EPA-600-4/88-029
HEATED PURGE AND TRAP METHOD
DEVELOPMENT AND TESTING
by
Samuel V. Lucas and Hazel M. Burkholder
BATTELLE
Columbus Division
Columbus, Ohio 43201-2693
Contract Number 68-03-3224
Work Assignment 1-09
Technical Project Monitors: James Longbottom
Thomas Pressley
Project Officer: Robert O'Herron
Environmental Monitoring and Support Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
REPRODUCEDBY
U.S. DEPARTMENT OF COMMERCE
NATIONAL TECHNICAL
INFORMATION SERVICE
SPRMGFIELD.VA 22161
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA/600/4-88/029
2.
3. RECIPIENT'S ACCESSION NO.
PBS 8 2426Q7I&S
4. TITLE AND SUBTITLE
Heated Purge and Trap Method Development and Testing
5. REPORT DATE
August 1988
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
S. V. Lucas and H. M. Burkholder
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Battelle Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201-2693
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO
68-03-3224
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Research and Development
Environmental Monitoring and Support Laboratory
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA 600/06
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The goal of this research was to develop a heated purge and trap method
that could be used in conjunction with SW-846 method 8240 for the analysis
of volatile, water soluble Appendix VIII analytes. The developed method was
validated according to a partial single laboratory method validation test to
determine its performance characteristics using mass spectrometric
detection. A group of 33 polar analytes comprising selected aldehydes,
ketones, alcohols, ethers, nitriles, a thiol and 9 nitrogen bases were
examined. All but eight of the compounds were eliminated from the method
due to:
1. Poor chromatographic performance,
2. Hydrolytic instability,
3. Purge and trap/desorb unsuitability, and
4. Compounds not commercially available.
The eight compounds found suitable for analysis for heated purge and
trap were: acrolein, methacrylonitrile, acrylonitrile, acetonitrile,
2-butanone, 1,4-dioxane, isobutanol and propionitrile.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
18. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (Tim Report)
Not Classified
21. NO. OF PAGES
81
20. SECURITY CLASS (Tins page/
Not Classified
22. PRICE
EPA. Form 2220-1 (R«v. 4-77)
PREVIOUS EDITION IS OBSOLETE .
-------
NOTICE
"The information in this document has been funded wholly or in part by
the United States Environmental Protection Agency under Contract Number 68-
03-3224 (Work Assignment 1-09) to Battelle Memorial Institute, Battelle
Columbus Division, Columbus, Ohio 43201-2693. It has been approved for
publication as an EPA document. Mention of trade names or commercial
products does not constitute endorsement or reconmendation."
-------
FOREWORD
Environmental measurements are required to determine the quality of
ambient waters and the character of waste effluents. The Environmental
Monitoring and Support Laboratory - Cincinnati, Ohio conducts research to:
o Develop and evaluate methods to measure the presence and
concentration of physical, chemical, and radiological
pollutants in water, wastewater, bottom sediments, and solid
waste.
o Investigate methods for the concentration, recovery, and
identification of viruses, bacteria and other
microbiological organisms in water; and, to determine the
responses of aquatic organisms to water quality.
o Develop and operate an Agency-wide quality assurance program
to assure standardization and quality control of systems for
monitoring water and wastewater.
o Develop and operate a computerized system for instrument
automation leading to improved data collection, analysis,
and quality control.
This report is concerned with one of the steps in an on-going program
to develop and evaluate analysis methods addressing organic analytes for
inclusion in EPA SW-846. In the present case, a method for the analysis of
polar, water soluble organic analytes has been developed and tested. The
developed method involves sample processing using a heated purge-trap-desorb
approach. The developed method was tested using gas chromatograph-mass
spectrometry techniques.
Robert L. Booth, Director
Environmental Monitoring and Support
Laboratory - Cincinnati
n i
-------
Abstract
A heated purge-trap-desorb (H-PTD) analysis method for polar, water
soluble volatile organic analytes was developed and tested. The research
goal was to obtain analysis sensitivities for the target analytes comparable
to those achieved presently for nonpolar analytes using the room temperature
PTD approach of EPA SW-846 Method 5030 while at the same time not deviating
significantly from the type of apparatus currently in use for Method 5030.
Thirty-three polar, water soluble analytes from Appendix VIII and the
Michigan list were included in the scope of work. Eight of those analytes
proved to be amenable to the developed method: acrolein, methyl ethyl
ketone, methylacrylonitrile, acrylonitrile, acetonitrile, propionitrile,
1,4-dioxane and isobutanol. The GC-MS method detection limits for the
analytes ranged from 2 to 9 ug/L and were about 5-fold higher than would be
obtainable with more optimized chromatography. Three alcohols, propargyl
alcohol, 2-chloroethanol and l,3-dichloro-2-propanol could not be recovered
to any degree by H-PTD even using 80 percent saturation of the sample with
salts and dramatically increased purge gas volumes. One analyte, 3-
chloropropionitrile, could not be recovered from the trap desorption step,
even though it was shown to be only slightly hydrolytically unstable. The
other 21 analytes had plausible chemical reasons for the absence of any H-
PTD recovery for them: nine were nitrogen bases, six were hydrolytically
unstable, two were mercaptans and four had very low volatility and/or very
high water solubility. The absolute H-PTD recoveries obtained ranged from
about 80 percent for analytes such as acrylonitrile and methylacrylonitrile
to about 30 percent for 1,4-dioxane. The developed method employs a
standard 5-mL purge vessel with an integral, very small volume water-cooled
condenser and the same all-Tenax trap as specified in Method 5030 for use in
Method 8030. Commercially available PTD apparatus can be used without
modification except to accommodate the 90 °C water bath used to heat the
purged sample.
-------
Foreword ill
Abstract iv
Figures vii
Tables viii
Acknowledgment ix
1 Introduction 1
2 Conclusions 4
3 RBccomendations 6
4 Experimental 7
Materials 7
Sources 7
Preparation and Use of Stock and Standard Solutions 7
Method Development Activities 8
GC Testing 8
Hydrolytic Stability Testing 9
H-FTD Recovery Pre-Screening 10
Purge Recovery Enhancement Through Salting-Out 11
Effect of Purge Temperature on Analyte Recovery 12
Purged Water Removal Using a Dessicant 13
High Retention Trapping Materials 14
Breakthrough Volume Testing for Tenax and Novel
Trapping Materials 15
H-PTD GC-MS Method Testing '. 17
Instrumentation and Equipment 17
Analysis Conditions 17
Spiking and Calibration Standards 18
Sanple Preparation 19
Data Processing 19
5 Results and Discussion 21
Overview 21
Preliminary Activities 23
Information Gathering 23
Analyte Screening 24
-------
CONTENTS
(Continued)
Method Development Activities 25
Selection of GC Conditions 25
Hydrolytic Stability Testing 26
H-FTD Recovery Pre-Screening 27
Purge Recovery Enhancement Through Salting-Out 28
Effect of Purge Temperature on Analyte Recovery 30
Refinement of the Condenser Design 31
Purged Water Control Options Not Involving a Condenser 33
Breakttirough Volume Testing for Tenax and
Novel Trapping Materials 36
H-PTD GC-MS Method Testing 39
Overview 39
H-PTD GC-MS Analysis Results 40
Bibliography 43
-------
FIGURES
1 Hierarchial approach for analytical methods development
for organic RCRA analytes 44
2 Gas chromatograms of H-PID analytes on: (A) 0.53 mn I.D.
Supelcowax-10 fused silica column and (B) 0.75 mn I.D.
borosilicate glass VDOOL column 45
3 Purge vessel-condenser unit used for H-FTD GC-MS method
performance testing 46
4 GO-FED chromatogram demonstrating band splitting of
analytes with a Carbotrap:Carboseive, 4:1 trap 47
5 GC-FID Ghronatogram for septum injections of 11 H-PTD
analytes (500 ng, each analyte) 48
6 GC-FID chranatogram for H-PID analytes spiked directly
onto the DADS/PMDA trap and then desorbed with 50°C
preheat (500 ng, each analyte) 49
7 GC-FID chromatogram for H-PID analytes spiked directly
onto the DADS/PMDA trap and then desorbed with 120°C
preheat (500 ng, each analyte) 50
8 GC-FID chranatogram for H-PID analytes spiked directly
onto the DAIH/PMDA trap and then desorbed with 50°C
preheat (500 ng, each analyte) 51
9 GC-MS chranatogram for H-PTO analysis of the highest
calibration level 52
10 FID chranatogram showing typical analyte resolution
for H-PID analysis obtained during the methods
development phase 53
vn
-------
TABLES
Number Page
1 Commercial sources and purities of analytes included in the
study 54
2 Temperatures used for trap testing 55
3 GC retention data for analytes using the Supelcowax-10 and
VOCOL columns 56
4 Summary of GC separation effectiveness in terms of k' values . 60
5 Recovery results for hydrolytic stability testing 61
6 Recovery results for H-PTD analyte screening 62
7 Concentrations and ionic strengths of salts used for H-PTD
enhancement 63
8 Summary of H-PTD recovery enhancement by salt addition .... 64
9 Dependence of H-PTD recovery on purge temperature 65
10 Breakthrough characteristics of acrolein and acetonitrile
on novel trapping materials 66
11 Breakthrough characteristics of acrolein and acetonitrile
on Tenax 67
12 Analyte calibration response factors for GC-MS quantification. 68
13 H-PTD GC-MS analysis method performance for the low spiking
level 69
14 H-PTD GC-MS analysis method performance for the high spiking
level 70
15 Comparison of chromatographic peak widths for H-PTD processed
and split septum injected samples 71
vi ii
-------
ACKNOWLEDGMENT
The U.S. EPA Project Officer, Janes E. Longbottom, Environmental
Monitoring and Support laboratory, Cincinnati, is acknowledged for his
conception and direction of this research program.
Also acknowledged are the following Battelle Columbus Division staff:
Dr. J. Scott Warner who provided valuable technical guidance and
consultation, Dr. Richard A. Kbrnfeld who was the Program Manager, Roxanne
Edwards who acquired and processed the GO-MS data for method testing, and
Laraine Porter, Barbara Poison and Gayle Pakrosnis who prepared the
manuscript.
-------
SECTION 1
INTRODUCTION
The Resource Conservation and Recovery Act, RCRA, requires predisposal
monitoring of about 370 organic species listed in Appendix VIII to 40 CFR
Part 261. In response to a petition by the State of Michigan, the U.S. EPA
has proposed the amendment of RCRA Appendix VIII by the addition of over 100
other organic compounds. Using the hierarchical approach shown in Figure 1,
these analytes were evaluated for suitability for gas chromatography-mass
spectrometry (GC-MS) determination and classified as volatile or
semivolatile for eventual inclusion in the analyte sets of available
standard methods. GC-MS suitability studies were performed in Work
Assignment 4 of the current contract. Purge-trap-desorb (PTD) methods such
as EPA SW-846 Method 5030 work well for the nonpolar, relatively water-
insoluble volatile analytes, while methods based on liquid-liquid extraction
followed by extract volume reduction, as described in EPA SW-846 Method
3510, typically are effective for the nonpolar semivolatile analytes. Work
Assignments 10 and 8 of the current contract yielded results on the
suitability of Methods 5030 or 3510, respectively, for selected analytes.
Some of the Appendix VIII or Michigan List organic compounds are not
amenable to either of these approaches because they are so highly volatile
that they are partially or totally lost on solvent reduction of organic
extracts and are too water soluble to be effectively recovered by liquid-
liquid extraction or effectively purged by the typical room temperature
purge-trap-desorp (RT-PTD) approach such as that of Method 5030. Typically,
such analytes can only be analyzed by direct aqueous injection which is
inherently about 1000-fold less sensitive than PTD analysis using a 5-mL
sample. Reasonable recoveries of a few highly water-soluble analytes have
been obtained with a heated PTD (H-PTD) approach analogous to Method 8030,
which has been validated for acetonitrile, acrylonitrile, and acrolein.
The goal of this research effort is to develop an H-PTD method to be
used in conjunction with Method 5030 for the analysis of a wide range of
volatile, water-soluble Appendix VIII and Michigan petition analytes. The
target characteristics for the desired method are:
o Provide analysis capability for analytes currently
-------
restricted to direct aqueous injection methods.
o Incorporate the advantages of fused silica capillary
chromatography.
o Be readily adaptable to the RT-PTD apparatus and detection
systems currently in use.
o Obtain sensitivity comparable to existing RT-PTD methods for
the determination of less water-soluble analytes.
o Avoid, if possible, analysis steps involving two-stage, trap
to trap desorptions or special cryofocusing apparatus.
In the work reported here, all of the organic Michigan petition and
Appendix VIII analytes which (1) had not proved amenable or had displayed
poor analysis characteristics in Methods 8240 and/or 3510 and (2) were
thought to be potentially amenable to an H-PTD approach, were included in
the study. Initially, the following 13 analytes of seven functional group
types were specified by the EPA Project Officer for inclusion in this
research program:
Nitriles Ketones
Acetonitrile Methyl ethyl ketone
Acrylonitrile Bromoacetone
Propionitrile
Malononitrile Alcohols
3-Chloropropionitrile Isobutyl alcohol
Propargyl alcohol
Aldehyde
Acrolein Ether
1,4-Dioxane
Nitrogen Base Thiol
Pyridine Trichloromethanethiol
Subsequently, based on the results of other heirarchical scheme-structured
Work Assignments which had, at that time, just been completed, the EPA
Project Officer added another 20 analytes which had not qualified for
inclusion in those methods. These 20 analytes could be classified in two
groups:
(1) Analytes eliminated from Methods 8010, 8015, and 8020 validation
due to poor RT-PTD recovery:
2
-------
Method 8010 Method 8015 Method 8020
Chloral hydrate Acrylamide 2-Picoline
2-Chloroethanol Beta-Propiolactone Thiophenol
l,3-Dichloro-2-propano1 Methacrylonitrile
(2) Analytes not detected or very poorly detected using GC-MS and the
packed column specified in Method 8240 with the performance deficit
attributable to the use of an inappropriate GC column:
Nitrogen Bases
Hydrazines: Nitriles
Methylhydrazine 2-Hydroxypropionitrile
1,1-Dimethylhydrazine 2-hydroxy-2-methylpropionitrile
1,2-Dimethylhydrazi ne
Amines: Miscellaneous
1°: n-Propylamine Tetranitromethane
2°: Aziridine Methyl mercaptan
Methylaziridine 2-Butanone Peroxide
3°: N-2-Hydroxyethylaziridine
The experimental design employed for this study directed attention to:
o use of chromatographic systems optimized for polar analytes
o novel trapping materials and approaches
o effective control of water vapor exiting the heated purge
vessel
o the effect of temperature on recovery of the purge step
o the effect of dissolved salts on recovery of the purge step.
The H-PTD experimental conditions which evolved from previous investigations
(See Bibliograhy) were tested using fused silica capillary column GC-MS
analysis for the determination of all of the originally included analytes
which proved recoverable to some degree with the H-PTD conditions developed.
