United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S7-82-063 Feb. 1983
Project Summary
Evaluation of Potential VOC
Screening Instruments
K. T. Menzies and R. E. Fasano
This report describes the evaluation
of potential fugitive source emission
screening instruments for analysis of
volatile organic compounds (VOCs). An
initial review of available portable VOC
detection instruments indicated that
detectors operating on several princi-
ples (i.e., flame ionization, catalytic
combustion, photoionization, infrared
absorption, and thermal conductivity)
might be useful for VOC analysis.
However, flame ionization and catalytic
combustion devices evaluated previ-
ously showed poor sensitivity for highly
substituted aliphatic and aromatic
organic compounds. Instruments util-
izing photoionization and infrared may
be able to meet necessary criteria for
practical and accurate VOC analysis of
highly substituted organics. Therefore,
three commercially available instru-
ments (HNU PI-101, AID 580, and
Foxboro Miran 80) were modified and
evaluated for 32 such compounds at
concentrations of 100-10,000 ppmv.
Results indicate that photoionization
may be suitable for general VOC screen-
ing, but a reliable instrument/dilution
system does not exist. Infrared absorp-
tion will apparently not provide suitable
general VOC screening, but may be
useful for analysis of some classes of
organic compounds.
This Project Summary was developed
by EPA's Industrial Environmental
Research Laboratory. Research Triangle
Park. NC, to announce key findings of
the research project that is fully docu-
mented in a separate report of the same
title (see Project Report ordering
information at back).
Introduction
The U.S. EPA has issued performance
standards and guidelines to limit emis-
sions of volatile organic compounds
(VOCs) from several stationary source
categories; e.g., surface coating opera-
tions. It is apparent that sources other
than classical point sources may also
emit large amounts of VOCs into the
workplace and surrounding atmosphere.
As described in EPA Method 21, Deter-
mination of Volatile Organic Comppund
Leaks, technically and economically feas-
ible devices suitable for monitoring such
fugitive sources as valves include only a
few that are portable. These devices can
be placed near possible points of emis-
sions and will respond to releases of the
organic compounds. Instruments suit-
able for this purpose include, but are not
limited to, catalytic oxidation, flame
ionization, infrared absorption, and photo-
ionization detectors.
Unfortunately, due to the chemical
complexity of many fugitive sources and
the lack of universal sensitivity of these
detectors, the detectors previously eval-
uated cannot adequately measure the
concentration of all chemicals likely to be
released. This fact has been documented
for two commercially available detectors
using flame ionization and catalytic
combustion. Among 168 compounds
tested, 23 showed sufficiently poor
response that the actual and measured
concentrations differed by a factor of
greater than five.
The classes of compounds which show
the poorest agreement with the actual
concentration generally incorporate func-
tional groups such as halides, hydroxyl
(alcohols), carbonyl (aldehydes, ketones)
and carboxylate (acid) and include both
substituted aromatic hydrocarbons and
low-molecular-weight, highly substituted
aliphatic compounds.
Additional portable devices which
respond accurately to these compounds
-------
are needed for VOC screening. Instru-
ments other than flame ionization or
catalytic oxidation detectors which might
meet this goal operate on the principles of
infrared absorption, photoionization, and
thermal conductivity.
The first step in this task was to procure
one or more units of detectors that meet
the specifications of Method 21. The VOC
instrument must be rugged, reliable,
relatively inexpensive, portable, and easy
to operate. Of course, it must respond to
the organic compounds of interest and be
able to measure the leak definition concen-
tration specified in the regulations.
According to Method 21, the instrument
must be intrinsically safe for operation in
explosive atmospheres as defined by the
applicable National Electric Code. Few
detectors are now "approved" for such an
environment.
The second step in this task was to set
up a laboratory system capable of mixing
known volumes of vapors with air and
delivering the mixtures of known concen-
tration to the detectors. Tedlar bags and a
volumetric mixing system were selected
for sample preparation since they provide
adequate accuracy/precision and require
little cost or time to set up.
The third step in this task was evalua-
tion of the detectors for response to the
compounds of interest. The response
factors (RFs) were determined at several
concentrations from 100 to 10,000 ppmv:
Rp_ Actual Concentration
Concentration Calculated from
Instrument Response
Measurements were limited to concen-
trations approaching about 90% of the
saturation concentration of 75% of the
lower explosive limit (LEL). To permit
statistically valid interpretation of the
measured response factors, five replicate
measurements at three concentrations
were conducted. Data analysis included
calculations of mean response factors
and confidence intervals.
