Application of the O2-doped ECD to isomer differentiation
by James Allen Campbell
A thesis submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in
Chemistry
Montana State University
© Copyright by James Allen Campbell (1984)
Abstract:
The addition of oxygen to the nitrogen carrier gas of a constant current electron capture detector (BCD)
is shown to provide increased sensitivity and isomer distinction for environmentally important
compounds such as polychlorinated biphenyls, polycyclic aromatic amines and hydroxides with
appropriate EC-enhancing tags, chloroanthracenes and chlorophenanthrenes, methylanthracenes and
methylphenanthrenes, and 2,3,7-trichlorodibenzo-p-dioxin. In most cases, the isomers of these
particular compounds can be distinguished from the other isomers based solely on the measured
response enhancements.
The oxygen doped BCD has been applied to compound identification where only partial resolution
with capillary column gas chromatography is obtained. In addition, in instances where several
compounds coelute and have the exact same retention times, the mole fraction of each component in
the unresolved peak can be determined.
Atmospheric pressure ionization mass spectrometry (APIMS) gives an indication of the ions formed
with and without the presence of oxygen in the source. For the chloroanthracenes, methylanthracenes,
and methylphenanthrenes, actual oxygen incorporation is observed when oxygen is present. In contrast,
for the polycyclic aromatic amines derivatized with trifluoroacetic anhydride (TFAA), the reaction with
oxygen to produce an induced response involves a charge transfer between the oxygen anion and the
analyte molecule.
Several TFAA derivatives of the aminoanthracenes and aminophenanthrenes were examined using
electron impact mass spectrometry and chemical ionization mass spectrometry. Negative chemical
ionization mass spectrometry with methane and isobutane as the reagent gases was evaluated as a
method of isomer differentiation of these compounds and shows considerable promise.
In order to improve the precision involved in the measurement of response enhancements, two methods
with parallel and series arrangement of the ECDs were utilized. The parallel arrangement with dual
columns represents a significant improvement in the reproducibility of response enhancement
measurements. The series detector arrangement seems to be confusing in view of possible additional
reactions in the detectors.
The addition of ethyl chloride to an electron capture detector increased the response of anthracene and
similar molecules with a low normal BCD response and those that react with the gaseous electron
through a resonance type of reaction mechanism. The addition of ethyl chloride to the detector does not
significantly increase the baseline frequency, in contrast to the addition of oxygen. The negative ions
formed in this reaction have been identified. 
©  1985
JAMES ALLEN CAMPBELL
All R ights Reserved
APPLICATION OF THE Og-DOPED ECD 
TO ISOMER DIFFERENTIATION
by
Janes Allen Campbell
A thesis submitted in partial fulfillment 
of the requirements for the degree
of
Doctor of Philosophy 
in
Chemistry
MONTANA STATE UNIVERSITY 
Bozenan, Montana
August 1984
ii
APPROVAL
of a thesis submitted by
James Allen Campbell
This thesis has been read by each member of the thesis 
committee and has been found to be satisfactory regarding content, 
English usage, format, citations, bibliographic style, and 
consistency, and is ready for submission to the College of Graduate 
Studies.
Chairperson, Graduate Committee
Approved for the Major Department
Head, Major Department
Approved for the College of Graduate Studies
Date Graduate Dean
iii
STATEMENT OF PERMISSION TO USE
In presenting this thesis in partial fulfillment of the I
requirements for a doctoral degree at Montana State University, I 
agree that the Library shall make it available to borrowers under 
rules of the Library. I further agree that copying of this thesis 
is allowable only for scholarly purposes, consistent with "fair 
use" as prescribed in the U.S. Copyright Law. Requests for 
extensive copying or reproduction of this thesis should be referred 
to University Microfilms International, 300 North Zeeb Road, Ann 
Arbor, Michigan 48106, to whom I have granted "the exclusive right 
to reproduce and distribute fcy abstract in any format."
Signature
Date
iv
DEDICATION
I would like to dedicate this thesis to the memory of my 
father who recently passed away. He had his own particular 
definition of love and will always be remembered in my thoughts and 
actions. In addition, this is dedicated to the living, my mother, 
my wife, and son whose constant encouragement made this all 
possible.
VITA
James Allen Campbell was born December 14, 1948 in Wolf Point, 
Montana, the younger of two sons of Donald W. and Harriet V. 
Campbell. He attended public schools in Montana and graduated from 
Stanford High School in May, 1966.
He majored in Chemistry at Montana State University where 
he received the Bachelor of Science degree in Chemistry in June, 
1970. He then received a Fulbright-Hays Scholarship to study at 
the University of Heidelberg, Heidelberg, West Germany. He then 
returned to Montana State University in 1973 and completed the 
Masters of Science Degree in Chemistry in August of 1974.
After working at Battelle Northwest for five years, he 
returned again to Montana State University in 1979, where he has 
been supported by graduate teaching and graduate research 
assistantships.
He is married and has one child.
He has accepted a position as a Senior Research Scientist with 
Dow Chemical Company, Midland, Michigan, beginning in December 
1983.
vi
ACKNOWLEDGEMENT
The author would like to express his sincere appreciation to 
his colleagues, committee members, and others whose support has been 
invaluable. In particular, special thanks are due my advisor. Dr. 
Eric Grimsrud, for his patience, persistence, technical assistance, 
and guidance during this study.
Financial assistance from the National Science Foundation is 
greatly appreciated.
Special thanks are due Dr. Abbott and Dr. P. Donahoe for 
many hours of thought provoking discussions.
In addition I would like to thank Rob Ekeland and Kari 
Dawson for their friendship and use of their computer.
TABLE OF CONTENTS
Page
List of Tables......................... .................. .. . %
List of Figures......................... xii
Abstract........................... xviii
Introduction..................................................   I
Analytical Techniques............................... .. . 8
Electron Capture Detector History.......................  25
APIMS and Plasma Chromatography. . . . . . . . .  ........  30
Oiygen as a Dopant . .............  32
Model for Instrument Response. .....................   38
Other Dopants..................    47
Research Objectives ..........................................  49
Experimental..................................................  50
Instrumentation...........................   50
Chrcmatographic Parameters .............................  54
Oxygen Addition. . . . ..................................  56
Measurement Procedures .................................. 58
Sample Preparation ......................................  58
Parallel Detectors ............... . . . . . . . . . . .  62
Series Detectors................................... .. . 63
Mass Spectrometric Parameters..............      65
Atmospheric Pressure Ionization Mass 
Spectrometry............................................  68
vii
viii
Results and Discussion. ...........................   71
Polychl orobiphenyIs...................................... . 71
Isomer. Differentiation..................   71
Chloroanthracenes and Chlorophenanthrene ..........  . . .  74
Isomer Differentiation. .................    74
Mixture Analyses. . ..................................  79
1
Atmospheric Pressure Ionization Mass 
Spectrometry..............       92
Polycyclic Aronratic Amines..............   98
Normal ECD Response .....................................102
Temperature Dependence...............................   102
Isoner Differentiation.........................  104
Mixture Analysis........................................ 112
Atmospheric Pressure Ionization Mass 
Spectrometry................  116
Mass Spectral Analyses.................................. 124
Polycyclic Aronatic Hydroxides..........    152
Isomer Differentiation. . . . . . . . . . . . . . . .  153
Methylanthracenes and Methylphenanthrenes................... 156
Atmospheric Pressure Ionization Mass
Spectrometry............................................ 161
Dioxin....................... .............................• 164
Simultaneous Electron Capture and Oxygen-Sensitized 
Electron Capture Detection.................................. 166
TABLE OF CONTENTS (continued)
Page
ix
Page
Parallel E C D s ..............   167
Series Detectors. .............................   176
Tanden Cells.............................   185
Summary.............................   189
Future Applications.....................................   194
Literature Cited.........................................   201
TABLE OF CONTENTS (continued)
XLIST OF TABLES
Table
1 Dioxins Lethality Ccmpared to Other Poisons...............
2 Response Enhancements (RE) for the Chlorinated Biphenyls.
3 Relative Abundances for the Chloro Isaners of
Anthracene and Phenanthrene................................
4 APIMS Results for the Chloroanthracenes and 9-chloro-
anthracene from Single Ion Monitoring.....................
5 Oxygen Induced EC Response Enchancements for the Tri-
fluoroacetic Anhydride (TFAA) and Perfluoropropionic 
Anhydride (PFPA) Derivatives of Various Polycyclic 
Aromatic Amines. ................... ......................
6 Relative Parent Negative Ion Intensities by APIMS for the 
TFAA Derivatives of Several Polycyclic Arcmatic Amines . .
I  Oxygen-Induced ECD and APIMS Response Enchancement for
the Trifluoroacetic Anhydride Derivatives of Amino- 
anthracenes and Aminophenanthrenes .......................
8 Ion Intensities from the Electron Impact Mass Spectra
of the TFAA derivatives of the Aminoanthracene and 
Phenanthrene ..............................................
9 Ion Intensities from the PCI Isobutane Mass Spectra 
of TFAA Derivatives of the Isomers of Aminoanthracene
and Phenanthrene ..........................................
10 Ion Intensities from the NCI Isobutane Mass Spectra of
the TFAA Derivatives of the Isoners of Aminoanthracene 
and Phenanthrene................... .......................
II Ion Intensities from the PCI (Methane) Mass Spectra of
the Derivatives of the Iscmers of Aminoanthracene and 
Phenanthrene ..............................................
12 Ion Intensities from the NCI Methane Mass Spectra of the 
TFAA Derivatives of the Isomers of Aminoanthracene and 
Phenanthrene................... .................... .. . .
Page'
9
. 73
77 
94
107
117
119
128
131
134
134
139
LIST OF TABLES (continued)
Page
13 Oxygen induced Response Enhancements for Methoxy
Derivatives for Several Polycyclic Aronatic Hydroxides. . . 155
14 Response Enhancenents for Methylanthracenes and
Methylphenanthrenes ........................................  160
xii
1 Capillary column gas chromatogram of coal tar........ 13
2 Electron impact mass spectra of triphenylene, 
chrysene, benz(a)anthracene, napthacene, and
benzo (c) phenanthrene .  ............................. 16
3 Methane chemical ionization mass spectra of anthracene
and phenanthr e n e ...........................  21
4 Argon-methane chemical ionization mass spectra of
anthracene and phenanthrene .........................  22
5 Typical chromatograms in the studies of oxygen's
effects on ECD responses.......... ................... 36
6 Effect of oxygen on the baseline noise at several
detector tenperatures.......................  45
7 Schematic illustration of the displaced coaxial
cylindrical ECD used in this study ...................  51
8 Block diagram of the electronic components of a
constant current B C D .................................... 53
9 Diagram indication positions for oxygen addition . . .  57
10 Schematic of the Varian capillary column effluent
splitter................................  64
11 Tandem cell configuration.............................   66
12 TWo sources used in the APIMS studies.................... 69
13 The response of BCD to 3-chlorobiphenyl and
3,5-dichlorobiphenyl .................................. 72
14 The electron impact mass spectra for l-,2-, and
9-chloroanthracene .................................... 75
15 The electron impact mass spectrum for 9-chloro-
phenanthrene . . . . . .  ...................  . . . . .  76
LIST OF FIGURES
Figure Page
xiii
16 ECD responses to 1-, 2-, and 9-chlor©anthracene, 
and 9-chlorophenanthrene (bottom) and chromatograms
with 0.30% oxygen present in the detector . . . . . . .  78
17 Oxygen-induced response enhancements for the
chior©anthracenes as a function of concentration. . . .  80
18 Normal ECD responses for I- and 9-chlorcanthracene 
and a mixture of the two (bottom chromatograms) and 
with the addition of 0.30% oxygen in the detector (top)
at a temperature of 300°C.......... .. 82
19 Response enhancements o f . two component mixtures as
a function of the molar fraction of one of the 
components...................................... 83
20 ECD responses to pure isomers of chior©anthracene 
and a mixture of isomers without oxygen and with
0.30% oxygen in the detector at 3OO0C . . . . . . . . .  85
21 ECD responses to pure isomers of chloroanthracene 
and a mixture of all three isomers without oxygen
and with 0.30% oxygen in the detector at 350°C. . . . .  86
22 Capillary column gas chromatograms of the three-
component mixture shown in Figure 20............... . . 89
23 Capillary column gas chromatograms of the three- 
component mixture in Figure 20 with on-column
injector. . . ................... .. 91
24 Negative API mass spectral scans taken during elution 
of 9-chloroanthracene without and with added oxygen at
a source temperature of 250°C . . .  ............. .. 93
25 Flame ionization detector response to underivatized
amines of 9-, 1-, and 2-aminoanthracene............... 99
26 Flame ionization detector response to TFAA derivatives
of 9-, 1-, and 2-aminanthracene . . ................... 100
27 Normal ECD response to 9-aminophenanthrene (TFAA)
and underivatized 9-aminophenanthrene ............. , . 101
LIST OF FIGURES (continued)
Figure page
;{
xiv
LIST OF FIGURES (continued)
Fiaure Paae
28 Normal ECD responses as a function of detector 
temperature for 3-aminophenanthrene (PFPA) and 
4-aminophenanthrene (TFAA)......................... . 103
29 Capillary column gas chromatograms of the TFAA deri­
vative of 2-aminofluorene . ........................... 105
30 Capillary column gas chromatograms of the PFPA deriva­
tives of 9-aminophenanthrene, 2-aminoanthracene, and 
1-aminoanthracene ...................................... 106
31 ' Structure and numbering scheme for biphenyl, naphtha­
lene, anthracene, phenanthrene, fluorene, benz(a)- 
anthracene, chrysene, pyrene, and benzo(c)phenan­
threne. . ............................'.................. 109
32 Capillary column gas chromatograms of a three 
component mixture of 9-aminoanthracene (TFAA), 1- 
aminoanthracene (TFAA), and 2-aminoanthracene (TFAA). . 114
33 Capillary column gas chromatograms of the two-component 
equimolar mixture of the TFAA derivatives of 1-amino­
anthracene and 9-aminophenanth rene. ...................
■ I
us
34 Negative ion API mass spectral scans taken during . 
elution of TFAA derivative of 9-aminoanthracene 
without and with .20% added oxygen at a source 
temperature of 250°C....................................
-
I': ■,
120
35 Single-ion APIMS monitor of the negative ion intensity 
at m/e 289 for the 1-aminophenanthrene (TFAA) derivative 
without and with added oxygen (.20%) at a source temp­
erature of 250°C........ ............................... 122
36 Negative ion API mass spectral scans taken during 
elution of derivative of 3-aminophenanthrene (PFPA) 
with and without .20% added oxygen..................... 125
37 Electron impact mass spectrum of 1-aminoanthracene 
(TFAA). .'.............................................. ' 126
38 Electron impact mass spectrum of 4-aminophenanthrene 
.(TFAA).................................................. 127
n'
XV
LIST OF FIGURES (continued)
■Figure Page
39 Isobutane positive chemical ionization mass spectrum
of 1-aminoanthracene (TFAA)..............  129
40 Isobutane positive chemical ionization mass spectrum
of 4-aminophenanthrene (TFAA)............................130
41 Isobutane negative chemical ionization mass spectrum
of 1-aminoanthracene (TFAA)....................... .. . 132
42 Isobutane negative chemical ionization mass spectrum
of 4-aminophenantarene (TFAA) . .......................... 133
43 Methane positive chemical ionization mass spectrum of
1-aminoanthracene (TFAA)............    135
44 Methane positive chemical ionization mass spectrum of
4-aminophenanthrene (TFAA)................................ 136
45 Methane negative chemical ionization mass spectrum of
1-aminoanthracene (TFAA).................................. 137
46 Methane negative chemical ionization mass spectrum of
4-aminophenantarene (TFAA)................................ 138
47 B/E scan on m/e 290 of 4-aminophenanthrene
(TFAA). ................. 140
48 B/E scan on m/e 272 of 4-aminophenanthrene
(TFAA)....................... ......................... . 141
49 B/E scan on m/e 272 of 4-aminophenanthrene
(TFAA) with He as the CAD g a s ................  143
50 Single-ion traces of m/e 289 and 112 for 9-aminoan-
thracene (TFAA) with isobutane negative chemical ioni­
zation mass spectrometry..............    147
51 Electron impact mass spectrum of 3-aminophenanthrene
(PFPA)..................   148
52 Electron impact mass spectrum of 4-aminophenanthrene
(PFPA). . . . . . .  .................................... 149
xvi
53 Isobutane positive chemical ionization mass spectrum
of 4-aminophenanthrene (PFPA........................... 150
54 Methane positive chemical ionization mass spectrum
of 4-aminophenanthrene (PFPA)......................... 151
55 Capillary column gas chromatograms of 4-hydroxy-
phenanthrene (methoxy derivative) .....................  154
56 Electron impact mass spectrum of 2-methylanthracene . . 157
57 Electron impact mass spectrum of 2-methylphenanthrene . 158
58 Electron impact mass spectrum of 1-methylphenanthrene . 159
59 Response enhancement for 9,10-dime thy!anthracene using
series detectors........................................ 162
60 Capillary column gas chromatograms of 2,3,7-tri­
chi orodibenzor-p-dioxin. . . ............................165
61 Dual ECD responses with dual columns of sample contain­
ing 2 ng of 2-chloroanthracene, repeated three times 
with both detectors operated in the normal, oxygen-free 
mode.................................................... 169
62 Dual ECD and oxygen-sensitized detection of sample 
containing 12 ng 2-chloroanthracene, repeated ten
times.................................................. 170
63 Dual ECD and oxygen-sensitized ECD detection as
a function of concentration ...........................  172
64 Dual ECD and oxygen-sensitized EC detection of samples 
containing 5 ng of 1-chloroanthracene, 12 ng of 2-
chloroanthracene and 6 ng 9-chloroanthracene...........173
65 Dual ECD responses with a capillary column effluent
splitter of sample containing 8 ng of 9-chloroan- 
thracene................................................ 175
66 Schematic diagram of the fused silica capillary
column effluent splitter............................... . 177
LIST OF FIGURES (continued)
Figure Page
xvii
67 Dual ECD responses of ECDs in series detection of
1-, 2-, 9-chloroanthracene and 9-chlorophenanthrene . . 178
68 Dual ECD responses and oxygen-sensitized ECDs in 
series detection of 1-, 2 -, and 9-chloroanthracene
and 9-chlorophenanthrehe................................179
69 Dual ECD responses and O^-sensitized ECD in series 
of sample containing 9-chloroanthracene repeated ten
times.................................................. 181
70 Dual ECD responses of series detectors of separate 
samples containing hexachlorobenzene, 9-nitro-
benzene, and 2 , 1 -dini tr of I uorene....................... 182
71. Dual ECD and oxygen-sensitized ECD responses of
series detectors of a mixture of 9-chlorophenanthrene 
and 2-chloroanthracene..................................183
72 Series detectors in which the bottom chromatograms 
are normal ECD responses and the top chromatograms
are with 0.30%. oxygen in the second detector...........184
73 Series detectors in which the chromatograms on the
left are with the normal EXD response and the ones 
on the right are with 0.30% oxygen present in both 
detectors.............................................. 186
74 Dual ECD responses and oxygen-sensitized response of
50 ng sample of 2-chloroanthracene............... 187
75 Chemical class separation scheme for synthetic fuel
products............... ................................ 195
76 ECD capillary gas chromatogram of TFAA derivatized
APH fraction in SRC II H D ............................. 196
77 Capillary gas chromatograms of 7 ng sample of anthra­
cene with ethyl chloride doping....................... 198
78 Single-ion APIMS of the negative ion intensity at m/e
35 with a sample of anthracene with and without added 
ethyl chloride at a source temperature of 250°C . . . .  199
LIST OF FIGURES (continued)
Figure Page
xviii
ABSTRACT
The addition of oxygen to the nitrogen carrier gas of a 
constant current electron capture detector (BCD) is shown to 
provide increased sensitivity and isomer distinction for 
environmentally important compounds such as polychlorinated 
biphenyls, polycyclic aromatic amines and hydroxides with 
appropriate EC-enhancing tags, chloroanthracenes and 
chiorophenanthrenes, methy!anthracenes and methylphenanthrenes, and 
2,3,7-trichlorodibenzo-p-dioxin. In most cases, the isomers of 
these particular compounds can be distinguished from the other 
isomers based solely on the measured response enhancements.
The oxygen doped ECD has been applied to compound 
identification where only partial resolution with capillary column 
gas chromatography is obtained. In addition, in instances where 
several compounds coelute and have the exact same retention times, 
the mole fraction of each component in the unresolved peak can be 
determined.
Atmospheric pressure ionization mass spectrometry (APIMS) 
gives an indication of the ions formed with and without the 
presence of oxygen in the source. For the chloroanthracenes, 
methylanthracenes, and methylphenanthrenes, actual oxygen 
incorporation is observed when oxygen is present. In contrast, for 
the polycyclic aromatic amines derivatized with trifluoroacetic 
anhydride (TFAA), the reaction with oxygen to produce an induced 
response involves a charge transfer between the oxygen anion and 
the analyte molecule.
Several TFAA derivatives of the aminoanthracenes and 
aminophenanthrenes were examined using electron impact mass 
spectrometry and chemical ionization mass spectrometry. Negative 
chemical ionization mass spectrometry with methane and. isobutane as 
the reagent gases was evaluated as a method of isomer 
differentiation of these compounds and shows considerable promise.
In order to improve the precision involved in the measurement 
of response enhancements, two methods with parallel and series 
arrangement of the ECDs were utilized. The parallel arrangement 
with dual columns represents a significant improvement in the 
reproducibility of response enhancement measurements. The series 
detector arrangement seems to be confusing in view of possible 
additional reactions in the detectors.
The addition of ethyl chloride to an electron capture detector 
increased the response of anthracene and similar molecules with a 
low normal ECD response and those that react with the gaseous 
electron through a resonance type of reaction mechanism. The addi­
tion of ethyl chloride to the detector does not significantly in­
crease the baseline frequency, in contrast to the addition of oxygen. 
The negative ions formed in this reaction have been identified.
IINTRODUCTION
Polycyclic aromatic hydrocarbons (PAHs) and substituted 
polycyclic aromatic hydrocarbons (SPAHs) are present as 
environmental pollutants formed from both natural and anthropogenic 
sources. Natural sources include forest and prairie fires and in 
situ synthesis from degraded biological material. Anthropogenic 
sources of PAHs and SPAHs include the burning of coal refuse banks, 
residential fireplaces, and commercial incinerators. .
PAH mixtures can be extremely complex, since they can contain 
numerous isomeric compounds. A  case in point is a simple sample 
containing the isomeric compounds with four-fused rings. These 
would include benz[a]anthracene (A), triphenylene (B), 
benzo[c]phenanthrene (C), chrysene (D), and naphthacene (E). The 
structure and numbering scheme for these compounds are shown below.
D E
2It follows that for example, a monosubstituted (amine, chloride, 
methyl, etc.) naphthacene has three isomers. Accounting for all of 
the isomeric possibilities for the five compounds for 
monosubstitution, the total number is 29. For a disubstituted 
naphthacene, the number of isomers is 20, and a disubstituted 
benz [a] anthracene has 66 isomers. The total number of 
disubstituted isomers for the five compounds then probably exceeds 
100. As the number of substituted positions increases, the number 
of possible isomer combinations becomes enormous. This can pose 
quite a challenging problem for the analytical chemist to identify 
possible isomers in a sample. Even if the sample contained a 
single component, one isomer of the sixty-six possible 
disubstituted isomers of benz [a!anthracene, the problem of 
identification is still complex. Mass spectrometry, which is 
normally used for this type of analysis, alone is frequently unable 
to provide positive identification because of the large number of 
possible isomers for a given molecular mass. In addition, for some 
substituted PAHs, a strong dependence of carcinogenicity and 
mutagencity on the position of substitution has been found. In 
view of these observations,' identification of specific isomers is 
extremely important. Therefore, a technique which provides 
assistance in the identification of substituted PAHs will be a 
welcomed addition to the arsenal of analysis techniques presently 
available to the analytical chemist.
In the next section the background and history of some of the 
more environmentally important substituted polycyclic aromatic
3hydrocarobons which include the poly chiorobiphenyls, polycyclic 
amines and hydroxides and dioxins will be discussed followed by a 
short review of the typical analytical techniques for the 
identification of PAHs and SPAHs. The history and development of 
the electron capture detector (EOD) and the use of oxygen as a 
dopant will be reviewed. A discussion of the model for 
instrumental response with oxygen doping and the application of 
> other dopants will follow.
Polychlorobiphenyls (KBs)
Chlorinated biphenyls have been available as commercial 
products with a variety of applications for over fifty years. They 
were marketed under the trade name of Aroclors by Monsanto in North 
America and in other parts of the world under various names, such 
as Clophen (G.F.R., Phenoclor (France), and Kennechlor (Japan). 
These PCB formulations, since their development in 1929, have been 
utilized chiefly as electrical coolant and insulating fluids in 
transformers and capacitors, but also in heat exchange and 
hydraulic systems, as plasticizers in paints, adhesives, caulking 
compounds and dye-carriers in carbonless copy paper. Their escape 
from these systems or their disposal in wastes has led to worldwide 
environmental contamination with PCBs. The presence of PCBs in the 
environment became apparent only after the ECD had been extensively 
applied to the monitoring of organochlorine pesticide residues, 
during the early 1960s. Chromatograms of wildlife sample extracts 
were found to exhibit many electron-capturing responses which could
4not be attributed to organochlorine pesticide residues. These were
/
first identified as PCBs by Jensen (I) who confirmed their presence 
with mass spectrometry.
Because of their resistance to breakdown and their other 
physical properties which are similar to those of the 
organochlorine pesticides, particularly DET and its metabolites, 
the PCBs tend to bioaccumulate towards the top of the food web, 
namely in the fatty tissues of predatory fish,. mammals and birds.
PCBs have been found in all parts of the environment. Their 
presence and distribution have been reviewed by Risebrouah et al. (2), 
Peakall (3), Jensen (4) and Nisbet and Sarofin (5). Pal et al. (6) 
have reviewed the fate of PCBs in soil-plant systems as well as in 
other environmental substrates. Fishbein (7) has reviewed.some of 
the chromatographic and biological properties of the PCBs.
