United States
Environmental Protection
Agency
Atmospheric Research and
Exposure Assessment Laboratory
Research Triangle Park, NC 27711
Research and Development
EPA/600/SR-93/115 August 1993
EPA Project Summary
Development of an Analysis
Method for Total Nonmethane
Volatile Organic Carbon
Emissions from Stationary
Sources
James F. McGaughey and Samuel C. Foster II
The accurate measurement of the to-
tal nonmethane volatile organic carbon
emissions from stationary sources is
critical to characterizing many indus-
trial processes and for regulating ac-
cording to the Clean Air Act. Current
methods are difficult to use and the
ability to do performance audits has
been marginal, especially at low con-
centrations (<50 parts per million of
carbon, ppmc).
One of the key elements for an ideal
measurement technique would be a de-
tector that responds to all classes of
organic compounds equally, based on
the number of carbon atoms present. A
commercially available catalytic flame
lonizatlon detector (CFID) has shown
promise In this area. Laboratory stud-
ies with a CFID were performed to de-
termine the response of compounds
with various functional groups. These
classes included brominated, chlori-
nated, nltrogenated, oxygenated, aro-
matic, and non-aromatic compounds.
The response of each compound is
compared to the response of an alkane
with the same number of carbon at-
oms. This report discusses the experi-
mental work with this detector and an
approach for sampling, sample recov-
ery, and field tests for comparison to
ithe EPA Method 25.
The information In this document has
been funded wholly or In part by the
United States Environmental Protection
Agency (EPA) under contract 68-D1-
0010 to Radian Corporation. It has been
subjected to Agency review and ap-
proved for publication. Mention of trade
names or commercial products does
not constitute endorsement or recom-
mendation for use.
This Project Summary was developed
by EPA's Atmospheric Research and
Exposure Assessment Laboratory, Re-
search 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 accurate measurement of the total
nonmethane volatile organic carbon emis-
sions from stationary sources is critical to
the characterization of many industrial pro-
cesses. Current methods are difficult to
use especially at low concentrations (<50
ppmc). One of the key elements for an
ideal measurement technique would be a
detector that responds to all classes of
organic compounds, equally based on the
number of carbon atoms present. The flame
ionization detector (FID, the detector of
choice for most of the analytical methods)
responds to unsubstituted alkanes in this
manner. However, when functional groups
are added or when the structure (aromatic
or cyclic) changes, the response no longer
follows this pattern. A commercially avail-
able catalytic flame ionization detector
(CFID) has shown promise in this area.
The CFID uses a ceramic source coated
with a nickel/aluminum oxide compound to
act as a combination ignitor, polarizer, and
catalytic surface in an hyair flame environ-
ment. The CFID ceramic catalyst tempera-
ture is controlled through a power supply
that is adjustable from 0.0 to 4.0 amperes
Printed on Recycled Paper
-------
(amps). Increasing the current to the cata-
lyst raises the source temperature. A bal-
ance between the catalyst temperature and
the detector temperature is essential to the
complete combustion of organic com-
pounds. Generally, the catalyst tempera-
ture can be varied from 400 to 800°C, and
the detector temperature can be varied
between 100 and 400°C.(1)
The detector's performance was evalu-
ated by analyzing organic compounds with
various functional groups (halogen, oxy-
gen, nitrogen, and aromatic). Functional
groups were evaluated at different cur-
rents and fuel ratios until an optimal cur-
rent and fuel ratio was found that gave a
universal response. Once the optimal con-
ditions were determined, the performance
of the CFID was compared to the perfor-
mance of an FID. The overall performance
of the CFID was evaluate by analyzing 61
organic compounds. The response ratio
for each compound was compared to the
response ratio of straight-chained alkanes.
All of the response ratios are based on the
number of nanomoles (nmoles) that were
injected on column.
Experimental Method
A CFID and power supply were installed
on a Varian 3600 gas chromatograph (GC).
The power supply current was variable
from 0.0 to 4.0 amps. The FID was a
Varian FID installed on a Varian 3400 GC.
