QUANTITATIVE CAPILLARY COLUMN
GAS CHROMATOGRAPHY-MASS SPECTROMETRY METHODS OF
ANALYSIS FOR TOXIC ORGANIC COMPOUNDS
R. L. Harless and R. G. Lewis
Health Effects Research Laboratory (MD-69)
United States Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Presented at
Symposium on "Practical Solutions to
Quantitative Capillary Column Gas Chromatography"
The Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy
Atlantic City, New Jersey
March 10-15, 1980
-------
ABSTRACT
High resolution glass capillary column gas chromatography (GC)
coupled with low or high resolution mass spectrometry (HRMS) is one of
the most powerful tools in analytical chemistry. Low concentrations,
nanogram- and picogram-per-gram (ng/g and pg/g) levels, of toxic organic
compounds in complex sample media can be unambiguously identified and
quantified utilizing this technique.
A capillary column GC-HRMS multiple ion selection method of analysis
was developed and utilized for the quantitative determinations of ng/g
and pg/g levels of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) residues
in human, biological and environmental samples. Approximate minimum
limits of detection were: human milk and adipose tissue - 0.5 to 4 ppt;
beef, deer and elk adipose tissue - 2 to 5 ppt; fish - 2 to 20 ppt;
water, soil and sediment - 0.02 to 5 ppt. The applications of this
technique to the quantitative determination of TCDD, higher chlorinated
dioxins, and other toxic compounds in needle biopsy samples of human
tissue (ca. 100 milligrams) are also described. Variables that influence
quantitative analyses, criteria for confirmation of toxic residues,
detection limits, quality assurance programs for validation of results,
and typical results are discussed in the text.
The GC-MS interface is a very critical part of this system. A
versatile and unique GC-MS interface (1,2), which utilizes a positive
helium atmosphere at the coupling point (atmospheric pressure to 1 x
10-6 Torr vacuum) and eliminates the need for a gas tight connection
requiring graphite ferrules, was devised. The GC-MS interface has been
in constant use for three years and has eliminated or minimized many of
the problems associated with the use of glass capillary columns; e.g.,
vacuum leaks, and glass breakage. Additional advantages are: (1)
direct GC-MS coupling for maximum transfer efficiency; (2) atmospheric
pressure at the GC column exit to enhance column lifetime and efficiency;
and (3) single valve isolation of the MS from the GC.
2
-------
INTRODUCTION
Mass spectrometry (MS) was essentially created in 1898 when Wien*
showed that a beam of positive ions could be deflected by electric and
magnetic fields. Thompson produced crude mass spectra in 1905 and
documented the existence of two neon isotopes through the use of a
M
single magnetic deflection instrument in 1912. Dempster and Aston in
1918 and 1919, respectively, designed more elaborate instruments for the
measurement of relative abundance of isotopes. Reliable mass spectro-
meters became available in the 1940s and were used extensively then by
the petroleum industry. During the 1960s, and especially in the 1970s,
the field developed rapidly and has today become an indispensable tool
for the qualitative and quantitative characterization of organic compounds.
While the mass spectrum is uniquely characteristic of the pure
organic molecules being ionized, in most samples these compounds are
encountered as only components of complex mixtures. Gas chromatography
(GC) came into use in 1952 and was recognized to be an ideal separation
tool for such complex mixtures. It was obvious that the combination GC
and MS techniques would yield a very powerful analytical device. Unfor-
tunately, the coupling of the instruments is quite difficult because of
their incompatibility - the GC operating at atmospheric pressure, while
the MS requires a high vacuum source. However, many types of GC-MS
interfaces have been developed to meet the need; e.g., the frit separator
introduced by Watson and Bieman^, jet separators of Becker** and Ryhage^,
O Q
si 1 cone membrane and direct coupling.
3
-------
The GC-MS interface is a very critical part of this system described
here. A versatile and unique GC-MS interface was developed^ which
utilizes a positive helium atmosphere at the coupling point (atmospheric
pressure to less than 1 x 10 ® Torr vacuum) and eliminates the need for
a gas tight connection requiring graphite ferrules. The interface has
been in constant use in the authors' laboratory for three years and has
eliminated most of the problems associated with the use of glass capillary
columns; e.g., vacuum leaks and glass breakage. Additional advantages
of the system are:
(1) direct GC-MS coupling insures maximum transfer efficiency;
(2) atmospheric pressure at the GC column exit enhances column
lifetime and efficiency;
(3) single valve isolation of the MS from the GC;
(4) a polar and non-polar column is constantly ready for use and
may be easily and rapidly (1 min) coupled to provide the
desired chromatographic separation.
A direct-coupled capillary column GC-MS system is one of the most
powerful tools in analytical chemistry. Polar and non-polar glass
capillary GC columns provide optimal chromatographic resolution, the
high sensitivity and maximum specificity. Nanogram- to picogram-per-
gram (ng/g to pg/g) concentrations of toxic compounds present in complex
mixtures may be unambiguously identified and quantitatively determined
with this technique.
4
-------
The application of HRMS techniques for the determination of parts-
per-trillion (ppt) levels of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD)
in various sample types has involved the following HRMS principles: (1)
direct probe, single ion monitoring HRMS analysis;11 (2) GC-HRMS double-
12
ion monitoring; and (3) the method of analysis described herein, which
13
is based on glass capillary GC-HRMS techniques. These techniques are
quite versatile and may be easily and rapidly applied to the determination
of other toxic compounds of interest; e.g., polychloroinated dibenzofurans
(PCDFs), polychlorinated biphenyls (PCBs), etc.
