United States
Environmental Protection
Agency
Office of
Toxic Substances
Washington DC 20460
EPA-560/5-86-020
Octooer, 1986
Toxic Substances
vvEPA
ANALYSIS FOR
POLYCHLORINATED
DSBENZO- p-DIOXINS (PCDD)
AND DIBENZOFURANS (PCDF)
IN HUMAN ADIPOSE TISSUE:
METHOD EVALUATION STUDY
s
c
(3
08
"c
O
C
O
•Q
95 % Confidence Limits
for Individual Analyses
50
Spiked Concentration (pg/g)
-------
ANALYSIS FOR POLYCHLORINATED DIBENZO-g-DIOXINS (PCDD) AND DIBENZOFURANS (PCDF)
IN HUMAN ADIPOSE TISSUE: METHOD EVALUATION STUDY
by
John S. Stanley, Randy E. Ayling, Karin M. Bauer, Michael J. McGrath,
Thomas M. Sack, and Kelly R. Thornburg
FINAL REPORT
EPA Prime Contract No. 68-02-3938
Work Assignment No. 46
MRI Project No. 8501-A(46)
and
EPA Prime Contract No. 68-02-4252
Work Assignment No. 24
MRI Project No. 8824-A(01)
Field Studies Branch, TS-798
Office of Toxic Substances
U.S. Environmental Protection Agency
401 M Street, S.W.
Washington, DC 20460
Attn: Ms. Janet Remmers, Work Assignment Manager
Dr. Joseph J. Breen, Project Officer
-------
DISCLAIMER
This document has been reviewed and approved for publication by the
Office of Toxic Substances, Office of Pesticides and Toxic Substances, U.S.
Environmental Protection Agency. The use of trade names or commercial prod-
ucts does not constitute Agency endorsement .or recommendation for use.
-------
PREFACE
This report provides a summary of the results from a method eval-
uation study for the determination of 2,3,7,8-substituted polychlorinated
dibenzo-p_-dioxins (PCDD) and dibenzofurans (PCDF) in human adipose tissues
at the parts-per-trillion (ppt) level. This method evaluation is an integral
part of a collaborative program between the U.S. Environmental Protection
Agency's Office of Toxic Substances and the Veterans Administration to deter-
mine if significant differences exist in the 2,3,7,8-substituted PCDD and/or
PCDF levels in human adipose tissues for Vietnam veterans compared to the
general adult male population. The study design will focus on specimens
within EPA's National Human Adipose Tissue Survey (NHATS) repository. The
method evaluation described in this report was necessary to establish method
performance (accuracy and precision) before proceeding with actual sample
analysis.
This method evaluation study was completed under EPA Contract Nos.
68-02-4252, Work Assignment 24 and 68-02-3938, Work Assignment 46, "Analysis
for Dioxins and Furans in Human Adipose Tissue," Ms. Janet Remmers, Work
Assignment Manager, and Dr. Joseph Breen, Project Officer.
MIDWEST RESEARCH INSTITUTE
aul C. Constant
Program Manager
Approved:
Jack Balsinger
Quality Assurance Coordinator
Joffin E. Going, Director
Chemical Sciences Department
October 30, 1986
-------
TABLE OF CONTENTS
I. Introduction 1
II. Summary 3
III. Recommendations 4
IV. Experimental 5
A. Preparation of Homogenized Tissue 5
B. Analytical Standards. _ 5
1. Calibration Standards 6
2. Spiking Solutions 6
C. Analytical Procedure 10
D. HRGC/MS Analysis 12
E. Data Interpretation 17
1. Qualitative 17
2. Quantitation 17
F. Quality Assurance/Quality Control (QA/QC) 20
G. Preliminary Method Studies 20
1. Gravimetric Studies 21
2. Carbon-14 Recovery Studies 21
V. Results 23
A. Analytical Results 23
B. Statistical Analysis 23
1. Recovery of Internal Quantitation Standards. . 57
2. Estimation of Background Levels of PCDDs and
PCDFs 57
3. Day-to-Day HRGC/MS Analysis Precision 59
VI. Quality Assurance/Quality Control (QA/QC) 63
A. Initial Calibration 63
B. Daily Verification of Response Factors 65
C. Blanks 65
D. Absolute Recoveries of the Internal Quantitation
Standards 77
VII. Glossary of Terms 82
VIII. References 84
Appendix A - Analytical Protocol for Determination of PCDDs and
PCDFs in Human Adipose Tissue A-l
-------
LIST OF FIGURES
Figure
1 Comparison of the HRGC/MS-SIM reconstructed ion chromato-
gram (RIC) from the analysis of unspiked homogenized
human adipose tissue matrix and a calibration standard
for PCDDs and PCDFs 30
2 Example of the TCDF (m/z 304) and TCDD (m/z 320)
HRGC/MS-SIM elution profiles in unspiked and spiked
human adipose 31
3 Example of the PeCDF (m/z 338) and PeCDD (m/z 354)
HRGC/MS-SIM elution profiles in unspiked and spiked
human adipose 32
4 Example of the HxCDF (m/z 374) and HxCDD (m/z 390)
HRGC/MS-SIM elution profiles in unspiked and spiked
human adipose 33
5 Example of the HpCDF (m/z 408) and HpCDD (m/z 424)
HRGC/MS-SIM elution profiles in unspiked and spiked
human adipose 34
6 Examples of the OCDF (m/z 442) and OCDD (m/z 458)
HRGC/MS-SIM elution profiles in unspiked and spiked
human adipose 35
7 Measured concentrations versus concentrations of 2,3,7,8-
TCDD spiked into the homogenized human adipose lipid
matrix 36
8 Measured concentrations versus concentrations of
1,2,3,7,8-PeCDD spiked into the homogenized human
adipose lipid matrix 37
9 Measured concentrations versus concentrations of
1,2,3,4,7,8-HxCDD spiked into the homogenized human
adipose lipid matrix 38
10 Measured concentrations versus concentrations of
1,2,3,6,7,8-HxCDD spiked into the homogenized human
adipose lipid matrix 39
11 Measured concentrations versus concentrations of
1,2,3,7,8,9-HxCDD spiked into the homogenized human
adipose lipid matrix 40
12 Measured concentrations versus concentrations of
1,2,3,4,6,7,8-HpCDD spiked into the homogenized human
adipose lipid matrix 41
-------
LIST OF FIGURES (continued)
Figure
13 Measured concentrations versus concentrations of OCDD
spiked into the homogenized human adipose lipid matrix. . 42
14 Measured concentrations versus concentrations of
2,3,7,8-TCDF spiked into the homogenized human
adipose lipid matrix 43
15 Measured concentrations versus concentrations of
1,2,3,7,8-PeCDF spiked into the homogenized human
adipose lipid matrix 44
16 Measured concentrations versus concentrations of
2,3,4,7,8-PeCDF spiked into the homogenized human
adipose lipid matrix 45
17 Measured concentrations versus concentrations of
1,2,3,4,7,8-HxCDF spiked into the homogenized human
adipose lipid matrix 46
18 Measured concentrations versus concentrations of
1,2,3,6,7,8-HxCDF spiked into the homogenized human
adipose lipid matrix 47
19 Measured concentrations versus concentrations of
2,3,4,6,7,8-HxCDF spiked into the homogenized human
adipose lipid matrix 48
20 Measured concentrations versus concentrations of
1,2,3,7,8,9-HxCDF spiked into the homogenized human
adipose lipid matrix 49
21 Measured concentrations versus concentrations of
1,2,3,4,7,8,9-HpCDF spiked into the homogenized human
adipose lipid matrix 50
22 Measured concentrations versus concentrations of
1,2,3,4,6,7,8-HpCDF spiked into the homogenized human
adipose lipid matrix 51
23 Measured concentrations versus concentrations of OCDF
spiked into the homogenized human adipose lipid matrix. . 52
24 Method accuracy estimates as determined from the slopes
of the least squares regression lines for the 17
target PCDD and PCDF analytes 54
25 Control charts showing response factors by date for
2,3,7,8-TCDF and 2,3,7,8-TCDD 66
vi i
-------
LIST OF FIGURES (continued)
Figure Page
26 Control charts showing response factors by date for
1,2,3,7,8-PeCDF and 2,3,4,7,8-PeCDF 67
27 Control chart showing response factors by date for
1,2,3,7,8-PeCDD 68
28 Control charts showing response factors by date for
1,2,3,4,7,8-HxCDF and 1,2,3,6,7,8-HxCDF 69
29 Control charts showing response factors by date for
2,3,4,6,7,8-HxCDF and 1,2,3,7,8,9-HxCDF 70
30 Control charts showing response factors by date for
1,2,3,4,7,8-HxCDD and 1,2,3,6,7,8-HxCDD 71
31 Control chart showing response factors by date for
1,2,3,7,8,9-HxCDD 72
32 Control charts showing response factors by date for
1,2,3,4,6,7,8-HpCDF and 1,2,3,4,7,8,9-HpCDF 73
33 Control chart showing response factors by date for
1,2,3,4,6,7,8-HpCDD 74
34 Control charts showing response factors by date for OCDF
and OCDD 75
vm
-------
LIST OF TABLES
Table Page
1 Analytical Standards Available for the Method
Evaluation Studies 7
2 Concentration Calibration Solutions 8
3 Nat-ive PCDD and PCDF Spiking Solution 9
4 Internal Standard Spiking Solutions 11
5 HRGC/LRMS Operating Conditions for PCDD/PCDF Analysis ... 13
6 Ions Monitored for HRGC/MS Analysis of PCDD/PCDF 14
7 Typical Daily Sequence for PCDD/PCDF Analysis 16
8 Ion Ratios for HRGC/MS Analysis of PCDD/PCDF 18
9 Summary of the Results of the Sample Preparation Method
Evaluation Using Carbon-14 PCDDs 22
10 Spiked Versus Measured Concentrations of 2,3,7,8-TCDF
and 2,3,7,8-TCDD in Homogenized Human Adipose Lipid
Samples 24
11 Spiked Versus Measured Concentrations of 1,2,3,7,8-PeCDF,
2,3,4,7,8-PeCDF, and 1,2,3,7,8-PeCDD in Homogenized
Human Adipose Tissue Samples 25
12 Spiked Versus Measured Concentrations of 1,2,3,4,7,8-;
1,2,3,6,7,8-; 2,3,4,6,7,8-; and 1,2,3,7,8,9-HxCDF
in Homogenized Human Adipose Lipid Matrix 26
13 Spiked Versus Measured Concentration of 1,2,3,4,7,8-;
1,2,3,6,7,8-; and 1,2,3,7,8,9-HxCDD in Homogenized
Human Adipose Lipid Samples 27
14 Spiked Versus Measured Concentrations of 1,2,3,4,6,7,8-
HpCDF, 1,2,3,4,7,8,9-HpCDF, and 1,2,3,4,6,7,8-HpCDD
in Homogenized Human Adipose Lipid Samples 28
15 Spiked Versus Measured Concentrations of OCDF and OCDD
In Homogenized Human Adipose Lipid Samples 29
16 Regression Line Slopes with 95% Confidence Limits 55
17 Results of the Analysis of the Low and High Level
Native Spike Solutions 56
-------
LIST OF TABLES (continued)
Table
18
19
20
21
22
23
24
25
Background Level Estimates with 95% Confidence Limits . . .
Day-to-Day Precision of Analysis of Specific Sample
Extracts for Tetra- and Pentachloro PCDF and PCDD ....
Day-to-Day Precision of Analysis of Specific Sample
Extracts for Hexa- and Heptachloro PCDF and PCDD
Day-to-Day Precision of Analysis of Specific Sample
Extracts for OCDF and OCDD
Relative Response Factors (Grand Means) Determined from
Multipoint Concentration Calibration Standards
Summary of Results from the Analysis of a Laboratory
Method Blank
Recovery of Radiolabeled PCDDs from Precleaned Activated
Alumina
Absolute Recoveries of the Internal Quantitation Standards
from the Human Adioose Li Did Matrix
Paqe
58
60
61
62
64
76
78
79
26 Recovery of Carbon-14 Labeled 2,3,7,8-TCDD, 1,2,3,4,7,8-
HxCDD, and OCDD as a Function of Final Concentration
Conditions 81
-------
I. INTRODUCTION
The Environmental Protection Agency Office of Toxic Substances
(EPA/OTS) and the Veterans Administration (VA) have established an interagency
agreement to study the level of polychlorinated dibenzo-p_-dioxins (PCDDs) and
dibenzofurans (PCDFs) in human adipose tissues. The occurrence and levels of
PCDDs and PCDFs with chlorine substitution in the 2,3,7,8 positions (especially
2,3,7,8-TCDD) of the parent molecules are of primary interest.
As part of this interagency effort, it has been proposed to use
selected adipose tissue samples that were collected for the Field Studies
Branch (FSB) of EPA's Office of Toxic Substances (OTS) through the National
Human Adipose Tissue Survey (NHATS) to determine exposure to PCDDs and PCDFs.
The available adipose tissues include specimens obtained from young men whose
age indicates that they could have served in Vietnam and could have been ex-
posed to Agent Orange. The tissues were originally collected as part of a
broadly based and statistical random sampling of the continental United States.
The analysis of these tissues may provide information on the differences of
exposure of the general adult male population and Vietnam veterans to the
2,3,7,8-substituted PCDDs and PCDFs.
The overall objectives of the proposed EPA/VA collaborative studies
are:
1. Evaluate the reliability, accuracy, precision, and sensitiv-
ity of a proposed method for the determination of 2,3,7,8-
substituted PCDDs and PCDFs (tetra- through octachloro
homologs) in human adipose tissue at the parts-per-tril1 ion
(ppt) level.
2. Determine if these compounds can be detected in adipose tissues
of the American male adult population; and
3. Determine if individuals with military service in Vietnam have
significantly different levels of 2,3,7,8-substituted PCDDs
and PCDFs (particularly 2,3,7,8-TCDD) than other American men.
As a prelude to this work assignment, MRI conducted an extensive
literature review of applicable analytical methods and conducted a meeting
with recognized experts in this field to identify critical aspects of analyt-
ical methodology.1'2
Based on the information gathered through the literature review and
the meeting with the recognized experts, a special report was prepared for
OTS proposing a framework for an analytical method for analysis of human adi-
pose tissues.3 Several studies have been completed since the issuance of that
report which reflect the advances in analytical techniques for adipose tissue
analysis.4 16 The salient features of these methods have been combined into
a single protocol for the routine analysis of tetra- through octachloro PCDDs
and PCDFs at the low-parts-per-tril1 ion level for the EPA/VA tissue study.
This report focuses on a method evaluation study that was conducted
to achieve the first objective of the interagency agreement. Clarification
-------
of method performance is necessary before proceeding with the analysis of
actual samples retrieved from the NHATS repository.
This report includes a summary of the method evaluation study re-
sults (Section II). Recommendations to be implemented before proceeding with
the actual tissue samples from the NHATS repository are presented in Section
III. A description of the actual experimental procedures is provided in Sec-
tion IV. Results of sample analyses are summarized in Section V, and quality
assurance/quality control (QA/QC) aspects of the study are detailed in Section
VI. Pertinent references are listed in Section VII. Appendix A contains the
detailed analytical protocol that will be followed for the analysis of the
NHATS specimens designated in the study design to be provided by EPA/VA.
-------
II. SUMMARY
The results of the replicate analysis of spiked and unspiked homog-
enized human adipose tissue matrix demonstrate that the analytical method
produces accurate and precise data for 17 specific 2,3,7,8-substituted PCDD
and PCDF (tetra- through octachloro homologs) compounds. Accuracy of the
analytical method was demonstrated to range from 90 to 120% for the 17
2,3,7,8-substituted PCDD and PCDF compounds. Data are reported for three or
four replicate analyses of samples spiked at three different concentration
levels. The endogenous or background levels of the PCDD and PCDF congeners
in the homogenized adipose lipid matrix were estimated through regression
analyses of measured (found) versus spiked concentrations for each compound.
The analytical method is capable of providing quantitative data for
tetra- through octachloro PCDD and PCDF congeners to concentration levels as
low as 1 pg of the tetrachloro congeners per gram of adipose tissue. However,
an interference was noted at m/z 304 which coeluted with 2,3,7,8-TCDF, result-
ing in a detection level of approximately 4 pg/g.
Average absolute recoveries of the internal quantisation standards
ranged from 52% for 13C12-TCDD up to 89% for 13C12-OCDD. The agreement of
the measured concentrations versus the spiked concentrations for each PCDD
and PCDF congener demonstrates that the internal standard quantitation proce-
dure provides an accurate measure of concentration which is independent of
the absolute recovery.
Final concentration conditions were noted to have pronounced effect
on the absolute recoveries of the lower chlorinated compounds, particularly
2,3,7,8-TCDD. Experiments with carbon-14 labeled 2,3,7,8-TCDD demonstrated
that final concentration at temperatures of 55 to 60°C resulted in recoveries
as low as 54% while the same procedure conducted at ambient conditions re-
sulted in greater than 90% recovery.
Analysis of method and reagent blanks provided information on po-
tential artifacts in the sample preparation scheme. Additional experiments
were conducted with carbon-14 labeled PCDDs to evaluate the cleanup efficiency
and recovery of PCDDs from chromatographic materials, particularly acidic
alumina.
-------
III. RECOMMENDATIONS
Some minor modifications have been made in the written protocol
(Appendix A) that were not included in this phase of the method validation.
These include:
a cleanup procedure for activated acidic alumina prior to
fractionation of sample extracts to remove artifacts; and
final concentration of the sample extracts using nitrogen
blowdown at room temperature rather than heating to 55-60°C.
The spiking solutions used to prepare the spiked quality control
samples should be submitted for replicate (minimum of three/per spike level)
HRGC/MS analysis to assist the interpretation of positive or negative bias in
the accuracy of QC sample data.
The accuracy bounds should be extended to 50-130% from 50-115% as
specified in the draft quality assurance program plan.
The method should include additional internal quantisation standards
to pair with the HpCDF and OCDF congeners. Also, an additional internal recov-
ery standard, possibly 13C12-l,2,3,4,7,8-HxCDD, is required to provide better
estimates of absolute method recovery. These additional compounds, if avail-
able, will be incorporated into the method before initiating sample analyses.
Analysis for 2,3,7,8-TCDF may require high resolution mass spec-
trometry to avoid interferences occurring at m/z 304. This will require
modification of the HRGC/HRMS portion of the protocol to include the spe-
cific acquisition parameters for the characteristic ions of 2,3,7,8-TCDF.
-------
IV. EXPERIMENTAL.
A. Preparation of Homogenized Tissue
A bulk lipid sample was prepared from the extracts of human adipose
tissue samples collected through the NHATS program. The adipose tissue sam-
ples have been stored in a deep freezer at approximately -10°C since collec-
tion. The homogenized tissue extract or bulk lipid was used in this method
evaluation study for preparation of replicate samples spiked with varying
levels of specific PCDD and PCDF isomers. This homogenized matrix will also
be used for preparing control and spiked quality control samples for the
actual NHATS sample analysis phase of the program.
A total of 2,465 g of adipose tissue was extracted, dried, and
concentrated to yield 1,652 g (62% of original weight) of homogenized lipid.
Specific procedures for preparing this matrix are described below.
The adipose tissue samples were thawed at room temperature for I to
2 h. Portions of the samples were added to a blender cup of a Waring® blender
and covered with methylene chloride. The volume of methylene chloride was
approximately equal to the sample volume (100 to 200 ml). This mixture was
blended at high speed for approximately 10 min, and the contents were trans-
ferred to a 500-mL Erlenmeyer flask and further blended with a Tekmar®
Tissumizer, also at high speed for 10 min. A powder funnel was plugged with
a wad of glass wool (silanized, methylene chloride extracted) and filled with
~ 50 g of sodium sulfate (heated overnight to 600°C in a muffle furnace).
The sodium sulfate was wetted with methylene chloride prior to elution of the
sample extract. The dried effluents were refiltered in the same way using a
fresh bed of sodium sulfate to remove particulate and residual water.
The samples were transferred to 1-L round bottom flasks, and the
solvent was removed by rotary evaporation. The water bath on the rotary evap-
orator was kept at 60°C using a thermostatted heating element. Once the sol-
vent appeared to have been removed (constant volume in flask, no visible con-
densation in condenser), the heating and evaporation process was continued
for at least 2 h. The flask and contents were removed and stored in a refrig-
erator. The extracted lipid solidified upon refrigeration and was visually
checked for homogeneity. No precipitates or phase separation was observed.
The lipid residue was allowed to liquify at room temperature and was trans-
ferred to a 4-L glass bottle with a Teflon®-lined lid.
The lipid residue was brought to room temperature and heated just
enough to allow the lipid to achieve an oily state prior to aliquotting por-
tions for the method evaluation studies.
B. Analytical Standards
Analytical standards including native PCDD and PCDF congeners,
stable isotope (carbon-13) labeled standards and radiolabeled (carbon-14)
standards were purchased from Cambridge Isotope Laboratories, Woburn,
Massachusetts, and Pathfinder Laboratories, St. Louis, Missouri. The 2,3,7,8-
TCDD was received from the EPA Reference Materials Branch as a solution in
-------
isooctane. The other native PCDD and PCDF congeners were received as 1-mg
neat standards. The stable and radiolabeled isotopes were received as solu-
tions in n-nonane or isooctane and toluene, respectively. Table 1 provides a
summary of the standards used for this study.
Stock solutions of the individual PCDD and PCDF congeners were pre-
pared from the neat standards. The neat materials were weighed using a Cahn
27 electrobalance calibrated versus a 1-mg (Class M) standard. The neat com-
pounds were transferred to glass vials and were dissolved in 2.0 to 3.0 ml of
toluene (Burdick and Jackson, distilled in glass). Toluene was added to each
standard using volumetric pipettes (Class A). The OCDD required dilution to
10.0 mL using a 50:50 mixture of toluene and anisole.
A working solution consisting of the 17 native PCDD and PCDF con-
geners was prepared in toluene at a concentration of 2 ug/mL for the TCDD,
TCDF, PeCDD, and PeCDF congeners, 5 vq/ml for the HxCDD, HxCDF, HpCDD, and
HpCDF congeners, and 10 ug/mL for the OCDD and OCDF. The working solution
was used to prepare both the lipid matrix spiking solution and the calibra-
tion standards.
The stable isotope labeled internal standards were obtained as so-
lutions in ji-nonane or isooctane at 50 ug/mL concentration with the exception
of the 13C12-OCDD, which was provided at 10 ug/mL. Separate working solutions
containing mixtures of the carbon-13 labeled PCDDs and PCDFs were prepared in
isooctane for use in the calibration standards and the sample spiking solutions.
The carbon-14 radiolabeled PCDDs were used for preliminary method
evaluation studies. The specific activity of the 14C-2,3,7,8-TCDD (117.56
mCi/mmole) was high enough to allow recovery studies at spike levels equiva-
lent to 10 pg/g for a 10-g sample.
1. Calibration Standards
Eight concentration calibration standards containing the 17 native
and the 9 carbon-13 labeled internal standards were prepared for determining
the consistency of response factors for the native PCDDs and PCDFs versus the
corresponding carbon-13 congeners. Table 2 presents a summary of the calibra-
tion standards prepared for the method calibration study. The solution con-
centrations (pg/uL) can also be considered as equivalent to residue levels in
picograms per gram of adipose. For example, a 1 pg/nL concentration standard
corresponds to a tissue concentration of 1 pg of PCDD or PCDF congener per
gram of adipose assuming a 10-g sample is available for analysis.
2. Spiking Solutions
a. Native PCDD and PCDF
A solution containing the 2,3,7,8-substituted PCDD and PCDF
congeners was prepared in isooctane for spiking the homogenized lipid ma-
terials for the method evaluation study. Table 3 specifies the levels of
each of the native PCDD and PCDF congeners present in this solution.
-------
Table 1. Analytical Standards Available for the Method Evaluation Studies
Compound
Native
2,3,7,8-TCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDO
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,4,6,7,8-HpCDD
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDD
OCDF
l3C12~Internal standar
2,3,7,8-TCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDD
1,2,3,7,8-PeCDF
1,2,3,6,7,8-HxCDD
1,2,3,4,7,8-HxCDF
1,2,3,4,6,7,8-HpCDD
OCDD
37Cl-Internal standard
37C14-1,2,3,4,6,7,8-
Source
EPA QA Reference Materials
Branch
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
ds
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
Cambridge Isotope Laboratories
KOR Isotopes
Lot/Code
20603
AWN 1203-74/EF-903C
MLB-706-53/ED-950C
AWN-729-21/EF-953C
AWN-729-45/EF-956C
830244/ED-961C
MLB-706-47/ED-960C
MLB-706-73/ED-969C
AWN-729-20/EF-964C
MB 13106-7/EF-962-C
MB 13106-47/EF-967-C
MB 13106-3/EF-968-C
MLB-706-21/ED-971C
AWN-729-22/EF-973C
MB-13-106-77/EF-975C
8465-F-982-C/EF-982C
F2832/ED-980C
R00208/ED-900
R00236/EF-904
R00241/ED-955
R00221/EF-952
R00249/ED-966C
R00234/EF-963C
R00248/ED-972
R00263/ED-981
580012/SSY-4-32
HpCDD
14C12~Radio1abe1ed standards
2,3,7,8-TCDD
1,2,3,4,7,8-HxCDD
OCDD
Pathfinder Laboratories
Pathfinder Laboratories
Pathfinder Laboratories
S.A. = 117.56 mCi/mmole
S.A. = 24.16 mCi/mmole
S.A. = 20.50 mCi/mmole
S.A. = specific activity.
-------
Table 2. Concentration Calibration Solutions'
Concentration in calibration solutions in pg/pl
Compound CS1 CS2 CS3 CS4 CSS CS6 CS7 CSS
Native
2,3,7
2,3,7
1,2,3
1,2,3
2,3,4
1,2,3
1,2,3
1,2,3
1,2,3
1,2,3
1,2,3
2,3,4
1,2,3
1,2,3
1,2,3
OCDD
OCDF
,8-TCDD
,8-TCDF
,7,8-PeCDD
,7,8-PeCDF
,7,8-PeCDF
,4,7,8-HxCDD
,6,7,8-HxCDD
,7,8,9-HxCDD
,4,7,8-HxCDF
,6,7,8-HxCDF
,7,8,9-HxCDF
,6,7,8-HxCDF
,4,6,7,8-HpCDD
,4,6,7,8-HpCDF
,4,7,8,9-HpCDF
200
200
200
200
200
500
500
500
500
500
500
500
500
500
500
1,000
1,000
100
100
100
100
100
250
250
250
250
250
250
250
250
250
250
500
500
Internal quantisation standards
13C12-2,3,7,8-TCDD
13C12-2,3,7,8-TCDF
13C12-l,2,3,7,8-PeCDD
13C12-l,2,3,7,8-PeCDF
13C12-l,2,3,6,7,8-HxCDD 125
13C12-l,2,3,4,7,8-HxCDF 125
13C19-1,2,3,4,6,7,8-
12
HpCDD
13C12-OCDD
250
Internal recovery standard
13C12-1,2,3,4-TCDD 50
250
50
50
50
50
50
50
125
125
125
125
125
125
125
125
125
125
250
250
25
25
25
25
25
62.5
62.5
62.5
62.5
62.5
62.5
62.5
62.5
62.5
62.5
125
125
10
10
10
10
10
25
25
25
25
25
25
25
25
25
25
50
50
5
5
5
5
5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
25
25
2.5
2.5
2.5
2.5
2.5
6.25
6.25
6.25
6.25
6.25
6.25
6.25
6.25
6.25
6.25
12.5
12.5
1
1
1
1
1
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
5
5
50
50
50
50
25
25
25
50
50
50
50
125
125
125
50
50
50
50
125
125
125
50
50
50
50
125
125
125
50
50
50
50
125
125
125
50
50
50
50
125
125
125
50
50
50
50
125
125
125
50
50
50
50
125
125
125
250
50
250
50
250
50
250
50
250
50
250
50
Prepared in tridecane.