-------
SECTION 2
CONCLUSIONS
Some of the conclusions regarding the potential analytical performance
of an H-PTD-based analysis method for polar, volatile organic analytes are
the following:
o Eight of the 33 analytes included in the study were
determined to be amenable to H-PTD. Those eight analytes
are: acrolein, methyl ethyl ketone, methacrylonitrile,
acrylonitrile, acetonitrile, propionitrile, 1,4-dioxane and
isobutanol.
o Twenty-one of these 33 analytes tested were not amenable to
H-PTD because they are nitrogen bases (9 analytes),
hydrolytically unstable (6 analytes), mercaptans (2
analytes), or had both very low volatility and very high
water solubility (4 analytes).
o A small volume condenser was effective for controlling the
amount of water vapor exiting the heated purge vessel and
subsequently condensing in the PTD device plumbing. The use
of this condenser enabled a reliable achievement of adequate
analysis precision, i.e., less than 15 percent relative
standard deviation.
o The use of salt enhanced purge recovery of all analytes that
were not quantitatively recovered without salt. Salting out
at the levels used, 80 percent saturation at 85 C, is
difficult using conventional purge vessels and probably
promotes poorer reproducibility. Since highly soluble salts
with both mono- and di-valent cations and anions were used,
other salts are not expected to prove more effective than
those used in this study.
o H-PTD GC-MS method detection limits (MDLs) for the eight
analytes tested were 5- to 10-fold higher than could have
" been expected for an optimal case based on Method 8240
performance for nonpolar analytes. The higher than desired
MDLs result from a combination of three factors: (1)
variable and incomplete purge recovery, (2) water-broadened
chromatographic peaks and (3) MS fragmentation
characteristics resulting in lower detection response
factors.
o As expected, H-PTD recovery was predictably related to the
purge temperature: the higher the purge temperature, the
higher the recovery.
-------
The use of zeolite dessicants such as molecular sieve to
remove water from the purge stream was shown to have no
possibility for further development.
None of the novel trapping materials tested showed promise
for overcoming the technical problems of H-PTD. While Tenax
adsorbent traps have almost no reserve capacity for the most
polar and volatile analytes, they nevertheless are expected
to remain the trapping material of choice for H-FTO.
Hie five hydrazines and aziridlnes studied displayed poor
chromatographic characteristics or were not amenable to GC
on either of the GC columns tested, even when injected in
non-aqueous or solvent-free media, i.e., gaseous or neat.
These analytes were: 1,1-dimethylhydrazine, 1,2-
dimethylhydrazine, ethyleneimine, 2-methylaziridine and N-2-
hydraxyethylaziridine.
Chloral hydrate, 2-butanone peroxide, 2-hydroxypro-
pionitrile, and 2-methyl-propionitrile are thermally labile
and were found to be not amenable to GC analysis.
Beta-Propiolactone was apparently solvolyzed by both
methanol and water used as injection solvent, and could be
eluted from the GC only when injected neat.
Bromoacetone was unacceptably hydrolytically unstable with
60 to 90 percent loss over a 15-minute exposure to pH 7 and
85°C. Tetranitromethane and thicphenol were quantitatively
lost under these conditions.
l,3-Dichloro-2-propanol, which was apparently stable toward
the heated purge conditions, could not be recovered when
injected directly on the trap and then desorbed after a
normal purge sequence. Propargyl alcohol and 2-
chloroethanol could be adequately recovered from the trap
but could not be purged from an aqueous sample.
Other compounds which failed to qualify for inclusion in the
method due to nonpurgeability were: acrylamide, malo-
nonitrile, pyridine, and 2-picoline.
-------
SECTION 3
RECCflMENDATIONS
Based on the results obtained in this study, the following
ccranendations can be made:
o Further evaluation of less polar columns should be conducted
to improve the chromatographic performance of the H-PID
method developed in this study. Ihe choice of the polar,
Carbowax-type column was driven principally by extremely
polar analytes which later had to be excluded from the
method due to nonpurgeability. It is very likely that other
stationary GC phases, e.g. DB-624, might reduce the peak
broadening caused by the reverse solvent effect attributed
to desorbed water.
o Further work on the eight analytes which qualified for the
developed H-FTD method and similar analytes should be
directed toward modifying Method 8240, using the standard
Tenax/coconut charcoal trap, to enable the inclusion of
these analytes in that method. A separate method for the
eight analytes may not be cost-effective. The use of the
combination heated purge vessel-condenser unit developed in
this work should dramatically improve Method 5030
performance for partially recovered analytes such as
naphthalene, tri- and tetrachlorobenzene and
hexachloropropene without any detrimental effect on the more
volatile Method 8240 analytes.
o Salt addition and/or purge temperatures higher than 90 °C
should not be used for H-PTD analyses.
o Non-PTD method development work should be conducted for the
determination of most of the alcohols of interest. Some of
the approaches that might be successful for these types of
analytes include fractional distillation, fractional
crystallization, solid phase extraction, and development of
large volume, i.e., 50 to 250 uL, direct aqueous injection
methods
o Further method development activity for the volatile
nitrogen bases of interest (hydrazines, alkyl amines,
aziridines and pyridines) should involve solid phase
extraction using ion exchange resins, ion chromatography,
fractional crystallization, and/or large volume direct
aqueous injection.
-------
SECTION 4
EXPERIMENTAL
Materials
Sources
Commercial sources and purities of the 33 analytes originally included
in this study are shown in Table 1. Reagent water was prepared by boiling
for 15 minutes 1-L quantities of commercial distilled water which was first
processed through a Milli-Q water purification system modified by placing
activated carbon modules in the last two cartridge positions. The boiled,
cooled water was stored until use at which time it was passed through a 2.5
cm I.D. by 45 cm long column of activated carbon (Barneby and Cheney, type
PC, Lot 2547).
Preparation and Use of Stock and Standard Solutions
For the method development phase of this study, stock solutions and
analyte mixture spiking standards were prepared in aqueous solution to avoid
possible interference in FID chromatograms from any H-PTD-recovered methanol
that is usually used as a spiking solvent. In the; early stages, individual
aqueous analyte stocks were prepared at 100 mg/mL and then used to prepare
spiking and septum injection calibration standards. Initially, these stock
solutions were prepared gravimetrically using 2.0 or 10.0 mL volumetric
glassware with analyte dispensed from an appropriately sized micro!iter
syringe according to its known or determined density. In later work,
volumetric dispensing of neat analytes to prepare stock solutions was used
without checking weights because that approach had been shown to result in
gravimetric errors of less than 2 to 3 percent. Finally, the intermediate
step of preparing individual stock solutions was omitted and analyte spiking
stock mixtures were prepared directly from neat analytes. Dilutions to
working standards were performed in 10.0-, 5.0-, or 2.0-mL volumetric
glassware using appropriately sized syringes. Final concentration working
standards were prepared in 10.0-mL quantity and immediately bottled in five
7
-------
seals. These working standards were stored at - 4 °C when not in actual
use, and a previously unused vial of spiking and calibration standard was
used for each day of experimental trials. Typically, these standards were
prepared at about one-week intervals, and analyte degradation during storage
over that tiny period was never found to be a problem.
Since all of the method development work was evaluated on the basis of
absolute H-PTD recovery, the preparation of a working standard always
involved two concentration levels: (1) purged sample spiking standard of
which 5.0 uL was dispensed from a 5-uL syringe into the 5.0 mL aqueous
sample and (2) septum injection calibration standard at 2.5-fold higher
concentration than (1), above, of which 2.0 uL was injected directly onto
the GC column, again, using a 5.0-uL syringe. Thus, the 2.0 uL septum
injection corresponded to a 100 percent H-PTD recovery standard. To insure
reproducibility of the septum-injected 100 percent recovery standard, the
clean, dry 5.0-uL syringe was primed with reagent water and indexed to 0.0
uL. The 2.0 uL sample was then drawn into the syringe so that the needle
full of reagent water would act to flush all of the calibration standard
from the syringe needle.
Method Development Activities
GC Testing
Analytes were tested individually for chromatographic performance on
both a polar and nonpolar magabore capillary GC column. The GC conditions
and columns used were:
Gas Chromatograph Carlo Erba Model 4160
Detector Flame lonization
Columns 1. 30 m x 0.53 mm ID fused silica with
1 uM SupelcoWax-10 film (Supelco)
2. 60 m x 0.75 mm ID borosilicate glass
with 1.5 uM proprietary film, VDCOL
(Supelco)
Cven Program 40°C (2 min), 8°C/min, 225°C (10 min)
Injector Temperature 230°C
-------
Detector Temperature 250°C
Carrier Gas Helium at 10 ml/min, flew oontrolled
Injection Volume 1.0 uL by autosampler using 1.0 or
10 ug/uL solutions
Initially, analytes were tested in aqueous solution since trap-desorbed
samples and calibration standards in later work were expected to be
equivalent to an aqueous injection. Analytes that were not detected
satisfactorily were retested first in methanol injection solvent. If the
analytes were still not detected, pentane injection solvent was used or
analytes were injected as neat liquids or headspace vapors above neat
reagent. Except in the subsections that follow, the GC conditions shown
above and the Supelcowax-10 column were used in all subsequent testing.
Hvdrolytic Stability Testing
The following 13 analytes were tested for hydrolytic stability:
Acrolein 3-chloropropionitrile Methyl mercaptan
Acrylamide 1,1-dimethyl hydrazine Propargyl alcohol
Acrylonitrile Methylaziridine Tetranitromethane
Bronoacetone Methylhydrazine Thiophenol
N-2-Hydrc«yethylaziridine
Each analyte was tested individually in duplicate by direct septum
injection of an aqueous standard held at 85°C. The aqueous mcrfjg was 10 mM
phosphate buffer at pH 6.8. For each test, sufficient neat analyte, except
for acrylamide in 100 mg/mL methanol, was injected into a 42-mL septum
sealed serum vial filled nearly headspace free with room temperature buffer
for a final concentration of 100 ug/mL. The sample was quickly miy»j by
multiple inversion aided by glass beads, and then a 2-uL aliquot was
injected into the GC. This zero time sample functioned as the 100 percent
recovery sample. The sample was then submerged in an 85°C water bath and
sampled a second time after 15 min to determine the amount of analyte
remaining intact. The GC conditions used were isothermal at 10°C below the
elution temperature of the analyte in question. If necessary to eliminate
-------
antoiguity in GC peak identity assignment, the same isothermal conditions
were vised for analyte injections as a neat liquid, using a syringe needle
wet with analyte, or as gas, using headspace above the neat analyte.
H-PIP Recovery Pre-Screening
The compounds that performed adequately or marginally during the GC
testing and hydrolytic stability testing were divided into three analyte
sets:
Set 1 Set 2 Set 3
Methyl mercaptan Methyl ethyl ketone n-Propylamine
Methacrylonitrile Acrolein Acrylonitrile
Propionitrile Aoetonitrile 2-Methylaziridine
Propargyl alcohol 1,4-Dioxane Pyridine
Thiophenol Isobutanol 2-Picoline
Acrylamide Bromoacetone N-2-hydoxy-
Malononitrile 2-Chloroethanol ethylaziridine
3-Chloropropionitrile
1,3-Dichloro-2-propanol
Analytes were assigned to sets on the basis of GC retention time and
functional group types. Preliminary testing of H-PTD was done using the GC
conditions given above in the GC Testing subsection except that a Varian
3700 GC was used. The H-PTD conditions were:
Purge-Trap-Desorb System Tekmar LSC-2
Purge Line and Valve Temp. 130°C
Purge Temperature 85°C
Desorb Temperature 180°C
Trap Bake Temperature 210°C
Purge Time 15 min
Desorb Time 4 min
Trap Bake Time 10 min
Purge Gas Helium at 35 ml/min
Post Condenser Make-up Gas Helium at 10 mL/min
Condenser Cooling Water
Temperature 4°C
Recoveries were determined by comparing peak areas from the 250 ng
10
-------
septum injection corresponding to 100 percent PTD recovery to the
corresponding peak areas f ran the 5 mL H-FID samples which also contained
250 ng of each analyte (50 ug/L). Peak areas of two replicates of septum-
injected 100 percent recovery standards were averaged and used to quantify
the recoveries obtained for the three replicates of H-PTD analysis. The
septum-injected recovery standards were always the first and last runs of a
given set. The condenser used in this work was the first design (c.f.
Section 5. Results and Discussion. "Refinement of the Condenser Design")
which was a commercially available condenser with 1/4 inch CD by 8 cm long
cold zone modified to attach to the purge vessel outlet with a 1/4 inch
Swagelok union. The condenser cold zone was packed with 3 mm glass helices.
Rirge Recovery Enban"cimont Through Saltincr-Out
The four anhydrous salts used, sodium chloride, sodium sulfate,
magnesium chloride, and magnesium sulfate, were heated in a muffle furnace
at 450°C overnight to remove any trace organic impurities. Using Chemical
Handbook data (CRC, 65th Edition), the amount of each salt required to
produce 80 percent saturation at 85°C was computed, assuming a linear
relationship for solubility versus temperature, using the two solubility
data points closest to (above and below) 85°C. For 5 mL aqueous samples,
those salt amounts were: Nad, 1.5 gm; Na2SO4 1.5 gm; MgCl2, 2.8 gm; MgS04,
2.7 gm. Samples were prepared for analysis by: (1) the premeasured amount
of the tested salt was added to a clean, dry purge vessel, (2) the purge
vessel/condenser unit was attached to the FID and condenser water flow
lines, (3) the 5-mL aqueous sample, which had immediately before been spiked
with the analyte mixture, was added to the purge vessel, (4) the salt was
dissolved by repeated inversions of the purge vessel/condenser unit, and (5)
the purge vessel/condenser unit was installed in the 85°C water bath and the
purge flow initiated. The analytes used in this study were: acrolein,
methyl ethyl ketone, acrylonitrile, acetonitrile, propionitrile, 1,4-
dioxane, isobutanol, 2-chloroethanol, l,3-dichloro-2-propanol, 3-
cnloroprapionitrile, and cyclohexanone. These analytes were spiked at the
50 ug/L level (250 ng in a 5-mL water sample). The condenser used in this
work was the second design (c.f. Section 5. P'pgults and Disoiggjon.
"Refinement of the Condenser Design") which was an open, Vigreux-type, with
n
-------
a 6-mm ID by 3 on long cold zone operated with 20°C condenser water. Hie
purge flow was 30 ml/min for 15 min. A desorb preheat temperature of 150°C
was vised in initial experiments but changed to 50°C in later experiments due
to apparent partial decomposition on desorption of acrolein and
acrylonitrile. Otherwise the GC and H-PTO conditions were identical with
those given or referenced above in the "H-FID Recovery Pre-Screening"
subsections.
Effect of Purge Temperature on Apaiyte Recovery
The following nine analytes were used to examine H-PID recovery
variation as a function of the purge temperature: acrolein, methyl ethyl
tetone, methylacrylonitrile, acrylonitrile, aoetonitrile, propionitrile,
1,4-dioxane, iscbutanol and cyclohexanone. The purge temperatures tested
were room tenperature (22°C), 40, 60, 85 and 99°C. All experiments were
performed in duplicate except for the 60°C trial which was in triplicate.