It is imperative that the methodology utilized by U.S. Government
regulatory agencies be validated. Therefore, great measures have been
taken to address this requirement. The TCDD methodology utilized by the
U.S. Environmental Protection Agency and its collaborating laboratories
has been validated by scientifically acceptable methods which have
included: (1) validation studies (blind samples fortified with known
37
amounts of Cl^-TCDD and 0.2 to 100 ppt TCDD); (2) the incorporation of
quality assurance (QA) samples into the analysis of actual samples (10%
to 20% of total); (3) multiple laboratory participation; and (4) the
application of specific criteria for confirmation of TCDD residues.
Analyses for TCDD residues in human and environmental samples must
be performed at the picogram-per-gram (ppt) concentration level because
of its extreme toxicity and its occurrence as a trace contaminant (parts
per billion) in specific chemical products. These analyses are complicated
by the presence of naturally occurring compounds and chlorinated industrial
pollutants such as PCBs, chlorinated benzylphenyl ethers, etc. Therefore,
"5
-------
extremely efficient and specific sample preparation procedures, coupled
with highly sensitive and specific GC-HRMS detection techniques, are
pre-requisites for TCDD analysis at the low ppt concentration range.
The comprehensive sample preparation procedures and capillary column
GC-HRMS methods of analysis for TCDD in various media are described
14
elsewhere.
The application of these techniques, and the results obtained for
the quantitative determinations for TCDD in fresh water fish, fly ash
from coal-fired power plants, and in 250 milligram samples of human
tissue are discussed. Practical solutions to specific types of problems
encountered in these quantitative analysis are described. The versatility
of these techniques is shown in the determinations for polychlorinated
dibenzofurans in ambient air samples collected near the incineration of
PCBs.
6
-------
EXPERIMENTAL
Instrumentation: A Varian MAT 311A mass spectrometer directly
coupled to a Varian Model 2700 gas chromatograph (GC) was utilized for
these analyses. The GC was equipped with a 30-m x 0.25-mm (i.d.) 0V-101
WCOT glass capillary column capable of resolutions of 70,000 to more
than 100,000 effective plates. The MS was equipped with a combination
chemical ionization (CI) and electron impact (EI) ion source (operated
in the EI mode), and an eight-channel hardware (manual control) multiple
ion selection (MIS) device. Each MIS channel was individually controlled
for selection of acceleration voltage, measurement of range output,
signal bandwidth, compensation for background contamination, and integration
rate. The intensities of the selected masses were monitored in a time-
division multiplex system, alternatively set to each of the selected
masses and recording their intensities simultaneously on an eight-channel
Soltec recorder. The adjustable integration rate (0.01 to 1 sec) was
sufficient to accurately reproduce capillary column peaks of 2-sec width
at half peak-height. The electrostatic analyzer voltage (ESA) was
monitored and used to calculate the exact acceleration voltage required
for the specific masses to be monitored. The sensitivity by mass specific
detection was dramatically increased compared to normal spectrum scanning
because the specific masses were measured over a longer and integrated time
basis.
The magnet current was tuned to perfluorokerosene (PFK) m/z 318.9793
and the MS adjusted for 7,000 to 8,500 mass resolution. The ESA voltage
was monitored and used in calculating the exact acceleration voltages
7
-------
required for TCDD masses 327.8847 (C12H40237C14), 321.8936 (C12H40235C1337C1)
and 319.8965 (C^H^O^Cl^). The calculated values were introduced into
MIS channels 2, 3, and 4.
COCL Loss Analysis: The exact acceleration voltage required for
TCDD masses m/z 256.9327, m/z 258.9298, m/z 319.8965, m/z 321.8936, and
m/z 327.8847 were introduced into respective MIS channels utilizing PFK
m/z 254.9856 as the reference. The analysis was performed through
adherence to the previously described MIS procedure time schedule. The
GC-HRMS MIS five channel simultaneous responses for TCDD and 37C1-TCDD
were recorded on an eight channel Sol tec recorder.
Elemental Composition Analysis: The MS was adjusted for 10,000
mass resolution utilizing PFK m/z 318.9793 as the reference. The peak
matching analysis in real time was initiated under the exact time schedule
of events utilized in the GC-HRMS MIS analysis. The PFK reference mass
and TCDD masses (m/z 319.8965 or 321.8936) are observed to be exactly
superimposed on the MS oscilloscope at the correct retention time of
TCDD. The presence of TCDD isomers eluting before or after TCDD may
also be confirmed in this analysis.
GC-HRMS Operating Parameters: Injection port temperature - 260°C;
GC transfer line into MS source - 255°C; column temperature - 70°C;
programmed at 34°C/min to 270°C, beginning exactly 6 minutes after
injection of sample; GC-MS interface isolation valve closed at 12 minutes;
MS ion source temperature - 235°C; electron energy - 70 eV; filament
emission - 1 mA; variable acceleration voltage - 3 KV maximum; mass
resolution - 7,000 to 10,000; multiplier gain - 10®; MIS analysis initiated
37
16 minutes after injection of sample; TCDD and CI-TCDD retention
8
-------
time - 20 minutes ± 15 seconds. MIS technique, electric peak jumping
37
mode; TCDD integration rate, 100 milliseconds; CI-TCDD integration
rate 30 milliseconds.