The analytical column, used for all analy-
ses, was a DB-5, 0.54 millimeter x 30
meter, fused silica capillary column. A PE
Nelson 3000 series Chromatography Data
system was used for data acquisition and
processing.
The detector tower temperature was set
at 310°C for all of the experiments. The
temperature limit for the column, as indi-
cated by the manufacturer, was 350°C.
The operating conditions were well below
the limits of the column.
The fuel/air ratio, as recommended by
the manufacturer for the CFID, was a 1:10
mix of hydrogen and air. To minimize
source deterioration, the flow of hydrogen
did not exceed 25 mL/min and the flow of
air did not exceed 250 mL/min. The maxi-
mum flows were chosen for the initial stud-
ies, and a different ratio was later evaluated.
A mix of four aliphatic hydrocarbons was
prepared at a concentration of 1.0 milli-
moles (mmole) each in dichloromethane.
This mix was used as the baseline for
evaluating the detector response to the
number of carbon atoms present. A solu-
tion of dichloromethane, chloromethane,
and tetrachloromethane (single carbon
chloroalkanes) in nonane was prepared
with each compound at 0.12 mmole. The
chloroalkane solution was analyzed on the
CFID with the current set at 0.0 and on the
FID for comparison. The chloroalkane so-
lution was then analyzed on the CFID at
six different currents: 0.0,2.0,2.4,2.8,3.2,
and 3.6 amps, to find the optimal current
for the chlorinated compounds.
Different mixtures containing compounds
of specific functional groups were then
prepared. The standards were prepared at
a nominal concentration of 500 ug/mL An
internal standard (IS), nonane, was added
to each solution at a concentration of 115
ug/mL The standards were analyzed at
the optimal current, and at a higher current
to determine the effects on the different
functional groups.
The response factor (RF) for each com-
pound was calculated using Equation 1.
The response factor to nmole was plotted
against the number of carbons in each
compound.
RF = (Compound area/I S area)
* (1/ nmoles of compound injected)
0)
A "least-squares-fit" was applied to the
data points from each functional group
with the slope, intercept, and correlation
coefficient calculated for each of the gen-
erated lines. The linear regression infor-
mation was compared to the results for the
aliphatic hydrocarbons. The number of car-
bons that each compound deviated from
the aliphatic line was calculated using Equa-
tion 2.
No. Carbons Deviated = No.C
RF-I
(2)
where
No. C = the actual number of car-
bon atoms found in the com-
pound
RF = the response factor for the
compound Equation 1
I = the intercept from the least-
squares-fit of the data
S = the slope of least-squares-
fit
The average number of deviated car-
bons was calculated for each class of com-
pounds for comparison to the aliphatic
hydrocarbons.
Results and Discussion
A mixture of four straight-chained al-
kanes (heptane, octane, nonane, and de-
cane) was analyzed on the CFID and
compared to the FID as a preliminary test
of detector linearity. The CFID was compa-
rable to the FID, with both detectors show-
ing linearity with increasing carbon number
for the aliphatic hydrocarbon mix.
Chlorinated compounds were chosen for
the initial experiments because of their low
response on FID, as compared to alkanes.
Single carbon compounds (dichlorometh-
ane, trichloromethane, and tetrachloro-
methane) were selected so that the only
difference between the compounds was
the number of chlorines present. With 0.0
amps of current applied to the detector,
the chloroalkanes responded similarly, on
a molar basis, when analyzed on the CFID.
When the chloroalkanes were analyzed on
the FID, the response decreased as the
number of chlorines increased. The chlo-
roalkane standard was then analyzed at
0.0, 2.0, 2.4, 2.8, 3.2, and 3.6 amps to
determine the optimal current for this class
of compounds. As the current was in-
creased, the sensitivity increased, but the
baseline became increasingly noisy. The
best compromise between sensitivity, uni-
form response, and baseline stability was
found to be a current setting of 2.4 amps.