Results and Discussion
Photoionization Detection
The photoionization technique was
evaluated for a limited number of com-
pounds due to both chemical and, more
significantly, equipment problems. The
PI-101 was calibrated with dichloro-
methane to permit direct comparison
with response factors previously report-
ed. The response factors observed for the '
16 compounds tested on the PI-101
range from 0.50 to 48. Those for the 13
compounds evaluated at 1000 ppmv,
summarized in Table 1, range from 0.5 to
8.0. Of the compounds tested, 75% (12)
have response factors of less than 5.0
and greater than 0.2. There appears to be
no obvious trend of response factor with
molecular weight (carbon number) or
functionality within this group. It is
interesting to note that, for both alcohols
tested (methanol and ethanol), the re-
sponse factors are inversely proportional
to carbon number. This is consistent with
the large ionization potential and fewer
ionizable electrons in methanol, as
compared to ethanol. It also appears that
non-bonding electrons on the oxygen
atom of the alcohols do not provide a
much greater photoionization yield than
other sigma-bonded electrons in com-
pounds with similar carbon numbers.
Although the response factors for the
limited number of compounds tested do
not unequivocally confirm the suitability
of photoionization as a general VOC
screening technique, an important but
cautious observation can be made. That
is, based on this small sample of com-
pounds tested, which includes an aro-
matic compound (toluene), an ether
(acetal), an alcohol (ethanol) and chlo-
rinated alkanesftrichloroethane and chlo-
roform), the response factor at concen-
trations of 500 - 10,000 ppmv may be
within a factor of five. This result is
consistent with an expectation of more
similar photoionization yield from sigma
and pi electrons when the compound is
influenced by UV radiation of approx-
imately 12eV rather than 10eV. The
expectation that photoionization yield for
aliphatic and aromatic compounds may
be similar indicates the potential use-
fulness of photoionization in VOC screen-
ing.
In terms of suitability as a potential
VOC detector, the most significant result
with respect to the photoionization
detector (HNU Systems, Inc. PI-101 and
AID, Inc. 580) is the difficulty observed in
operating the prototype dilution system.
Both dilution probes were designed and
Table 1. Response Factors with 95% Confidence Intervals*
OCPDB*
ID No.
-•
790
810
930
-
-
1660
1236
2500
-•
3349
3291
3395
Compound Name Response Factor
Acetal
Carbon Disulfide
Carbon Tetrachloride
Chloroform
Diketene
Dimethylsulfide
Ethanol
Ethylene Dichloride. trans 1.2
Methanol
Pentanethiol, 1-
Toluene
Tetrachloroethane, 1, 1,2,2-
Trichloroethane, 1,1,1 -
1.1
0.50
0.94
1.3
6.8
0.85
2.8
0.96
8.0
0.79
0.85
1.4
0.98
Confidence Interval
1.0 •
0.45-
0.77-
1.3 •
5.9 -
0.80-
2.4 -
0.84-
6.3 •
0.68-
0.67-
1.3 -
0.74-
1.1
0.57
1.20
1.4
7.9
0.90
3.4
1.1
11.0
0.96
1.2
1.4
1.4
* Concentration = 7000 ppmv; a/I are of the light liquid (LLj volatility class.
b Organic Chemical Producers Data Base.
-------
fabricated by the respective manufac-
turers under severe time limitations.
Neither probe was designed to permit
reliable independent measurement of
dilution ratio or reproducible adjustment.
Thus, the absolute dilution ratio is in
some doubt. The ability to adjust the
dilution ratios was practically nonexis-
tent.
Infrared Detection
A total of 32 compounds were analyzed
on the Miran 80. Prior to testing, the
instrument was calibrated with individual
span gases at eight analytical wave-
lengths which correspond to individual
functional groups; e.g., C-H; C-C1; C-OH.
Test compounds were then run, and the
instrument response calculated on the
basis of the response indicated by the
specific span gas used at individual
analytical wavelengths.
An analysis of the data indicates that
the response factors for most compounds
with a particular functional group,
determined at an analytical wavelength
which corresponds to that functional
group, are generally less than 20. This is
consistent with the general observation
that the functional group is more
important than the remainder of the
molecule in determining the IR extinction
coefficient of the compound at the
wavelength of interest.
For example, three of the four aromatic
compounds tested have reasonable re-
sponse factors «5) at 6.35 urn as shown
below. This wavelength is within a broad
aromatic ring stretch area.
Compound
Response Factor
Range
Diisopropyl Benzene 2.42 - 3.75
Dimethyl
Styrene,2,4-
Methyl Styrene
0.185-0.394
0.229-0.718
to incomplete resolution. Also, some shift
of the C-H stretch wavelength probably
occurs due to nearby oxygen or halogens.
