CD
O
••fc
Report Ho. 78-CKO-12
TEST
ANALYSIS OF POLYNUCIEAR'AROMATIC HYDROCARBONS
FROM COKE OVEN EFFLUENTS
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Air Quality Planning and Standards
Emission Measurement Branch
Research Triangle Park. North Carolina
-------
AIR POLLUTION EMISSION TEST
ANALYSIS OF POLYNUCLEAR AROMATIC HYDROCARBONS
FROM COKE OVEN EFFLUENTS
By
R.G. Beimer
September 26, 1978
TRW Defense and Space Systems Group
One Space Park
Redondo Beach, California 90278
-------
TABLE OF CONTENTS
PAGE
1.0 INTRODOCTION 1
2.0 SAMPLE HANDLING AND PREPARATION 3
2.1 Procedures for GC/MS Sample Preparation 3
2.2 Sample Preparation for Thin Layer
Chromatographic (TLC) Plates . 4
2.3 Liquid Chromatography Clean Up Procedure
for Polynuclear Aromatic Hydrocarbons 4
3.0 SAMPLE ANALYSIS AND RESULTS 5
3.1 GC/MS Analysis of Liquid Samples 5
3.2 GC/MS Analysis of Thin Layer
Chromatographic (TLC) Plates 32
3.3 Results of GC/MS Analysis of Coke
Oven Samples 40
4.0 METHODS OF IDENTIFICATION & QUANTITATION 44
4.1 Methods for PAH Identification 44
4.2 Methods for PAH Quantisation 45
5.0 OTHER TECHNIQUES FOR THE DETERMINATION OF POLYNUCLEAR
AROMATIC HYDROCARBONS AND RECOMMENDATIONS FOR
IMPROVEMENTS 46
Appendix 1
Liquid Chromatography Separation Procedure 50
Appendix 2
Table of PAH Standards Available 53
Appendix 3
Conditions used for GC/MS Analysis 54
-------
LIST OF FIGURES
PAGE
Figure 1. Reconstructed GC from 60 Meter OV-17 Glass
Capillary Column GC/MS (Sample 367) . ; 7
Figure 2. Reconstructed GC from 60 Meter OV-17 Glass
Capillary Column GC/MS, Sample 367
(Time Scale Expanded) 8
Figure 3. 60 Meter OV-17 Glass Capillary Reconstructed
GC and Mass Chromatogram for M/e = 202 Region .... 10
Figure 4. Spectrum for Fluoranthene from OV-17
Capillary Run (Scan 1736) 11
Figure 5. Library Search for Scan 1736 12
Figure 6. Spectrum for Pyrene from OV-17 Capillary
Run (Scan 1834) 13
Figure 7. Library Search for Scan 1834 . 14
Figure 8. Reconstructed GC from 30 Meter OV-101 Glass
Capillary Column GC/MS, Sample 367 16
Figure 9. Reconstructed GC from OV-101 Glass Capillary ,,
Column Sample 367 (Time Scale Expanded)
Figure 10. Molecular Weight 202 Region of OV-101
Capillary GC/MS Run 20
Figure 11. Molecular Weight 252 Region of OV-101
Capillary GC/MS Run 22
Figure 12. Mass Spectrum Scan 1716 from OV-101
Capillary GC/MS Run 23
Figure 13. Library Search Scan 1716 24
Figure 14. Spectrum 1618 from OV-101 Capillary
GC/MS Run 25
Figure 15. Library Search Scan 1618 26
Figure 16. Reconstructed GC from Dexil 300 Packed
Column GC/MS Run 27
Figure 17. Reconstructed GC from Dexil 300 Packed
Column, Sample 367 (Time Scale Expaned) 28
Figure 18. Molecular Weight 202 Region of Dexil
300 Packed Column GC/MS 31
Figure 19. Molecular Weight 252 Region of Dexil
300 Packed Column GC/MS 33
Figure 20. Spectrum 1730 for Dexil 300 Packed
Column GC/MS 34
Figure 21. Library Search Spectrum 1730 35
Figure 22. Spectrum 1778 from Dexil 300 Packed
Column GC/MS Run 36
Figure 23. Library Search Scan 1778 37
ii
-------
(Continued)
LIST OF FIGURES
PAGE
Figure 24.
Figure 25.
Figure 26.
Figure 27.
Figure 28.
Figure 29.
GC/MS of TLC Channel 8 . .
Mass Spectrum TLC Channel 8
Scan 1753
GC/MS Results from Coke Oven Samples ,
Coke Oven Samples Analyzed by GC/MS . ,
HPLC Separation of PAH Standards Using
UV (Lower Trace) and Selected Wave
Length Fluorescence (Upper Trace) . . ,
HPLC Separation of PAH Standards Using
UV (Upper Trace) and Selected Wave
Length Fluorescence (Lower Trace) . . ,
38
39
41
43
48
49
m
-------
1.0 INTRODUCTION
This document is the final report for the Analysis of Polynuclear
Aromatic Hydrocarbons From Coke Oven Effluents. The program was conducted
by TRW for the Environmental Protection Agency, Emission Measurement Branch,
Research Triangle Park, North Carolina under Contract Number EPA 68-02-2812.
Four samples taken from a coke oven at Republic Steel Corporation in
Gadsden, Alabama were used for methods development in polynuclear aromatic
hydrocarbon determination.
Polynuclear aromatic hydrocarbons (PAH), also referred to as polycyclic
organic matter (POM), are suspected of being serious health hazards since
some of the individual compounds are known carcinogens. This class of
compounds condense on particulate matter and can be inhaled into the lungs
or contacted with the skin. When coal is pyrolyzed by a burning or coking
process, significant amounts of PAHs are produced. In the burning process
most of the organic material is destroyed through oxidation, producing
water and carbon dioxide. In the coking process however, significant amounts
of organics are evolved and emitted to the atmosphere. Many techniques
exist for the determination of polynuclear aromatic hydrocarbons. Specific
compound analysis for PAHs of particular interest or potency have been
developed and are generally better than survey analysis techniques. As an
example, the determination of benzo(a)pyrene can be accomplished quickly
and accurately by thin layer chromatography (TLC) with quantitation using
variable wavelength fluorescence, a method developed by the EPA . Other
specific PAHs have been determined by liquid chromatography using UV and/or
fluroescence detection.
For general purpose determination of PAHs in complex mixtures the use
of gas chromatography-mass spectrometry is unexcelled. The gas chromato-
graph is used to separate the complex mixture into individual component
peaks and the mass spectrometer is used as a detector for both qualitative
identification of the PAH and for quantitative determination. In simple
cases where only a few PAHs are present, the use of packed gas chromatographic
]New Benzo(a)pyrene Analytical Method; EPA EMSL ACB SFMCS RTP N.C.; Don
Swanson et.al. (no date).
-------
columns are recommended. Packed columns are generally easier to use, re-
quire a shorter time for analysis, and are cheaper. When highly complex
mixtures of PAHs are present or significant interferences occur, the use
of capillary columns becomes necessary. Capillary column selection is
usually dependent on the volatility and polarity of the PAHs to be determined.
PAHs typically show very strong molecular ions (the mass representative
of molecular weight) in their mass spectra because of the stability of the
aromatic system. The strong melecular ion allows for a very simple
identification of molecular weight even in complex mixtures. PAHs produce
very sensitive spectra when ionized in a mass spectrometer and can be
identified in a complex mixture quite easily. The disadvantage of this
strong molecular ion production is that isomeric PAHs produce spectra which
are virtually identical. The mass spectrometer is a rather poor device for
distinguishing isomers since once ionized the compound will seek its most
thermodynamically stable form. Since isomeric PAHs produce similar or
identical mass spectra the determination of structure is accomplished by
comparison of GC retention time using standards.
The major limitation of the gas chromatograph is the volatility of
compounds to be analyzed. In order to be determined, PAHs must produce
good chromatographic peaks at temperatures below 300°C. Above 300°C even
the most stable GC column liquid phases produce high background in the
mass spectrometer. This GC temperature limitation means that high molecular
weight PAHs (above MW = 300) cannot be reliably determined by this technique.