A mixture of six aliphatic hydrocarbons
(hexane, heptane, octane, decane,
tetradecane, hexadecane) was prepared
from stock standards four times and ana-
lyzed in duplicate using the CFID with the
current set at 2.4 amps (Figure 1). The
RFs were averaged and a "least-square-
fit" was applied to the data points (Table
1). The aliphatic hydrocarbons responded
linearly on the CFID with a correlation
coefficient of 0.992, and the resulting line
was used as the baseline for comparison
with the other compound classes.
Separate mixtures of compounds from
five functional groups (aromatic, bromi-
nated, chlorinated, nitrogenated, oxygen-
ated) with nonane as the IS were prepared
and analyzed at 2.4 amps. The RF for
each compound was compared to the RF
for the aliphatic hydrocarbons. Some of
the compounds could be placed in several
of the functional groups, but they were
grouped together based on the predomi-
nate functional group. Additional studies
were performed at higher currents for the
aliphatic, aromatic, chlorinated, and oxy-
genated compounds in an attempt to im-
prove linearity and sensitivity.
Figures 1 through 6 provide a graphical
representation of the CFID response ver-
sus carbon number for the six functional
groups studied at 2.4 amps. For compari-
son purposes, a "least-squares-fit" was
performed on each data set that generated
a value for the slope and correlation coeffi-
cient. The two values for each data set
were compared to those generated for the
aliphatic compounds, which was used as
the target or theoretical situation. The data
from the "least-squares-fit" for the aliphatic
compounds and the RF calculated for each
-------
0.8
06 —
•s
(b
^
B Aliphatic
Target Line
/
/'
^r \
s
i
X
,
X"
,
, x
X'
xx
024
Figure 1. Aliphatics at 2.4 Amps.
6 8 10 12
Number of Carbons
14
16 18
Table 1. Linear Regression for Compounds at 2.4 Amps.
Functional
Group
Aromatic
Brominated
Chlorinated
Aliphatic
Nitrogenated
Oxygenated
No.
Compounds
Analyzed
7
3
20
6
8
17
Slope
0.0611
0.0501
0.0401
0.0435
0.0477
0.0456
Intercept
-0. 1208
-0.001 1
0.0461
0.0154
-0.0380
-0.0349
Corr.
Coef.
0.9872
0.9990
0.9642
0.9964
0.9771
0.9711
Avg. Deviation
From Target
(No. Carbons)
0.41
0.07
0.40
0.01
-0.50
-0.90
compound associated with the other func-
tional groups were used to calculate the
number of carbon atoms for each com-
pound. This experimentally determined
value for the number of carbon atoms was
then compared to the actual number of
carbon atoms in each compound (Table
1).
The plotted slopes for the nitrogenated
and oxygenated compounds (Figures 5
and 6) are similar to that for the aliphatic
compounds, which indicates that the re-
sponses increase with the number of car-
bon atoms (as expected for normal al-
kanes). However, the magnitude of the
responses was less than that for the
aliphatics, making the experimentally de-
termined carbon number for the nitroge-
nated compounds, on the average, low by
approximately 0.5 carbon and the oxygen-
ated compounds low by approximately 1
carbon.
The slopes for the aromatic and bromi-
nated (Figures 2 and 3) compounds were
greater than that for the aliphatics. This
shows that the CFID response increases
as carbon number increases but at a
greater magnitude than for aliphatic com-
pounds. The experimentally determined
carbon number for the aromatics was found
to be high, on the average, by 0.4 carbons,
whereas, the experimental number of car-
bons for the brominated compounds was
found to be equal to the number of actual
carbons.
The slope for the chlorinated compounds
was less than for the aliphatics, indicating
that the CFID response increases as the
number of carbons increase, but at a mag-
nitude less than that for the aliphatics. The
experimentally determined number of car-
bons was high, on the average, by 0.4
carbons. The correlation coefficients were
all greater than 0.93. This indicates that all
of the data points lay on or near the result-
ing line.