A list of aliphatic compounds and corres-
ponding response factor ranges at this
wavelength is shown in Table 2. If four
alkylated aromatic compounds are in-
cluded in the list of compounds with
response factors less than 5 at 3.3 - 3.4
um, the percentage of compounds tested
with suitable response factors increases
to 62%.
Ten chlorinated hydrocarbons tested in
this program yielded measurable re-
sponse factors at 13.5 um; 70% yielded
response factors less than 5 at this
wavelength.
Since the ultimate goal of this instru-
ment evaluation is to assess the suit-
ability of IR as a general VOC screening
technique, an assessment of the useful-
ness of a single wavelength for measure-
ment of organic compounds of varied
molecular weight and functionality is in
order. A review of the data indicates that
32 test compounds yield response factors
of 5.0 - 0.2 at each analytical wavelength:
Number of
Wavelength (//m) Compounds
Table 2. Substituted Aliphatic Compounds
with Response Factors Less than
20 at 3.3 yum
3.3
3.4
3.6
5.7
6.35
8.8
9.5
13.5
12
4
3
1
3
4
15
7
Within this group, the addition of the
large aliphatic group (isopropyl) on the
benzene ring appears to reduce the
sensitivity (larger response factor) at the
aromatic C — C stretch wavelength as
compared to less alkylated aromatics.
For aliphatic and substituted aliphatic
compounds, the C-H stretch wavelength
of 3.3 jum yields suitable response factors
(<5) for about 52% of those tested. The
classical aliphatic C-H stretch is observed
at 3.4/um, but some overlap of 3.3 and 3.4
/Km IR bands may occur in the Miran due
•fr U S GOVERNMENT PRINTING OFFICE^ 1983-
The results indicate that only 3.3, 9.5,
and 13.5 /urn analytical wavelengths
respond acceptably for a large number of
compounds. However, in any case, fewer
than 50% of the compounds are reliably
detected. Thus, there is apparently no
useful agreement in response factors
between, for example, a large number of
aromatic compounds and aliphatic com-
pounds (e.g., 50% of those tested) at
analytical wavelengths specific to each
compound class. This observation indi-
cates that infrared spectrophotometry is
not particularly suitable for general VOC
screening.
On the other hand, the fact that the
response factors do not vary by large
values (i.e., greater than 5.0) for some
classes of compounds (e.g., halogenated
aliphatics at 13.5/urn and aliphatic and
alkylated aromatics at 3.3 - 3.4 /um)
corroborates the suitability of infrared
spectrophotometry for VOC screening of
compounds belonging to one functional
group.
- 659-O17/O89B
Compound
Response
Factor Flange
Acetyl-1 -propanol,3-
Chloro-acetaldehyde
Dichloro-1 -propanol.2,3-
Dichloro-2-propanol, 1,3-
Diketene
Dimethylsulfide
Ethanol
Ethyleneglycoldimethyl
Ether
Ethyleneglycolmonoethyl
Ether Acetate
Formaldehyde
Formic Acid
Glycidol
Methanol
Methylene Chloride
Pentanethiol. 1 •
Propylene Chlorohydrin
Tetrachloroethane, 1.1.2,2-
Trichloroethane. 1.1,1-
1.23 - 2.02
2.73 - 8.62
18.5
5.29
8.06 -14.1
0.488- 0.495
0.261 - 0.292
0.196- 0.296
0.280- 0.488
1.09 • 1.88
0.529- 0.722
0.382
0.294- 0.410
2.67 - 2.87
0.314- 0.633
0.334- 0.4Q3
8.59 - 9.90
1.69 - 3.76
Conclusions
In summary, based on the results of
this evaluation, it appears that:
1. Infrared (IR) spectrophotometry is not
suitable for general VOC screening,
except for analysis of VOC emissions
of a single organic functional group
character.
2. IR screening of organic compounds
of a single functional class (e.g., C-
C1) may be suitable for as many as
80% of compounds in the class.
3. IR screening at a wavelength corres-
ponding to both aliphatic and aro-
matic stretches may be suitable for
as many as 50% of organic com-
pounds.
4. A portable photoionization device is
not available for VOC screening at
concentrations of 100 - 10,000
ppmv.
5. Development of a reliable dilution
probe for use on a photoionization
device is readily achievable.
6. With such a dilution probe, it appears
that a photoionization device with an
11.7 or 11.8 eV UV lamp may be, used
for reliable analysis of VOC fugitive
emissions.
-------
K. T, Menzies and R. E. Fasano are with Arthur D. Little, Inc., Cambridge, MA
02140.
Merrill D. Jackson is the EPA Project Officer (see below).
The complete report, entitled "Evaluation of Potential VOC Screening Instru-
ments, " (Order No. PB 83-139 733; Cost: $11.50. 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:
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
------- |