A limited amount of work has been done in the determination of high
molecular weight PAHs using liquid chromatography. With a liquid chromato-
graph, the need for volatility does not exist and materials of virtually
any molecular weight can be determined. With the proper selection of
columns and the incorporation of a variable wavelength excitation and
detection fluorometer, PAHs can be determined with accuracy. Chromatographic
separation is not always requried provided the excitation and detection wave-
length can be selected specifically for PAHs of interest. An example of a
liquid chromatographic determination of polynuclear aromatic hydrocarbons is
given in Section 5.0
-------
2.0 SAMPLE HANDLING AND PREPARATION
Due to the potentially hazardous nature of polynuclear aromatic hydro-
carbons, all samples and standards used in this study were handled with
gloved hands and in a properly functioning fume hood. PAH compounds are
also sensitive to ultraviolet radiation which cause a slow degradation in
compound structure. For this reason the samples and standards were kept
refrigerated and in the dark except when used, and then only under GE
F96T12-GO gold light.
2.1 Procedures for GC/MS Sample Preparation
The samples for gas chromatography/mass spectrometry were initially
screened using packed column GC/MS; chromatographic conditions for the
various GC columns used in this study are given in Appendix 3. The purpose
of this initial screening was to determine approximate PAH content and to
obtain a measure of sample complexity. The screening process showed the
samples to be highly concentrated and to contain mainly aromatic hydro-
carbons. It was also determined at this point that the use of packed gas
chromatographic columns would not provide the separation power required for
the identification of the various isomeric species present in the samples
since GC retention time would be used for specific isomer identification.
Internal standards to be used for quantitation were selected so as to
elute in the regions of interest, but not to interfere with peaks of interest.
The internal standard mixture was made by disolving 50 mg of each internal
standard in 100 ml of acetone. 2 ml of this internal standard solution
and 2 ml of individual samples were diluted to 10 ml and this solution was
used for subsequent analysis. 3 to 5 u/1 portions of the sample were in-
jected into the gas chromatographic column for the determination of PAH
content.
2.2 Sample Preparation for Thin Layer Chromatographic (TLC) Plates
The TLC procedure discussed briefly and referenced in Section 1.0
was used by the EPA for the determination of benzo(a)pyrene. This TLC
plate was analyzed by GC/MS to determine if benzo(a)pyrene was the only
component present since anthanthrene coelutes with benzo(a)pyrene under
the conditions of the method. The TLC plate was first studied by
-------
illuminating the surface with ultraviolet light and finding the location of
the benzo(a)pyrene standards in channels one through six. Using the loca-
tion provided by the standards, the sample channels were scraped at the same
distance from the start. With the use of a UV light it is impossible to
distinguish a specific spot in the channels containing sample since the
entire area from top to bottom was fluorescent.
The scraped area material was placed in a vial with one milliliter of
methylene chloride. The vial was shaken for one minute, then allowed to
stand for 30 minutes to let the solid material settle. A portion of this
Q
solution was analyzed by 6C/MS. A Dexil 300 packed GC column was used for
the determination. Complete operating conditions for the column are pro-
vided in Appendix 3. Channel 5, the twenty nanogram standard of benxo(a)-
pyrene, was used in an attempt to quantitate the benzo(a)pyrene in channels
eight and twelve. This method of quantitation was necessary since it was
not known how efficiently the benzo(a)pyrene was extracted from the TLC
substrate.
It was obvious that the TLC separation was extremely good since only
benzo(a)pyrene was detected by GC/MS in the region of the TLC plate sampled.
It was requested that, due to a specific interference, anthanthrene be
looked for in the TLC samples. No anthanthrene was detected under the con-
ditions used for this analysis. Anthanthrene has a much longer retention
time on this GC column than benzo(a)pyrene and would easily be detected even
though no standard for this material was available in our laboratory.
2.3 Liquid Chromatography Clean Up Procedure for Polynuclear Aromatic
Hydrocarbons
As a means of preparing the coke oven effluent samples, a portion of
the internal standard prepared material, made up as described in Section 2.1,
was chromatographed using the Level 1 liquid chromatograph separation pro-
cedure. Details of this procedure are given in Appendix 1. Level 1
procedures are designed for semi-quantitative (- a factor of 2 to 3)
determination of pollutants and to provide screening data to assess pol-
lution hazards. The procedures are not necessarily adaptable for quantita-
tive determination of emissions from a specific source when accuracies of
better than a factor of 2 are required. By using internal standards it was
-------
possible to determine the procedure's effect on the absolute as well as
relative PAH concentration. The results of the LC procedure show there is
about a factor of 2 loss of lower molecular weight pounds relative to the
high molecular weight ones. In addition, an overall decrease in PAH concen-
tration occurs. The purpose of the LC procedure is to clean up samples so
that subsequent analysis will be based on similar compounds. The separation
is based primarily on polarity. It is therefore recommended that LC prepara-
tion not be used for concentrated effluents from coke ovens, since the samples
are already high in polynuclear aromatic hydrocarbons and the trace amounts
of other materials are not believed to interfere.
3.0 SAMPLE ANALYSIS AND RESULTS
3.1 GC/MS Analysis of Liquid Samples
The four samples for gas chromatography/mass spectrometry were analyzed
/B)
together with their solvent blanks by initial screening using a Dexil 30(r
packed GC column. Conditions for chromatograph columns are given in Appendix
3. Capillary columns were investigated as an alternative to the use of packed
columns after it was determined that significant concentrations of poly-
nuclear aromatic hydrocarbons were present and that various isomeric forms
would have to be determined.
After preparation with the internal standard, as discussed in Section 2.1,
a portion of each sample was chromatographed on a 60 meter glass wall coated
open tubular OV-17 column. Very good chromatographic resolution resulted.
Qualitative as well as quantitative determination of various compounds were
2
made. Standards, where available, and published relative retention times
were used to identify many of the isomeric compounds.
The application of the OV-17 glass capillary column was limited and
compounds above molecular weight 220 could not be determined due to long
retention times. Compounds with molecular weights below 220 were well
resolved. An example for sample #367 of the chromatographic resolution
using the glass OV-17 capillary column is given in Figure 1.
2R.C. Lao, et.al., Anal. Chem. 45.6, May 1973, 908-915.
-------
The remaining three samples produced similar chromatograms with the only
difference being relative concentration. The compounds identified by GC/MS
from this chromatographic run are presented in Figures 2 and 2A which are
simply a repeat of Figure 1 with the time scale expanded for easier reading.
To illustrate the chromatographic resolution obtained using the OV-17
glass capillary column, Figure 3 shows a narrow portion of the chromatogram
displaying the molecular weight 202 region. In addition to the reconstructed
gas chromatogram, a mass chromatogram (display of mass 202 versus time) is
also presented. Two large peaks can been seen to elute in this region. The
use of standards has confirmed that the first peak is fluoranthene and the
second is pyrene. An unknown molecular weight 202 polynuclear aromatic
hydrocarbon elutes between these two materials. Figure 4 and 6 show the mass
spectra obtained for fluoranthene and pyrene respectively. In examining
these two spectra, it is obvious that the molecular weight of the compounds
are both 202 due to the prominant peak at this mass and a half mass peak or
doubly charged molecular ion peak at mass 101. Differences in these two
spectra are very subtle, such as small variations in the relative intensity
of mass peaks. This similarity of isomers is a common problem when dealing
with the mass spectra of polynuclear aromatic hydrocarbons. It is fortunate
that the difference in retention time on this OV-17 column is dramatic and
that this difference can be used to determine which compound of molecular
weight 202 is fluroanthene and which is pyrene.