Precision and Accuracy of GC-HRMS Technique: The reproducibility
of MIS peak-height responses for TCDD quantification standards during
daily operation was calculated to be ± 20% relative to 5 or 10 pg injected.
The GC-HRMS peak matching accuracy in real time for known exact masses
was determined to be ± 2 millimass units at 9,500 mass resolution with
PFK as the reference. A recent modification to the MS provided an
additional gain of 10 or 100 to the maximum MS sensitivity. Peak matching
analysis in real time at 9,500 mass resolution can be performed on 1 pg
TCDD with this modification.
GC-MS Interface: The platinum capillary coupling point (vacuum to
15
atmosphere) is similar to the technique published by Neuner-Jehle et al.
It provides an excellent point for establishing the desired pressure
drop, good mechanical stability, and presents an inert surface to most
organic compounds.
The interface diagram is shown in Figure 1. A positive helium
atmosphere is maintained in the interface and surrounds the atmosphere-
to-vacuum coupling point. The platinum capillary tubing and glass
capillary column are butted together inside a glass sleeve located
within the helium atmosphere. The pressure of the helium atmosphere can
be varied to accommodate capillary column flow rates of 0.5 to 5.0
mL/min helium. The chromatograph end of the interface contains the
helium atmosphere and is open to the atmosphere. Therefore, it can be
9
-------
ELECTRIC VALVE
POSITIVE
i/r* • i/w
RIDUCING UNION
WARNING: A NT* TO 1/1C
SWAGCLOK^UNMN MUST BC
INSTALL!0 TO CONNECT TMt
CAMLLARV COLUMN TO THE
INTERFACE TO FROVIOC A
-------
considered a unique open system utilizing a helium atmosphere for
connecting the glass capillary column (operating at atmospheric pressure)
to the mass spectrometer ion source (operating at 10 ® Torr).
A bypass/vent line equipped with an electric and manual valve
connects the interface to the MS rough vacuum pump. The activation of
either valve effectively permits the solvent peak, or any unwanted peak
to be diverted away from the ion source. For example, 2 pi of benzene
injected into the capillary column at 100°C raises the source pressure
to 2 x 10 6 Torr for thirty seconds if either valve is open. The electric
valve is used during operation and the manual valve is used for standby
operation and allows the ion source to be maintained at the desired
pressure of 1 x 10 ® Torr for extended periods of time. After closure
of the valves, the original source pressure of 9 x 10 ® Torr is reached
in ca. 90 sec. This equilibration time can be reduced by relocating the
valves and reducing the length of the bypass/vent line. The activation
of either valve will reduce the source pressure from 9 x 10 ® Torr to
less than 1 x 10 ® Torr in 4 sec. WARNING: A 1/8" to 1/16" Swagelok®
union or equivalent must be installed to connect the capillary column to
the interface to provide a gas tight connection before using flammable
and explosive reagent and carrier gases.
10
-------
Sample Preparation: The sample preparation procedure involved:
(1) fortification of 10-g sample with 2.5 to 10 ng of ^Cl-TCDD* for
determination of extraction and clean-up efficiency; (2) saponification
with hot alkali, followed by extraction with hexane; (3) washing with
concentrated sulfuric acid; (4) chromatographic clean-up on alumina; and
(5) concentration of the column extract to 60 pi for GC/HRMS analysis.
In addition, a "neutral" clean-up procedure (acetonitrile partitioning)
was utilized for the specific isolation of TCDD from highly contaminated
sources for confirmation, but it has not been validated by multiple laboratory
collaboration.
37
Fly ash samples (10 g) were fortified with 5 ng Cl-TCDD and
subjected to Soxhlet extraction with 100 mL benzene for 24 hours. The
benzene was concentrated just to dryness on a steam bath. The residue
was dissolved in 100 mL hexane and subjected to the previously described
acid/base extraction and clean-up procedures.
Quality Assurance Program: The QA program was initiated prior to
sample preparation and consisted of: (1) fortification of real and QA
37
samples with 2.5 to 10 ng of Cl-TCDD; (2) fortification of QA samples
with 0 to 1250 pg (0-125 ppt) TCDD; and (3) submission of real samples
and QA sample extracts (60 pl)to the GC/MS laboratories in a blind
fashion (i.e., there was no way to distinguish between QA and actual
samples). The efficiency, accuracy, precision and validity of TCDD
analyses were dependent upon the incorporated QA program.
*^C1-TCDD = labeled 2,3,7,8-^^Cl^-TCDD, isotopic purity greater than 98%.
12
-------
Criteria for Confirmation of TCDD Residues: The capillary column
GC/HRMS MIS analysis for 2,3,7,8-TCDD residues had to satisfy the criteria
(I-V) shown in Table 1 to be considered a confirmed positive sample.
The supplemental criteria (A) and (B) were occasionally applied to 10 to
100 ppt TCDD analyses.
TABLE 1
CRITERIA UTILIZED FOR CONFIRMATION OF 2,3,7,8-TCDD RESIDUES
I. Correct capillary column GC/HRMS retention time of 2,3,7,8-TCDD.
II. Correct responses for the co-injection of samples fortified with
Cl-TCDD and TCDD standard.
III. Correct chlorine isotope ratio of the molecular ion (m/z 320 and
m/z 322).
IV. Correct capillary column GC/HRMS multiple ion monitoring response
for TCDD masses (simultaneous response for elemental composition of
m/z 320, m/z 322, and m/z 328).