Polycyclic Aromatic Amines and Polycyclic Aromatic Hydroxides
The occurrence of polycyclic aromatic amines (PAA) in coal- 
derived liquids and residues was reported as early as 1958 by Karr 
St Si*. (8). Other workers have implied or demonstrated the presence 
of PAAs having one or two aromatic rings in synfuel products 
(9,10). More recently, however, even more highly mutagenic 
polycyclic aromatic amines (PAAs) have been detected in some 
synfuels materials (11-14) and have been found to be the 
determinant mutagens in some coal-derived materials (13,15-17).
These species include recognized or suspected carcinogens such as 
1-aminonaphthalene, 2-aminonaphthalene, 4-aminobiphenyl, I-
5anthracene, and 2-aminoanthracene (18-20). Lee, Later, and Wilson 
have found that mutagenic activity within an isomer group is 
structure dependent (21).
. . Many hydroxy type compounds have been shown to be present in
synthetic fuel products and several methods have been developed for 
their isolation (22-25). Compounds such as hydroxyfIuorenes, 
hydroxynaphthalenes, and hydroxybiphenyls have been identified in 
these synfuels (23). These compounds like the polycyclic aromatic 
amines may possess mutagenic and carcinogenic characteristics (25).
Chlorinated Dioxin
Chlorinated dioxins were first synthesized as early as 1872 by 
German scientists, but only recently has an enormous interest been 
focused on the polychlorodibenzo-p-dioxins (EDDs) and in particular 
one of the 22 isomers of the tetrachlorodibenzo-p-dioxin (TCDD) 
group, 2,3,7,8-TCDD, as a result of an explosion of a safety valve 
in an industrial plant producing trichlorophenol at Sevesco in 
northern Italy in July 1976. As the explosion was the result of 
overheating of the plant reactor, 2,3,7,8-TCDD, the most toxic (26) 
compound of the series, was produced in much larger amounts than in 
the normal process, where its concentration is restricted to low 
parts per million levels (27).
Tines Beach, Missouri, has a most unenviable reputation as the 
town too poisoned to live in. Just a few miles west of St. Louis, 
its fate has been sealed as the result of seme poor waste disposal 
practices, insufficient environmental laws in the early 1970s, and
6political pressures for action. Because of the contamination and 
the federal government's decision to buy the town. Times Beach is 
expected to virtually disappear.
The story traces back to the 1960s and begins with agent 
orange. Hoffman-Taft, one of the original defendants in the agent 
orange trial, made 2,4,5-trichlorophenoxyacetic acid (2,4,5-T) for 
the Department of Defense for a while, but ceased production in 
1969, about the time ecological concerns led the military to halt 
spraying. In November of 1969, the plant in Verona, Missouri, was 
leased to North Eastern Pharmaceutical and Chemical company, and 
then later sold to Syntex Agribusiness, which let the 
pharmaceutical company stay to produce hexachlorophene.
According to EPA1s records, wastes from the plant were being 
disposed of properly by shipping them to a waste facility owned by 
what is now Rollins Environmental Service near Baton Rouge,
Louisiana. But in early 1971, allegedly to save money. North 
Eastern contracted with a firm called Independent Petrochemical to 
haul away its sludge bottoms. Independent, in turn, subcontracted 
the job to Russell Bliss, a waste oil hauler in Missouri. The 
records show that Bliss hauled away 18,500 gal of waste bottoms 
containing dioxin from the Verona plant, which he apparently stored 
in waste oil tanks near Fronenac, Missouri, between February and October 
1971. But Bliss used some of this contaminated waste oil to spray 
horse arenas in May of 1971. Three stables apparently were 
sprayed, and the consequences were severer Over the next few days 
and weeks, hundreds of animals got sick and died, including at
7least 65 horses. CSie six-year old child, the daughter of one of 
the stable owners, developed an inflamed and bleeding bladder after 
playing in the soil of the arena, and three other children and one 
adult complained of skin lesions after exposure to the stables.
All of the symptoms disappeared after exposure was halted and have 
not recurred.
State of Missouri investigators, reasoning that something must 
have been in the oil that was sprayed, sent samples for analysis. 
The Center for Disease Control, with few clues to go on, took until 
1974 to identify dioxin as the toxic compound in the oil. They
' I
eventually determined that the oil was contaminated at about 33 
ppm, a level far higher than ary that occurred in Vietnam from 
agent orange. Crystals of triclorophenol found during the soil 
analysis led the investigators to the Hoffman-Taft plant in Verona, 
and the thinking was that 2,4,5-T production was the culprit. But 
it was then discovered that the hexachlorophene wastes made by 
North Eastern had been disposed of improperly.
Spraying began in Vietnam in January 1962, using a variety of 
herbicide mixtures. Small amounts were used at first, but the 
amounts jumped at the end of 1965 and heavy use continued until 
1969. Mounting concerns about damage to Vietnam's ecology led to a 
tapering off and finally a halt to the spraying of 2,4,5-T in 1970. 
Of the seven or so herbicide formulations used, the most 
significant was called agent orange, an oily liquid that was a 50-
850 mixture of the n-butyl esters of 2,4,5-T and 2,4- 
dichloropheno^acetic acid (2,4-D).
The major source of PCDDs is the massive production and use of 
pentachlorophenol (PCP) and other chlorophenols that are widely used 
as herbicides, insecticides, and wood preservatives. These 
compounds contain a variety of chlorine-containing contaminants and 
by-products that have different structures tut a behavior that is 
similar to that of the PCDDs, and also their molecular weights lie 
in the same range. In addition, chlorinated dioxins have been 
found on fly ash particulate matter generated by municipal 
incinerators (28) and chimney particulate matter from wood combustion 
(29).
The overall picture indicates dioxins are extremely toxic and 
carcinogenic based on animal studies, but the picture is much less 
clear about human health effects. Dioxin's lethality is compared to 
other poisons in Table I. It is clear that an analytical technique 
that would distinguish isomers of the dioxins, as well as the 
previously described compounds, would be of tremendous importance.
Analytical Techniques for PAH and SPAH Identification
As previously indicated, PAH mixtures can be extremely 
complex, since they contain numerous isomeric compounds. The 
success of the chemical analysis, whether it be quantitative or 
qualitative (of even a single component), hinges on the resolving 
power and sensitivity of the analytical method employed.
9Table I. Dioxin's lethality compared to other poisons
Minimum lethal 
dose (moles per
Substance Animal kg body weight)
Botulinum toxic a Mouse 3.3 X IQ-l?
Tetanus toxin Mouse ■1.0 X 10-15
Diphtheria toxin Mouse 4.2 X 10-12
2,3,7,8-TCDD Guinea Pig 3.1 X H %
Bufotoxin Cat 5.2 X 10-7
Curare Mouse 7.2 X
r—I
Strychnine Mouse 1.5 X
Or—<
Muscarin Cat 5.2 X 10-6
Diisopropylfluoro-
phosphate
Mouse 1.6 X 10-5
Sodium cyanide Mouse
OCN X
H
Source: EPA
10
In the following sections, the most important analytical 
techniques for the analysis of PAHs and substituted PAHs will be 
discussed. The enphasis will be placed on chromatographic methods, 
mass spectrometry, and spectroscopic methods.
Chromatographic Methods
High Performance Liquid Chromatography
A major advantage of HPLC for the determination of PAHs is the 
general applicability to even completely involatile compounds of 
interest. Ultraviolet (UV) absorption and fluorescence detectors 
are generally used in series; the UV detector is universal for 
PAHs, whereas, the fluorescence detector provides high specificity.
Several workers (30-32) have described the use of variable 
wavelength UV detection to achieve some degree of selectivity.
Thons and Zander (33) used complete UV absorption spectra of PAHs 
for the identification in HELC elluents. Other researchers (30,34) 
have employed absorbance ratios at several wavelengths for 
qualitative analyses.
Selective fluorescence quenching of certain PAHs in the 
presence of nitromethane has been investigated as a selective 
detection system for HELC. Sawicki jst al. (35) and later Dreeskamp 
et al. (36) reported that in the presence of nitromethane, the 
fluorescence emission of six-membered ring PAHs were quenched to a 
much greater degree than those containing a fluoranthenic 
structure. Recently, Konansh et al. (37) investigated further the 
potential of this phenomenon for the selective detection of the
11
benzofluoranthene isomers in the presence of the benzopyrene 
isomers and perylene.
A recent development in spectroscopic detection of PAHs in 
chromatographic effluents is the use of multichannel rapid scanning 
spectrometers. These detectors permit the recording of 
fluorescence spectra " on-the-fly", thereby eliminating stop flow 
or valving to trap the chromatographic peak in the flew cell. 
Jadamec £t sJU (38) described the use of such a system for the 
characterization of petroleum fractions for the determination of 
the source of oil spills.
The potential of the combination of liquid chrcmatography-mass 
spectrometry (LC/MS) for the separation and" identification of 
organic compounds has generated considerable research interest in 
the past few years. At present, two LC/MS approaches are 
available, i.e., direct liquid injection and depositing the 
effluent onto a moving belt from which the mobile phase is 
evaporated prior to introduction into the mass spectrometer. Both 
of these approaches transfer only a small portion of the LC solute 
into the MS. Christensen sfc aL. (39) recently described a new 
LC/MS interface that combines several of the advantages of both the 
direct liquid injection and moving belt techniques. The LC 
effluent is concentrated by evaporation as it flows down an 
electrically heated wire. The concentrated effluent then flows 
through a small needle valve and is sprayed into the MS ion source.
Dark £t al. (40) using the moving belt method, demonstrated 
the LC fraction of coal liquid samples with structural
12
characterization by LC/MS. Using LC retention data on two 
different columns and the molecular weight data, they identified a 
number of aromatic compounds with molecular weights up to 250 in 
the coal liquid. Other applications of LC/MS involving PAH 
separation and identification have been reported (41,42). A review 
of the LC/MS techniques by McFadden (43) emphasized the current 
applications of this technique. Many other applications of LC/MS 
have been reported involving PAHs (43-45). For isomer distinction, 
this suffers from the same problems associated with the use of 
the mass spectrometer which will be discussed in more detail later. 
In the future, LC/MS probably will find greater application in the 
determination of higher molecular weight PAHs that are not 
particularly amenable to GC analysis.
Gas Chromatography
The extreme complexity of the PAH mixtures demands the 
greatest resolution possible in their, analysis, and in this 
respect, as chromatography with packed columns has fallen far short 
of the capabilities of glass capillary columns. Capillary column 
gas chromatography was refined by a number of workers and the 
present standard was reached in 1975 when Lee (46) showed that 
acid-leaching of Lewis acids from the glass from which capillaries 
were made greatly improved the deactivation and efficiency of 
columns. Figure I shows a chromatogram of coal tar PAH obtained 
with an acid-leached column, subsequently coated with SE-52. The
13
-_ J-JL L . Lll XaIi ^ I I d J  Il-IJ' lli/K JiIul -^ lL  ^L ——-
I____ !—
40 75 100 150 175TEMPERATURE (0C)
Figure I. Capillary Column Gas Chromatogram of Coal Tar. Peak 
(I) Naphthalene, (2) Phenanthrene, (3) Anthracene, (4) 
Fluoranthene, (5) Pyrene, (6) Benz(a)anthracene, (7) 
Chrysene, (8) Benzo(a)pyrene, (9) Benzo(e)pyrene, (10) 
Perylene. (Reference 46).
14
performance of this column, as measured by the resolution between 
the isomer pairs phenanthrene-anthracene, benz[a]anthracene- 
chrysene, and benzo[elpyrene-benzo[alpyrene, represents the best 
resolution currently available. Fused silica columns are now used 
almost universally. Liquid crystal phases show pronounced 
selectivity for PAHs in column chromatography and allow the 
resolution of a variety of iscmer pairs such as anthracene and 
phenanthrene.
The most widely used gas chromatographic detector for PAHs is 
the flame ionization detector (FID). This is a result of its" 
universally accepted characteristics of excellent response 
linearity, sensitivity, and reliability. Response increases with 
molecular weight and response factors are similar for most 
structural isaners.
As early as 1965, Cantuti et al. (47,48) showed that the 
response of the electron capture detector (BCD) for PAHs was 
dependent on the structure of the compound, and that the detector 
could selective for PAHs in hydrocarbon mixtures. Bjorseth and 
Ekland (49) measured the ECD/FID response ratios for 29 PAHs, and 
found that many isomers could be differentiated by measurements of 
these ratios.
Mass Spectrometry
In the last decade, mass spectrometry (MS) has gained wide 
acceptance for the analysis of PAHs. New rapid scan techniques 
coupled with high resolution chromatography and improved data
X15
systems have greatly surpassed ary other method or combination of 
methods used for EAH analysis.
Electron Impact
The electron impact mass spectra of PAHs are generally quite 
simple, mainly consisting of an intense molecular ion and lower 
intensity ions due to the loss of m e  to four hydrogen atoms.
Doubly charged molecular ions are quite common and are usually near 
twenty percent of the abundance of the molecular ion. In most 
cases, differentiation of PAH isomers by electron impact mass 
spectra alone cannot be achieved. Even in the cases of isomers 
with very different structures, such as fluoranthene and pyrene, 
the mass spectra are most often indistinguishable. Figure 2 
compares the EI mass spectra of the four-ring isomers, 
triphenylene, chrysene, benz[a3anthracene, naphthacene, and 
benzo tc 3 phenanthrene. All the mass spectra are essentially 
identical except for the benzo[c3phenanthrene. In this case, 
steric interaction between the protons on and I and 12 carbons 
facilitates the loss of these two protons with the subsequent 
formation of the benzo [ghil fluoranthene ion. A large number of 
electron impact mass spectra for PAHs have been compiled and may be 
found in several reference books (50,51).
Mass spectrometry coupled with gas chromatography has been used 
to confirm the presence of PCB residues and characterizing the 
molecular composition of PCB formulations. Coupling of the gas 
chromatograph and mass spectrometer is normally carried out in such
50
100
0
100
50
0
100
50
0
100 - 
50 -
100 
50 
0
20 40 60 80 100 120 140 160 180 200220 240 260 280 300 320 340 360 
m/e
Figure 2. Electron Impact Mass Spectra of Triphenylenef 
Benz (a)anthracenef Napthacenef and Benzo(c)- 
Phenanthrene. (Reference 64).
Chrysene,
17
a way that a chromatogram-type readout analogous to that produced 
by conventional GC detectors is produced. The usual source of the 
GC signal is a total-ion-current monitor located between the ion 
source and mass—analyzing magnet (52). An extensive review of MS 
applications to pesticide residue analysis has been written by 
Biros (53). Bonnelli (54) reports that cleanup of K B  samples and 
separation from pesticides are not as important with GC/MS 
techniques as with conventional gas chromatography. Detailed 
investigations of primary ion spectra and fragmentation patterns of 
synthesized individual K B  isomers have been made (55,56). It was 
concluded (57) that the primary ion spectra of different isomers 
are virtually indistinguishable with the result that the use of 
mass spectrometry for structural studies of PCBs is limited.
Several researchers have recently reported procedures for 
isolating an amine-enriched fraction from synthetic crudes (58,59). 
This procedure, while effective, is not suitable for rapid 
analytical scale determinations. Derivatization of amines with 
fluorinated anhydrides, which is effective for all sterically 
unhindered amines (60,61) is a very useful step in any amine 
analysis. For example, trifluoroacetylation introduces a foreign 
element which is appropriate for selective spectroscopic 
identification procedures (62) and renders the compound more EC 
responsive. In addition, derivatized amines have characteristic 
fragmentation patterns. The parent ion gives the molecular weight 
of the derivatized compound. Ttoo other major ions are present: (I) 
the (M - 97)+ ion corresponds to the loss of the trifluoroacetic
(anhydride group and (2) the other prominent ion in the mass 
spectrum results from the loss of HCN from the (M - 97)+ fragment. 
Similar fragmentation patterns have been observed with the 
perfluoropropionic anhydride derivatives with the major fragment 
ion corresponding to the loss of the PFP group, (M - 147)*. Later, 
Lee, and Wilson (59) have identified several PAAs in a solvent 
refined coal liquid with the use of EC-enhancing derivatization 
procedures with subsequent GC/MS and GC/ECD analyses. Aromatic 
amines have also been determined in cigarette smoke using column- 
chromatographic isolation procedures followed by derivatization 
with pentafluoropropionic anhydride and subsequent GC/MS analyses 
(63).
The analysis of dioxin is usually done with gas chromatography 
and mass spectrometry. But there's considerably more to it than 
that. Normally, dioxins are present, if at all, at levels ranging 
from parts per million down to the vanishing point. They coexist 
with many other compounds, and many of these are present in much 
larger amounts and capable of interfering with the analysis. The . 
first step is to transfer the dioxins (and other chlorinated 
organics) frcm the sample matrix to an organic liquid, by a series 
of extractions. An isotopically labeled internal standard is added 
to help determine how much sample is lost in later steps, and to 
assist in quantitation. The organic extract is cleaned up with 
aqueous base and acid solutions and distilled water. Then the 
organic extract undergoes a sequence of preliminary liquid 
chromatographic separations, using a variety of columns and
. 18
19
eluents. All the fractions are recombined and concentrated for 
GC/MS analysis.
Usually, the sample first goes through a GC/low-resolution MS 
system for preliminary screening. This can show that TCDDs,. for 
example, are present, but it isn't sensitive enough to distinguish 
among the various isomers. If TCDD or other dioxins of interest 
are revealed by this preliminary analysis, the sample then goes to 
a second GC/high-resolution MS analysis. The present and amount of 
specific isomers like 2,3,7,8-TCDD then can be determined from the 
ratios of certain key mass fragments. With such techniques, and 
depending on the nature of the sample, it's possible to detect and 
quantify dioxins down to low parts-per-trillion levels with 
reasonable confidence. In the case of a simple sample like water, 
one can go even lower, down to the parts-per-quadrillion level, by 
taking a very large sample and concentrating the dioxins into a 
much smaller volume by solvent extraction. In the case of more 
complex samples, like soils, this approach is probably beyond the 
capabilities of today's analytical labs. Other techniques such as 
MS/MS and LC/MS have been utilized for dioxin analysis.
The analytical problems involved in searching for, identifying 
and determining PCDDs in environmental, biological and industrial 
matrices cannot be solved using an ECD alone because of the 
presence of interfering compounds. However, if highly selective 
and reliable purification and clean-up of the sample is possible, 
an ECD can be of immense value as a auxiliary detector in routine 
work.
20
Once the sample has been purified, the use of a very good 
capillary column with an ECD can be competitive with, and is perhaps 
better than, an inefficient packed column followed fcy a low- 
resolution mass spectrometer.
Chemical Ionization (Cl)
Conventional chemical ionization (Cl) mass spectrometry, using 
methane or isobutane as reagent gases, produces mass spectra of 
PAHs that also appear quite simple. The most abundant ion is the 
(M+l)+ or protonated molecular ion, and various adduct ions.
Recent studies (64,65) indicate that the mass spectra of different 
isaners appear in most cases to be practically identical. A 
typical example is shown in Figure 3 which illustrates the methane 
Cl mass spectra of phenanthrene and anthracene.
The use of a mixed charge exchange-chemical ionization reagent 
gas for mass spectrometry of PAHs which differentiates between 
isomers has recently been reported (66,67). The relative rates of 
the charge exchange and proton exchange reactions, and hence the 
ratio of the abundance of the (M+l)+ ion to be abundance of the M+ 
ion will vary according to the proton affinity and/or ionization 
potential of each PAH. Since the ionization potentials of PAH 
isomers are dependent on the specific structure of the molecule, 
the argon-methane reagent can produce quite different spectra for 
different isomers. This can be seen in Figure 4 , in which a 
comparison of the argon-methane Cl spectra for anthracene and 
phenanthrene is illustrated.
(M+1)
(M+29)
(M+41)
(M+1)
(M+29)
(M+41)
Figure 3. Methane chemical ionization mass spectra of anthracene 
and phenanthrene. (Reference 64).
22
100 -
80 -
(M+29)60 -
40 -
100 —
(M+29)
60 -
Figure 4. Argon-methane chemical ionization mass spectra of 
anthracene and phenanthrene. (Reference 64).
23
Hunt (68) has also shown that careful selection of reagent 
gases based on proton affinities can provide a means for 
differentiating PAH isomers. For example, the use of CgHgOD as a 
reagent gas can produce different spectra for phenanthrene and 
anthracene.
Shushan sfc. ala. (69) recently described a method in which both 
B and E are scanned so that (B/E) is held at the constant value 
required to transmit stable daughter ions of designated parent ions 
with preselected (m/e) ratio; in this way fragment ions formed in 
the first field-free region, from the preselected parent ions, are 
successively transmitted. The daughter-ion spectra of the four 
isomers, chrysene, triphenylene, benz[a!anthracene, and naphthacene 
showed considerable differences in the relative abundances of the 
Mf, (M-H)+, and (M-2H)+ ions.
Spectroscopic Methods 
UV Absorbance
The limitations of relatively low sensitivity and the poor 
specificity of spectra containing broad lines severely restrict the 
applications of ultra-violet (UV) absorption spectroscopy in the 
analysis of PAHs. Recent approaches to the luminescence analysis 
of PAHs have been aimed at increasing the sensitivity and 
selectivity of the methods through the use of new excitation 
sources and dispersive systems, more sophisticated detectors, and 
data systems (70,71).
24
Nuclear magnetic resonance
Conventional single-scan NMR may confirm the presence of a 
given class of compounds and may be applied to the identification 
of specific individual compounds. Keefer st al. (72) investigated 
the methods of identification of metiy!-substituted PAH from 
environmental samples by continuous wave %-NMR. The identities of 
4-methylpyrene and 2-methylpyrene were identified from their 
different chemical shifts. Pulse Fourier-transform %-NMR can be 
applied with much more facility to small samples. Bartle st al.
(73) were able to identify by FT %-NMR PAHs in mixtures 
separated from atmospheric dust and from the condensates of tobacco 
and marijuana smoke; pulse FT spectra at 9OMHz allowed 
identification of both parent PAH and their methyl derivatives.
- IO
C-NMR spectroscopy is a promising technique when sufficient 
sample is present. Sensitivity is relatively poor at the 
concentration levels normally found in environmental samples for 
NMR in the analysis of PAHs.
Infrared
The chief disadvantages in the conventional IR spectroscopy of 
PAHs in the environment are the absence of unique features in the 
spectra and the lack of proportionality between band strengths and 
concentration, and until the introduction of matrix isolation (MI) 
FT-IR there were few applications reported. Wehry st al. (74,75) 
have conducted detailed investigations of the applicability of MI 
FT-IR in the quantitative and qualitative analysis of PAHs.
25
Most of the techniques discussed have several limitations for 
PAH differentiation. These include sensitivity or sufficient 
sample size as in the spectroscopic methods and in general, 
inability to distinguish isomeric compounds as in the case of both 
El and Cl mass spectrometry. Even in the case of capillary column 
gas chromatography (see Figure I), the retention times for sane of 
the isomer sets of PAHs are almost identical. Based strictly on 
retention times alone, positive identification would be difficult. - 
As will be pointed out, the C^-doped BCD provides a very sensitive 
means for compound identification and provides assistance in the 
clarification of complex chromatograms where peaks are incompletely 
resolved.
Electron Capture Detector History and Development
The electron capture detector, along with the flame 
ionization detector, provide gas chromatography with its most 
commonly used detectors. For the ECD, a radioactive isotope is 
used as a source of beta electrons which bombard the carrier gas 
resulting in the formation of a plasma of positive ions, radicals, 
and thermal electrons. The application of a potential difference 
to the electron capture cell allows the collection of thermal 
electrons, which defines the standing current. Ihe electron 
capturing species introduced in the carrier gas stream reacts with 
the thermal electrons to produce negative ions . The decrease in 
detector current due to the renoval of thermal electrons by
26
reaction with electron capturing compounds forms the basis of the 
detector response.
Lovelock first reported the use of the electron capture 
detector (BCD) for the detection of solutes in the effluent of a 
gas chromatography column in 1958 (76). The first ECD consisted of 
a simple enclosed ionization chamber containing a radioactive 
source (tritium as the beta particle emitter) and a pair of D.C. 
electrodes to monitor the cell current at a constant low potential. 
The response to vapor concentration was nonlinear and the response 
factors varied greatly for different compounds and in an 
unpredictable manner. In spite of these disadvantages, the 
electron absorption method using a simple ion chamber was utilized 
by Goodwin st al. (77) and Watts and Klein (78) for environmental 
problems, particularly the analysis of trace pesticides.
Lovelock introduced the pulse sampling technique to reduce the 
sources of anomalous responses (79) which were shown to be related 
to space charge effects, high applied voltages, contact potentials, 
and unpredictable changes in electron energy. Brief pulses of 0.5 
to 1.0 microsecond width and 100 to 200 microsecond period of . 
sufficient magnitude to collect all of the electrons present in the 
plasma are used, rather than a steady D.C. potential. The current 
measured reflects the chemical events occurring within the gaseous 
plasma during the field-free period between pulses. If the 
difference in standing current which accompanies sample elution is 
taken as the ECD response, the linear dynamic range of the constant 
frequency pulsed ECD varies from 50 times the detection limit to
27
500 times the detection limit, depending upon the type of 
radioactive source used. This linear range corresponds to only 
about 10% reduction in the standing current. For certain 
substances the sensitivity may be three to four times higher in the 
constant frequency pulsed mode than it is in the D.C. mode.
The pulse sampling procedure had several advantages over the 
D.C. mode. Most, of the time no field is applied to the detector 
and the free electrons are in thermal equilibrium with the gas 
molecules. In addition, the electrons don't drift out of the 
plasma so that negative ion formation occurs in the region where 
the positive ions are also present and where recombination can most 
efficiently take place. The electron sampling period was so brief 
that no significant movement of the negative ions can occur; and as 
a result, the accuracy of the measurements are not affected by 
space charge effects or by the collection of negative ions at the 
anode. Also, a pulse amplitude of 30 volts is sufficient for the 
complete collection of the electrons. With the pulsed mode, the 
effects of contact potential effects are rarely if ever encountered 
if reasonably fast pulsing is used. Another important advantage of 
pulsed electron capture detection is that it provides an inherently 
more stable baseline than D.C. operation. This allows the use of 
column oven temperature programming to improve peak shape and 
reduce analysis time. For the reasons discussed, the pulse 
sampling technique provided for an increase in reproducibility over 
the previous D. C. mode.