Figures 5 and 7 show the computer library searches of the two spectra
displayed in Figures 4 and 6 respectively. It is interesting that the
library search which takes into account variations in mass peak relative
intensity, correctly identifies fluoranthene and pyrene as the first choices
in these two cases. This is, however, just coincidental since in many
other cases an incorrect assignment by the library has been made, such as
identifying fluoranthene as pyrene. In all cases, these two materials are
the first two compounds identified and have the highest probability of
match, minimizing the chance of error. Several other compounds which have
molecular weights of 202 and also have mass 202 as their largest peak are
listed in the two library search outputs.
-------
FIGURE 1 . RECONSTRUCTED GC FROM 60 METER OV-17 GLASS CAPILLARY COLUMN
GC/MS (SAMPLE 367)
u
500
16:40
T
33:20
1500
50:00
225288.
2000
66:40
SCAN
TIME
-------
FIGURE 2. RECONSTRUCTED GC FROM 60 METER OV-17 GLASS CAPILLARY COLUMN GC/MS,
SAMPLE 367 (TIME SCALE EXPANDED)
-188.0
01
00
800
26:40
33:20
-------
FIGURE 2. RECONSTRUCTED GC FROM 60 METER OV-17 GLASS CAPILLARY COLUMN GC/MS,
SAMPLE 367 (TIME SCALE EXPANDED)
Continued
Page 2
225280.
J=
O-
1397
(U
O
JD
i-
03
O
O
2271
2039
2178
1400
46:40
1600
53:20
1800
60:00
2000
66:40
—I
2200
73:20
SCAN
TIME
-------
.0-1
202 J
RIG.
1700
56:40
FIGURE 3
60 METER OV-17 GLASS CAPILLARY RECONSTRUCTED GC AND
MASS CHROMATOGRAM FOR M/e = 202 REGION
1735
! Fluoranthene
MW = 202
! \
1723
MASS CHROMATOGRAM M/e = 202
$833
Pyrene
| MW = 202
Unknown MW = 202
PAH
1750
58:20
1800
60:00
1850
61:40
32832.
202.860
± 0.500
81280.
1900 SCAN
63:20 TIME
-------
t . orcuinun run rLUUKHm ntiNt hKUM UV-I/ LAKiLLAKY RUN (SCAN 1736)
1W0.0-I
202.2
r 32512.
50.0-
Fluoranthene
MW = 202
101
78.2
88.2
67.2
'I'l'
tt/E
60
80
.0
111,1 12.6. 1
i ~t^i 'I ..... j'l'l'l'l'i I'
1^.2 |h; 187.3. .. li
I'f'ft i [-rM'J'l I i i ci I'f'lldi
100
J20
160
r
180
260
220
240
-------
LIBRARY SEARCH DflTfi: S77802367 *l?3b BfiSE M/E: 202
84/17/78 9:44:08 + 57:52 CALI: DC0417 * 2 RIC: 77855.
SAMPLE: 5UL
25409 SPECTRA IN LIBRARYNB SEARCHED FOR MAXIMUM PURITY
62 MATCHED AT LEAST 4 OF THE IS LARGEST PEAKS IN THE UNKNOWN
REDUCTION: PKS/100 AMU: 40; WINDOWS: 50, 7
PRE-SEARCH: ENTRIES TO PASS: 53; SAMPLE PEfiKS: 16
MAIN SEARCH: MASSES: 50 - 550; NORM INT: 25; RATIO FACTORS: 2.8, 1.0
RANK IN NAME
1 1841 FLUORANTHENE
2 1646 PYRENE
3 5813 7H-FURO\3,2-GVM\BENZOPYRAN~7~ONE,9-HYDROXY-
4 21242 DIBENZOTHIOPHENE,l,2,3,4-TET2AHYDRO-8-METHYL-
'5 2408 7H-FURO\3,2-G\\l\BF.NZOPYRAIi-?-OME,4-HYDROXY-
RANK FORMULA M.UT B.PK PUR IT/ FIT RFIT
1 C16.H1B 202 282 898 973 916
2 C16.H10 . 202 202 896 974 915
3 C11.H6.04 202 202 624 852 647
4 C13.H14.S 262 202 591 722 686
5 C11.H6.04 202 282 559 775 634
FIGURE 5. LIBRARY SEARCH FOR SCAN 1736
12
-------
FIGURE 6 . SPECTRUM FOR PYRENE FROM OV-17 CAPILLARY COLUMN RUN (SCAN 1834)
50.0-
n/E
10
78.1
88.0
,67.2 ,,, . _, I
1 1 1 1 . 1 J 1 | i I ill | I . • L
60 80 Id
&.V
.0
| I6f2 122.1 135.2 15Jr,! 163-2 if/ * 1&?-2 il
_, 1 1 , | j , 1 . | 1 1 1 1 1 , 1 , . | . . 1 J 4_l / . 1 j J 1 . . , . ij 1 j , . 1 1 J 1 J J_ _
10 120 140 160 180 20(
Pyrene
MW = 202
218.2
1 I
a ' l .1 1 i 1 1 1 1 1 1 1
» 220 240
18816.
-------
LIBRARY SEARCH ifiTP.; S7760236? *;834 BASE M/E: 282
04/17/78 9:44:60+61:88 CALi:?C0417* 2 RIC: 51567.
'SAMPLE: SUL
25409 SPECTRA IN LIBRARYNB SEARCHED FOR MAXIMA PURITY
74 MATCHED AT LEAST 4 OF TK~. 16 LARGEST PEfif-'.S IN THE UNKNOWN
REDUCTION: PKS/100 AMU: 42; yiriLOUS: SB, 7
PRE-SEARCH: ENTRIES TO PASS: 50s SAfFLE PEAKS: 16
MAIN SEARCH: MASSES: 53 - 559; NORM INT: 25; RfiTIO FACTORS: 2.8, 1.8
RANK IN NAME
1 1646 PYRENE
2 1841 FLUORANTHENE
3 9082 PYRENE,4,5-DIHYDRQ-
4 5813 7H-FUROx3,2-G\MxDENZGPYRAN-r-OHE,9-hVDROXv/-
5 21242 DIBENZOTH10PHENE, l,2,3,4-TETRftHYDRO-8-H£THYl.-
RANK FORMULA M.UT B.PK FURHY FIT RFIT
1 C16.H10 232 262 871 991 872
2 C16.H10 202 282 863 982 835
3 C16.H12 204 232 632 ?26 663
4 C11.H6.04 202 2S2 573 829 589
5 C13.HI4.S 232 232 572 742 657
FIGURE 7 . LIBRARY SEARCH FOR SCAN 1834
14
-------
Figure 3 illustrates the high resolving power of this OV-17 glass
capillary column; however, it can be seen that after the elution of pyrene
the peaks begin to broaden and no significant concentration of materials
is detected. The region where chrysene, molecular weight 228,. should elute
was examined using data gathered on the OV-17 column and only broad chroma-
tographic peaks were detected. No positive identification or quantitative
determination could be made on the chromatographic peaks which
resulted from chrysene eluting from the OV-17 capillary column.
To improve the molecular weight range which could be determined, a
shorter and less polar OV-101 capillary column was chosen. In this case
a 30 meter glass wall coated OV-101 column was used in an attempt to elute
higher molecular weight polynuclear aromatic hydrocarbons. Figures 8 and 9
show the reconstructed gas chromatogram produced using the OV-101 capillary
column on sample #367. All samples were run with #367 used for illustration.
Figure 8 shows the total reconstructed chromatogram on a single page and
Figure 9 shows the time expanded scale chromatogram to better illustrate
the chromatographic resolution and to allow for the identification of the
various compounds detected. The early part of the chromatogram shows the
column resolution to be less than that achieved with the OV-17 capillary
column. The latter part of the chromatogram shows that the molecular weight
228 components (chrysene et.al.) as well as the molecular weight 252 components
(benzopyrene et.al.) can be determined.
In an attempt to make a direct comparison of the chromatographic
resolution, Figure 10 shows the same molecular weight region for fluor-
anthene and pyrene as illustrated in Figure 3. The OV-101 capillary column
gives adequate resolution for fluoranthene and pyrene and is sufficient for
identification and quantitation. Low concentration components in the early
part of the chromatogram are not as well resolved and if their determination
is desired the use of the longer OV-17 capillary column would be required.