V. Response of m/z 320 and m/z 322 greater than 2.5 times noise level.
Supplemental criteria which were applied to highly contaminated sample
extracts:
(A) C0C1 loss indicative of TCDD structure, and
(B) Capillary column GC/HRMS peak-matching analysis of m/z 320 and m/z
322 in real time to confirm the TCDD elemental compositions.
13
-------
DISCUSSION AND RESULTS
The basic requirements for reliable quantitative trace organic
analyses are accurate fortification and quantification standards and
incorporated QA program, validated sample preparation procedures, and
GC-HRMS detection methods. The analyses are improved by multiple laboratory
participation, especially in the ppt concentration range. A primary
fortification and quantification standard must be used by all participating
laboratories to insure reasonable agreement in the quantitative results.
The performance of a laboratory or a group of laboratories may be then
determined by statistical evaluation of the analytical results.
Glass WCOT capillary columns provide total enhancement in the
qualitative and quantitative aspects of GC-MS analyses. Unique problems
can be encountered with WCOT capillary columns, compared with those
associated with packed or SCOT columns, but they can be easily resolved
or at least minimized. These problems are concerned with column load
capacity, the size of the sample injected, dead space, deterioration of
resolution caused by water or contamination, and other factors.
The direct-coupled GC-MS interface described in the Experimental
section insures maximum transfer efficiency and can effectively by-pass
1 to 10 pi of benzene solvent away from the MS. Therefore, the MS can
be optimized on 1- to 10-pg quantification standards and the sample can
be diluted or concentrated to fall within the selected quantification
range. This method provides the most efficient, accurate, and reliable
analysis because the peak heights of the standard and internal standard
14
-------
(co-injected with the sample) are constantly being evaluated during daily
operation. Sample matrix effects can cause a significant decrease in
sensitivity for the specific component because of gross amounts of
contamination, co-eluting components, etc. The utilization of 1- to
10-pg internal quantification standards requires a minimum amount of
sample and also corrects for any decrease in sensitivity due to sample
matrix effects.
Injection of large sample volumes (1- to 10-pl) is confined to cold
capillary columns (75°C or lower) in order to prevent damage to the
column. Hundreds of 1- to 10-pl sample dilutions have been injected
onto cold capillary columns without obvious deterioration of chromato-
graphic resolution or column life in these specific types of analyses.
A small amount (0.5 pi) of n-tetradecane is incorporated (co-injected)
as a "keeper" in all sample injections and improves the chromatographic
resolution and bandwidth of components.
Contamination, which may result in peak broading, can be minimized
by the use of silanized glass wool in the injection port. Discarding
the first 1 to 2 m of column and periodic cleaning the injection port
also minimizes and/or resolves this type of problem. Trace amounts of
water in the helium carrier gas also can destroy the chromatographic
resolution. The latter problem can be eliminated by use of a molecular
seive, which should be frequently changed, and the column can be rejuvinated
by repeated injections of methanol.
Quantification of TCDD: Peak height measurements for the mass m/z
328 of ^Cl-TCDD internal standard and the m/z 320 and m/z 322 (TCDD)
15
-------
37
mass of sample and the sample fortified with a Cl-TCDD and TCDD quanti-
fication standard are used to determine: (1) the percent recovery of
37
Cl-TCDD (sample preparation efficiency); (2) the residue level of
TCDD; and (3) the limit of detection for TCDD. Typical concentrations
37
of quantification standards are 50 to 250 pg/yl for Cl-TCDD and 1 to 5
37
pg/pl for TCDD. The Cl-TCDD % recovery value is used to correct the
TCDD residue value and the limit of detection for sample preparation
procedure efficiency.
Limit of Detection: The minimum limit of detection is defined as
the amount of TCDD that would provide clearly defined peak shapes for
the masses m/z 320 and m/z 322 in the proper isotopic ratio and with a
signal-to-noise ratio greater than 2.5:1. The sample weight and sample
size used in analysis, % recovery, sample matrix effects and the noise
present in the time frame of measurement will affect the minimum detection
limit.
Quantitative Determinations for TCDD in Fish: The fish samples
were collected in the state of Michigan (Tittabawassee, Grand and Saginaw
rivers and Saginaw Bay) under the direction of EPA Region V. The species
of fish were channel catfish, carp, yellow perch, small mouth bass, and
sucker. A chain-of-custody procedure was maintained from time of collection
through completion of analyses.
The edible portions (2.5 to 10 g) of the respective fish samples
were fortified with 2.5 to 10 ng of ^Cl-TCDD and subjected to acid/base
extraction and the normal clean-up procedure. A typical GC-HRMS MIS
analysis of an extract of ocean perch and the extract fortified with
16
-------
37
Cl-TCDD and TCDD is shown in Figure 2. Unusual high mass intensities
from high concentrations of chlorinated organic contaminants were encoun-
tered in samples of fish collected from polluted waters. These masses
differed from the exact mass of TCDD and were detected at m/z 320 and
m/z 322. The contaminant masses were not found in the analysis of fish
taken from other locations. The high concentrations of co-extractable
chlorinated components caused problems, such as capillary column overload,
co-elution of components and decreased MS sensitivity. To minimize
and/or cancel these effects, a very high MS sensitivity and small sample
size were necessary.
A small number of highly contaminated fish extracts were subjected
to the GC-HRMS analytical methods described previously and to the neutral
clean-up procedure to provide additional confirmation for the presence
of TCDD. The following criteria (Table 2) were used for verification of
results.