28
Wentworth (80) derived a response relationship based on steady
I
state kinetic relationships. In addition, Wentworth and coworkers 
showed two major mechanisms of electron attachment, dissociative 
capture and resonance capture which could be distinguished based on 
their temperature dependence (81). In the first case, the negative 
ion dissociates and forms radicals and ions or a radical ion which 
can subsequently react further (Reaction I), and in the second 
case, a stable negative ion is formed which will only recombine 
with the positive species or react with the radicals (Reaction 2). 
The reactions are listed as follows:
AB + e" -*• A + B“ (I)
AB + e" 2 AB“ (2)
For compounds that react with thermal electrons by a dissociative 
mechanism, the response is higher at higher detector temperatures.. 
Compounds capturing electrons by Reaction 2 have their highest 
response at low detector temperatures and are thereby easily 
differentiated from those capturing electrons through Reaction I.
Simmonds developed the coaxial ECD with a gold foil plated 
with a 6% i  source capable of temperatures up to 400oC (82). 
Previously used titanium tritide sources could not be heated above 
225°C without producing a radioactive hazard. In 1971, a high 
temperature ^H-Sc source became available (83). The temperature 
limit of this source was 325°C and had a high sensitivity to
29
pesticides. A 147Bm-Au foil capable of operating at temperatures 
to 400°C has also been used as an ECD source with properties 
comparable with those of ^ N i  (34). one of the major advantages 
of using ® % i  as the radioactive source is that with the sustained 
use of oxygen over long periods of time (several years), no 
decrease in sensitivity or detector performance is noticed.
Maggs sfc al. (85) described a hew mode of operation of the 
pulsed BCD in which the detector current is held constant while the 
frequency of the applied pulses is varied. When a compound enters 
the detector and absorbs electrons, the pulse frequency will 
increase to collect more electrons and keep the current at the 
predetermined level. The average current Icell is proportional to 
the concentration CneJ of free electrons in the cell, and the 
frequency f of the applied pulses as follows:
1Cell0^ IneIf (3)
Hence, when only carrier gas is in the cell, [Ne] is relatively 
large and f is low. This pulse frequency with pure carrier gas can 
be defined as the base frequency, fD, of the system. Typically, fc 
is so low that the time between pulses is about 1000 times longer 
than the width of each pulse. Hence, for most of the time there is 
no polarization voltage across the cell. When an electronegative 
sample enters the EC cell, it converts some of the free electrons 
in the cell to negative ions. As a result, [ne] decreases and f 
increases in order to keep Icell constant. The passage of an 
electronegative sample through the electron capture cell produces
Ian output signal with peak height proportional to the frequency 
difference(fA - f0) where f& is the frequency corresponding to a 
sample concentration [A] within the EC cell. Maggs al. (85) 
have shown that (fA - f0) is proportional to [na3, so that there 
is an inherently linear relationship between peak height signal and 
sample concentration.
Patterson (86) has reported a displaced coaxial ^3Ni ECD which 
uses a narrow pulse width(0.64 usee) and has a 0.3 ml cell volume 
with the carrier gas flow counter to the electron flow. This 
arrangement permits sufficient electron collection to ensure 
satisfactory detector response. A linear dynamic range of nearly 
five orders of magnitude can be obtained from this cell. A major 
advantage of the wide linear dynamic range of this constant current 
ECD is its general tolerability to modest levels of detector 
contamination (86). Patterson also found that the response of this 
ECD is linear until the electron density is reduced to less than 
0.5% of its original magnitude. The Varian ECD that is used in our 
oxygen-doping studies is designed after the one reported by 
Patterson (86).
AEIMS and Plasma Chromatroaraphy
The recent development of two analytical techniques: plasma 
chromatography (87) and atmospheric pressure ionization mass 
spectrometry (88,89) has added to the understanding of reactions 
occurring in the ECD and the oxygen-doped ECD . Both of these 
instruments have ^3Ni ionization sources very similar to the ECD in
30
31
terms of pressure and temperature. Instead of measuring electron 
density as in the ECDf these devices allow measurement of the ions 
produced frcm electron-molecule and ion-molecule reactions.
Karasek and Kane (90) found that at 125°C 90% of the 
electrons in their ion source of a plasma chromatograph react with 
oxygen and water in the air to frcm water clusters of the form 
(H20)n°2~ where n varies as a function of water concentration and 
temperature. They also observed that the addition of H2O to 
nitrogen containing appreciable oxygen resulted in a much larger 
decrease in electron concentration resulting in the formation of 
more (HgO)^Og" than the presence of oxygen fcy itself.
Spangler and Collins (91) similarly identified O2" and 
(HgO)Og"" as the predominant negative ions in plasma chromatography 
using zero grade air as the carrier gas. They also studied the 
effects of water concentration upon the identity and concentration 
of reactant ions. Dzidic st al. (92) , with an APIMS, identified 
Og" and its hydrated forms in high purity nitrogen carrier gas 
thought to contain approximately 100 ppm O2 and 10 ppm HgO. In 
addition, they studied the reactions of m-chloronitrobenzene and p- 
chloronitrobenzene in an API source. The m-iscmer reacts through 
electron attachment, and this leads to a stable M~ radical ion.
The reaction sequence for p-chloronitrobenzene leads to two 
negative ions: chloride ion and the nitrophenoxide ion of mass 
equivalent to (M-C1+0). Grimsrud sfc al° (93) also observed the (M- 
Cl-K)) anion of p-chloronitrobenzene and other chlorinated compounds 
when using their GC/APIMS. Horning and coworkers (94) observed O2"
32
and (H2O)C^- to be the predominant ions in the negative ion spectra 
at IOO0C for nitrogen carrier gas containing approximately 2 ppt 
oxygen (V/V). In a study of negative ion formation for 
polychlorinated biphenyls, using a plasma chromatograph, Karasek 
(95) found that the presence of a single chlorine atom in one or 
both rings was not sufficient to form M- ions, but that the 
presence of three or more chlorine atoms always led to negative 
organic ions.
Use of Oxygen As An BCD Dopant
Grimsrud and Stebbins (96) reported the effect of oxygen on 
the response of a constant-current ® % i  electron-capture detector. 
Because of the greatly increased linear dynamic range and the high 
temperature capabilities of this instrument, oxygen contamination 
of the carrier gas was found to be much less harmful to the 
chromatogram baseline than had been previously reported for earlier 
ECD models. They also found that adding oxygen to the carrier gas 
increased the baseline frequency and the baseline frequency 
increased with lower detector temperatures. This is consistent 
with a resonance electron capture mechanism proposed for oxygen and 
similar compounds (Reaction 2). Water itself produced very little 
or no effect when the oxygen concentration was low, but at high 
concentrations the baseline frequency becomes significantly 
affected.
For the highly chlorinated molecules (e.g. tetrachloroethane, 
pentachloroacetone, and tetrachloroethylene), the effect of oxygen
33
was observed only at high levels of oxygen doping and at lower 
detector temperatures. However, 1—chlorobutane showed large 
response enhancement factors, 14 and 55, with 800 to 2000 ppm added 
oxygen, respectively. Anthracene exhibited somewhat different
response characteristics, even at low concentrations of Og (150 ppm) 
the response enhancement was 15.
Other researchers have utilized 02-doping, primarily for the 
purpose of increasing the sensitivity for a specific compound. 
Simmonds has used high purity nitrogen carrier gas containing added 
oxygen (100 ppm V/V) for the analysis of carbon dioxide and nitrous 
oxide in the atmosphere (97). The addition of oxygen to the 
carrier gas produced a signal for carbon dioxide while preserving 
the usual response to nitrous oxide, thereby allowing analysis of 
both gases from a single sample of ambient air. Simmonds 
postulated a charge transfer mechanism involving O2- to be 
responsible for the increased response to COg using oxygen doped 
carrier gas. His mechanism is almost identical to that discussed 
by Grimsrud and Stebbins (96) and Miller and Grimsrud (98).
An oxygen doped GC/ECD is currently being used for the 
continuous on-line monitoring of bis(chlorcmetbyl)ether in the 
workplace environment of Dow Chemical Company, Midland, Michigan 
(99). This fully automated system performs the monitoring tasks 
previously done by operator-assisted GC/MS with considerably less
34
expensive equipment and significant financial savings in operator 
expense.
Pasmussen and coworkers (100) recently reported the use of the 
use of oxygen doped nitrogen carrier gas to measure latitudinal 
distribution of methyl chloride in the atmosphere with a constant 
current ^ N i  electron capture detector. These data were obtained 
from any analysis procedure very similar to that reported by 
Grimsrud and Miller (101,102). More, recently, Rasmussen and Khalil 
(103) have reported the first measurement ever of ambient 
atmospheric concentrations of CHClFg ( Freon 22 ). These 
measurements were obtained with an oxygen doped ECD operated at 
275°C using the techniques previously described by Rasmussen 
St Si. (100).
A mechanism for oxygen doping was proposed by Grimsrud and 
Stebbins (96) and was similar to that mentioned by Van de Weil and 
Tcmassen (104). Through an ion molecule reaction O-f and the 
sample molecule coupled to the electron attachment-detachment 
equilibria with Og. If the rate of this Og- route of electron 
attachment is faster than the rate of direct electron attachment, 
an enhanced ECD response would be expected. The following 
reactions were proposed to account for this mechanism.
(4)
(5)
O2 + e“ t  O2- 
O2- + A  ^  Products 
A  + e- Products (6)
35
The last reaction indicates the reaction of the analyte molecule 
with the gaseous electron, the basis of the normal BCD response. A 
measured response enhancement is thought to reflect the ratio of 
the rates of Reactions 5 and 6 and may depend strongly on small 
variations in the structure of the substrate molecule.
Using published thermodynamic data, Grimsrud and Stebbins (96) 
calculated the relative concentrations of the ions, Og^fOg(HgO)"", 
and O f . The clear results of this consideration is that in dry (I 
ppn of water) carrier gas, Og- is considered to be the dominant 
negative ionic species at the temperatures studied. Even with 2000 
ppm of oxygen, added at the lowest detector temperature examined,
150°C, O f  will be only 0.002 as abundant as Og-. Kinetic 
considerations of the electron attachment-detachment equilibrium 
with Og have enabled them to explain the oxygen dependence of the 
baseline frequency.
The enhancement of response to 32 simple chlorinated molecules 
of a constant-current electron capture detector caused by the 
addition of oxygen to its carrier gas was determined by Miller and 
Grimsrud (98). The effect of size and isomeric differences for 
alkyl, vinyl, allyI, and phenyl chlorides were examined. As shown 
in Figure 5 , the response enhancements for several isomeric 
dichloro compounds are illustrated. This was probably the first 
clear indication that oxygen addition could be used for isomer 
identification. The response enhancements obtained for the trans 
isomer of dichloroethylene was 3.9, the cis by 5.5, and the 1,1- 
dichloro isomer was enhanced by only 1.1 at 300°C with an oxygen
36
8 10
Time (min.)
Figure 5. Typical chromatograms in the studies of oxygen's
effects on ECD responses. The bottom chromatogram is 
with normal BCD conditions and the top ones are with 
0.20% oxygen in the carrier gas. (Reference 98).
37
concentration of 2.0%. In that sane study, they found that the 
number of chlorine atoms per molecule had a large effect on the BCD 
response and on the response enhancements. As an example, the BCD 
responses increased by a factor of nearly IO4 in going from CHgCl 
to CCI4, while the corresponding response enhancements decrease 
from 113 to 1.2. In addition, they showed a correlation between 
the bond dissociation energy and measured relative rates of 
reaction with C^- for several selected chlorinated hydrocarbons.
The enhancement of electron capture detector response with the 
addition of oxygen to the carrier gas was extended to polycylic 
aromatic hydrocarbons and related chior©hydrocarbons (105). Large 
aromatic molecules possess EC sensitivities and exhibit oxygen 
induced response enhancements of magnitudes sufficient to allow 
their measurement in trace analysis schemes involving gas 
chromatography. The measured RE values are dependent on subtle 
structural variations, and in fact, can be used to detect 
structural differences among isomers that can't be resolved by 
existing forms of mass spectrometry. For example, triphenylene and 
chrysene have identical mass spectra and have the same retention 
times even on capillary columns, but their RE values are 3.0 and 
140, respectively. Another example includes the isomers 
phenanthrene and anthracene. Based solely on response 
enhancements, these two compounds can be distinguished. The RE 
value for phenanthrene was determined to be 4.6, while that of 
anthracene was 62.
38
Model for Instrument Response with Oxygen Doping (97)
The formulation of the model for instrument response with 
oxygen doping is similar to that described by Cobby, Grimsrudf and 
Warden (106) of a clean gas condition of the ECO. The model tor 
©2"doping by Grimsrud and Miller (98) differs in that oxygen is 
always present and the removal of a constant fraction of electrons 
to the electrometer during each sampling pulse and the acheivement 
of a steady-state condition for all charged particles during the 
period between pulses is assumed. All neutral and ionic species 
are assumed to be uniformly mixed within the ECD so that their 
presence can be described by one concentration expression. While 
an assumed even distribution of electrons may be suspect in the 
undoped, pulsed ECD where a significant space-charge, positive ion 
field may exist following the removal of electrons by a sampling 
pulse (93), this assumption is much more valid in the Q2-doped BCD, 
because most of the negative charge carriers are ions which are not 
removed with each pulse and overall charge neutrality within the 
ECD should be quickly reestablished (107). The assumption that a 
steady-state condition is acheived during the time between pulses 
is reasonable in this instance because the rates of electron 
attachment to Og and detachment from Og- will be faster than the 
pulse frequency.
39
The following reactions might be expected to occur in an oxygen- 
doped ECD during the period between sampling pulses:
P+ + e"radiation. 
k aA
e + A —> B- + neutrals
ko
6 + G2 O2
— v A
O2 + A- C- + neutrals
, P
P + e“ — — neutrals
, R
P + negative ions ----> neutrals
(7a)
(Tb)
(7c)
(7d)
(7e)
(7f)
The ®^Ni ionization of the carrier gas molecules to positive ions 
and electrons is presented in Reaction Ta. ^Reaction Tb shows the 
direct electron capture of the analyte molecule A. Reaction Tc is 
electron attachment and detachment with oxygen and Td is an ion- 
molecule reaction between O2- and A. Reactions Te and Tf are the 
recombination reactions of positive ions with electrons and with 
negative ions. It is assumed that the recombination Reactions Te 
and Tf occur with the same rate R. Siegel and McKeown (107) have 
considered this point in detail for conditions within the ECD.
They have also shown that other possible loss mechanisms for 
charged particles in a ECD. at equilibrium, such as diffusion to the 
walls or ventilation, contribute very little to the total loss rate 
which is dominated by recombination. These loss mechanisms are not 
considered to be important during the steady-state portion of the 
period between pulses.
Grimsrud and Miller (98) assumed that the total positive ion 
density remains fairly constant and independent of the amount of O2
40
or sample present. It is also assumed that during the field-free 
period between the BCD sampling pulses that charge neutrality 
within the cell will be obtained.
where S is the ion-electron production rate from the ^ N i  source, R 
is the rate constant for recombination, V  is the volume of the cell, 
R is the recombination rate, and n+ is the number density of the 
positive species. For a 7.SmCi ® % i  source of 0.3 ml volume, S is 
about 1.3 X IO11 s-1 assuming 17-keV average primary energy and one 
ion pair per 35-eV primary energy (107). Taking R to be about 3 X 
IO-6 mLs-1 (107), Hf is calculated to be about 3.8X10® mlT1. 
Grimsrud .&L. (93) have obtained experimental support for the 
assumptions leading to the previous equation using the ECD/APIMS 
instrumentation described above and elsewhere (93). They noted 
that in the saturated condition (all negative species are ions), 
that the total positive ion signal is equal to the total negative 
ion signal. Also, upon changing the source from a condition of 
electron predominance to saturation, the total positive ion signal 
of our APIMS does not change drastically, but increases only about 
50%. With high levels of oxygen doped into the carrier gas, the 
total positive ion signal is altered almost negligibly as sample 
enters the ion source, because in this case ions are the 
predominant negative charge carriers at all times.
41
In the absence of added sample and relating the rate equations 
with the fact that each approaches a steady state condition at the 
end of each field-free period, an expression for the electron 
density can be obtained.
' Stk-Q + Rn+)
6 v<k-0Rr|+ + kon0„Rn+ * r X 2) <9)2
For a frequency-modulated ECD the pulse frequency, f, is thought to 
be inversely related to the population of electrons existing at the 
instant of the sampling pulse (86,108). Also, the frequency will 
be proportional to the cell current demanded by the electrometer. 
Therefore, the following equation can be written:
f J<LVtl (10)
For the instrument we normally use, the electrometer current, I, is 
held constant at 0.3nA when nitrogen is used as the carrier gas. K 
is an empirical proportionality constant. This term includes 
unknown factors for the electron collection process, such as the 
fraction of electrons collected during the 0.6p_s, 50-V pulses in 
nitrogen (86), and the contribution to the overall current of 
positive ion migration to the collection electrode during the period 
between pulses (93). Substitution of Equation 9 for ne and 
arithmetic reduction yields the following equation:
T0
KI Rn+ko
S k„0 + Rnt  nO2 (U )
42
where the frequency is given as f0 to indicate this is the 
corresponding baseline frequency with oxygen-doped carrier gas and 
no sample (nA  = 0).
With sample present in the oxygen-doped ECD and application of 
a steady-state condition, an expression for f can be obtained.
f T (keAnA + RrU
KI kOkOAnA *  Rnt ko 
5 k-o  + kOAnA + Rn+
(12)
The response of a constant-current ECD is then taken as the increase 
in frequency, f-fQ, caused by the presence of sample molecules.
An expression for this response can be derived from Equations 11 and 
12 and reduced to the following expression :
L(keAnA) + f
(k-o +
k-o kokoAnA
kOAnA Rn+ )(k  + Rn+) 1O2'13'
The first term is the contribution to the response of the electron 
capture reaction Tb while the second term is the contribution due to 
the assistance provided by
' An expression for the oxygen-induced response enhancement (RE) 
is obtained in Equation 14 by dividing the total response, f-£0r by
43
the contribution of the normal EC response, KI/S(Rg aUa)
k R k
RE ~ 1 + keA(k-o + kOAnA + RV (k-o + rI+1 " \  <14)
If k-0 is much larger than R0AnA anc^  Rn+, the above equation can be 
simplified. This appears to be justified under the experimental 
conditions used. As an example, at 300° c R_0= lx IO5S-1 (calculated 
from its second order rate constant in nitrogen (109) of R_0 = 
1.9(+0.4) X 10-U(T/300)3/2e-4900/T ^ - 1  x ^ 1.3 X IO19 mL"1,
the number density of the buffer gas at I atm pressure, 300° C).
The product, Rn+, will be about 3 X 10“6' mL-1 = 1100 s"1. The 
product R0AnA will vary with the sample A and its concentration.
The very fastest ion-molecule reactions occur with a rate constant 
of about I X 10~9 InLs-I (HO). The concentration of the 
monochloroalRane samples were about O.lppm or 1.3 X IQl2 mL-1, at 
300° C. For these cases RoAnA will be less than 1.3 X 10^, 
considerably smaller than R.q. Equation 14 is reduced to Equation 
15
The ratio Ro/R-o in Equation 15 is the equilibrium constant for 
Reaction 7c. In an oxygen carrier gas, R0 = 1.4(+0.2)X 10~ 
29(300/T)e(-^00/T)mL2s-;L(109) and R_0 = 2.7(+0.3) X 10“
30 (T/300) l/2e (-5590/T) mLs“l (109). The ratio of these rates yields 
an equilibrium constant which will apply to Reaction 7c in any
44
buffer gas since the dependence of the third body cancels in the 
ratio. Equation 16 can be written
I. + 2.7 X TO"16 T"3^ 2 e ml *
eA
(16)
where T is the temperature in degrees Kelvin. Under conditions used 
by Miller £t sL. (98) with 2 ppt O2 ( 5.4 X IO16 273/T mL"1) the 
previous equation reduces to the following:
RE = ! + 8.8 at 250°C (17)
keA
RE = ! + 3.1 at 300°C (18)
keA
RE = I + 1.2 ^ at 350°C (19)
eA
Expressions for the overall response, f-f0. with oxygen doping can
be derived from Equation 13 in the same manner as above. At 300° C, 
the following relationship is obtained:
f - f o ' f  <keA + 3-° kOAlrlA at 300°C ,20>
The experimental characterizations of oxygen doping determined 
(98) thus far are in good agreement with the model developed. It 
has been shown that the base-line frequency, f0, increases linearly 
with oxygen concentration as required fcy Equation 10 (101) and is 
illustrated in Figure 6. It has also teen shown that with oxygen 
doping the response is linearly related to sample concentration 
throughout the linear range of our instrument as required by 
Equation 20. In addition it has teen shown (see reference 98,
B
as
el
in
e 
Fr
eq
ue
nc
y 
(K
H
z)
7  200
0 0.5 1.0 1.5 2.0 25
Concen tra t ion  (ppth)
Figure 6. Effect of oxygen on the baseline noise at several detector 
temperatures. (Reference 101)
46
Figure 2) that the relationship between R E - I  and oxygen 
concentration is a nearly linear one as required by Equation 15.
The general observation that these plots show some curvature from 0 
to 1% oxygen is possibly due to small but notable changes in the 
total positive ion concentration caused by small changes in the 
effective recombination rate, R, as the predominant negative species 
present are changed from electrons to negative ions.
The temperature dependence of response enhancements predicted 
by Equations 17-19 also appear to be in harmony with experiments 
which might be reasonably used to test this relationship. For 
example, 1,2-dichloroethane has an ECD response which appears 
relatively insensitive to temperature change (98). Its relative 
oxygen-induced response value is high suggesting a large value of 
koA for it. It is reasonable to assume that this rate constant will 
also be relatively insensitive to temperature , as fast ion-molecule 
reactions generally are (110). Since the ratio of koA/keA for 1,2- 
dichloroethane might then be assumed to be insensitive to 
temperature change, the RE - I values as predicted by Equations 39a- 
c should have relative values of approximately 8.5:3.0:1.2 at 250°
C, 3 OO0C, and 350°C, respectively. The response enhancements for 
12-dichloroethane are 161, 69, and 22 at those temperatures, which 
yield relative values for RE - I of 8.8:3.7:1,2, in good agreement 
with the above expectations. The ratio of the ion-molecule and 
electron capture rate constants, Ro aA sa, which caused these 
relatively large enhancements is deduced to be equal to about 20.
47 V
The experimental results in general give credence to the model tor 
oxygen doping proposed by Grimsrud and Miller (98).
Other dopants
Fehsenfeld and coworkers (111) reported that the sensitivity of 
the ECD to CO2, Hg and CH4 could be enhanced by the addition of 
nitrous oxide to the carrier gas. Shortly thereafter, Sievers and 
coworkers (112) further characterized the NgO enhancement process 
and demonstrated that a nitrous oxide chemically enhanced ECD 
responds to alkanes with about the same sensitivity as the flame 
ionization detector. The proposed mechanism for this enhancement 
process (112) postulates that a fast irreversible charge exchange 
ion-molecule reaction between O- and the sample is responsible tor 
the detector response. The proposed reaction mechanism is 
illustrated in the following equations :
e“ + N2O 0“ + N2 (21a)
CT + N2O -*• NO" + NO (21b)
r  + N2 . NO + N2 + e" (21c)
A ■*" stable negative ion (21d)
Electrons are attached (Equation 21a ) and released (Equation 21c). 
The concentrations of electrons, Cf, and NO- in the ECD depend 
on the rate constants for Equations 21a-d , the concentration of N2O 
in the N2 carrier gas, and the operating temperature of the 
detector. Thus, any compound, analyte A, that can react with 0“ or 
NO" to form a stable negative ion ( Equation 21d) will interrupt the
48
reaction cycle, causing a reduction in the electron density and a 
detector response. This technique has been termed " selective 
electron-capture sensitization ( SECS)”. Goldan si- (113) were 
able to detect part per billion levels of vinyl chloride using the 
N2O doped BCD. This represents a very significant improvement when 
compared to the normal ECD detection limit for vinyl chloride of 
approximately 1-5 ppm (V/V). In a recent publication, Sievers et al • 
(114) determined the response enhancements of water, phenols, amines 
and aromatic and heterocyclic compounds using the NgO doped BCD.
From a comparison of the NgO signal enhancements with those obtained 
by Miller et el- (105) several interesting features were noted. As 
an example, triphenylene and dibenzothiophene exhibited relatively 
large enhancements with NgO addition, although they were among the 
least enhanced compounds when O2 was added. Substitution of a 
methyl group at the 9-position of anthracene decreased signal 
enhancement slightly with N2O, while the response enhancement values 
increased by a factor of two with Og addition. Water exhibited a 
relatively large enhancement factor, 260 and could be sensitively 
detected with selective electron capture sensitization. In general 
terms, doping with N2O gives smaller enhancement values, and at the 
same time, is less dependent on isomer.
The reaction with 0“ is more exothermic than with O2- and not as
selective.
49
RESEARCH OBJECTIVES
The objective of this study is to apply the O2-^oped ECD to 
isomer differentiation of substituted polycyclic aromatic 
hydrocarbons , particularly in instances where no assistance for 
identification is provided by conventional forms of mass 
spectrometry. Also, the use of oxygen-induced response enhancement 
for the quantitative analysis of completely unresolved or partially 
resolved components of gas chromatographic peaks was evaluated.
For compounds whose normal EXD response is very low, the use of 
EC-enhancing chemical tags was examined. Again, isomer 
differentiation with the 02-doped ECD was studied.
Also worthy of examination was the use of the APIMS for 
indication of the ions formed with and without the presence of 
oxygen in its source for different compounds. A comparison of 
response enhancements determined from the APIMS measurements and the
O2-doped ECD was made. In addition, conventional forms of mass 
spectrometry were evaluated for isomer distinction.
In order to improve the technique of oxygen enhancement 
measurements, the use of two ECDs for simultaneous electron capture 
and oxygen-sensitized electron capture detection was examined. Both 
a parallel and series arrangement of the two ECDs was tested in 
order to improve the precision of the enhancement measurements.