It was determined by the use of standards, that the OV-101 capillary
column could not separate chrysene and triphenylene. This separation and
determination may be possible by using a shorter OV-17 capillary column
15
-------
FIGURE 8. RECONSTRUCTED GC FROM 30 METER OV-101 GLASS CAPILLARY COLUMN
GC/MS, SAMPLE #367
188.0-1
278
58432.
i
CTl
BIC I
178
373
445
588
16:40
636
1045
WJUVVA~~»
1337
M,72 1285
1618
1716
33:28
1508
58:80
SCAN
TINE
-------
FIGURE 9. RECONSTRUCTED GC FROM OV-101 GLASS CAPILLARY COLUMN
SAMPLE #367 (TIME SCALE EXPANDED)
1-169.8
-------
FIGURE 9 RECONSTRUCTED GC FROM OV-101 GLASS CAPILLARY COLUMN
' SAMPLE #367 (TIME SCALE EXPANDED)
Continued
Page 2
CO
QJ
0)
u>
(U
o
(13
S-
O)
E
08
O)
c
-------
FIGURE 9. RECONSTRUCTED GC FROM OV-101 GLASS CAPILLARY COLUMN
SAMPLE #367 (TIME SCALE EXPANDED)
Continued
Page 3
58432.
1460
46:40
1500
56:00
1600
53:20
1700
56:40
1800
60:00
SCAN
TIME
-------
FIGURE 10. MOLECULAR WEIGHT 202 REGION OF OV-101 CAPILLARY GC/MS RUN
202 _
ro
o
BIC_
1045
Fluoranthene
MW = 202
MASS CHROMATOGRAM M/e = 202
1S84
)
1045
1035
1863
Pyrene
MW = 202
RECONSTRUCTED GC
1030
34:20
1040
34:40
1050
35:00
35:20
1070
35:40
1080
36:00
1090
36:20
2*2.06*
* 9.
11616.
1104
1100
36:40
lilt
-------
The very important molecular weight 252 region of the chrbmatogram is
shown in Figure 11. PAH compounds such as benzo(a)pyrene and benzo(e)pyrene
have molecular weights of 252. As can be seen in Figure 11 and confirmed
with standards, benzo(e)pyrene and benzo(a)pyrene are well resolved on this
chromatographic column. The benzof1uoranthene isomers are apparently not
resolved and an unknown molecular weight 252 PAH was detected. The reten-
tion time of this unknown was not comparable with any available standards.
Figures 12 and 14 show the spectra of benzo(a)pyrene and benzof1uoranthene
respectively. Like all PAH compounds, both materials give large molecular
ions and strongly doubly charged molecular ions. The library search output,
Figure 15, for benzof1uoranthene gave the correct assignment rank; however,
examination of the library search output for benzo(a)pyrene, Figure 13
which was confirmed by standard addition, gave the wrong PAH as the best
fit. In fact benzo(a)pyrene is not in the list of possibilities. It should
be noted that all of the compounds identified by the library are PAHs and
have molecular weights and base peaks of 252. It is necessary in cases
such as this that the operator make a judgement as to the adequacy of the
library search results and to use all of the information available in-
cluding such things as relative retention time and the results of standard
addition of known materials to confirm identification. In this case each
sample was spiked with a known concentration of benzo(a)pyrene to confirm
identify.
The 30 meter OC-101 capillary column gave good chromatographic resolu-
tion up to molecular weight 252; however, for higher molecular weight com-
pounds the peaks were very broad and no determination was possible. To
obtain quantitative data on materials of higher molecular weight a Dexil
D
300 glass packed chromatographic column was used. Figure 16 shows an
example of the reconstructed gas chromatogram for sample number 367 on
the Dexil column and Figure 17 shows a time expanded GC with the identifica-
tion of the various components. As expected, the chromatographic resolution
of the packed column was far inferior to that achieved using the capillary
columns. The higher molecular weight compounds (i.e., 250-300) could be
eluted with reasonable peak shape. As a comparison of the chromatographic
resolution, Figure 18 shows the molecular weight 202 region for the Dexil
column which was previously shown for the capillary column runs. Even
though the chromatographic resolution of the packed column is less, the
21
-------
FIGURE 11. MOLECULAR WEIGHT 252 REGION OF OV-101 CAPILLARY GC/MS RUN
1619
' vl
: x n
MASS CHROMATOGRAM M/e = 252
\ Benzo Fluoranthene
MW = 252
1Z17 Benz(a)Pyrene
252
1660
53:20
1650
55:00
1700
56:40
1750
58:20
1852.
Unknown
1S49 MW = 252
A
V
252.075
0.500
5920.
SCAN
TIHE
-------
FIGURE 12. MASS SPECTRUM SCAN 1716 FROM OV-101 CAPILLARY GC/MS RUN
100.0-
252.0
Benz (a) Pyrene
MW = 252
rv>
co
126.4
50.0-
112.4
54.9
73
.1
10
M/E
50
.2
207.0
134.2
l""l " [' ' I' "I1 t "'['"'i '—i—i-Y"—| >-| 1 1 p
100 150
228.1
217.1
239.2
200
779.
268.1
281.1
i ...... i — ' — i ' ' ' f-"» — i
; 250
•
-------
LIBRARY SEARCH MTA: 5776623678 *17t6 BASE M/E: 252
05/07/78 9:04:00 + 57:12 CALI: DC0506 * 1 RIC: 4711.
SAMPLE: 5UL ORIGINAL SAMPLE
25409 SPECTRA IN LIBRARYNB SEARCHED FOR HflXIMUM PURITY
62 MATCHED AT LEAST 4 OF THE 16 LARGEST PEAKS IN THE UNKNOUN
REDUCTION: PKS/100 AMU: 48; UINDOUS: 50, 7
PRE-SEARCH: ENTRIES TO PASS: 50; SAMPLE PEAKS: 16
MAIN SEARCH: MASSES: 50 - 550.; NORM INT: 25; RATIO FACTORS: 2.0. 1.0
RANK IN NAME
1 1840 BENZ\E\ACEPHENANTHRYLENE
2 1842 BENZO\K\FLUORANTHENE
3 1839 BENZO\J\FLUORANTHENE
4 1832 PERYLENE
5 1828 BENZO\E\PYRENE
RANK FORMULA M.UT B.PK PURITY FIT RFIT
1 C20.H12 252 252 713 963 724
2 C20.H12 252 252 709 . 945 727
3 C20.H12 252 252 707 951 723
4 C20.H12 252 252 597 947 715
5 C20.H12 252 252 680 926 714
FIGURE 13. LIBRARY SEARCH SCAN 1716
24
-------
FIGURE 14. SPECTRUM 1618 FROM OV-101 CAPILLARY GC/MS RUN
190.0-1
ro
en
50.0-
252.0
Benzo Fluoranthene
MW = 252
903.
56.
69
.1
112.4
83.0
130.
207.
I5
f'2
'1
228.1
JJ
281.2
M/E
50
| •
150
200
250 '
-------
LIBRARY SEARCH DhTfl: 3770023670 *1613 BASE M/E: 252
05/07/78 3:04:00 + 53:56 ML I: DC0506 * 1 RIC: 5343.
SAMPLE: 5UL ORIGINAL SAMPLE
25409 SPECTRA IN LIBRARYHB SEARCHED FOR MAXIMUM PURITY
54 MATCHED AT LEAST 4 OF THE 16 LARGEST PEAKS IN THE UNKNOWN
REDUCTION: PKS/100 AMU: 48; UINDGLJS: 50, 7
PRE-SEARCH: ENTRIES TO PASS: 50; SAMPLE PEAKS: 16
MAIN SEARCH: MASSES: 50 - 558; NORM INT: 25; RATIO FACTORS: 2.0, 1.0
RANK IN NAME
•;1 1842 BENZONKXFLUORANTHENE
*2 1828 BENZO\E\PYRENE
.3 1840 BENZ\E\ACEPHENANTHRYLEHE
,4 1832 PERYLENE
,•5 1839 BENZONJXFLUORANTHENE
RANK FORMULA M.UT B.PK PURITY FIT RFIT
1 C20.H12 252 252 734 955 741
2 C20.H12 252 252 729 956 737
3 C20.H12 252 252 725 959 735
4 C20.H12 252 252 725 956 734
5 C20.H12 252 252 721 955 736
FIGURE 15. LIBRARY SEARCH SCAN 1618
26
-------
106.0-!