17
-------
s
_ <
"ill
1 1 '
©
1 ¦ 1 1 1 ¦ 1
© '
TCOD
Q-m/e 328
0We 322
_ ©-m/« 320
©.
-
,f « 1
fwuw/V*1 i/i, i
Av««g
o\
-r-r-T-r-i' 1—r-t—i—t— -i—»-•r*t
i!
•• ;i
--v-H >
1 11 , i (
' *>/'•'"X'At
-T—i—1—--T-T-J—
12 14 16 18 20 22 12 14 16 18 20 22
TIME (min)
Figure 2. Capillary column GC-HRMS Analysis of a QA extract of ocean
perch. (A) Sample. (B) Sample fortified with TCOD
quantification standard. The experimental results were: 52%
recovery of 37C1-TCDD; 34 ppt TCDD detected; 4 ppt TCDD detection
37
limit. The sample had been fortified with 5 ng CI-TCDD and
37 ppt TCDD.
18
-------
TABLE 2
SUPPLEMENTAL CRITERIA FOR CONFIRMATION OF RESULTS
(1) MIS simultaneous response of m/z 320, m/z 322, and m/z 324 to
confirm the tetrachloro isotope ratio.
(2) MIS simultaneous responses of m/z 320, m/z 322, m/z 257, and m/z
259 to confirm the M+-C0C1 loss, which is indicative of TCDD structure.
(3) Capillary column GC-HRMS peak-matching analyses at 10,000 mass
resolution in real time were used to confirm the elemental composition
of TCDD masses m/z 319.8965 and m/z 321.8936. Three exact masses
corresponding to TCDD isomers were also confirmed in these analyses.
High concentrations of contaminant masses differing from the exact
mass of TCDD were observed and measured to determine the elemental
composition. Tentative identifications were assigned to a number
of the components, PCBs (321.8677), chlorinated benyzlphenyl ethers
(319.9329), etc.
Four highly contaminated samples were subjected to the neutral
extraction and cleanup procedure. Capillary column GC-HRMS MIS analyses
37
yielded positive CI-TCDD and TCDD responses which were essentially
free from contamination. The neutral clean-up procedure was very effective
for this specific isolation of TCDD from other chlorinated compound
contamination; however, it has not been validated for quantitative TCDD
analysis by multiple laboratory participation.
A summary of analytical results for the quantitative measurement of
2,3,7,8-TCDD in fish is given in Table 3.
19
-------
TABLE 3
SUMMARY OF RESULTS FOR THE QUANTITATIVE DETERMINATION
OF TCDD RESIDUES IN FISH SAMPLES
1. Twenty-six of 35 samples contained detectable quantities of TCDD.
2. Eleven samples contained concentrations greater than 40 ppt.
3. Concentrations ranged from 4 ppt to 690 ppt.
4. The highest concentrations were detected in catfish and carp (bottom
feeders).
5. The lowest concentrations were detected in perch and bass (game
fish).
6. The highest concentration was detected in a catfish from the Titta-
bawassee River and the lowest concentration was detected in a perch
from the Saginaw Bay (commercial fishing waters).
7. A river dilution effect was evident from results obtained for fish
samples collected at specific points on the Tittabawassee River,
which flows into the Saginaw Bay.
8. Three components which satisfied the analytical criteria for the
other three TCDD isomers, with the exception of GC-HRMS retention
time, were detected in several fish samples at very low concentrations.
Typical analytical results for quality assurance samples used in
the analyses of the fish samples are shown in Table 4.
20
-------
TABLE 4
TYPICAL ANALYTICAL RESULTS FOR QUALITY ASSURANCE SAMPLES
GENERATED DURING THE ANALYSIS OF FISH FOR 2,3,7,8-TCDD RESIDUES
Sample Weight
Cg) and IDa
37C1-TCDD
% Recovery^
TCDD Detection
Limit (ppt)c
TCDD Detected
(ppt)c'd
TCDD
(pg)
Fortification
Level
(ppt)
5(1)
62
2
22
100
20
5(1)
52
4
34d
185
37
5(1)
82
3
ND
0
0
5(4)
100
3
ND
0
0
5(3)
54
1
19
70
14
5(3)
100
2
ND
0
0
5(3)
78
2
ND
0
0
5(1)
92
2
19
55
11
5(1)
97
5
45
240
48
10(1)
100+
1
8
130
13
10(1)
100+
4
43
600
60
10(2)
100+
7
ND
0
0
10(2)
100+
3
ND
0
0
10(1)
100+
4
76
1250
125
10(4)
67
1
ND
0
0
10(1)
93
3
56
650
65
10(1)
84
4
73
620
62
aKey: (1) Ocean Perch, (2) Lake Trout, (3) Beef Liver, (4) Method Blank
bEach sample had been fortified with 5 or 10 ng 37C1-TCDD.
Corrected for % recovery losses (^Cl-TCDD Mean % Recovery, 86%; TCDD Mean
% Accuracy, ± 15%).
^ND = Not Detected.