50
EXPERIMENTAL
Instrumentation
A  Varian 3700 gas chromatograph equipped with an Aerograph 
Flame Ionization Detector (FID) and a ® % i  constant current 
electron capture detector (ECD) was utilized to obtain the response 
enhancement data reported here. The EC ceil is a ceramic-metal 
assembly capable of operation at temperatures in excess of 400°C, 
for short periods of time. Figure 7 presents an illustration of 
the unique cell geometry of the Aerograph ECD (86). The cell has 
displaced coaxial cylinder geometry with the larger cylinder 
containing a 7.5 mCi ® % i  foil and the second cylinder serving as 
the electron collector. Carrier gas from the GC passes through the 
collector cylinder first, and then up through the foil cylinder. 
This particular arrangement minimizes the diffusion and convection 
of electrons into the collector cylinder (86). In addition, the 
number of long-range beta particles striking the collector is 
reduced. In effect, this cell minimizes the "field-free" 
background current, which is defined as the contribution to the 
cell current which does not vary with the pulse frequency (86).
The cell is polarized fcy applying negative voltage pulses of 50V 
amplitude and 0.64 microsecond width to the walls of the large 
cylinder while grounding the collector cylinder. The net result is 
that the movement of electrons within the foil cylinder is against
51
Figure 7
Radioactive
Foil
Negative 
Voltage Pulse
Collector Electrode
Gas Flow
Schematic illustration of the displaced coaxial cylindrical 
BCD used in this work.
52
the gas flow. A fraction of electrons are collected during the 
time the pulse is on. As a result, the unique geometry has allowed 
the cell volume to be made very low (0=3 ml) for increased 
sensitivity.
The cell current is held constant at 0.3nA using nitrogen as 
the carrier gas through the use of the electronic circuit 
illustrated in Figure 8. The cell current, Icell, is combined 
with an external reference current such that their difference is 
the input to the electrometer. The electrometer output determines 
the magnitude of Icell • The system works by electronically 
varying the pulse frequency to maintain the relationship Icell = 
Iref . Since the pulse frequency is the quantity which is always 
changing in this method of operation, the output signal is a 
voltage proportional to that frequency.
Using the GC/ECD system just described, the data for this 
study were collected. In addition, a small portion of the data 
were obtained using the flame ionization detector to compare the 
retention times in instances where more than one peak was eluted. 
The Aerograph FID has a linear range greater than IO^ and a 
detection limit of less than 2.7 X 10"-^ g-o/sec (115). This FID 
uses a ceramic flame tip to minimize the effects of sample 
degradation at metallic surfaces. The small contribution to the 
system noise by the electrometer allows the FID system to operate 
at a maximum senstitivity of IXlO-^  a  full scale. Retention times 
were determined for all the compounds discussed by using the flame 
ionization detector prior to the ECD analysis.
ECD Cell
Radioactive
Foil
Electron
Collector
Figure 8.
I T
Ionized
Gas
Negative 
Voltage Pulse
-Icell
^ref ~*cel|)
Variable
Frequency
Pulser
A
Eltictr Oivitittir
Signal Out 
(Volts a freq)
Reference
Current
Pulse Frequency Varied To Maintain IceU ■ I re^
Block diagram of the electronic components of a constant 
correct BCD.
54
Chromatograms were recorded on a Varian Model 9176 strip chart 
recorder or a BBC Goerz Servogar 303 three pen recorder.
Chromatographic Parameters
For the chlorinated biphenyls, a SE-52 capillary column,
.25mm i.d. X 15m,. was used at a constant oven temperature of ISO0C. 
The detector temperature was 30O0C and the injector was maintained 
at 210°C.
A 1.5-ft length by 1/8 inch stainless steel column packed with 
4% OVlOl on Chromosorb W was used for the mixture analyses of the 
chloroanthracenes. The temperature of the oven for packed column 
separations was 170°C and the injector was at 200°C. The detector 
was maintained at a temperature of 3 OO0C. A SE-52 capillary 
columndSm X 0.25mm LdJ was also used in the study of the 
chloroanthracenes . The oven was temperature-programmed as 
follows: 90°C for 3 min, 20°C increase per minute to 200° C, and 
hold at 200%  for 10 min.
A  SE-52 column of 0.25mm Ld. and 15m length was used with He 
as the carrier gas and nitrogen as the make-up gas for the amine 
and hydroxy derivatives. The detector was maintained at either 
250°C or 300°C and the injector temperature was 220°C 
Isothermal conditions of the column were generally used for both 
the amine derivatives and the methoxy compounds. For the 
naphthalenes the column temperature was 130%, anthracenes 180%, 
phenanthrenes 230%, fluorenes 190%, pyrenes 230%, and the 
benzo(c)phenanthrenes and chrysenes 2 3 0 %  A temperature program
55
of the oven was used for the mixture analysis and is as follows: 
ISO0C for I min, 10 C increase per minute to 230°C, and then 
hold at 23O0C for 10 minutes.
The methylanthracenes and methylphenanthrenes were analyzed 
with series detectors which will be discussed in more detail later 
at a detector temperature of 250°C, injector was at 210°C, and 
the oven at 180%. Again, a SE-52 column of 0.25 m m  i.d. X 15m 
was used .
The sample, of 2,3,7-trichlorodibenzo-p^-dioxin was analyzed at 
a detector temperature of 250°C, injector 210°C, and the oven 
isothermally programmed at 180%. The capillary column previously 
discussed was used.
The make-up gas for the capillary column studies was nitrogen 
and was passed through two traps containing 13X molecular sieve 
and CaSO4 . Two oxygen scrubbers (Alltech Oxy-Trap) were added to 
the carrier gas line to remove residual oxygen from the nitrogen 
gas stream. When heated to 200°C , this scrubber is specified to 
reduce oxygen content to less than 0.1 ppm. The addition of the 
scrubbers and traps eliminated the need for using ultra-high purity 
nitrogen as the carrier gas in packed column work or the make-up 
gas in capillary column studies.
An SGE on-column injector model 0C1-2 was utilized for 
analyzing the chloroanthracenes and the temperature used was as 
follows: 50°C for 3 minutes, 10°/min to 200°C, and then hold at 
200°C for three minutes. The detector temperature was maintained 
at 300%.
56
For capillary column analyses, the injection port is modified 
to reduce its volume by placing a glass liner inside a piece of 
stainless steel tubing. As a result, the chromatographic quality 
is definitely improved.
Oxygen Addition
Z -
With the packed, column experiments, oxygen was normally added 
in the carrier gas.prior to the detector after the column. In 
earlier studies, the oxygen was added prior to the injector, and 
oxygen has had no detrimental effects on the lifetime or 
performance of the columns used since the oven was not elevated in 
temperature. Either technique provided identical response 
enhancements.
For the experiments utilizing the capillary columns, oxygen 
was added as a make-up gas after the column just prior to the 
detector by combining the carrier gas flow with nitrogen gas 
containing 2.4% oxygen. This has keen accomplished by using the Hg 
flow system into the universal detector base upon which our ECD is 
mounted. The hydrogen inlet is located at the back of the 
instrument and is normally used for H2 addition to the flame 
ionization detector at the base of the detector. This is 
illustrated in Figure 9. By adjusting the flow rate of the 
nitrogen containing oxygen make-up gas to produce a baseline 
frequency indicative of some known oxygen concentration, 
measurements can be made at a constant oxygen level. From one 
determination to the next, a precise level of oxygen is adjusted fcy
57
^  ECD cell
i
INSULATED 
ECD TOWER
AIR LINE
IONIZATION 
DETECTOR BASE
"H 3"  GAS LINE
B Oxygen addition as 
known concentration 
of entire makeup gas
Fused silica column
Figure 9. Diagram illustrating positions for addition of oxygen.
58
fine-tuning the oxygen flow until a preselected magnitude of 
baseline frequency is obtained. A known concentration of nitrogen 
containing oxygen can also be added as the entire make-up gas at 
position labelled B in Figure 9. Either experimental design or 
method of adding oxygen has produced accurate response enhancement 
values for the compounds studied.
Measurenent procedures
The ECD chromatograms of individual compounds or mixtures 
which were easily separated from each other by gas chromatography 
were first obtained using nitrogen carrier gas or make-up gas as in 
the case of capillary columns. Oxygen was added to the make-up gas 
and the analysis of each compound was repeated. A signal 
enhancement value is calculated by dividing the response of the Pg- 
doped ECD by the response observed with normal electron capture for 
the same amount of compound.
peak height with oxygen
Response enhancement (RE) = ■ •__________________  (22)
peak height without oxygen
All compounds were analyzed at least three times both with and 
without oxygen.
Sample Preparation
The following chemicals were purchased from Aldrich Co: 3,4,5-,
2,4,5-, 2,4,6-, 2,3,6-, 2,3,5-, 2,3,4-, 3,5,4'- 
trichlorobiphenyl, 1-, 2-, and 9-chloroanthracene, 9- 
chlorophenanthrene, 1-, 2-, and 9-aminoanthracene,
59
9-aminophenanthrene, 1-, 2-, and 4-aminoEyrene, 2-, 3-, and 
4-aminobenz(c)phenanthrene, 2 ,1 -, 1,8-, 1,5-diaminonaphthalene, 2,4-,
1.5- , 1,4-, and 1,7-dimethoxynaphthalene, 9-metho2Qrf*ienanthrene,
4.5- and 1,9-dihydroxyphenanthrene, 9-amino-3-methoxyphenanthrene, 
12-methoxy-7-methylbenz (a)anthracene, 5-methoxy-7-
methyIbenz (a)anthracene, 8-methoxy-7-‘methylbenz(a)anthracene, 2- 
methylanthracene, 9-methy!anthracene, and 9,10-dimethylanthracene.
The hydroxy isomers of chrysene and 2,3,7-trichlorodibenzo-p- 
dioxin were obtained from the National Cancer Institute’s Standard 
Chemical Carcinogen Reference Repository. Samples of 1-, 2-, 3-, 
and 4-aminophenanthrene, 1-, 2-, 3-, 4-, and 9-aminofIuorene, 2-,
3-, 4-aminobiphenyI, and the I- and 2-aminonaphthalene and the 1- 
methylanthracene, 1-, 2-, 3-, 4-, and 9-methylphenanthrene were all 
kindly supplied fcy Doug Later and Milton Lee of Brigham Young 
university. The 2-, 3-, and 4-chl or obi phenyl and 2,3-, 3,5-, 2,4-, 
4,4'-, 2,6-, 3,3’, 2,5-, and 2,2-dichlorobiphenyl were all 
supplied by Richard Geer of Montana State University.
Standards for Uie chloroanthracenes, chlorophenanthrenes, 
methylanthracenes, methylphenanthrenes and chi or obiphenyl s were 
prepared in the following manner. One to five mg of the solid 
samples were dissolved in 10.0 ml of thiophene-free toluene. This 
mixture was subsequently diluted with a known volume of toluene.
If necessary, the second mixture was further diluted such that the 
resulting sample size was sufficient to produce small, but easily 
measurable peaks. Those sample sizes varied considerably and 
encompassed a wide range of concentrations. For example, sample
60
sizes range from I ng to .01 ng per injection produced easily 
measured peaks for most compounds.
The polycyclic aromatic amines (PAA) were derivatized by using 
a procedure similar to that utilized by Later sfc al. (116). 
Approximately 10-20 mg of each was dissolved in 10.0 ml of benzene. 
A 2-ml aliquot was placed in a 5-ml flask with I ml of 0.05M 
triethylamine in benzene. Next, 30 pi of trifluoroacetic 
anhydride (Aldrich Chemical Co.) or perfluoropropionic 
anhydride (Fluka) was added to the mixture. The flask was sealed 
and heated in a water bath at 60° C for 15 minutes. The reaction 
mixture was then cooled for several minutes , I ml of 10% aqueous 
ammonia solution was added to terminate the reaction, and the 
mixture was then shaken for 5 minutes. Finally, the organic and 
aqueous phases were allowed to separate and the benzene layer 
containing the derivatized amines was recovered and diluted with a 
known volume of benzene for subsequent GC/ECD analyses. The 9- 
isomer of fluorene (HCL salt) was treated with base before 
undergoing the reaction with either trifluoroacetic anhydride or 
perfluoropropionic anhydride. The general reaction for the amine 
with either trifluoroacetic anhydride or perfluoropropionic 
anhydride is shown in the following reaction.
+
O
HOcq1Fm
61
Masuda and Hoffman (117) used PFPA derivatization to identify and 
quantify I- and 2-aminonaphthalene in cigarette smoke condensate 
and reported yields exceeding 95% of several standard polycyclic 
amines to PFPA-amide derivatives (118). Similar yields were 
observed by Later st al. (116) for the PFPA-amides, however, 
somewhat lower conversion yields, 85-95%, were obtained for the 
PAA-TFAA derivatives.
Due to the limited solubility in benzene, the procedure for 
preparing derivatives of the diaminonaphthalenes and 2,7- 
diaminofluorene differed somewhat from the previous procedure. 
Approximately 100 mg of the diamino compound was dissolved in 0.5 
ml of tetratydrofuran (!HF) and diluted to 2 ml with benzene in a 5- 
ml flask. Next, IOOyuuL of TFAA and 0.5ml of .SM trie thy lamine in 
benzene was added to the solution. The flask was sealed and heated 
in a water bath at 60° C for 30 minutes. The reaction mixture was 
then cooled and allowed to evaporate under a steady stream of 
nitrogen. The solid residue remaining was then dissolved in 0.5 ml 
of THF and diluted with 10 ml of benzene for subsequent ECD 
analysis.
The methoxy and dimethoxy derivatives were prepared in the 
following manner. Two test tubes (20 X 150mm) were connected in 
series with glass tubing through neoprene stoppers so that incoming 
nitrogen bubbled through the solutions in the test tubes. In the 
first test tube 10 ml of ether was added. To the second test tube, 
I ml of ether, I ml of 2-(2-ethoxyethoxy) ethanol, 1.5 ml of 37%
62
. aqueous potassium hydroxide solution, and approximately 0.2mg of N- 
methyl-N-nitroso-p-toluenesulfonamide(Diazold, Aldrich Chemical 
CO.) were added. The nitrogen flow was adjusted to about lOml/min. 
The vial containing the hydroxide in either ether or THF was 
positioned so that the gas bubbled through the sample. The 
reaction was allowed to proceed for about 10 min until the yellow 
color of diazomethane persists in the sample vial. The reaction 
solution was then allowed to evaporate to dryness and the solid 
residue remaining was redissolved in toluene for subsequent 
analysis by GC/ECD. Standard precautionary procedures associated 
with the use of diazomethane were strictly adhered to.
All organic solvents were of pesticide grade, the water was 
deionized and doubly distilled, and the derivatization reagents 
were used without additional purification. The stock ammonia 
solution (Baker) was extracted three times with equal portions of 
methylene chloride to remove any organic contaminants.
Parallel Detectors .
Dual Columns
Two fused silica columns, SE-30, of 0.25 mm Ld. and 6m  length 
were placed in the injector port and used with He as the carrier 
gas and nitrogen as the make-up gas. When desired, oxygen was 
added by combining the nitrogen make-up gas with about 7 ml/min of 
nitrogen containing 2% oxygen. An oxygen concentration of about 
3 ppt is thereby maintained in the doped detector. From one 
determination to the next, a precise level of oxygen is adjusted by
63
fine-tuning the oxygen flow until a preselected magnitude of 
baseline frequency is observed, The two detectors were maintained 
at 300% and the injector temperature was 210%, Isothermal 
conditions of the two columns were generally used to determine the 
enhancements for the individual compounds. The column temperature 
was held constant at 1 4 0 %  A splitless injection of 0.4/xl was 
used where each injection contained an amount of sample in toluene 
to produce small, but easily measured peaks.
Splitter
Using a commercially available capillary column effluent 
splitter (Varian) as illustrated in Figure 10, an experimental 
design was utilized with one column (SE-52, 0.25 mm Ld, X 15m) and 
parallel detectors. The temperature of the column was 210%, the 
injector was 220%, and both detectors were maintained at 3 0 0 %  
Helium was used as the carrier gas. with nitrogen as the make-up 
gas. As discussed previously, the oxygen was added to one detector 
and adjusted to a preselected baseline frequency.
Series Detectors
In this experiment, an SE-52 capillary column 0.25 mm Ld. X 
15m was used with helium as the carrier gas and nitrogen as the 
make-up gas. The effluent stream from the first ECD is connected 
to the second one with a portion of glass-lined stainless steel 
tubing(1/16 inch ad. X 1.5 ft,). The tubing was wrapped with heat 
tape and heated to a temperature of 2 1 0 %  Both detectors were
64
To Detector A fo Detector B
Channel A> Channel B
Makeup gas
c Capillary Column
Split Point
Flow Stabilizer
Mixing Chamber
Figure 10, Schematic of the Varian capillary column effluent splitter.
65
maintained at a temperature of 300%. Isothermal conditions were 
used to determine individual response enhancement values. For the 
chi or oanth racenes and 9-chlorophenanthrene the oven temperature was 
ISO0C. The temperature program for the separation of 9- 
chlorophenanthrene and 2-chlor©anthracene was as follows; 140% 
for 2 min, 5° per minute to 200%, and hold at 2 0 0 %  for 5 
minutes. In this type of arrangement oxygen can be added to both 
detectors simultaneously.
Tandem Cells
This unique cell design is shown in Figure 11. Instead of 
having a coaxial pin design, this cell has the anodes coming in 
from the side of the cell and the pins are encompassed by high 
temperature ceramic material. Each volume contains a 63Ni foil of 
approximately 7.5 mCi and both cells are contained within the same 
block with a small transfer space between the cells. The effluent 
from the first cell flows into the second cell. . In addition, each 
cell has its own electronics package patterned after previous work 
by Knighton and Gfimsrud (119). This allows both cells to be 
monitored simultaneously and the second cell has the capability of 
adding oxygen. In other words, both the normal ECD response can be 
recorded as well as the oxygen-induced response simultaneously. 
This design is in its preliminary stage, but shows significant 
promise for analyses.
Mass spectrometric parameters
The samples were dissolved in benzene to prepare I yg/yl
66
Carrier gas outlet 
/T'
High Temperature
Ceramic Insulators
Ni foils
Additional 
Gas Inlet
T
Carrier gas inlet
Figure 11. Tandem cell configuration
67
solutions. One to five Pl of the solution were introduced into a 
VGmmlS GC/MS/DS equipped with a 4% OVlOl column (1/8 inch ad, X 
6ft) with a flow rate of 30 ml/min. The temperature of the column 
was 220%, injector was 230%, and the ion source temperature was 
also 230%. The transfer lines were maintained at 230%. The Cl 
spectra were measured using 99.5% methane and isobutane (Matheson) 
over the range of m/e 50 to 400, at 1.5 sec/scan and a I sec 
interval. The ionization chamber was maintained at a pressure of 
2X 10~5 mbar (.2 torr) for both methane and isobutane. The 
accelerating voltage was 4.0 kV and the emission current was 100 
amps. Mass calibration files were generated with 
perfluorokerosene (PFK). The electron impact spectra were obtained 
with an electron energy of 70 eV.
The B/E scans were performed with the VG 7070 GC/MS/DS 
equipped with a 4% OVlOl column (1/8 inch ad. X 6ft.) with a flow 
rate of 30 ml/min. The temperature of the column was 2 0 0 %  and 
the injector was 2 1 0 %  The source was maintained at 200% and 
the transfer lines were heated to 2 0 0 %  The spectra were 
measured using 99.95% isobutane (Matheson) over the range of m/e 60 
to 500 at 1.0 sec/scan at a I sec interval. The ionization chamber 
was maintained at 2 X 10“^ mbar. The accelerating voltage was 6.0 
kV, electron voltage was 90 eV, and the emission current at 0.5mA. 
Helium was used as the collision gas and the spectra were recorded
with a UV recorder.
X68
Atmospheric Pressure Ionization Mass Spectrometry (APIMS)
An atmospheric pressure ionization mass spectrometer (APIMS) was 
used for certain aspects of this study to identify ionic species 
with and without the presence of added oxygen. This homebuilt 
APIMS has an BCD as the ion source of internal volume (I ml) 
typical of ® % i  BCD's and includes a coaxial pin which serves as 
the BCD anode. A five-eighth inch nickel disk of 25 ym aperture in 
its center provides a controlled leak of the ions source contents 
into the vacuum region of a quadrupole mass filter. Figure 12 
illustrates the ion source and schematic of the APIMS used in these 
studies. With this instrument, the BCD response to an electron 
capturing compound can be monitored along with mass spectral 
measurements of the ions simultaneously formed in the API source.
In addition, the ions formed in the presence of oxygen can also be 
monitored. The APIMS was mass calibrated with SFg and 
hexachlorobenzene. An Aerograph gas chromatograph with a 1.5ft 
packed column containing 4% OVlOl on Chromosorb W was used to 
introduce selected samples into the APIMS. The temperature of the 
ion source was 250°C, and the carrier gas was nitrogen at a flow 
rate of 40 ml/miru By combining the carrier gas with about 7 
ml/min of nitrogen containing 2% oxygen after the column and before 
the cell, an oxygen concentration of approximately 3 ppt was 
obtained when desired. Negative ions with and without oxygen 
present in the carrier gas can be identified both by single-ion 
monitoring of chromatographic effluent and also by performing mass 
scans during the elution of the substances of interest. Usually,
69
20 m m
/ / /
movab le  
pi n
gas in
0.2 mC i
63Ni 
on d isk  
o f  3 m m  
d ia m e te r
25u apertu re
same 
d isk
PLUS
15 mCi
cylindrica l
foil on 
cell walls
Figure 12. IVzo sources used in the APIMS studies.
70
higher column temperatures were used for single-ion monitoring than 
mass scanning. In the scanning mode, one normally scans a 100 mass 
range in approximately 10 sec. In using lower column temperatures, 
the peak elution time is long enough to allow to scan the desired 
range. For single ion monitoring, one needs a high a signal as 
possible and that is why a higher column temperature is desired.
71
RESULTS AND DISCUSSION
The first portion of this study will be discussed in sections 
corresponding to the various sets of isomeric compounds examined: 
polychlorobiphenyls, chloroanthracenes and chlorophenanthrene, 
polycyclic aromatic amines, polycyclic aromatic hydroxides, 
methy!anthracenes, and methylphenanthrenes, and 2,3,7-trichlorodi- 
benzo-p-dioxin. The second portion will be involved with the 
techniques utilized in optimizing the precision with which 
individual response enhancement measurements can be made.
Polvchlorobiphenvls (PCBs)
Isomer Differentiation
The numbering scheme and structure for the PCBs is shown as
follows: / , •
3 2 2 3
ez 5
and a typical chromatogram indicating the response enhancements is 
indicated in Figure 13. The response enhancements for the 
monochloro, dichloro, and trichlorobiphenyls are listed in Table 2. 
For the monochlorobiphenyl compounds studied, the 4-isomer can be 
distinguished from the other two. Based solely bn response 
enhancement values, the 3—isomer could not be differentiated from 
the 2-isomer. The 3,5-isomer can easily be distinguished from the
72
Figure 13. The response of ECD to 3-chlorobiphenyl (bottom) and 
3,5-dichlorobiphenyl (top). fHie chromatograms on the 
left are with normal BCD conditions and the ones one 
the right are with 0.20% oxygen in the detector at 
2 50 0C.
73
Table 2. Response Enhancements 
Chlorinated Bipheryls
2
3
4
2, 3
3, 5
2, 4
4, 4'
2, 6
3/ 3'
2, 5
2, 2'
CO 4, 5
2, 4/ 5
2, 4, 6
CN 3, 6
to 3, 5
2, 3, 4
3/ 5, 4'
„E.) for the Chlorinated Biphenyls
R.E.
1.1 
1.6 
. 2a
4.0
2.3
2.5
1.5
1.0
1.9
1.0
2.1
2.2
2.6
2.3
2.1
2.2
2.1
74
others, in view of its relatively high response enhancement. The 
4,4,-isomer can be differentiated from the 2,6-, 3,3'-, 2,3-, .3,5- 
and 2,2'-chlorobiphenyls but not the rest of the dichlorobiphenyls. 
The 2,2' and 3,3' dichlorobiphenyls can be distinguished from the 
others based on their low enhancement values. The 2,4,6- 
trichlorobiphenyl isomer can be differentiated from the others, but 
based solely on enhancement values, the others could not be 
distinguished. As one increases the number of chlorine atoms, the 
response enhancement does not decrease. Even as the number of 
chlorine atoms increases to three and four, the response 
enhancements remain clustered around two. In an examination of the 
alkylchlorides, Miller sfc al, (98) found that as the number of 
chlorine atoms increased, the normal ECD response increased and the 
response enhancement usually decreased. It is also interesting to 
note that the response enhancements for all the biphenyl compounds 
are very low as compared to other compounds studied. This is 
probably due to the arrangement of the rings.
Chloroanthracenes and Chlorophenanthrene 
Isomer Identification
The three isomers of chloroanthracene and 9-chlorophenanthrene 
have identical mass spectra which are shown in Figure 14 and 15, 
respectively and the relative ion intensities are listed in Table 
3. The ions observed with the largest intensities are the parent 
ion at m/e 212 and .the ion at m/e 176 which represents a loss of 
HCl from the parent ion. Therefore, no distinction is expected to
75
Figure 14.
C l
C C O
I. J J • J  t ...I. — . J l  I. .
5 0  1 0 0 1 5 0  2 0 0
C C C ^ '
■ • J I . jl . ' U i  I . ■. fc Jl i I
5 0  1 0 0 1 5 0  2 0 0
c S o
} l J L I
5 0  I d o
H  l  ■.
1 5 0  2 d o
The electron impact mass spectra for l-,2-„ and 
9-chloroanthracene.
Figure 15. The electron impact mass spectrum of 9-chlorcphenanthrene.