FIGURE 16. RECONSTRUCTED GC FROM DEXIL 300 PACKED COLUMN GC/MS RUN
(SAMPLE 367)
245768.
r\>
BIC
•4^
500
16:40
~i 1 r
1600
33:20
1500
50:00
2000 '
66:40:
SCAN I
TIME j
-------
FIGURE 17 RECONSTRUCTED GC FROM DEXIL 300 PACKED COLUMN, SAMPLE 367
(TIME SCALE EXPANDED)
-------
FIGURE 17. RECONSTRUCTED GC FROM DEXIL 300 PACKED COLUMN, SAMPLE 367
(TIME SCALE EXPANDED)
CONTINUED
PAGE 2
ro
vo
/
900
30:00
-------
FIGURE 17. RECONSTRUCTED GC FROM DEXIL 300 PACKED COLUMN, SAMPLE 367
(TIME SCALE EXPANDED)
CONTINUED
PAGE 3
24576*.
CO
o
CD
C
o>
cu
Q-
CM
00
CM
CM
T3
O)
to
-0
CO
r>,
JC
4->
CU
1624
"""X
co
CM
II
O)
o
IQ
N
cu
cu
c
cu
CO
•t
CM
O
C
cu
-a
c
-a
cu
o
to
to
cu
cu
I
o
N
C
cu
.0
1668
2*377
2163
2228
2290 2338 2374
1700
1908
63:20
2080
66:40
2168
76:3(3
2260
73:29
2360:
76:40
SCAN
TIME
ID
-------
FIGURE 18. MOLECULAR WEIGHT 202 REGION OF DEXIL 300 PACKED COLUMN GC/MS
$88.0-i
J
202 _
1
20416.
Fluoranthene
MW = 202
MASS CHROMATOGRAM M/e = 202
1267
/I
Pyrene
MW = 202
Unknown PAH
MW = 202
V
202.660
± 0.500
sue
1203
1220
.••*"'''*-
—I
1200
40:00
1220
40:40
RECONSTRUCTED GC
1240
41:20
1260
42:00
1230
42:40
66048.
1300 SCAN
43:20 TIHE
-------
fluoranthene and pyrene are clearly separated. If one studies the molecular
weight 252 region of the packed column chromatogram, Figure 19, it is
obvious that the benzo(a)pyrene and benzo(e)pyrene which were resolved on
the OV-101 capillary column, are now fused producing a peak which has only
a slight shoulder. The benzof1uoranthene and perylene are still resolved.
The lack of separation for B(a)P and B(e)P would not allow for identifica-
tion or quantitative determination of the individual components. Using
the packed GC column it was also found that chrysene, triphenylene and
benz(a)anthracene all coelute. Spectra obtained from the packed column
are similar to those obtained using the capillary columns. Figure 20
shows the spectrum for benzof1uoranthene, molecular weight 252, obtained
from the packed column analysis and Figure 22 shows the spectrum for a
mixture of the benzo(a) and benzo(e)pyrenes. The library search results
for the respective spectra are given in Figures 21 and 23. The library
search shows the compounds to have molecular weights of 252 and base peaks
of 252; however, it is difficult for the library to distinguish the isomeric
forms of these polynuclear aromatic hydrocarbons. It is imperative that
the gas chromatographic retention time be accurately measure, preferably
relative to an internal standard, to confirm the identification of a given
PAH.
3.2 6C/MS Analysis of Thin Layer Chromatographic (TLC) Plates
After preparation, by extraction with one milliliter of methylene
chloride (see Section 2.2 for details), the TLC samples were chromatographed
ff\
on a Dexil 3QQP'column using the conditions detailed in Appendix 3. The
reconstructed gas chromatogram for the extract, Figure 24, shows a Chromato-
graphic peak which results from benzo(a)pyrene, confirmed by standard addi-
tion. The chromatogram itself is relatively PAH free, since all of the other
peaks detected are either phthalate or adipate esters, probably impurities
from the TLC plate or the solvents used for extraction. Figure 25 shows
the mass spectrum obtained at scan 1753 which is consistent with benzo(a)-
pyrene.
The ability to quantitate a TLC spot using the procedures presented
previously are somewhat questionable. In an attempt to quantitate the TLC
data, a sample was prepared from the 20 ng standard B(a)P channel and
chromatographed using the same conditions as for samples from channels 8
32
-------
FIGURE 19. MOLECULAR WEIGHT 252 REGION OF DEXIL 300 PACKED COLUMN GC/MS
1730
Benzof1uoranthene
1416
252
CO
00 '
1778
MASS CHROMATOGRAM M/e = 252
Benzo Pyrene
B(a)P & B(e)P
1587
RIC_
RECONSTRUCTED GC
5936.
252.075
t 0.500
71040.
1446
1400
46:40
1500
50:00
1600
53:20
1700
56:40
J
1800
60:00
1900 SCAN *
63:20 TIHE *
-------
dOT.
100.0-1
25
.9
5936.
Benzofluoranthene
MW = 252
CO
50.0-
13-® 150,0 163.1
ll I 1 ... li III. ..I .
ll i I » ~.i—f~
224.0
niii.
*
M/E 50
150
200
250
-------
LIBRARY SEARCH DATA: S77S02367 *1?38 ErtSE rt-£: 252
04/21/78 13:34:88+57:48 CALI: DC9421 * 1 RIC: 24319.
SAMPLE: 3UL UITH 3 INTERNAL STANDARDS
25499 SPECTRA IN LI9RARYNB SEARCHED FOR ^ftXIWI PURITY
72 NOTCHED AT LEAST 4 OF THE 16 LARGEST PEAKS IN THE UNKNQUN
REDUCTION: PKS/108 AMU: 43; UINl'DUS? 50, 7
PRE-SEARCH: ENTRIES TO PASS: 59; SArfi'LE PEAKS: 16
MAIN SEARCH: MASSES: 53 - 558; NORM INT: 25; RATIO FACTORS: 2.0, 1.8
RANK IN NAME
'1 1842 BEMZO\K\FLUORANTHENE
2 1839 BENZO\J\FLUORANTHENE
.;3 1840 BENZ\E\ACEPHENANTHRYLENE
'4 1832 PERYLENE
5 1828 BENZO\E\PYRENE
RANK FORMULA M. UT B.PK PURITY FIT RFIT
1 C20.H12 252 252 835 969 824
2 C20.H12 252 252 302 963 827
3 C20.H12 252 Z'52 f'83 951 817
4 C20.H12 252 252 783 945 817
5 C20.H12 252 252 775 941 882
FIGURE 21. LIBRARY SEARCH SPECTRUM 1730
35
-------
FIGURE 22. SPECTRUM 1778 FROM DEXIL 300 PACKED COLUMN GC/MS RUN
25 .9
r
4544.
Benzo Pyrene MW = 252
B(a)P & B(e)P
to
126.1
113.1
54.7
lui
69.1
Ji
tt/E 50
82.9 99.2
100
134.1 149.1 163.2
150
224.1
.
lln, III
200
250
.llill
-------
LIBRARY SEARCH DATA: S77602367 *1778 BASE M/E: 252
04/21/78 13:34:09 +59:16 CftLI: DC0421 * 1 RJC: 20575.