Septum injection 100 percent recovery calibration standards were analyzed in
triplicate. Except for the 22°C trial, 500 ng of each analyte, i.e. 100
ug/L in the 5.0 mL purged sample, was the spiking level. The 22°C data set
was part of the salting out study, and, thus, the spiking level was 250 ng
(50 ug/uL) and methacrylonitril6 vas omitted from the analyte set. The
analysis conditions for the 22°C data set were identical to those given in
the preceding subsection on salting-out. The analysis conditions for the
other temperatures tested were different from those for 22°C as follows: an
OI Model 4460 PTO unit was used and the third condenser design was employed
(c.f., Section 5, Results and Discussion. "Refinement of the Condenser
Design"). This condenser was of a Vigreux design with 9 mm CD by 18 cm long
cold zone and 9 mm purge vessel outlet and condenser inlet which was
connected with a 3/8 inch Swagelok union. The trap was maintained at 25°C
during the purge step for which the flow was 40 ml/min for 15 min. No post-
condenser make-up flow was used. The PTD unit maintained the valve and
transfer lines at 90°C and employed a 50°C desorption pre-heat. The
experimental design called for a highest purge temperature trial at 100°C,
but, due to the barometric pressure on the day the experiment was performed,
the boiling temperature of the water bath used to heat the purge vessel was
99°C.
12
-------
Purged Water Rmoval Using a Desiocant
type 3A molecular sieve in 1/16 inch spherical pellets were
obtained from MCB Chemicals. The pellets were ground with a mortar and
pestal and size selected vising 14 and 30 mesh standard sieves (1.41 and
0.59 mm, respectively) . A 30 cm length of 1/4 inch stainless steel tubing
was cleaned, acid washed, packed with 3 gm of the prepared molecular sieve
and installed in the oven compartment of a Varian 1400 GC. The dryer unit
was conditioned overnight at 200°C with 40 ml/min helium. The amount, 3 gm,
of molecular sieve used was ten percent in excess of that estimated to be
necessary based on the computed quantity of water expected to exit the purge
vessel and isotherms provided in Union Carbide specification sheets on 3A
molecular sieve. The molecular sieve drying unit was then attached to the
H-PID system between the purge vessel and trap by means of an external
6-port valve and heated transfer lines, both of which were maintained at
103°C. Water breakthrough was determined by observing oondensate formation
in a Pasteur pipette attached to the dryer outlet which was maintained at
room temperature. Using a 15 minute, 40 ml/min purge with the sieve at
100°C, water breakthrough was observed at 12 minutes, i.e. , 480 roL of purge
flow. By determining the weight of water collected in the Pasteur pipette,
the following dryer operating conditions were adopted: the dryer module was
held at 90°C during the 15 minute purge followed by a 20 minute bake cycle
at 230°C and 150 ml/min helium in the same flow direction as that for the
purge cycle. The analyte throughput of the dryer module was tested by
spiking 2 uL of an aqueous standard containing 500 ng/uL of each analyte
into the inlet of the drying module and then executing a normal 85°C H-PTO
cycle. This experiment corresponded to 100 percent purge recovery with all
analytes exposed to the dryer module. The 100 percent dryer throughput
standard was similarly produced by identically spiking the outlet of the
dryer module, i.e., the transfer line leading to the trap. These
experiments, including a septum injection of the spiking standard, were
performed in duplicate. In a second experiment, the spike was applied to
the dryer module inlet at the tenth minute of the 15 min purge cycle to pre-
load the dryer module with water. The analytes used in this study were
acrolein, methyl ethyl ketone, methyacrylonitrile, acrylonitrile,
13
-------
aoetonitrile, propionitrile, 1,4-dioxane, isobutanol, cyclohexanone,
propargyl alcohol, 2-chloroethanol, l,3-dichlcro-2-propanol and acrylamide.
Except where noted otherwise, the H-PTD and GC instrumentation and
conditions for these experiments were identical to those cited in the
preceding subsection.
High Retention Trapping Materials
High retention trapping materials which might be able to retain
analytes at temperatures above the dew point of the heated purge vessel
effluent were tested. The rationale for this approach was the complete
elimination of the formation of condensate which could recapture analytes
purged from the sample.
Two trapping materials with relatively strong retention characteristics
were obtained from Supelco. (1) Carbotrap, a carbonaceous adsorbent in
common use for air sampling, for example, workspace sampling, of volatile
materials such as benzene, and (2) Carboseive, a new proprietary adsorbent
with uncoranaonly high retention characteristics (i.e. a specific breakthrough
volume for methyl chloride of - 12 1/gn.). A standard configuration Tekmar
trap was packed with 24 cm of Carbotrap and installed in the GC oven as a
substitute for the column. A splitter valve ahead of the detector was used
to reduce the flow into the detector to a reasonable amount. Trap
breakthrough for methyl ethyl ketone was tested using carrier which was a 30
mL/roin purge flow through a purge vessel containing reagent water and held
at 85°C. This carrier was diluted upstream of the trap with 150 ml/min of
helium make-up flow giving a dew point of 48.7°C. The breakthrough volume
of 10 ug of methyl ethyl ketone with the trap at 50°C was approximately 900
mL for the peak top and about 600 mL for the one set corresponding to only
150 and 100 mL, respectively, of purge vessel flow. A similar experiment
was performed with Carbosieve, and the breakthrough volume for methyl ethyl
ketone was greater than 6300 mL at 200°C, about 1900 mL at 250°C and about
800 mL at 285°C. Based on these results, a miypri bed trap containing 4:1,
w/w, of Carbotrap and Carbosieve, respectively, was prepared with the
Carbotrap at the inlet end. This trap was conditioned overnight at 300°C
with 30 mL/min of dry helium. The ability of this trap to adsorb and desorb
analytes under H-PTD conditions and enable adequate GC analysis of the
14
-------
desorbed analytes was tested using the «»"₯=» 13 analytes cited in the
preceding subsection. An Oceanographies International (01) Model 4460 FID
unit was used and the transfer line between the purge vessel and OI 4460 was
heated to 87°C. In duplicate experiments the analytes were spiked into the
transfer line between the purge vessel and trap using 2.0 uL of an aqueous
standard containing each analyte at 500 ng/uL. The GC conditions were
identical to those described in preceding subsections. The H-PTD conditions
were:
Condenser none used
Purge bath 85 °C
Trap during purge step 90 °C
OI 4460 valve and transfer line 90 °C
Purge flow 30 ml/min, helium, 15 min
Post-purge vessel make-up none
Trap desorption preheat variable, see text
Trap desorption 300 oC for 4 nun
Trap bake 300 oC, 20 min, dry helium
These trials produced cfaromatograms with split analyte peaks (see Section
5) , and desorption preheat temperatures of 90, 100, 120, 250 and 300°C were
tried to alleviate this problem. Since this attempt was not successful,
samples with analytes spiked into the 5inL of reagent water were not
analyzed.
Breakthrough Volume Testing for
Tenax and Novel
Tenax trap analyte breakthrough testing employed commercially available
all-Tenax traps (Supelco, No. 2-0295) . Identical empty traps were packed
with two types of novel trapping materials which are chemically described in
the corresponding subsection of Section 5. Results and Discussion. For
breakthrough testing, these traps were installed as a GC column in a Varian
3700 GC with an FID. The GC oven control was used to thermostat the trap
for above-ambient or carefully temperature controlled breakthrough volume
testing. The GC injector carrier supply was 40 miymin consisting of
30 mlyfaLn of purge flow from a standard 5-mL purge vessel containing reagent
15
-------
water which was held at 85°C plus 10 mL/min of dry helium added at the
condenser outlet. A condenser was used at the purge vessel outlet exactly
as for the analysis of H-PTO samples during the various method development
phases. Trap breakthrough experiments were conducted twice: preliminary
testing early in the study and a second time late in the study using precise
trap temperature control. In early work, the condenser used was the first
design, and in later work the third condenser design was employed. These
condensers are discussed in the subsection, "Refinement of the Condenser
Design" in Section 5. The temperature of the condenser cooling water used
with the earlier work was 4°C and that used for the later work was 20°C.
In the earlier work, which involved only Tenax traps, seven analytes
were tested that were thought, on the basis of boiling point and polarity,
to be potential breakthrough candidates: acetonitrile, acrylonitrile,
acrolein, methyl ethyl ketone, isobutanol, methyl mercaptan and
n-propylamine. Analytes were injected individually in the transfer line
connecting the purge vessel to PID unit just after the point where the
10 mL/min make-up flow was added. The trap temperature for these trials was
the ambient temperature, estimated to be between 19 and 24°C. The injected
sample was 5 uL of a 0.2 ug/mL aqueous standard (1.0 ug total) and the
breakthrough trials were performed in triplicate for analytes that broke
through in less than 1200 mL (30 min of purge flow) or two replicates for
analytes with breakthrough volumes exceeding 1200 mL. Between each run, the
trap was heated to 220°C until a stable baseline was once again achieved.
The breakthrough point was taken as the intercept of the extrapolated
recorder baseline and the up-slope of the triangulated breakthrough peak.
This intercept typically represented 0.1 to 1.0 percent of analyte
breakthrough.
In the later experiments, the temperature was much more carefully
controlled, and only two analytes were investigated, acrolein and
acetonitrile, which were known to be the only analytes of the set for which
trap breakthrough might be an issue. Tenax and the two novel trapping
materials described above were investigated, and the experimental approach
used was very similar to the one described above for the earlier work. The
only differences were the following:
o trap temperatures below 30°C (Tenax only) were measured to
the nearest 0.1°C using a calibrated mercury thermometer
16
-------
o the sairples were gaseous injection of headspace diluted to
achieve 10 + 2 ug of injected analyte
o samples were injected through the septum of the GC injector
o the H-PID carrier gas flow was 40 mL/min and no post-
condenser make-up dilution was used.
The tested trapping temperatures are given in Table 2.
H-PTD GC-MS Method Testing
Instrumentation and Equipment
The GC-MS system used was a Firmigan 3200 quadrupole unit interfaced to
a Carlo Erba Model 4160 GC. A Finnigan INCOG MS data system with version
5.5 software was used for data acquisition and processing. The GC column
was interfaced to the mass spectrometer through a glass jet separator for
which 50 mL/min helium make-up gas was required for ir^xlTBirf analyte
throughput across the jet. The GC column used was a 30m by 0.53 mm I.D.
fused silica capillary coated with 1.0 urn Supeloowax-10 (Supelco Co.). A
Tekmar LSC-2 FID unit equipped with an all-Tenax trap (Supelco No. 2-2095)
was used in its normal configuration with the sample transfer line connected
to the GC injector carrier inlet line. This configuration permitted septum
injection of samples directly onto the GC column. The purge vessel-
condenser unit is described in detail in a pertinent part of Section 5. The
purge vessel was heated with a magnetically-stirred 1.5-L water bath using
an ordinary stirring hot plate. Condenser cooling water was circulated from
a Haake Model FK2N refrigerator bath with Model FN circulating heater
control unit.
Analysis Conditions
The H-PID conditions used were the following:
Purge: 40 mL/min for 11 min at 90 °C
Trapping Temp: Less than 25 °C
Trap Desorption: 180 °C for 4 min with 50 °C preheat
17
-------
Trap Bake: 210 °C for 10 min, dry helium
Heated Zones: Valve and transfer lines at 100 to 120 °C
Post-Condenser Makeup: None
The chronatographic conditions were:
Carrier: Helium at 10 ml/min, flow controlled giving
a head pressure of 5.6 psi at 40 °C
Helium make-up gas: 50 ml/min, between column and jet
separator, to maintain maximum analyte
throughput across the jet.
Oven Program 40 °C for 4 min (during trap desorption)
then program to 130 °C at 8 °C/min. and
hold for 2 min.
Temp. Zones: Injector, 200 °C
Separator oven, 210 °C
Separator to MS transfer line, 150 °C
Ihe MS operating conditions were:
lonization: 70 eV, electron impact
Ream current: 200 uA, total
Ion source tuning: Adjusted to meet EPA criteria for
brcnof luorobenzene mass spectrum
Scan: 35 to 250 ami, 1 sec per scan
Preamplifier: 10~7 amp/volt
Electron multiplier; Gain sufficient for 180K to 220K area for
m/e 84 of the quantification internal
standard, benzene-D6, 250 ng
iteration
Ihe eight analytes included in the study were: acrolein, methyl
ethyl ketone, methylacrylcnitrile, acrylonitrile, acetonitrile,
propionitrile, 1,4-dioxane and isobutanol. Four surrogate standards were
also used: methyl ethyl ketone-Dg, acetonitrile-EXj, 1,4-dicxane-Dg, and
18
-------
bromofluorobenzene (BFB) . The quantification internal standard was benzene-
D6. Stocks and spiking standards of analytes were prepared in reagent
water. Acrolein and acrylonitrile aqueous stocks were always prepared from
reagent on the f^me day as calibration or sample spiking were performed to
avoid any loss of these analytes to decomposition. The other six analytes
were prepared as a mixture as long as one week prior to use and stored at
4 °C without headspace in 1.8-mL screw cap GC autosanpler vials with Teflon-
faced silicone septa seals. A new vial was used each time calibration or
spiking standards were prepared. Stock and spiking standards of the
surrogate and internal standards were prepared in methanol and stored at
-10 °C when not in use.
Surrogate and internal standards were spiked as 10-uL aliquots into
each 5.0-mL calibration sanple just prior to analysis using a 10-uL syringe.
The levels used were: benzene-D6 and HFB, 50 ug/L; methyl ethyl ketone-Dg,
60 ug/L; acetonitrile-D3, 400 ug/L; and l,4-dioxane-D3, 300 ug/L. For the
high and low level spiked samples, the deuterated analytes were spiked into
the reagent water before bottling the individual analysis replicates and
only benzene-Dg and BFB were spiked into the 5.0 mL sanple just prior to
analysis.
Sanle
The reagent water used for calibration samples and spiked replicate
samples was prepared as specified for the method development activities. To
prepare the high and low level spiked sample replicates, reagent water in a
500 mL volumetric flask was spiked with an aqueous, freshly prepared analyte
mixture and a methanolic mixture containing the three deuterated analyte
surrogate standards. After diluting to the calibration mark and thoroughly
mixing, standard screw-cap 40-mL VC& vials were filled and sealed headspace
free with Teflon-faced silicone septa. Typically, eleven such replicate
samples could be produced from the 500-mL spiked sample. Three 40-mL blank
reagent water samples were similarly bottled for use as GC-MS analysis
blanks. The samples were stored inverted on ice for analysis on the
following day. The analyte spiking levels for the low and high spiking
level replicates are given in the pertinent results tables in Section 5.
19
-------
Data Processing
All GC-MS data processing was performed using standard Finnigan
INOOS version 5.5 software. Search libraries vised the three most intense
ions in the scan range that were above 20 percent relative abundance.
Calibration response factors were generated by the data system using the
best linear fit of the data over the five calibration levels used.
Verification of proper integration of the quantification ion peaks was
performed by examining each integrated peak on the CRT screen during final
data processing. All of the calibration and replicate analysis data was
archived on 9-track magnetic tape.