-------
Table 3. Native PCDD and PCDF Spiking Solution3
Concentration
Compound (pg/uL)
2,3,7,8-TCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDD
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2, 3,7,8, 9-HxCDD
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,4,6,7,8-HpCDD
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDD
OCDF
5
5
5
5
5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
25
25
Prepared in isooctane.
-------
b. Internal Standards
Two different internal standard spiking solutions were prepared
for quantisation of native PCDD and PCDF congeners. The compositions of each
of the spiking solutions are presented in Table 4. The internal quantisation
standards were spiked into the lipid aliquots prior to any cleanup procedures
and hence were carried throughout the method exactly as the corresponding
native congeners. The internal recovery standard was added in 10 uL of a
keeper solution (tridecane) during final extract concentration prior to analy-
sis. The recovery standard was used to measure the absolute method recoveries
of the internal quantitation standards.
C. Analytical Procedure
The homogenized human adipose lipid matrix was allowed to come to
room temperature and then warmed in a water bath until the matrix changed to
an oily state. Approximately 10.0 g of the oily material was transferred by
pipette to preweighed glass vials, and the actual weight of the lipid was
determined to the nearest 0.01 g by difference using an analytical balance.
Four 10.00-g aliquots were spiked with 20 pL of the native spiking solution
presented in Table 3, another four aliquots were spiked with 50 |j|_ of the same
solution, and three additional aliquots were spiked with 100 pL of native PCDD
and PCDF solution. These spikes were equivalent to concentrations ranging
from 10, 25, and 50 pg/g in the lipid matrix for the tetra- and pentachloro
PCDD and PCDF congeners up to 50, 125, and 250 pg/g for the OCDD and OCDF for
the low, medium, and high level spikes.
In addition to the spiked samples, three aliquots of the lipid
material were transferred for determining the endogenous levels of each of
the PCDD and PCDF congeners in the control matrix.
Each of the sample aliquots was fortified with 100 (jL of the in-
ternal quantitation standard spiking solution (Table 4). The spiked samples
were each quantitatively transferred to 500-mL Erlenmeyer flasks using hexane.
The residues were diluted with a total of 200 ml of hexane, and
100 g of sulfuric acid (H2S04) modified silica gel (40% w/w) was added to each
solution with stirring. The mixtures were stirred for approximately 2 h, and
the supernatants were decanted and filtered through filter funnels packed with
anhydrous sodium sulfate (Na2S04). The H2S04 modified silica adsorbents were
washed with at least two additional aliquots of hexane and dried by elution
through Na2S04.
The combined hexane extracts for each sample were eluted through a
column consisting of the 40% H2S04 modified silica gel (4.0 g) and silica
gel (1.0 g). The eluates were concentrated to approximately 15 ml_ and added
to columns of acidic alumina (Bio-Rad, AG-4, 6.0 g). The acidic alumina col-
umns were eluted first with 20 mL of hexane, which was collected but not ana-
lyzed, followed by elution with 30 ml of 20% methylene chloride in hexane.
The PCDDs and PCDFs were eluted from the acidic alumina using the 20% methyl-
ene chloride in hexane. The PCDDs and PCDFs in the eluates were isolated from
other chlorinated planar aromatics using columns (5-mL disposable pipettes
containing 500 mg of 18% Carbopak C and Celite-545). The Carbopak C/Celite
10
-------
Table 4. Internal Standard Spiking Solutions
Concentration
Compound (pg/uL)
Internal quantisation standard
13C12-2,3,7,8-TCDD 5
"13C12-2,3,7,8-TCDF 5
13C12-l,2,3,7,8-PeCDD . 5
13C12-l,2,3,7,8-PeCDF 5
13C12-l,2,3,6,7,8-HxCDD 12.5
13C12-l,2,3,4,7,8-HxCDF 12.5
13C12-l,2,3,4,6)7)8-HpCDD 12.5
13C12-OCDD 25
Internal recovery standard
13C12-1,2,3,4-TCDD 50
^Solution prepared in isooctane.
Solution prepared in tridecane.
-------
columns were pre-eluted with 2 ml of toluene, 1 mL of 75:20:5 methylene
chloride/methanol/benzene, I mL of 1:1 methylene chloride/cyclohexane, and
2 ml of hexane. The sample extracts (30 ml) were added to the columns, which
were eluted with 2 mL of hexane, 1 mL of 1:1 methylene chloride/cyclohexane,
and 1 mL of the 75:20:5 methylene chloride/methanol/benzene. These eluents
were collected and combined but were not analyzed. The Carbopak C/Celite
columns were turned upside down, and the PCDDs and PCDFs were eluted with
20 mL of toluene. The toluene was concentrated to less than 1 mL using flow-
ing nitrogen and a heated water bath (55-60°C) and transferred to 1.0-mL
conical vials using a solution of 1% toluene in methylene chloride. Tridecane
(10 uL) containing 500 pg of the internal recovery standard 13C12-1,2,3,4-TCDD
was added as a keeper when the solution had -concentrated to approximately
200 pL. The extracts were concentrated to final volume using nitrogen and
the heated water bath.
D. HRGC/MS Analysis
The analyses of the spiked and unspiked lipid samples were completed
using a Kratos MS50TC double-focusing magnetic sector mass spectrometer. The
determination for the tetra- through octachloro homologs was achieved in a
single analysis using the conditions described in Table 5. Table 6 provides
the characteristic ions monitored for each PCDD and PCDF homolog. As noted
from Table 6, the analysis requires five different parameter descriptions that
were switched automatically during the course of the analysis. Parameters
monitored included two characteristic molecular ions for each PCDD and PCDF
homolog and the corresponding carbon-13 labeled internal standard. In addi-
tion, a fragment ion of perfluorokerosene (PFK), m/z 380.976, was monitored
throughout each analysis to ensure that proper mass calibration was maintained.
The parameter descriptors also included an ion characteristic of specific
homologs of chlorinated diphenyl ethers to demonstrate that responses meeting
the qualitative criteria for specific PCDF congeners were not due to these
potential interferences.
Triplicate analyses of six of the eight calibration solutions
(Table 2) were completed, and the variability in relative response factors
across this range was calculated. The analyst was required to demonstrate on
a daily basis that the relative response factors (RRF) were in agreement
within ± 20% of the established averages for 2,3,7,8-TCDD and 2,3,7,8-TCDF
and within ± 30% of the average RRF values for the other congeners. The
equation used to calculate the relative response factors for each PCDD and
PCDF congener are discussed later in this report (Section E - Data Interpre-
tation, 2 - Quantitation, p. 17). The analyst was also required to determine
column performance by analyzing a mixture of TCDD isomers before proceeding
with sample analysis. Table 7 gives an example of the typical daily sequence
for PCDD/PCDF analysis.
12
-------
Table 5. HRGC/LRMS Operating Conditions for PCDD/PCDF Analysis
Mass spectrometer
Accelerating voltage:
Trap current:
Electron energy:
Electron multiplier voltage:
Source temperature:
Resolution:
Overall SIM cycle time:
8,000 V
500 |jA
70 eV
-1,800 V
280°C
^ 3,000 (10% valley definition)
I s
Gas chromatograph
Column coating:
Film thickness:
Column dimensions:
He linear velocity:
He head pressure:
Injection type:
Split flow:
Purge flow:
Injector temperature:
Interface temperature:
Injection size:
Initial temperature:
Initial time:
Temperature program:
DB-5
0.25 pm
60 m x 0.25 mm ID
~ 25 cm/sec
1.75 kg/cm2 (25 psi)
Splitless, 45 s
30 mL/min
6 mL/min
270°C
300°C
1-2 uL
200°C
2 min
200°C to 330°C at 5°C/min
13
-------
Table 6. Ions Monitored for HRGC/MS Analysis of PCDD/PCDF
Descriptor ID
Al TCDF
13C12-TCDF
TCDD
13C12-TCDD
HxCDPE
PFK (lock mass)
A2 TCDF
TCDD
PeCDF
13C12-PeCDF
PeCDD
13C12-PeCDD
PFK (lock mass)
HpCDPE
A3 HxCDF
PFK (lock mass)
13Cl2-HxCDF
HxCDD
13C12-HxCDD
OCDPE
Mass
303.902
305.899
315.942
317.939
319.897
321.894
331.937
333.934
373.840
380.976
303.902
305.899
319.897
321.894
337.863
339.860
349.903
351.900
353.858
355.855
365.898
367.895
380.976
407.801
373.821
375.818
380.976
385.861
387.859
389.816
391.813
401.856
403.853
443.759
Nominal dwell
time (sec)
0.090
0.090
0.090
0.090
0.090
0.090
0.090
0.090
0.090
0.090
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.035
0.035
0.080
0.080
0.080
0.080
0.080
0.080
0.080
0.080
0.080
0.080
14
-------
Table 6 (continued)
Descriptor ID
A4 PFK (lock mass)
HxCDD
HpCDF
13C12-HpCDF
HpCDD
13C12-HpCDD
37Cl4-HpCDD
NCDPE
A5 PFK (lock mass)
OCDF
13C12-OCDF
OCDD
13C12-OCDD
DCDPE
Mass-
380.976
389.816
391.813
407.782
409.779
419.822
421.819
423.777
425.774
435.817
437.814
429.768
431.765
477.720
380.976
441.743
443.740
453.784
455.781
457.738
459.735
469.779
471.776
511.681
Nominal dwell
time (sec)
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.060
0.070
0.070
0.070
0.070
0.070
0.070
0.070
0.070
0.060
15
-------
Table 7. "Typical Daily Sequence for PCDD/PCDF Analysis
1. Tune and calibrate mass scale versus perfluorokerosene (PFK).
2. Determine column performance by injecting the TCDD isomer mixture.
3. Inject concentration calibration solution 2.5 to 12.5 pg/ul_ (CS-7)
solution.
4. Inject blank (tridecane).
5. Inject samples 1 through "n".
6 Inject concentration calibration solution 2.5 to 12.5 pg/uL (CS-7)
solution.
16
-------
E. Data Interpretation
1. Qualitative
The HRGC/MS elution profiles of the tetra- through octachloro PCDD
and PCDF homologs were established through the analysis of environmental sam-
ple extract (fly ash from a municipal waste incinerator). The characteristic
ions for each homolog were plotted within the retention window established
using this mixture. The criteria for identification of a response as a PCDD
or PCDF were the coincidental response of the characteristic ions monitored
within the established retention window, and within ± 20% of the theoretical
ion ratio. Table 8 presents the range of ion ratios used for the qualitative
criteria for the specific PCDD and PCDF homologs and internal standards.
2. Quantitation
Quantisation of the specific PCDD and PCDF congeners was achieved
using the respective internal quantitation standards. For example, TCDD was
quantitated versus the 13C12-2,3,7,8-TCDD; PeCDD versus the 13C12-1,2,3,7,8-
PeCDD, etc. The HpCDF and OCDF responses were quantitated versus the carbon-13
labeled hepta- and octachlorodibenzo-p_-dioxin internal standards since the
corresponding dibenzofuran internal standards were not available for this
study. The absolute recovery of the internal quantitation standards was
achieved using 13C12"1,2,3,4-TCDD. A second internal recovery standard,
37Cl4-l,2,3,4,6,7,8-HpCDD, was evaluated but was not used for recovery mea-
surements due to interference arising from the corresponding native HpCDD.
Relative response factors (RRF) were calculated for each of the
native PCDD and PCDF compounds listed in Table 2. The RRF values were calcu-
lated as shown in Equation 1.
DDC ASTD X CIS c .
RRF = A—;TT— tq. i
IS STD
where A<-Tn = the sum of the area responses for the two characteristic ions
of the standard compound;
Aj<. = the sum of the area responses for the two characteristic ions
of the corresponding internal quantitation standard;
Cjc- = concentration (pg/pL) of the internal quantitation standard; and
= concentration (pg/nL) of the standard compound.
The relative response factors for the internal quantitation stan-
dards (RRF,.<0 were calculated as shown in Eq. 2.
17
-------
Table 8. Ion Ratios for HRGC/MS Analysis of PCDD/PCDF
Compound
TCDF
13C12-TCDF
TCDD
13C12-TCDD
PCDF
13C12-PeCDF
PeCDD
13C12-PeCDD
HxCDF
13C12-HxCDF
HxCDD
13C12-HxCDD
HpCDF
13C12-HpCDF
HpCDD
13C12-HpCDD
OCDF
13C12-OCDF
OCDD
13C12-OCDD
Ions monitored
304/306
316/318
320/322
332/334
338/340
350/352
354/356
366/368
374/376
386/388
390/392
402/404
408/410
420/422
424/426
436/438
442/444
454/456
458/460
470/472
Theoretical ratio Acceptable range'
0.76
0.76
0.76
0.76
0.61
€.61
0.61
0.61
1.22
1.22
1.22
1.22
1.02
1.02
1.02
1.02
0.87
0.87
0.87
0.87
0.61 -
0.61 -
0.61 -
0.61 -
0.49 -
0.49 -
0.49 -
0.49 -
0.98 -
0.98 -
0.98 -
0.98 -
0.82 -
0.82 -
0.82 -
0.82 -
0.70 -
0.70 -
0.70 -
0.70 -
0.91
0.91
0.91
0.91
0.73
0.73
0.73
0.73
1.46
1.46
1.46
1.46
1.22
1.22
1.22
1.22
1.04
1.04
1.04
1.04
Acceptable range is ± 20% of the theoretical value.
18
-------
AT<- * '•'DC
RRFIS = AIS x c Eq. 2
Ib ARS X UjS
where AIS and CJS are defined as in Equation 1 and
CDC = concentration (pg/uL) of the internal recovery standard,
'RS
13C12-1,2,3,4-TCDD, and
AR<- = the sum of the area responses for the two characteristic ions
• (m/z 332 and 334) corresponding to the internal recovery
standard.
A calibration curve was established using six concentration levels
of standards; for example, the calibration curve for 2,3,7,8-TCDD was ini-
tially established with standards at concentrations of 1, 2.5, 5, 10, 50, and
100 pg/uL. The 2.5 pg/uL standard was analyzed daily to verify response
factors and instrument sensitivity. The RRF values for each of the internal
quantitation standards were calculated versus the internal recovery standard,
13C12-1,2,3,4-TCDD, using Equation 2.
The concentration of a PCDD or PCDF congener in a composite sample
was calculated as shown in Equation 3.
r _ Asamp1e x QIS F ~
Vr AIS x RRF x Wt tq' J
where CWT = wet tissue concentration of the PCDD or PCDF congener in each
tissue (pg/g);
A , = sum of the area responses for the two characteristic ions of
sample the pc[JD Qr pCDF congener;
AIS = sum of the area responses for the two characteristic ions of
the respective internal quantitation standard;
QT(- = amount of the internal quantitation standard added to the
ib sample (500 pg of 13C12-TCDD, 13C12-TCDF, 13C12-PeCDD,
and 13C12-PeCDF; 1,250 pg of 13C12-HxCDD, 13C12-HxCDF,
and 13C12-HpCDD; or 2,500 pg of 13C12-OCDD);
RRF = the relative response factor for the PCDD or PCDF congener
from Equation 1; and
Wt = mass of the sample (grams).
The absolute recovery of the internal quantitation standard was
calculated using Equation 4.
19
-------
A x Q
Recovery (%) = T L* ™ x 100 Eq. 4
where ARS = sum of the area responses for the two characteristic ions of
the internal recovery standard, 13C12-1,2,3,4-TCDD;
QR<- = amount of the internal recovery standard (13C12-1,2,3,4-TCDD)
added to the final extract (500 pg); and
RRFTr = response factor of the internal quantitation standard relative
to the internal recovery standard. These values are calcu-
lated as defined in Equation'2. The RRFJC. values were all
calculated versus 13C12-1,2,3,4-TCDD.
All data were qualified to reflect that the response for a particu-
lar compound was a positive quantifiable parameter, present as a trace value
only, or was not detected. Positive quantifiable values were identified for
responses greater than 10 times background signal to noise. Trace (TR) values
were assigned to responses which were in the range of 2.5 to 10 times back-
ground signal to noise. A value of not detected (ND) was used to reflect that
a response was not detected at greater than 2.5 times signal to noise. A limit
of detection (LOD) was calculated for all trace and not detected values using
the peak height response of the respective internal standard and the average
measured signal to noise for the characteristic ions of the PCDDs and PCDFs.
F. Quality Assurance/Quality Control (QA/QC)
The QA/QC procedures included analysis of multipoint calibration
concentration standards to establish relative response factor (RRF) curves
for each of the 17 native PCDD and PCDF congeners. Triplicate analyses of
6 concentration calculation standards (Table 2; CS2, CSS, CS5, CS6, CS7, and
CSS) were determined to vary by less than ± 20% for TCDD and TCDF and ± 30%
for all other PCDD and PCDF congeners. The mean RRF values were also deter-
mined to vary by less than this criteria over the entire calibration range.
The mean RRF (RRF) values and instrument sensitivity were verified daily by
bracketing the sample analyses with an injection of a standard that ranged
from 2.5 pg/pL for 2,3,7,8-TCDD and 2,3,7,8-TCDF up to 12.5 pg/iA for OCDD
and OCDF. The criterion for continuing with the sample analysis was agree-
ment of the measured RRF value with the mean RRF within ± 20% for 2,3,7,8-
TCDD and TCDF and ± 30% for all other PCDD and PCDF congeners.
Other activities included the analysis of laboratory method blanks
and reagent blanks and measurement of the absolute recoveries of the internal
quantitation standards. Laboratory method blanks were samples that were han-
dled exactly as an adipose sample except no lipid matrix was used.
G. Preliminary Method Studies
Prior to analysis of the homogenized human adipose lipid matrix by
HRGC/MS, several experiments were conducted to confirm that the sample prepa-
ration scheme was feasible.
20
-------
1. Gravimetric Studies
The first concern was the efficient removal of up to 10 g of lipid
matrix from extracted adipose tissue. A series of experiments was conducted
with 10-g lipid aliquots to demonstrate removal of lipid using the H2S04-Si02
slurry technique. Initially, 50 g of the acid modified silica was added to
the lipid extract in 100 mL of hexane. The acid modified silica was noted to
turn dark brown immediately on contact with the lipid solution. The hexane
was recovered and the adsorbent was extracted with additional hexane. The
extracts were combined and concentrated to 5 ml with Kuderna-Danish evaporators.
The extract was eluted through a column of 4.0 g of acid modified silica and
1.0 g of silica with 45 mL of hexane. The acid modified silica was noted to
be highly discolored throughout, and the extract required a second slurry
treatment of the eluent with an additional 50 g of acid modified silica gel.
The adsorbent from the second slurry procedure was noted to discolor signifi-
cantly, indicating that lipid materials had not been efficiently removed from
the first step of the procedure. The hexane supernatant from the second
slurry cleanup was reduced in volume and taken to dryness in a preweighed
glass vial. The final residue was measured at approximately 10 mg for dupli-
cate samples, which translates into a removal efficiency of 99.9% based on
the initial 10-g aliquot.
This lipid cleanup procedure was modified such that 100 g of acid
modified silica gel was used in the initial slurry cleanup, followed by elu-
tion of the resulting extract through a column containing 4.0 g of acid modi-
fied silica and 1.0 g of silica gel. The lipid removal efficiency of dupli-
cate samples through the cleanup procedure was determined to average 99.8%
(20 to 30 mg of the initial 10-g lipid remained after cleanup). The column
cleanup step in this procedure did not exhibit any significant color change.
Thus this step of the procedure was incorporated into the method as a check
of the efficiency of lipid removal to prevent overloading of the acidic alu-
mina fractionation column.
2. Carbon-14 Recovery Studies
The carbon-14 radiolabeled PCDD standards listed in Table 1 were
used to estimate overall method recoveries for the tetra- through octachloro
homologs prior to proceeding with the HRGC/MS evaluation. Triplicate analyses
(10-g aliquots of lipid materials) were conducted with each of the three
radiolabeled standards. The first experiment addressed the recovery of the
compounds from bulk lipid cleanup. Triplicate analyses using 10-g aliquots
were completed for the three compounds at the following concentrations:
14C-2,3,7,8-TCDD, 10 pg/g; 14C-l,2,3,4,7,8-HxCDD, 100 pg/g; and 14C-OCDD,
250 pg/g. The results of these analyses indicated that all compounds were
recovered in the range of approximately 70 to 80%. Following this experiment,
the total sample preparation procedure described earlier in this report was
evaluated using triplicate analysis of 10-g lipid samples. Table 9 provides
a summary of the results from this study. These data indicate that overall
method recovery is limited by the initial bulk lipid removal procedure. This
assumption is based on the similar recoveries of the carbon-14 labeled com-
pounds noted for evaluation of the bulk lipid removal step as compared to the
total sample preparation scheme.
21
-------
Table 9. Summary of the Results of the Sample Preparation
Method Evaluation Using Carbon-14 PCDDs
Analytes
14C-2,3,7,8-TCDD
14C-l,2,3,4,7,8-HxCDD
14C-OCDD
Spike
levels
(pg/g)
10
100
250
Total method
recovery (%)
68
79
82
Bulk lipid
removal,
recovery
75
66
76
Average value for triplicate analyses taken through the total sample
preparation scheme. Precision of measurements varied by less than ± 10%
(relative standard deviation).
3Average value for triplicate analyses taken through bulk lipid cleanup
only. Precision of measurement varied by less than ± 6% (relative stan-
dard deviation).
22
-------
V. RESULTS
A. Analytical Results
The analytical results for the quantitation of the 17 target PCDD
and PCDF 2,3,7,8-substituted congeners in the spiked and unspiked homogenized
human adipose lipid samples are presented in Tables 10 to 15. These data
demonstrate that 13 of the 17 congeners were definitely detected in the un-
spiked lipid matrix. Although 2,3,7,8-TCDF is reported as not detected, re-
sponses for the characteristic ions (m/z 304 and 306) greater than 10 times
signal to noise were noted to be coincident with the internal standard, 13C12~
2,3,7,8-TCDF. The ratio of the responses (m/z 304/306) in each of the trip-
licate analyses of the unspiked matrix were- well outside the acceptable ratio
of 0.90 to 0.61 established in Table 8. Figure 1 provides a comparison of
the HRGC/MS-SIM responses noted for the unpsiked human adipose lipid matrix
as compared to a concentration calibration standard. Figures 2 through 6
provide examples of the individual PCDD and PCDF responses observed for the
unspiked lipid samples as compared to fortified matrices.
In general, the precision of the replicate measurements at each
spike level is good (relative standard deviations typically less than 10%)
for PCDD and PCDF congeners that were detected with responses greater than 10
times signal to noise. The precision of the measurements for the unspiked
matrices for 1,2,3,7,8-PeCDF (Table 11), 1,2,3,7,8,9-HxCDF (Table 12), and
OCDF (Table 15) ranges from 21.6% to 43.1% as a result of little or no re-
sponse at the specified retention windows.
B. Statistical Analysis
The regression results for each of the 17 specific PCDD and PCDF
congeners are plotted separately in Figures 7 to 23. These plots provide the
results for the individual sample analyses, a line defining the results of a
least squares regression analysis, and boundaries that depict the confidence
limits for the range of spiked concentrations. The regression lines were ob-
tained by the method of least squares using the sample measurements at the
three spiking levels and the unspiked level.
Two types of upper and lower 95% confidence limits or bounds were
calculated for the least square regressions of measured (found) concentrations
versus spiked levels. The first set of confidence limits (defined by the
inner pair of curves closest to the regression line) is the 95% confidence
bounds for the regression line. These bounds are interpreted as follows:
The true regression line (as would be determined if the experiment were re-
peated a countless number of times at the same spiked levels) lies within
these confidence limits unless the analytical results are sufficiently unusual
to be among those expected to occur less than 5% of the time.
The second set of confidence bounds, depicted by the outer pair of
lines, constitutes the 95% confidence limits for the result of a single analy-
sis at a particular spiking level. The interpretation is as follows: the
result (reported value) of a single analysis of a sample spiked at a given
level can be predicted to fall between these 95% confidence bounds unless the
analytical result is among those sufficiently unusual to be expected less than
5% of the time.
23
-------
L_J
^^
C_J
1 —
1
CO
f
f^.
».
ro
CM
T3
C
re
u_
o
£_}
r~~ ~ */)
1 Ol
CO r—
- 0.
p^. E
» re
CO CO
r.
CM TJ
• r-
<4- Q.
0 ••-
1
in
c 01
o in
•r- 0
4-) a.
re •(-
s- ~a
4-> <
c
ai c
u re
C E
O 3
O 1C
™U "CJ
ai CD
i- N
3 "r-
in c
re a)
CU
•r—
Q.
CO
0
rH
ai
t—
-Q
ro
h-
O ^§
O QJ ^^
0 4->
H- 3 >,
1 r- S-
CN O QJ
r-i in >
0 .Q 0
co re u
rt 0)
t-
c
a o
Q T-
C-> 4-^
\— IQ <~*.
i s- en
CO 4-* ^^
- c en
r^.
** U *»-•'
CO C
- O
CM U
a
O i—
(_3 CD
1— >
1 0) ^>
co r- en
** "^s,
r^ cu en
'^ a.
CO T- ^>
- a.
CM m
^*\
u_ s^
o cu ^—^
0 4->
H- 3 >,
1 •— i-
M o ai
T-* in >
O J2 O
M re u
rt OJ
s_
c:
u_ o
Q T-
0 4-)
H- re <^-
i i- en
CO 4^ ^v
- c. en
1 — ni o
- (J •»— <
CO C
- o
CM CJ
u_
r~~\ r—
O 0)
r- >
1
co i— en
*• ^x
p~- a> en
"J^ Q.
00 -i- • — '
- Q.
CM in
-
co co m
• • •
co CM CD o co r»»
m m ^j- m
r*^- ^ i — ( r^> CM in
O rH CO rH rH O
rH rH rH rH i— t
O O O
CO CO CO
• • •
en co en o CM co
in co in co
re re re
/*^s /"^ ^*^ /"^
rH rH O rH rH 00
<-)-,-]- ,-j. «^- O i— 1
^*s *<*^ ^^ ^x-/
o o o a
~Z.-Z.-Z. -ZL
000
/*^>
^
^^
c >
re a Q
O) 1 — CO
s: co a:
CO in rH
• • •
co co co CM m «3- co
m m m CD m
- m co o rH
CO CM <3- CM CO rH *J-
CM CM CM CM CM
0 0 0 O
rH rH rH rH
CO rH CO
• . •
rH I"-. rH 00 CD f~» O
CD CO f** r~« CD rH
co oo co oo rH m en
^" ^^ CO CO ^J1 G3 CO
rH rH rH rH rH
O O O O
rH rH rH rH
/^^,
^S
*^s/
c
re a a
0) h- CO
2: co cc.