77
Table 3. Relative Abundances for the Chloro Iscmers of Anthracene and 
Phenanthrene
% Abundance
M 2A m BE
75.1 6.0 5.0 6.0 4.9
88.1 10.5 10.6 8.9 11.8
106.0 7.5 6.9 6.3 4.1
150.1 8.0 ■ 7.5 7.9 4.1
151 8.2 7.8 7.9 6.0
175 5.6 5.3 5.9 5.6
176 31.1 30.7 32.6 23.9
177 12.3 12.6 12.0 8.1
212 100 100 100 100
213 16.2 15.8 16 16.7
214 31.1 30.6 29.2 34.9
78
be provided,By conventional forms of mass spectra. In Figure 16, 
typical chromatograms are shown from which determinations and 
response enhancements are obtained. The first chromatograms are 
for the pure isomers of 1-, 2-, and 9-chloroanthracene and 9- 
chlorophenanthrene. For each pair of chromatograms the lower one 
is obtained with approximately 3.0 parts per thousand oxygen 
continuously present in the detector at 300° C. It is seen thait 
the response of each of the pure isomers is enhanced by oxygen's 
presence, but by differing magnitudes. The response to the 1- 
isomer is enhanced by 4.0, to the 2-isomer by 6.9, to the 9-isomer 
by 21.0, and the 9-chl or ophenanthr ene by 1.8. In addition the data 
in Figure 17 indicate that the magnitude of the response 
enhancement observed for the isomers of chloroanthracene is a 
relatively constant value over at least two orders of magnitude 
change in concentration above the lowest concentration levels used 
here. This fact is extremely important and eliminates determining 
calibration curves for these compounds. The response enhancement 
values can be utilized to provide positive identification in cases 
where mass spectrometry fails.
Mixture Analysis (120)
If the composition of mixed peaks are to be determined from 
the measured response enhancement, it is necessary that the 
enhancement of a mixed peak, REmix , be expressible as the sum of 
the contributions of the individual components as shown in Equation
23.
79
Figure 16. EXD responses to 1-, 2-, and 9-chloroanthracene, and 
9-chloro£*ienanthrene (bottom) and chromatograms with 
0.30% oxygen present in the detector (top). Detector 
temperature is 300°C
Re
sp
on
se
 E
nh
an
ce
m
en
t
80
- o ---=----- o - o -
9 -  chloroanthracene
O —
y
10
5
oi
2 - chloroanthracene
_o___O
%
O 0”0 " O “ O
I - chloroanthracene ^
w " • ..................» ’ ......................- »“
10 100 1000
Relative Concentration
Figure 17. Oxygen-induced response enhancenents for the chloro- 
anthracenes as a function of concentration. A 
relative concentration value of 1.0 corresponds to 10 
ng injected.
81
(23)
where RE^ is the oxygen-caused response enhancement of the 
individual components, , is the molar fraction of substance i in
mixture is indicated in Figure 18. In order to test the validity 
of this relationship the response enhancements of many different 
mixtures of the chloroanthracenes have been measured. For the 
three possible sets of two-component mixtures, the results are 
shown in Figure 19 where the measured enhancement is plotted as a 
function of the molar fraction of one of the sample components. 
These data indicate that the measured enhancements are linearly 
related to the molar fraction of each component and that Equation 
23 does, indeed, describe the response enhancement expected for a 
binary mixed peak. It would be relatively straight forward, 
therefore, to determine the relative amounts of two isomers known 
to be present in a mixed binary peak of unknown composition. For 
these cases the unknown quantities sought, X1 and X2 , are 
obtained from the application of Equation 23 and the trivial 
relationship that the sum of the mole fractions is equal to one,
I • e.,
It is reasonable to expect that Equation 23 will also be 
applicable to three component peaks. The last pair of 
chromatograms in Figure 20 is seen to support this expectation, 
since response enhancements measured for the individual isomers and
I. An example of a two-component
X1 + X2 = I (24)
82
Figure 18. Normal BCD responses for I- and 9-chloroanthracene and 
a mixture of the two (bottom chromatograms) and with 
the addition of 0.30% oxygen in the detector (top) at 
a temperature of 300°C Mole fraction of 1- 
chloroanthracene is 0.19 and for 9-chloroanthracene 
is 0.81 in the mixture.
83
X  O
molar  f ra c t io n
Figure 19. Response enhancements of two-component mixtures as a 
function of the molar fraction of one of the 
components. The mixtures are composed of I- and 2- 
chl or ©anthracene (GU), I- and 9-chloroanthracene (0), 
and 2- and 9-chloroanthracene (x).
84
from the known molar ratios of each in the mixture, a value of 11.1 
is predicted (0.35 X 4.0 + 6.9 X 0.29 + 0.37 X 21.1= 11.1). 
Unfortunately, for the three component mixture system, the relative 
composition of an unknown mixed peak is not uniquely determined by 
a single enhancement measurement as in the binary case. In this 
case a single REm^x value has many possible solutions when applied 
to Equation 23 since there are then three unknowns and only two 
equations relating them. For the chloroanthracenes under 
consideration here a solution to this problem is provided by 
repeating all enhancement measurements at a different detector 
temperature. The above measurements have indicated that at 300° c, 
the measured enhancements are REj = 4.0, REg = 6.9, REg = 21.1 and 
REltlix = 11.2. With a detector temperature of 3500C, these 
enhancement measurements are found to be 2.5, 3.4, 15.6, and 7.6, 
respectively. Figures 20 and 21 illustrate the chromatograms 
obtained at 3OO0C and 350°C, respectively. With the data at 350°
C, Equation 23 can be used a second time to provide the third 
equation necessary to determine uniquely the molar ratios in the 
three component mixtures. Applying this procedure to the 
synthesized mixture of three isomers, the data indicate that the 
molar ratios are.X1 = 0.32, X2 = 0.31, and X 9 = 0.37. These 
values are in good agreement with the known values of 0.35, 0.28, 
and 0.37, respectively. The equations needed to solve these three 
component systems are as follows:
85
X 1
I
ii
L / - 4
Xl
Figure 20. ECD responses to pure isomers of chloroanthracene and 
a mixture of isomers without oxygen (lower 
chromatograms) and with 0.30% oxygen (upper 
chromatograms) in the detector at 300°C. Samples 
amounts are 2.9, 2.3, and 1.5 ng for the 1-, 2-, and 
9-chi or anthracene isomers respectively. The mixture 
contains molar fractions of these components of 0.35, 
0.28, and 0.37, respectively.
Figure 21. ECD responses to pure isomers of chloranthracene and 
a mixture of all three isomers without oxygen (lower 
chromatograms) and with 0.30% oxygen (upper chromato­
grams) in the detector at 350°C. Sample amounts are 
2.9, 2.3, and 1.5 ng for the 1-, 2-, and 9-isomers, 
respectively. The mixture contains molar fractions 
of these of 0.35, 0.28, and 0.37, respectively.
87
reMIX (iji^  } = X1RE1 + X2RE2 + XgREg (25)
r eMIX(T2) = X1RE1 ' + X2RE2 ' + XgREg' (26)
X1 + X2 + X g  = I (27)
In solving the three-component problem it is essential to 
recognize that the new set of RE values obtained at a second 
detector temperature must not be proportionately related to the 
second set. That is, if the new set were merely half, for 
instance, of the original set, no additional information will be 
obtained by the analysis at the second temperature. The reason for 
this is made clear by inspection of Equation 26 which is seen to 
be unchanged if all the RE values are simply altered by a 
proportional factor. Also, for this reason, no new information 
concerning the sample is expected by altering the amount of oxygen 
used in the detector for the enhancement measurement. We have 
previously shown that for the constant-current ECD that this change 
merely increases or decreases all RE values in proportion.
Also, with this detection scheme, it is important to recognize 
the situations where peak height rather than peak area can be used 
as the measure of detector response. With the short packed column 
used for these studies, the retention times of the three 
chloroanthracenes were indistinguishable. In this case their peak 
height or peak area could be used as a measure of response and the
88
same values for REfjjx an^ RE^ are detained (this is true as long 
as the normal and oxygen-sensitized responses are linearly related 
to sample concentrations over the concentration range of interest.) 
However, for the cases where the components of a mixed peak are 
partially separated by the column, care must be taken to use the 
peak area, only, as the measure of response. Then the method as 
outlined can be used for the quantitation of partially resolved 
mixed peak. For these cases, the use of the peak height as the 
measure of response will cause systematic error because at the 
instant of the peak maxima, the molar ratio of components in the 
detector differs from the molar ratio of these components in the 
original sample.
For chromatographic separations where only partial resolution 
of a mixed peak is observed, the use of the oxygen-sensitized EGD 
can be helpful in providing an indication of the nature of that 
partial separation. Consider, for example, the two chromatograms 
shown in Figure 22. The same mixture of the three 
chi or ©anthracenes as was previously described (Xj = 0.35, Xg= 0.29, 
Xg= 0.37) is now partially separated into a doublet by the use of a 
capillary column. The question one might ask is to which portions 
of this doublet do the individual isomers contribute. This 
question could be answered by very precise measurement of the 
retention times of the individual components or by the analysis of 
several prepared standards containing varied and known amounts of 
each of the three isomers. Alternatively, this information is 
provided very clearly and simply by repeating the analysis once
89
Figure 22. Capillary column gas chromatograms of the three- 
component mixture shown in Figure 20. The arrows 
indicate the point of elution of the chloranthracene 
isomers. The lower chromatogram is with normal EC 
detection and the upper is with 0.30% oxygen in the 
makeup gas. Detector temperature is 300°C.
Iwith oxygen in the detector. By a comparison of the two 
chromatograms in Figure 22, it is evident that the second peak of 
the doublet is due to the 9-isomer, alone, since its enhanced by 
about 21 times in the presence of oxygen. The first peak is 
enhanced by 5.4 and, therefore, is a composite of approximately 
equimolar quantities of the I- and 2-isomers. In correct terms 
= 0.55 and Xg = 0.45. The calculated response enhancement for 
this mixture of components should be 0.55 (4.0) + 0.45 (6.9) = 5.3 
which agrees quite well with the experimental value. Since the 
same sample analyzed here is the same three-component mixture shown 
previously in Figure 19, the above deductions are correct.
In addition, the same three-component mixture was analyzed 
with the oxygen-sensitized ECD using an on column injector. The 
sample is actually injected on to the column through a cold injector. 
The results are shown in Figure 23. Again, the enhancements give 
an indication of the identity of the components. In using the type 
of logic discussed previously, it is obvious that the first peak is 
approximately an equimolar mixture of I- and 2-isomers. In view of 
the large enhancement value obtained for the last peak, the 
identity of that peak is definitely the 9-isomer. The results 
indicate that for compounds that may be extremely unstable in a hot 
injector (thermally labile), this type of system will still lend 
itself for isomer identification.
It is anticipated that the detector scheme described here 
could be useful for the analysis of many electron capture-active 
compounds where the relative amounts of potentially present
90
91
Figure 23. Capillary column gas chromatograms of the three- 
component mixture in Figure 20 with an on-column 
injector. The arrows indicate the point of elution of 
the chloroanthracene isomers. Hie lower chromatogram 
is with normal EC detection and the upper is with 
0.30% oxygen in the makeup gas. Detector temperature 
is 300° C.
92
structural isomers under consideration meet certain minimum 
requirements of the method demonstrated here for the 
chi or ©anthracenes. First, the measured response enhancements of 
the pure isomers must be unique for each isomer. Second, the
I ,
normal and the oxygen-sensitized responses of the ECD must be 
linearly related to the concentration of each analyte over the 
concentration range of interest.
Atmospheric Pressure Ionization Mass Soectraneter (APIMS)
,1
In order to examine the ions formed by the chi or ©anthracenes 
with and without the presence of oxygen, the APIMS was used. An 
example of the negative ion API mass spectral scan taken during the 
elution of 9-chloroanthracene with and without added oxygen is 
illustrated in Figure 24.
The APIMS results for the chloroanthraeenes and 9- 
chlorophenanthrene are shown in Table 4 obtained from single-ion 
monitoring. Some interesting features come immediately to 
attention. Without any oxygen present in the system, both the 2- 
isomer and 1-isomer form primarily the M- ion.
The ratio of the M-/(M + Og)- is 6 and 3, respectively for the 
I- and 2-isomers. The 9-chloroanthracene and 9-chlorophenanthrene 
show primarily the (M-Cl+Og)- species. In the presence of 
oxygen (approximately 3 ppt), both the 2-isomer and 1-isomer form 
primarily the (M+Og)- species while both the 9-chloroanthracene and 
9-chlorophenanthrene form almost entirely (M-Cl+Og)-. These
1 'i
results are consistent with other researchers who have found
Y
93
Xl
I
i
I—  
150
U JVAl
250
I
50
M/E
Figure 24. Negative API mass spectral scans taken during elution 
of 9-c hi or ©anthracene without and with added oxygen at 
a source temperature of 250°C.
94
Table 4„ APIMS Results for the Chloronthracenes and 9-.
chlorophenanthrene from Single-Ion Monitoring a
without Oxygen
M/E JA 2A 9A 9P
212(M-) 2800 3000 200 100
244 (M-H)2)- 500 1000 1500 1000
194(M-CL+0)“ 100 100 100 100
209 (M-CL-K)2)- 250 100 4000 4000
228 (M-K))- 500 100 500 . 100
TOTAL 5000 6000 6500 6000
M/E ' with Cbrygen &
212 50
244 600 4300 100 100
194 100 50 50
209 5Q 6500 3500
228 100 100
TOTAL 1000 4000 6500 4000
irce temperature is to 8 p
b Oxygen concentration is about 3ppt.
95
similar reactions with other chlorinated species (91,92,94). it 
is reasonable to be able to distinguish the 9-isomers from the 
other two based on their propensity to form the (M-Cl-HD2)- species 
in the presence of oxygen in the APIMSL One could also possibly be 
able to do binary mixture analysis (for example, 9-chloroanthracene 
and 2-chl or ©anthracene} based on the ratio of the (M-Cl-HD2)- 
/(MHD2)- species.* The following reaction sequence is postulated 
for the chloroanthracenes and chlorophenanthrene and their reaction
with O2 in the API source. This is similar to that of Horning and 
coworkers (92,94).
( I )
O3 - M
I (3) (4) C I t2)
O2' +  M -----H>[M02] ' f = 5  O2 + M '
(5)
(6)
(M-ci+(^r
For the isomers of chloroanthracene and chlorophenanthrene, an 
intermediate (M-HD2)- is formed by reactions I and 3 or 2 and 4.
This intermediate may remain intact, as in the case for I- and 2- 
chlor©anthracene, or dissociate, through reaction 6 to form (M— 
Cl+O2)- ions which are observed for 9-chloroanthracene and 9- 
chlorophenanthrene. Another result of the API studies indicates 
that the position of attack by O2- occurs at the 9 or 10 position of 9 
chi or ©anthracene as well as the other chloroanthracene isomers. In
96
both the I- and 2-isomer, the presence of the (M-K)2)- species and 
for the 9-isomer, the presence of (M-Cl-K)2)- species with oxygen 
clearly indicates that the attack by O2- takes place at the 9 or 10 
positions of chloroanthracene. The nucleophilic attack of O2- on 
chloroanthracene probably occurs by the following reaction.
These predictions are also consistent with results calculated 
from simple Huckel theory for the simple molecules with no 
substitituents. The values of El u m o  an^ E-/ which are readily 
available (121) or easily calculated for polycyclic aromatic 
hydrocarbons and provide a useful guide to the behavior of all PAHs 
in the ECD, with and without O2 doping. El u m o is defined as the 
calculated energy of the lowest unoccupied molecular orbital and L- 
is the calculated resonance energy lost by removal of a specific 
atom from the "fi'-electron system. In a simple molecule like
97
anthracene there are essentially only three positions for a single 
substituent. These positions with their corresponding L- values 
are:
These values indicate that probable nucleophilic attack by 0%- 
would occur at either the 9 or 10 position. In other words, the 
earlier computer predictions actually form the foundation of the 
proposed mechanism and the ions observed in the APIMS give 
additional support. Use of inductive and resonance parameters in 
the framework of Huckel theory satisfactorily accounts for the 
electron affinity for different substituents and positions of the 
substituents of several compounds. As an example, Wentworth sfc al«_ 
(122) utilized Huckel calculations to predict the difference in 
normal BCD responses for simple molecules as I- and 2- 
napthaldehyde. Grimsrud sfc al. (123) have used El u m q  and L- 
values to correlate normal BCD response and reactivity with O f t 
respectively. Their results correlate reasonably well with 
compounds such as anthracene, phenanthrene, triphenylene, and 
pyrene. It is postulated at this point that computer 
calculations (Huckel or extended Huckel) might be precise enough 
to predict differences in response with C^- as a function of isomer 
and substituent position.
2.238
2.018
98
The positive ions were also examined by APIMS and they 
consisted of MH+ for the chloroanthracenes and 9-chlorophenanthrene 
'compounds.
Polycyclic Aromatic Amines (124)
Many compounds of biomedical and environmental interest, 
particularly those containing polar functional groups, are 
thermally labile at the temperatures required for their separation 
and cannot be determined directly by gas chromatography. To 
permit these compounds to be separated by gas chromatography the 
technique of derivatization was developed. The classic reasons for 
derivative formation are: to improve the stability of the compound; 
to improve the chromatographic performance of the compound by 
reducing undesirable column interactions? and to change the 
separation properties of the compound by a purposeful adjustment of 
its volatility, thus eliminating peak overlaps. In addition, 
derivatization includes the use as a means of introducing a 
detector-oriented tag into a molecule and as the means for 
maintaining the chromatographic integrity pf the compound during 
its passage through the gas chromatographic column. The amines are 
derivatized not for the purpose of improving the thermal stability 
of the compound, but to increase its detector sensitivity. Figures 
•25 and .26 illustrate chromatograms of the under ivatized amines of 
anthracene and the derivatized amines, respectively. The thermal 
stability is not altered and the retention times are very much the 
same in the two chromatogram. Figure 27 shows the ECD response of
J Ui-- L
Figure 25.
Flame ionization detector response to underivatized amines of 
y / 1-7 and 2-amxnoanthracene.
1
iV
J---—
Figure 26. Flame ionization detector response to TFAA derivatives of 9-,
1-, and 2-aminanthracene.
1
0
0
1 0 1
Figure 27. Normal ECD response to 9-aminophenanthrene (TFAA) 
(top Chromatogram) and underivatized (bottom) 9- 
aminophenanthrene. Concentration of underivatized 
approximately 100 times derivatized compound. 
Detector is at 300°C.
102
the derivatized and underivatized amine of 9-phenanthrene. The 
concentration of the underivatized amine is approximately 100 times 
that of the derivatized amine. As a result, the derivatization 
procedure increases the detector sensitivity for the specific 
derivative.
Normal ECD Response of Derivatives
Martin and Rowland (125) have compared the response of the 
electron capture detector to various derivatives of 
pharmaceutically important amines. Clarke st si. (126) have 
assumed that electron attachment occurs at the carbonyl group of 
the amide but is destabilized by a contribution to the resonance 
form by the loan pair of electrons in nitrogen. This contention is 
further supported by the increased sensitivity of aromatic amines 
in which the electron density on the amino nitrogen is lowered by 
the electron-withdrawing ability of the benzene ring. Clarke also 
stated that the CFg group by itself does not appear to hold any 
great advantage for electron-capture. Its power lies in enhancing 
the polarity of an adjacent carbonyl group.
Temperature Dependence
The effect of detector temperature on detector response to a 
particular compound is of considerable analytical significance.
The response as a function of detector temperature gives an 
insight into the mechanisms involved in electron capture.
Figure 28 illustrates the response for the TFAA derivative of
4-aminophenanthrene and the PFPA derivative of 3-aminophenanthrene
Pe
ak
 A
re
a(
mm
)
I
103
TFPA
PFPA
Detector Temperature (°C )
Figure 28. Normal BCD responses as a function of detector
temperature for 3-aminophenanthrene (PFPA) and 4- 
am inophenanthrene (TFAA).
104
for detector temperatures of 250°, 300°; and 350°C. As one can 
see, for the 4-aminophenanthrene (TFAA), the response decreases as 
the temperature decreases over the designated temperature range.
In contrast, the response for the PFPA derivative of 3- 
aminophenanthrene has its highest response at the highest detector 
temperature. This is indicative of a dissociative type of 
mechanism. In other words, this type of derivative follows an 
electron capture response similar to Reaction I. This is similar 
to results obtained by Zlatkis and coworkers (127,128) for the 
pentafluoropropionate derivatives of n-hexylamine and 
cyclohexylamine. In addition, this difference in response for the 
two derivatives might give an insight into the characteristics of 
the two derivatives discussed in more detail later.
Isomer Differentiation
In Figures 29 and 30, typical chromatograms from which the 
response enhancements for the TFAA and PFPA derivatives have been 
determined. The chromatogram on the left is obtained by normal EC 
detection at 3 OO0C detector temperature, and the one on the right 
is obtained with 3 ppt oxygen continuously present in the detector, 
again at 300°C, The magnitude of the enhancement is again 
determined by the ratio of the peak heights, with and without added 
oxygen. Table 5 lists the response enhancements determined in this 
way for all of the aromatic amines studied where either the 
trifluoroacetic anhydride(TFAA) derivative, the perfluoropropionic 
anhydride (PFPA) derivative, or both were examined. Table 5 also
105
Figure 29. Capillary column gas chromatograms of the TFAA 
derivative of 2-aminofluorene. Tfte chromatogram on 
the left is with normal EC detection and the one on 
the right is with 0.30% oxygen in the makeup gas. 
Detector temperature is 300°C
106
Figure 30. Capillary column gas chromatograms of the FFPA 
derivatives of 9-aminophenanthrene, 2- 
aminoanthracene, and 1-ami noanthracene. The top 
^roinatograns are with 0.30% oxygen in the makeup 
gas. The bottom chromatograms are with normal EC 
detection. Detector temperature is 300°C.
107
Table 5. Oxygen Induced EC Response Enhancements for the
Trifluoroacetic Anhydride (TFAA) and Perfluoropropionic 
Anhydride (PFPA) Derivatives of Various Polycyclic Aromatic 
Amines a
Response Enhancement
Amine TFAA PFPA
2-aminobiphenyl 106 1.3
3- 124
4- ' . 102 1.4
1-aminpnaphthalene 125 1.9
2- H O 1.7
2,7-diaminonaphthalene 2.4
IrS- 1.6
1,5- 1.0
1-aminoanthracene 5.5 2.1
2- 1.3 1.7
9- 10.2 2.7
1-aminophenanthrene 32 1.1
2- 62
3- 74 1.4
4- 40 1.3
9- 32 2.1
1-aminofluorene 56 1.6
2- 96 2.6
3- 96 1.2
108
TABLE 5 (cont'd.)
4-
9-
2,7-diaminof luorene
1- aminopyrene
2-  
4-
2- aminobenzo(c) 
phenanthrene
3-
71 1.4
1.7 1.1
9.8
2.8
4.5
2.9
6.4
5.3
4- 4.8
~ EC detector temperature is 300°C Oxygen concentration in the doped 
detector is 3.0 parts per thousand.
109
I
3 2
4 m
2 3
5 4 5 10 4
4
Figure 31. Structure and numbering scheme for biphenyl (I), 
naphthalene (2), anthracene (3), phenanthrene (4), 
fluorene (5), benz(aJanthracene (6), chrysene (7), 
pyrene (8), and benzo(cJphenanthrene (9).
no
lists the enhancements for the TFAA derivatives of several diamino 
compounds. The results listed in Table 5 are obtained from the 
average of at least three determinations. The structure and 
numbering scheme for each group are shown in Figure 31.
In considering the data listed in Table 5, a prominent feature 
immediately apparent is that the response enhancements of the FFPA 
derivatives are consistently much smaller than those of the TFAA 
derivatives. Laterf Leef and Wilson (116) found that the normal 
ECD response to a PFPA derivative was nearly an order of magnitude
i
greater than that of a TFAA derivative for a given PAA. Our 
observations of relative EC sensitivity concur with this earlier 
report and this undoubtedly causes the large differences in RE 
values between the two derivative groups. That is, the oxygen 
enhancements of the PFPA derivatives are smaller because their 
normal electron capture rates are greater. This leaves less 
opportunity for response enhancements by oxygen addition. While 
the PFPA results show some variation among isomer sets, it appears 
that the TFAA derivatives are superior for the purpose of isomer 
distinction and each specific isomer group will be discussed in 
more detail below.
In considering the TFAA derivatives of the PAAs listed in 
Table 5, it is evident the RE differences among members of isomeric 
sets frequently exceed the +6% estimated standard deviation 
associated with the method. This value was determined by ten 
repetitive analyses for one of the isomers with and without the 
presence of oxygen in the detector. Considering each group in the
Ill
order listed, the enhancement value for the 2-biphenyl isomer is 
106+7, for 3-biphenyl the value is 124+9, and the 4-biphenyl is 
102+7. From enhancement data, one could distinguish the 3-biphenyl 
isomer from the other two. For the two monosubstituted naphthalene 
isomers, the enhancements, 125+8 and 110+7, differ somewhat more 
than the uncertainty of individual measurements. The three 
anthracene isomers are easily distinguished with the enhancements, 
1.3+0.1, 5.5+0.3, and 10.2+0.6. For the five phenanthrene TFAA 
derivatives, only the I- and 9-isomers do not have clearly unique 
enhancement values. The 3-aminophenanthrene isomer, one of the 
most mutagenic tested by Later jet al. (129) can easily be 
distinguished from the 1-, 4-, and 9-isomers of phenanthrene and is 
barely distinguishable from the two isomer. Among the five 
aminofluorenes, all except the 2-and 3-derivatives can be be' 
differentiated. It is interesting to note that the lowest 
enhancement is obtained with the 9-isomer of this group. Of the 
aminopyrene TFAA derivatives, perhaps only the 2-isomer is 
distinguishable from the other two. The isomers of 
benzo(c)phenanthrene also indicate barely distinguishable response 
enhancement values. It appears that as the size of the fused-ring 
structure is increased, the RE values become smaller and less 
isomer dependent. This is probably because the normal response is 
then too large and little increase in response is possible with 
oxygen addition. For the same reason the response enhancement
112
values for the diamino compounds are expected to be considerably 
less than those obtained for the monoamine derivatives . 
Nevertheless, in the case of the diaminonaphthalenes, the 2,7- 
isomer can easily be discerned from the 1,8- and the 1,5-isomers. 
For the cases where the RE values are low due to a normal EC 
response which is made too large by derivatization, the use of 
another, less EC sensitive chemical tag may be warranted.
Mixture Analysis
A  readily apparent and powerful application of oxygen-induced 
response enhancements will be to support the assignment of 
individual peaks in gas chromatography with EC detection.