20
-------
SECTION 5
RESULTS AND DISCUSSION
OVERVIEW
Based on the results of the information gathering activities, described
below, the central difficulty for H-PTD analysis concerns the problems
created by the large amount of water vapor which exit the heated purge zone
with the purge gas. For a typical 11 minute purge at 40 ml/min, standard
temperature and pressure, and a 90°C purge temperature, the amount of water
vapor exiting the purge vessel is about 0.75 g. Condensation of this
relatively large amount of water in the PTD device, especially the trap, has
been observed by other investigations to cause extremely poor
reproducibility and even prevent quantitative analysis. Three main effects
were cited: (1) promotion of unacceptable memory effects due to analytes
sequestering in water droplets in the connecting lines and 6-port valve, (2)
degradation of chronatcgraphy through the introduction of very large amounts
of water onto the GC column, and (3) loss of analyte dissolved in liquid
water exiting the outlet of the trap during the purge step. Thus, one of
the main focuses of the work reported here was concerned with minimizing the
effect from this water during the trap and desorb cycles. Three main
approaches were investigated:
o The use of a condenser at the purge vessel outlet to
return most of the evaporated water to the purge vessel
o The use of traps having substantially greater retention of
organic analytes than Tenax so that these traps could
operate at an elevated temperature during the purge and
trap step. Post-purge make-up gas was used to reduce the
purge carrier dew point and the trap temperature was
maintained above the resulting dew point, thereby
eliminating any possibility of condensation of water vapor
from the purge carrier
o The use of a Zeolite-type dessicant held at a temperature
above the dew point between the purge vessel and the trap
to adsorb water vapor from the heated purge carrier, while
permitting purged analytes to pass through without
retention.
The experimental design for the key elements of the method development
phase addressed the parameters that were expected to affect H-PTD
21
-------
performance, and those parameters were examined in a hierarchial scheme as
follows:
o Polar and nonpolar fused silica capillary GC columns were
tested using direct injection of all candidate analytes.
The column providing the best GC performance was used in
all subsequent work
o Analytes were screened for hydrolytic instability; those
which could not be detected by direct injection of a pH
7.0 buffered standard solution held at 85°C for 15 minutes
were not included in subsequent work
o Analytes which were hydrolytically stable were tested
using preliminary purging conditions similar to H-FTD
Method 8030 for acetonitrile, acrolein, and acrylonitrile.
In most cases, analytes for which no recovery could be
detected at levels above 1000 ug/L were eliminated from
further testing
o The enhancement of purge recovery by 80 percent saturation
salting-out was examined
o The effect of variation of the temperature during purging
was examined
o Alternatives not involving condensation for control of the
water evaporated from the purge vessel were examined.
The method performance parameters used to evaluate the progress of the
method development activities were the percent recoveries and analysis
precisions for the analytes. Each day that experimental trials were
performed, two or more replicates of calibration standards which represented
100 percent PTD recovery were analyzed by direct septum injection of 2.0 uL
of the same analyte mixture used, to fortify the 5-mL purged samples. In
the later case, use of a 2.5-fold dilution was made to enable a 5.0 uL
volume for spiking 5-mL samples. Typically, these 100 percent calibration
analyses contained 250 ng of each analyte which corresponded to a 50 ug/L
concentration in the 5.0 mL purged sample. Experimental trials were always
performed in at least duplicate and usually in triplicate or higher
replication with all experimental trials and calibration standards analyzed
on the gaiff* day. These method development experiments all employed flame
ionization detection (FID), so all stock and spiking standards had to be
prepared in water to avoid methanol which interfered chromatographically
22
-------
with sane of the analytes.
Based en the results obtained in these method development activities,
H-PTD conditions were adopted and tested using MS detection. The H-FTD
GC/MS testing was designed to define accuracy, precision, and MDL values for
all of the analytes for which any level of FID recovery had been obtained
during the method development phase.
PRELIMINARY ACTIVITIES
Information Gathering
Two information gathering activities were completed before the
experimental plan for H-PID method development was completed; (1) computer-
assisted search of recent scientific literature and (2) direct telephone
inquiry of scientists known or thought to have conducted related research
activities.
Scientific Literature
The Chemical Abstracts Service (CAS) on-line data base was searched for
the period 1967 to the present. The Registry file was searched using CAS
numbers and compounds names linked with a logical "and" to key words and
word stems and an initial 811 references were selectively reduced by further
search criteria limitations to 369 references for which keyword listings
and/or abstracts were used to select nine pertinent publications.
Photocopies of these nine references, listed in the Bibliography, were
obtained and reviewed.
Direct contact of Other FJesearchers
The scientists who were contacted directly by phone and queried about
their experience and expectations regarding the H-FID analysis approach
were:
Tom Bellar, EPA Cincinnati
Bob Westendorf, Tekmar
Eric Johnson, Finnigan
Dennis Gere, Hewlett Packard
Phil wylie, Hewlett Packard Avondale Applications Lab
23
-------
Bernie Barnard, 01
Neil Mosesman, Supelco
Robert Freeman, J & W Scientific
Some of the findings from these telephone discussions that were of
interest to this H-PTD method development program were:
Purge temperatures above 60 C are reported to coat the Tenax
trap with liquid water which can move as a bulk liquid phase
and apparently substantially reduce trapping efficiency of
polar analytes dissolved in the liquid water exiting the
trap outlet.
Desorbed water causes chromatographic peak broadening,
tailing, and retention time shift.
Condensers at the outlet of the purge vessel, used to reduce
the amount of water reaching the trap, result in significant
reduction in PTD recovery of the more water soluble
analytes.
A glass lined open split GC/MS interface had been
substituted for a jet separator and was effective with flow
rates greater than 7 mL/min.
Major carry-over problems had been attributed to unheated
areas in post purge vessel plumbing which provide liquid
water sites for analyte reabsorption.
Recovery problems attributed to metal tubing in PTD units
was alleviated by acid washing which appears to deactivate
the surface similar to silylation of glass surfaces.
Needle sparge was found to be effective for H-PTD foam
control with recoveries of moderately polar analytes within
1 to 2 percent of recoveries obtained with a frit purge.
A permeation dryer was being developed to strip moisture
from the purge flow before entering the trap.
Nafion permeation dryers had been observed to be totally
ineffective for polar analytes such as acetone and alcohols
since these highly water soluble analytes are removed along
with the water.
Analyte Screening
One analyte on the lists originally included by the EPA Project
24
-------
not commercially available, based on the 1985 edition of Chem Sources. In
addition, the CAS No., 594-42-3, was found to correspond to the industrial
chemical perchloromethylmercaptan, ClsCSCl, which is highly reactive with
water, rather than the listed analyte. Therefore, trichloromethylmercaptan
was dropped from the study. A second analyte, ethyleneimine, was obtained
from two commercial suppliers and, in both cases, the material received was
completely polymerized. Therefore, ethyleneimine was also dropped from the
study. During the method development phase, GC data quantification by the
internal standard method was attempted, and one of the internal standards
tried was cyclohexanone. This compound was not 100 percent purged, so it
was ruled out as an appropriate internal standard. However, it's H-PTD
behavior and elution position in the chromatogram made it desirable to add
to the set of tested analytes even though the use of internal standard
quantification was not incorporated in the remaining method development
activities. Thus, for the studies in which data were generated for
cyclohexanone, it has been included in the corresponding results tables even
though it was not one of the originally included analytes.
METHOD DEVELOPMENT ACTIVITIES
Selection of GC Conditions
Only megabore fused silica columns were tested in the selection of the
GC system to be used for this work. There were two reasons for this
limitation:
o The greater resolving power and greater elution temperature
range of capillary, open tubular columns were thought to be
of substantial value to the future usefulness and
flexibility of the method for solid/hazardous waste
matrices.
o The desire to use existing trap technology without
additional prefocusing absorptive or cryogenic traps
required a very wide bore capillary to accommodate the
minimum 10 mL/min desorption flow rate for these traps as
the flow rate for the GC column.
GC column testing was limited to only two types of stationary phases:
25
-------
(1) a highly polar phase essentially equivalent to Carbowax (30 meter by
0.53 ran I.D. fused silica with 1.0-um Supeloowax-10 coating) and (2) a
relatively nonpolar phase which is often used for Method 8240 analysis (60
meter by 0.75 ran I.D. glass with 1.5-um proprietary siloxane coating,
VOOOL) . Both of these columns were obtained from Supelco. The GC retention
characteristics for the analytes on these two columns are presented in
Table 3. The significantly greater selectivity of the polar GC column for
these polar analytes is obvious from the distribution of the analytes in k1
ranges shown in Table 4. As a more convincing demonstration of the
superiority of the polar GC column for these analytes, the Figure 2
chromatcgrams were generated using 17 analytes from Table 3 for which
reasonable GC characteristics were obtained with aqueous injection solvent.
Figure 2 compares the chromatographic performance obtained on these two
columns and clearly demonstrates the superior separating power of the polar
Carbowax column over the nonpolar siloxane column. The Figure 2
chromatograms were obtained using identical conditions for both columns
which are shown below:
(1) same gas chromatograph,
(2) 10 ml/min helium carrier,
(3) identical injected samples,
(4) identical injector cavity conditions,
(5) identical vertical scale factors for the FID signal, and
(6) identical GC oven programs (40°C for 4 minutes, then
8°C/min until final analyte elutes) .
Because the polar GC column was, as predicted, superior for these highly
polar analytes, all subsequent work was performed using the Supeloowax-10
column.
Hvdrolvtic gt-Ahi i it
Analytes which were thought potentially labile were tested for
hydrolytic stability toward neutral pH (0.01 M phosphate in reagent water,
pH 6.8) at 85°C. Although some of the analytes tested had relatively poor
cdiromatcgraphic characteristics, as indicated in Table 3, the
chromatographic characteristics were good enough for the assessment of
26
-------
analyte stability. The results for the 13 analytes selected for hydrolytic
stability testing are given in Table 5. Based on these results, six
analytes were dropped from all subsequent work: bromoacetone, N-2-
hydroxyethylaziridine, methylaziridine, methylhydrazine, tetranitromethane,
and thiophenol. Although acrolein and 3-chloropropionitrile displayed a
significant degree of hydrolytic instability, they were clearly candidates
for inclusion in subsequent method development activity. The greater than
100 percent recovery indicated for 1,1-dimethylhydrazine probably indicates
that the GC peak assigned for it is actually a decomposition product of it.
Nevertheless, it was also included in subsequent work.
H-PTD Recovery Prescreeninq
Twenty-two analytes were screened for their potential to be recovered
by H-PTD. Nine analytes not tested were ones for which chromatographic
testing results, reasonable chemical expectation, and/or hydrolytic
instability testing results indicated essentially no possibility of ultimate
inclusion in the method: 2-methyl-2-hydroxypropionitrile,
2-hydroxypropionitrile, 2-butanone peroxide, chloral hydrate, 1,2-
dimethylhydrazine, 1,1-dimethylhydrazine, methylhydrazine, Beta-
propiolactone, tetranitromethane. The 22 analytes were tested in three
groups of six to nine analytes per set with all of the nitrogen bases tested
in the same group to avoid potential chemical reactivity problems. The H-
PTD conditions used were ones thought a priori to be reasonable but not
optimized. To control the water vapor exiting the purge vessel, a small,
water-cooled condenser (4 mm ID by 10 cm long) was packed with glass helices
and attached to the purge vessel outlet using 1/4-inch Swagelok fittings.
The condenser water temperature ranged from 20 C to 5 C.
The H-PTD recovery results for the prescreening tests are tabulated in
Table 6. The recoveries obtained for some of the analytes tested were
significantly lower than those previously obtained in other research
programs. For example, on Work Assignments 3 and 5 of US EPA Contract
68-03-1760, absolute H-PTD recoveries of 40, 60 and 85 percent were obtained
for acetonitrile, acrolein and acrylonitrile, respectively. In more recent
work on Work Assignment 1 of US EPA Contract 68-03-3224, recoveries of
37+3, 56 ± 2 and 78.7 + 0.3 percent were obtained for these three analytes
27
-------
respectively. Methyl ethyl ketone was recovered at 81 + 2 percent in that
sane Work Assignment and at 92 ± 6 percent in Work Assignment 9 of Contract
68-03-1760. Also found in the Work Assignment 9 study were absolute H-PTD
recoveries for 1,4-dioxane and allyl alcohol of 15 ± 8 and 12 + 6,
respectively. Although allyl alcohol is not an analyte in the present work,
its H-PID recovery might be expected to be similar to propargyl alcohol
which is one of the Table 6 analytes.
In the earlier work, the purge vessel was immersed in the heated bath
only to the point of the liquid-headspace interface during purging so that
the upper part of the purge vessel and outlet tube functioned as an air-
cooled condenser. In the present work, the helix-packed water-cooled
condenser (first condenser design) held up so much condensate that the purge
gas bubbled through it and, apparently, a significant amount of the purged
analyte was lost to this low temperature aqueous phase. Because of these
results, the condenser design was changed to an open, Vigreux design of 6-mm
CO glass which was attached to the purge vessel outlet with a 1/4-inch
Swagelok union for subsequent tests. This and other aspects of the
evolution of the purge vessel condenser design are more fully discussed in a
later section. Based on the prescreening results, 8 of the 22 Table 3
analytes were omitted from any further method development testing:
thiophenol, malononitrile, bromoacetone and all five of the nitrogen bases
of analyte Set 3. Although both acrylamide and malononitrile seemed to
behave similarly, only malononitrile was eliminated since there was
additional experimental data indicating malononitrile is highly reactive to
one or more of the other analytes that qualified for continued inclusion in
the study.
Rjrge Recovery 'Enhancement Through Saltincr-out
In theory, H-PTD recovery enhancement should be possible through a
shift in the aqueous phase-vapor phase equilibrium constant toward the vapor
phase by using high salt levels to lower the activity coefficient of the
aqueous solvent. Salts were selected to examine this salting-out effect on
the basis of solubility, commercial availability, and mono- and divalent
options to achieve high ionic strength. A salt concentration of 80 percent
saturation at 85 °C was chosen to achieve maximum ionic displacement without
28
-------
the inevitable dissolution problems expected when attainting to work near
100 percent saturation. The salts chosen were Nad, Na2SO4, MgCl2 and
MgSO4. The quantity of salt needed for an 80 percent saturation level was
approximated assuming linearity of the solubility versus temperature
relationship for the two literature solubility values for temperatures
closest to the 85°C purge temperature. The amounts of anhydrous salt used
in each of the four cases are shown in Table 7.
Initial salt dissolution testing showed that the chlorides of sodium
and magnesium for 80 percent saturation at 85°C dissolved fairly rapidly and
completely with good possibility of reasonable reproducibility in an
analytical protocol. Ihe sulfate salts were highly exothermic on
dissolution, were subject to caking in the purge vessel, and did not always
fully dissolve. Magnesium sulfate resulted in a highly viscous liquid that
apparently facilitated the formation of a liquid plug at the purge vessel
outlet/condenser inlet through which the purge flow would percolate. This
phenomenon apparently resulted in the highest variability in recovery of all
four salts. Thus, in spite of magnesium sulfate having the highest ionic
strength of the four salts tested, it was clearly the least desirable from
the aspect of practical use.
Typical recoveries for the four salts at 80 percent saturation are
shown in Table 8. The recovery enhancement was, as predicted, the greatest
for the two most polar, water soluble analytes. The recovery of dioxane
increased from 21 percent without salt to 56 percent with sodium sulfate,
and aoetonitrile increased from 59 percent without salt to 88 percent with
magnesium chloride. Perhaps the most noteworthy result was that 2-
chloroethanol, which had been shown to be well recovered from traps when
spiked directly onto them, was still not detected in any of the salting-out
experiments. The use of salt increased the recovery of all analytes which
were not quantitatively recovered without salt. Based on the increased
difficulty of sample handling with salt, this recovery enhancement, which
appeared to bring with it a considerable deterioration in reproducibility,
was judged not to be an advantage for the method, and no further near-
saturation salting-out experiments were performed.