O rH CO
* • •
oo co rH o co in en
^1- m in CD m
oo co co co o CM en
O O O 00 O rH CM
^" ^ ^" CO ^"
m in in un
CM CM CM CM
o r-- in
* • •
en m CM CM r>» r^ i—t
m r-. CD r~ co rH
oo oo r-~ r~ CM rH m
O O O 00 O rH CO
CO CO CO CM CO
in m m in
CM CM CM CM
/"s
^
^^
c.
re Q Q
ai \— co
2! CO Q£
r-~- o «^"
in co oo oo co CM
CO rH CO in CO 00
m CM r»- oo co «^-
CD r*» co co
o o o
m m m
1 — rH CO
• . •
«3- 00 00 CD 00 4-
p^ ^" 00 CD OO rH
i — en in r- rH co
m in m m
000
m in m
,^— >
^
s^^/
C
re Q Q
Q) I— CO
C
re
1 1
ai
i. re re
re -i- 4->
ai s- a.
t- Q. CU
en o u
i- U
4- a. re
0 Q.
ro ai
in aj 4->
C .f~
O 4-> C
Q. ro
m 4^ f~
cu ro 4->
/— s S-
co ro
ai
c -o s-
0 C 0)
•r- ro
O <3" i —
cu o -Q
4-> co ro
a» s-
T3 N Ol
•~x "O
4- E •"-
O •*-* in
C
4-> in o
•r- C U
E 0
•r— -r- in
i— ro
u 3
T3 --
ai +-> o
4-> in -i—
re •<- 4->
s s- ro
•r- (U S-
4-> 4J
in u c
cu ro o
S- •!-
cu re
.c -c cu
4-) CJ -C
in f~
•r— -4— > **
O S-
m .a ai
O) >
in i. a>
CD 0 2
^Z 4— O
c -a
QJ Cl)
S- 4-> •
ro o LJ_
a. c en
Cin i
i/j ^~
•i— ro i
5 oo
3 CD I--
i— in
re "— co
> o -o
c CM en
i
. o s- o
•a 4-> o
a> i 4- o
u re ai
O) C E r- 1
4-> O) •*— CO
OJ •!— -t->
T3 in o
C
4J in o 4—
O OJ ••- O
C E 4->
II 4-> ai 01
4J C
Q o cu ro
Z !—* S- i-
ro
24
-------
0
a
a»
a.
CO
ro
CM
c
CO
u_
Q
ai
a.
CO
. a>
ro •—
- a.
CM E
u,
O OJ
(_J 3
a. IA
CO *—
- I/I
ro O
. a.
CSJ •—
-•3
<*- c
°1
C X
o
•^ T3
•*-> a>
*J C
c ai
a> en
u o
C £
O O
CJ I
•a c
u
3
(TJ
a>
^
W)
3
U)
L.
a>
•o
a>
Q.
00
t— 4
1-H
ai
t—
0 ^
o «
a. 3 >»
i f— f*-
N o a>
U A O
O (O (J
t-
0
a c
o o
i S'S
P"* C O>
- c
CM O
- (J
o
a
ai
i a> --^
co — o>
r*T oi"^
- O."™'
CM Wt
i— t
U. -— v
o x
a- 3 >>
i r— t.
— W) >
<_> J3 O
n co U
^ OJ
Uu
5§
a» "-
oo 2 "oi
P^ C O)
- u
CSJ
u.
o
a> a>
a. >
CO f— 'oi
r*. ai Q)
-^ a.
* Q.S^
CSI
fee
tJ O
& *J
i co *— N
00 1- 0
P^- C Ol
ro u •— '
* C
CSJ O
- O
U.
o *—
ai 41
a. >
I 4) /-».
00—0)
r*- o> o>
CSI U)
i-H
OOUJ
in to in in
o 91.00 en r-t in
CM t— t i— 1 1—4
000
O VO *tf"
•
S r5 CD f^ CS*
ooico m rom
O i-t Cn G t~t i£)
CSJ CM -H CSI
000
(C
^OO O^O f-f
iii §
000
s
c
CD o a
01 (— tO
mr-4ro
o r- LT» CM co ro m
U3 in in vc in
CSJ f^ O O O rH ID
o o o o
P-l ,-4 pH f-4
o r»- f-i
•
en ^D oo r*» r** ^ ^^
r«* r^. 00 to CM ^ tn
CM ro CM ro ro I-H
0000
r-4 rH i— t fH
SSSS Som
0000
f-4 rH f-t iH
s
C
moo
s: J> c£
o CTI ys
tSSSS S"^
0 rHCOl/1 I~- in «•
co VJD en ro VJD CM in
CM CM «SJ CSI
m ro ^~
VD CO CO CO P*» rH i-H
V CO i-4 CO CO in rH
in in in in
CSJ CSJ CSI CSJ
en o co tn oo csj r*»
CM ro CM esj csi
in in in in
CM CS) CSI CM
S
c
at h- i/>
Z to ce
OOlC
CO h- to
25
-------
u_
a
(j a; >,
X ** fc.
z D ai -•— •
22 S*
c_> .a 01
n ro t-
1 C
en o
CO -*->
r** ia i- O)
»CJ *J "*s
m x c en
-z ai a
CM U «-- '
- c
rH O
U
en .—
at
00 >
-u_ ai '-»
r-. a r- en
m x a» O)
-z .* a.
CM -r- «»-•
a.
rH M
I C
00 O
r- *J
-u_ a ^
tD Q t- O)
Wx "c *a>
•z ai o.
m u *— '
c
CM O
U
1
CO i—
a*
r*» >
• ^ CU ^«v
«• x ai en
-z .* a.
a.
CM i/>
I C
CO O
r-. 4->
"U_ IO <"•"*
*£> a t- CT
«CJ •*-> "^v
m x c CT
-z ai a
CM U *W
rH O
U
i -
a»
r^ >
-u_ a> <*^
10 ca r- en
-O *"».
ro x o» en
-z ^ o.
CM •!- S«. '
- g-
I C
CO O
r*. ^
-u. *o *-N
T a t- en
-o *-> ^
m x c en
*z a> a.
CM (J <**
C
rH 0
U
CO r-
a,
f"^ >
-U. Ol -—
^" O r- C?)
-tJ -X.
ro x ai o»
-z ^ a.
CM — <— •
a.
rH U)
r*. m <&
in rx CM in CM *f
in ^n in in
as
m r*- en r** CM ro
o o o o o tn
s^"^»s_^ -w CM
§ii i
000
* * V «T O CO
O O O
ro * r«* ^- CM fx.
CM CM CM CM O rH
rH rH rH rH
O O O
O rH <• CM CM rH
CM CM CM CM O rH
CM CM CM CM
O O O
X
c
•c a o
Ol r- m «*> n
in in m tn
CM CM CM CM
us ȣ> CO co o ao r-
to co o en en rH ^r
m ro ^ PO m
in in m in
CM CM CM CM
r*. O CO O rH CM ^~
LD CM co en en CM *r
^- in v ^ v
in in in vn
CM CM CM CM
M
C
rtj a a
O> h- l/>
Z to ee
CD CM *£>
lO i£> iD r^ ^D
r*^ ^- CM m P"- «o- m
CM en CM us CM in 10
co co oo rx. oo
m m m m
CM CM CM CM
VD U3 U3 iO
^- m «• rH m CM CM
o* r«- r«» p«* rx.
m m m m
CM CM CM CM
•>£> VO U3 U3
r«* in u3 o en CM en
fn o en ID en m ro
oo co r- r-. r-.
in in in in
CM CM CM CM
ID U3 U3 *£>
CM rH m r- rH r>- r*.
o en o o o o o
en co en en en
m in m m
CM CM CM CM
U3 «3 U3 VO
W
C
m o a
O> r- trt
£ CO CC
O P-. rH
t*- ^- P^. \O rH CO
in in m m
O rH «$• rH ^- \Q
<*> «• rn O rH r>-
•«• ^- U3 in rH
rt rH rH rH
in in in
CM CM CM
rH rH rH
rH rH rH ^- in PO
-«• en us
r-, in us u? O o
m in m in
r- 1 rH rH rH
in m in
CM CM CM
rH rH rH
X
C
moo
Ol r- (/I
Z w> ce
26
-------
o
o
o
x
X
cn
00
r-.
CM
rH
C
re
• -
CO
r-;
U3
* t/>
m o»
~ 1
i -a
CO •'•*
- Q.
- O>
m vi
- o
CM a.
r-Tjg
O C
re
c E
o =
"- IT
re -o
(- a>
c •*-
C CT
*
t/i
3
t/t
t-
•0
^
Q.
l/l
rH
at
re
t—
o w
a ^
cj a>
X 4-> >»
X 3 S-
i •— A>
IN 0 >
— VI O
O J3 U
n re ai
Q
0
*X 0
i -t->
- i~ cn
CO +•> "••».
- C CT
piv flj Q,
- O
CM U
0
X •—
X Hi
1 >
-r- en
• V Cn
r~ j£ Q.
f*l CL
CM
O
0
0 C
X O
X "-
co re *~*
- i- cn
- c cn
10 a* a.
- U •— '
CO C
- o
CM U
rH
O
a
X r—
X (U
1 >
CO Or --^
-r- Ol
» a> C3i
m o.
CM"
rH
0
O
X O
1 4->
co re *-*
- s. cn
^ e'en
•3- a> Q.
- LI *-^
» o
CM W
rt
o
o
X r—
X Oi
1 >
co ai *-^
-r- OI
f a> cn
•q- j<: a.
ro Q.
- i/t
r-"
r* CM O
CO O 00 CO rH CM
in vo in in
i— i o r*-. CM r*. co
rH CM CM CM rH
O O O
ooo r* o ^
r- CM *r r«. ^- CM
m ^o in in
rH rH rH rH
000
\D f** m in CM in
CM CM CM CM
OOO
S
c
re a o
OJ t— OO
m P" UD
in rH m o m m m
in rH co f) cn «f o
\£> V *T \T> m rH CM
in m in m
CM CM CM CM
o o O O o in cn
^ m co ro in *»• f-
in in in in
CM CM CM CM
CO O P«- f*- kO tiO rH
<• in in in m
m m m in
CM CM CM CM
2
c
re o a
(U h- to
Z m C£
rocnr-
SS1SP: S1"*
o cn CM ^ ic in rH
rH cn O O -O
rH rH rH rH
m m m m
CM CM CSJ CM
o o o o co in CM
o cn o o ^- en ^~
CM m CM CM CM
CM CM CM CM CM
m m in m
CM CM CM CM
cr> (*•"• f) eft cn o cn
co cn cn co co
in m m in
CM CM CM CM
s
c
to o o
a* r- i/>
X l/)CC
0 tf)^-
CM P- rH O CM ^
mr-* rncM«
^ m CO VD CM rH
rH rH rH rH
m in m
CM CM CM
rH rH rH
O O O cn co o~i
Sen vc ^- 10 w>
cn ID co »H
CM CM CM CM
in in >n
CM CM CM
rH rH rH
r-t co o cn cn m
rH O ^D in *T PO
rH rH rH rH
m m in
CM CM CM
rH rH rH
M
_ **-'
10 Q Q
Ol r- V)
Z VI CC
27
-------
o
a
CJ
a.
x
CO
fvf
**.
^r
CM
rH
•a
c
„
U.
a
en
CO
ps.
- I/I
f) a>
csj CL
rH «C
U- "O
o a.
a.—
x — t
CO 0)
- IS)
-•o
ro" C
- m
CM E
**- -a
U> -r^
C C
O QJ
•r- Ol
*-> O
fO E
fc. O
••-> X
C
O* C
(J •(-
o
•o
V
f-
3
in
(0
OJ
3
I/)
ai
•a
a.
rH
ai
(0
0 &>
o a/
Q. •*-" >i
f ^ %
W 0 >
— vfl O
tj J3 U
§
(ci c
X O
co •£ ^
^ u ai
v£> c en
- a> a.
«3- u 's—
- c
f> 0
- u
CM
rH
a
o
CJ
£-
i u
CO >
<& o> en
-^ a.
m" S"
CSI
rH
LL.
0
*ci c
x _o
en "-^
co t- en
p*. c en
-wo.
*f U ^->
-. c
d O
* u
CM
a
CJ
a.
x ^~
i Ol
* OJ >-v
CO •— O)
-^ a.
- Q.
ro wi
CM
u_
U
CL C
X O
CO 4->
- m ^-s
r*. i- en
ixi c en
- QJ CL
f) O
- o
CM
rH
U.
0
CL
i a*
CO > ^
r-. «— en
•xi ai en
.^ Q.
- o.
CM
rH
<*i in m
r- r«- iC r^
PS. co m r-, m CM
**• o if) en CM f— •
f- 1 rH rH rH
CM CM CM CM
O O O
IQ to ia (o
fl rH O ^ CM
cd.d ct°s
§§§ §
o o o
ujmvc OMD I-.
o r~ co oo 1-1 in
ro es» cvj eg
0 0 O
«
c
QJ h- l/l
£ LO QC
0 10 U5
SfCSS g"1^"
PS- o en in fi o r*>
*r en <*i co ro ^9- rH
CM CM CM CM CM
m m m m
CM CM CM CM
rH ^ fs.
sO rn n in ^* rH in
CM CM CSI CSJ CSJ
m m m in
CM CM CM CM
VDsDOrr) rH \£> O
??SS £"""
in in in m
CM CM CM CM
2
c
(DO Q
r-* en en CD o en
rH co en ^ co co CM
CO CO VD P*» r*-
CM CM CM CM CM
in in m m
CM CM CM CM
O CO f>
So en co en o rH
vo m m m
in in in in
CSJ CM CSJ CM
us m o f) «*• CM CM
m m in in
csj csi csi CM
*?
c *"""'
•a a o
at h- to
z to ce
«r co
CS( rH cn ^ ^ >O
CM "«• PS. CO CM CO
f> in fl O t£» rH
m m <• m
m m in
CM CM CM
r-H l-H rH
^- rs. o
cn to rH CM ro m
•-H csj CM CM
rH i-H rH rH
m m m
CM CM CM
in in in in
in in in
CM CM CM
•g
28
-------
Table 15. Spiked Versus Measured Concentrations of OCDF and OCDD
in Homogenized Human Adipose Lipid Samples
OCDF
spike level
(pg/g)
0
0
0
Mean
STD
RSD (%)
50
50
50
50
Mean
STD
RSD (%)
125
125
125
125
Mean
STD
RSD (%)
250
250
250
Mean
STD
RSD (%)
OCDF
concentration
(pg/g)
4.9
2.3
2.6
3.2
1.4
43.1
44.2
45.0
45.9
49.6
46.2
2.4
5.2
111.1
107.8
110.7
113.1
110.7
2.2
2.0
227.8
231.0
228.6
229.1
1.7
0.7
OCDD
spike level
(pg/g)
0
0
0
50
50
50
50
125
125
125
125
250
250
250
OCDD
concentration
(pg/g)
804
833
781
806.1
26.1
3.2
849
856
876
860
860.4
11.4
1.3
932
934
944
907
929.1
15.7
1.7
1,080
1,140
1,070
1,096
34.7
3.2
13C12-OCDD
abolute recovery
(%)
91
88
94
91.0
3.0
3.3
100
87
91
91
92.3
5.5
6.0
90
96
102
104
98.0
6.3
6.5
69
67
74
70.0
3.6
5.2
29
-------
RIC.
Umpilted Human Adipote TIMUP
3 7,8-r.CDF
IVCDD
8-M,CDf
3.V;."-HxCr>f
cor
COD
con
7 P-M.CDO
CDl>
CDF
P-HpCDF
8-HpCOn
,4 « 7,B-HpCDD
5 OCDD
'3CI7-OCDD
OCDF
It. "
21,22,23
***dk«*t
10M
9.10
(I
,
:> '*
1286
isee
2688 SCAN
Colibralinn Slatvlord
1,2
6.7
LLJUU
1,12
15.16.17
21,22.23
25,26
ree
12W
tew
2*98 SCAN
Figure 1. Comparison of the HRGC/MS-SIM reconstructed ion chromatogram (RIC)
from the analysis of unspiked homogenized human adipose tissue
matrix and a calibration standard for PCDDs and PCDFs.
30
-------
Urnpiked Human Adipote
3B4 .
384 .
Spik.d Human Adipou
2,3,7,8-TCDF
ieee
IBM
use
1296 1258 SCAN
320 .
32« .
Unspiked Human Adipoie
A JV ^
^
2.3.7.8-ICDD
Spiked Human Adipoie
l3C(2-2.3.7,8-ICDF
A r
A
2.3,7.8-KDD
9BO
1858
12M
125C
SCAN
Figure 2. Example of the TCDF (m/z 304) and TCDD (m/z 320) HRGC/MS-SIM
elution profiles in unspiked and spiked human adipose. The spiked
concentrations for TCDF and TCDD in these chromatograms were 25 pg/g
each.
31
-------
338 I
Unspik«d Human Adipose
1,2.3,7.8-PeCDF
^
2.3.4.7.8-PeCDF
338 J
Spiked Human Adipoie
1,2.3,7,8-PeCDF-
2,3.4.7,8-PeCDF
use
1350
354 I
354 J
Umpiked Human Adipose
3C]2-l,2,3,7,&-PeCDF
1.2.3,7,8-PeCDD
Spiked Human Adipose
3C|2-I.2.3.7,8-F«CDF
1.2.3,7.8-PeCDD
/U.. AJ
1280 I:M 1300 13^0
SCAN
Figure 3. Example of the PeCDF (m/z 338) and PeCDD (m/z 354) HRGC/MS-SIM
elution profiles in unspiked and spiked human adipose. The spiked con-
centrations for each isomer of PeCDF and PeCDD in these chromatograms
were 25 pg/g.
32
-------
374 .
Unspiked Human Adipose
|l,2,3,4.7.8-HxCDF
|l.2,3.6.7.8-HxCDF
2.3.4,6.7,8-HxCDF
374 .
Spiked Humai« Adipose
,2,3.4,7,8-H«CUF
1345 1363 1483
2,3.6,7.8-HxCDF
2.3,4,6.7.6-
,H,CDF
l.2.3.7,e.9-M«CDF
1238 13*8 1358 1488 1458 1580 1550 1688
330 .
sse .
Unspiked Human Adipose
1,2.3.4,7.8-MxCDD-
13Cl2-l,2.3,4.7,8-HxCDF
1,2,3,6,7.8-HxCDD
.7.6,9-HxCBD
Spiked Human Adipose
l.2.3.4.7.8-HxCDD
'3C|2-l.2.3,4.7,8-HxCDFl
1.2,3.6.7.8-HxCDD
1,2.3,7,8.9-HxCDD
13*8 1358 1468 1458 1580 1558 1698 1658 SCON
Figure 4. Example of the HxCDF (m/z 374) and HxCDD (m/z 390) HRGC/MS-SIM
elution profiles in unspiked and spiked human adipose. The spiked con-
centrations for each isomer of HxCDF and HxCDD were 62.5 pg/g.
33
-------
488 .
«24 .
Unspiked Human Adipose
,2.3,4,6.7.8-HpCDf
Spiked Human Adipose
,2,3,4,6.7,8-HpCDF
I 1.2,3,4.7,8,9-HpCDF
MSB ISM 1998 ISM 1698 1708 17W IBM SCAN
Untpiked Hunan Adipose
,2,3.4,6,7,8-HpCDD
Spiked Human Adipose
1.2,3,4.6,7.8-HpCDO
1908 \SX 1688 1639 1708 1798 1888 1656 SCAN
Figure 5. Example of the HpCDF (m/z 408) and HpCDD (m/z 424) HRGC/MS-SIM
elution profiles in unspiked and spiked human adipose. The spiked con-
centrations for each isomer of HpCDF and HpCDD were 62.5 pg/g.
34
-------
442 .
Unspiked Human Adipou
ll3c,2-OCDD
OCDF
442 .
Spiked Human Adipote
13C|2-OCDD|
17M
1758
isee
1958
2eee SCAN
456 .
Untpiked Human Adipose
OCDD
45« .
Spiked Human Adipoie
OCOO
1799
1S89
IB58
1388
20W SCAN
Figure 6. Examples of the OCDF (m/z 442) and OCDD (m/z 458) HRGC/MS-SIM
elution profiles in unspiked and spiked human adipose. The spiked con-
centrations for OCDD and OCDF were 125 pg/g each.
35
-------
CkD
M
ft
ft) 04
o
o
i
oo
ro S_
C\J ro
4-
o -a
(/) Q.
C. -I—
O i—
•(-> OJ
ro (/>
S- O
-(-> a.
OJ "O
(J ro
c
o c
<_) ro
E
i/l 3
3 -C
1/5
s- -a
OJ OJ
I/) C
C (D
O O)
4J
fO
i-
a
CO
O
E
O
a)
a
O)
o
c
o o
U -P
c
T3 -r-
OJ
s- -a
3 ai
to -^
ro •!-
O) Q.
ai
s-
3
p-unoj
36
-------
o
to
o
CO
k- °
W)
.2
(H
(H-^
O C
Is
0 ft
'1-4 i-H
D
h- O
o
a
o
ai
o.
i
oo
oo x
CSJ si
« -!->
i— I (T3
4-
O T3
in a_
C •!-
o i—
-P a)
re in
s- o
•!-> Q.
C -I—
o en
• r- O
->-> E
m o
u -u
c
o o
-a •!-
O)
S- T3
^ a>
in -^
(T3 •!-
0) Q.
S in
00
OJ
3
O5
37
-------
o
cu
o
o
o
00
Ml
o
CO
a a
_^ w ^^^
o
CD
O
in
o
T»«
O
n
o
CM
O
o
o
o>
o
00
o
I*.
o
CO
o
in
o
T|<
o
m
o
CM
Q
C2
O
X
:c
i
00
00 X
oo si
-+J
I— I H
in c
c ai
o O5
•i- O
4J e
n3 O
S- _C
4->
c a;
a> jz
L) 4->
c
o o
U -(->
c
"a •!-
O)
in
(TJ
O)
a.
01
0)
s_
cn
38
-------
o
CM
O
O
o \
00 00
ft
o *
." ai
ro in
s_ o
-p a.
C -I-
aj -a
u ro
C
O C
U 03
e
in rs
t- T3
ai ai
> (vl
c ai
o ai
•i- O
-i-1 £
(0 O
S- JC
u -P
c
o o
U -P
c
-a -r-
cu
S- T3
3 O)
in -^
TO ••-
ai a.
-------
CUI
eu>
a,
f* n
Is
0) A
o d
d *
o *
CJ °°
•a "2
•f-4 i—t
a
D
o
o
C_3
X
O1
00
*-.
r~-
PO X
CSJ i-
"-!->
rH (O
E
>v-
O T3
i/> a.
c ••-
o i—
+-> ai
Q.
C •!-
a) T3
U n3
c
o c
u m
in
i- -a
ai (U
> N
in
c
O
O)
o
C
o o
U 4->
c
•a -I-
ai
t-
in
re
a>
O)
3
C7>
OJ
(S/S
-------
a
a
a.
i
CO
00 X
CM i-
-4->
I— I H3
4-
o -a
•r—
cn Q-
C -I-
O r—
+-> 0)
ro in
s- o
•p a.
C •!-
N
•r—
in c
C 01
O 05
•i- O
•P E
ra o
S- ^
•p
C OJ
ai -c
u -P
C
o o
U -P
c
T3 -r-
OJ
i- T3
rj ai
in -^
TO -r-
QJ a.
c\j
OJ
S-
en
41
-------
00
a
So.
•»
a
CO
o
a>
a.
in
Q
Q
O
O
O X
S- -r-
•I-J Q.
C •!-
CU •—
(_)
c CD
O (/)
u o
a.
CO -r-
3 T3
t/> (T3
S_
0) C
> m
to ^
C -C
o
•i- -a
-t-> a>
(TJ N
t. •!-
+-> C
C 01
0) CD
U O
C E
O O
3
tO O
(0 +J
OJ C
OJ
1-
42
-------
cm
\
ss
t-.
•s »
CT;
CD CU
D
_ o
I
00
r«- x
on i-
oo re
e
4-
o -a
in Q_
c •<-
O r—
•r-
-t-> Q)
fO U)
S- Q
-(-> a.
C -r-
ai "o
(J M
•i—
in c
c ai
o 01
(T3 O
s_ ^:
-P
c a;
OJ £1
(J 4->
c
o o
u +->
c.
-D -r-
ai
S- T3
3 O)
in ^
TO •!-
O) O-
<3-
rH
OJ
43
-------
o
CJ
0)
Q-
I
00
oo
00
ft
d w
Is
^. I •
£ ft
ss
o
O
at
"s
a
CM
•4J
rH n3
4-
o -a
I/) Q.
C •!-
O r—
->-> ai
(O (/I
i- O
•(-> Q.
a) -a
u N
in c
c a;
O O)
•i- O
-M E
(C O
S- .C
4J
C 0)
<1J ^
U +->
c
o o
c
-a -i-
QJ
i- -o
OJ Q.
OJ
S-
o>
O
CO
o
in
o
<*
o
n
o
CM
pxinoj
44
-------
0$
O
H.2
gft
^S
a
o
0
O)
Q_
00
OO S-
--M
CM fB
o -a
in Q.
C -i-
O r—
•P 0)
N
•r—
in c
C 0)
o a>
•r- O
•P e
(O O
t- -C
C 0)
U -P
C
o o
U -P
T3 •<-
CD
i- -a
in
to
OJ
Q.
ui
o
a>
o
00
o
CO
o
in
o
CO
o
CM
OJ
s_
3
O)
•r~
LU.
45
-------
o
CM
o
o
o
00
M
a
.2
o
CO
0)
>
0
CO
o .*.;
^tt -i-l r
a
CO
a
o
CM
a
o
x
z
i
03
CO X
CSJ ll
"4->
i— I N
C 0)
O U3
•i— O
+-> e
ID O
S- -C
C 01
ai -c
u -i->
c:
o o
U -(->
-a T-
ai
S- TJ
^ ai
to -^:
(13 T-
ai Q.
s: to
QJ
S-
CD
O
CO
O
IO
o
CO
o o
CM -*
O
O
O
CD
O
CD
o
CO
o
CO
o
CM
46
-------
D
x
:c
i
00
n x
c\j tl
' -i->
r-l ai
ro u)
5- O
+-> a.
c -P-
OJ T3
u m
c
o c
U ra
E
1/5 ^
S_ T3
HI OJ
> N
tn C
C CD
o a>
•i- O
4J E
rD O
S- -C
+J
c ai
ai JZ
u -(->
c
o o
U 4->
CU
-a
ai
O> Q.
s: ui
06
rH
cu
CT!