Moreover, the technique should also provide assistance in the 
clarification of complex chromatograms where peaks are incompletely 
resolved (120). For example in Figure 32, two of the three 
anthracene TFAA derivatives are only partially separated by a 
temperature programmed capillary column. Knowing the enhancement 
expected for each isomer, one can quickly associate each isomer 
with its contribution to the partially resolved doublet by 
reanalyzing the sample with oxygen in the detector make-up gas.
The partially resolved peaks are almost twice as much as the second 
component. These paired chromatograms establish that the first 
peak of interest is the 9-isomer and the second peak is the 1- 
isomer. By retention time alone, this assignment would have been 
difficult with this chromatographic system. Another more
113
X  I
Figure 32. Capillary column gas chromatograms of a three-
component mixture of 9-aminoanthracene (TEAA), 1- 
Miinoanthracene (TFAA)f and 2-aminoanthracene (TFAA). 
Ine chromatogram on the left is with normal EC 
detection and the one on the right is with 0.30%
S o P c 1 10 ^  makeUp 9aS" Dstector temperaure is
114
challenging problem occurs in the case of exactly coeluting peaks. 
For example. Figure 33 shows chromatograms of a sample containing 
equal quantities of the T F M  derivatives of 9-phenanthrene and I- 
anthracene for which no separation is provided by the column. The 
response enhancement observed for the single peak of interest which 
contains both compounds is 19.3. The composition of this mixed 
peak can be obtained by the. following relationship:
REm j x = -^l^'l XgREg (28)
where REM j x is the response enhancement of the mixture, REj and REg 
are those of the pure isomers, and Xj and Xg are the mole fractions 
of the two isomers, where
Xj + Xg = I (29)
For this case, REj and REg are 5.5 and 32, respectively. The 
measured value of REm IX thereby indicates that Xj = 0.48 and Xg = 
0.52, which agrees very well with the known composition of the 
mixture (Xj = 0.50 and Xg = 0.50). In addition, linearity studies 
have shown that both the normal ECD response as well as the oxygen- 
induced responses are linear over a concentration range of 100. In 
other words, the response enhancement as a function of 
concentration is fairly linear over this range for the T F M  and 
PFPA derivatives examined.
115
Figure 33. Capillary column gas chromatograms of the two-
component equimolar mixture of the T F M  derivatives of 
1-aminoanthracene and 9-aminophenanthrene. The arrows 
indicate the point of elution of the compounds. The 
chromatogram on the left is with normal BC detection 
and the one on the right is with 0.30% oxygen in the 
makeup gas. Detector temperature is 300°C
116
Atmospheric Pressure Ionization Mass Spectrometry (APiMS)
In order to g a m  further insight into the chemical basis of 
the measurements of response enhancements with the TFAA and PFPA 
derivatives, we have identified the negative ions formed in the 
detector by the use of an atmospheric pressure ionization mass 
spectrometer (APIMS) where the ion source is, in fact, an ECD (92). 
The salient feature of the API mass spectrum of every PAA 
derivatized with TFAA was that the only negative ion observed with 
or without added oxygen is the parent M“ ion. Additional APIMS 
results for one set of isomers are shown in Table 6. This table 
indicates the relative M- ion intensities observed for the TFAA 
derivatives of 1-aminoanthracene, 2-aminoanthracene, 9- 
aminoanthracene, and 9-aminoanthracene in a series of single-ion 
mass chromatograms under the identical oxygen-free and oxygen-doped 
conditions. It should be noted that the temperature of the EG 
detector was 300° C and that of the APIMS source was 250° C. If 
these measurements are compared with! the corresponding RE 
measurements indicated in Table 6, it is seen that the relative 
increase in ion intensities caused by addition of oxygen in the 
APlMS experiment correlate quite well with the RE determinations by 
the ECD for this set of isomers. For example, of the four 
compounds the APIMS response to the 9-aminophenanthrene TFAA 
derivative increases most strongly with added oxygen (X9) and the 
measured RE value for this isomer is also the largest (X32). At the 
other extreme, no increase in the M- intensity for the
117
Table 6. Relative Parent Negative Ion Intensities by APIMS for the 
TFAA Derivatives of Several Polycyclic Aromatic Amines (a
Amine HQ JQ2 .with Q2 (k)
1-aminoanthracene 1.0 . 2.5
2-aminoanthracene 1.0 1.0
9-aminoanthracene 1.0 4.0
9-aminophenanthrene 1.0 9.0
(a) Analyte concentrations were adjusted to produce similar negative
ion intensities without adding O2.
(b) Oxygen, concentration is 3.0 parts pen thousand.
Source temperature is 300°C
2-aminoanthracene derivative was caused by oxygen and the RE value 
for this compound was only 1.3. In all cases, it is noted that 
oxygen causes a greater ECD enhancement than API mass spectral 
enhancement. The larger EC enhancement is expected and is 
attributed to the mode.of signal processing with the constant- 
current, frequency-modulated ECD as has been previously described 
(98,130).
In view of the preliminary results discussed above of the 
oxygen-doped ECD and APIMS studies of the TFAA derivatives of 1-, 2-, 
9-aminoanthracene and 9-aminophenanthrene, the study was extended 
to include the other isomers aminoanthracene and aminophenanthrene 
were studied in much tne same way with 0.20% oxygen in both the 
ECD and API at the same temperature of 250° C (131). The structures 
of the eight compounds examined are illustrated in the following:
;
118
anthracene phenanthrene
substituent
1
The ECD responses to each of these were obtained and were similar to 
that illustrated in Figure 29. Chromatograms such as these were 
recorded without and tiien with 0.20% oxygen in the detector at a 
temperature of 250° C. The ratio of ECD responses with and without 
oxygen to each P M  derivative was thereby determined for at least 
three separate sets of measurements. The average values of these
oxygen-induced response enhancements (RE) are shown in Table 7. It 
is seen that a very wide range of RE values are obtained, from a low 
value of 1.0 (no enhancement at all) for 2-aminoanthracene (TFM) to 
an enhancement of 107 for 3-aminophenanthrene (TFM). Furthermore, 
the RE values measured for the other members of this isomeric group 
are spread almost evenly throughout the range so that each isomer is 
uniquely indicated by its RE fingerprint.
Each P M  was then analyzed by GC/APIMS where conditions within 
its ion source were set nearly to those of the BCD. In Figure 34, 
for example, API mass spectral scans during the elution of the 9- 
aminophenanthrene derivative with and without added O2 are shown.
As in this case and the preliminary data discussed earlier, the 
striking feature of all mass spectra observed is that under these 
conditions, with or without added oxygen, the only negative ion
.119
Table 7. Oxygen-Induced ECD and APIMS Response Enhancements for the 
Trifluoroacetic Anhydride Derivatives of Aminoanthracenes 
and Aminophenanthrenes S
Amine BCD APIMS
2-aminoanthracene 1.0 I
1-am inoanth racene 3.1 2
9-aminoanthracene 8.1 4
1-am inophenanthr ene 19 9
4-aminophenanthrene 34 13
9-am inophenanthr ene 43 16
2-aminophenanthrene 72 20
3 -am inophenanth r ene 107 29
9 BCD and APIMS source temperature is 250°C Oxygen concentration in 
doped detector is 0.20%.
~ Reproducibility of ECD and APIMS enhancement measurements are 
approximately ±  7% and ±  15%, respectively.
observed for this entire set of molecules were due to the parent, M“, 
ion at m/e 289 and its . isotope peak at m/e 290. The isotope 
peak at m/e 290 is not resolved from the base peak at 289 in Figure 
34 because the response time of the pulse counting detector is too 
slow to follow the relatively fast mass scan rate used here. With 
slower scanning over the parent ion mass region, alone, these two 
peaks are resolved and exhibit the expected 5.7 to 1.0 ratio. No 
fragment or oxygen adduct ions were seen for these compounds in 
spite of their frequent appearance elsewhere in the atmospheric 
pressure ionization of other aromatic hydrocarbons (92,123). The 
API results suggest that the ionization of these molecules under 
oxygen free conditions occur by simple electron capture
1 2 0
w ith  O
w ithou t O
Figure 34 Negative ion API mass spectral scans taken during 
elution of TFAA derivative. of 9-aminoanthracene 
without and with .20% added oxygen at a source 
temperature of 2 SO0C
1 2 1
e + a  ->•
and with added oxygen occurs by charge
02- + a  -> A" + O2 (31)
to form the parent negative ion in both cases. Clearly, the mass 
spectra, themselves, provide no information from which isomer 
identity can be determined. It should be noted that the charge 
transfer reaction of 02“ with the analyte molecule , in this 
particular instance the TFAA derivatives, is in contrast to the 
actual oxygen incorporation observed with the chioroanthracene and 
chlorophenanthrene molecules. In addition, the temperature 
dependence of the TFAA derivatives indicated a resonance type 
mechanism for electron capture. This is supported by the fact that 
in the APIMS, the only ion observed without oxygen is the M~ ion.
Ey performing paired single ion monitoring experiments such as 
that shown in Figure 35, the enhancement of the APIMS parent 
negative ion signal caused by oxygen was measured for each isomer 
studied. The magnitude of the APIMS enhancements (average of at 
least three determinations) is also shown in Table 7 alongside with 
the ECD data discussed previously. As with the ECD results, the 
magnitude of the APIMS enhancements of the parent negative ion 
intensities were also very dependent on the structural differences 
within this isomeric group. It is also evident in Table 7 that the 
relative order of the observed ECD and APIMS enhancements are the 
same. For both methods, the 2-aminoanthracene derivative shows the
A“ (30)
transfer.
122
X I
without 
jZ  ° 2
with
I
O 2
x 50
T I M E
Figure 35. Single-ion APIMS monitor of the negative ion intensity 
at m/e 289 for the 1-aminophenanthrene (TFAA) 
derivative without and with added oxygen (.20%) at a 
source temperature of 250°C
Ilowest enhancement, the 3-aminophenanthrene derivative shows the 
greatest enhancement, while the enhancements of the other six 
compounds are relatively evenly spread throughout the observed 
ranges. The fact that the oxygen-induced enhancements measured by 
the ECD are uniformly greater than those measured by the APIMS is 
expected. This is thought to be largely due to the nature of the 
frequency-based response of the constant-current ECD which causes 
the observed ECD response enhancement to exceed the actual ratio of 
rates in Reaction 30 and 32. This effect has been previously 
described in a detailed model of the 02-doped EGD. The fact that 
the order of reactivity of the compounds listed in Table 7 are the 
same for both methods provides additional support for that model.
The experiments reported here suggest that Reactions 30 and 31 
provide the basis of the responses observed in the atmospheric 
pressure ionization cells and that the only measurable 
characteristics of these reactions that provides an indication of 
compound identity is the ratio of the rate constants for Reactions 
30 and 31. This rate constant ratio appears to more easily and 
reproducibly measured by the ECD than by the APIMS. The 
reproducibility of the ECD and APIMS enhancement measurements are 
approximately ±7% and ±15%, respectively. These results further 
support our earlier suggestion that an analysis scheme which 
provides normal and oxygen-enhanced ECD responses to a 
chromatographic effluent provides a very simple, but powerful aid 
for the identification of the components of complex mixtures by gas 
chromatography.
123
124
Several EFPA derivatives were also examined with the APIMS 
system. For those examples, without oxygen the ions observed were 
M at m/e 339, an ion at m/e 319 which corresponds to (M-HF)- , and 
an ion at m/e 309 . In the presence of oxygen the same species were 
observed with the M- ion increasing slightly in intensity. Figure 
36 illustrates the API mass spectral scans during the elution of the 
PFPA derivative of 4-aminophenanthrene with and without added 
oxygen. These results are consistent with the results obtained of 
the detector response as a function of detector temperature. Those 
results suggest a dissociative type mechanism for electron capture, 
and the ions observed without oxygen in the APIMS support this 
evidence.
Mass JSRectfal Analysis s£ TFAA and PFPA Derivativps (132)
The TFAA derivatives of the isomers of aminoanthracene and 
aminophenanthrene and several PFPA derivatives of the same group 
were examined with electron impact (EI) and both negative and 
positive chemical ionization (NCI and PCI) mass spectrometry with 
isobutane and methane as reagent gases to disclose the 
characteristics and evaluate the possibility of isomer distinction.
The EI spectra of the TFAA derivatives of 1-aminoanthracene and 
4-aminophenanthrene are shown in Figures 37 and 38, respectively and 
the relative intensities of the common ions observed in the EI 
spectra for all eight isomers are listed in Table 8. Figures 39 and 
40 illustrate the isobutane PCI for 1-aminoanthracene and 4- 
aminophenanthrene TFAA derivatives. The relative intensities of the
125
O 2
X1
X1
r
300 400
M/E
Figure 36. Negative ion API mass spectral scans taken during 
elution of derivative of 3-aminophenanthrene (PFPA) 
with (right) and without (right) .20% added oxygen. 
Source temperature is 250°C.
Figure 37. Electron impact mass spectrum of 1-aminoanthracene (TFAA).
126
100 J£S
Figure 38. Electron impact mass spectrum of 4-aminophenanthrene 
(TFM).
127
128
Table 8. Ion Intensities from the Electron Impact Mass Spectra of the 
TFM Derivatives of the .Aminoanthracenes and Phenanthrene. -
M/E (% Abundance)
Amine 231 220 i n 203 192 121 120 155
IA 100 6 18 0 14 4 11 55
2A 100 4 21 0 2 0 9 55
9A 100 0 32 0 87 11 22 46
IP 100 8 15 0 23 3 . io 48
2P 100 3 11 0 20 0 7 39
3P 100 0 11 0 24 3 7 . 41
4P 100 30 12 33 23 42 19 0
9P 100 5 10 0 6 3 8 77
common ions observed are shown in Table 9. The isobutane NCI for
the same two derivatives are illustrated in Figures 41 and 42, and 
the intensities for all eight isomers are listed in Table 10.
Figures 43 and 44 illustrate the methane PCI for 1-aminoanthracene 
and 4-aminophenanthrene, respectively with the intensities ligted in 
Table 11. The methane NCI for the same two compounds is shown in 
Figures 45 and 46 and the relative intensities for all eight isomers 
are listed in Table 12.
The EI spectra for the I- and 2-aminoanthracene (TFM) and the 2- 
and 3-aminophenanthrenes derivatives are quite similar. The 
intensities of the parent m/e 289 and the fragment ions of m/e 192 
and 165 are almost equal in intensity. The fragmentation pattern
10011
Figure 39. isobutane positive chemical ionization mass spectrum of 1-
aminoanthracene (TFAA).
129
100] 2 1 2
80.
80. 
XI  . 
NO.
20.
IOO r  T
Figure 40. isobutane positive chemical ionization mass spectrum of 4-
amino^ienanthrene (TFAA).
130
131
giving rise to these ions and others is shown in the following 
scheme:
165 C A  2 2 0
The ion intensities of m/e 192 for the 9-aminoanthracene derivative, 
m/e 202 for 4-aminophenanthrene and m/e 165 for the 9- 
aminophenanthrene would distinguish those isomers from the others.
Table 9. Ion Intensities from the PCI (isobutane) Mass Spectra of
the TFAA Derivatives of the Isomers of Aminoanthracene and 
Phenanthrene
M/E (% Abundance)
Amine 292 291 220 299
IA 22 23 100 21
2A 12 15 100 21
9A 4 18 100. 20
IP 2 18 100 15
2P 2 17 100 16
3P 6 17 100 17
4P 2 21 100 20
9P 3 17 100 20
222
0
0
0
0
0
0
20
0
I 001 _20S
BtL
Bd 
XI  . 
Md
26
2d
2 i
d I 7
127
I . I . 1
isa -Tx Tio
Figure 41. isobutane negative chemical ionization mass spectrum of 1-
aminoanthracene (TFAA).
132
100
112
263
2YiV
250
I
311
'-C'
Figure 42. Isobutane negative chemical ionization mass spectrum of 4-
aminophenantarene (TFAA).
133
134
Table 10. Ion Intensities from the NCI (isobutane) Mass Sepctra of the 
TFAA Derivatives of the Isomers of Aminoanthracene and 
Phenanthrene
M/E (% Abundance)
Amine 299 288 231 271 252. 219 203 112
IA 100 8 I I 3 3 2 2
2A 100 20 5 0 17 4 20 O
BA 100 22 6 I 8 22 2 50
IP 20 52 48 4 100 0 5 22
2P 13 53 24 3 100 I 2 8
3P 8 46 63 5 100 0 3 14
4P 25 84 6 100 35 0 0 4
9P 21 63 54 0 100 0 5 38
Table 11. Ion Intensities from the PCI (methane) Mass Spectra of the 
TFAA Derivatives of the Isomers of Aminoanthracene and 
Phenanthrene
M/E (% Abundance)
Amine 292 291 220 289 272 220 219 124
IA 20 23 100 42 I 7 8 5
2A 25 24 100 3B 0 20 8 21
BA 5 20 100 38 0 30 24 18
IP 5 23 100 18 5 13 8 13
2P 2 12 100 23 0 10 10 25
3P 2 14 100 23 4 14 13 35
4P I 21 100 32 44 21 . 0 2
BP 2 15 100 18 I 15 8 15.
IOOTT
80.
290
Bd 
I l . 
Md
2d
4 00 4-i-L J
Figure 43. Methane positive chemical ionization mass spectrum of 1-
aminoanthracene (TFAA).
135
Figure 44. Methane positive chemical 
aminophenanthrene (TFAA).
ionization mass spectrum of 4-
136
i oon
Figure 45. Methane negative chemical ionization mass spectrum of 1- 
aminoanthracene (TFAA).
137
271
Figure 46. Methane negative chemical ionization mass spectrum of 4-
aminophenantarene (TFAA).
138
139
Table 12. Ion Intensities from the NCI (methane) Mass Sepctra of the 
TFAA Derivatives of the Isomers of Aminoanthracene and 
Phenanthrene
M/E (% Abundance)
Amine 289 288 287 273 272 211 269 219 m m
IA 100 12 40 16 ' 0 2 2 85 8 8
2A 100 15 21 23 0 2 15 34 55 9
9A 6 7 38 3 0 I I 100 2 6
IP 8 12 8 100 0 4 24 0 9 13
2P 15 13 10 100 0 8 40 I 4 9
3P 9 8 5 100 0 3 25 0 6 8
4P 6 7 4 20 20 100 13 0 I 4
9P 15 27 5 100 0 4 58 I 7 2
The electron impact mass spectra allows differentiation of only 
three of the eight isomers.
The positive ion isobutane CIMS for all eight isomers are all 
quite similar. The major ions formed include (M+l) , (M+2), and 
. (M+3). Howverf the 4-aminophenanthrene TFAA isomer can be easily 
differentiated from the others by a fragment ion in its PCI spectrum 
at m/e 272. The gas phase ion that gives rise to this fragment is 
unique to the 4-isomer because of its substitution position, termed 
the bay region of the molecule. B/E linked scans were done on m/e 
290 and the fragment ions observed were m/e 272, 220, and 202. The 
spectrum is shown in Figure 47. Furthermore, B/E linked scans were 
done on m/e 272 ,and the major fragment ion observed was . m/e 202 and 
the spectrum is illustrated in Figure 48. There are probably
140
several mechanisms that could account for the ions observed in the 
spectrum for 4-aminophenanthrene (TFAA) and one of them is shown as 
follows:
m/e 289 m/e 290 m/e 272
290 272 220 202
M/E
Figure 47. B/E scan on m/e 290 of 4-aminophenanthrene (TFAA).
141
M/E
Figure 48. B/E scan on m/e 272 of 4-aminophenanthrene (TFAA).
-142
m/e.- 272
Following the initial protonation of the molecule at the carbonyl 
oxygen, there is a loss of H2O which leads to the formation of a 
six-membered ring bridging at the carbonyl carbon atom of the T F M  , 
side chain. Subsequent loss of HCF3 gives rise to the observed 
fragment ion of m/e 202. Alternatively, the proposed structure of 
m/e 290 could lose the fragment HCF3 to form the ion m/e 220 and 
then lose the water molecule to again form the cyclized structure of 
m/e 202. Another indication of the formation of the cyclized 
structure of m/e 272 was obtained from the results of B/E linked 
scans with and without collisional activation with helium gas. As 
shown in Figure 49, m/e 272 seems to be a relatively stable ion.
The CA spectrum shows loss of HF (m/e 252) and HCF3 (m/e 202). If the 
ion m/e 272 did not have the proposed cyclic structure, it might be 
expected that more daughter ions would result in the CA spectrum of 
this fragmentation. It should be emphasized that the proposed 
mechanism for fragmentation of the 4-isomer may be one of many 
possible to explain the observed fragmentation pattern.
143
272 252
M/E
I I I I I I I 
200
Figure 49. B/E scan on m/e 272 of 4-aminophenanthrene (TFM) 
with He as the CAD gas.
144
The formation of the cyclic structure similar to m/e 272 is also 
found in solution chemistry. The following cyclization of m- 
methoxy- -phenylamine. occurs quite easily (133).
OMe R
As a similar reaction, the origin of m/e 202 could be studied by 
preparing and isolating the proposed structure of m/e 272 by the 
following reaction.
This reaction is actually a modification of the Pictet-Gans 
synthesis (134,135). This compound could then be used as the 
starting point for further EI and Cl studies. This synthesis has 
not been attempted in our lab because of the lack of adequate sample 
and in view of the carcinogenic and mutagenic properties of these 
derivatives. Our laboratory is simply not equipped for this type of 
procedure.
The positive ion methane CIMS indicates much more fragmentation 
and could be utilized to identify 9-aminoanthracene and 4- 
aminophenanthrene. Again, the propensity of the TFAA derivative of 
4-aminophenanthrene to form the fragment ion. at m/e 272 clearly 
differentiates that particular isomer. The TFAA derivative of
145
9-aminoanthracene can also be distinguished by the ions at m/e 219 and 
179. The ion at m/e 179 corresponds simply to the protonated 
anthracene ion.
The negative ion Cl (isobutane) spectra for the 1-, 2 -, and 9- 
aminoanthracene (TFM) derivatives are very similar, but based on the 
intensity of the fragment ion at m/e 112 which corresponds to the 
actual charge remaining with the amide substituent, the 9- 
aminoanthracene can be distinguished from the other two. The 4- 
aminophenanthrene derivative is easily differentiated from the 
others based on the fragment ion at m/e , 271. This structure may be 
similar to the four-fused ring structure postulated for the PCI at 
m/e 272. The NCI (methane) seems to be more informative, (fhe 9- 
aminoanthracene can easily be distinguished from the others, due to 
the formation of the ion at m/e 219 which corresponds to loss of 
HCFg from the parent ion. The 4-aminophenanthrene can easily be 
differentiated by the fragment ion at m/e 271. Also, the 9- 
aminophenanthrene derivative can be identified, primarily based on 
the intensities of m/e 112 and 269.
The results obtained by negative ion Cl in methane and isobutane 
appear to be more promising than those with either EI or PCI. The 
negative ion Cl of these compounds indicate that the relative 
intensities of the negative fragment ions are dependent on the 
concentration of the analyte. Apparently reactions in addition to 
simple electron attachment are reasonable for the total negative ion 
Cl spectra observed. An example of typical concentration dependence 
observed in the NCI of several of the derivatives is illustrated in
146
Figure 50. This phenomenon may be due in part to ion source 
chemistry or possible reaction with other reagent ions present in 
the methane or isobutane reagent gases. We have observed ions 
including C CH-, CH2 , CHg , O and GH- when either methane or 
isobutane were used as the reagent gases. Other researchers have 
found similar reagent ions (136) in the negative ion CIMS when 
methane was used as the reagent gas. The concentration dependence 
is presently under further study.
In summary, negative ion Cl shows considerable promise for the 
differentiation of isomeric PAAs but appears to be complicated by 
presently unknown ionization processes. Positive ion EI and Cl 
provide only partial differentiation of the isomers.
The El spectrum for the PFPA derivatives of 3-ajninophenanthrene 
and 4-aminophenanthrene are shown in Figures 51 and 52, 
respectively. The isobutane and methane PCI of 4- 
aminophenanthrene (PFPA) are shown in Figures 53 and 54. The PFPA 
derivative also forms the (M + I), ( M + 2), and ( M + 3), similar 
to the TFAA derivatives. However, the PFPA derivative does not form 
the (M + I) species as the major ion in the positive ion Cl mass 
spectrum as does the TFAA compounds. In addition, there is evidence 
of the postulated ring structure m/e 202 (similar, tp that observed 
for the TFAA derivative). The peak at m/e = 202 might be formed as 
a result of the following proposed reaction scheme:
147
2 : 3 21 : 1 0l: 2
Figure 50. Single-ion traces of m/e 289 and 112 for 9-
aminoanthracene (TFAA) with isobutane negative 
chemical ionization mass spectrometry.
:  ' I
Figure 51. Electron impact mass spectrum of 3-aminophenanthrene (PFPA) .
148
I OO 339
8 0
298
212 270
289
•4— 4
Figure 52. Electron impact mass spectrum of 4-aminophenanthrene (PFPA).
149
Figure 53. isobutane positive chemical ionization mass spectrum
of 4-aminophenanthrene (PFPA).
150
Figure 54. Methane positive chemical ionization mass spectrum of 4-
aminophenanthrene (PFPA).
151
152
m/e 339 m/e 340
m/e 220
m/e 202
It would be interesting to do B/E scans on m/e = 220, 339, and 340 
to see the formation of the ion at m/e = 202. In addition, it would 
be interesting to react the derivative with PCI5, similar to that 
discussed for the TFAA derivative, analyze the mixture by mass 
spectrometry to see if the cyclized product can be made in solution.
The negative Cl of the 4-aminophenanthrene (PFPA) derivative 
indicates fragment ions at m/e 201 and 219, probably indicating the 
formation of the proposed ring structure discussed previously. The 
negative ion mass spectra (both methane and isobutane) show many 
other fragment ions of lower intensity that have not been 
identified.
Polycyclic Aromatic Hydroxides (124)
The hydroxy compounds were derivatized for the reasons described 
for the amine compounds as well as for increasing the thermal 
stabililty of the parent compound.