Effect of Puroe Temperature on Analvte Recoverv
29
-------
A fundamental assumption underlying the experimental design of the
method development plan was that the distribution ratio for polar, water
soluble analytes between the aqueous and gas phases would be uniformly
shifted toward the gas phase as the purge temperature was increased. Thus,
initially no experiments were planned to demonstrate this thermodynamic
assumption. At the EPA Project Officer's request, a purge temperature
variation study was performed near the end of the method development work
since there were reports of many labs already using 40 to 60 "C purge
temperatures to achieve more reproducible results in Method 8240. H-PID
recovery results for five purge temperatures 22, 40, 60, 85, and 100°C, are
suranarized in Table 9. In these experiments, a 15-minute purge at 40 ml/min
(total volume 0.6 L) was used (third condenser design) with the condenser
water maintained at 20°C and the trap at 23-25°C during the purge step.
Three replicates were obtained at 60°C and 2 replicates were obtained for
the other three temperatures. Because of a slightly below standard
atomospheric pressure on the day these experiments were performed, the 100 "C
trials were actually at 99°C which was the boiling temperature of the water
bath. The most noteworthy conclusions that can be made from Table 9 are the
following:
o So far as H-PTO recovery is concerned, there is no
optimal temperature between room temperature and 100°C
since the recoveries of all analytes increase with
temperature.
o The most volatile/least water soluble analytes, i.e.,
those with the highest recoveries at 22 and 40°C, were
recovered essentially quantitatively at or above 85 °C:
methyl ethyl ketone, methacrylonitrile, and
acrylonitrile.
o The least volatile/most water soluble analytes, 1,4-
dioxane, acetonitrile, and cyclohexanone, cannot be
expected to approach quantitative recovery at any
temperature using a 5-mL sample and a purge gas volume
of 600 mL.
o Although more replicates at each temperature are
required to make a final conclusion, the precision
30
-------
obtainable is in the 1 to 5 percent BSD range and is
apparently not a function of the purge bath temperature.
o Since methyl ethyl ketone, acrylcnitrile, and
methacylonitrile are known from prior experiments to be
essentially quantitatively recovered by H-PTD at 85 °C,
the Table 9 results indicate the presence of a leak in
the system between the purge vessel and trap during
purging, between the trap and head of the column during
desorption, or both. Because of the precision obtained
and rational recovery change with temperature, the leak
was apparently constant throughout the series of
experiments, and has no impact on the conclusions made
from the H-PID recovery results.
On the basis of these results, 90°C was adopted as the purge
temperature for the GC-MS method performance evaluation study. Since a
water bath is the most likely means for purge vessel thermostating, adoption
of a 90°C purge temperature should accumulate the use of a water bath in
all but the most extreme conditions.
Refinement of the Condenser Design
As noted above, the first condenser design used a small commercially
available condenser with 1/4-inch O.D. by 8-cm length cold zone which was
packed with 3-nm glass helices. The condenser was attached to the purge
vessel outlet using a 1/4-inch Swagelok union with Teflon ferrules. For
this and all other variants of purge vessel/condenser design, the unit was
immersed in the heated bath up to the beginning of the condenser cold zone.
The packed condenser retained essentially all of the condensate causing the
purge flow to percolate through the liquid and become well equilibrated with
such a large volume of cold aqueous pha.se that purged analytes were
effectively recaptured and overall H-PID recoveries were significantly
reduced.
The second condenser design was fabricated in the Battelle glass-
blowing shop and retained the 1/4-inch O.D. tubulations and Swagelok
attachment to the purge vessel outlet, but substituted a 6-ram I.D. by 3-cm
31
-------
long Vigreaux type cold zone. Improved analyte H-PTD recoveries were
obtained with this second design, but the 1/4-inch tabulation and/or
Swagelok union allowed the formation a plug of water in the connecting
tubing after about 3 to 5 minutes of purge flow. Thus, the purge flow
percolated through this liquid water at the hot-cold zone transition point,
i.e., the inlet of the condenser, creating good conditions for recapture of
purged analytes. Poor analysis precision was correlated with the severity
of this condensate plug problem. Careful acid washing of both the condenser
and purge vessel to promote clean draining of the condensate and prevention
of water plug formation was not reliably effective.
The third design involved (1) modification of the purge vessel outlet
and condenser inlet to 9-nm O.D. glass, (2) the use of a 3/8-inch Swagelok
union connector, and (3) lengthening of the Vigreux cold zone to 18 cm.
Except for some condensate hold-up in the Swagelok union, this design
eliminated the problem of condensate plugs at the condenser inlet, and, as a
result, analysis precision for all analytes and recoveries for some of the
analytes were improved.
The final purge vessel-condenser design, used only for the GC-MS
testing described below, is shown in Figure 3. The only changes from the
previous design were the reduction in length of the condenser cold zone to
10 cm and the elimination of the 3/8-inch Swagelok union by changing to a
one-piece, glass-blown design.
A clean, dry purge vessel-condenser unit was required for each analysis
to avoid carryover problems. Since thorough rinsing of the purge vessel in
place did not eliminate carryover, the problem was probably due to analyte
left in the small volume of cold water on the surfaces of the condenser. If
the condenser were also rinsed, the relatively large amount of water left
behind wetting its entire length apparently recaptured purged analytes in
the next run thereby reducing their recovery from that obtainable with a dry
purge vessel-condenser unit. Thus, in all of the trials at each stage of
the condenser design, multiple units were cycled through cleaning and drying
steps so that each analysis was done with a clean, dry purge vessel and
condenser. Although typically only the bottom 2 to 3 cm of the condenser
cold zone appeared wet with water droplets at the end of the purge step, no
experiments were performed to establish the effect of the length of the cold
zone on the analysis result. The 10-cm length used for the final
32
-------
experiments is thought to be somewhat longer than actually necessary. One
aspect of condenser operation that definitely had an effect on H-PID
recovery was the temperature of the water circulated through the condenser
cooling jacket. Temperatures between 0°C and room temperature were used and
the results from a number of trials all indicated that lower cold zone
temperatures gave lower analyte recoveries/ as would be expected.
Purged Water Control Options Not Involving a Condenser
Two noncondenser approaches to the control of the water vapor exiting
the heated purge vessel were investigated. These activities were pursued
due to the total lack of recovery in the prior experiments for three
analytes that were thought to be at least partially recoverable by H-PTD:
propargyl alcohol, 2-chloroethanol, acrylamide. All three of these analytes
had been shown to be well-recovered when spiked directly onto traps prior to
a blank heated purge cycle followed by desorption in the normal sequence.
Their nonreoovery was thought to be due to the recapture of purged analyte
by the small amount of cold condensate present in the condenser. The two
noncondenser approaches involved (1) the use of a desicant to remove water
and (2) the use of highly retentive trapping materials which could be
operated at temperatures above the dew point of the purge stream.
Water Removal Using Zeolite Desicant
Removal of water vapor from the purge flow was accomplished using a
zeolite-type molecular sieve dessicant between the purge vessel and the
trap. No post-purge purge carrier dilution was used, so the molecular sieve
(Linde 3A) was maintained at 100 "C to preclude condensation of water
anywhere in the system. Type 3-A molecular sieve was chosen since the other
types would definitely be expected to retain the analytes in the set with
smallest molecular size. Zeolite-based molecular sieve was chosen since an
easily recyclable system would be required for routine operation.
Additionally, if most of the water absorptive capacity was inside the
nominally 3-A cavities, the loss of analyte to adsorption on surface sites
would be minimized. Isotherms for water adsorption provided by T.-inA* were
used to predict the quantity of molecular sieve necessary to just adsorb all
of the purged water at a sieve temperature of 100°C. Spherical 1/16-inch 3A
33
-------
sieve was ground and sized to 14-30 mesh and 3 g were packed into a 1/4-inch
stainless tube. Using a 450-mL purge at 85°C, water breakthrough was found
to occur at sieve temperatures higher than 95°C, so the trials were
performed at 90°C to ensure essentially complete water removal. The use of
a near-stoichionetric quantity of dessicant was expected to minimize the
loss of analytes by adsorption onto the molecular sieve surface.
The ability of the molecular sieve trap to remove water but not
analytes was tested by spiking the analyte mixture into the line between the
purge vessel and the molecular sieve module (i.e., recovery) and proceeding
through the usual H-FTD sequence. This experimental design examined the
efficiency of analyte throughput for the 100 percent purge recovery
situation. A 100 percent molecular sieve throughput standard was generated
by spiking the same mixture onto the head of the trap followed by the usual
H-FTD sequence. Between runs, the molecular sieve dryer was cycled through
a 230°C, 20 minute bake step using dry helium at 150 ToL/vdn. Under these
conditions, the dryer throughput performance was totally unacceptable with
less than 1.0 percent throughput of methyl ethyl ketone, methacrylonitrile
and acrylonitrile and no throughput of the other analytes which were
acrolein, acetonitrile, propionitrile, 1,4-dioxane, isobutanol, proparagyl
alcohol, 2-chloroethanol, l,3-dichloro-2-propanol, acrylamide, and
cyclohexanone. The experiment was repeated by spiking the analytes upstream
of the dryer at the 12th minute of the 15 minute purge to preload the dryer
with water in an attempt to deactivate its analyte adsorptive capacity/ but
the throughput results were only slightly improved; less than a factor of 20
increase in the three analytes detected in the first experiment with none of
the other analytes detected.
Highly Retentive Trapping Materials
The strategy of this approach was to operate the trap at a high enough
temperature to prevent water condensation. By partially diluting the purge
flow at its exit from the heated purge zone, the trap can be maintained at a
temperature that is possibly low enough to prevent thermal degradation of
analytes but that is slightly above the dew point of the resulting purge
flow carrier. Thus, the trapping effectiveness boundary conditions were to
prevent breakthrough or chemical decomposition at a 90 °C trapping
temperature with 450 mL of total flow (no dilution) or to exhibit larger
34
-------
roan temperature breakthrough volumes of less than 13,000 mL for the most
volatile analytes (purge carrier diluted 28-fold to achieve a dew point of
18 °C). Since the very high flow required for room tenperature operation
would far exceed acceptable velocities for analyte trapping, only a 90 "C and
70°C trapping tenperature with no dilution and 1:1 dilution make-up flow,
respectively, were tested.
The two high-retention trapping materials tried were Carbosieve and
Carbotrap, both supplied by Supelco. Carbotrap was not retentive enough for
the most volatile analytes at the elevated trapping temperature. While
Carbosieve was apparently effective at preventing breakthrough at the
elevated trapping temperature, poor desorption characteristics resulted in
essentially no separation of the early eluting analytes. This result may
have been partially due to adsorption and desorption of excessive amounts of
water since excellent chromatograms could be produced with analytes spiked
onto these traps at room temperature followed by a 4-min dry purge flow and
then identically desorbed. A combination trap consisting of 4:1
Carbopack: Carbosieve with the Carbotrap at the inlet was used with the 90°C
trapping temperature and no purge carrier dilution. Good chromatograms were
obtained from this trap when analytes were spiked directly onto it followed
by the normal H-PID sequence using reagent water. However, the GC peaks for
the early eluting analytes, especially acrolein and methyl ethyl ketone,
were severely split and the mid elution range analytes were broadened due to
the significantly different desorption characteristics of these two trapping
materials. An example chromatogram is shown in Figure 4. Chromatographic
performance of the ffiv*^ trap could possibly have been improved using a two-
step desorption sequence, i.e., desorbing the Carbosieve onto the Carbotrap
followed by Carbotrap desorption onto the GC column. However, this type of
instrumentation and modification was judged outside the scope of work and
the high temperature trapping investigations were terminated without
proceeding to recovery experiments in which analytes were spiked into a
5.0-mL aqueous purge sample.
Breakthrough Volume Testing for Tenax and
Novel Trapping Materials
Traps containing activated carbon and silica gel were not considered in
this work since both of those materials are known to result in the
35
-------
adsorption and desorption of significantly greater amounts of water than
Tenax. In agreement with earlier work, initial tests for Tenax breakthrough
problems for the most poorly retained analytes indicated that, with only one
exception, breakthrough would not occur for traps held at room temperature
and total purge flow volumes of less than 450 mL. The one exception was
methyl mercaptan which broke through Tenax in only 200 mL of purge gas
volume. Minor to significant degrees of trap breakthrough was observed to
occur for acrolein and acetonitrile in a series of experiments which tested
the effect on H-PTD recovery of varying the temperature of the condenser
cooling water. In those experiments, the trap was held at the same
temperature as the condenser cooling water and breakthrough of acetonitrile
and acrolein was clearly indicated by the recovery results for 40 °C and
possibly indicated as near onset at 30 to 33 °C. Because of these results,
it was decided to (1) test some novel trapping materials prepared at
Battelle under an earlier research program for the US EPA and (2) more
carefully determine the temperature dependency of acetonitrile and acrolein
trap breakthrough for Tenax.
Testing of Novel Trapping Materials
Two traps were packed with Battelle-synthesized trapping materials.
Standard traps as supplied by Supelco for the Tekmar FID apparatus were
packed in the usual 24-cm bed length. These materials were copolymer
polyamides of pyromellitic dianhydride (FMDA) with 4,4'-
diaminodiphenylsulfone (DADS), and 4,4'-diaminodiphenylinethane (DMM). The
/
resulting polyamides are thus designated DADS/FMDA and DAEH/HCA. A third
copolymer with 2, e-dicJiloro-p-phenyleriediamine (DCPDA), designated
DCH&/RG&, had been shown in Work Assignment 1-05 of the present contract
to be equal to DAEM/M& jj, retention of vinyl chloride, so it was not
tested.
Breakthrough volumes for the two analytes of highest interest, acrolein
and acetonitrile, were determined over a temperature range high enough to
give reasonable elution, i.e., breakthrough, times. The resultant data were
then reduced using linear plots of log retention vs. 1/T(°K) which enabled
extrapolation to 25 °C for the predicted elution volume. This
experimentation was essentially equivalent to using the trap as a GC column.
Because of the asymmetry of the eluted peak, the elution volume at the peak
36
-------
top corresponded to approximately a 25 percent breakthrough fraction. These
"peak height" elution volumes were then corrected to give a breakthrough
onset volume. The results are given in Table 10.
Since DADS/TMD& and DAEH/TMDA were clearly retentive of these analytes,
further testing was performed to check the desorption efficiency of 11
analytes which had been shown to desorb effectively from Tenax: acrolein,
methyl ethyl ketone, inethacrylonitrile, acrylonitrile, acetonitrile,
propionitrile, 1,4-dioxane, isobutanol, cyclohexane, proparagyl alcohol, and
2-chloroethanol. A mixture containing 500 ng of each analyte in 2 uL of
water was applied in duplicate experiments directly onto the inlet of each
trap and then the usual H-PTD sequence was initiated to create a trap-desorb
experiment corresponding to 100 percent purging efficiency. Each trap was
evaluated in duplicate. Ihe resulting chromatograms showed very poor
separation due apparently to poor desorption characteristics of these
strongly polar trapping materials. Ihe use of a 120°C preheat rather than
50 "C, which should have provided narrower bands of desorbed analyte,
afforded no improvement in chromatographic quality. The chromatograms
illustrating these results are given in Figure 5, which is a septum-injected
reference chromatogram obtained from a septum injection; Figure 6, obtained
from desorption from DADS/HCA with 50°C preheat; Figure 7, obtained from
desorption from DADS/RCA desorption with 120'C preheat; and Figure 8,
obtained from desorption from DMH/BCA with 50°C preheat. For the desorbed
samples, chromatographic separation was so poor that quantification versus
the septum-injected 100 percent recovery standards was judged not to be a
worthwhile activity. Furthermore, the results found for the two-component
trap of Carbosive and Carbopack, described in the preceding section,
indicated that using Tenax and one of these polar trapping materials would
probably result in split GC peaks for the early eluting components, as
illustrated in Figure 4, which is caused by the time lag for desorbing the
more retentive back-up section of the trap. Therefore, it was decided to
continue to use the all-Tenax trap and to choose purge flow and purge time
conditions that would accommodate the less than optimal retention
characteristics of Tenax.
