RPCA reduces the  di-
mensionality of the  data set  by creating
composite variables that are  linear func-
tions of the original variables. This reduc-
tion  in dimension  is  based  upon  the
observed correlations between the origi-
nal variables. The correlation coefficients
obtained  by this technique are shown in
Table 2.
  The RPCA results indicate that the mea-
sured variability in concentration  among
the eight  variables in the experiments can
be compactly expressed in terms of three
composite variables:
  • Variable 1 is  principally a  weighted
 •  function of the  compounds  toluene,
    octane, tetrachloroethene, and chlo-
    robenzene; 1,1,1,2-tetrachloroethane
    also figures into this composite, al-
    though  it  is  apparent  that  its
    intercorrelations with other high boil-
    ing point compounds are not as strong
    as those exhibited between toluene,
    octane, tetrachloroethene, and chlo-
    robenzene.
  • Variable 2 is dominated by  benzene,
    although  some contribution from
    1,1,1,2-tetrachloroethane also is indi-
    cated.
  • Variable 3 is  a weighted  function of
    chloroform and carbon tetrachloride.
  RPCA analyses indicated a  tendency
for  low boiling  point compounds  to  be-
have differently from compounds with rela-
tively  high boiling  points. Differences in
standard scores between the control and
experimental trains  are far  more  pro-
nounced for the high boiling  point com-
posite  variable  than for  the  low  boiling
point variable.

Conclusions
  The presence of soot on the glass wool
particulate filter within a VOST probe ap-
pears to have the potential to bias  the
amount of certain VOCs  collected in  the
VOST traps, in particular, octane and chlo-
robenzene. The extent of the bias does
generally correlate with the amount of soot
on the filter. There is some evidence that
the higher boiling  point compounds  are
more likely to be biased by the  presence
of soot,  but other chemical or physical
properties may  also  be  involved.  The
cause of the negative bias is not known
for  certain;  it could be at  least partially
based  upon pressure differences in  the
test VOST console as compared to a con-
trol VOST (due to a soot-based pressure
drop in the probe), in addition to a poten-
tial  sorption of the VOC upon the soot.
Loss of VOCs due to soot on the appara-
tus surfaces was carefully controlled, which
thus clarified the effect of soot in the par-
ticulate filter. Also, a high degree of sam-
pling reproducibility was achieved  in the
experiments,  and  this strengthened  the
above  conclusions.
   In addition to a negative bias  (i.e., loss
of VOCs known to be present in the gas
phase), a positive bias (i.e., desorption of
VOCs  from the soot into the gas phase)
may be indicated so that compounds that
may not always be present in the sampled
gases  may show up in  the VOST  analy-
sis. A positive bias was observed for ben-
zene, which was presumably formed as a
PIC  during soot generation and tempo-
rarily trapped in the soot.  (Alternatively,
the soot may have partially decomposed,
emitting benzene as a by-product.)
   If the presence of soot in the filter of the
VOST  probe does affect the VOCs with
boiling points above 120° C, as this study
indicates, it is highly unlikely that the ef-
fect would have a  major impact on  previ-
ous hazardous waste incinerator data. This
study  examined moderately  heavy soot
loadings, which are not typical of properly
operating  hazardous waste incinerators.
Also, only two  VOCs (octane and chlo-
robenzene) showed a  significant  effect
from the soot loadings. Both of these  VOCs
have boiling points above 120° C, indicat-
ing that  only VOCs with boiling  points
above  that level would be affected.
   Since more than one  POHC is chosen
for ORE determinations and four  nines
(99.99%) must be achieved for all POHCs,
it is not likely that the VOST-Soot  effect
on this one POHC  (chlorobenzene)  would
have any impact on permitting decisions.
However, this study does indicate that cau-
tion should be exercised in future VOST
sampling when potentially sooty conditions
are present or are expected in incinerator
emissions.
Table 2. Pearson Correlation Coefficients Between Test Compounds for Phase 3

CHCI3
CCI4
CSHSCH3
CaH10
CCI2CCI2
C6HSCI
CCI3CHpl
W.
CHCI3
1.000
0.775
0.001
0.001
0.250
0.055
0.518
0.239
CCI4

1.000
0.001
0.090
0.056
0.131
0.142
-0.031
CACW3


1.000
0.817
0.865
0.874
0.535
-0.164
CMo



1.000
0.557
0.969
0.268
-0.401
CC/2CC/2 CeHsCI CCI3CH.pl CgHg




1.000
0.669 1.000
0.852 0.360 1.000
0.053 -0.266 0.206 1.000
                                                                                    •U.S. Government Printing Office: 1993 — 750-071/60217

-------
  The information in this document has
been funded wholly or in part by the U.S.
EPA under Contract No. 68-02-4463 to
Midwest Research Institute, Kansas City,
MO. It has been subjected to Agency re-
view and approved for publication. Men-
tion of trade names or commercial products
does not constitute endorsement or  rec-
ommendation for use.
  GaryD. Hinshawis with Midwest Research Institutef Kansas-CityfMO 64110^
    EPA authors, Frank W. Wilshire (also the EPA Project Officer, see below) and
    Larry D. Johnson, are with the Atmospheric Research and Exposure Assess-
    ment Laboratory, Research Triangle Park, NC 27711.
  The complete report, entitled "Volatile Organic Sampling Train—Soot Interference
    Report,' (Order No. PB93-144145; Cost: $19.50; 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
     BULK RATE
POSTAGE & FEES PAID
         EPA
   PERMIT No. G-35
EPA/600/SR-92/208

-------