47
-------
48
-------
cut
oo
a>
s
o
8ft
I-* •<-<
GO
D
a
CJ
X
^
I
00
I—
CO X
CSJ 5-
i—* O3
e
o -a
•r—
u) a.
c ••—
O r—
(T3 (/)
i- O
-P CL
C •!-
3
3 ^
I/)
S- T3
ai at
> N
in c
c ai
O O)
•i- O
-*-> E
(0 O
a; _c
u +->
c
o o
U -i-J
C
T3 T-
ai
s- -a
n ai
iy) ^
(0 ••-
ai o.
o
S-
3
O!
49
-------
D
o
Q.
I
oo
r-
«*•
CO X
CM il
i—I ro
E
1/1 Q.
C •!-
O i—
(O i/l
s- o
+J Q.
O) T3
O fO
c
o c
O ft)
5- T3
Oi oi
> N
to C
C
•i- O
+-> E
rO O
C O)
oj -c
U -4->
o o
O +J
S- T3
3 O)
in J*i
fO ••-
O) Q.
S 10
OJ
-------
\
eo
A
st
(H
•s a
cj I-H
£ 0,
o
u
D
O
C_>
a.
I
00
ro X
c\j si
- +->
rH (T3
in Q.
C T-
O •—
+-> HI
(0 U5
s- o
+-> a.
C 'f—
aj ~a
L> m
c
o c
CJ n3
£
10 3
3 .C
l/)
s- -a
ai ai
> N
10 C
c a>
o en
•i- O
+-> E
(T5 O
i- ^z
+J
C OJ
ai ^
u -p
c
o o
U 4->
c
T3 •!-
OJ
t- T3
3 OJ
O) Q.
s: m
CM
CM
ai
i.
o
CO
o
10
o
CD
o
o
o
en
o
00
o
CD
o
in
o
CO
o
CM
51
-------
cut
Ml
rj n
Is
o
D
T3
OJ
o
o
l«-
O X
in il
C -P
O ro
(D
S-
tu •—
u
c 01
O v)
u o
Q.
IS> -i-
3 -a
in re
s_
ai c
> ro
E
U1 3
C -C
o
•r- -a
4-> OJ
(T3 N
S- •!-
-U C
c ai
O) O5
u o
C E
o o
(J ^z
T3 OJ
<1J ^Z
S- +->
3
in o
03 -P
O) C
no
CM
OJ
s_
en
52
-------
The slopes of the calculated regression lines from the data points
in each of the 14 analyses can be used as an indication of the accuracy of
the analytical method for the 17 target analytes. Figure 24 is a plot of the
slope of regression lines versus the 17 individual compounds. Table 16 pro-
vides a key to specific compounds associated with a number on the x-axis of
this plot. The plot presents the estimated slope from each least squares re-
gression line as well as the upper and lower 95% confidence limits for the
slope. The slope of the regression line can be interpreted as a measure of
accuracy with a value of 1.00 equivalent to 100% agreement of the measured
concentration with the theoretical values (background plus spike level). The
plot of the 95% confidence limits presents some confirmation on the precision
of measurements across the four spike levels. These confidence bounds can
also be used to determine whether the accuracy of the measurements (slope of
regression line) is significantly different from 100% (or 1.00). If the ver-
tical line connecting the lower and upper 95% confidence limits intersects
with the horizontal line at 1, then the accuracy of the method (as determined
from the regression line) is not significantly different from 100% (slope =
1.00). The results plotted in Figure 24 demonstrate that the method accura-
cies for 7 of the 17 analytes are not significantly different from 100%.
On the other hand, if the upper and lower confidence limits are both
greater than or both less than 1.00, then the accuracy of the method is sig-
nificantly different from 100%. The data presented graphically in Figure 24
indicate that some positive bias (greater than 100%) is associated with the
method accuracies for 9 of the 17 analytes while the measurements for a single
analyte (OCDF) result in a slightly negative (less than 100%) bias.
Table 16 provides a key to the compound identification in Figure 24
and tabulates the slope of the regression lines and the upper and lower 95%
confidence limits for each of the 2,3,7,8-substituted PCDD and PCDF analytes.
As noted from Table 16, method accuracy (as defined by the regression line
slope) ranges from 90% for OCDF up to 121% for 1,2,3,7,8,9-HxCDF. The accu-
racies for all other measurements fall within a range of 97 to 115%. The
overall method accuracies meet the initial accuracy objective of 50-115%
identified in the project quality assurance program plan. However, the pre-
dicted accuracy results for individual analysis as defined by the 95% upper
confidence limits indicate that this range should be adjusted to 50-130%.
The bias in the accuracy of the measurements may be a result of
slight differences in the concentration calibration standards and the internal
quantitation standard and native PCDD and PCDF spiking solutions. As a pre-
liminary check on these differences, solutions of the low level and of the
high level native spike combined with the internal quantitation standards were
analyzed. The results of these analyses are provided in Table 17. Accuracy
was calculated as measured/spiked x 100.
The results of these analyses suggest that bias observed in overall
method accuracy is attributed to the differences in the spiking solutions ver-
sus the calibration standards. For instance, the four HxCDF isomers demon-
strated a consistent positive bias to method accuracy based on the least squares
regression analysis. The analysis of the spiking solutions, submitted as sam-
ples, also indicates a definite positive bias for the same four HxCDF isomers.
53
-------
ACCURACY ESTIMATES
1.4
o
c
c
o
'55
«
£
o>
£
-------
Table 16. Regression Line Slopes with 95% Confidence Limits
Compound
no.
01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
Compound
2,3,7,8-TCDF
2,3,7,8-TCDD
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,8-HxCDD
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
1,2,3,4,6,7,8-HpCDD
OCDF
OCDD
Slope
1.08
1.13
1.07
0.98
1.04
1.07
1.12
1.12
1.21
0.98
1.01
1.12
1.01
0.97
1.08
0.90
1.15
Significantly
different
from 1.00?
yes
yes
yes
no
no
yes
yes
yes
yes
no
no
no
no
no
yes
yes
yes
Lower 95%
confidence
limit
1.04
1.08
1.02
0.82
0.98
1.06
1.09
1.09
1.12
0.92
0.86
0.94
0.95
0.95
1.01
0.88
1.00
Upper 95%
confidence
1 imit
1.11
1.19
1.12
1.13
1.11
1.09
1.16
1.15
1.29
1.05
1.16
1.30
1.07
1.00
1.16
0.92
1.30
55
-------
in
C
0
•r~
^_>
3
r—
0
CO
CU
-^
Q.
CO
cu
.^
i ^
re
z
^_
CU
>
cu
x:
S
•o
c
re
3
O
5
>+_
o
in
• r—
in
>.
^—
re
c
cu
^—
Ol
•r"
a:
.*:
•r—
CL
tfl
r—
QJ
^—
3
o
i
u
re
3 ^«^
(J >»^
O
C
O
-O 4-> /— s
ai re — i
1- S- 3.
3 •!-> \
in C O)
re cu cx
CU D ^^
s: c
o
u
c
o
-p /->
cu re — i
•^ S- 3.
•^ ^-J ^^
Q. C O)
co cu a.
(J s_^
c
o
u
(J
re
s- /-\
^ ^?
u
cu re _ i
s- s- a.
3 4J \
in C D)
re cu a.
cu o ^^
2! c
o
o
c
o
• r-
-)-> /-^
cu re — i
^ J- 3.
•r- -K> ^^
Q. C D5
CO CU Q.
O '•^-^
c
o
u
•^
c
3
O
O-
o
o
CO *t
O rH
rH rH
^f r***
LO LO
O 0
LO LO
o o
CO CM
rH rH
CO CM
rH rH
O O
rH rH
LL- O
Q 0
1— I—
00 00
r^ r^
CO CO
CM CM
^ co LD
o en o
rH rH
CM cn co
LO «3- LO
0 0 O
LO LO LO
O O O
rH O CM
rH rH rH
rH O CM
rH rH rH
000
rH rH rH
LJ_ LL. O
000
O O O
ai cu cu
a. a. a.
i i i
CO CO CO
r-.r-.r-~
CO «S- CO
CM CO CM
rH CM rH
r-- o CM co co ID co
O rH CM <- Cn O rH
rH rH rH rH rH rH
*d- r-~ CM LO co CM oo
CO CO LO 00 CM CO <3"
rH rH rH rH rH rH rH
LO LO LO LO LO LO LO
CM CM CM CM CM CM CM
rH rH rH rH rH rH rH
CO GO O O CO ID ID
o CM ^- ID o en LO
rH rH rH rH rH rH
r^ CM LO o r-« *^* cn
CM CO CO «*• CM CM CO
LO LO LO LO LO LO LO
CM CM CM CM CM CM CM
U_ LL. U_ U_ O O O
o o o o o o o
o co o o c_j o o
X X X X X X X
^ i 3; a: ^ a: ni
i i i i i i i
oo oo co cn co oo cn
r-- r— . > r-. oo r-* r** co
«t ID ID r- ^j- ID r->
CO CO ^" CO CO CO CO
CM CM CO CM CM CM CM
rH rH CM rH rH rH rH
co r~* ^~
cn cn o
rH
ID rH O
rH CM CO
rH rH rH
LO LO LO
CM CM CM
rH rH rH
00 O «3-
CO O O
__J __J
r^n m
CM LO ID
CM CM CM
LO LO LO
CM CM CM
u. u_ o
O O 0
o co o
Q. Q. Q.
3: ^ 3i
i i i
co cn oo
»* *. *.
1 — CO 1 —
ID r*~ ID
«fr *fr «*
co co co
CM CM CM
rH rH rH
cn o
cn o
rH
f*1^ f*"^
•51- LO
CM CM
0 0
LO LO
CM CM
00 ID
co cn
t 00
o o
LO LO
1 1 f~*\
LJ_ LJ
0 O
f J f *
0 0
56
-------
Similar trends are noted for other compounds in Table 17 compared to the data
presented in Figure 24 and Table 16.
The limited number of analyses of the spiking solutions does not
provide an adequate comparison with the sample data to confirm the bias.
However, it is recommended that at least triplicate measurements of the spik-
ing solutions at each fortification level should be analyzed at the outset of
the actual NHATS sample analysis program. This will be necessary to account
for any biases that will be observed from the determination of PCDD and PCDF
residue levels in spiked QC samples. It should be noted that additional
homogenized spiked samples will be prepared prior to initiation of the NHATS
sample analyses.
1. Recovery of Internal Quantisation Standards
The absolute recoveries for the carbon-13 labeled internal quanti-
tation standards were determined for each sample by comparing the responses
to the internal recovery standard, 13C12~1,2,3,4-TCDD. The average recoveries
of the compounds in Tables 10 to 15 range from 52.1% for 13C12-2,3,7,8-TCDD
up to 88.9% for 13C12-OCDD. The results for the absolute recoveries com-
pared to the overall method accuracy for each compound indicate the importance
of the internal standard quantisation technique for analysis of the PCDDs and
PCDFs in human adipose.
2. Estimation of Background Levels of PCDDs and PCDFs
The estimated background levels of the various PCDD and PCDF con-
geners were determined as the intercept obtained from the least squares linear
regression analyses. Table 18 provides a comparison of the average measured
values for the unspiked matrix and the background concentration estimates from
linear regression analysis of the data. In general, the measured and esti-
mated background levels are in good agreement. However, several analytes with
concentrations of less than 5 pg/g, particularly OCDF, demonstrate some dis-
agreement in the measured versus estimated concentrations. This apparently
arises from the fact that the first spike level is significantly greater than
the actual background concentration. In the case of OCDF, the first spike
level estimated by linear regression was 50 pg/g compared to an average mea-
sured value of 3.2 pg/g. In order to provide a better estimate of the back-
ground level based on the linear regression analysis, additional spike levels
between 5 and 50 pg/g would be required. Table 18 also provides the upper
and lower 95% confidence limits for the background levels estimated by the
linear regression analysis of the data. These estimated background levels
and confidence limits can be viewed as the intersections of the regression
line and its upper and lower 95% confidence bounds, respectively, with the
y-axis (measured or found concentration). These values will be used as the
initial data points for developing control charts of the unspiked lipid ma-
trix which will be analyzed with each batch of samples throughout the EPA/VA
study.
57
-------
a>
&« u
Lf) C /->
cn ai -p CD
•a •<- ^^
S- ••- E Ol
OJ M- -r- Q.
Q. C r- v-/
Q. o
Z5 (J
cu
u
c
HI
•a
c
o
cn
.c
-P
•r~"
3
in
cu
•P
ro
01
>
ai
o
s-
O)
(0
ca
00
OJ
&« U
in c
OJ
T3
S_ -i-
OJ >+-
3 c
o o
—I U
O)
O)
a.
•P -P
c c
r— (O OJ OJ
OJ U
Oj
N
l(- O
C -I- i.
CTT3 M-
•D
CU
-P
ro
in (0
O ' — O)
S- OJ >x
cn > O)
.* OJ Q.
U r— ^^
•a
T3 C
OJ
S-
(0
-3 r— ^->
O OJ O5
i- > \
OJ OJ O)
(0 -* r— Q.
0) U •^-<
in
TJ
C
3
O
a.
o
•a
c
3 .
O O
a. c
o
o
ur>
oo o
CMCNJCsJi— I
CVJ^DCOCM
CSJ
ro o
CVJ
00
cn
CM
CvJ
00
O
CM
inin ininininin ininwin
ojajocuojojojojocycuojcu
o
c
in
cu
in
cu
tD rH CO
• oo •c\joo<3-c\jro -cn^Lnr-v
CO-rH CSJ....
^-^ i—l **** CM O^ CSJ rH ^ ~<-s ^ CTl ID LO
i-H CMrHCMrH CMuDCMCM
Q Q Q rH
ro
o
CM
cn
U
Q
CD
LT)
CM
CM
o • . • •
- — -i— ii — rooo
CMLDCMCM
Q rH
ro
i—I
CM
CM
ro
O
00
U_ LL. U_ U_ O O O
O O O O O Q O
U_U_QC_}C_3OCJOC_)O
aaaxxxxxxxi
000x0:^x0:^x00
U-OCUCUOJIIIIIII
aoa.a.a.oooococnoooocnr^
I
cn
oo
oo
- -Q -a -a
r~-(^rO'«i-rocoro«^-rororororoocooooc_5
M«.MM#i*t»sA*sAM*s«. Q_ m (^ «i (~t I I f^^t
OOfOCMrOCMCMCMrOCMCSJCMCMCMXCMXCMXQQ
----«"»••-»-'•- « « OO
CMCMi—ICMrHrHrHCMrH—IrHrHrH rH r-1 OO
OOOOOOOOO
UD
i.
•P
ro
•a
cu .
.* ro
•r- -P
a. ro
c
3 t-
O
OJ
O i—
ro
m c T3
cu ro o
m _c
>, c -P
i— O 0)
ro *r" E
c in
ro in "o
OJ CU
J- -P
cj) ro
ro cu E
O J- •!-
•<- -P
r- S- U5
a. ro cu
•r- OJ
S- C OJ
-P -r- £
01
CU
£ aj in
•P U
<4— OJ
o e i—
O if—
in s. a;
0) ^ S-
O)
ro "o in
S- OJ OJ
oj > m
> •<- OJ
CU -P
cu -a c
J= (D
-P OJ S-
s- ro
cu cu Q.
ro c:
in -I—
C/5 i—
i— cu ai
cu > 3
•> cu i—
Q)I— ro
•a
T3 C OJ
C 3 .C
301—
O S-
U CT
O)-^ •
-*Z U "O
-------
3. Day-to-Day HRGC/MS Analysis Precision
In addition to the analysis of the replicate spiked samples, four
extracts were analyzed by HRGC/MS on two different dates. The results of the
duplicate HRGC/MS analyses of these four samples for the 17 target compounds
are presented in Tables 19 to 21. Concentration values from the second analy-
sis date were included in the statistical analysis of data presented earlier
in this section.
59
-------
Table IB. Day-to-Day Precision of Analysis of Specific Sample Extracts
for Tetra- and Pentachloro PCDF and PCDD
Analysis
date
4/22/86
4/28/86
RPD (%)b
4/22/86
4/28/86
RPD (%)
4/22/86
4/28/86
RPD (%)
4/22/86
4/28/86
RPD (%)
Spike
level
(pg/g)
0
0
0
0
0
0
10
10
2,3,7,8-
TCDF
(pg/g)
ND (3.0)a
ND (4.1)
29
ND (3.1)
ND (4.1)
28
ND (3.3)
ND (4.0)
19
12
14
18
2,3,7,8-
TCDD
(pg/g)
10
11
10
10
11
10
10
13
26
21
23
9
1,2,3,7,8-
PeCDF
(pg/g)
ND (0.84)
ND (1.12)
29
ND (0.78)
ND (0.75)
4
ND (0.81)
ND (0.75)
8
11
12
8
2,3,4,7,8-
PeCDF
(pg/g)
24
21
13
23
22
4
16
19
17
26
28
7
1,2,3,7,8-
PeCDD
(pg/g)
18
20
11
17
20
16
17
18
6
27
19
35
,ND = not detected. Value in parentheses is the estimated limit of detection.
Relative percent difference. Calculated as the difference of the two values
divided by the mean of the two values times 100%.
60
-------
o
o
Q_
TJ
C
TO
U_
O
Q_
o
S-
0
J:E
u
TO
CL
OJ
x
"O
C
TO
X
0)
X
t-
o
V*-
Ul
^_>
U
TO
S-
X
LU
OJ
CL
E
TO
00
u
£
• r—
U
OJ
Q.
i/)
0
in
t/l
>^
CO
c
kf_
o
c
o
in
u
OJ
Q_
>^
TO
O
i
O
t
TO
Q
O
CM
CU
.a
TO
h—
*f l Q ^"*«
- oo o cn
ro - o "x.
-r-. a. O)
CM -x a.
r-H
^" 1 U_ *~^
» cn o cn
-CO Q. C3J
CM -x a.
* r*- •--'
rH
.
-CO Q O)
CO -O X.
-r*. a. cn
CM -X QL
rH
- I Q /->
CM 00 X Cn
- -x a.
rH P- ^-s
• i a — s
ro cO ^3 CT
- -0 X.
CM f* x cn
. -X Q.
rH V£> ^
- t 0 ^
ro CO Q C?)
- "O^x.
CM r- x CT
* "X CL
- 1 U_ '•"N
ro cn ^3 CJ)
- -x a.
- I U_ *~*
*r co o OT
ro r^ x cn
- *x a.
" 1 U_ **^
ro 00 O CT
- - C_) X.
CM f^» X CJ)
- *x a.
rH UD ^— '
- 1 U. •—*
ro CO Q CD
- -ox.
CM r^ x rj>
- -x a
cu i— cn
-ii CJ X,
•- > o>
a. cj Q.
r- S^
Ul
tn flj
r— fQ
c
<=:
O U3 ro
rH rH
CM CM
ro vo
CO ro
0 0
cn
o a r*-
Z Z
CO rH 0
CM ro rH
rH rH rH
O^ O ro
rH t— (
O CM O
CM CM rH
TO
ro in
O O ro
O O
CM cn
*3" ^ ^n
rH
CM CM O
rH i— (
rH CM in
CM CM
O O
_Q
CO CO &-S
CM CO
CM CM £3
\ X. Q-
*t vf ce
r»^rH CM rot^ ro mrH ro
O rH CM rH PO *3"
Csj CM CM CNJ CM CNI
cn «3~ <^ co
O rH O CM
r-i rH rH O
Q- ^D ro m CO
Q O Q O rH CM CM
Z Z Z Z rH
in r** co r^- CTI r^- ^ cn ic
CM CM CM CM ^ ^
co^ouo cOtDro usrorH
i-HCMro rHCMro ^-voro
OOOf^ uDrHrO rHrH^O
tn r** rH r^ r**- r*» 00
rH rH rH rH rH rH
rH ro O"> rH o in ro cn ro
CM CM C\l CM ^- ^ rH
CNJ ^" rH ^
ro r^» ^ CO
OOCn OOrH OLDLO
Q O O O
z z z z
O CM cn ro
^' ^ in ro ^r o CO CM ro
rH CM ro rH
CMCMO CNjroco uor^-ro
r-l rH rH rH ro ro
CM CM O O rO «J" *J" r** {D
CSJ CM CM CM rH Q- ^>
o o o o m LO
CM CM
OOCO&€ COCO^ COCO&^
CM CO CM CO CM CO
CMCM Q CMCM O CMCM Q
•x~xa_ •^v-^,a_ x. x. cx
^" * ^ ^ pf*
8
O
rH
in
OJ
£
+J
t/t
OJ
3
TO
O
OJ
Jj
>*—
o
c
a>
E
0)
XI
-Q
at
c >
•r- T3
(j in
-P D
OJ r-
"O fO
(4—
o o
•p-
E OJ
•t- jr:
l~"
•o «*-
OJ O
ro CU
£ U
•r- C
4-* CU
in u
flJ OJ
M—
OJ M—
5 '-5
in cu
•r- ,C
in
CU >/)
in ro
OJ
£ -D
4-J OJ
C 4->
OJ ro
I. r-
CL U
C fO
•1- (_)
aj
3 .
r- OJ
> cz
4-> C
OJ U
CU OJ
•0 D.
o •>
4-1
II (0
Q CU
TO ^
61
-------
Table 21. Day-to-Day Precision of Analysis of Specific Sample Extracts
for OCDF and OCDD
Analysis
date
4/22/86
4/28/86
RPD (%)a
4/22/86
4/28/86
RPD (%)
4/22/86
4/28/86
RPD (%)
4/22/86
4/28/86
RPD (%)
Spike
level
(pg/g)
0
0
0
0
0
0
50
50
OCDF
concentration
(pg/g)
4.9
3.5
33
2.2
2.3
4
1.9
2.6
31
43
44
2
OCDD
concentration
(pg/g)
811
810
0.1
819
836
2
788
784
1
834
848
2
aRelative percent difference. Calculated as the difference of the two
values divided by the mean of the two values times 100%.
62
-------
VI. QUALITY ASSURANCE/QUALITY CONTROL (QA/QC)
As discussed in the experimental section of this report, the QA/QC
activities included the analysis of a multipoint calibration curve, daily
verification of relative response factors for each analyte, analysis of a
method blank and reagent blanks along with the samples, and determining the
absolute recoveries of each of the internal quantitation standards for every
sample. Each of these QA/QC activities is discussed below.
A. Initial Calibration
At the outset of sample analysis activity, six calibration concen-
tration standards containing each of the target PCDDs and PCDFs at varying
levels and constant concentrations of the internal quantitation and recovery
standards were analyzed in triplicate. The relative response factors (RRF)
for each native compound and internal quantitation standard were determined
for each standard analysis. An average RRF and relative percent standard
deviation (RSD) were determined for each concentration level. The average
RRF values from each of the six concentration calibration standards were then
used to calculate a grand mean RRF value for each compound in the calibration
solution. Table 22 presents a summary of the grand mean RRF values for each
component in the standards. As noted from Table 22, the average RRF values
for native PCDDs and PCDFs generally varied by less than ± 10% (RSD) with the
exception of the pentachloro congeners. These results fall well within the
criteria established in the draft quality assurance program plan which re-
quired the variability of RRF values for the tetrachloro homologs to be within
± 20% (RSD) while the RRF criterion on all other compounds was set at ± 30%
(RDS).
The variability of the RRF values for the internal quantitation
standards, on the other hand, was noted to increase with the degree of chlo-
rination. This is a result of the measurement of all internal quantitation
standards versus the single internal recovery standard, 13C12-1,2,3,4-TCDD.
A second internal recovery standard, 37Cl4-l,2,3,4,6,7,8-HpCDD, was evaluated.
However, problems resulting from contribution of native HpCDD to the charac-
teristic ions of this internal standard resulted in variabilities in the RRF
value up to 50%. Hence, this internal standard was not used for any calcula-
tions. It is anticipated that an additional internal recovery standard, such
as 13C12-l,2,3,4,7,8-HxCDD, will reduce the variability in the RRF values of
the higher chlorinated internal quantitation standards. This compound will
be incorporated into the method if available.
The sensitivity of the Kratos MS-50TC to the tetra- through octa-
chloro PCDDs and PCDFs was demonstrated through the triplicate analysis of
the low level standard (CS-8, Table 2) that ranged in concentration from
1 pg/(jL for the tetra- and pentachloro congeners up to 5 pg/uL for the octa-
chloro congeners. These solution concentration values of 1 pg/uL and 5 pg/pL
correspond to residue levels in tissue of 1 pg/g and 5 pg/g, respectively.
Table 22 provides an indication of the observed signal-to-noise ratio for
each of the native PCDD and PCDF congeners. These data demonstrate that the '
low level standard is well above the instrument detection limit, which is de-
fined as the amount of a particular compound necessary to give a signal 2.5
times the background signal to noise for each of the characteristic ions while
meeting the qualitative criteria for ion ratios.
63
-------
Table 22. Relative Response Factors (Grand Means) Determined from
Multipoint Concentration Calibration Standards
Compound
RRFa RSD (%)
RRF
contro^
limits
Signal-to-noise
ratio for low
level standard0
Calibration
range
(pg/pL)
2,3,7,8-TCDF
2,3,7,8-TCDD
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
1,2,3,4,6,7,8-HpCDD
OCDF
OCDD
13Ci2-l,2,3,4-TCDDd
laCi2-2,3,7,8-TCDF
13Ci2-2,3,7,8-TCDD
13Ci2-l,2,3,7,8-PeCDF
iaCi2-l,2,3,7,8-PeCDD
13Ci2-l,2,3,4,7,8-HxCDF
13Ci2-l,2,3,6,7,8-HxCDD
iaCurl,2,3,4,6,7,8-
HpCDD
a'Cl4-l,2,3,4,6,7,8-
1.00
0.80
0.98
1.06
1.33
0.94
0.93
0.86
0.86
.31
.44
1.
1.
1.61
2.
1.
1.
1.
,33
.89
19
,38
1.04
1.
1.
1.
1.
,00
,98
73
36
0.70
1.28
0.41
0.33
0.12
5.7
6.2
5.2
10.1
11.3
3.1
2.4
2.7
6.9
4.9
3.0
1.0
4.0
3.6
4.7
3.3
2.5
7.6
4.7
4.5
7.7
15.8
19.6
25.8
51.8
0.24 28.0
0.80-1.20
0.64-0.96
0.68-1.28
0.74-1.38
0.93-1.73
0.66-1.22
0.65-1.21
0.60-1.12
0.60-1.12
0.92-1.70
1.01-1.87
1.13-2.09
1.63-3.03
1.32-2.46
0.83-1.55
0.97-1.79
0.73-1.35
1.58-2.38
1.38-2.08
0.95-1.77
0.49-0.91
0.90-1.66
0.29-0.53
0.23-0.43
0.17-0.31
12
6.5
11
8.9
5.7
32
30
29
17
13
14
14
35
26
13
31
21
2.
2.
2.
2.
2.
1-100
1-100
1-100
1-100
1-100
2.5-250
2.5-250
2.5-250
2.5-250
2.5-250
,5-250
.5-250
.5-250
5-250
5-250
5-500
5-500
50
50
50
50
50
125
125
125
125
250
?RRF = grand mean RRF.