Several functional groups were analyzed
at a higher current to possibly improve
linearity and sensitivity. Aliphatic, aromatic,
oxygenated, and chlorinated compounds
were analyzed at 3.2 amps. Table 2 shows
the resulting linear regression data for the
compounds that were analyzed. Linearity
was not improved with the correlation coef-
ficients being less than 0.96. Sensitivity
toward increasing carbon numbers in-
creased slightly for the oxygenated com-
pounds, compared to the slope of the lines
at 2.4 amps and 3.2 amps, and the experi-
mentally determined number of carbons,
on average, increased by 0.4 carbons.
Sensitivity did not increase for the chlori-
nated compounds, with the slope increas-
ing and the experimentally determined
number of carbons, on average, increase
to 2 carbons.
The oxygenated compounds were of
special interest, since they are a major
component of many Method 25 analyzes.
They showed a reduced response, as com-
pared to the aliphatic hydrocarbons, there-
fore it was important to closely examine
this class of compounds. As noted above,
increasing the current did not change the
overall response of the compounds. The
fuel-gas mixture was changed to 40 mL
min for hydrogen and 250 mL/min for air
The CFID did not behave well at this fuel
ratio. The baseline was erratic, and the
signal dropped below the baseline after
the solvent peak passed through the col-
umn. The CFID behaves better at a 1:10
gas ratio; therefore the ratio cannot be
changed to achieve better sensitivity to-
wards a functional group.
As a confirmation of the response of the
CFID toward the oxygenated compounds,
the oxygenated compounds were analyzed
with aliphatic hydrocarbons on the CFID at
-------
0.8 ~
05
04 —
01 —
B Aromatic
Target Line
x
X
X
X
Xi
X
Bx
,
1
X"
i
V
u
y
X
X
s
X
X
X
6 8 10 12
Number of Carbons
14 16
18
Figure 2. Aromatic at 2.4 Amps.
0.8'
0.7
0.6
o 0.5
•2
o>
^
§ 0.4
0.3
0.2
0.1
,*
B Brominated
Target Line
/
1
X
X
X
X"
X
x
/
X
X
X
X
X
X
X
X
2.4 amps and the FID. The CFID response
to the oxygenated compounds was the
same as the FID response. Table 3 lists
the compounds analyzed and the response
on the CFID and the FID. The number ot
carbons deviated from the target aliphatic
line was calculated and the results are
listed in Table 3. The average number of
carbons deviated from the target response
was -1.25 for both the CFID and the FID.
There were 61 compounds analyzed on
the CFID. Figure 6 shows all of the com-
pounds analyzed on the CFID, and Table
4 lists all of the compounds that were
analyzed in order of increasing response
factor. The compounds show that the CFID
response increased as the number of car
bons increased. The response for the com-
pounds that are showing a low response
are only low, on average, by 1 carbon
atom.
Conclusions
The CFID is a detector that acts as a
carbon counter, in that the response to
compounds increases linearly as the num-
ber of carbons increases. Oxygenated com-
pounds did not respond as well as the
other functional groups but did respond
linearly with increasing carbon number
For halogenated compounds, the CFID
outperformed the FID with a response that
was unaffected by the number of chlorine
atoms and responded linearly with increas-
ing carbon number. The CFID (at 2.4 amps)
results averaged one carbon number or
less deviation when compared with aliphate
compounds. The CFID remained stable for
6 months of continuous use. The CFID is a
versatile detector that is able to overcome
some of the selectivity problems of the
FID. The CFID appears to be a good choice
as a universal detector that may increase
the overall detection limit of current sta
tionary source analyses methods.
Reference
1. Theory and Operation of the TID/
CFID Detectors, FTD Detector, Re
mote FID Detector, Tandem TID
Detector, FID Detector; DETector
Engineering & Technology, Inc.
1991,ppl-2-l-6.
6 fl ro /2 M re te
Number of Carbons
Figure 3. Brominated at 2.4 Amps.
-------
0.8
06 —
% 0.5
1
8. 0.4 -
0.1 -
0 —
m Chlorinated
Target Line
1 / '
x
'-
i
//
•x
f
x
X
/
'
X
. '
'
,'
024
F/gu/v 4. Chlorinated at 2.4 Amps.