21
-------
Evaluation of QA results shown in Table 4 indicate that reasonably
accurate TCDD values were obtained in the presence of high concentrations
of chlorinated contamination. The number of QA samples with zero forti-
fication levels of TCDD were very important in these analyses because of
the large number of positive fish extracts and the unusual and high
concentrations of chlorinated contaminants. Accurate results were
obtained for a positive fish sample fortified with a known amount of
TCDD. The results shown in Table 3 were in reasonable agreement to
those reported in the Dow Chemical report "Trace Chemistries of Fire"^
for fish from similar locations. Kuehl et al.^ used negative chemical
ionization HRMS for the analysis of fish from various rivers and found a
unique distribution of polychlorinated dibenzo-p-dioxins, including
TCDD, in the Tittabawassee River. The estimated concentration of TCDD
in fish from a similiar location on the Tittabawassee River was within
experimental error to those reported in Table 3.
Based on results from three independent laboratories, it can be
concluded that the Tibbawawasee River watershed and Saginaw Bay commercial
fishing waters are contaminated with TCDD residues. The toxicological
significance of the trace levels of TCDD in fish is not known. The
source or sources of this contamination has been postulated but these
theories need to be validated by other studies; e.g., additional monitoring
studies on the Tittabawassee River and other similiar rivers in the U.S.
Quantitative Determinations for TCDD in Human Adipose Tissue: A
preliminary validation study was performed to determine the feasibility
22
-------
for the quantitative measurement of low ppt levels of TCDD in 250 milligram
samples of human tissue (equivalent to a needle biopsy). The samples
used were for quality assurance purposes and had been previously analyzed
for TCDD. However, this fact was not known in the sample preparation
laboratory prior to undertaking of the study. The results shown in
Table 5 suggest that reasonably accurate results may be generated from
250 mg samples. One obvious major discrepancy is apparent in the case
of the second sample listed. The small sample analysis indicated a
concentration of 16 ppt TCDD instead of the 5 ppt previously detected in
a 5 g sample of the sample tissue. Investigations are under way to
determine the reasons for this discrepancy. These include confirmation
of TCDD by multiple laboratory analysis and tracing the history of the
sample.
TABLE 5
FEASIBILITY STUDY FOR THE QUANTITATIVE DETERMINATION
OF TCDD IN QA TISSUE SAMPLES
TCDD Detection TCDD Detected TCDD Fortification
Limit (ppt) (ppt) Level
(pg) (ppt)
3 8 14.
5 16 0 0^
1 2 0.5 2C
1 3 2 8_
1 6 1.6 6
a
37
Cl-TCDD mean % recovery - 75%. Values are not corrected for
% recovery losses.
^Each 0.250 g sample was fortified with 0.5 ng ^Cl-TCDD.
cStandard solutions.
23
-------
Two microliters of quantification standard, consisting of 12.5
37
pg/pl Cl-TCDD and 0.5 pg/pl TCDD, were utilized in connection with the
human fat analyses. An MIS response is shown in Figure 3. Capillary
column resolution of components and a mass resolution of 8000 was sufficient
to resolve 2,3,7,8-TCDD from other chlorinated compounds (e.g., DOE and
PCBs). Future analyses will utilize 0.25 pg quantification standards
and a mass resolution of 9,500 as a result of contamination encountered
in this specific sample and the need to lower the detection limit.
Quantitative Search for TCDD in Fly Ash: Representative samples of
fly ash from coal-fired power plants were subjected to a specific extraction
(24 hour Soxhlet extraction utilizing benzene) and clean-up procedure
prior to analysis for TCDD residues. The extraction of organic compounds
from fly ash (carbon) is extremely difficult. The results shown in
Table 6 show that TCDD was not detected at an average detection limit of
2 ppt. These results are in agreement with the findings reported by
18
Kimble and Gross and suggest that TCDDs are not formed in detectable
quantities (>2 ppt) in highly efficient coal-fired combustion processes.
24
-------
ilpfWci-TCOO
¦ S pt^Ct—TCDO
TCOO
1aW*32B
2m/< 322
Snfe 320
Figure 3. GC-HRMS Multiple Ion Selection Chromatogram for a Quantification
Standard, 1 pg TCDO and 25 pg 3^C1-TCDD.
25
-------
TABLE 6
QUANTITATIVE DETERMINATIONS FOR TCDD IN FLY ASH
37ci-tcdd
TCDO Detection
TCDD Detected0
% Recovery
Limit0 (ppt)
(ppt)
59
2.5
10d
67
2.2
ND
60
2.8
ND
80
1.3
ND
80
1.0
ND
79
1.5
ND
74
1.3
ND
71
2.2
ND
aEach 5 g sample was fortified with 2.5 ng 37C1-TCDD.
^Mean % recovery = 71%.
Corrected for % recovery losses.
^QA sample fortified with 8 ppt TCOD.
26
-------
Contamination: Contamination from chlorinated compounds is a
constant problem in ppt TCDO analyses. Examples of the types of chlori-
nated compounds, their elemental compositions, exact masses, and the
approximate mass resolution required to effect separation from TCDD
masses are shown in Table 7.