153
Isomer Differentiation
Figure 55 indicates an example of the chromatograms from 
which the response enhancements for the hydroxy compounds were 
determined. For the methoxy derivatives listed in Table 13, the RE 
differences among members of isomer sets in most instances again 
exceed the ±6% level for estimated uncertainty. The I- and 2- 
methoxynaphthalene isomers can. be distinguished with RE values of 
81 and 62, respectively. The dimethoxy derivatives of naphthalene 
are also easily discerned with enhancement values ranging from 24 
to 105. The only two methoxyphenanthrenes available for study here 
show great variation of their enhancements, as do the three isomers 
of dimethoxyphenanthrene. While the RE values of 6-methoxy-12- 
methylbenz (a)anthracene and 5-methoxy-7-methylbenz(a)anthracene are 
too similar to be distinguished, the other members of this isomer 
group can be differentiated. In the case of the six 
methoxychrysene isomers, an almost continuous range of enhancements 
from 3.6 to 11.1 are observed. The enhancement value of 9-amino-3- 
methoxyphenanthrene (TFAA) also included in Table 13 is 31. This 
compound was prepared by using the two derivatization methods 
sequentially with methylation of the hydroxy functional group 
performed first. This result enables one to stepwise derivatize 
groups on a single compound to make them more EC responsive and 
oxygen sensitive.
154
Figure 55. Capillary column gas chromatograms of 4-hydroxy-
phenanthrene (methoxy derivative). The chromatogram 
on the left is without oxygen and the one on the right 
is with 0.30% oxygen in the makeup gas. The detector 
temperature is 300°C.
155
Table 13. Oxygen Induced Response Enhancements for the Methoxy 
Derivatives for Several Polycyclic Aromatic Hydroxides
Response
Hydroxide Enhancement
1- hydroxynaphthalene 81
2- 62
1.4- dihydroxynaphthalene 24
1.5- 74
1,7- 105
4-hydroxyphenanthrene 11.5
9- 80
4.5- dihydroxyphenanthrene 13.9
1,9- 3.0
2,7- 63
1- hydroxychrysene
2-
3-
4-
5-
6-
3.6 
8.0 
9.2 
3; 8 
6.8 
11.1
12-hydroxy-7-methyl 
benz (a) anthracene 
6-hydroxy-12-methyl 
benz (a)anthracene 
5-hydroxy-7-methyl
16.0
22
156
TABLE 13 (cant'd.)
benz (a)anthracene 20
8-hydroxy-7-methyl ,
benz(a)anthracene 11.1
9-amino-3-hydroxy 
phenanthrene - 31
a EC detector temperature is 300°C. Oxygen concentration in the doped 
detector is 3.0 parts per thousand.
b Methoxyr TFAA derivative.
Atmospheric Pressure Ionization Mass Spectrometry
The only ions observed by APIMS for the methoxy derivatives of 
2-hydroxy naphthalene and 9-hydroxyphenanthrene with and without 
oxygen was the (M-15)- ion. The dimethoxy compounds of 1,7- 
dimethoxynaphthalene and 2,7-dimethoxyphenanthrene were also 
examined and ions observed included both the M” and the (M-15)"" ion. 
Both of these were formed with and without the presence of added 
oxygen. Without oxygen the M“ ion is the predominant ion and with 
oxygen the (M-15)"" is the predominant ionic species.
Methylanthracenes and Methylphenanthrenes 
Isomer Differentiation
The methyl isomers of anthracene and phenanthrene present 
another interesting set of isomeric compounds in which typical forms 
mass spectrometry, fail to discern them. The electron impact mass 
spectra (EI) for 2-methylanthracene, 2-methylphenanthrene,
1001—
1/6 I
Figure 56. Electron impact mass spectrum of 2-methy!anthracene.
157
ion
8 0.
6 0,
X I
MO.
191
9'
4— K-
I 65
I 8  Cl '
Figure 57. Electron impact mass spectrum of 2-methyphenanthrene.
158
100 qia.
6 0_ 
11 
'10
9 6
9 i
Jso' J 8 0
Figure 58. Electron impact mass spectrum of I-methylphenanthrene.
159
160
and 1-methylphenanthrene are shown in Figures 56, 57, and 58, 
respectively, and for all practical purposes, are almost identical.
The response enhancements for these compounds are listed in 
Table 14. The oxygen concentration was 0.2% with a detector 
temperature of 250° c and column temperature of 180%. The RE 
values for this isomeric group allow these compounds to be 
distinguished based solely on the oxygen—induced response. It 
should also be noted that the RE values were, determined using the
Response Enhancements of the Methylanthrancenes 
Methylphenanthrenes
Name R.E.
1-methy !anthracene 3.4
2- 5.7 '
9- 13.6
9,I0-dimethy!anthracene 10.8
1-methy lphenanthrene 14.1
2- 20.2
3- 1.0
4- 11.0
9- 3.5
161
series detector design that will be discussed in more detail later.
An example is shown in Figure 59. Again, the values range from 3.4 
for 1-methylanthracene, 13.6 for 9-methy!anthracene, and 20.2 for 2- 
methylphenanthrene. The only two isomers that can not be, 
differentiated are 9-methylphenanthrene (3.5) and 1- 
methylanthracene(3.4), because their RE values are too similar.
Atmospheric Pressure ionization Mass Spectrometry (APIMS)
In order to establish the ions formed with and without the 
presence of oxygen, the APIMS was again utilized. For the 9- 
methylanthracene isomer, the major ion formed with and without added 
oxygen was the (M-CH3-K)2)- species. The results are quite similar 
to that obtained for the 9-chl or ©anthracene molecule, Without 
oxygen for the 9-methy !anthracene, the ions formed were (M-K)2)" and 
(M-CH3+O2)"" in a ratio of about 1:4. With oxygen present, the major 
ion formed is the (M-CH3-HD3)-* For the 2-methylphenanthrene and 
without oxygen, the major ionic species present were M- and (M-K)2)- 
in a ratio of approximately 5:1. With the addition of oxygen, the 
major ion formed is the (M-K)2)" with a small amount «5%) of (M-KD)". 
The reaction mechanism for 9-methylanthracene is similar to that 
postulated for the 9-ehloroanthracene and 9-chlorophenanthrene .
M
V
(M-CH3+ O2V
162
Figure 59. Response enhancement for 9, 10-dimethy!anthracene
using series detectors. The chromatogram on the left 
is without oxygen and the one in th top right is with 
0.20% oxygen in the detector at 250°C
163
For 9-methylanthracene, the intermediate (MtO2)- is formed by 
reactions I and 3 or 2 and 4. This intermediate dissociates through 
reaction 6 to form (M-CHgtO2)- • In the case of 2—methylanthracene, 
the intermediate again is formed through reactions I and 3 or 2 and 
4, but remains intact. Another compound examined was the 9,10- 
dimethy!anthracene molecule, and the major ion formed both with and 
without oxygen was a species at m/e 209 which corresponds to a 
species of the type (M^CHgtO2)-, (M^CHgtHtO2)-, or possibly (M- 
2CHgt2Hto2)-. In view of the resolution problem or inability to 
differentiate mass differences of 1-2, it is difficult to 
distinguish the differences between these postulated species. The 
structure and exact mass of the ionic species could possibly be an 
oxygen bridged system similar to that observed with the 
photooxidation of anthracene on atmospheric particulate matter
(137). The following products have been identified
I
hv
. Particulate  
matter A
2 to 7 %
i?
O
12 to 19 %
B
8 to 13 %
2 to 5 %
164
The species formed for the 9,10-dimethylanthracene in the presence 
of oxygen in the APIMS might be similar to structure or product A 
or B. If one used (a) labelled oxygen the question might be 
resolved or (b) if species A could be isolated and then put in the 
source of the APIMS, one might be able to deduce the structure of 
the oxygenated species of this particular compound. Both of these 
experiments would be interesting to perform and be very informative 
at the same time. In addition, it might be informative to examine a 
molecule such as 9,10-dibromoanthracene with the APIMS and added 
oxygen to determine the species formed. If a ion of the form ( M -. 
ZBr + Og) is produced in the presence of oxygen, there should be no 
trouble in resolving that from the parent ion.
Dioxin
The structure for the dioxins is shown in the following 
illustration.
For the single dioxin compound examined, 2,3,7-trichlorodibenzo-p- 
dioxin, the response both with and without the presence of oxygen in 
the detector is shown in Figure 60. At a detector temperature of 
250° C, column temperature of 180° C, and approximately 2 ppfc oxygen 
in the detector, the response enhancement was determined to be 14.1. 
In addition. Miller sfc si. (105) determined the response enhancement
165
Figure 60. Capillary column gas chromatograms of 2,3,7-trichloro- 
dibenzo-p-dioxin. The chromatogram on the left is the 
normal BCD response and the one on the right is with 
0.20% oxygen in the detector at a temperature of 
250°C Sample contains approximately 2 ng of dioxin.
166
of dibenzofuran and found it to be 175. From the analysis of these 
two compounds some interesting implications might be made. As an 
example, in using the series detector scheme which will be discussed 
in the next section, the uncertainty in determining response 
enhancement values was determined to about 2% and could be applied 
to the dioxin isomers for possible differentiation. Utilizing a 
HPLC for prior cleanup and other forms of extraction schemes, it 
would be interesting to see if the oxygen-doped ECD could be used 
with actual environmental samples containing trace quantities of 
dioxins. It should be noted that precautionary procedures for 
handling and disposal of the dioxin waste should be strictly adhered 
to. Handling the samples with gloves and venting the ECD effluent 
out of the laboratory are highly recommended safety precautions.
Simultaneous Electron Capture and Oxygen-Sensitized Electron Capture 
Detection
The ECD is used extensively for the trace analysis of organic 
compounds and frequently provides the only or best means of sensing 
pollutants at their environmental concentrations. Hie oxygen 
sensitized ECD may add to this detector's proven worth by also 
offering a means for compound verification in the trace analysis of 
environmental samples. It has been previously shown that the method 
is greatly assisted by the large variations of RE values among 
various sets of related compounds, but the precision with which an 
individual RE measurement needs to be improved. A  decrease in the
167
uncertainty of individual RE measurements will correspondingly 
increase the discriminating power of the method.
Previously, all oxygen RE measurements have been made using a 
single ECD and a given chromatographic separation which was 
performed twice, once without oxygen and again with oxygen in the 
detector. With great care in reproducing manual, splitless 
injections to a capillary column, the relative standard deviation of 
previous RE measurements have typically been +7%. The precision ot 
this technique was determined by analyzing one of the T F M  
derivatives of the amines ten times both with and without oxygen.
For these measurements, the relative standard deviation of the 
measured response enhancements was +7%. Miller's reproducibility 
was estimated at 10% for primarily packed column chromatography.
Parallel ECDs (138) .
Dual Columns
Improved reproducibility as well as increased ease of
■
measurement might be expected if two ECDs are simultaneously used 
for each analysis where one detector is operated normally and the 
other is doped with oxygen. The first new technique using two 
parallel ECDs that will be discussed involves the use of two 
identical capillary columns placed in the same injector port with 
one column going to one ECD and the other one going to ECQ number 
two. In order to determine the split ratio which occurs in the 
injector port and to ensure that the split ratio is approximately 
1:1, the chromatograms shown in Figure 61, j were first obtained in
168
the normal, oxygen-free mode. The relative peak heights observed 
suggest that the split ratio to column A and B is 1.20. The 
possibility that the paired ECD responses may also differ due to 
differing sensitivities of the two detectors was tested by reversing 
the columns at the detectors. As shown in part b of Figure 61, an A 
to B response ratio of 0.82 is obtained. This result indicates that 
our two detectors have essentially identical sensitivities and the 
small difference in normal EC responses seen here, is entirely due 
to the split ratio of the sample at the detector and not due to any 
differences in the detector response. Also, it is noted that in 
this preliminary test that the retention times of 2-chloroanthracene 
in the two columns are essentially identical for all practical 
purposes.
In Figure 62, a single 2-chloroanthracene sample has- been 
analyzed ten times in repetition. While the peak heights vary 
noticeably from one injection to the next (due to the difficulty in 
reproducibly delivering the 0.4 yl sample by syringe), the ratio of 
the oxygen-sensitized to the normal responses remains very constant. 
The ratios of the ten dual responses shown are 6.2, 6.2, 6.2, 5.9, 
5.8, 6.3, 5.9, 6.0, 6.1, and 6.1. The relative standard deviation 
of these values is 2.6%. The average of these values, 6.1, times 
the split ratio, 1.20, indicates that the oxygen-induced RE value 
for 2-chloroanthracene is 7.3. This value correlates well with the 
value of 6.9 previously observed in a study using 0.3% oxygen and a 
300° C detector temperature, but by two separate, paired analyses 
with only one detector (119). In that study unusually great care
EC
D
 
(B
) 
EC
D
 (
A)
169
Figure 61.
t im e
Dual EOD responses with dual columns of sample 
containing 2 ng of 2-chloroanthrcene, repeated three 
times with both detectors operatd in the normal, 
oxygen-free mode. The last chromatogram is repeated 
with column-to-detector connections reversed. 
Detector temperatures are 300°C.
EC
D
 (
A)
 
O
5-
E
C
D
(B
)
170
I
A l
x l
i
L
t im e  ->
Figure 62. Dual ECD and oxygen-sensitized detection of sample 
containing 12 ng 2-chloroanthracene, repeated ten 
times. Detector temperature is 300°C and oxygen 
concentration in detector is .30%.
171
was taken to reproducibly inject samples and the uncertainty 
(relative standard deviation) was estimated to be +7%. The use of 
the two parallel detectors as described has reduced this 
uncertainty a factor of two and has greatly relaxed the requirement 
for precise injections.
In Figure 63 chromatograms are shown which indicate that the 
response ratios are quite constant over a range of different analyte 
concentrations. In these analyses, RE values of 6.9, 7.0, 7.0, and 
7.4 are observed as the analyte concentration is repeatedly doubled. 
In a previous study (119), the RE values of this compound and its 
two isomers were found to be independent of concentrations over a 
change of approximately two orders of magnitude.
Also, in Figure 64, the simultaneous dual ECD responses to all 
three isomers of chior©anthracene are shown. From these analyses, 
the injection split ratio is again 1.20, and RE values for 1-, 2-, 
and 9-chloroanthracene are determined to be 3.4, 7.3, . and 26.0.
These values are in reasonable agreement with those previously 
observed under similar conditions fcy the single detector method
(119). Since the mass spectra of the three chloroanthracenes are 
identical (See Figure 14), their ECD responses as shown in Figure 64 
would seem to offer a particularly useful means for their 
identification. The improved precision accompanying the use of the 
two parallel ECDs is expected to proportionately increase the 
reliablity of the method. In addition, another advantage of this 
method involves the time required for analyses. Using the 
previously described method, one analyzed the samples without
EC
D 
(A
) 
O0
-E
C
D
 (
B)
172
Figure 63. Dual BCD and O9-Sensitized BCD detection of four samples 
containing 1.5, 3.0, 6.0, and 12 ng 2-chloroanthracene.
173
t im e
Figure 64. Dual ECD and oxygen-sensitized EC detection of samples 
containing 5 ng of 1-chloroanthracene (a), 12 ng of 2- 
chloroanthracene (b), and 6 ng 9-chloroanthracene (c).
174
oxygen, added oxygen to the detector and waited approximately 2-3 
hours for equilibration, and then analyzed the samples again in the 
presence of oxygen. Using this technique with two parallel 
detectors the analysis time is reduced by a factor of two.
Splitter
The response enhancements have been examined where the split was 
performed with a post-column effluent splitter. A  Varian capillary 
column effluent splitter was utilized. As far as the detector 
performance is concerned, this scheme functions as well as y/hen the 
split occurs in the injection port. Chromatograms are shown in 
Figure 65 which clearly indicates the only disadvantage of the 
effluent splitter is some deterioration of the chromatographic 
resolution. In addition, 9-chloroanthracene was analyzed ten times 
in repetition. The peak heights vary noticeably from one injection 
to the next, but again the reproducibility is still quite good. The 
relative standard deviation of these measurements is 5.4%, which is 
still an improvement over the old method of using single detector. 
The response enhancement as determined for the 9-chloroanthracene 
was . 3 4, which is slightly higher than previously determined (119), 
but comparable with Miller's results (105). The response 
enhancements for the other isomers agree quite well with previous 
studies. The values obtained are 3.8 and 7.1 for the 1- 
chloroanthracene and 2-chloroanthracene molecules, respectively.
The major factor contributing to the poor resolution is the effluent 
splitter itself. The major problem associated with this splitter is
175
Figure 65. Dual EXD responses with a capillary column effluent 
splitter of sample containing 8 ng of 9- 
chloroanth racene. The chromatogram on the left is the 
normal EXD responses and the one on the right is with 
0.30% oxygen in the second cell. Both detectors are 
at 300°C.
Ithe fact that it has stainless steel tubing in places(this was not 
known at the time of the experiment). It has been shown that for 
some molecules that stainless steel will absorb some quantity of 
sample. A splitter constructed as to reduce the dead volume 
involved, as well as removing the stainless steel, would be very 
beneficial and probably solve the problem in resolution. An 
efficient fused silica capillary column effluent splitter has been 
reported by Later, Wright, and Lee (138). A schematic diagram of 
this splitter is illustrated in Figure 66, and would probably
improve the peak broadening problems as experienced with the Varian 
splitter.
The use of either dual columns or a splitter could be used in a 
parallel combination of the O2^oped BCD and a FID or nitrogen or 
phosphorous specific detectors. Other combinations of detectors 
could be possible with this type of arrangement.
Series Detectors (138)
In this experiment, the effluent from one undioped BCD is . 
connected to the inlet part of the second detector with a section of 
glass-lined stainless steel tubing(1.5ft). The glass-lined tubing 
and the BCD exit line was connected by a small piece, of Teflon 
tubing. Typical chromatograms indicating the difference in detector 
response is shown in Figure 67 for the chi or ©anthracenes and 9- 
chlorophenanthrene. It is interesting to note that the response in 
the two detectors is similar for all the compounds except 9- 
chlorophenanthrene. Figure 68 illustrates the enhancements obtained
176
177
B
Figure 66. Schematic diagram of the fused silica capillary column 
effluent splitter. A: fused silica capillary qolumn; 
B: AgCl cement; C: glass sleeve. (Reference 139).
178
Dual BCD responses in series detection of 1-, 2-, 9- 
chlor©anthracene and 9-chlorophenanthrene. Detector 
temperatures are 3000C.
Figure 67.
179
Figure 68. Dual ECD responses and oxygen-sensitized ECDs in
series detection of 1-, 2-, and 9-chlor©anthracene and 
9-chlorophenanthrene. The bottom chromatograms are 
normal ECD responses and the upper ones are with 0.30% 
oxygen present in the second detector. Both detector 
temperatures are 300°C
180
with oxygen present in the second detector. The enhancements 
obtained are 2.5, 6.1, 16.6, and L8 for the 1-chloroanthracene, 2- 
chloroanthracene, 9-chloroanthracene, and 9-chlorophenanthrene. 
These values are somewhat lower than those obtained with a single 
detector (119). A sample of 9-chl or oanthracene was analyzed in the 
same repetitive fashion as reported previously. The results of 
these analyses are shown in Figure 69. The response enhancements . 
are 16.7, 16.7, 16.6, 17.2, 16.8, 17.0, 17.2,17.4, 17.4, and 17.2. 
The results of the response enhancement measurement indicate that 
the relative standard deviation for repetitive analyses is 2.1%.
For those compounds which have large normal BCD responses, one 
would expect that the second undoped detector would see slightly 
less material, and as a result, the response in the second detector 
would be somewhat less. This is in fact what is observed in Figure
70 which is an example of the two undoped detector responses to 9- 
nitroanthracene and hexachlorobenzene. Both of these compounds are 
very strongly responding compounds (140).
Mixture analysis can be accomplished with little or no loss of 
resolution from one detector to the other. This is shown in Figure
71 and illustrates no loss of chromatographic resolution. In this 
illustration, it should be noted that the response of 9- 
chlorophenanth rene is larger in the second detector when both are 
undoped. This might simply reflect the second detector's increased 
sensitivity for that particular compound. Figure 72 shows the 
response enhancement as a function of concentration of the analyte 
molecule using the series detector system of analysis.
Figure 69. Dual ECD and O^-sensitized ECD responses in series of sampling
containing 9-cnl or ©anthracene repeated ten times. The chromatogram 
on the left is the normal ECD response in both detectors. The con­
centration of oxygen is 0.30% and temperatures are 300°C.
181
182
Figure 70. Dual ECD responses of series detectors of separate 
samples containing (A) hexachlorobenzene, (B) 9- 
nitrobenzene, and (C) 2,7-dinitrofluorene. The arrows 
indicate the peaks of interest in the two cells.
183
Figure 71. Dual ECD and oxygen-sensitized BCD responses of series 
detectors of a mixture of 9-chlorophenanthrene and 2- 
chloroanthracene. The chromatogram on the left is 
without oxygen and the upper right is with 0.30% 
oxygen in the second detector. Detector temperature 
is 300°C
184
Figure 72. Series detectors in which the bottom chromatograms are 
normal ECD responses and the top chromatograms are 
with 0.30% oxygen in the second detector. The sample 
analyzed is 2-chl or ©anthracene with 100 ng, 8 ng, and 
3 ng, respectively. Detector temperature is 300°C
185
Also, this type of experimental design allows the opportunity to 
to have oxygen present in both detectors, and the results are shown 
in Figure 73. The response enhancements from both detectors being 
doped for the 9-isomer are 24.6 and 19.8 for an average of 22.2.
For the 2-isomer, the enhancements obtained are 6.4 and 8.1 for an 
average of 6.9. The 1-chlor©anthracene exhibits response 
enhancements of 3.7 for detector I and 5.4 from detector 2 for an 
average of 4.5. These results compare quite favorably with results 
obtained previously with parallel detectors and a single detector. 
Another interesting aspect of doping both detectors is shown in 
Figure 74 in which the concentration is fairly large. For this 
particular instance, the second detector shows a larger enhancement 
than the first detector. A possible explanation is that several 
oxygenated species are formed in the first detector which in turn 
react with oxygen in the second detector. This type of system is 
again a much more precise method of determining response 
enhancements and allows for much faster analysis times.
In order to more accurately determine relative detector 
responses, a much better and shorter transfer line might be used.
Part of the problem here might involve some absorption in the 
transfer, line. This technique definitely has some interesting and 
varied applications.
Tandem Cells
Another example of series detectors is shown in Figure 11. This 
is a specialized dual detector in which two EC volumes are placed in
186
Figure 73. Series detectors in which the chromatograms on the
left are with the normal ECD response and the ones on 
the right are with 0.30% oxygen present in both 
detectors. The sample is 2-chloroanthracene at a 
detector temperature of 300°C.
187
I
X4
X1
V
X1
I
X4
J
Figure 74. Dual ECD responses and oxygen-sensitized response of
50 ng sample of 2-chloroanthracene. The chromatograms 
on the left are with normal EC detection and the ones 
on the right are with oxygen present in both 
detectors.
188
series and the second volume contains added oxygen when applicable. 
Except for the most strongly responding molecules, the ECD can be 
considered a nondestructive detector. With this tandem arrangement, 
one can arrange conditions so that each cell receives the same 
quantity of analyte, and the reproducibility of oxygen—induced 
enhancement measurements should be significantly improved. At the 
time of this writing, the tandem cells are being characterized and 
evaluated for analysis capabilities^ The tandem cell presents an 
interesting cell design (i.e. the anode pins in the side of the cell 
rather than coaxial as in the Varian ECD). This type of design 
would allow oxygen doping to take place in the second cell, and, in 
addition, studies of strongly responding compounds can be 
accomplished. For those strongly responding compounds, one should 
see a much smaller response in the second cell. Also, since the 
response enhancement measurements can be done with one injection, 
and thereby increasing the precision of measurements and at the same 
time, decreasing the amount of time for analysis. In addition, 
oxygen can be added to both cells by adding it as a portion of the 
total makeup gas prior to the first detector. <
The best arrangement for reproducibility is the parallel 
detectors using dual columns in the same injector port. The series 
detector arrangement should be used with caution, in view of the 
fact that additional reactions seem to be occurring in the second 
detector, particularly when both are doped.
189
SJMMARY
This study has clearly shown that oxygen-induced enhancements 
in the ECD analysis of various isomeric sets of compounds vary with 
the chemical structure of the molecule.
For the poly chi or oina ted biphenyls, the response enhancements 
were small but several isomers could be differentiated in some 
instances. The response enhancements seem to be useful for isomer 
distinction for these compounds up to the tetrachloro series. For 
this particular set of isomers, the normal .ECD response is high, 
and as a result, the enhancements were small and very similar.
Most of the enhancement values were approximately 2, regardless of 
the number of chlorine atoms attached to the ring.
The chloroanthracenes and chiorophenanthrene isomers can 
easily be differentiated based solely on their response enhancement 
values. Binary mixtures of coeluting compounds can be 
quantitatively determined by utilizing the oxygen-doped ECD and the 
mole fraction of each component in the unresolved peak can be 
determined. For mixtures of three coeluting components, the 
response enhancement measurements must be repeated at another 
detector temperature. In cases where only partial resolution of 
the components is observed and where the retention times are much 
too close for distinction, the 02-doped ECD can be extremely 
beneficial for identification.
The APIMS was used to identify specific ionic species formed 
with and without the presence of added oxygen for the 
chloroanthracene and phenanthrene isomers. For these compounds, 
actual oxygen incorporation is observed and a reaction scheme has 
been proposed to account for the species observed in the API MS.
For the I-chi oroanthracene and 2-chloroanthracene molecules, the 
major ionic species observed without oxygen was the M~ parent ion 
and with oxygen present in the source, the major ion was (M+C^)- 
The 9-chlor©anthracene and 9-chlorophenanthrene form (M-Cl+C^)- 
species both with and without the presence of oxygen in the source. 
In addition, the results indicate that oxygen attack actually 
occurs at the 9-position in anthracene. Several of the isomers can 
be differentiated based on the ratios of observed ionic species.
The polycyclic aromatic amines and hydroxides provided an 
interesting set of isomers that were examined. Both of these are 
extremely important from an environmental standpoint, in view of 
their reported mutagenic and carcinogenic characteristics. Their 
normal ECD responses are very low; therefore, an electron enhancing 
tag is attached to the substituent( trifluoroacetic anhydride or 
perfluoropropionic anhydride for the amine and methyl for the 
hydroxy compounds). The response enhancements observed were 
determined for various sets of isomers including biphenyl, 
naphthalene, anthracene, phenanthrene, fluorene, pyrene, and 
chrysene. In most cases, isomers could be distinguished based 
solely on the enhancement values. In addition, stepwise 
derivatization can be done on two different substituents and the
190
Iresults indicate the enhancement values may be isomer dependent.