RRF control limits designate the acceptable range of values based on the criteria for
± 20% of the MF for 2,3,7,8-TCDD and 2,3,7,8-TCDF and ± 30% of the RRF for all other
PCDD and PCDF compounds.
Value for signal-to-noise ratio based on observed response for the major
characteristic ion for each native PCDD or PCDF congener (Data File
,8501017X02).
Internal recovery standard.
64
-------
B. Daily Verification of Response Factors
Before proceeding with analysis of samples, the analyst was required
to verify the existing response factor calibration through the analysis of a
calibration standard (CS-7, Table 2). Criteria for proceeding with sample
analysis required that the measured RRF value for 2,3,7,8-TCDD and 2,3,7,8-TCDF
were within ± 20% (and all other congeners within ± 30%) of the mean RRF es-
tablished from the calibration curve. This standard was also analyzed at the
end of each working day to demonstrate that the calibration had been maintained.
All RRF values were tabulated to generate RRF control charts for each specific
PCDD and PCDF congener.
Figures 25 through 34 are plots (control charts) of the RRF values
established for the 17 individual target analytes. The RRF data are plotted
versus time of analysis. These plots contain 28 individual data points, 18
of which were generated for triplicate analysis of 6 concentration calibration
solutions from initial calibration and 10 analyses of solution CS-7 (Table 2)
injected over the 5 days for which actual samples were analyzed. The upper
and lower boundaries (dashed lines) represent a relative standard deviation
of approximately ± 10% with the exception of the plot for 1,2,3,7,8-PeCDD,
for which the boundaries are plotted as ± 20%.
It should be noted that the actual control limits as specified in
the project QAPP were set at ± 20% for 2,3,7,8-TCDD and 2,3,7,8-TCDF and ± 30%
for all other target analytes. Boundaries of ± 10% have been used in Figures
25 through 34 as a means to provide the reader with a better perspective in
the actual distribution of the measured calibration points. The values for
the acceptable ranges of each PCDD and PCDF compound based on the initial cal-
ibration are presented in Table 22. The data presented for the RRF values in
Figures 25 through 34 are well within these established control limits. The
average RRF values and corresponding standard deviations reported in each of
these plots are calculated from the total 28 standard analyses.
C. Blanks
As specified in the quality assurance program plan, a laboratory
method blank was prepared along with the 14 human adipose lipid samples. The
method blank was taken through all procedures as if it were an actual sample,
although no lipid matrix was introduced. The analysis of the method blank
resulted in the data reported for each of the target analytes reported in
Table 23. As noted in Table 23, 1,2,3,4,6,7,8-HpCDD and OCDD were detected
at concentrations equivalent to 4.0 and 30 pg/g (equivalent to a 10-g lipid
sample), respectively.
The contribution of these PCDDs were not subtracted from the observed
responses for the spiked and unspiked samples. These background levels account-
ed for less than 2% of the 1,2,3,4,6,7,8-HpCDD and less than 4% of the OCDD
measured in the unspiked lipid samples. In addition to these compounds,
responses that correspond to the elution of two TCDD isomers (1,3,6,8- and
1,3,7,9-) and a PeCDD (isomer not determined) were detected in the method
blank.
65
-------
P! 2,3,7.8- TETRflCHLCKW-RlRHll
F!C13-2/3,7,9-TETRACH.ORO-FIJRftN
Efl/fiEF.f*£fl)/
-------
CMPil,2,3,7.8-PEN
-cc.m-1 7 T.7iS
(ftREA/KEF.flKEflVC
1.299
1 . 100
1.009
e.see
0.300
A ?no
rftCHUORO-fURAH
-FQffftCHUKO-FURftM
WT./REF.AMT.) (AUi 9.971)
^
! 4
X '
i i
T i •*
n i
II vj,
II f.
II «.
II ft V
4> II V X*
* • ,T ?!
H x x r
if * x
ii ^
Tx M
! x
i
x
ST.OEU.»
9.935
Z ST.OEU.=
8.759
DftTE
4/13/86
4/23/86
5/ 3X86
CMFs 2,3,4,7,8-FEHTflCHLORO-FliRftH
,:EF:C13-1,2,3.7,8-FEHTrtCHLORO-rlJKflH
/(-EF.(W
1.4W
1.300
1.290
1.1W
1.9W
3.308
8.800
I.B53)
v "
»
I
X
ST.OEU.=
0.113
7. ST.DP.'.'
10.601
OftTE
4/13/86
4/23/85
5/ 3/86
Figure 26. Control charts showing response factors by date for 1,2,3,7,8-PeCDF
and 2,3,4,7,8-PeCDF. Dotted lines represent approximately ± 10% of the mean.
67
-------
CN>! 1,2.3,7,8-PEHTflCHLORO-OIOXIN
SEFiCl5-I,2,S-7.8-FEHTflCHLORO-OIOXIN
/(l»1T./SEF.fll1T.> ((W:
3.060
1.375)
2.069
ST.DEU."
8.233
7. ST.DEU."
DATE
4/13/86
4/23/86
V 3/8S
Figure 27. Control chart showing response factors by date for 1,2,3,7,8-PeCDD.
Dotted lines represent approximately ± 20% of the mean.
68
-------
CMP! 1 - 2,3,4,7.8-HEXACH.Of?0-FL'Rf»l
*EF! C13-1.2 • 3,4,7,3-HEXACHLORO-FJF. «N
<«Eft'P.EF.f*EflV(flf1T./REF.lWT.>
-------
CTPs 2,3,4.6,7.8-«EXACHUTRO-FURftN
.5EF:C13-I,2,3,4,7,8-HEXflCHLOFO-FURAN
-------
CWiUl-HEXACHLQRO-OIOXIM
-EF:C13-I,2,3,6,7,8-HEXftCHLOEO-OIOXIN
«W?Eft/REF.f«B»Aftl1T./REF.flt1T.) (flVs
1.509
1.394)
1.4PB
1.139
DATE
4/13/85
I
4/23/85
ST.DEV.*
9.877
2 ST.OEir..
5.31"
5/ 3/85
C.1P:b2,3,6,7.
SEFiC13-l,2,3,
iIOXIH
" "I-OIOX1H
(«REfl/REF.(«EA)/(fl«T./REF.«1T.)
1.788
1.433)
X
0.036
Z ST.OBJ.-
6.015
i.see
*Y X *
11
f
• « I *
F*
1.408
'X
1.300 *
1.208
DftTE 4/13/86 4/23/86 5/3/86
Figure 30. Control charts showing response factors by date for 1,2,3,4,7,8-HxCDD
and 1,2,3,6,7,8-HxCDD. Dotted lines represent approximately ± 10% of the mean.
71
-------
1.657)
2.200
2.1??
2.808
i.sw
1.79?
1.508
I. SOT
1.499
MTE 4/13/86
4/23/9S
Figure 31. Control chart showing response factors by date for 1,2,3,7,8,9-HxCDD.
Dotted lines represent approximately ± 10% of the mean.
72
-------
i.see
1.489
OflTE 4/13/95
X
CMPi!, 2,3- 4,6,7,8-HEPTflCHLORO-f UR«I
SEP:C13-1.2,3,4,6,7,8-HEPTftCHLORO-OIOXIN
dWEA/REF.(*EA)/X«1T./REF.«1T.) ~
-"V-^f rr-| r-
2.289 H 'i ...
IX £ *
I
2.100 -I *
i. see
y
1.330
DHTE 4/13.-8S 4/23.35
CMF!l.2.3,4,",8,3-HEPTftCHLORO-RJRfiN
.:EF!C13-1,2,3,4,6.?,8-HEPT«CHLORO-OIOX!N
2. ICO
X ST.DEU.'
' 9.159
' 7. ST.DP.'.»
•? 9.42C
^{_
•X *
ID I
.'? xW
1.3??
1.39?
i
i
i
NT „,
i.ree -| x
^ X
1.S00 -I S x-
4/23/95
5/ 3/86
Figure 32. Control charts showing response factors by date for 1,2,3,4,6,7,8-HpCDF
and 1,2,3,4,7,8,9-HpCDF. Dotted lines represent approximately ± 10% of the mean.
73
-------
CMPil,2,3,4,6,7,8-HEPTimORO-OI
.?EFiC13-l,2,3,4,6,7,8-HEPTflCHt.r-
<(*EA/R£F.ftREfl)/(flf1T.'-REF.fl«T.)
1.466
IIOXIH
1.175)
i.see
i.ioo
1.M9
ST.OEV.'
8.971
z ST.DF.'.=
6.343
*t
!x
e.sea
OflTE 4/l3.'8S
4/23/85
3'35
Figure 33. Control chart showing response factors by date for 1,2,3,4,6,7,8-HpCDD.
Dotted lines represent approximately ± 10% of the mean.
74
-------
I!P:OCTftCHLORO-FURftH
F:Cl3-OCTflCaORO-OIOXIH
(ftREfl/REF.flREfl)/
-------
Table 23. Summary of Results from the Analysis
of a Laboratory Method Blank
Compound
2,3,7,8-TCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3, 4,6,7, 8-HpCDF
1,2,3,4,7,8,9-HpCDF
1,2,3,4,6,7,8-HpCDD
OCDF
OCDD
Concentration
(pg/g)
ND (0.50)
ND (2.2)
ND (0.5)
ND (0.5)
ND (1.2)
ND (0.5)
ND (0.5)
ND (0.5)
ND (0.5)
ND (1.0)
ND (0.9)
ND (0.8)
ND (0.5)
ND (0.5)
4.0
ND (0.5)
30
aAt least three other PCDD compounds were de-
tected but not quantitated in the laboratory
method blank. These included 1,3,6,8- and
1,3,7,9-TCDD and an unidentified PeCDD
, isomer.
Concentration based on assumption of 10.0 g
equivalent lipid sample. The background
concentration of 1,2,3,4,6,7,8-HpCDD and
OCDD were not subtracted from the measured
concentration for the spiked and unspiked
lipid matrix.
76
-------
Further analysis of individual reagents used for preparation of
the samples identified the activated acidic alumina as the source of the arti-
facts. Acidic alumina that had been cleaned by Soxhlet extraction but not
activated at 190°C was analyzed, and the artifacts were not detected. This
indicates that the artifacts are generated during activation of acidic alumina
at elevated temperatures (190°C). Similar background problems from the same
PCDD congeners have recently been reported by the Center for Disease
Control.17'18
An experiment was designed to evaluate a procedure for cleaning the
activated acidic alumina immediately prior to the fractionation of the sample
extract. The acidic alumina (6.0 g) was packed in hexane. The packed column
was eluted with 40 ml of methylene chloride/hexane (1:1) solution followed by
80 to 100 mL of hexane. The sample extract was added to the column and was
eluted with 20 ml of hexane followed by 30 mL of 20% methylene chloride in
hexane which was reserved for PCDD and PCDF analysis.
The carbon-14 radiolabeled 2,3,7,8-TCDD, 1,2,3,4,7,8-HxCDD, and OCDD
were used to evaluate recovery of PCDDs from the cleaned alumina. Recoveries
of the radiolabeled PCDDs from the activated acidic alumina precleaned by the
procedure described above are detailed in Table 24. These data demonstrate
that the selected PCDDs are quantitatively (greater than 90%) recovered from
the precleaned acidic alumina. This procedure for cleanup of activated acidic
alumina was not initiated for the analysis of the lipid samples described in
this report. However, it has been integrated into the analytical protocol
(Appendix A) for routine application with sample preparation activities.
D. Absolute Recoveries of the Internal Quantitation Standards
The absolute recoveries of the carbon-13 labeled internal quantita-
tion standards were determined by comparing responses with the internal recov-
ery standard, 13C12~1,2,3,4-TCDD, which was added during final concentration
prior to HRGC/MS analysis. A summary of the average and range of recoveries
of the 8 internal quantisation standards from the 14 human adipose lipid
samples is provided in Table 25.
These data indicate that recoveries ranged from an average of 52.1%
for 13C12-2,3,7,8-TCDD up to 88.9% for 13C12-OCDD. The average recoveries
for the lower chlorinated internal standards were lower than the preliminary
method studies with carbon-14 radiolabeled standards had indicated. This re-
sulted in a closer evaluation of the final concentration step prior to mass
spectrometry. The first extracts for the human adipose lipid extracts were
concentrated with a nitrogen evaporation system equipped with a water bath at
approximately 55°C. Final blowdown of the samples required addition of the
internal recovery standard in 10 uL of tridecane as a keeper solution. How-
ever, it was noted at the elevated temperature final volumes from nitrogen
evaporation were generally on the order of 2 to 5 uL. This required addition
of another 10 uL of tridecane prior to HRGC/MS analysis.
In an effort to assess the effect of reducing the final volume of
tridecane at elevated temperatures on absolute recoveries of the internal
quantisation standards, an experiment using the radiolabeled TCDD, HxCDD, and
OCDD standards was conducted.
77
-------
Table 24. Recovery of Radiolabeled PCDDs from
Precleaned Activated Alumina
Spike
level Recovery
Compound (pg) (%)
14C-2,3,7,8-TCDD 100 92
300 96
300 97
14C-l,2,3,4,7,8-HxCDD 1,000 103
3,000 101
3,000 100
14C-OCDD 2,500 102
7,500 99
7,500 97
78
-------
Table 25. Absolute Recoveries of the Internal Quantitation Standards
from the Human Adipose Lipid Matrix
Internal
quantisation Average
standard recovery (%)
13C12-2,3,7,8-TCDF
13C12-2,3,7,8-TCDD
13C12-l,2,3,7,8-PeCDF
13C12-l,2,3,7,8-PeCDD
13C12-1,2,3,4,7,8-
HxCDF
13C12-1,2,3,6,7,8-
HxCDD
13C12-1,2, 3, 4,6,7,8-
HpCDD
13C12-OCDD
64.0
52.1
76.1
57.1
59.4
63.6
76.3
88.9
Standard
deviation
7.9
5.0
8.9
3.5
5.0
5.3
10.3
11.4
Relative
standard
deviation (%)
12.3
9.6
11.7
6.1
8.4
8.4
13.6
12.9
Range of
recovery (%)
48-78
43-62
62-90
51-64
52-70
57-77
61-99
67-104
aValues based on 14 analyses of human adipose lipid samples.
79
-------
Four solutions of the same spike level were prepared with each
radiolabeled compound in 1 ml of toluene. Two of the spiked solutions were
heated at 55-60°C and the solvent was reduced under a gentle stream of pre-
purified nitrogen. The toluene solution was concentrated to 100 pL, 500 uL
of 1% toluene in methylene chloride was added, and the solution was concen-
trated to 200 uL. At this time 10 uL of the keeper tridecane was added and
the solution was allowed to concentrate further. The remaining two solutions
for each radiolabeled compound were taken through a similar solvent exchange
and concentration procedure except the solution was allowed to concentrate at
room temperature.
One of the most obvious results was the observation that solutions
held at elevated temperatures could be reduced to dryness even when tridecane
had been added as a keeper. On the other hand, solutions for which tridecane
had been added but remained at room temperature could only be concentrated to
a 10-pL final volume. The recoveries of the radiolabeled standards from each
of the solutions in this study are presented in Table 26.
The results from this study indicate that the final concentration
condition may have a pronounced effect on the absolute recoveries of the PCDDs
and PCDFs, especially for the lower chlorinated congeners such as 2,3,7,8-TCDD.
These conclusions are supported by an independent study in comparison of con-
centration techniques for 2,3,7,8-TCDD.19 However, it should be noted that
the approach to target analyte quantisation based on the internal standard
method (isotope dilution for 8 of the 17 target analytes) is not affected by
absolute recoveries as low as 50%. The procedure for final concentration in
the analytical protocol (Appendix A) for the analysis of the NHATS samples
for the EPA/VA study has been modified to specify room temperature conditions.
80
-------
IULJIC C-\J . t\CV-\J V t 1 Jf
and OCDD as
Compound
14C-2,3,7,8-TCDD
14C-l,2,3,4,7,8-HxCDD
14C-OCDD
\s i \> u i *• w i i j
a Function
Spi ke
level
(pg)
300
300
300
300
3,000
3,000
3,000
3,000
7,500
7,500
7,500
7,500
U 1 ^\*t-r \~ t *~\A t^ y <~S y I ) ^^ IV
of Final Concentrati
Concentration
conditions
55-60°C
55-60°C
20°C
20°C
55-60°C
55-60°C
20°C
20°C
55-60°C
55-60°C
20°C
20°C
"" , — , >_,»,.,.
on Conditions
Observed
final
volume
1-2 pL
dryness
10 pL
10 pL
1-2 pL
5 pL
10 pL
10 pL
1-2 pL
2-3 pL
10 pL
10 pL
Observed
recovery (%)
78
54
98
93
94
102
105
107
94
94
100
97
'Each solution was concentrated under a gentle stream of flowing nitrogen.
-------
VII. GLOSSARY OF TERMS
Accuracy - A measurement of the bias of a system, which for this study, is
based on the agreement of the 2,3,7,8 substituted PCDD and PCDF to an ac-
cepted reference standard.
Batch, sample - A sample batch consists of up to 10 human adipose tissue sam-
ples, one method blank, 2 internal quality control (QC) samples (spiked and
unspiked), and an external performance audit sample (blind spike).
Blank, laboratory method - This blank is prepared in the laboratory through
performing all analytical procedures except addition of a sample aliquot to
the extraction vessel. A minimum of one laboratory method blank will be
analyzed with each batch of samples.
Calibration standards (concentration calibration solutions) - Solutions con-
taining known amounts of the native analytes (unlabeled 2,3,7,8-substituted
PCDDs and PCDFs), the internal quantisation standards (carbon-13 labeled
PCDDs and PCDFs), and the recovery standard, 13C12-1,2,3,4-TCDD. These cal-
ibration solutions are used to determine instrument response of the analytes
relative to the internal quantisation standards and of the internal quanti-
tation standards relative to the internal recovery standard.
Lipid - The organic solvent extractable constituents of adipose tissue con-
sisting of fatty oils, proteins, and carbohydrates. The concentrations of
PCDDs and PCDFs are reported on the lipid content bases.
Instrumental mass calibration - An internal instrumental systems check and
tuning standard, perfluorokerosene (PFK), is introduced automatically by the
instrument. The mass ion 380.976 is monitored by the analyst as an instru-
mental systems check and is also used to tune the instrument.
Internal quantitation standards - Carbon-13 labeled PCDDs and PCDFs, which
are added to every sample and are present at the same concentration in every
method blank and quality control sample. These are added to the adipose
tissue prior to extraction and are used to measure the concentration of each
analyte. The concentration of each internal quantitation standard is measured
in every sample, and percent recovery is determined using the internal recovery
standard.
Internal recovery standard - 13C12-1,2,3,4-TCDD and
which is added to every sample extract just before the final concentration
step and HRGC/MS-SIM analysis.
Limit of detection (LOD) - A value, derived from the noise to signal re-
sponse, which is equal to 2.5 times the average instrumental noise level
is the limit of detection.
Limit of quantitation (LOQ) - A value, derived from the noise to signal re-
sponse, which is equal to 10 times the average instrumental noise level is
the limit of quantitation.
82
-------
Mass resolution check - Standard method used to demonstrate static resolution
of 10,000 minimum (10% valley definition).
Not detected (ND) - A nonresponse or a response which is less than the limit
of detection is reported as not detected.
Precision - The results from analysis of replicate samples (spiked and un-
spiked) provide the measure of method precision. The precision of the method
is reported as standard deviation or relative standard deviation.
Performance check mixture, HRGC column - A mixture containing known amounts
of selected TCDD standards; it is used to demonstrate continued acceptable
performance of the capillary column, to separate (g 25 % valley on a 50-m CP
Sil 88 or 60-m SP-2330 HRGC column and 30 to 60% for a 60 m DB-5 HRGC column)
2,3,7,8-TCDD isomer from all other 21 TCDD isomers, and to define the TCDD
retention time window.
PCDD - Polychlorinated dibenzo-p_-dioxins.
PCDF - Polychlorinated dibenzofurans.
Relative response factor - Response of the mass spectrometer to a known
amount of an analyte relative to a known amount of an internal standard
(quantitation or recovery).
Trace (TR) - A response which is greater than the limit of detection but less
than the limit of quantitation is reported as a trace value. An estimated
method detection limit is provided for trace value.
83
-------
VIII. REFERENCES
1. Stanley JS. 1984. Methods of analysis of polychlorinated dibenzo-p_-
dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs) in biological
matrices--!iterature review and recommendations. EPA-560/584-00.
2. Stanley JS, Going JE, Redford DP, Kutz KW, Young AL. 1985. A survey of
analytical methods for measurement of polychlorinated dibenzo-p_-dioxins
(PCDD) and polychlorinated dibenzofurans (PCDF) in human adipose tissues.
In: Chlorinated dioxins and dibenzofurans in the total environment II.
Keith LH, Rappe C, Choudhary G, eds. Butterworth Publishers, pp. 181-195.
3. Stanley JS. 1984 (March 28). Proposed analytical method for analysis
of PCDDs/PCDFs in human adipose tissue: special report. Washington, DC:
Office of Pesticides and Toxic Substances, U.S. Environmental Protection
Agency. Contract 68-02-3938, Work Assignment 8.
4. Stanley JS. 1986 (April 23). Broad scan analysis of human adipose tissue:
polychlorinated dibenzo-p_-dioxins (PCDDs) and polychlorinated dibenzofurans
(PCDFs). Draft final report. Washington, DC: Office of Pesticides and
Toxic Substances, U.S. Environmental Protection Agency. Contract 68-02-
3938, Work Assignment 8.
5. Albro et al. 1985. Methods for the quantitative determination of multi-
ple, specific polychlorinated dibenzo-p_-dioxin and dibenzofuran isomers
in human adipose tissue in the parts-per-trillion range. An interlabora-
tory study. Anal Chem 57: 2717-2725.
6. Patterson DG et al. 1986. High-resolution gas chromatographic/high-
resolution mass spectrometric analysis of human adipose tissue for
2,3,7,8-tetrachlorodibenzo-p_-dioxin. Anal Chem 58: 705-713.
7. Schecter A, Tiernan TO, Taylor ML, VanNess GF, Barrett JH, Wagel DJ,
Gitlitz G, Bogdasarian M. 1985. Biological markers after exposure to
polychlorinated dibenzo-p_-dioxins, dibenzofurans, biphenyls, and bi-
phenylenes. Part I: Findings using fat biopsies to estimate exposure.
In: Chlorinated dioxins and dibenzofurans in the total environment II.
Keith LH, Rappe C, Choudhary G, eds. Butterworth Publishers, pp. 215-246.
8. Schecter A, Ryan JJ. 1985. Dioxin and furan levels in human adipose
tissue from exposed and control populations. 189th National ACS Meeting
Symposium on Chlorinated Dioxins and Dibenzofurans in the Total Environ-
ment III, Miami, Florida. Preprint Division of Environmental Chemistry,
ACS 25:160-163, Paper No. 56.
9. Ryan JJ, Williams DT, Lau BPY, Sakuma T. 1985. Analysis of human fat
tissue for 2,3,7,8-tetrachlorodibenzo-p_-dioxin and chlorinated dibenzo-
furan residues. In: Chlorinated dioxins and dibenzofurans in the total
environment II. Keith LH, Rappe C, Choudhary G, eds. Butterworth Pub-
lishers, pp. 205-214.
84
-------
10. Ryan JJ, Schecter A, Lizotte R, Sun W-F, Miller L. 1985. Tissue dis-
tribution of dioxins and furans in humans from the general population.
Chemosphere 14: 929-932.
11. Nygren M, Hansson M, Rappe C, Domellof L, Hardell L. 1985. Analysis of
polychlorinated dibenzo-p_-dioxins and dibenzofurans in adipose tissue
from soft-tissue sarcoma patients and controls. 189th National ACS
Meeting Symposium on Chlorinated Dioxins and Dibenzofurans in the Total
Environment III, Miami, Florida, 1985. Preprint Division of Environ-
mental Chemistry, ACS 25:160-163, Paper No. 55.
12. Smith LM, Stalling DL, Johnson JJ. 1984. Determination of part per
trillion levels of polychlorinated dibenzofurans and dioxins in environ-
mental samples. Anal Chem 58: 1830-1842.
13. Rappe C, Nygren M, Linstrom G, Hanson H. 1985. Dioxins and dibenzo-
furans in human tissues and milk of European origin. 5th International
Symposium on Chlorinated Dioxins and Related Compounds, Bayreuth, FRG,
September 16-19, 1985.
14. Ryan JJ. 1985. Variation of dioxins and furans in humans with age and
organ by country. 5th International Symposium on Chlorinated Dioxins
and Related Compounds, Bayreuth, FRG, September 16-19, 1985.
15. Graham M, Hileman FD, Wendling J, Wilson JD. 1985. Chlorocarbons in
adipose tissue samples. 5th International Symposium on Chlorinated
Dioxins and Related Compounds, Bayreuth, FRG, September 16-19, 1985.
16. Patterson DG, Holler JS, Smith SJ, Liddle JA, Sampson EJ, Needham LL.
1985. Human tissue data in certain U.S. populations. 5th International
Symposium on Chlorinated Dioxins and Related Compounds, Bayreuth, FRG,
September 16-19, 1985.
17. Patterson DG, Holler JS, Groce DF, Alexander LR, Lapeza CR, O'Conner RC,
Liddle JA. 1986. Control of interferences in the analysis of human
adipose tissue to 2,3,7,8-tetrachlorodibenzo-p_-dioxin (TCDD). Environ
Toxicol Chem 5: 355-360.
18. Holler JS, Patterson DG, Alexander LR, Groce DF, O'Connor RC, Lapeza CR.
1985. Control of artifacts and contamination in the development of a
dioxin analytical program. 33rd Annual Conference on Mass Spectrometry
and Allied Topics, San Diego, CA, May 26-31, 1985.
19. O'Keefe PN, Meyer C, Dillon K. 1982. Comparison of concentration tech-
niques for 2,3,7,8-tetrachlorodibenzo-p_-dioxin. Anal Chem 54: 2623-
2625.
85
-------
APPENDIX A
ANALYTICAL PROTOCOL FOR DETERMINATION OF PCDDs AND PCDFs
IN HUMAN ADIPOSE TISSUE
A-l
-------
TABLE OF CONTENTS
Section Description Page
1 Scope and Application A-3
2 Summary of Method A-3
3 Definitions A-6
4 Interferences A-7
5 Safety A-7
6 Apparatus and Equipment A-8
7 Reagents and Standard Solutions A-ll
8 High Resolution Gas Chromatography/Mass Spectrometry
Performance Criteria A-13
9 Quality Control Procedures A-31
10 Sample Preservation and Handling A-33
11 Sample Extraction A-33
12 Cleanup Procedures A-35
13 Analytical Procedures A-38
14 Date Reduction A-43
15 Reporting and Documentation A-50
A-2
-------
ANALYTICAL PROTOCOL FOR DETERMINATION OF PCDDs AND PCDFs
IN HUMAN ADIPOSE TISSUE
1. SCOPE AND APPLICATION
I.I This method provides procedures for the detection and quantitative
measurement of polychlorinated dibenzo-p_-dioxins (PCDD) and poly-
chlorinated dibenzofurans (PCDF) at concentrations ranging from 1
to 100 pg/g for the tetrachloro congeners up to 5 to 500 pg/g for
the octachloro congeners in 10-g aliquots of human adipose tissue.