TABLE 7
CONTAMINATION
COMPOUND
ELEMENTAL
COMPOSITION
EXACT
MASS
REQUIRED
MASS
RESOLUTION
TCDD
ELEMENTAL
.COMPOSITION
TCDD
EXACT
MASS
TETRACHLORO-
BENZYLPHENYL-
ETHER
OOE
C13H0CI4
cunfWc*
319.9329
319.9321
9,000
9,000
Ci2H40235Cl4
319.8965
PCB HEPTACHLORO-
BIPHENYL FRAGMENT
C12H33Sci5
321.8678
12,000
Ci2H40235CI337CI
321.8935
PCB PENTACHLORO-
BIPHENYL
C12H535Cl3^Cl2
327.8775
45,000
Ci2H40237CU
327.8847
RESOLUTION « —
AM
In general, PCBs cause the most serious problems because of distortion
of the m/z 320 and m/z 322 chlorine isotope ratios. Fortunately, the
sample preparation procedure is quite specific for removal of PCBs. In
cases of gross PCB contamination, the MS mass resolution can be easily
increased to ca. 14,000 to effect separation at m/z 322. In addition,
the PCB m/z 326 mass can be utilized to calculate the contribution to
27
-------
m/z 328, which is a mixture of PCB and TCDD mass ions. The minimum
detection limit in this analysis is higher than normal, however.
Utilizing the criteria shown in Table 1 for confirmation of TCDD
residues, the chlorinated compounds shown in Table 7 do not present
serious problems in TCDD analysis except in rare cases of gross contami-
nation. Neutral clean-up procedures, mass resolution of 14,000, and
polar or non polar capillary columns effectively resolve this type of
problem.
The occasional and unexpected sample containing high ppt to ppm
levels of TCDD may cause serious contamination problems with other TCDD
samples prepared in the laboratory. Erroneous or low ppt TCDD residues
may be detected in succeeding samples as a result, even when extremely
meticulous glassware cleaning procedures are exercised. Laboratory
records, good quality assurance practices, and multiple laboratory
participation should detect and eliminate these problems.
TCDD ISOMERS: Recent reports^'have shown that chlorinated
dioxins, including 2,3,7,8-TCDD and other TCDD isomers, may be formed in
combustion processes. The toxicological properties of the various TCDD
isomers are known to be significantly different; however, the mass
19
spectra are almost identical. Therefore, it is extremely important
that TCDD isomers be resolved by utilization of glass capillary columns
prior to their introduction into the MS to assure the most conclusive
identification of a specific isomer. 2,3,7,8-TCDD and all other available
isomers may be separated, detected and/or quantified by the technique
described here. However, only a limited number of the 22 theoretical
28
-------
TCDD isomers are commercially available. Compounds utilized in this
laboratory for co-injection purposes are the 1,2,3,4-, 1,3,6,8-, 2,3,6,8-,
and 2,3,7,8-tetrachlorodibenzo-p-dioxins. Confirmation of 2,3,7,8-TCDD
in the presence of other TCDD isomers also requires analysis on a second
on o-i
capillary column of different polarity to differentiate between isomers. '
Most toxicological studies have been devoted to the 2,3,7,8-TCDD
isomer and the other commercially available isomers. Therefore, the
toxicological properties of all of the other TCDD isomers are not known.
Unidentified TCDD isomers have been detected and confirmed in specific
types of samples through the use of capillary column GC-HRMS techniques.
The significance of these findings (ppt to ppm of unidentified TCDD
isomers) should be determined in the future.
Analysis of PCB Incineration Emissions for Chlorinated Piben2ofurans:
The ambient air samples were collected in the vicinity of a high temperature
incineration during a test burn of PCBs. These samples were collected
22
with the high-volume sampler developed by Lewis et al. and were subjected
22
to an extraction and cleanup procedure specific for PCB residues. The
concentrated extracts were then submitted to capillary column GC-HRMS
analysis for polychlorinated dibenzofurnan (PCDFs). Theoretically, 135
PCDF isomers are possible. 2,3,7,8-tetrachlorodibenzofuran (2,3,7,8-TCDF)
is the most toxic of the 38 TCDF isomers and is one of the major TCDF
isomers formed in combustion of the PCBs, Aroclors 1254 and Aroclor
23
1260. Only one TCDF isomer, 2,3,7,8-TCDF, was available to this
laboratory. Therefore, the following procedures were utilized to rapidly
screen the complex extracts for TCDF and PCDF residues.
29
-------
The GC-HRMS parameters were established and optimized with 10- to
20-pg 2,3,7,8-TCDF standards and the procedures discussed earlier for
TCDD. The criteria utilized for confirmation of TCDF residues and/or
establishment of minimum detection limits were:.
1. Capillary column GC-HRMS retention time of 2,3,7,8-TCDF.
2. Correct responses for the co-injection of sample and 2,3,7,8-TCDF
standard.
3. Capillary column GC-HRMS multiple ion simultaneous response
(m/z 303.9016 and m/z 305.8986).
4. Molecular ion chlorine isotope ratio (m/z 304 and m/z 306).
5. M/z 304 and m/z 306 MS response greater than 2.5 times noise
level.
The capillary column GC-HRMS retention time of 2,3,7,8-TCDF was 13
minutes ±15 seconds. The MS mass resolution (8,500) was sufficient to
resolve TCDF from contamination.
The sample extracts (300 pi each) were analyzed utilizing the
capillary column GC-HRMS multiple ion selection technique. These analyses
35
were performed on the TCDF molecular ion (m/z 303.9016, C^H^O Cl^) and
M + 2 ion (m/z 305.8986, C12H4035C1337C1) using PFK m/z 292.9825 as the
reference standard. The sequence for each sample analysis was: actual
sample, then sample fortified with 10 or 20 pg of 2,3,7,8-TCDF. The
3
isomers of TCDF and 2,3,7,8-TCDF were not detected at 1 to 10 pg/m
detection limits.