The APIMS was utilized to examine the negative ions of the TFAA 
and PFPA derivatives with and without the presence of oxygen. The 
salient feature of the API mass spectrometry of every TFAA 
derivative examined was that the only negative ion observed both 
with and without added oxygen was the M- ion. This suggests that 
the ionization of these molecules under oxygen free conditions 
occur by simple electron capture and with added oxygen occurs by 
charge transfer to form the parent negative ion in both cases. The 
APlMS results support a resonance type of electron capture 
mechanism for the TFAA derivatives which is substantiated by a 
temperature vs. response for this type of derivative with the 
normal ECD response. The PFPA derivatives enhance very little and 
the APIMS results indicate for dissociative type of electron 
capture mechanism which is again supported by ECD studies.
The TFAA derivatives of anthracene and phenanthrene were 
examined using the same experimental conditions in both the ECD and 
the APIMS. The response enhancement values obtained from the two 
methods correlate quite well. The relative order of the observed 
ECD and APIMS enhancements are the same, although the exact numbers 
determined are not the same.
Several of the TFAA and PFPA derivatives were examined by 
electron impact and chemical ionization mass spectrometry. For the 
TFAA derivatives of aminoanthracene and aminophenanthrene, negative 
Cl shows considerable promise for isomer distinction. The 4- 
am inophenanthrene( both PFPA and TFAA) can easily be distinguished
■ 19I
192
from the other isomers based solely on the propensity of this 
derivative to form the following proposed ring structure:
The precision of the enhancement measurements was primarily 
determined by the ability to reproducibly inject a given volume of 
sample. Improved reproducibility as well as increased ease of 
measurement was accomplished by using two ECDs simultaneously where 
one detector is operated normally and the other is doped with 
ojygen. By using dual columns in the same injector port, the 
relative standard deviation in repetitive measurements was 2.6%. 
This was a tremendous improvement over the previous method of two 
separate, paired analyses, with only one detector. The use of the 
parallel detectors also has reduced the uncertainty by a factor of 
three and has greatly relaxed the requirement for precise 
injections. In addition, series detectors were utilized in which 
the first ECD is connected to the second one with a. transfer line. 
By doping the second detector the response enhancements can then be 
determined with one injection. The relative standard deviation for 
repetitive analyses was 2.1%. Both of these techniques represent a 
major improvement in determining response enhancements.
The 02~doped ECD has become a very powerful analytical for 
distinguishing isomers even in cases where mass spectrometry fails. 
In addition, the technique can be used in quantitative analysis of
193
coeluting components and in instances where only partial resolution 
is obtained.
194
FUTURE APPLICATIONS
The next logical step in utilizing the 02-doped ECD would be 
to extend its application to actual environmental samples. An 
example of this would be to examine the amine fraction from a SRC 
II process. By the following separation procedure modeled after 
Later et al. (115), the enriched amine fraction can be isolated. 
The separation procedure is illustrated in Figure 75. After the 
derivatization procedure with either TFAA or PFPA and utilizing the 
oxygen-doped BCD, the derivatized amines could be identified based 
on their specific response enhancements. Using Figure 76 as an 
example of very good chromatographic separation of TFAA 
derivatives, the response enhancements would allow identification.
The application of the C^-doped ECD could also be extended to 
environmental samples such as dioxins or dibenzofurans. With the 
preliminary separation and clean up done by HPLC, the doped ECD 
should be very helpful in this area.
By using dual columns or a fused silica splitter, the oxygen- 
doped ECD can be used in parallel with other detectors such as the 
FID or with nitrogen, sulfur, or phosphorous specific detectors.
In addition, Andy Valkenburg in our laboratory has recently 
discovered that the addition of ethyl chloride to an electron 
capture detector increased the response for molecules similar to 
anthracene, molecules that have a low normal BCD response and
195
I-------
Hexane
I
Aliphatic
Hydrocarbons
SAMPLE
Neutral 'Alumina
Benzene
Neutral
PAC
--- 1
Chloroform
I
N PAC
Silicic Acid
Benzene-ether
I
Hexane-Benzene
I
Benzene
I
V
3' PANH 2' PANH
-y
Enriched
APAH
Derivatization
Fluorpamides
APAH
Figure 75. Chemical class separation scheme for synthetic fuel 
products. Key: polycyclic aromatic compounds (PAC), 
nitrogen polycyclic aromatic compounds (N-PAC), 
secondary nitrogen polycyclic aromatic heterocycles 
(2° -PANH), amino polycyclic hydrocarbons (APAH), and 
tertiary nitrogen polycyclic aromatic heterocycles (3° 
-PANH). (Reference 59).
196
Figure 76. ECD capillary gas chromatogram of TFAA derivatized
APAH fraction in SRC II HD. (I) aminoaphthalene, (2) 
2-aminoaphthalene, (3) Cj-aminoaphthalene, (4) Co- 
aminoaphthalenes, (5) 3-aminobiphenyI, (6) 4- 
aminobiE*ienyl, (7) Cj-aminobiphenyl, (8) C2- 
aminobipjienyl, (9) aminofluorene, (10-12) 
aminoanthracenes, (13) aminofluoranthrene and (14) 
aminopyrene. (Reference 59).
197
capture electrons by a resonance type of mechanism. It should also 
be pointed out that the addition of ethyl chloride to the detector 
does not significantly effect the baseline frequency, in contrast 
to the presence of ojygen.
Figure 77 illustrates the normal ECD response for anthracene 
with an without the presence of ethyl chloride. The response to 
anthracene is enhanced by approximately 70.
In much the same way that the APIMS was utilized to determine 
the ions formed with and without the presence of oxygen, the API 
was also used to observe the negative ions formed in the presence 
of ethyl chloride in the source. The ions observed with ethyl 
chloride included Cl- and m/e at 206. Both of these ions were not 
observed with the absence of ethyl chloride. The corresponding 
scanning mass spectra and single-ion monitoring is illustrated in 
Figure 78.
The following mechanism is proposed for the reaction of ethyl 
chloride with anthracene to produce an enhancement in the observed 
detector response.
e~ + A A™ (33)
PT + CH3CH2Cl Cl" + B (34)
where A is the anthracene molecule. This mechanism is similar to 
the mechanism proposed for the oxygen-induced response enhancement.
O2 + e~ O2"
O2" + A Products
198
cn—
Figure 77. Capillary gas chromatograms of 7 ng sample of 
anthracene with ethyl chloride doping. The 
chromatogram on the left is with normal EC response 
and the one on the right is with 100 ppm ethyl 
chloride in the detector. Detector temperature is 
250°C.
199
M/E 35
M/E 206
Figure 78.
■d/
— jutiW
/
V
1K
Single ion APIMS monitor of the negative ion intensity 
at m/e 35 (top) with a sample of anthracene with and 
without added ethyl chloride at a source temperature 
of 250°C The bottom chromatograms are single ion 
APIMS monitor of the negative ion intensity at m/e 206 
with a sample of anthracene with and without added 
ethyl chloride at source temperature of 250°C.
200
In the second case, the anthracene molecule is actually the 
substrate with the oxygen anion the catalyst. However, in the 
first example, the anthracene molecule actually substitutes for the 
oxygen an now becomes the catalyst.
In view of the fact that the addition of ethyl chloride 
actually decreases the baseline frequency in contrast to oxygen 
doping, there are many applications for this technique. Other 
compounds will be examined for response enhancements in the 
presence of ethyl chloride.
Another application with the addition of ethyl chloride doping 
and using parallel detectors, one could be doped with oxygen and 
the other with ethyl chloride. The ratio of responses might also 
be isomer dependent.
2 0 1
LITERATURE CITED
2 0 2
(1) S. Jensen, New Scientist. 32, 612(1966).
(2) P.W. Risebrough, P. Reicha, S. G. Herman, D. B. Peakall and 0. 
N. Kerivan, Nature(London), 220. 1098(1968).
(3) D. B. Peakall, Res. Rev... 44, 1(1972).
(4) S. Jensen, Ambla. I, 123(1972).
(5) I. C. T. Nisbet and A. F. Sarofin. Environ. Health Persoect. . 
I, 21(1972).
(6) D. Pal, J. B. Weberr, and M. R. Overcash, Res. Rev. , 4, 
45(1980).
(7) L. Fishbein, J. Chromatoar. , 58. , 345(1972).
(8) C. Karr and T. L. Chang, J. Inst. Fuel. 31. 522(1958).
(9) D. Braun, D. G. Earrishaw, F. R. McDonaldland and H. G. Jensen, 
Anal. Chem.. i2, 146(1970).
(10) C.-h. Ho, B. R. Clark, M. R. Guerin, C. Y. Ma, and T. K. Rao, 
Prep. Pap-Am. Chem. Soc., Div. Fuel Chem., 2A, 281(1979).
(11) M. R. Guerin, C.-h. Ho, T. K. Rao, B. R. Clark and J. L.
Epler, Hnviron. Res.. 23. 42(1980).
(12) B. W. Wilson, R. A. Pelroy and J. T. Cresto, Mutat. Res.. 79. 
193(1980).
(13) M. R. Guerin, I. B. Rubin, T. K. Rao, B. R. Clark and j. L. 
Epler, £iiel, M f  282(1981).
(14) M. V. Buchanen, C. -h. Ho, M. R. Guerin and B. R. Clark. In " 
Chemical Analysis and Biological Fate: Polynuclear Aromatic 
Hydrocarbons.", M. Cooke and A. J. Dennis, Eds, Battelle 
Press: Columbus, Ohio, 133(1981).
(15) R. A. Pelrpy and A. Gandolfi, Mutat. Res.. 22, 329(1980).
(16) B. w. Wilson, R. A. Pelroy, M. L. Lee and D. W. Later, In " 
Proceedings of the 29th Conference on Mass Spectrometry and 
Allied Topics, " Abstract No. MPM085(1981).
(17) B. w. Wilson, J. C. Craun, R. A. Pelroy, M. R. Peterson and w. 
a  Felix, In " Proceedings, 2nd USDOE Environmental Control 
Symposium," I, 415(1980).
203
(18) D. Hoffmann and E. L. Wynderf In " Chemical Carcinogenesis," 
C. E. Searlef Ed., American Chemical Society, Washington, D. 
C., p. 324(1976).
(19) D. G. Clayson and R. C. Garner, In " Chemical 
Carcinogenesis," C. E. Searle, Ed., American Chemical 
Society: Washington, a  C., 366(1976).
(20) B. Shaikh, M. R. Hallmark, R. K. Hallmark, W. G. Manning, A. 
P m n o c k  and J. C. Kavalik, J. Chromatoar. 195. 392(1980).
(21) a  W. Later, Battelle Northwest Labs, Richland, Washington, 
personal communication, June, 1983.
(22) F. K. Schweighardt, C. M. White, S. Friedman and J. L. 
Schultz, In " Organic Chemistry of Coal," American Chemical 
Society: Washinbgton, a C, 240(1978).
(23) P. C. Uden, A. P. Carpenter, H. M. Hackett, P. E. Henderson 
and S. Siggia, Anal, Chem.. 51, 38(1979).
(24) J. F. Schabron, R. J. Hurtebise and H. F. Silver, Anal. Chem. 
51 , 1426(1979).
(25) a  W. Wilson, Battelle Northwest Labs, Richland, Washington, 
personal communication, June 19882. .
(26) B. A. Schwertz, J. M. Norris, G. L. Sparschen, U. K. Rowe, P. 
J. Gehring, J. L. Emerson and G, C. Gerbig, Environ. Health 
Perspect. , 5, 87(1973).
(27) H. Buser, J, Chromatoar.. 107. 275(1975).
(28) R. E. Clement and F. W. Karasek, J. Chromatoor. , 234, 
395(1982).
(29) T. J. Nestrick and L. L. Lamparski, Anal. Chem.. 54. 
2292(1982).
(30) A. M. Krstulovic, D. M. Rosie and P. R. Brown, Anal. Chem..
42, 1383(1976).
(31) D. Fechner and B. Seifert, Z. Anal. Chem. 292. 199(1978).
(32) R. D. Smillie, D. T. Wang and 0. Meresz, J. Environ. Sci. 
Health. A13. 47(1978).
(33) R. Thoms and M. Zander, A. Anal. Chem.. 282. 443(1976).
(34) R. K. Sorrell and R. Reding, J. Chromatoor.. 185. 655(1979).
y
204
(35) E. Sawickif T. W. Stanley, and W. C. Elbert, Talanta. 11. 
1433(1964).
(36) H. Dreeskamp, E. Keck, and M. Zander, A. Naturforsch. 30a. 
1311(1975).
(37) P. L. Konash, W. E. May and S. H. Wise, Washington 
Chromatography Discussion Group Meeting, Gaithersburg, 
Maryland, May 21, 1980.
(38) J. R. Jadamec, W. A. Saner and Y. Ta!mi. Anal. Chem.. 49. 
316(1977).
(39) R. G. Christensen, H. S. Hertz, S. Meiselmen and E. White, 
Anal. Chem., in press.
(40) W. A. Dark, W. H. McFadden, and D. L. Bradford, J. 
Chromatoor. Sci.. 15. 454(1977).
(41) T. Takeiichi, Y. Hirata and Y. Ohumura, Anal. Chem.. 50. 
660(1978).
(42) W. H. McFadden, J. Chromatoor. Sci.. JJ, 2(1979).
(43) W. H. McFadden, J. Chromatoor. Sci.. 18. 97(1981).
(44) T. Takeuchi, Y: Hirata and Y. Ohumura, Anal. Chem.. 50. 
289(1978).
(45) W. H. McFadden. J. Chromatoor. Sci., 18. 97(1980).
(46) M. L. Lee, K. D. Bartle, and M. V. Novotny, Anal. Chem.. 47. 
540(1975).
(47) V. Cantuti, G. P. Cartoni and U. Cantuti, J. Chromatoor.. 17. 
60(1965).
(48) G. Alberini, V. Cantuti, and G. P. Cartoni, In "Gas 
Chromatography 1966," A. B. Littlewood, Ed, Institute of 
Petroleum, London, 258(1967).
(49) A. Bjorseth and G. Eklund, JJED A £Qr 2, 22(1979).
(50) S. Safe and 0. Huntzinger, In " Mass Spectrometry of 
Pesticides and Pollutants, "CRC Press, Cleveland, Ohio, 
77(1973).
(51) E. Stenhagen, S. Abrahams and F. W. McLafferty, In "Registry of 
Mass Spectra Data", John Wiley and Sons, New York, New York, 
1-4, 1974.
205
(52) R. A. Flathf In "Guide to Modern Methods of Instrumental 
Analysis," T. H. Gauenf ed.f Wiley-Intersciencef New York, 
Chapter 9, 1972.
(53)
(54)
(55)
(56)
(57)
(58)
F. J. Biros, Residue Rev.. 4Q., 1(1971).
E. J. Bonelli, Am. Lab. February 1971.
S. Safe and 0. Huntzingerf Chenu Comm.. 446(1971).
S. Safe and 0. Huntzihgerf J. Chem. Soc. Perkin Trans.. 
686(1972).
X/
S. Safe and 0. Huntzingerf J. Chem. Soc. D. 446(1971).
C.-h. Ho, B. R. Clark, 
Prepr. Bap.-Amer. Chem. 
281(1979).
M. R. Guerin, C. Y. Maf and T. K. Raof 
Soc. , Division of Fuel Chemistry, 24,
(59) D. w. Later, M. L. Lee and B. w. Wilson, Anal. Chem., 54, 
117(1982).
(60) K. Blau and G. S. King, In "Handbook of Derivatives for 
Chromatography", Heyden:London, P.104-151, 1978.
(61) D. R. Knfapp, In "Handbook of Analytical Derivatization 
Reaptions", WileyrNew York, p.65-103,1979.
(62) B. A. Tomkins and C. Feldman, Anal. Chip. Acta. 119.,
203(1980).,
(63) C. Patriankos and D. Hoffmann, J. Anal. Toxicol.. 3,
150(1979).
(64) M. L. Lee, D. L. Vassilaros, w. S. Pipkin and w. L. Sorenson, 
In " Trace Organic Analysis; NBS Special Publication 519, U. 
S. government Printing Office, Washington, D. C., p.731 (1979).
(65) J. S. Warner, Anal. Chem.. 48, 578(1976).
(66) M. L. Lee and R. A. Hites, J. Am. Chem. Soc.. ££, 2008(1977).
(67) R. A. Hites and G. R. Duboy, In "Carcinogensis-A 
Comphrehensive Survey: Polynuclear Aromatic Hydrocarbons", P.
W. Jones and R. I. Freudenthal, Eds., Raven Press, New York, 
vol. 3, 1978.
(68) D. F. Hunt, P. J. Gale and S. K. Sethi, In" 26th Annual 
Conference of Mass Spectrometry and Allied Topics", St.
Louis, Missouri, May 1978., page 151.
206
(69) B. Shushan, S. H. Safe and R. K. Boyd, Anal. Chem. , 51. 
156(1979).
(70) K. R. Naqvi, A. R. Holzworth and U. P. Wild, Appl. 
Spectroscopy Rev... 12. 131(1977).
(71) M. Sautif, Pure Ajpplied Chem.. 48 , 99 (1976).
(72) L. K. Keefer, L. Wallcave, J. Loo and R. S. Peterson, Anal. 
Ghent.. 50. 936A(1978).
(73) K.D. Bartle, M. L. Lee, and M. Novotny. Analyst. 102. 
731(1977).
(74) E. L. Wehry and G. Mamantoo, Anal. Chem.. 51. 643AQ979).
(75) P. Tokousbalides, E. L. Wehry and G. Manantoo, J. Phys. Ghent. 
SI, 1769(1977).
(76) J. E. Lovelock, J. Chromatoar.. I, 25(1958).
(77) E. S. Goodwin, R. Goulden and J. G. Reynolds, In " Gas 
Chromatography with Electron Capture Ionization Detection for 
Rapid Identification of Pesticide Residue in Crops'1, 18th lnt. 
Congress on Pure and Applied Chemistry, Montreal, August 1961.
(78) j. D. Watts and A. K. Klein, In" Determination of Chlorinated 
Pesticides and Residues by Electron Capture Gas . 
Chromatrogaphy",. 75th Annual Meeting of the Association of 
Official Agricultural Chemists, Washington, D. C., October, 
1961.
(79) J. E. Lovelock, Anal. Chem.. 35. 474(1963).
(80) W. E. Wentworth, E. Chen and J. E. Lovelock, J. Phys. Chem., 
IQ, 445(1966).
(.81) W. E. Wentworth, R. S. Becker and R. Tung, J. Phys. Chem.. 71. 
1652(1967).
(82) P. G. Simmonds, D. C. Fenimore, B. C. Pettit, J. E. Lovelock 
and A. Zlatkis, Anal. Chem., 39. 1428(1967);.
(83) D. C. Fenimore,'Pi R. Loy and A. Zlatkis, Anal. Chem.. 43. 
1972(1971).
(84) J. A. Lubowitz and W. C. Parker, J. Chromatoar.. 62.
53(1971).
(85) R. J. Maggs, P. L. Johnes, A. J. Davies and J. E. Lovelock, 
Anal. Chem., 43. 1966(1971).
207
(86) P. L. Patterson, J. Chronatogr., 134, 25(1977).
(87) F. W. Karasek, Anal. Ghent.. 46. -710AQ974).
(88) D. I. Carroll, I. Dzidic, R. N. Stillwell, M. G. Horning and 
E. C. Horning, Anal. Chem.. 46. 706(1974).
(89) M. W. Siegel and M. McKeown, Am. Tab.. Novenber 1975.
(90) F. W. Karasek and D. M. Kane, Anal. Chem. 45. 576(1973).
(91) G. E. Spangler and C. I. Collins, Anal. Chem.. 47. 393(1975).
(92) I. Dzidic, D. I. Carroll, R. N. Stillwell and E. C. Horning, 
Anal. Chem.. 42, 1308(1975).
(93) E. P. Grimsrud, S. H. Kim, and P. L. Gobby, Anal. Chem., 51. 
223(1979).
(94) E. C. Horning, D. I. Carroll, I. Dzidic, and R. N. Stillwell, 
In "Electron Capture Theory and Practices in Chromatography", 
Elsevier Scientific Publishing Company, New York, p.359, 1981.
(95) F. W. Karasek, Anal. Chem.. 43. 1982(1971).
(96) E. P. Grimsrud and R. G. Stebbins, J. Chromatogr^ 155. 
19(1978).
(97) P. G. Simmonds, J. Chromatogr.. 166. 593(1978).
(98) D. A. Miller and E. P. Grimsrud, Anal. Chem.. 51. 851(1979).
(99) L. A. Shadoff, Analytical Research Laboratories, Dow Chemical 
Company, Midland, Michigan, personal communication, August, 
1983.
(100) R. A. Rasmussen, L. E. Rasmussen, M. A. Khalil and R. W. 
Dalluge, J. Geophys. Res.. 2, H O  (1976).
(101) E. P. Grimsrud and D. A. Miller, Anal. Chem., JiQ,
1141 (1978). ,
(102) E. P. Grimsrud and D. A. Miller, 9th materials Research 
Symposium, Trace Organic Analysis: A New Frontier in 
Analytical Chemistry, National Bureau of Standards, 
Gaithersburg, Maryland, April 10-13, 1978.
208
(103) R. A. Rasmussen and M. A, Khalil, ,Seophvs. Res. Letters.. 
Ifif 240(1978).
(104) H. J. Van de Wiel and P. Tomassen, J. Chromatnar.. 71.
I (1971). - '
(105) D. A. Miller, K. Skogerboe and E. P. Grimsrud, Anal. Chem.. 
51, 464(1981).
(106) P. L. Gobby, E.P. Grimsrud and S.W. Warden, Anal. Chem.. j52.
(107) M. W. Siegel and M. Cl McKeown, J. Chromatoar.. 122, 
397(1976).
(108) W. A. Aue and S. Kapilla, 
255(1973).
11,
(109) M. A. Biondi, M. H. Bortner and T. Baurer, Ed., "Defense 
Nuclear Agency Reaction Rate Handbook," March 1975, 4th 
revision, Chapter 12.
(HO) S. GL Lias and P. Ausloos, "Ion Molecule Reactions", American 
Chemical Society, Washington, D. C., Chapter 4, 1975.
(111) M. P. Phillips, R. E. Sievers, P. D. Goldan, W. C. Kusten and 
F. C. Fesenfeld. Anal. Chem.. 51. 1819(1979).
(112) R. S. Sievers, M. P. Phillips, R. M. Barkley, M. A. Wizner,
M. J. Bollinger, R. S. Hutte and F. C. Fesenfeld, Anal.
-Chem.. 31, 1819(1979).
(113) P. D. Goldan, F. C. Fesenfeld, W. C. Kuster, M. P. Phillips 
and R. S. Sievers, Anal. Chem.. 52. 1751(1980).
(114) M. A. Wizner, S. Singhawangcha, R. M. Barkeley, and R. E. 
Sievers, Inn " Advances in Chromatography", A. Zlatkis, Ed., 
Proceedings of the 17th International Symposium, Las Vegas, 
Nevada, April 5-8, 1982.
(115) F. R. McDonald and U. Christian, M*. Iab., November 1976.
(116) D. W. Later, M. L. Lee and B. W. Wilson, Anal. Chem.. 54. 
117(1982).
(117) Y. Masuda and D. Hoffman. Anal. Chem.. 41, 650(1969).
(118) Y. Masuda and D. Hoffman, 
695(1969).
I., I,
209
(119) W. B. Knighton and E. P. Grimsrud7 Anal. Chem.. 54. 1892 
(1982).
(120) J. A. Campbell and E. P. Grimsrud7 J. Chromatnar.. 243, I 
(1982).
(121) C A .  Coulson and A. Streitweiser In "Dictionary of - 
Electron Calculations", Walt Freeman and Company7 San 
Francisco, California, 1965.
(122) W. E. Wentworth, L. W. Kao, and D. S. Beckner, J. Phys. 
Chem., 79. 1161 (1975).
(123) E. P. Grimsrud, D. A. Miller, R  G. Stebbens, and S. H. Kim, 
ZL of Chromatoqr.. 197. 51 (1980).
(124) J. A. Campbell, E. P. Grimsrud, and L. R. Hageman, Anal. 
Chem., 55, 1325 (1983).
(125) S. B. Martin and M. Rowland, J. Pharm. Sci.. 61. dl235 
(1972).
(126) D. D. Clarke, S. Wills and S. E. Gitlow, J. Gas Chromatoor.. 
I, 310 (1966).
(127) B. C. Pettitt, P. G. Simmonds and A. Zlatkis, J. Chromatogr. 
Sci., 2, 645 (1969).
(128) A. Zlatkis and B. C  Pettitt, Chromatoaraphaics. Zr 484 
(1969).
(129) D. W. Later, Battelle Northwest Laboratories, personal 
communication, June, 1983.
(130) E. P. Grimsrud, S. W. Warden and R. G. Stebbins, Anal. Chem.. 
51, 716 (1981).
(131) J. A. Campbell and E. P. Grimsrud, accepted J. of Chrcmatogr.
(132) J. A. Campbell, M. J. ConnalIy, L. L. Sears, and E. P. 
Grimsrud, in preparation.
(133) S. Sugasawa and H. Shigahara, Berichte. JAr 459 (1941).
(134) A. Pictet and A. Gams, Chem. Berichte. AAr 2030 (1911).
(135) G. Todor, J. Gal, and B. H. Phillips, Berichte. 74. 459 
(1941).
(136) Y. Iida and S. Daishina, Chem. Letters. 273 (1983).
2 1 0
(137) M. A. Fox and S. Olive, Science, 205 , 582(1979).
(138) J. A. Campbell and E. P. Grimsrud, submitted J. Chrcmatogr.
(139) D. W. Later, B. W. Wright, and M. L. Lee, UEC  ^  ££, 406 
(1981).
(140) D. A. Miller, Michigan Institute of Technology, personal
communication, June, 1983. ,
MONTANA STATE UNIVERSITY LIBRARIES
IlllilllllliillWlIIlI
3 1762 10005423 6
D378
C152 Campbell, J. A. 
cop.2 Application of the 02-dope< 
ECD to isomer differentiatiox
DATE ISSUED TO
D378 
C152 
cop.2