1.2 The minimum measurable concentration is estimated to range from
1 pg/g (1 part per trillion) for 2,3,7,8-TCDD and 2,3,7,8-TCDF up
to 5 pg/g for OCDD and OCDF. However, these detection limits
depend on the kinds and concentrations of interfering compounds
in the sample matrix and the absolute method recovery.
1.3 The method will be used to determine PCDDs and PCDFs, particularly
congeners with chlorine substitution in the 2,3,7,8 positions.
Table 1 lists the specific PCDDs and PCDFs and target method
detection limits.
2. SUMMARY OF METHOD
Figure 1 presents a schematic of the analytical procedures for deter-
mination of PCDDs and PCDFs in human adipose tissue. The analytical
method requires extraction and isolation of lipid materials from human
adipose samples. This is accomplished using sample sizes ranging up to
10 g. The tissue is spiked with known amounts of the carbon-13 labeled
PCDDs and PCDFs (e.g., 500 pg of 13C12-TCDD/F to 2,500 pg of 13C12-OCDD/F)
as internal quantisation standards. Extraction and homogenization are
accomplished using methylene chloride and a Tekmar Tissuemizer®. The
extract is filtered through anhydrous sodium sulfate to remove water.
The extraction procedure is repeated (three to five times) until the
tissue sample has been thoroughly homogenized. The final extract is
adjusted to a known volume (100 ml) and the extractable lipid is
determined using a minimum of 1% of the final volume. The methylene
chloride in the remaining extract is concentrated until only an oily
residue remains. The residue is diluted with hexane (^ 200 ml), and
100 g of sulfuric acid modified silica gel (40% w/w) is added to the
solution with stirring. The mixture is stirred for approximately 2 h,
and the supernatant is decanted and filtered through anhydrous sodium
sulfate. The adsorbent is washed with at least two additional aliquots
of hexane.
The combined hexane extracts are eluted through a column consisting of a
layer of sulfuric acid modified silica gel, and a layer of unmodified
silica gel. The eluate is concentrated to approximately 1 ml and added
to a column of acidic alumina. The PCDDs and PCDFs are eluted from the
alumina using 20% methylene chloride/hexane. This eluate is concentrated
to approximately 0.5 mL and is added to a 500-mg Carbopak C/Celite column.
The PCDDs and PCDFs are eluted from the column using 20 mL of toluene.
A-3
-------
Table 1. Target PCDD and PCDF Congeners and Target Method
Detection Limits
Compound
2,3,7,8-TCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDD
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,4,7,8-HxCDD
1,2,3,6, 7, 8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,7,8-HxCDF
1,2,3,6, 7, 8-HxCDF
1,2,3,7,8,9-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,4,6,7,8-HpCDD
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDD
OCDF
CAS no.a
1746-01-6
51207-31-9
40321-76-4
57117-41-6
57117-31-4
39227-28-6
57653-85-7
19408-74-3
70648-29-9
57117-44-9
72918-21-9
60851-34-5
35822-46-9
67562-39-4
55673-89-7
3268-87-9
39001-02-0
Target method detection
limit (pg/g)
1.0
1.0
1.0
1.0
1.0
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
5.0
5.0
Chemical Abstract Services number.
""pg/g = parts per trillion.
A-4
-------
Initial Sample Preparation
Isolation of Extractable Lipid Materials
Add Internal Quantitation Standards
(13c-PCDDs/PCDFs)
Homogenization in Methylene Chloride
i
Lipid Determination
Solvent Exchange
Bulk Lipid Removal
Acid Modified Silica Gel
Slurry Technique
I
Provides Cleanup of Oxidizable Compounds
with Rapid Sample Turnaround, Improved
Cleanup Efficiency and Recovery
Removal of Chemical Interferences
Acidic Silica/Silica
Acidic Alumina
Provides Seperation of PCBs and Other
Potential Interferences from PCDDs and PCDf
Carbopak C/Celite
Selective Adsorption and Isolation of PCDDs/PCDFs
Add Internal Recovery Standards
HRGC/MS-SIM Analysis
i
LRMS
Identification/Quantitation
of Tetra-Ocfa PCDDs/PCDFs
1
HRMS
Confirmation of 2,3,7,8-TCDD
Figure 1. Schematic of the sample preparation and
instrumental analysis procedures for determination
of PCDDs and PCDFs in human adipose tissue.
A-5
-------
The toluene is concentrated to less than 1 ml and transferred to conical
vials. Tridecane (10 uL) containing 500 pg of an internal recovery stan-
dard is added as a keeper, and the extract is concentrated to final vol-
ume.
The HRGC/MS analysis is completed in the selected ion monitoring mode
(SIM). Analysis of the tetra- through octachloro PCDD and PCDF congeners
is achieved using low resolution mass spectrometry. Separation of the
tetra- through octachloro PCDD and PCDF congeners is achieved using a
60-m DB-5 column. Verification of the 2,3,7,8-TCDD is achieved using
either a 50-m CP Sil 88 column or 60-m SP-2330 column and HRGC/MS-SIM
analysis in the high resolution mode (R = 10,000).
3. DEFINITIONS
3.1 Concentration calibration solutions -- Solutions containing known
amounts of the native analytes (unlabeled 2,3,7,8-substituted
PCDDs and PCDFs), the internal quantisation standards (Carbon-13
labeled PCDDs and PCDFs), and the recovery standard, 13C12-
1,2,3,4-TCDD. These calibration solutions are used to determine
instrument response of the analytes relative to the internal
quantisation standards and of the internal quantisation standards
relative to the internal recovery standard.
3.2 Internal quantisation standards -- Carbon-13 labeled PCDDs and
PCDFs, which are added to every sample and are present at the
same concentration in every method blank and quality control
sample. These are added to the adipose tissue and are used to
measure the concentration of each analyte. The concentration
of each internal quantisation standard is measured in every
sample, and percent recovery is determined using the internal
recovery standard.
3.3 Internal recovery standard -- 13C12-1,2,3,4-TCDD and 13C12-
1,2,3,7,8,9-HxCDD which is added to every sample extract just
before the final concentration step and HRGC/MS-SIM analysis.
3.4 Laboratory method blank -- This blank is prepared in the labora-
tory through performing all analytical procedures except addition
of a sample aliquot to the extraction vessel. A minimum of one
laboratory method blank will be analyzed with each batch of sam-
ples.
3.5 HRGC column performance check mixture -- A mixture containing
known amounts of selected TCDD standards; it is used to demon-
strate continued acceptable performance of the capillary column,
to separate (S 25% valley on a 50-m CP Sil 88 or 60-m SP-2330
HRGC column and 30 to 60% for a 60-m DB-5 HRGC column) 2,3,7,8-
TCDD isomer from all other 21 TCDD isomers, and to define the
TCDD retention time window.
3.6 Relative response factor -- Response of the mass spectrometer to
a known amount of an analyte relative to a known amount of an
internal standard (quantisation or recovery).
A-6
-------
3.7 Mass resolution check -- Standard method used to demonstrate
static resolution of 10,000 minimum (10% valley definition).
3.8 Sample batch -- A sample batch consists of up to 10 human adipose
tissue samples, one method blank, 2 internal quality control (QC)
samples (spiked and unspiked), and an external performance audit
sample (blind spike).
4. INTERFERENCES
Chemicals which elute from the HRGC column with ± 10 scans of the inter-
nal and/or recovery standards and which produce within the retention time
window ions at any of the masses used to detect or quantify PCDDs, PCDFs,
or the internal quantitation and recovery standards are potential inter-
ferences. Most frequently encountered potential interferences are other
sample components that are extracted along with the PCDDs and PCDFs, e.g.,
PCBs, chlorinated methoxybiphenyls, chlorinated hydroxydiphenyl ethers,
chlorinated benzylphenyl ethers, chlorinated naphthalenes, DDE, DDT, etc.
The actual incidence of interference by these chemicals depends also
upon relative concentrations, mass spectrometric resolution, and chro-
matographic conditions. Because very low levels (pg/g) of PCDDs and
PCDFs are anticipated, the elimination of interferences is essential.
High purity reagents and solvents must be used and all equipment must be
scrupulously cleaned. Laboratory method blanks must be analyzed to demon-
strate absence of contamination that would interfere with measurement of
the PCDDs and PCDFs. Column chromatographic procedures are used to remove
coextracted sample components; these procedures must be performed care-
fully to minimize loss of PCDDs and PCDFs during attempts to increase
their concentration relative to other sample components.
5. SAFETY
5.1 The toxicity or carcinogenicity of each reagent used in this
method has not been precisely defined; however, each chemical
compound should be treated as a potential health hazard. The
2,3,7,8-TCDD is a known teratogen, mutagen, and carcinogen. In-
gestion of microgram quantities can result in toxic effects. The
other 2,3,7,8-substituted PCDDs and PCDFs may exhibit teratogenic,
mutagenic, and carcinogenic effects. From this viewpoint, expo-
sure to these chemicals must be reduced to the lowest possible
level by whatever means available. Only experienced personnel
will be allowed to work with these chemicals.
5.2 All laboratory personnel will be required to wear laboratory
coats or coveralls, gloves, and safety glasses. The neat stan-
dards, stock, and working solutions will be handled only in a
Class A fume hood or glove box. When manipulating stock stan-
dards or working solutions, the analyst is advised to place the
solution vials in a secure holder (sample block or glass beaker)
to prevent accidental spills.
A-7
-------
5.3 If these standards are spilled, absorb as much as possible with
absorbent paper and place in a container clearly labeled as PCDD
or PCDF waste. Solvent-wash all contaminated surfaces with tolu-
ene and absorbent paper followed by washing with a strong soap
and water solution. Dispose of all contaminated materials in
sealed steel containers labeled as contaminated with PCDD and/or
PCDF residue and indicate the approximate level of contamination.
As a final precaution, prepare a wipe sample of the exposed sur-
face area and include the wipe as part of the sample analysis
batch. This will be used to confirm that the work area is free
of contamination.
5.4 If handling of these compounds results in skin contact, immedi-
ately remove all contaminated clothing and wash the affected skin
areas with soap and water for at least 15 min.
5.5 Disposal of laboratory wastes -- All laboratory wastes (solvents
and absorbents) will be disposed of as hazardous wastes. The
laboratory personnel should take care to dispose of the sodium
sulfate, silica gel, and alumina in separate containers. Excess
solvents should be disposed of in gallon polyethylene jugs con-
taining a layer of activated charcoal. Excess solvent that is
known to be contaminated with PCDDs or PCDFs should be kept at a
minimum by evaporating the solvent with a stream of air.
6. APPARATUS AND EQUIPMENT
6.1 High Resolution Gas Chromatograph/Mass Spectrometer/Data System
(HRGC/HRMS/DS)
6.1.1 The GC must be equipped for temperature programming,
and all required accessories must be available, such as
syringes, gases, and a capillary column. The GC injec-
tion port must be designed for capillary columns. The
use of splitless injection techniques is recommended.
When using this method, a l-(jL injection volume is used.
The injection volumes for all extracts, blanks, calibra-
tion solutions, and the performance check sample must
be consistent.
6.1.2 High Resolution Gas Chromatograph-Mass Spectrometer
Interface
The HRGC/MS interface is directly coupled to the mass
spectrometer ion source. All components of the inter-
face should be glass or glass-lined stainless steel.
The interface components should be compatible with
300°C temperatures. The HRGC/MS interface must be
appropriately designed so that the separation of the
PCDDs and PCDFs which is achieved in the gas chromato-
graphic column is not appreciably degraded. Cold spots
and/or active surfaces (adsorption sites) in the HRGC/MS
A-8
-------
interface can cause peak tailing and peak broadening.
It is recommended that the HRGC column be fitted directly
into the MS ion source. Graphite ferrules should be
avoided in the HRGC injection port since they may ab-
sorb PCDDs or PCDFs. Vespel or equivalent ferrules
are recommended.
6.1.3 Mass Spectrometer
The mass spectrometer must be capable of maintaining a
minimum resolution of 10,000 (10% valley) for high reso-
lution confirmation analysis. The mass spectrometer
must be operated in a selected ion monitoring (SIM)
mode with total cycle time (including voltage reset
time) of 1 s or less.
6.1.4 Data System
A dedicated hardware or data system is required to con-
trol the rapid multiple ion monitoring process and to
acquire the data. Quantification data (peak areas or
peak heights) and SIM traces (displays of intensities
of each m/z (characteristic ion) being monitored as a
function of time) must be acquired during the analyses.
Quantifications may be reported based upon computer-
generated peak areas or upon measured peak heights.
6.2 HRGC Columns
For isomer-specific determinations of 2,3,7,8-TCDD, the following
fused silica capillary columns are recommended: a 50-m CP-Sil 88
column and a 60-m SP-2330 (SP-2331) column. However, any capil-
lary column which separates 2,3,7,8-TCDD from all other TCDDs may
be used for such analyses, provided that the minimum acceptance
criteria in Section 8 are met.
6.3 Miscellaneous Equipment
6.3.1 Nitrogen evaporation apparatus with variable flow rate.
6.3.2 Balance capable of accurately weighing to ± 0.01 g.
6.3.3 Balance capable of accurately weighint to ± 0.0001 g.
6.3.4 Water bath -- equipped with concentric ring cover and
capable of being temperature-controlled.
6.3.5 Stainless steel spatulas or spoons.
6.3.6 Magnetic stirrers and stir bars.
6.3.7 High speed tissue homogenizer -- Tekmar Tissuemizer®
equipped with an EN-8 probe or equivalent.
6.3.8 Vacuum dessicator.
A-9
-------
6.4 Glassware
6.4.1 Erlenmeyer flask -- 500 ml_.
6.4.2 Kuderna-Danish apparatus -- 500-mL evaporating flask,
15-mL graduated concentrator tubes with ground-glass
stoppers, and three-ball macro Snyder column (Kontes
K-570001-0500, K-503000-0121, and K-569001-0219 or
equivalent).
6.4.3 Minivials -- 1-mL borosilicate glass with conical-shaped
reservoir and screw caps lined with Teflon®-faced sili-
cone disks.
6.4.4 Powder funnels -- glass.
6.4.5 Chromatographic columns for the silica and alumina
chromatography -- 1 cm ID x 10 cm long and 1 cm ID x
30 cm long with 250-mL reservoir and equipped with TFE
stopcocks.
6.4.6 Chromatographic column for the Carbopak cleanup --
disposable 5-mL graduated glass pipets, 6 to 7 mm ID.
6.4.7 Glass rods.
6.4.8 Carborundum boiling chips -- Extracted for 6 hr in a
Soxhlet apparatus with benzene and air dried.
6.4.9 Glass wool, silanized (Supelco) -- Extract with methylene
chloride and hexane and air dry before use.
6.4.10 Glassware cleaning procedure -- All glassware used for
these analyses will be cleaned via the following proce-
dure. Wash the glassware in soap and water, rinse with
copious amounts of tap water, distilled water, and
disti1led-in-glass acetone, in that order. Immediately
prior to use, the glassware should be rinsed with
distilied-in-glass quality solvents: methylene chloride,
toluene, and hexane. The glassware should be allowed
to dry fully.
As an added precuation, all glassware will be marked
with a unique code that should be noted in the extrac-
tion and cleanup procedures for each sample. This
glassware tracking will allow background results from
specific glassware to be documented.
After use, each piece of glassware should be rinsed
with the last solvent used in it, followed by a rinse
with toluene, then acetone, before transferring it to
the glassware washing facility.
A-10
-------
7. REAGENTS AND STANDARD SOLUTIONS
7.1 Column Chromatography Reagents
7.1.1 Alumina, acidic (Biorad, AG-4) -- Extract the alumina
in a Soxhlet apparatus with methylene chloride for 18 h
(minimum of two cycles per hour). Air dry and activate
it by heating in a foil-covered glass container for 24 h
at 190°C.
7.1.2 Silica gel — High purity grade, type 60, 70-230 mesh;
extract the silica gel in a Soxhlet apparatus with
methylene chloride for 10 h (minimum of 2 cycles per
hour). Air dry and activate it by heating in a foil-
covered glass container for 24 h at 130°C.
7.1.3 Silica gel impregnated with 40% (by weight) sulfuric
acid -- Add two parts (by weight) concentrated sulfuric
acid to three parts (by weight) silica gel (extracted
and activated) (e.g., 40 g of H2S04 plus 60 g of silica
gel) in a glass screw-cap bottle. Tumble for 5 to 6 h,
shaking occasionally until free of lumps.
7.1.4 Sulfuric acid, concentrated -- ACS grade, specific
gravity 1.84.
7.1.5 Graphitized carbon black (Carbopack C, Supelco), sur-
face of approximately 12 m2/g, 80/100 mesh -- Mix thor-
oughly 3.6 g of Carbopack C and 16.4 g of Celite 545®
in a 40-mL vial. Activate at 130°C for 6 h. Store in
a desiccator.
7.1.6 Celite 545® (Fischer Scientific), reagent grade, or
equivalent.
7.2 Desiccating agents -- Sodium sulfate; granular, anhydrous. Before
use extract with methylene chloride for 16 h (minimum of two cy-
cles per hour), air dry and then muffle for ^ 4 h in a shallow
tray at 400°C. Let it cool in a desiccator and store in oven at
130°C.
7.3 Solvents -- High purity, distilled in glass: methylene chloride,
toluene, benzene, cyclohexane, methanol, acetone, hexane; reagent
grade: tridecane. High purity solvents are dispensed from Teflon®
squirt bottles.
7.4 Concentration Calibration Solutions (Table 2)
Eight tridecane solutions containing native calibration standards,
13C12-labeled internal quantitation standards, and two internal
recovery standards are required. The complete compound list is
A-11
-------
Table 2. Concentration Calibration Solutions
Compound
Native
2,3,7,8-TCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDD
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,4,6,7,8-HpCDD
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDD
OCDF
Internal Quantitation
Standards
13C12-2,3,7,8-TCDD
13C12-2,3,7,8-TCDF
13C12-l,2,3,7,8-PeCDD
13C12-l,2,3,7,8-PeCDF
13C12-l,2,3,6,7,8-HxCDD
13C12-l,2,3,4,7,8-HxCDF
13C12-l,2,3,4,6,7,8-HpCDD
13C12-l,2,3,4,6,7,8-HpCDF
13C12-OCDD
Internal Recovery
Standard
13C12-1,2,3,4-TCDD
l3C1?-l,2,3,7,8,9-HxCDD
Concentration in cali
CS1
200
200
200
200
200
500
500
500
500
500
500
500
500
500
500
1,000
1,000
50
50
50
50
125
125
125
125
250
50
125
CS2
100
100
100
100
100
250
250
250
250
250
250
250
250
250
250
500
500
50
50
50
50
125
125
125
125
250
50
125
CS3
50
50
50
50
50
125
125
125
125
125
125
125
125
125
125
250
250
50
50
50
50
125
125
125
125
250
50
125
CS4
25
25
25
25
25
62.
62.
62.
62.
62.
62.
62.
62.
62.
62.
125
125
50
50
50
50
125
125
125
125
250
50
125
bration solutions in
5
5
5
5
5
5
5
5
5
5
CSS
10
10
10
10
10
25
25
25
25
25
25
25
25
25
25
50
50
50
50
50
50
125
125
125
125
250
50
125
CS6
5
5
5
5
5
12.
12.
12.
12.
12.
12.
12.
12.
12.
12.
25
25
50
50
50
50
125
125
125
125
250
50
125
5
5
5
5
5
5
5
5
5
5
CS7
2.
2.
2.
2.
2.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
12.
12.
50
50
50
50
125
125
125
125
250
50
125
PQ/ML
5
5
5
5
5
25
25
25
25
25
25
25
25
25
25
5
5
CSS
1
1
1
1
1
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
5
5
50
50
50
50
125
125
125
125
250
50
125
\-12
-------
7.5
given in Table 2. The native 2,3,7,8-TCDD is supplied as a cer-
tified standard solution from the U.S. EPA QA Reference Materials
Branch. All other native compounds were supplied in crystalline
form by Cambridge Isotope Laboratories (Woburn, MA). 13C12~
Labeled internal quantisation standards were supplied in solution
in ji-nonane by Cambridge Isotope Laboratories. Portions of the
native standards were accurately weighed to the nearest 0.001 mg
with a Cahn 27 electrobalance and dissolved in toluene.
Column Performance Check Mixture
The column performance check mixture consists of several TCDD
isomers which will be used to document the separation of 2,3,7,8-
TCDD from all other isomers. This solution will contain TCDDs
(A) eluting closely to 2,3,7,8-TCDD, and the first- (F) and last-
eluting (L) TCDDs.
Analyte
Unlabeled 2,3,7,8-TCDD
13C12-2,3,7,8-TCDD
1,2,3,4-TCDD (A)
1,4,7,8-TCDD (A)
1,2,3,7-TCDD (A)
1,2,3,8-TCDD (A)
1,3,6,8-TCDD (F)
1,2,8,9-TCDD (L)
Approximate amount per ampule
10 ng
10 ng
10 ng
10 ng
10 ng
10 ng
10 ng
10 ng
7.6 Spiking Solutions
Three solutions are prepared using the same stock as in Section
7.4. A native standard solution and a 13C12 internal quantita-
tion standard solution are prepared in isooctane (Tables 3 and
4). A recovery standard solution is prepared in tridecane (Ta-
ble 4). Samples are spiked with 100 pL of internal quantisation
standard solution and final sample extracts are spiked with 10 pL
of internal recovery standard solution.
8. HIGH RESOLUTION GAS CHROMATOGRAPHY/MASS SPECTROMETRY PERFORMANCE CRITERIA
Samples and standards are analyzed by using a Carlo Erba MFC500 gas chro-
matography (GC) coupled to a Kratos MS50TC double-focusing mass spectrom-
eter (MS) to be operated in the electron impact mode. The HRGC/MS inter-
face is simply a direct connection of the fused silica HRGC column to
the ion source of the MS via a heated interface oven. Data acquisition
and processing are controlled by a Finnigan-MAT Incos 2300 data system.
A-13
-------
Table 3. Native Spiking Solution1
Compound
2,3,7,8-TCDD
2,3,7,8-TCDF
1,2,3,7, 8-PeCDD
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
1,2, 3,7,8, 9-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,4,6,7,8-HpCDD
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDD
OCDF
Concentration
(pg/uL)
5
5
5
5
5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
25
25
Prepared in isooctane.
A-14
-------
Table 4. Internal Standard Spiking Solutions
Concentration
Compound (pg/uL)
Internal Quantitation Standards
13C12-2,3,7,8-TCDD 5
13C12-2,3,7,8-TCDF 5
13C12-l,2,3,7,8-PeCDD 5
13C12-l,2,3,7,8-PeCDF 5
13C12-l,2,3,6,7,8-HxCDD 12.5
13C12-l,2,3,4,7,8-HxCDF 12.5
13C12-l,2,3,4,6,7,8-HpCDD 12.5
13C12-l,2,3,4,6,7,8-HpCDF 12.5
13C12-OCDD 25
Internal Recovery Standard
13C12-1,2,3,4-TCDD 50
13C12-l,2,3,7,8,9-HxCDD 125
.Prepared in isooctane.
Prepared in tridecane.
A-15
-------
8.1 HRGC/MS Analysis of PCDD/PCDF
Single run selected ion monitoring (SIM) analysis of the tetra-
chloro through octachloro-dioxins and furans is carried out with
the instrumental conditions and parameters outlined in Table 5.
For each HRGC/MS run, five distinct groups of ions, which corre-
spond to each chlorine level, are sequentially monitored. These
ion descriptors are shown in Table 6. The masses of the two most
abundant ions in the molecular ion cluster of each dioxin and furan
and isotopically labeled standard are monitored. In addition,
the masses corresponding to the molecular ions of the hexachloro
through decachlorodiphenyl ethers (PCDEs) are monitored to aid in
the confirmation of positive furan results. Interference from
the presence of PCDE is noted by coincident response to the char-
acteristic ions for PCDFs. A lock mass, m/z 381 from PFK (per-
fluorokerosene), is used to observe and correct any magnet/instrument
drift during the analysis.
8.1.1 Tuning and Mass Calibration
The mass spectrometer is tuned on a daily basis to
yield optimum sensitivity and peak shape using an ion
peak (m/z 381) from PFK. The resolution is visually
monitored and maintained at fe 3,000 (10% valley defini-
tion) to provide adequate noise rejection while main-
taining good ion transmission.
Mass calibration of the mass spectrometer for the HRGC/MS
analysis of PCDD/PCDF is carried out on a daily basis.
The magnetic field is adjusted to pass m/z 300 at full
accelerating voltage. PFK is admitted to the MS and an
accelerating voltage scan from 8,000 to 4,000 V is ac-
quired by the data system. This corresponds to an effec-
tive mass range of 301 to 593 amu. Upon completion of
a successful calibration step, the five ion descriptors
shown in Table 6 are updated to reflect the new mass
calibration.
8.1.2 Ion Descriptor Switching
The ion descriptors shown in Table 6 are sequentially
monitored during a PCDD/PCDF analysis to cover the re-
tention windows of each chlorination level. The reten-
tion windows and hence the descriptor switch points are
determined initially and whenever a new HRGC column is
installed by injection of a mixture of PCDD and PCDF
congeners. Daily adjustment of the descriptor switch
times are performed when careful monitoring of the stan-
dard retention times shows this to be necessary. The
descriptors are designed to ensure acquisition of all
isomers of each homolog.
A-16
-------
Table 5. HRGC/LRMS Operating Conditions for PCDD/PCDF Analysis
Mass spectrometer
Accelerating voltage:
Trap current:
Electron energy:
Electron multiplier voltage:
Source temperature:
Resolution:
Overall SIM cycle time:
8,000 V
500 |JA
70 eV
-1,800 V
280°C
^ 3,000 (10% valley definition)
I s
Gas chromatograph
Column coating:
Film thickness:
Column dimensions:
He linear velocity:
He head pressure:
Injection type:
Split flow:
Purge flow:
Injector temperature:
Interface temperature:
Injection size:
Initial temperature:
Initial time:
Temperature program:
DB-5
0.25 pro
60 m x 0.25 mm ID
~ 25 cm/sec
1.75 kg/cm2 (25 psi)
Splitless, 45 s
30 mL/min
6 mL/min
270°C
300°C
1-2 uL
200°C
2 min
200°C to 330°C at 5°C/min
A-17
-------
Table 6. Ions Monitored for HRGC/MS of PCDD/PCDF
Descriptor ID
Al TCDF
13C12-TCDF
TCDD
13C12-TCDD
HxCDPE
PFK (lock mass)
A2 TCDF
TCDD
PeCDF
13C12-PeCDF
PeCDD
13C12-PeCDD
PFK (lock mass)
HpCDPE
A3 HxCDF
PFK (lock mass)
13C12-HxCDF
HxCDD
13C12-HxCDD
OCDPE
Mass
303.902
305.899
315.942
317.939
319.897
321.894
331.937
333.934
373.840
380.976
303.902
305.899
319.897
321.894
337.863
339.860
349.903
351.900
353.858
355.855
365.898
367.895
380.976
407.801
373.821
375.818
380.976
385.861
387.858
389.816
391.813
401.856
403.853
443.759
Nominal dwell
time (sec)
0.090
0.090
0.090
0.090
0.090
0.090
0.090
0.090
0.090
0.090
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.035
0.035
0.080
0.080
0.080
0.080
0.080
0.080
0.080
0.080
0.080
0.080
A-18
-------
Table 6 (continued)
Descriptor ID
A4 PFK (lock mass)
HxCDD
HpCDF
13C12-HpCDF
HpCDD
13C12-HpCDD
37Cl4-HpCDD
NCDPE
A5 PFK (lock mass)
OCDF
13C12-OCDF
OCDD
13C12-OCDD
DCDPE
Mass
380.976
389.816
391.813
407.782
409.779
419.822
421.819
423.777
425.774
435.817
437.814
429.768
431.765
477.720
380.976
441.743
443.740
453.783
455.780
457.738
459.735
469.779
471.776
511.681
Nominal dwell
time (sec)
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.06
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.06
A-19
-------
8.1.3 HRGC Column Performance (60-m DB-5)
The HRGC column performance must be demonstrated at the
start of each 12-h analysis period.