Acrolein and Acetonitrile Breakthrough Characteristics for Tenax
In similar fashion to the breakthrough experiments described above for
37
-------
novel trapping materials, breakthrough volumes for acrolein and acetonitrile
from Tenax were determined, and the results are shown in Table 11. At the
trap temperature proposed for the GC-MS method testing, 25°C, the average
breakthrough onset for acetonitrile and acrolein was about 510 and 525 mL,
respectively, providing a 13 and 17 percent margin, respectively, against
breakthrough for a 450 mL purge volume. This margin against breakthrough is
considered adequate because the elution volume corresponding to the top of
the peak for acetonitrile and acrolein at 25 "C are essentially equal at 590
and 600 mL, respectively. This latter purge volume corresponds to
accidental overpurging by 4.5 min (i.e., 30 percent) at 30 ml/min.
Alternatively, the linear plot of log (retention) vs. 1/TCK) can be used to
predict the same breakthrough situation at 450 mL of purge flow if the trap
temperature were accidently at 29°C instead of 25°C. Using these
hypothetical analysis procedure errors and the observed breakthrough profile
asymmetries, 35 percent and 15 percent of the acetonitrile and acrolein,
respectively, that exited the purge vessel at a purge time of 0.0 min would
be lost to breakthrough. Since breakthrough onset precedes the peak top by
3 min or less, an average of 18 and 8 percent of the acetonitrile and
acrolein, respectively, exiting the purge vessel during the first 3 min of
purging would potentially be lost to breakthrough under these conditions.
Assuming that 40 percent of the total analyte purged is recovered in the
first 20 percent, i.e., 3 min, of the purge time, the amount of analyte lost
in each case can be estimated to be:
Acetonitrile: 18 percent of 40 percent = 7 percent
Acrolein: 8 percent of 40 percent = 3 percent
Thus, a reasonable margin of error of 5 percent on purge flow rate/purge
time (the example above is a 33 percent error) which would require purge
flow to be no higher than 31.5 ml/min or the purge time to be no greater
than 15 min and 45 sec, would not result in any breakthrough loss of these
two analytes. With regard to trap temperature, a 2°C error, i.e., the trap
at 27°C during the first 3 min of purging, can be expected to result in
breakthrough losses for these two analytes of less than 1 and 2 percent for
acrolein and acetonitrile, respectively. Thus, specification in the H-PID
method for use of the standard all-Tenax trap at a temperature less than or
38
-------
equal to 25.0°C (77°F). A purge flew of 30.0 + 0.5 M/min and a purge time
of 15 min + 10 sec is adequate for ensuring good trapping of the two
analytes with greatest potential for breakthrough.
H-PID GC-MS METHOD TESTING
Overview
Using the H-PID conditions resulting from the method development
activities, only the following eight analytes of the original set could be
recovered to some extent: acrolein, methyl ethyl ketone, methacrylonitrile,
acrylonitrile, acetonitrile, propionitrile, 1,4-dioxane and isobutanol. The
final H-PID conditions are presented in detail in the corresponding portion
of Section 4, Experimental. The experimental design for the H-PID GC-MS
testing involved the following elements and features:
o Preliminary testing to approximate H-PID GC-MS detection
limits and, thereby, plan spiking levels
o Establishment of GC-MS calibration curves for each analyte
from a level 3- to 5-fold above the estimated MDL to a 100-
fold higher level in half orders of magnitude increments,
i.e., IX, 3X, 10X, 30X and 100X calibration levels
o Analysis of 10 replicates of a reagent water sample spiked
at a level approximately 5-fold above the estimated MDL (low
spiking level)
o Analysis of 10 replicates of a reagent water sample spiked
10-fold higher than the low level spike
o Computation of recoveries and analysis precision from the
quantified replicate analyses, and computation of an
estimated MDL value for each analyte from the low spiking
level analysis precision
H-PID GC-MS Analysis Results
Calibration Results
Calibration standards were analyzed in duplicate at the 3X, 10X, 30X and
39
-------
100X levels and in quadruplicate at the IX level. The lowest calibration
level used for each analyte (the IX level) is given in Table 12 along with
quantification response factors and the percent deviation of the calibration
curve from the best straight line fit. The response factors and linear
model fit results are given for the case of inclusion of all five
calibration levels as well as that for which the lowest level (IX level) ,
which had the poorest reproducibility, was omitted.
The lowest response factors found were for acrolein and 1,4-dioxane.
The base peak and major fragment ions for both of these analytes are in the
m/e 27 to 30 range, partially explaining their low response factors. MS
detection sensitivity in the best case, for acrylonitrile, was still only 28
percent of the MS sensitivity for the quantification internal standard,
benzene-Dg. This unfortunately low inherent MS sensitivity is a combination
of relatively poor MS ionization and detection characteristics as well as
incomplete H-FID recovery.
Analsis Results for Siked Reagent
The spiked reagent water analysis results for individual replicates
plus the average percent recoveries and precisions are shown in Table 13 for
the low spiking level and in Table 14 for the high spiking level. Table 13
also shows MDL values conputed from the precisions found for the low spiking
level. Without exception, the analysis precisions are very good, typically
about 5 percent relative standard deviation for the high spiking level and
about 7 percent relative standard deviation for the low spiking level.
Average recoveries at both spiking levels are typically between 85 and 100
percent for all analytes at both spiking levels, the only exception being
acrylonitrile with an average recovery of 66 percent at the high spiking
level. Since the three high spiking level calibration check standards also
gave an average recovery of 67 percent for acrylonitrile and since the low
spiking level gave an 83 + 4 percent recovery for acrylonitrile, the low
recovery for the high spiking level is probably due to an improperly made
acrylonitrile stock solution for either the calibration data set or the
spiked reagent water/calibration check standard data set.
The MDL values shown in Table 13 are 10- to 30-fold higher than those
reported for the most well-behaved nonpolar analytes of Method 8240.
40
-------
Certainly, part of this sensitivity difference is attributable to less than
quantitative H-PID recovery for these polar analytes as well as inherently
less sensitive MS detection due to lower ionization cross sections and
unfavorable fragmentation pathways. The major cause of the high MDlB,
however, is the excessive width of the GC peaks for the analytes. A GC-MS
chrcmatogram for the highest level H-PID calibration standard is shown in
Figure 9. The most significant feature of this chromatogram is the
substantial difference in peak widths for the group of analytes as compared
with that for EFB. Water was found to elute in the range of 5.8 to 8.8
minutes as a very heavily fronted peak under the chromatographic conditions
used. Thus, all of the method analytes elute ahead of the desorbed water
which is, in effect, the injection solvent, and the BFB elutes significantly
after the water "solvent", dearly, what Figure 9 indicates is a reverse
solvent effect chrcmatographic band broadening of these pre-solvent analytes
eluting ahead of the "solvent": the peak widths at half height in Figure 9
are about 15 sec for isobutanol and about 4 sec for BFB. For ccnparison to
the GC-MS data, Figure 10 shows a typical FID chrcnBtogram obtained during
the method development phase. The isobutanol peak width at half height for
this FID chromatogram is about 10 sec. Thus, reverse solvent effect band
broadening was occuring in both cases. The more severe band broadening of
Figure 9 for GC-MS may be due to a higher amount of desorbed water in that
case. The FID chromatogram was from H-FTD conditions which employed a lower
temperature condenser and a 20 percent past-condenser dilution of the purge
carrier, and these conditions may have resulted in less adsorption and
desorption of purged water.
To further ellucidate this band broadening phenomenon, septum
injections were made on the GC-MS system using a 10:1 injector split ratio.
The peak widths obtained for the analytes with H-PID sample processing and
the split septum injection are shown in Table 15. The injector cavity of
the GC-MS was about 1.0 mL. Assuming a 10 ml/min column flow, and a 2-uL
sample vaporizing to 2.0 mL volume, the time-width of the sample band
entering the column would be about 1.2 sec. Adjusting for injector cavity
dilution and non instantaneous evaporation kinetics but neglecting the
column head pressure sample compression factor, a ingx-intm sample band width
of about 1.5 sec can be assumed. Table 15 shows for acrolein and methyl
ethyl ketone the band broadening expected for the normal chromatographic
41
-------
performance on the type of column used. Progressively later eluting
analytes are progressively broadened until, for isobutanol, which elutes
during the water elution, the band widths are equal to the H-PID case. This
behavior clearly indicates that only 0.1 uL of injected water can have a
serious bard-broadening effect using this chronatographis system. The
conclusion is that a nonpolar chronatographic system such as DB-624 that is
less vulnerable to polar solvent band broadening might provide better
chromatographic performance and, consequently, lower MDIs for these
analytes.
42
-------
BIBLIOGRAPHY
Baddings, H. T., deJocng, C., Dapper, R. P. M., "Autanatic System for
Rapid Analysis of Volatile Compounds by Purge-and-Cold-
Trapping/Capillary Gas Chrooaatography," HRC & PC 8,755-763 (1985)
Bertsch, W., Anderson, E., Holzer, G., "Trace Analysis of Organic Volatiles
in Water by Gas Qironatography-Mass Spectrcmetry with Glass Capillary
Columns, "J. Ghronatoar. 112. 701-718 (1975)
Blanchard, R. D., Hardy, J. K., "Continuous Monitoring Device for the
Collection of 23 Volatile Organic Priority Pollutants," Apal Chan. 58,
1529-1532 (1986)
Dreisch, F. A., Munson, T. 0., "Purge-and-Trap Analysis Using Fused Silica
Capillary Column GC/MS," J Chromatccir Sci, 21, 111-118 (1983)
Ramstad, T., Nestrick, T. J., "Determination of Polar Volatiles in Water
by Volatile Organics Analysis," Water Res. 15, 375-381 (1981)
Ramstad, T., Nestrick, T. J., Peters, T. L., "Application of the Purge-
and-Trap Technique," Am lab, 13, 65-73 (1981)
Ramdtad, T., Nicholson, L. W., "Determination of Sub-Parts-per-Billion
Levels of Acrylonitrile in Aqueous Solution," Anal Chan. 541. 1191-1196
(1982)
Trussell, A. R., Moncur, J. G., Lieu, F., Leong, L. Y. C., "Simultaneous
Analysis of All Five Organic Priority Pollutant Fractions," HRC & CC,
4, 156-163 (1981)
Warner, J. M., Beasley, R. K., "Purge and Trap Chronatographic Method for
the Determination of Acrylonitrile, Chlorobenzene, 1,2-Dichloro-
ethane, and Ethylbenzene in Aqueous Samples," Apal Cham. 56, 1953-1956
(1984)
43
-------
(LI
s.
o
a>
a
XJ
o
re
u
_c co
o 01
Q. (O
Q. C
(O Qi
(J O
o u
(O C
s- re
-------
0)
c
o
_ Ol
3
2
s
-C
QJ -^-
r J
o
c
19
TIIIlp"Tlr~T|Ir
B. 0 2.4 4.0
B
7.2
o
o
1
t
o
u
^
Ol
.5 *
c o
O 1-
a o 10
.- o.
i. 0. O
>> I t-
X CVJ O.
*
J
01
^~
u
«-*
\
o
£
* - Artifact ?
E
L.
U
<£
1
I
at
e g
° s.
JS
u
CM 01
s-
o
o.
s
a.
£
O
^
1
K . 1
o
1.
1
CM
1
O
O
.c
u
a
f*>
^
^^^
1 , , i i | i i i , | i i i i | , i r | . i i r | r-r-r-
8.8 12. B 14.4 18.8 10.2 21.6 2
01
C <
3 i -
s z
- Artifact
fT T-'^T-'-}^]f")r~r [ir
B.B l.S 3.6
. e
Figure 2. Gas Chromatograms of H-PTD Analytes on (A) 0.53 mm ID
Supelcowax-10 Fused Silica Column and (B) 0.75 mm ID
Borosilicate Glass VOCOL column. See text for GC
conditions.
45
-------
Vigreux Condenser
10 cm Made From 8-mm OD
Glass Tubing
Temperature-Controlled
Condenser Cooling
Water
4_evel of Immersion
in Heated Water Bath
Inlet Port Shortened
1.7 cm to Completely Submerge
in Heated Water Bath
Glass-Blown
3/8-inch OD
Glass Tubing
Connection
Standard 5-mL
Purge Vessel
(Supelco No. 2-2424)
Figure 3. Purge vessel - Condenser Unit Used for H-PTD
GC/MS Method Performance Testing.
46
-------
CL>
C
o
X
O)
_c
o
(J
o'
I I I I
0.0
2.4
4.3
7.2
9.6
12.0
Figure 4. GC-FID Chromatogram Demonstrating Band Splitting of
Analytes With a Carbotrap: Carbosieve, 4:1, Trap.
47
-------
61
*- in'
in
en
o
MOTE
i/i
CD
4-J
cu
a
vx
re
CU
C.
CD
i
C
I
cu
r
O
s_
(J
^
1 t
cu
c
o
cu
_C
0)
r
>1
f"
4^
CU
2:
CM
o
r*-
S-
4J
0
>fc
s.
u
re
.£!
O)
s:
ro
cu
4->
f~
c
0
*>,
s_
u
^
f
C
0
4->
CU
(J
«c
IT)
cu
El
1
c:
o
5.
o
s_
Q.
vo
cu
c
re
X
o
f~^
i
»
i i
i-»
"o
re
4^
3
-C
0
I/)
1 t
CO
anone
X
cu
f
o
"u
>J
0
CTi
1 Alcohol
>^
CD
S-
rt3
C.
O
s_
Q.
O
o
S-
o
-C
o
1
CM
» t
O
in
(Li
(J
(O
O)
o
o
IT)
01
0)
«3
D--
I
c
o
(J
Ol
co
LU
o
o
I I I I I'TTT TTT-TTTI
J TTT-TJT
|"T TTT]'! I I I ] I TTT ym I I
If)
O
cn
O
o
5-3
o
(LI
3
CD
d
48
-------
a>
-a
ro
O)
D_
s-
o
OJ
3
C7I
QJ
O>
OO
o
o
i i i rTT~rn~r"p"i" n-j-rm~]
n T T
M i i j~n i [|T"t~T"ry
i- ro
f- O)
Q JC
0)
o s-
O> Q.
i- <_}
CLO
00 O
in
in
ro
C TD
«Z O)
O S-
r O
O. oo
i QJ
= Q
S- C
O OJ v
**- JZ 0)
i^S
S- C ro
cn ro c
O ro
*-> c.
ro ro JZ
E S- (J
O r TJ
i- O)
-C <
o o «
21 O^
Q Q- C
U- OO O
I Q O
t-5 «=£ LD
CJ O
IO
OJ
s_
- 69
49
-------
ta
!- in
in
- o5
(D
T3
OJ
0.
s_
o
O)
s-
CT1
d)
OJ
CO
a
C9
OJ
c
o
U 4->
QJ OJ
S- (II
i- -C
o oi
s-
a Q-
a>
^t <_>
--o
C.O
CO CSJ
nj
c -a
<: a)
a
i
a.
in
UJ
s_ c
O 0) .