SAMPLE: 3UL UITH 3 INTERNAL
25489 SPECTRA IN LIBRARYNB SEP.RCHED FOR MAXIMUM PURITY
72 MATCHED AT LEAST 4 CF THE 15 LARGEST PEAKS IN THE UNKNOljN
REDUCTION: PKS/103 AMU: 40; WINDOWS: 50, 7
PRE-SEARCH: ENTRIES TO PASS: 58; SAMPLE PEAKS: 16
MAIN SEARCH: MASSES: 50 - 550; NORM INT: 25; RATIO FACTORS: 2.0, 1.0
RANK IN NAME
1 1839 8ENZO\JNFLUQRANTHENE
2 1842 BENZOXKNFLUQRANTHENE
3 1832 PERYLENE
4 1840 BENZ\ENACEPHENANTHRYLEHE
5 1828 BENZO\E\PYRENE
RANK FORMULA M.UT B.PK PURITY FIT RFIT
1 C20.H12 252 252 799 365 810
2 C20.H12 252 252 782 967 809
3 C20.H12 252 252 774 950 803
4 C20.H12 252 252 773 954 802
5 C20.H12 252 252 741 926 791
FIGURE 23. LIBRARY SEARCH SCAN 1778
37
-------
100.
1
co
CO
FIGURE 24. GC/MS OF TLC CHANNEL 8
MASS CHROMATOGRAM M/e =252
Benz(a)pyrene
252.W5
RIC
RECONSTRUCTED GC(DEXIL PACKED COLUMN)
A I
1
560
16:40
1000
33:20
1500
50:00
2000
66:40
-"—i
5(
83:»
-------
FIGURE 25. MASS SPECTRUM TLC CHANNEL 8 SCAN 1753
189.0-1
252
1786.
CO
10
58.8-
Benz(a)pyrene
MW = 252
M/E
63 -TC 87
^Jgi^HlL^Ll'
50 108
i
1
,
JJ08
inill ; ;l
.3
1
i i
26
*y\c
'?? i7\7 1?1 T • nlli
1
158
288
258
-------
and 12. A persistent clogging of the syringe needle with the finely divided
particles from the TLC plate occured for both the standard and track 12.
In an attempt to correct this problem, the sample was filtered through
a millipore syringe filter. The procedure resulted in some apparent loss
of sample and unknown dilution effects. In the future, it would be
advisable to add an internal standard to the extraction solvent so that
subsequent losses of solvent or sample due to processing can be com-
pensated for.
B(a)P was detected in both of the TLC samples and the standard. It
was not detected in a blank taken from an equivalent region of the TLC
plate in an empty track.
3.3 Results of GC/MS Analysis of Coke Oven Samples
Figure 26 is a tabulation of the GC/MS results from the coke oven
samples. The compounds listed were taken from the suggested list of
PAH components to be measured, contained in the original task scope of work.
Figure 26 gives the amount of each compound in mg/ml of the "as received"
solution. The techniques used for quantisation are detailed in Section 4.0.
The following information relates to Figure 26, explaining the various
results presented therein. As previously discussed, chrysene and tripheny-
lene, both molecular weight 228, were unresolved on the 30 meter OV-101
column. Due to this lack of separation, only the total of these compounds
is given. One molecular weight 228 component which was found in the samples
could not be identified using standards or published relative retention
times. Currently, no standard is available for benzo(c)phenanthrene,
molecular weight 228; however, published relative retention time data in-
dicates that it should be chromatographically resolved using a
packed column.
Benzo(j)aceanthrylene, molecular weight 252, was not available as a
standard and no reliable relative retention time data could be found;
therefore, this material could not be confirmed. Other PAH compounds such
as 7, 12-dimethylbenz(a)anthracene and 3-methylcholanthrene were not
available as standards, however their mass spectra should be quite unique.
The molecular weights of these two materials are relatively free of inter-
ferences and the methyl substitution should show some loss of 15 mass units from
molecular ion giving unique mass spectra. To confirm the presence of
40
-------
FIGURE 26. GC/MS RESULTS FROM COKE OVEN SAMPLES
Compound
Naphthalene
Fluoranthene
Unknown
Pyrene
Chrysene
Triphenylene
Unknown
Benz (a) anthracene
*Benzo (c) phenanthrene
Benzo (a) pyrene
Benzo (e) pyrene
Benzofluoranthene
Unknown
*Benz (j) aceanthrylene
Perylene
*7,12-Dimethylbenz (a) anthracene
Dibenzo (c, g) carbazole
*3-Methy 1 chol anthrene
Indeno (1,2, 3-cd) pyrene
Unknown
Unknown
*Dibenzo (a, h) anthracene
*Dibenzo (a, j) anthracene
Unknown.
*Dibenz acridine
*Dibenzo (a, h) pyrene
*Dibenzo (a, i) pyrene
Molecular
Weight
128
202
202
202
228
228
228
228
228
252
252
252
252
252
252
256
267
268
276
276
276
278
278
278
279
302
302
(.Concentration (mg/ml}
367 371 366 364
4.2 4.6 6.8 0.34
1.3 4.2 2-9 0.075
0.17 0.55 0.55 T
0.9 3.0 1.8 T
0.3 1.2 0.8 0.028
0.007 0.09 0.1 T
0.30 1.0 0.8 0.024
0.044 0.14 0.18 T
0.24 0.7 0.65 0.02
0.12 0.3 0.38 0.01
0.47 1.7 1.2 0.043
0.07 0.27 0.24 T
_
0.14 0.17 0.27 ND
ND ND ND ND
ND ND ND ND
ND ND ND ND
0.10 0.22 0.42 T
0.055 0.14 0.19 T
0.01 0.055 0.05 ND
0.055 0.16 0.23 ND
0.055 0.80 0.12 ND
0.018 0.022 0.048 ND
ND ND ND ND
0.12 0.36 0.50 ND
Comments
Chrysene/Tripheny-
lene unresolved
RRT used for ID
No RRT to confirm
MS should be unique
MS should be unique
Have 1, 2, 3, 4 and
1 , 2, 5, 6 Standards
MS unique
Unresolved
*Standard not available (identification based on published RRT and MS)
ND = Not detected < 0.005 mg/ml
T - Trace < .01 mg/ml but detected
41
-------
these species mass chromatograms of their molecular ions were displayed
together with the loss of methyl (M-15). The displays were compared to
determine if the two masses gave an inflection at the same scan number
and at a reasonable retention time. No such coincident maximization of
intensities occurred and therefore it was determined that 7, 12-dimethylbenz
(a) anthracene and 3-methylcholanthrene were not present in any of the
samples.
At molecular weight 278, three individual chromatographic peaks were
identified. Two isomer standards of dibenzoanthracene were available, the
1, 2, 3, 4 and the 1, 2, 5, 6 substituted materials. Chromatographically
it was found that these two isomers could be resolved; however, since the
exact AH and AJ standards were not available, it was impossible to determine
which isomers are actually present.
Nitrogen containing PAH compounds produce very unique mass spectra
because if an odd number of nitrogen atoms are present in the molecule
it will have an odd molecular weight. Only low molecular weight nitrogen
containing compounds were detected in the samples, the higher molecular
weight species were not found to be present.
Dibenzopyrene, being a high molecular weight PAH, elutes from the Dexil
packed column as a very broad peak. Due to the broad nature of the
chromatographic peak and the lack of standards to determine if indeed a single
dibenzopyrene would give such a peak, it is impossible to determine which
isomers are present. It is doubtful from this data that dibenzopyrene in
its various isomeric forms could be determined using gas chromatography.
It may be necessary to use some other technique, such as high pressure
liquid chromatography, to determine dibenzopyrenes.
Figure 27 shows a list of the samples analyzed from the Republic Steel
coke oven. Blanks for each of the solvents and the filter media were run
using the same conditions as employed for the samples. The blanks were
extremely clean with only very low levels of contaminants detected. The
levels of contamination would not contribute to any of the polynuclear
aromatic hydrocarbon measurements made.
42
-------
Figure 27. Coke Oven Samples Analyzed by GC/MS
Sample No.