8.1.3.1 Inject 1 uL of the column performance check
solution (Section 7.5) and acquire selected
ion monitoring (SIM) data for m/z 320, 322,
332, and 334.
8.1.3.2 The chromatographic peak separation between
2,3,7,8-TCDD and the peaks representing
any other TCDD isomers should be resolved
with a valley of 30-60%, where
Valley % = (x/y)(100)
x = measured height of the valley between
the chromatographic peak correspond-
ing to 2,3,7,8-TCDD and the peak of
the nearest TCDD isomer; and
y = the peak height of 2,3,7,8-TCDD.
Figure 2 is an example of the separation of
a TCDD isomer mixture and the calculation
of isomer resolution.
It is the responsibility of the laboratory
to verify the conditions suitable for the
appropriate resolution of 2,3,7,8-TCDD from
all other TCDD isomers. The column perfor-
mance check solution also contains the TCDD
isomers eluting first and last under the
analytical conditions specified in this
protocol, thus defining the retention time
window for total TCDD determination. Any
individual selected ion current profile or
the reconstructed total ion current
(m/z 320 + m/z 322) consititutes an accept-
able form of data presentation.
8.1.4 Initial Calibration for PCDD/PCDF Analysis
Initial calibration is required before any samples are
analyzed for PCDD/PCDF. Initial calibration is also
required if any routine calibration does not meet the
required criteria listed in Section 8.1.7.
8.1.4.1 Tune and calibrate the instrument with PFK
as outlined in Section 8.1.1.
A-20
-------
o
CM
m o
in o
co in
— o'
CM
CO -tt
in o
oo m
oo • •
m — O
OO CO
S
00
-o
m
CM
CM
p o
cb m
co o
co
CO -H
Q
Q
U
l^J
Q
Q
O
S_
A-21
-------
8.1.4.2 Six of the eight concentration calibration
solutions listed in Table 2 will be analyzed
for the initial calibration phase. These
must include solutions CS4 through CSS
(Table 2). The analyst may select any of
the remaining solutions for demonstrating
calibration at the upper concentration
range.
8.1.4.3 Using the HRGC and MS conditions in Ta-
ble 5 and the SIM monitoring descriptors
in Table 6, analyze a 1-uL aliquot of each
of the six concentration calibration solu-
tions in triplicate.
8.1.4.4 Compute the relative response factors (RRFs)
for each analyte in the concentration cali-
bration solution using the criteria for
positive identification of PCDD/PCDF's
given in Section 14.1 and the computa-
tional methods in Section 14.2.
8.1.4.5 Compute the means and their respective
relative standard deviations (% RSD) for
the RRFs from each triplicate analysis for
each analyte in the standard.
8.1.4.6 Calculate the grand means (RRF) and their
respective RSDs using the six mean RRFs
for each analyte.
8.1.5 Criteria for Acceptable Initial Calibration
8.1.5.1 The % RSD for the response factors for each
triplicate analysis of a single concentra-
tion calibration standard for each analyte
must be less than ± 30% except for the TCDD
and TCDF, which must be less than ± 20%.
8.1.5.2 The variation of the mean RRFs for the
six concentration calibrated standards
(Section 8.1.5.1) must be less than 30%
except for the TCDD and TCDF which must
be less than 20%.
8.1.5.3 The SIM traces for all ions used for quan-
titation must present a signal-to-noise
(S/N) ratio of ^ 2.5. This includes ana-
lytes and isotopically labeled standards.
A-22
-------
8.1.5.4 Isotopic ratios must be within ± 20% of
the theoretical values (see Table 7).
NOTE: If the criteria for acceptable cali-
bration listed above have been met, the
RRF can be considered independent of the
analyte quantity for the calibration concen-
tration range. The grand mean RRF from
the initial calibration for unlabeled PCDD/
PCDFs and for the isotopically labeled
standards will be used for all calcula-
tions until routine calibration criteria
(Section 8.1.7) are no longer met. At such
time, new mean RRFs will be calculated from
a new set of six triplicate determinations.
8.1.6 Routine Calibrations
Routine calibrations must be performed at the beginning
of every day before actual sample analyses are performed
and as the last injection of every day.
8.1.6.1 Inject 1 uL of the concentration calibra-
tion solution CS 7 (see Table 2) as the
initial calibration check on each analysis
day. It is recommended that the analyst
select a concentration calibration solu-
tion that brackets the sample concentrations
observed on a single analysis date as the
last injection of each analysis date.
8.1.6.2 Compute the RRFs for each analyte in the
concentration calibration solution using
the criteria for positive identification
of PCDD/Fs given in Section 14.1 and the
computational methods in Section 14.2.
8.1.7 Criteria for Acceptable Routine Calibration
8.1.7.1 The measured RRF for all analytes must be
within ± 30% of the grand mean values es-
tablished by triplicate analysis of the
calibration concentration solutions, ex-
cept for TCDD and TCDF, which must be
within ± 20% of the mean values established
in the initial calibration step.
8.1.7.2 Isotopic ratios must be within ± 20% of the
theoretical value for each analyte and iso-
topically labeled standard (see Table 7).
A-23
-------
Table 7. Ion Ratios for HRGC/LRMS Analysis of PCDD/PCDF
Compound
TCDF
13C12-TCDF
TCDD
13C12-TCDD
PeCDF
13C12-PeCDF
PeCDD
13C12-PeCDD
HxCDF
13C12-HxCDF
HxCDD
13Cl2-HxCDD
HpCDF
13C12-HpCDF
HpCDD
13C12-HpCDD
OCDF
13C12-OCDF
OCDD
13C12-OCDD
Ions monitored
304/306
316/318
320/322
332/334
338/340
350/352
354/356
366/368
374/376
386/388
390/392
402/404
408/410
420/422
424/426
436/438
442/444
454/456
458/460
470/472
Theoretical ratio
0.76
0.76
0.76
0.76
0.61
0.61
0.61
0.61
1.22
1.22
1.22
1.22
1.02
1.02
1.02
1.02
0.87
0.87
0.87
0.87
Acceptabl
0.61 -
0.61 -
0.61 -
0.61 -
0.49 -
0.49 -
0.49 -
0.49 -
0.98 -
0.98 -
0.98 -
0.98 -
0.82 -
0.82 -
0.82 -
0.82 -
0.70 -
0.70 -
0.70 -
0.70 -
e range
0.91
0.91
0.91
0.91
0.73
0.73
0.73
0.73
1.46
1.46
1.46
1.46
1.22
1.22
1.22
1.22
1.04
1.04
1.04
1.04
A-24
-------
8.1.7.3 If any of the above criteria is not met,
a second attempt may be made before re-
peating the entire initialization process.
8.2 HRGC/HRMS Analysis (Isomer Specific TCDD Analysis)
Isomer specific analysis for 2,3,7,8-TCDD is carried out with the
instrumental conditions and parameters shown in Table 8. In addi-
tion to monitoring the masses of the most abundant molecular ions
of TCDD, an ion corresponding to the loss of COC1 from the molecu-
lar ion is monitored for verification purposes. Mass spectrometer
resolution is maintained at or above 10,000 (10% valley definition)
in order to increase the specificity of the analysis.
8.2.1 Tuning and Mass Calibration
8.2.1.1 The mass spectrometer must be operated in
the electron (impact) ionization mode.
Static resolving power of at least 10,000
(10% valley) must be demonstrated before
any analysis of a set of samples is per-
formed. Static resolution checks must be
performed at the beginning and at the end
of each 12-h period of operation. How-
ever, it is recommended that a visual
check (i.e., not documented) of the static
resolution be made before and after each
analysis.
8.2.1.2 The MS shall be tuned daily using PFK to
yield a resolution of at least 10,000 (10%
valley) and optimal response at m/z 254.986.
This step is followed by calibration of an
accelerating voltage scan of PFK beginning
at m/z 254 (typical calibration range is
255 to 493 amu). Other voltage scans from
the same data file are used to establish
and document both the resolution at m/z
316.983 and the mass measurement accuracy
at m/z 330.979.
8.2.1.3 Following calibration, the SIM experiment
descriptor is updated to reflect the new
calibration. Six masses (see Table 8) are
monitored by scanning ~ m/10,000 amu (atomic
mass units) over each mass. The total cycle
time is kept to I s. The m/z 280.983 ion
from PFK is used as a lock mass because it
is the most abundant PFK ion within the
range of m/z 255 to 334 and therefore per-
mits the use of low partial pressures of
PFK, which minimizes PFK interferences at
the analytical masses.
A-25
-------
Table 8. HRGC/HRMS Operating Conditions
Mass spectrometer
Accelerating voltage:
Trap current:
Electron energy:
Electron multiplier voltage:
Source temperature:
Resolution:
8,000 V
500 MA
70 eV
2,000 V
280°C
10,000 (10% valley definition)
SIM Parameters
Identity Mass
TCDD-COC1 258.930
TCDD 319.897
TCDD 321.894
13C12-TCDD 331.937
13C12-TCDD 333.934
PFK (lock mass) 280.983
Overall SIM cycle time = 1 s
Nominal dwell times (s)
0.15
0.15
0.15
0.15
0.15
0.10
Gas chromatograph
Column coating:
Film Thickness:
Column dimensions:
Helium linear velocity:
Helium head pressure:
Injection type:
Split flow:
Purge flow:
Injector temperature:
Interface temperature:
Injection size:
Initial temperature:
Initial time:
Temperature program:
CP-Sil 88
0.2 urn
50 m x 0.22 mm ID
^ 25 cm/s
1.75 kg/cm2 (25 psi)
Splitless, 45 s
30 mL/min
6 mL/min
270°C
240°C
2 ML
200°C
1 min
200°C to 240°C at 4°C/min
A-26
-------
8.2.2 Mass Measurement and Resolution Check
Using a PFK molecular leak, tune the instrument to meet
the minimum required resolving power of 10,000 (10% valley)
at m/z 254.986 (or any other mass reasonably close to
m/z 259). Calibrate the voltage sweep at least across
the mass range m/z 259 to m/z 334 and verify that m/z
330.979 from PFK (or any other mass close to m/z 334)
is measured within ± 5 ppm (i.e., 1.7 mmu, if m/z 331
is chosen) using m/z 254.986 as a reference. Documenta-
tion of the mass resolution must then be accomplished
by recording the peak profile of the PFK reference peak
m/z 318.979 (or any other reference peak at a mass close
to m/z 320/322). The format of the peak profile represen-
tation must allow manual determination of the resolution;
i.e., the horizontal axis must be a calibrated mass scale
(amu or ppm per division). The results of the peak width
measurement (performed at 5% of the maximum which corre-
sponds to the 10% valley definition) must appear on the
hard copy and cannot exceed 100 ppm (or 31.9 mmu if m/z
319 is the chosen reference ion).
8.2.3 HRGC Column Performance (50-m CP Sil 88/60-m SP-2330)
Prior to any HRGC/HRMS analysis of calibration solutions
or samples for 2,3,7,8-TCDD, the resolution of the HRGC
columns must be documented to be within allowable limits
in order to provide conditions adequate for unambiguous
isomer-specific analysis of 2,3,7,8-TCDD. This column
performance check must be demonstrated at the start of
each 12-h analysis period.
8.2.3.1 Inject 2 uL of the column performance check
solution and acquire selected ion monitor-
ing (SIM) data for m/z 258.930, 319.897,
321.894, 331.937, and 333.934 within a
total cycle time of g 1 s (Table 8).
8.2.3.2 The chromatographic peak separation between
2,3,7,8-TCDD and the peaks representing
any other TCDD isomers must be resolved
with a valley of ^ 25%, where
Valley % = (x/y)(100)
x = measured height of the valley between
the chromatographic peak correspond-
ing to 2,3,7,8-TCDD and the peak of
the nearest TCDD isomer; and
y = the peak height of 2,3,7,8-TCDD.
A-27
-------
8.2.3.3 If the above resolution requirement is not
met, corrective action must be taken and
acceptable resolution documented prior to
any further analyses. Corrective action
may include removal of the first meter of
the HRGC column, replacement or clearing
of the injector port, or complete replace-
ment of the GC column.
8.2.3.4 The column performance check solution also
contains the TCDD isomers eluting first
and last under the analytical conditions
specified in this protocol, thus defining
the retention time window for total TCDD
determination. The peaks representing
2,3,7,8-TCDD and the first and the last
eluting TCDD isomer should be labeled and
identified as such on the chromatograms (F
and L, respectively). Any individual se-
lected ion current profile or the recon-
structed total ion current (m/z 259 + m/z
320 + m/z 322) constitutes an acceptable
form of data presentation.
8.2.4 Initial Calibration for HRGC/HRMS 2,3,7,8-TCDD Analysis
Initial calibration is required before any samples are
analyzed for 2,3,7,8-TCDD. Initial calibration is also
required if any routine calibration does not meet the
required criteria listed in Section 8.2.6.
8.2.4.1 At least six of the concentration calibra-
tion solutions listed in Table 2 must be
utilized for the initial calibration.
These must include solutions CS4 through
CSS. The analyst may select any of the
remaining solutions for demonstrating cal-
ibration at the upper concentration range.
8.2.4.2 Tune and calibrate the instrument with PFK
as described in Section 8.2.1.
8.2.4.3 Inject 1 uL of the column performance check
solution (Section 8.2.3) and acquire SIM
mass spectra data for m/z 258.930, 319.897,
321.894, 331.937, and 333.934 using a total
cycle time of S 1 s (see Table 8). The
laboratory must not perform any further
analysis until it has been demonstrated
and documented that the criterion listed
in Section 8.2.3.2 has been met.
A-28
-------
8.2.4.4 Using the same GC and MS conditions (Ta-
ble 8) that produced acceptable results
with the column performance check solu-
tion, analyze a 1-uL aliquot of each of
the six concentration calibration solu-
tions in triplicate.
8.2.4.5 Calculate the RRFs for unlabeled 2,3,7,8-
TCDD relative to 13C12-2,3,7,8-TCDD and
the RRF for 13C12-2,3,7,8-TCDD relative to
13C12-1,2,3,4-TCDD using the criteria for
positive identification of TCDD by HRGC/
HRMS given in Section 14.1 and the computa-
tional methods in Section 14.2.
8.2.4.6 Calculate the six means (RRFs) and their
respective relative standard deviations
(% RSD) for the response factors from each
of the triplicate analyses for both un-
labeled and 13C12-2,3,7,8-TCDD.
8.2.4.7 Calculate the grand mean RRFs and their
respective relative standard deviations
(% RSD) using the six mean RRFs.
8.2.5 Criteria for Acceptable Initial Calibration
The criteria listed below for acceptable calibration
must be met before analysis of any sample is performed.
8.2.5.1 The percent relative standard deviation
(RSD) for the response factors from each
of the triplicate analyses of a single con-
centration calibration standard for both un-
labeled and 13C12-2,3,7,8-TCDD must be
less than 20%.
8.2.5.2 The variation of the mean RRFs from the
six concentration calibration standards
unlabeled and 13C12-2,3,7,8-TCDD must be
less than 20% RSD.
8.2.5.3 SIM traces for 2,3,7,8-TCDD must present a
signal-to-noise ratio of ^ 2.5 for m/z
258.930, m/z 319.897, and m/z 321.894.
8.2.5.4 SIM traces for 13C12-2,3,7,8-TCDD must
present a signal-to-noise ratio ^2.5 for
m/z 331.937 and m/z 333.934.
8.2.5.5 Isotopic ratios for 320/322 and 332/334
must be within the allowed range (0.61 to
0.91).
A-29
-------
NOTE: If the criteria for acceptable cali-
bration listed above have been met, the
RRF can be considered independent of the
analyte quantity for the calibration con-
centration range. The grand mean RRF from
the initial calibration for unlabeled
2,3,7,8-TCDD and for 13C12-2,3,7,8-TCDD
will be used for all calculations until
routine calibration criteria (Section 8.2.6)
are no longer met. At such time, new mean
RRFs will be calculated from a new set of
six triplicate determinations.
8.2.6 Routine Calibrations
Routine calibrations must be performed at the beginning
of a 12~h period after successful mass resolution and
HRGC column performance check runs and before analysis
of actual samples. The response factor calibration
must also be verified at the end of each analysis date.
8.2.6.1 Inject I pL of the concentration calibra-
tion solution (CS7, Table 2) which contains
2.5 pg/nL of unlabeled 2,3,7,8-TCDD, 50.0
pg/|jL of 13C12-2,3,7,8-TCDD, and 50 pg/pL
of 13C12-1,2,3,4-TCDD. Using the same HRGC/
MS/DS conditions as used in Table 8, deter-
mine and document acceptable calibration
as provided below.
8.2.7 Criteria for Acceptable Routine Calibration
The following criteria must be met before further analy-
sis is performed. If these criteria are not met, cor-
rective action must be taken and the instrument must be
recalibrated.
8.2.7.1 The measured RRF for unlabeled 2,3,7,8-TCDD
must be within 20% of the mean values estab-
lished in the initial calibration by trip-
licate analyses of concentration calibra-
tion solutions.
8.2.7.2 The measured RRF for 13C12-2,3,7,8-TCDD
must be within 20% of the mean value estab-
lished by triplicate analysis of the con-
centration calibration solutions during
the initial calibration.
A-30
-------
8.2.7.3 Isotopic ratios must be within the allowed
range (0.61 to 0.90).
8.2.7.4 If one of the above criteria is not satis-
fied, a second attempt can be made before
repeating the entire initialization process.
NOTE: An initial calibration must be car-
ried out whenever the routine calibration
solution is replaced by a new one from a
different lot.
9. QUALITY CONTROL PROCEDURES
9.1 Summary of QC Analyses
9.1.1 Initial and routine calibration and instrument perfor-
mance checks.
9.1.2 Analysis of a batch of samples with accompanying QC
analyses:
Sample batch — 10 NHATS adipose tissue samples plus
additional QC analyses including 1 method blank, a con-
trol tissue and a spiked tissue sample.
"Blind" QC (external QC) samples may be submitted by an
external source (quality assurance group or independent
laboratory) and included among the batch of samples.
Blind samples include spiked samples, unidentified dupli-
cates, and performance evaluation samples.
9.2 Performance Evaluation Solutions -- Included among the samples in
every third batch will be a solution provided by the quality con-
trol coordinator containing known amounts of unlabeled 2,3,7,8-
TCDD and/or other PCDD/PCDF isomers. The accuracy of measure-
ments for performance evaluation samples should be in the range
of 70-130%.
9.3 Column Performance Check Solutions
9.3.1 At the beginning of each 12-h period during which sam-
ples are to be analyzed, an aliquot of the HRGC column
performance check solution shall be analyzed to demon-
strate adequate HRGC resolution for selected TCDD isomers.
A-31
-------
9.4 Method Blanks
9.4.1 A minimum of one method blank is generated with each
batch of samples. A method blank is generated by per-
forming all steps detailed in the analytical procedure
using all reagents, standards, equipment, apparatus,
glassware, and solvents that would be used for a sample
analysis, but omit addition of the adipose tissue.
9.4.1.1 The method blank must contain the same
amounts of Carbon-13 labeled internal
quantisation standards that are added to
samples before bulk lipid cleanup.
9.4.1.2 An acceptable method blank exhibits no
positive response for any of the charac-
teristic ions monitored.
9.4.1.2.1 If the above criterion is not
met, solvents, reagents,-spik-
ing solutions, apparatus, and
glassware are checked to locate
and eliminate the source of
contamination before any samples
are extracted and analyzed.
9.4.1.2.2 If new batches of reagents or
solvents contain interfering
contaminants, purify or dis-
card them.
9.5 Control Samples -- Control samples are prepared from a bulk sam-
ple(s) of human adipose tissue or similar matrix (e.g., porcine
fat). This material is prepared by blending the tissue with
methylene chloride, drying the extract by eluting through anhy-
drous sodium sulfate, and removing the methylene chloride using
rotoevaporation at elevated temperatures (80°C). The evaporation
process should be extended to ensure all traces of the extraction
solvent have been removed. The resulting oily matrix (lipid) is
subdivided into 10-g aliquots which are analyzed with each sample
batch. The results of the individual analysis will be used to
give a measure of precision from batch to batch over an entire
program. Sufficient tissue should be extracted to provide a
homogeneous lipid matrix that can be used over the total analysis
program. Enough lipid matrix is necessary to prepare the spiked
samples describe in Section 9.6.
9.6 Spiked Samples -- Spiked lipid samples are prepared using a por-
tion of the homogenized lipid described in Section 9.5. Suffi-
cient spiked lipid matrix is prepared to provide a minimum of one
spiked sample per sample batch. It is recommended that a minimum
A-32
-------
of three spiked levels of the matrix are prepared ranging from 10
to 50 times the estimated limit of detection for each compound.
Each analysis of spiked sample must be accompanied by analysis of
a control sample in order to make the necessary corrections for
background contribution before determining the accuracy of the
method (Equation 9-1).
ft fo/\ -inno/ Cone, spiked sample-cone, control sample .. n ,
Accuracy (%) = 100% x * Spike level Eq' 9"1
9.7 Duplicate Sample Analysis -- When possible a duplicate analysis
of specific samples is included in the sample batch as an addi-
tional measure of method precision. It is suggested that the
total tissue sample is extracted to isolate lipids material and
then subdivided for duplicate analysis. Precision is calculated
as relative percent difference (RPD) where the differences in the
duplicate measurements (for each analyte) is divided by the aver-
age of the two measurements and multiplied by 100%.
9.8 External Samples -- Samples submitted as blinds to the analyst
may consist of either performance solutions of PCDD and PCDF con-
geners or spiked sample matrices. These performance solutions or
samples should be submitted by a source external to the analytical
program (QA unit of analysis laboratory or independent laboratory),
Performance audit solutions are intended to evaluate instrument
calibration and quantitation procedures. Spiked blind samples
must be accompanied by the corresponding unspiked samples to cor-
rect concentrations for background concentration. The blind
spiked samples are intended to evaluate the total analytical pro-
cedure. The analyst must keep in mind that it is necessary to
compare differences in standard sources for each type of external
sample.
10. SAMPLE PRESERVATION AND HANDLING
All adipose tissue samples must be maintained at less than -20°C from
time of collection. The analyst should instruct the collaborator col-
lecting the sample(s) to avoid the use of chlorinated materials. Sam-
ples are handled using stainless steel forceps, spatulas, or scissors.
Aliquots of samples removed from sample bottles not used for analysis
are disposed rather than returned to the sample vial. All sample bot-
tles (glass) are cleaned as specified in Section 6.4.10. Teflon®-!ined
caps should be used. As with any biological sample, the analyst should
avoid any undue exposure.
11. SAMPLE EXTRACTION
11.1 Extraction of Adipose Tissue
11.1.1 Accurately weigh to the nearest 0.01 g a 10-g portion
of a frozen adipose tissue sample into a culture tube
(2.2 x 15 cm).
Note: Sample size may be smaller, depending on avail-
abi1ity.
A-33
-------
11.1.2 Addition of internal quantisation standards
Allow the adipose tissue specimen to reach room tempera-
ture and then add the carbon-13 internal quantisation
spiking solution (Section 7.6) such that it delivers
500 to 2,500 pg of each of the surrogates specified in
Table 4 in a 100-uL volume.
11.1.3 Add 10 mL of methylene chloride and homogenize the mix-
ture for approximately 1 min with a Tekmar Tissuemizer®.
11.1.4 Allow the mixture to separate and decant the methylene
chloride extract from the residual solid material using
a disposable pipette. The methylene chloride is eluted
through a filter funnel containing a plug of clean glass
wool and 5 to 10 g of anhydrous sodium sulfate. The
dried extract is collected in a 100-mL volumetric flask.
11.1.5 A second 10-mL aliquot of methylene chloride is added
to the sample and homogenized for I min. The methylene
chloride is decanted, dried, and transferred to the
100-mL volumetric flask as specified in Section 11.1.3
11.1.6 The culture tube is rinsed with at least two additional
aliquots (10 mL each) of methylene chloride, and the
entire contents are transferred to the filter funnel
containing the anhydrous sodium sulfate. The filter
funnel and contents are rinsed with additional methylene
chloride (20 to 40 mL). The total eluent from the fil-
ter funnel is collected in the 100-mL volumetric flask.
Discard the sodium sulfate.
11.1.7 The final volume of the extract for each sample is ad-
justed to 100 mL in the volumetric flask using methylene
chloride.
11.2 Lipid Determination
11.2.1 Preweigh a clean 1-dram glass vial to the nearest
0.0001 g using an analytical balance tared to zero.
11.2.2 Accurately transfer 1.0 mL of the final extract (100 mL)
from Section 11.1.7 to the 1-dram vial. Reduce the vol-
ume of methylene chloride from the extract using a water
bath (50-60°C) gentle stream of purified nitrogen until
an oil residue remains.
A-34
-------
11.2.3 Accurately weigh the 1-dram vial and residue to the
nearest 0.0001 g and calculate the weight of lipid
present in the vial based on difference. Nitrogen
blow-down is continued until a constant weight is
achieved.
11.2.4 Calculate the percent lipid content of the original
sample to the nearest 0.1% as shown in Equation 11-1.
i R x FVT
Lipid content, LC (%) = rr^ rf^- x 100% Eq. 11-1
WAT x VAL
where: W,R = weight of the lipid residue to the
nearest 0.0001 g calculated from
Section 11.2.3;
VFYT = total volume of the extract in ml from
tAI Section 11.1.6 (100.0 ml);
W,- = weight of the original adipose tissue
samples to the nearest 0.01 g from
Section 11.1.1; and
V.. = volume of the aliquot of the final ex-
tract in mL used for the quantitative
measure of the lipid residue (1.0 ml).
11.2.5 Record the lipid residue measured in Section 11.2.3 and
the percent lipid content calculated from Section 11.2.4.