03 -O ,
S- C (O
cn m c
o ro
+-> a.
tv ns ^
s s- <_>
C 1 fO
S- O)
J= eC
(_} Q -
a Si c
U_ CO O
I Q O
O <£ LO
50
-------
- in
in
CTl
QJ
o
l+-
LO
o
O
I 1 1 I t~TT
1 I I IT I T1 7
1~T i i i I i i i i I I i I i I I 4 i i
in
ta
O
-t->
c
o
0
Ol 4->
i- (0
r- Q;
Q JC
CD
T3 S-
tt) Q.
^i
f- (_)
0.0
t/5 O
LO
00
Q) .C
51
-------
. _
6u OSZ 'auezuaqcuonuouicua
6u 0002
6u 0001
6u 0001 '
S> CM
t^ mm
OO CO
o
o>
OO
in co
o
s_
QJ
- U3
re
u
oo
o o o
O .. .^-
OJ
O)
c;
o c-j
. CD «
.§
00
o>
fO
c
Q
o.
(O r
S- O)
en >
o o>
-i-J I
re
E c
o o
s- -
-C -l->
o re
s_
O)
i-
r~
o
o
o
CJ
H-i
at.
-------
1
J
J
CD
'o
S-
u
«=C
s
o
01
/L
QJ
250 ng, each analyte
0)
0)
e
o
ra
X
o
o
,
4.0 S.0
Retention Time, min
3.0
10.0
Figure 10. FID Chromatogram Showing Typical Analyte Resolution For
H-PTD Analysis Obtained During the Method development
Phase.
53
-------
TABIE 1. COMMERCIAL SOURCES AND PURITIES OF ANALXTES INCDUDED IN THE STUDY
Compound
CAS No.
Source
Purity lot No.
2-Hydroxy-2-methylpropionitrile
Aoetonitrile
Acrolein
Acrylamide
Acrylonitrile
N-2-Hydroxethylaziridine
Bromoacetone
2-Butanone peroxide
Chloral hydrate
2-Chloroethanol
3-Chloropopionitrile
1, 3-Dichloro-2-propanol
1,1-Dimethyl hydrazine
1,2-Dimethyl hydrazine
hydrociiloride
1,4-Dioxane
Ethyleneinmine
2-Hydroxypropionitrile
Isobutanol
Malononitrile
Methacrylonitrile
2-Methylaziridine
Methyl ethyl ketone
Methyl hydrazine
Methyl mercaptan
2-Picoline
Propargyl alcohol
/3-Propiolactone
Propionitrile
n-Propylamine
Pyridine
Tetranitromethane
Thiophenol
Trichlorcraethanethiol
75-86-5
75-05-8
107-02-8
79-06-1
107-13-1
1072-52-2
598-31-2
1338-23-4
302-17-0
107-07-3
542-76-7
96-23-1
540-73-8
57-14-7
123-91-1
151-56-4
78-97-7
78-82-1
109-77-3
126-98-7
75-55-8
78-93-3
60-34-4
74-93-1
109-06-8
107-19-7
57-57-8
107-12-0
107-10-8
7291-22-7
509-14-8
108-98-5
594-42-3
Aldrich
B & J
Aldrich
Rebottled
Aldrich
Aldrich
Chan Service
Aldrich
Fisher
Eastman
Fluka
Aldrich
Chem Serv
Chemserv
B&J 087
Chemserv
Fluka
Aldrich
Aldrich
Aldrich
Fluka
B&J 247
Aldrich
Matheson
Fluka
Aldrich
Aldrich
Aldrich
Aldrich
Baker
Chem Serv
Aldrich
(b)
0.99
0.99
-
-
0.97
50 wt %
-
0.97
0.98
-
-
-
(a)
0.90
0.99
0.99
-
0.98
-
0.98
0.98
0.97
0.99
0.99
0.98
0.99
-
0.97
01321BP
AK526
4811BL
2020HK
01021PL
2297B
0209CL
792836
131
226860
0819CM
4155J
AK688
2108BK
1204AL
226923
AJ635
JL092597
20043882
2821PL
0347AM
KE032097
MK4428
644354
7968B
120907
(a) Ethyleneimine was polymerized when recieved and was therefore not used.
(b) The CAS number given by EPA is for perchlorcroethylmercaptan CC13SC1 which is
the precursor for trichloromethanethiol.
54
-------
TABLE 2. TEMPERATURES USED FOR TRAP TESTING
Trap Temperatures, °C, Used for Testing
Given Trap Material
Novel Material No.l
Analyte
Acrolein
Tenax
20.6
25.1
24.9
30.1
30.0
DACH/BmA
44
66
92
DADS/FMDA
70
72
91
121
Aoetonitrile
20.8
24.8
25.0
29.9
30.0
44
66
92
120
91
100
123
150
55
-------
Q£
O
U.
CO
o
1 1
r-
CO
-H CO
Qi Z
UJ £
r- 3
O 1
< O
O£ (_d
^£
z i
0 O
o
>- o
3: >
Q.
^C r"**
fy ^E
0
Z CO
^£
UJ
< 3:
a uj
z.
Z. i i
O CO
" ' H3
p
z co
UJ UJ
1 1
UJ >-
cc. _i
+J
CO
r-
s-
0)
u
re
s_
re
c*
o
>^
f
0.
re
S-
cn
o
re
E
O
S.
JZ
0
c
o
CO
fj
1=
1
0
o
1
o
o
0
0
1 1
1
X
ro
$
O
u
a.
3
co
_ ^
- re
c
f- C
E E
3
» r
CO O
E 0
^~ c
CO
c >
o -
f- UJ
c en
CO C
4-> -1-
CO CO
ex: =>
i
O
0
o
o
1 1
1
X
ra
2
O
U
f
CO
a.
3
CO
TJ
0
a.
E
O
O
en
c
>^»
f
r*
ra
r
ro
^»
4j>
C
ro
4_>
TJ CO
0 JO
O 3
UJ co
Cn
c
^
r
r
re
+J
p
ro
^
4^
C
re
+J
"o ja
0 3
UJ CO
O> i~H
CM tO
^f UD
r^ o
^H IO
r 1 r 1
C
re
a. co
re c
o -
S- E
CU re
E i
r Q.
>> o
J= S-
-t-> Q-
CO i
z: c
o
o
o
UJ
o
0
CO
0*
vo
(H
00
f 1
CO
c
o
CO
f
^1
JC
CO
r^
>>
+>
CO
s.
c
O I TJ
i- 01-
+> O- re
(J4-> E-0
CO re 0 C
r^ CO O re
C C CO *->
r- TJ CO
CO
E > co co
0 Or- 3
S- J3 JO O
4- ra re CO
J3 3
co co o cr
r O S- re
jQ re a.
ro a. c
O.TJ CO
CO ra 4-> C
O CO >> O
0 J=r -
re ro +->
C **- C -r-
Z3 O ro 00
*
CO i
c +> co o
o >>co c
i- 1 -re
+j re u-> jc
u c +>
CO re r CO
i5 rv g
C -4->
i- ro 0) en
CO > C
E C -i- -r-
O en oo
S- O) 3
(^ ^ 00
O TJ CO
>>J3 t- r-
r- re re J3
C TJ ro
O CO c -i-
U ro S.
CO re +J ro
r- O. CO >
ja oo
ro TJ CO r
)-> re 3 QC
O. Q) O
d) _C rjj .-
U 3 C
0 M- CT-r-
< 0 ro E
*
^"
00
m
00
i 1
a>
c
^
N
(O
&-
-o
.C
1
>^
r*
4-*
|
a
fH
«
« 1
TJ
O
O
UJ
§
o
UJ
at
o
^*
CO
o
CM
C
CO
'o
5-
o
TJ
O
0
C3
§
0
U!
00
CM
f^
^f
CO
CO
CO
1
f
L.
4^
^
C
O
I
>^
^
0
m
r_
>^
.jC
^J
re O C
re O "4- -r
- S- i
cn co J3 cn -i- s_
cn c r c re co
C 'r- JO (4>* -r- 4»> 4^
i- +J re O -t-> -t-> ra
i CC4-> 3 JC f » 4->
r- O O O. Q.I cn ro C
re S- -i- 0)OO).i-3OCO
4^ (4_ 4.1 ^) 4.) | ^_ (/) 4J> ^
O U 4-> CO 3 e
4-> 4J>COrOCCO CCOO
JZ J='i-)COO " 3 3 00
cn cn c 3 a. co TJ
r~ r- *r> J^ r^ * CZ
< r >> ro TJ JD CD >> O
co oo cn i coc reci -f-
C J3O.ro -t-J -r- JD 4->
ro O.4-> -r- O
O O O OO O O CO UOCO-F-)
O O O OO S- - S- (J S- O C
uj ui ui CD i i Q- c a.
CO
c
r-
TJ
CO CO v
Q) r 1 1.
I !- -1- CO .1-
r- J_ S_ c N i
S. 4-> 4-> rt» re O
+J !-!- X I C
i- C C O >> ro
C O O - J= 4-»
0 r- - Q 4J 3
4-> >» Q. i CO JO
CO S- O «* 2* O
O O J- i co
< < a. I-H CM "
56
-------
,*>^.
^^
CO
3
C
^j
C
o
u
CO
LU
CO
1
- o
^D
4- >
0
c/)
"f
4-*
C/)
^
J-
co
u
ro
re
^
jc
o.
re
s-
CT*
O
re
Q
J_
jC
0 O
F 1
C 1
O X
re
2 Z
** O
C U
co r
§CO
Q.
O 3
i
o
o
CO >
£ o
C re i
O -^ X
i- C ro
4J 3t
c E 0
0) O
CO O
0£ QL
o
C
3
O
Q.
E
0
0
^
Q
O
C3
§
0
C3
en
en
o
i <
0
(O
CO
c
*^>
o
r-
s-
^J
Q.
CO C 00
00 0 >» 3
+-> re O »- ^ O
0. 4-> 4-> -O CO
CO I/I O re 3
i TO c coco co..ocr
^3 ro O ^~ "'"J ^" CJi*^^ ro
rOCO-i- .O C 4-> O) C 00 >>
4->f4-> ro-i-f > 3 S- S-
3 O -t-> O) COOOCOO
^ECO 3 >, - 00 O (1) > <4-
r- 0 -> ^ C i O 3
4->(4--i- S- COrOCOro
4-> +Ji-.«- i 4-> e CO
rococo 4-> O CO .O 03 .O Q.
4-> 3 re(4-i recrOC
00 >s O ^ 4-> O) O 4^ O> O
j^ i co ^co.rocn a. c !- ex e -r-
roros 4-> -D +-> C 0) - +J o -i- +J
O re *-> U -i ro O *"> ro O '<^
O O S- OOCCOUre CS-CCS-C
^E 4^ O r ^ ^E ro E ^C 4^ r!3 ^^ *^ Z3 ^^ >fB
c co
>> i O U I
r CO I E O - (O "O CJ
CC^O 4-> -4-> Q. QJ i.
OrO3(J U oo S- 4-> oo
40 00 OJ COCO O-I-COO3
COC -O r -> r03i.CO
i 0) 00 ^i C COCT3 CO
.a ex-!- - s- ro -i- -c co -» > 3
re C7>co 4-> s- T3 O) C 4-> 3>) -C i- ra
O.CS.-1-rO XI C 4->4->COO
CO ro CO f S -t fO -r-3 Q.O ^
U.C-f- S- » ^2 E ^ 4->
O i +J aj C 4-> $- CO -i-
rOO 4-> -r- 4-> O CO «-4-> O S
CJ= rOH- i C 4-> 00
>>reui C XI O S- -O XJ
^.c-r-reo j^ co -C034->4-> O) >, O Q. U U -i- (_) -O
CO £ C -i- T3 ca., jCtt)tt)(U4->COC-O T3
CO i_reco O rO4->(J i->-<-5 U 4-» ra O O
OCO4->O 0 OCCO UCC3C04-> O O
Q_ -r- i+_ oo £3. OZrOE<-'--'-'O'OtD O
-~. in vo vo
o c c 4-> i co
>> i- O CO >> O
JC i C O O> S-
t O re re S- O
>> o 4-> o re ,_
J= i- 3 E Q. JC
4-> O- CO O O O
CO tit. S- i
Z: CM CM CO Q. CM
57
-------
^^*
Jp
»
c
*^
yT
C
u
^"^
^T
CO
,
OO
k^-.
^.H^
_Q
f»
E
f
O
O
c
O)
.,_
g IS
u-
0
CO
u
^
4«3
CO
r-
S-
O>
u
re
i.
JC
C 3
^}
JC
o.
re
s-
CJ>
0
re
g
o
s_
JC
o
e
o
CO
4->
c
d)
g
5
o
o
1
0
o
0
o
1 I
1
X
re
§
u
O)
CL
3
(/)
1
^^
C re
O **-'
i- c
^J !-
C E
O)
4->
o>
a:
1
o
o
o
1
X
z
o
u
r_ »
(U
o.
3
^y
c
o
o.
E
0
o
C to <4_ CO
O C Q) 3 O 3
4J -1- 0) S- O o O
(1) 4-> U re Q) > 3 O O 3
C Qj d) JC re \ U -O O !- O O"
roa>4J''->c04->il>E >> rooO) 4-1 i i re
4J 3 re C CO re +J 3 4^ O. 3 O
C ^ & *»~ ^3 re »" *i~ CO " O" O * O>
Q) T3 E i- i- -l-> S- "OWre -r-j >, C
o. 0) >> co .»-o ja jcs re+-> c r -i-
JO JC 3 ~ C >)- O>CL 0>>>C -i-Cco
JC 0-O-i-JCi -£ JCr'i- 03
r- ro +> 3 O <+- 3 C JC 1/1 *~ S.CC JO i C
» E -i- O" S- 00 O O--i- O.»SCOreo reo
X re jn d) s- re .»r^. .1- +- -i- c -i- .-
O>Jiit O O) 3 S- O) E +->!-> 0)t-4->.
i re O) > >>»-i > 0> T3 O> i > Jiire-r- O) (!) -r- c
ja O) T3 c i i-'O OJD oiesj re a> co co +> co o
re Q.-I- i-a»+J+J-i-*J4JreCi i IUCO- rOO-r-
-t-> S-EEreroS-Cro -(->- O. O. O) O)E O.-i->
Q. .»X7 0> O "O fi a) E i r> c > >>o\ v.j>£'i >tcoo a> C i a) > o Q. ~o *o -*-> o aj
Ui J= +-> re -O J= 't-xiwre o o ro a) a) c
<: o ro i a> Q.-I- re a. o <: <4- re z ro -o oo o u (_> c TS >-
C I C
o c a> c 0)
>>-i- E C > 0 -t->
S_ +J 3 -i- -i- C I re
ttlOr 1 +-> O C T3 X
>Q)O ^OC -i- -r- O C
.- C i - O N C - M- -i- .