Capillary column GC-HRMS peak matching analysis in real time was
utilized to monitor the concentrated extracts for the molecular ions of
30
-------
tri- through octa-chlorodibenzofuran. These PCDF residues were not
The sample collection and analytical extraction efficiency is not
known. Therefore the PCOF results must be considered as semi-quantitative.
Extremely high concentrations of chlorinated dicyclopentadienes (I
and II), shown in Figure 4, were identified (in low resolution mass
spectra and HRMS analysis) and determined to be major interfering compounds
in the PCOF analysis. Analyses are being performed on the incinerator
gases to confirm the PCOF results and the possible source of compounds I
and II.
Figure 4. Structures of Chlorinated Dicyclopentadines extracted from
3
detected at an estimated detection limit of 100 pg/m .
I
II
air.
31
-------
CONCLUSION
Glass capillary columns provide a practical and effective solution
to the separation and qualitative and quantitative characterization of
trace level components in complex media. These validated results indicate
that a specific portion of our environment and food chain are contaminated
with low concentrations of TCDD residues. The toxicological significance
of these results is not known at this time.
The future objectives of the dioxin program within this laboratory
include: (1) the application and improvement of the capillary column
GC-HRMS techniques and sample preparation procedures described here for
the quantitative determination of tetra-, penta-, hexa-, hepta-, and
octa-chlorinated dibenzo-p-dioxins in 100- to 200-mg needle biopsy
samples of human or animal tissue (bioaccumulation and toxicological
effects in animals may then be closely followed), and (2) application of
capillary column GC-HRMS negative ionization (chemical ionization and
electron impact) methods of analysis for the determination of ppt levels
of chlorinated dibenzo-p-dioxins and other toxic compounds of interest
to EPA.
ACKNOWLEDGMENTS
We wish to thank Mike Dellarco, EPA Washington, D.C. for his help
in the overall coordination of the dioxin program; Karl Bremer and Lyman
Condie, EPA Region V, for overseeing the collection of fish samples; and
Aubry Dupuy, James Gibson, and Henry Shoemaker, Toxicant Analysis Center,
Bay St. Louis, Mississippi, for extraction and clean-up of samples for
TCDD analysis.
32
-------
REFERENCES
1. Wien, W., Ann. Physik, 65, 440 (1898).
2. Thompson, J. J., Phil Mag., 10, 548 (1905).
3. Dempster, A. J., Phys. Rev. , 11, 316 (1918).
4. Aston, F. W. , Phil. Mag., 38, 707 (1919).
5. Watson, J. T. and Biemann, K., Anal. Chem., 36, 1135 (1964); 37,
844 (1965). ~~ ~
6. Becker, E. W., in "Separation of Isotopes," H. London, ed., George
Newnes, publisher, London, 1961, p. 360.
7. Ryhage, R., Anal. Chem., 36, 759 (1964).
8. Llewellyn, P. M. and Littlejohn, D. P., Pittsburgh Conference on
Analytical Chemistry and Applied Spectroscopy, 1966. Technical
Information Quarterly, Varian, Palo Alto, Calif., Spring 1966.
9. Gohlke, Anal. Chem., 31, 535 (1959).
10. Harless, R. L. and Oswald, E. 0., (Paper No. RP-7), 26th Annual
Conf. on Mass Spectrometry and Allied Topics, St. Louis, Mo., June
1978.
11. Ryan, J. F., Biros, F. J. and Harless, R. L., (Paper No. C-2) 22nd
Annual Conf. on Mass Spectrometry and Allied Topics, Philadelphia,
PA, June 1974.
12. Harless, R. L. and Oswald, E. 0., in "Dioxin, Toxicological and
Chemical Aspects," F. Cattabeni, A. Cavallaro and G. Galli, eds.,
John Wiley & Sons, publisher, 1977, p. 5157.
13. Harless, R. L. and Oswald, E. 0., 23rd Collaborative Intnat'l.
Pesticide Analytical Council, Baltimore, Md., June 1979.
14. Harless, R. L., Oswald, E. 0., Wilkinson, M. K., Dupuy, A. E., Jr.,
McDaniel, D. D. and Tai, Han, Anal. Chem., 52 (in press)
15. Neuner-Jehle, N., Etzwieler, F. and Zarske, G., Chromatoqraphia, 6,
211 (1973).
16. Kuehl, 0. W., Dougherty, R. C., Tondeur, Y., Stalling, D. L.,
Smith, L. M. and Rappe, C., Negative Chemical Ionization Studies of
Polychlorinated Dibenzo-p-Dioxins, Dibenzofurans and Naphthalenes
in Environmental Samples, Unpublished report, 1979.
33
-------
17. The Chlorinated Dioxin Task Force, The Michigan Division, Dow
Chemical USA. The Trace Chemistries of Fire - A Source of and
Routes for the Entry of Chlorinated Dioxins Into the Environment,
1978.
18. Kimble, B. J. and Gross, M. L., Science, 207, 59 (1980).
19. Buser, H. R., Anal. Chem., 49, 918 (1977).
20. Buser, H. R., Bosshardt, H. P. and Rappe, C., Chemosphere^ 7, 165
(1978).
21. Olie, K., Vermeulen and Hutzinger, 0., Chemosphere, 6, 455 (1977).
22. Lewis, R. 6., Brown, A. R. and Jackson, M. D., Anal. Chem., 49,
1668 (1977).
23. Buser, H. R., Bosshardt, H. P., 23rd Collaborative Intnat'l.
Pesticide Analytical Council, Baltimore, Md., June 1979.
34
------- |