Identification
S77-002-364
S77-002-366
S77-002-367
S77-002-371
S77-002-372
S77-002-373
S77-002-374
S77-002-378
Cyclohexane rinse of first
impinger and pipe
Methylene chloride rinse of
first impinger
Acetone rinse of first impinger
Filter extract in cyclohexane
Methylene chloride blank
Cyclohexane blank
Filter media blank extract
Acetone blank
43
-------
All samples were also analyzed by GC/MS after separation using the
liquid chromatography procedure detailed in Appendix 1. An overall loss
in sample concentration was observed as a result of the procedure as well
as preferencial loss of the lower molecular weight PAHs. It was deter-
mined that by improving gas chromatographic resolution using capillary
columns and avoiding sample preparation which involved liquid chromatography,
more accurate results are possible.
4.0 METHODS OF IDENTIFICATION & QUANTITATION
4.1 Methods for PAH Identification
Polynuclear aromatic hydrocarbons are best identified, using GC/MS as
isomer classes based on molecular weight. Once a chromatographic separation
has been achieved and the large quantity of mass spectral data is stored in
a computer based data system, individual mass chromatograms for specific
molecular weights can be displayed. After it is established which classes
of PAH compounds are present in a given sample standards of the individual
components or relative retention time data can be used to identify specific
isomers.
For qualitative identification of specific isomers standard addition
of known PAH compounds were made to each sample. An increase in GC peak
intensity without broadening was used as identification confirmation. Once
a single isomer in a given molecular weight region is identified, relative
retention time data can be accurately applied for identification of isomers
where standards are not available.
The standards which are currently available at TRW are listed in
Appendix 2. Since there are virtually an unlimited number of polynuclear
aromatic hydrocarbons and a very finite number of standards available,
interferences of unresolved coeluting compounds are very possible. This
coelution is the limiting factor in quantisation of these samples. Improve-
ment in chromatographic resolution by the use of capillary columns and
improvement in detector selectivity with the mass spectrometer have reduced
the possibility of interferences, but not eliminated them.
44
-------
4.2 Methods for PAH Quantitation
The method used for quantitation of individual polynuclear aromatic
hydrocarbon compounds was internal standardization. As previously
described a known quantity of three internal standards, 9-methylanthracene,
9-phenylanthracene and 9, 10-diphenylanthracene were added to each sample.
The mass spectrometer response factors for the various polynuclear aromatic
hydrocarbons vary slightly. When standards were available, response factors
were measured. When not available they were extrapolated by molecular
weight from those which were measured. For any given mass spectrometer
response factors should be individually determined using available
standards. Published information tends to encourage the "easy way" which
is not appropriate when quantitation is done using a mass spectrometer.
In those cases where standard addition was used to identify compounds,
a known amount of standard was added. Calculations based on standard
addition were compared with values calculated using internal standards.
The data in most cases agreed well; however, in some cases the numbers
were quite different. When this occurred the calculations based on internal
standards were reported. No explanation is currently available on why in
some cases standard addition and internal standard results differ.
The ability of the mass spectrometer to quantitate compounds is some-
what limited. To be assured that a peak area is accurately measured, at
least ten points must be obtained over the peak profile. Since the mass
spectrometer is used in the scanning mode and requires from 1.5 to 2
seconds to complete a single scan, narrow gas chromatographic peaks have
only three to five points accross their profile. The lack of points results
in inaccurate peak area measurement and errors in the computed concentration.
Based on statistics a precision of +30% is about all one can expect from
narrow gas chromatographic peaks. Using repeat experiments to determine
the actual precision of the data presented in this report, a scatter of
somewhat more than 30% was observed.
45
-------
5.0 OTHER TECHNIQUES FOR THE DETERMINATION OF POLYNUCLEAR AROMATIC
HYDROCARBONS AND RECOMMENDATIONS FOR IMPROVEMENTS
One of the weakest areas in the determination of polynuclear aromatic
hydrocarbons by GC/MS is high molecular weight compounds. As discussed
earlier, PAHs which have molecular weights above 300 are poorly resolved
chromatographically. Capillary columns cannot operate at high enough
temperature to elute high molecular weight materials without irreversable
damage. In these cases, an alternative technique is necessary to achieve
separation, identification, and quantisation. Several techniques are
available to the analyst, most noteable of these are thin layer chromato-
graphy and high pressure liquid chromatography.
Techniques have been developed by Athur D. Little Inc.-*, for the
general screening of suspected polynuclear aromatic hydrocarbon samples
using their inherent property of fluorescence. The technique can
suggest the presence of PAH compounds in complex mixtures but cannot
quantitate or identify individual components. An extension of this flucre-?
scence technique is a method developed by the Environmental Protection
Agency, Environmental Monitoring and Support Laboratory at Research Triangle
Park, The technique allows for the specific identification and quantitation
of benzo(a)pyrene in complex mixtures. It uses thin layer chromatography
to achieve at least crude separation of the polynuclear aromatic hydrocarbons
and then relies on the specific fluroescence of benzo(a)pyrene for quanti-
tation. The specific fluorescence of benzo(a)pyrene is necessary since
the TLC procedure does not totally separate all PAH compounds. The fluore-
scence method uses an excitation wavelength of 388 nm and emission wavelength
of 430 nm which provides for a very selective measure of benzo(a)pyrene.
Other techniques for individual polynuclear aromatics have also been
published. Many of these techniques are designed to identify and quantify
single components in complex mixtures. High pressure liquid chromatography
has been utilized by many workers for the determination of PAH compounds
in environmental samples. Most of this work has been done using UV detection
because all of the polynuclear aromatic hydrocarbons absorb strongly in the
ultraviolet. UV absorption is useful as a general detector for PAH com-
pounds however, it relies on good chromatographic separation which is not
always possible in complex mixtures.
o
Sensitized Fluorescence for the Detection of Polycyclic Aromatic Hydrocarbons
E.M. Smith and P.L. Levins, ADLINC, Draft Report
46
-------
A more recent development is the selected wavelength fluorescence
detector for the liquid chromatograph. Modern fluorescence detectors allow
for variation of both the excitation and detection emission wavelengths.
As an example Figure 28 shows the application of HPLC to the analysis of
benzo(a)pyrene. The chromatogram is the result of using heptane as the
mobile phase and y-porasil as the stationary phase. Peaks 9 and 10 represent
benzo(a)pyrene and benzo(e)pyrene respectively. The top trace in the
figure shows the chromatogram as detected using a fluorometer with an
excitation wavelength of 395 nm and a detection wavelength of 405 r\m. This
data shows that even when materials coelute and resolution is incomplete,
individual isomers can be identified by fluorescene.
There is a parallel between the use of fluorescence detection for HPLC
and the use of a mass spectrometer as a GC detector. The mass spectrometer
can be specific through the selection of individual.ions which are con-
sistent with a given component; similarly, the fluorometer can be quite
specific by selecting excitation and emission wavelengths specific for a
particular PAH. The gas chromatograph has an advantage in that it is
simpler to apply and provides good resolution for low molecular weight PAHs
(to about 280). The liquid chromatograph does not require the sample to be
volatile and therefore molecular weight and boiling point are not limiting
factors. LC separation is accomplished by some individual characteristic
of the compound such as solubility, polarity, and/or molecular size.
An alternative column for the LC determination of benzo(a)pyrene is
shown in Figure 29. The use of a y-Bondapak C,g column and gradient elution
(HpO -»• CH3CN) provides a very clean separation of benzo(a)pyrene. More work
needs to be done to establish the types of columns which can be used to
resolve various polynuclear aromatic hydrocarbons. The fluorescence
characteristics of various PAH compounds are required so that when specific
determinations are to be made, the detection system can be specific. In
summary, the liquid chromatograph appears superior to GC/MS for both
separation and quantisation of polynuclear aromatic hydrocarbons above
molecular weight 280.
47
-------
1 HEPTANE I.OJ2U
\ min
[JPORASIL
UV 254nm 0.02 A
FIGURE 28. HPLC Separation of PAH Standards Using UV (Lower
Trace) and Selected Wave Length Fluorescence(Upper Trace)
48
-------
UV 254nm.
0.05A .
_.