11.3 Extract Concentration
11.3.1 Quantitatively transfer the remaining extract volume
(99.0 ml) to a 500-mL Erlenmeyer flask. Rinse the volu-
metric flask with 20 to 30 ml of additional methylene
chloride to ensure quantitative transfer.
11.3.2 Place the Erlenmeyer flask on a hot plate at 40°C to
remove solvent until an oily residue remains.
12. CLEANUP PROCEDURES
12.1 Bulk Lipid Removal
12.1.1 Add a total of 200 mL of ri-hexane to the spiked lipid
residue in the 500-mL Erlenmeyer flask.
A-35
-------
12.1.2 Slowly add, with stirring, 100 g of the 40% w/w sulfuric
acid impregnated silica gel (Section 7.1.3). Stir with
a magnetic stir-plate for 2 h.
12.1.3 Allow solids to settle and decant liquid through a powder
funnel containing 20 g of anhydrous sodium sulfate and
collect in a 500-mL sample bottle.
12.1.4 Rinse solids with two 50-mL portions of hexane. Stir
each rinse for 15 min, decant, and dry by elution
through sodium sulfate combining the hexane extracts
from Section 12.1.3.
12.1.5 After the rinses have gone through the sodium sulfate,
rinse the sodium sulfate with an additional 25 ml of
hexane and combine with the hexane extracts from Sec-
tion 12.1.4.
12.1.6 Prepare an acidic silica column as follows: Pack a
1 cm x 10 cm chromatographic column with a glass wool
plug, add approximately 25 ml of hexane, add 1.0 g of
silica gel (Section 7.1.2) and allow to settle, then
add 4.0 g of 40% w/w sulfuric acid impregnated silica
gel (Section 7.1.3) and allow to settle. Pack a second
chromatographic column (1 cm x 30 cm) with a glass wool
plug, add approximately 25 ml of hexane, add 6.0 g of
acidic alumina (Section 7.1.1), and allow to settle and
then top with a 1-cm layer of sodium sulfate (Section
7.2). Elute the excess hexane solvent through the
columns until the solvent level reaches the top of the
chromatographic packing. Inspect columns to ensure they
are free of channels and air bubbles. Wash the alumina
column with 40 mL of 50% v/v methylene chloride/hexane.
Remove the methylene chloride from the adsorbent by
eluting the column with an additional 100 mL of hexane.
Elute the excess solvent from the column until the
solvent level reaches the top of the sodium sulfate layer.
12.1.7 Quantitatively transfer the hexane extract from the
Erlenmeyer flask (Sections 12.1.3 through 12.1.5) to
the silica gel column reservoir. Allow the hexane ex-
tract to percolate through the column and collect in a
KD concentrator.
12.1.8 Complete the elution of the extract from the silica gel
column with 50 ml of hexane in the KD concentrator.
Concentrate the eluate to approximately 1.0 ml, using
nitrogen blow-down as necessary.
A-36
-------
Note: If the 40% sulfuric acid/silica gel is noted to
be highly discolored throughout the length of the ad-
sorbent bed it is necessary to repeat the cleaning pro-
cedure beginning with Section 12.1.1.
12.2 Separation of Chemical Interferences
12.2.1 Transfer the concentrate (1.0 ml) to the top of the
alumina column. Rinse the K-D assembly with two 1.0-mL
portions of hexane and transfer the rinses to the top
of the alumina column. Elute the alumina column with
18 ml of hexane until the hexane level is just below
the top of the sodium sulfate. Discard the eluate.
Columns must not be allowed to reach dryness (i.e., a
solvent "head" must be maintained).
12.2.2 Place 30 mL of 20% (v/v) methylene chloride in hexane
on top of the alumina and elute the TCDDs from the col-
umn. Collect this fraction in a 50-mL culture tube.
12.2.3 Prepare an 18% Carbopak C/Celite 545® mixture by thor-
oughly mixing 3.6 g of Carbopak C (80/100 mesh) and
16.4 g of Celite 545® in a 40-mL vial. Activate at
130°C for 6 h. Store in a desiccator. Cut off a clean
5-mL disposable glass pipet (6 to 7 mm ID) at the 4-mL
mark. Insert a plug of glass wool and push to the 2-mL
mark. Add 500 mg of the activated Carbopak/Celite mix-
ture followed by another glass wool plug. Using two
glass rods, push both glass wool plugs simultaneously
towards the Carbopak/Celite mixture and gently compress
the Carbopak/Celite plug to a length of 3 to 3.5 cm.
Pre-elute the column with 2 ml of toluene followed by
1 mL of 75:20:5 methylene chloride/methanol/ benzene,
1 ml of 1:1 cyclohexane in methylene chloride, and 2 ml
of hexane. The flow rate should be less than 0.5 mL/min.
While the column is still wet with hexane, add the entire
eluate (30 ml) from the alumina column (Section 12.2.2)
to the top of the column. Rinse the culture tube which
contained the extract twice with 1 ml of hexane and add
the rinsates to the top of the column. Elute the column
sequentially with two 1-mL aliquots of hexane, 1 ml of
1:1 cyclohexane in methylene chloride, and 1 mL of
75:20:5 methylene chloride/methanol/benzene. Turn the
column upside down and elute the PCDD/PCDF fraction with
20 mL of toluene into 6-dram vial.
12.2.4 Using a stream of nitrogen, reduce the toluene volume
to approximately 1 mL. Carefully transfer the concen-
trate into a 1-mL mini vial and reduce the volume to
about 200 pL using a stream of nitrogen.
A-37
-------
12.2.5 Rinse the concentrator tube with three washings using
500 uL of 1% toluene in methylene chloride. Concen-
trate to 200-500 ML and add 10 uL of the tridecane
solution containing the internal recovery standard and
store the sample in a refrigerator until HRGC/MS analy-
sis.
12.2.6 Immediately prior to analysis, using a gentle stream of
nitrogen at room temperature, remove toluene and methylene
chloride. Submit sample to HRGC/MS once a stable 10 uL
volume of tridecane is attained.
13. ANALYTICAL PROCEDURES
13.1 HRGC/MS Analysis for PCDD/PCDF
13.1.1 Once routine calibration criteria are met, the instru-
ment is ready for sample analysis. Prior to the first
sample, a blank injection of tridecane should be analyzed
to document system cleanliness. If any evidence of sys-
tem contamination is found, corrective action must be
taken and another tridecane blank analyzed.
The typical daily sequence of injections is shown in
Table 9 and Figure 3.
Note: Syringe Technique -- Congeners of PCDD/PCDF in the
syringes used for HRGC/MS analysis can be problematic un-
less the syringes are properly handled between samples.
The following procedure has been found to be very effec-
tive for PCDD/PCDF removal from contaminated syringes
and will be used throughout these analyses.
Rinse the syringe 10 times with isooctane.
Fill the syringe with toluene and sonicate syringe
and plunger in toluene for 5 min and repeat at least
twice.
Rinse the syringe 10 times with tridecane and pull
up 1 uL of clean tridecane.
Syringe is ready for use.
At no time should air be introduced into the HRGC column
by using an air plug in the syringe. The oxygen present
in the air plug will quickly degrade a nonbonded GC phase.
13.1.2 Inject a 1-uL aliquot of the extract into the GC, oper-
ated under the conditions previously used (Section 8.1)
to produce acceptable results with the performance check
solution.
A-38
-------
Table 9. Typical Daily Sequence for PCDD/PCDF Analysis
1. Tune and calibrate mass scale versus perfluorokerosene (PFK).
2. Inject column performance mixture.
3. Inject concentration calibration solution 2.5 to 12.5 pg/ul_ (CS-7)
solution.
4. Inject blank (tridecane).
5. Inject samples 1 through "N".
6 Inject concentration calibration solution 2.5 to 12.5 pg/[A (CS-7)
solution or other concentration calibration solutions CS1 to CSS to
bracket observed sample concentration.
A-39
-------
INSTRUMENTAL ANALYSIS
Instrument Mass Calibration vs PFK
Mass Resolution Check
I
Column Performance Evaluation
Does Column
Performance Meet
Minimum Resolution
Requirements?
No
Adjust Column
Length or Install
New Column
Yes
Calibration Standard Analysis
Do Relative
Response Factors Meet
Criteria Based on Initial
Calibrations ?
No
Reanalyze or Prepare Fresh
Calibration Standards and
Calibration Curve
Yes
Proceed with Sample Analysis
Figure 3. Daily QA procedures for proceeding with sample analysis,
A-40
-------
13.1.3 Acquire SIM data according to the same acquisition and
MS operating conditions previously used (Section 8.1)
to determine the relative response factors.
13.1.3.1 Acquire SIM data for the characteristic
ions designated in Table 6.
13.1.3.2 Instrument performance shall be monitored
by examining and recording the peak areas
for the recovery standard, 13C12-1,2)3,4-TCDD.
If this area should decrease to less than
50% of the calibration standard, sample
analyses shall be stopped until the problem
is found and corrected.
13.2 HRGC/HRMS Confirmation of 2,3,7,8-TCDD
The presence of 2,3,7,8-TCDD observed through the general PCDD
and PCDF procedure should be confirmed using HRGC/HRMS (resolu-
tion 10,000).
13.2.1 Once the daily criteria of mass calibration, mass reso-
lution, HRGC performance, and routine calibration are
met and documented, the instrument is ready for sample
analysis. Prior to the first sample, a blank injection
of tridecane will be made to document system cleanliness.
The typical daily schedule for HRGC/HRMS analysis of
TCDD is shown in Table 10 and Figure 3.
13.2.2 Inject a 1-pL aliquot of the extract into the GC, oper-
ated under the conditions previously used (Section 8.2)
to produce acceptable results with the column performance
check solution.
13.2.3 Acquire SIM data according to Section 8.2.4.3. Use the
same acquisition and MS operating conditions previously
used to determine the relative response factors.
13.2.3.1 Acquire SIM data for the following selected
characteristic ions:
m/z Compound
258.930 TCDD - COC1
319.897 Unlabeled TCDD
321.894 Unlabeled TCDD
331.937 13C12-2,3,7,8-TCDD>
13C12-1,2,3,4-TCDD
333.934 13C12-2,3,7,8-TCDD,
13C12-1,2,3,4-TCDD
A-41
-------
Table 10. Typical Daily Schedule for HRGC/HRMS Analysis of TCDD
1. Tune and calibrate mass scale.
2. Perform mass measurement check and mass resolution check.
3. Inject column performance check solution.
4. Inject the routine concentration calibration solution (CS7) and confirm
response factor consistency.
5. Inject tridecane blank.
6. Inject samples 1 through "N".
7. Inject concentration calibration solution and confirm response factor
consistency.
8. Mass resolution check.
A-42
-------
14. DATA REDUCTION
In this section, the procedures for the data reduction are outlined for
the analysis of data from both the HRGC/MS method for PCDD/PCDF and the
HRGC/HRMS method for 2,3,7,8-TCDD. Figure 4 presents a schematic of the
qualitative criteria for identifying PCDDs and PCDFs.
14.1 Qualitative Identification
14.1.1 The ion current responses for each mass for a particular
PCDD/PCDF analyte must be within ± 1 s to attain posi-
tive identification of that analyte. For example,
m/z 338 and m/z 340 must have maximum peak responses
that are within ± 1 s to be positively identified as
a pentachlorodibenzofuran.
14.1.2 The ion current intensities for a particular PCDD/PCDF
must be i 2.5 times the noise level (S/N ^ 2.5) for
positive identification of that isomer.
14.1.3 The integrated ion current ratios of the analytical
masses for a particular PCDD/PCDF must fall within the
ranges shown in Table 7.
14.1.4 The recovery of the internal quantitation standards
should be between 50 and 115%.
14.2 Quantitative Calculations
14.2.1 Relative response factors for native PCDD and PCDF
analytes (RRF). RRFs are calculated from the data ob-
tained during the analysis of concentration calibration
solutions using the following formula:
A • r
STD IS
RRF = AblU. r Eq. 14-1
HIS STD
where A^., = the sum of the areas of the integrated
ion abundances for the analyte in question.
For example, for TCDD, A,.,-,, would be the
sum of the integrated ion abundances for
m/z 320 and 322;
AT<- = the sum of the areas of the integrated ion
abundances for the labeled PCDD/F used as
the internal quantitation standard for the
above analyte. For example, for 13C12"
2,3,7,8-TCDD, AJS would be the sum of the
integrated ion abundance for m/z 332 and 334.
= concentration of the analyte in pg/pL;
A-43
-------
HRGC/MS-SlMData
Response to
Characteristic Molecular
Ions within the Appropriate
Homolog Retention
Window?
Characteristic
Ion Ratios within ±20%
Theoretical ?
Response
Corresponds to Specific
Isomer Retention
Time?
Report Compounds as
Not Detected (ND)
Calculate Sample LOP
Response Due to
Coextracted Interference
Quantitate Compound
as Per Protocol
Report as Isomer Unknown
Quantitate Specific Isomer as per Protocol
Figure 4. Qualitative criteria for identifying
PCDDs and PCDFs,
A-44
-------
CT<- = concentration of the internal quantitation
standard in pg/pL; and
Table 11 provides the pairing of target analytes to
internal quantitation standards for determining RRF
values for PCDD and PCDF compounds.
14.2.2 Relative response factors for the internal quantitation
standards (RRFIS). The RRFJS values are calculated from
data obtained during the analysis of concentration cali-
bration solutions using the following formula.
- AIS X CRS Eq. 14-2
IS ~ A x C
ib ARS X UIS
where AJS and GIS are defined as given in Section 14.2.1
and
CRS = concentrations of the internal recovery
standard in pg/uL; and
ARC. = the sum of the areas of the integrated ion
abundances for the labeled PCDD (13C12-
1,2,3,4-TCDD or 13C12-1,2,3,7,8,9-HxCDD).
For example, for 13C12-1,2,3,4-TCDD, ARS
would be the sum of the integrated ion
abundance for m/z 332 and 334.
Refer to Table 11 for pairing of the internal quantita-
tion standards with the appropriate internal recovery
standard.
14.2.3 Concentrations of sample components. Figure 5 presents
a schematic for quantitation of PCDDs and PCDFs which
meet the criteria specified in Section 14.1. Calculate
the concentration of PCDD/Fs in sample extracts using
the formula:
^_«i« " XTC . -i nn
Eq. 14-3
where C , = the lipid adjusted concentration of PCDD or
sample pCDf. congener in pg/g.
A -, = sum of the integrated ion abundances deter-
sample mined for the PCDD/PCDF in question;
AT<- = sum of the integrated ion abundances deter-
mined for the labeled PCDD/F used as the
internal quantitation standard for the above
analyte;
A-45
C ,
sample
_ Asample " QIS •
Mjf. * KKr * "A~I
• 100
r " LC
-------
Table 11. Target Analyte/Internal Quantitation Standard and Internal
Quantisation Standard/Internal Recovery Standard Pairs
Internal standards
Target analyte
Quantitation
Recovery
2,3,7,8-TCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
1,2,3,4,6,7,8-HpCDD
OCDF
OCDD
13C12-2,3,7,8-TCDD
13C12-2,3,7,8-TCDF
13C12-l,2,3,7,8-PeCDF
13C12-l,2,3,7,8-PeCDF
13C12-l,2,3,7,8-PeCDD
13C12-l,2,3,4,7,8-HxCDF
13C12-l,2,3,4,7,8-HxCDF
13C12-l,2,3,4,7,8-HxCDF
13C12-l,2,3,4,7,8-HxCDF
13C12-l,2,3,6,7,8-HxCDD
13C12-l,2,3,6,7,8-HxCDD
13C12-l,2,3,6,7,8-HxCDD
13Cl2-l,2,3,4,6,7,8-HpCDF
13C12-l,2,3,4,6,7,8-HpCDF
13C12-l,2,3,4,6,7,8-HpCDD
13C12-OCDD
13C12-OCDD
13C12-1,2,3,4-TCDD
13C12-1,2,3,4-TCDD
13C12-1,2,3,4-TCDD
13C12-1,2,3,4-TCDD
13C12-1,2,3,4-TCDD
13C12-l,2,3,7,8,9-HxCDD
13C12-l,2,3,7,8,9-HxCDD
13C12-l,2,3,7,8,9-HxCDD
13C12-l,2,3,7,8,9-HxCDD
13C12-l,2,3,7,8,9-HxCDD
13C12-l,2,3,7,8,9-HxCDD
13C12-l,2,3,7,8,9-HxCDD
13C12-l,2,3,7,8,9-HxCDD
13C12-l,2,3,7,8,9-HxCDD
13C12-l,2,3,7,8,9-HxCDD
13C12-l,2,3,7,8,9-HxCDD
13C12-l,2,3,7,8,9-HxCDD
A-46
-------
QUANTITATION
HRGC/MS-SIM Data
Report as Not Detected
Calculate Sample LOD
Response
Meets Al I
Qualitative
Criteria
Response
>2.5times
S/N?
Response
>10 times
S/N?
Calculate as per Protocol
Report as Trace (tr) Value
Quantitate as per Protocol
Report as Positive Quantifiable Value
Figure 5. Procedure for quantitation of PCDDs and PCDFs
in human adipose tissue.
A-47
-------
QIS = the amount (total pg) of the labeled internal
quantisation standard added to the sample
prior to extraction;
RRF = relative response factor of the above
analyte relative to its labeled internal
quantitation standard determined from the
initial triplicate calibration;
W.,. = weight (g) of original adipose tissue
sample; and
LC = percent extractable lipid determined from
Eq. 11-1.
Refer to Table 11 for pairing of target analytes with
the appropraite internal quantitation standard.
Quantitative data should be classified to indicate the
intensity of the signal response. Suggested qualifiers
include: not detected, ND (signal-to-noise ratio is
less than 2.5); trace, TR (signal-to-noise ratio is
greater than or equal to 2.5 but less than 10); and
positive quantifiable, PQ (signal-to-noise ratio is
greater than or equal to 10).
14.2.4 Recovery of internal quantitation standards. Calculate
the recovery of the labeled internal quantitation stan-
dards measured in the final extract using the formula:
Internal Quant. Std. AIS ' QRS inn _ ... .
Percent Recovery AR$ • QJS • RRF " 1UU tq' ^'4
where A,,. = sum of the integrated ion abundances deter-
mined for the labeled PCDD/PCDF internal
quantitation standard in question;
Ap<; = sum of the integrated ion abundances deter-
mined for m/z 332 and m/z 334 of 13C12-
1,2,3,4-TCDD or m/z 390 and m/z 392 of
13C13-l,2,3,7,8,9-HxCDD (recovery standards)
Q = amount (pg) of the respective recovery
standard, added to the final extract;
QT<. = amount (pg) the labeled internal quantita-
tion standard added to the sample prior to
extraction; and
A-48
-------
RRF
IS
relative response factor for the labeled
internal quantisation standard in question
relative to the internal recovery standard.
This value shall be the RRF determined from
the initial calibration.
Refer to Table 11 for pairing of the internal quanti-
tation standards with the appropriate target analytes.
Note: The result of calculations as presented in Sec-
tion 14.2 may be off by as much as 1% due to the fact
that 1 ml of the final 100 ml volume from the extrac-
tion was used for lipid determination.
14.3 Estimated Method Detection Limit
Estimated method detection limits must be calculated in situations
where (1) no response is noted for a specific congener; (2) a re-
sponse is noted but ion ratios are incorrect; and (3) where a re-
sponse is quantitated as a trace value.
14.3.1 For samples in which no unlabeled PCDD or PCDF is de-
tected, calculate the estimated minimum detectable con-
centration. The background area is determined by inte-
grating the ion abundances for the characteristic ions
in the appropriate region and relating the product area
to an estimated concentration that would produce that
product area.
Use the formula:
2.5
CE =
sample
A
Eq. 14-5
IS
RRF • W
AT
where CE = estimated concentration of unlabeled
or PCDF required to produce A , ;
sample
PCDD
sample
IS
sum of integrated ion abundances or peak
heights for the characteristic ions of the
unlabeled PCDD or PCDF isomer in the same
group of S 5 scans used to measure A
IS'
and
sum of integrated ion abundances for the
appropriate ions characteristic of the re-
spective internal quantisation standard.
Qj<-, RRF, and WAT retain the definitions previously
stated in Section 14.2. Alternatively, if peak height
measurements are used for quantification, measure the
estimated detection limit by the peak height of the noise
in the 2,3,7,8-TCDD RT window.
A-49
-------
14.3.2 For samples for which a response at the retention time
of a specific PCDD or PCDF congener is noted, but the
qualitative criteria for ion ratios are outside the
acceptable range (Table 7), the estimated detection
level is calculated as given in Eq. 14.3 except the
values are qualified as not detected, ND, and the
concentration is reported in parenthesis.
14.3.3 If a response for a specific PCDD or PCDF congener is
qualified as a trace, TR, value (signal to noise is
greater than or equal to 2.5 but less than 10) the
analyst must also provide an estimated method detection
limit. This is accomplished by using the observed sig-
nal to noise on either side of the response and calcu-
lating as given in Eq. 14-5.
15. REPORTING AND DOCUMENTATION
All data should be reported on an individual sample basis using the data
report format shown in Figure 6. The analyst is required to maintain
all raw data, calculations, and control charts in a format as to allow a
complete external data review. Suggested data formats for tracing cal-
culations are provided in Figure 7.
A-50
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF TOXIC SUBSTANCES
EXPOSURE EVALUATION DIVISION (TS-79S)
WASHINGTON, DC 20460
Pig« 1 of 1
NATIONAL HUMAN ADIPOSE TISSUE SURVEY
ANALYSIS REPORT FORM
EPA SAMPLE NUMBER _
LAB NUMBER
BATCH NUMBER
ANALYSIS DATE
MS ANALYST
REPORT DATE _
REPORTED BY _
NATIVE
COMPOUNDS
23.78-TCDO
2 3.7 8-TCDF
1 2 3.7,8-PeCOO
'237 8-PaCOF
23.478-PeCOF
1 23,4,78-HtCDD
1 2 3.6 7,8-HxCOD
1 23.789-HxCDD
12347 S-HxCDF
1 2 3 6,7,8-HxCDF
i 2378.9-HxCDF
2.3 4 6.7 3-HxCOF
1,2,3.4673-HoCDD
1 23.4678-HoCDF
123478 9-HoCDF
OCDO
OCDF
CONCENTRATION
(pg/gxl/
1 , , , !•! , 1
1 , . , |«l , 1
1 , ,,!•!, 1
1 , , , !•! , 1
1 ,,,!•! i 1
1 , , , 1*1 , 1
1 ,,,!•! i 1
1 , L , 1*1 1 J
1 , , , Ul , 1
1 ,,,!•! . 1
1 . , , Ul , 1
1 1 ! , l«l_lJ
1 , , , !•) , J
1 .,,!•! i 1
1 .,,!•!. |
1 , , , 1*1 , I
1 ,,,!•!, 1
DATA
QUALIFIER^/
INTERNAL OUANTITATION
STANDARD
13C, 2-2.3,7 8-TCDO
l3C12-2.3.7,a-TCDF
I3C,2-I.2.3.7.8-P9CDD
13C,2-I 2378-PeCOF
'^Cij-l 2.3.8 7 8-H«CDD
13C12-l,2.3.478-HxCDF
13C, 2-1.2.3,4.6.7 8-HpCDD
13C12-I 2 3,4,6.7 8-HpCDF
' 30,2-0000
13C,2-OCOF
SPIKED LEVEL
ipg)
PERCENT (%>
RECOVERY
REMARKS
Concentration reponefl is oased on total extractabie iioid (g)
NO - Not Deiected TR - Trace PC * Positive Quantifiable
Figure 6. Analysis report form.
A-51
-------
cS
O ro
- N
.
•r- 00
jj VO
ro ro
C ID
O roi
N M
-«\
O E
O
c
o
s_
0)
-a
a>
o a;
M- to
o
+-> a.
tO •!-
E -0
S- (O
O
M- C
C 01
3 0. OJ
O I CU
12-
13
Q.
(13 »«
m 13 c
*j -p- ai x
u CL-U •
n) •<- c x
t — O X
*J U ^-
X
•p
rfl
1^.
CD
S_
3
CD
o
a.
Q)
A-52
-------
TECHNICAL REPORT DATA
(Please read Inurucnons on (lie reverse before completing)
1 a£?oaT NO. 2.
EPA 560/5-86-020
4. 7!TL = AND SUBTITLE
Analysis for Polychlorinated Dibenzo-p-dioxins (PCDD) and
Dibenzofurans (PCDF) in Human Adipose Tissue: Method
Evaluation Study
7. AUTHORS js Stanley, RE Ayling, KM Bauer, MJ McGrath,
TM Sack, and KR Thornberg
a, *E,R FORMING ORGANIZATION NAME AND ADDRESS
Thdwest Research institute
425 Volker Boulevard
Kansas City, MO 64110
12. SPONSORING AGENCY NAME AND ADDRESS
Field Studies Branch (TS-798) Washinqton, DC 20460
Exposure Evaluation Division
Office of Toxic Substances
U.S. Environmental Protection Agpnrv
3. RECIPIENT'S ACCESSlOONO.
5. REPORT OATc
September 17, 1986
6. PERFORMING ORGANIZATION COOS
Midwest Research Institute
8. PERFORMING ORGANIZATION REPORT MC.
8824-A(01)
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-3938
68-02-4252
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
J Remmers, Work Assignment Manager; J Breen, Project Officer
16. A3STRACT
This report focuses on the evaluation of an HRGC/MS analytical method for determi-
nation of 2,3,7,8-substituted polychlorinated dibenzo-p_-dioxins (PCDD) and dibenzofurans
(PCDF) in human adipose tissue. This method will be used for analysis of samples from
EPA's National Human Adipose Tissue Survey (NHATS) as part of a collaborative effort be-
tween EPA's Office of Toxic Substances and the Veterans Administration. The method was
evaluated using aliquots of a bulk lipid matrix that was extracted from human adipose
tissue. The results of the replicate analysis of spiked and unspiked homogenized human
adipose tissue matrix demonstrate that the analytical method produces accurate and pre-
cise data for 17 specific 2,3,7,8-substituted PCDD and PCDF (tetra- through octachloro
homologs) congeners. The endogenous or background levels of the PCDD and PCDF congeners
in the homogenized adipose lipid matrix were estimated through regression analyses of
measured versus spiked concentrations for each compound. This unspiked matrix will be
used as a control sample with each batch of samples analyzed.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
3olychlorinated dibenzo-£-dioxin (PCDD)
3olychlorinated dibenzofuran (PCDF)
Human adipose tissue
Method evaluation
2,3,7,8-Tetrachlorodibenzo-p-dioxin
(2,3,7,8-TCDD)
2 ,3 ,7 ,8-Tetrachl orodi benzofuran (2.3. 7 .8-TrnF
13. ;iiTSi3uT:GN STATEMENT
Release unlimited
b. IDENTIFIERS/OPEN ENDED TERMS
)
19. SECURi TY CLASS / IJiu Report)
Unclassified
20. SECURITY CLASS / /Tin pafti
Unclassified
c. COSATi Fisiit'Group
21 NO. OF PAGca
145
22. PRICE
EPA firm 2220-1 (9-73)
------- |