6 8 10 12 14 16 18
Number of Carbons
0.8 —
0.6-
% 0.5
£
03
1
8. 0.4 -
1
0.1 -
n Nitrogenanated
— — Target Line
x
/'
i
/!
' "
Xl
x
i
"
X
\ '
X'
/x
0 24 6 fi 10 12
Flgun 5. Nitrogenated at 2.4 Amps. Number of Carbons
14 16
18
-------
0.8 —
1
a- °-4
H Oxygenated
Target Line
<'
,
1
/ 1
i
|X4"
i
i
x
>
/
, '
'
,'
6 8 10 12
Number of Carbons
14
16
18
Figure 6. Oxygenated at 2.4 Amps.
Table 2. Linear Regression lor Compounds at 3.2 Amps.
Functional
Group
Aromatic
Chlorinated
Aliphatic
Oxygenated
No.
Compounds
Analyzed
7
15
3
4
Slope
0.0453
0.0328
0.0155
0.0490
Intercept
0.0159
0. 1345
0.2447
-0.0293
Avg. Deviation
Corr. From Target
Coef. (No. Carbons)
0.9284
0.8968
0.9538
0.7969
0.33
1.99
-0.10
-0.49
-------
Table 3. CFID at 2.4 Amps vs. FID for Oxygenated Compounds
Compound
4-Methyl-2-Pentanone
p-Tolualdehyde
2-Butanone
Acetone
Ethyl Ether
Methanol
Propanol
Ethyl Acetate
RF
CFID
0. 1482
0.3234
0.1438
0.0898
0.1410
0.0394
0.1114
0. 1253
RF
FID
0. 1479
0.3291
0. 1457
0.0902
0. 1480
0.0363
0. 1075
0.1215
CFID Dev.
Carbons
-3.0
-0.9
-1.1
-1.3
-1.1
-0.5
-0.8
-1.5
FID Dev.
Carbons
-3.0
-0.8
-1.0
-1.3
-1.0
-0.5
-0.9
-1.6
Table 4. Organic Compounds Analyzed on CFID
Aliphatic Aromatic Brominated
Hexane
Heptane
Octane
Decane
Tetradecane
Hexadecane
Benzene
Toluene
o-Xylene
Ethylbenzene
m-Xylene
p-Xylene
1,2,4-Thmethylbenzene
Dibromomethane
1,3-Dibromoethane
Bromobenzene
Chlorinated
Nitrogenated
Methylene chloride
Chloroform
Carbon tetrachloride
1,1-Dichloroethylene
1,2-Dichloroethane
1,1,2-Trichloroethane
1,1,2,2-Tetrachloroethane
Trichloroethylene
Tetrachloroethylene
1,2-Dichloropropane
1,2,3- Trichloropropane
Hexachlorocyclopentadiene
2-Chlorophenol
1,2,4- Trichlorobenzene
Dichlorobenzene
4-Chloro-3-methyl-phenol
o-Chlorophenol
1,4-Dichlorobenzene
4-Chlorotoluene
Chlorobenzene
4-Nitrophenol
2,4-Dinitroaniline
2,4-Dinitrotoluene
4-Nitroaniline
1,4-Dinitrobenzene
2,6-Dinitrotoluene
1 -Nitronaphthalene
Diphenylamine
Oxygenated
Methanol
n-Propyl Alcohol
Acetone
Methyl ethyl ketone
Ethyl acetate
Ethyl ether
2-Butanone
Valeraldheyde
Hexanal
1-Butanol
Phenol
4-Methyl-2-pentanone
Benzaldehyde
p-Tolualdehyde
Acetonphenone
1-Octanal
Isophorone
&V.S. GOVERNMENT PRINTING OITICE: 1993 - 750471/MNMO
-------
James F. McGaughey and Samuel C. Foster II are with Radian Corporation,
Research Triangle Park, North Carolina 27709.
Merrill Jackson is the EPA Project Officer (see below).
The complete report, entitled "Development of an Analysis Method for Total
Nonmethane Volatile Organic Carbon Emissions from Stationary Sources,"
(Order No. PB93-214419; 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, North Carolina 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-93/115
------- |