2-0 ml/min
"pBONDAPAK C
420 8 X
X 360n.m BP
FIGURE 29. HPLC Separation of PAH Standards Using UV (Upper Trace)
and Selected Wave Length Fluorescence (Lower Trace)
-------
APPENDIX 1
LEVEL 1 LIQUID CHROMATOGRAPHY SEPARATION PROCEDURE4
Column: 200 mm x 10.5 mm ID, glass with Teflon stopcock.
Adsorbent: Davison Silica Gel, 60-200 mesh, Grade 950, (Fisher Scientific
Company). This adsorbent is activated at 110°C for two hours
just prior to use. Cool in a desiccator.
PROCEDURE FOR COLUMN PREPARATION
Wet pack the chromatographic column, plugged at one end with glass wool,
with 6.5 grams of freshly activated silica gel. A portion of properly
activated silica gel weighing 6.0 +0.2 g occupies 9 ml in a 10 ml graduated
cylinder. Vibrate the column for a minute to compact the gel bed. Pour
pentane into the solvent reservoir positioned above the column and let the
pentane flow into the silica gel bed until the column is homogeneous
throughout and free of any cracks and trapped air bubbles.* The total
height of the silica bed in this packed column is 10 cm. The solvent
void volume of the column is 2 to 4 ml. When the column is fully prepared,
allow the pentane level in the column to drop to the top of the silica bed
so that the sample can be loaded for subsequent chromatographic elution.
PREPARATION OF THE SAMPLE
At room temperature evaporate the solvent from an aliquot of solution
containing at least 10 to 100 mg of sample. The preferred sample weight is
100 mg. Weigh this sample in a glass weighing funnel. In order to facili-
tate transfer of the sample, add 0.5 to 1.0 g of activated silica gel to
the sample and carefully mix this with the sample using a micro-spatula.
Table 1 shows the sequence for the chromatographic elution. In order
to ensure adequate resolution and reproducibility, the column elution rate
is maintained at 1 ml per minute.
*A convenient device for the elimination of gel bed cracks and air bubbles
is acetone coolant, which is subsequently referred to as the ACE B method.
It consists of a paper towel wound loosely around the glass column along
the region of the crack or bubble; the paper towel is periodically moistened
with acetone. The acetone evaporation cools the region and dissipates the
bubble or crack.
4IERL-RTP Procedures Manual: Level 1 Environmental Assessment EPA-600/2-76-160a
50
-------
Table 1. Liquid Chromatography Elution Sequence
No.
Fraction
1
2
3
4
5
6
7
Solvent Composition
Pentane
20% Methylene chloride in pentane
50% Methylene chloride in pentane
Methylene chloride
5% Methanol in methyl ene chloride
20% Methanol in methyl ene chloride
50% Methanol in methylene chloride
Volume
Collected
25 ral
10 ml
10 ml
10 ml
10 ml
10 ml
10 ml
LOADING SAMPLE ON THE COLUMN
Quantitatively transfer the sample into the column via the weighing
funnel used for sample preparation; a micro-spatula can be used to aid in
the sample transfer. Rinse the funnel* with a few ml of pentane to com-
plete the quantitative sample transfer. (Note: Do not rinse with methylene
chloride because this solvent will cause the aromatics to elute with the
paraffins.) Add the solvent slowly to minimize disturbing the gel bed and
eliminate the trapped air bubbles, particularly in the zone of the sample-
containing silica gel, by using the ACE B approach (see footnote, preceding
page). The chromatographic system is now ready for sample fractionation.
CHROMATOGRAPHIC SEPARATION INTO 7 FRACTIONS
The volume of solvents shown in Table 1 represents the solvent volume
collected for that fraction. If the volume of solvent collected is less
than volume actually added due to evaporation, add additional solvent as
necessary. In all cases, however, the solvent level in the column should
be at the top of the gel bed, i.e., the sample-containing zone, at the
end of the collection of any sample fraction.
*Save this weighing funnel for subsequent additional rinsing with the
solvents used at Interim fractions up to raethylene chloride.
51
-------
After the first fraction is collected, rinse the original sample
weighing funnel with a few ml of the Fraction Number 2 solvent (20%
methylene chloride/pentane) and carefully transfer this rinsing into the
column. Repeat as necessary for Fractions 3 and 4. Fractions 2-4 were
collected and combined, the remaining fractions were discarded.
52
-------
APPENDIX 2
Table of PAH Standards Available
Acenaphthene
Acenaphthylene
Anthracene
Azulene
Benz(a)anthracene-7, 12-dione
1, 2-Benzathracene
Benzo(a)anthracene
Benzo(b)f1uoranthene
Benzo(j)f1uoranthene
Benzo(k)fluoranthene
1, 2-Benzofluorene
2, 3-Benzofluorene
Benzo(g, h, i)perylene
Benzo(c)phenanthrene
Benzo(a)pyrene
Benzo(e)pyrene
Benzo(f)quinoline
5, 6-Benzoquinoline
Biphenyl
Carbazole
Chrysene
Coronene
1, 2, 3, 4-Dibenzanthracene
1, 2, 5, 6-Dibenzanthracene
3, 4, 5, 6-Dibenzocarbazole
Dibenzofuran
Dibenzo-p-dioxin
1, 2, 4, 5-Dibenzopyrene
Dibenzothiophene
9, 10-Dihydroanthracene
3, 6-Dimethylphenanthrene
9, 10-Diphenylanthracene
5-Dodecahydrotri phenylene
Fluoranthene
Fluorene
2-Methylanthracene
9-Methylanthracene
4, 5-Methylenephenanathrene
2-Methylphenanthrene
Naphthalene
cis, anti-4, 5, 6, 6a, 7, 8, 12b-
Octahydrobenzo(j)fluoranthene
Perylene
Phenanthrene
9-Phenylanthracene
o-Phenylenepyrene
2-Phenylnaphthalene
Pyrene
p-Quarterphenyl
Terphenyl
1, 2, 3, 4-Tetrahydroanthracene
1, 2, 3, 4-Tetrahydrofluoranthene
1,3, 5-Triphenylbenzene
Triphenylene
Truxene
53
-------
APPENDIX 3
Conditions used for GC/MS Analysis
The following is a list of the chromatographic and mass spectrometer
conditions used for the various PAH determinations discussed in the test.
Conditions for Mass Spectrometer Operation
Instrument Type: Dupont 321 accelerating voltage scanning,
magnetic sector mass spectrometer inter-
faced to an INCOS data system.
Scan Time: 1.9 seconds (0.1 second delay between
scans)
Mass Range Scanned: 50-550
Instrument Resolution: 600
Instrument Sensitivity: 1 ng of a PAH would be detected in a pure
sample
Ionizing Voltage: 70 eV
Source Temperature: 250°C
GC/MS Interface: Glass Jet at 295°C
Chromatograph Conditions for OV-17 Capillary Column Analysis
Column: Glass OV-17 wall coated capillary, 60M
long by 0.01" id.
Injector: Operated in a split mode (50:1 split ratio)
at 290°C
Temperature Program: 50 - 280°C at 2°C/min after a three minute
delay
Flow Rate: 20 cm/sec ('v-l cc/min) of Helium
Injection Size: 5 yl
Total Analysis Time: -^90 min
54
-------
Chroraatograph Conditions for OV-101 Capillary Column
Column:
Injector:
Temperature Program:
Flow Rate:
Injection Size:
Total Analysis Time:
Glass OV-101 wall coated capillary. 30M
long by 0.01" id.
Operated in a split mode (25:1 split ratio)
at 290°C
100 - 290°C at 2°C/min after a three minute
del ay
20 cm/sec (~lcc/min) of Helium
5 pi
^90 min
Chromatograph Conditions for Dexil 300 Packed Column
Column:
Injector:
Temperature Program:
Flow Rate:
Injection Size:
Total Analysis Time:
Glass packed with 3% Dexil 300won
Chromosorb W, 5' long by 2.1 mm id.
300 °C
100 - 300°C at 4°C/min
30 cc/min of Helium
1 yl
-^80 min
55
------- |