O) -r- C Q) M- >,, 0> CO O C 0)
i O i i O COOCOCC
_O i JO Ol OC Q. +-> O >- O ro
re r O) ro C > re tt) £ c -i- +-> -i- -i->
+J re r- +J -r- C i JC Jd O 3 (-> O -M C
o, ja Q.I o o+J« u o i- a> -i- a>
Qj f" *^ Q) >r- ( CO Q) 4^ ^>) Q} E CO "l~^ CO O_
O +-> oo O ro -(-> EC ro T3 ro O C O
O-r-cOC O4JO-H) E O--t- O--O
rOSOO ro O) 0) -o > C4->E EC
C O-'i- Ci 0 i Cr - O JC O i O re
3 -iiJ 4-> 3 ro C J3 re O O).i_o)oreO
re »»T- -i -i re co r 4-> 't H) < O)
>>a>COCO >54-> 4-" O JO O r "O 4-> "O r
r CL4J o i C CO O. CMC re O) CO C O
JO CO. JD ro 3 O>3=O 4-> >-> CreCC
ro TD O) E ro 4-> O O -t- Q. C » E +-> E ro
J3 ro > O Xl oo <1) OJC4-> d) -i- TD 3 oo 3 JC
O O i O O J3 3 ro -U O O Or JO 4->
S- S- O (1) S- 3 O- COO) UOOO3OO)
Q- X) oo T3 Q. 00 rO ^ J3 >-> t3 o t/1 <-» E
CO CM O O
oo ur> o> «* 10 co
. CM >-l iO CO CM
O> i-H i-H i-H i-H r \
in i i o in in
co o in ey> vo
*
O «-H I CM CO Q
-H i-H i-H r>H tt ^£
a>
c
"O Q)
S- ^ I
- S. i
Nl 4-> >»
re -i- jc
4-> ">> O) O 0)
re JE c i- E a>
1_ 4J O O. 1 1
-o a» 4J o CM !-
>> X 0 i S. i i.
jc o re o a. >>4->
I- i C 0 X i-
r T3 O Ol S- - O C
re >> - JC O i. O
i. I O. O. r T3 -i-
O i O O JC >-, Q.
r CM s- -<_> a: o
jc i a. jc i i s-
O Z ' r CO CM O.
58
-------
,«N
o
tt)
3
C
^
4«>
C
0
^*
CO
LU
_J
CO
^
.c
CL
re
S-
0>
o
re
E
O
c^
.C
C
O
t/)
+J
c
0)
C
o
u
^J
o
o
o
0
i-H
1
X
re
^
0
u
tt)
CL
3
y5
»
tt)
E
r-
h-
^^
c re
o ^
r- C
+^ »
C E
0)
4-J
tt)
O£
i
O
O
0
o
i-H
1
X
re
X
o
u
t
tt)
CL
3
OO
o
c
3
O
CL
0
O!
+> C
re
.C 0
X 0
tt) O
o
I/) O>
c
tt) V)
r- 3
JD
4-> tt)
tt) C
U 0
0 &.
< 4-
o
0
0
CO
o
o>
^f
1 1
Cft
in
f t
"g
C
re
CL
O
s-
CL
1
CM
I
O
0
^
.c
(J
1"
o
1
CO
> '
c
o
1 *
u
ft)
e
I
C 1- P-
(J tt) CO
£ O 4->
r- x re
D) oo
Or
co 4-> re
- c il
0) O 0)
( -i- +-> C
"S £ E E
4-> V)
CL o cncn
tt) CL c
o S - co
U O 4->
re o 3 -o
C tt)i C
=> -o tt> re
4-
o
tt)
u
c
tt>
o
ft) C
0
S- -r-
re +>
tt> re
r ^3
u re
&»
" 0)
tt) tt)
r~ TJ
^3
re c
M E
CL 3
tt) i
(J 0
U U
re i
c c
=> 0
CO
CM
t t
O
tn
IK*
t<-H
tt)
1
r-
S-
4-*
f
c
o
f
CL
o
t.
CL
^)
X
o
s-
-o
^^
a:
i
CM
^^
^_
±*
o at>
c c
*^
4-> O
CO O
- O
o ^H
«^ C7
0) C
r^ (^
re 3
CLTJ
a) at
U -4-1
u c
re o
<= s-
o
o
o
C3
00
^J.
(1
o
^^
00
1)
tt)
a
F
E
re
r^
^.
S-
U
<
CT>
C
^J
C
o
S-
(4^
c
c'i
o
- CM
4^
u in
tt)
*F^ <^J
C ro
r
4J*
3 -i-
0 S-
0) 3
O" E
re -i-
c
o
^> "O
re c
a>
O U 4->
- Q. 0
c o
OM-
.« re
O) O O)
r S- C
.0-0.-.-
re 4->
»-> re 3
CL r
tt) <4- tt)
u o i
(J 4->
re Q. oo
coo
=> -t-> CL
o
o
o
C9
o^
^^
^^
^H
^«
o
o
CM
tt)
r
r
J-
4J
C
o
c
o
"re
SET
M
3
O
tt)
3
o-
re
$_
0
^
re
a)
CL
0
c
O
^
4^
U
Q)
C
r-
4^
re
tti
c
S-
o
<4-
o
c
3
O
4-
^^
fO
a>
Q.
r- (
i-H
F-H
r 1
a*
o
CM
0)
c
re
.c
4J
tt)
O
s-
.Jj
c
re
s_
0)
I
E
O ft)
i. C
(«- re
4_>
o c c
01 0 ft)
C - 0.
F ^J
«J o -C
+-> O oo
_Q »r^ Q)
0 C J-
1 >>
Q: 4) "Or
4-> C C
- >> re o
c.
o re . .-o
i- c en i-
4-> re 3 re
o -o
ft) 4-> O C
r^ re c) re
c a) in 4-»
>>
c
0
(4^
-C O
^^ ^o
C t- O ft)
ft) X O U
^ O) C
O 4J ft)
o t- -o
ft) O -f-
o «+- >
tt) S. C ft)
r- ft) O
CL-i- J-
O 4-> ft) E
s- c ref 0)
^* fO & -^^ ^^
C JC tt) O oo
O 4-> C >>
U tt) J-
oo O) O
X oo O
O ft) ft) C C9
J= 0
co 4-> re
c re s- .- .e
r- X O 00 +J
E 4-0
^ ft) 0.0
-J CL ft) E C
E re 4-> o t-
^ re u to
O 0) 3 ft) 3
I"H O""^
*s ft) "O
4-> re T3 +-> tt)
re a) re c c
CL ft) -t-
E -O J- E
3 o a) re i-
i- CO i- CL ft)
^ ft) CL +?
ft) -o re d)
JC ft) -i- "O
I 00 -
- J3 C O> ft)
s- re o c .0.
ft) 4-» (J !-
r- CL +J 4->
S- ft) >> C 0
i- O i O C
re u a) s.
M- -O
C 3
CO - O> O
C -O 4- C U
E ft) tt) -r-
3 CL-O r tt)
o re T- 4-»
o -c co re >»
u in oo re -t-> P
CM X re
C CM >> - C
»-» r 4-> .c re
O O t 3 4->
o *-> re -Q ^ a>
O -r- .C
I- C -i- O> X *->
O -i- S- C
4- E 4-> - ^ co
>^ tt) t re tt)
oo o E >- ft) <->
c E re a. re
o oo >>+> o
1- CO tt) !-
4-* c 4- > -a
r- tt) "O O -i- C
O J= C CO !-
c +-> re a) oo
O ft) ft) ft)
u x *- u o
c o a> x c
o -i- &- ft) a> re
03 E S_ -0 E
re s-
T>CM c -a ft) o
- s_ oo re S s_
X o re re a>
O 4- X -C 4J CL
4-> a.
* o m -a a) o
o re -i- o co -o
4- O ft) X U tt)
* CL re a> +->
ft) O C 4-> O
-C »o *-> ^3 re a>
4J E CD 3 4->
re -o tro>
en i. J- ft) -o
c wo re re -o
- o o o>4-> re 4->
oo s- o at re c o
= CLCD t- t3 - Z
^N ^"^ «~-s
re ^ u
" ^f -^-
59
-------
TKBLE 4. SUMBVRY OF GC SEPARATION EFFECTIVENESS IN TERMS OF k' VMIJES
Number of Analytes
in Given k* Range Using Given Column fa)
k' Range (a)
0-1
1-2
2-3
3-4
4-5
5-6
6-7
7-10
10-15
15-20
20-25
25-30
30-35
ND(b)
Supelcowax-10
1
2
2
0
3
0
4
1
3
5
2
3
2
3
VDOOL
1
9
3
7
4
4
0
0
0
0
0
0
0
3
TOTAL 31 31
(a) The column void times, i.e., RT for k'= 0.0, were 0.60 and 2.40 min
for Supelcowax-10 and VOOOL, respectively.
(b) Not Detected.
60
-------
CJ
2.
H t
^
(/}
Lu
^
>
H»
H^
_J
^H
QO
^t
f
C^
O
HH
H"
>-
g
0
^"
^C
^^
QC
O
LL.
V7
h
«J
^
CO
LU
Q-
^
>
&*
OL
UJ
^>
O
^J
UJ
O£
*
Lr)
UJ
__l
CD
<
H~
CM
0)
flj
u
r
r
O
0)
,4
QJ
ro
U
^w
0
CO
oc
c
re
CO
4->
c
CO
u
s-
eo
Q.
I
i
ro S- C -r-
CO O CO 1
S-H- >
E
CO O -C O
O. U. 1 re
CO
Of
-(>
c
CO
u
s-
QJ
ex
*«.
CO
E
re S_ C "-
CO O CO f
i. **- >
< f- C
o cs o
-* C - C
re 3 0) ->-
co o x: u E
Q. u_ H- ro
CO
^^
fi
^^*
s
+1
^N
re
~^
i.
CO
o
u
CO
ex
in
f 1
o
o
^v
M^
Xl
i.
CO
o
u
CO
cc.
in
1 1
0
0
CO
4^
X)
ro
ro o> CM vo i en vo
oiocn ioO'~i r^ii t>-cM 11
V^^ C7> *^ i VO vo i i 00 vo i i
o o en a CM o > o o oo oo
«* r«. o z ^* CM co z z oo to zz
CM CM O I I CO
o o m ooo vo o a oo oo
** CT> o z in o »* z z * to zz
CO CM CO 1 CM CM tO
CM O tO ICMVO Oil O II
r->ovo i^-vo toil ooo ii
I-H CO *~a4 *""! VO
OOO O 00 O to O O OO OO
^H to CM z in co vo z z in in zz
CO CM ^- i 1 1 ^~
*»^
u
ooo ^- o o - o a oo oo
CM to en 0*0 <* z z into zz
* CM «9- Z < CM r-IVO
CO
c
"O CO CO
- r C
i_ -.- !-
- S- N
IM +> ro r VV
re <- s. c o c
^ C ^3 CO CO ro *"" re
>> O >>CC +->O -C
co x: - jz f- -i- a. o -i->
r +JCOCX ,_ -o M ro , CO
- COCO >,--<» Ore E
COS- X O S- .C&-S. i- Oi
T3-t-> O (-> Q. 4J-1--O COt i-O
C -i- -r- S.COO QJ M >, E>> -»->C
-EC -DOS- Erex: oi -i-eo
QJ ro O >, ro O - i i r i- CJC
r i i rroi o>>>i >>re reo.
o>>>) IEX: ix:j= xzo. s- o
OOO ll-l ^COCO COS- CO *"
<£ Q> -
s- -» c
o o
CO -»->
X) 0
C CO
>, 0 -}
i -i- C
.0 I/I -1-
vt 3 a>
VO O C
o &
CX -f- S-
0 3
o -o
C CO
re QJ S-
CO o
m *o^_
C --T3
r- +J C
S- 0 re
3 3
o -o o
O 4->
o s.
QJ 0. S-
0 0
3 C -i-
-o o s-
o o - ex
s. s. »-»
re ex -i- -o
"O VO CO
C * . O E
re c/) r^ &_
vJ O
>> oo 3 e co -
s- co -a e~a u>
CO -M O 3 <4->
> re S. r rO O
o o ex o 3
o - o s- -o
co c o o
s. ex o o s.
eo - v\3 *-> ex
»-> S- 4-> O
C -i- CO re C
CO O 00 .C M- O
OS O +> >- f-
j- +j ex +J -4J
CO E C S- >-
ex co o o re to
x: o o
o -t-> w c c ex
O -O O ro E
r-H S- !- O
O S. 4-> >> O
re M- O >- i CO
H- CO J3 -0
,
co o E o--
4-> E C O S. C
r co u ex o
3 CO "O CO
CO »-> > --«-»
S. CO oo re
E ro -*: O
O S re "
C S- -O O i CO T3
- n- co JE r ex c
E *J to CO v-
c o s -o
O O QJ oo CO to
-r- +> £ COrere<*-re'O
re T3 S- t- S. CO
QJ -i O> C +> O»+J
x:>4->ooco3
<-> CO O -t-> - re -4-> i
o Z re *J 3 re co
en E o o* E
C II II O CO O CO
- s- -> co s- s.
oo 2: Q j= cx:x: co
=) Q Z O -- 1 <_) S
re X) O *O CO ^
61
-------
S.
"HJ
<«
CVJ
O ff O o evj
in in
VO VO O O O O CVJ
00 S3 »«^ jjtpooo^-joooo- o co co o o i
O -3 (
CNJ
O CVJ O O O CVJ
iI CVJ <
I iI «M
B!
r^
ii in I
co
, 8 . . - °°
ii i i i
t i 181 i i
2s 21
1 o oo iC o co r--t
. ^-» ^-i o in en ID
» p% CM co u? o h^>
co r^. CM *i I-H
CM
I 00 O> '
I rt II <
Ot eft
I CO I
! cQ 88 8 as ffl
CO f*» 00 CO
O)
o>
0)
0)
-'£
O O C
CM
^y S
r 0)
t- , 01 -r
itli;
' I CVJ CO '
^1 GO ^^ CS ^^ ^^J CO ^tf* ^O ^D fl^ll"1*^ 00 O^ ^5 ^~^ CSJ
Cn\ i11iiiii ^H ic i oo|r-1 ^i ii ?vj cvj cvj
62
OJ
IS>
V)
111
q}
o 85 w OJ
LT> C? C "O OJ
O _ < 4J
O> "Q "C3
£ c
) ^S +J .JS
-------
TABLE 7. CONCENTRATIONS AND IONIC STRENGTHS OF SALTS USED FOR H-FTD
RECOVERY ENHANCEMENT
Salt
NaCl
Na2S04
MgCl2
MgS04
Amount
Used, g(a>
1.5
1.5
2.8
2.7
Weight
Percent (a)
23
23
36
35
Approximate
Molarity
4.7
2.0
5.5
4.2
Ionic
Strength (b)
5.1
6.3
17.6
18.0
(a) Amount used per 5-mL aqueous sanple
(b) Ionic strength = 1/2 CiZi2 where Ci is the concentration, in molality/
of the 1th ion of the salt and Zi is the charge on that ion.
63
-------
CM
*«
-2
. s
D
3
i
5
CSJ
"o
CM
[«5
"%
as
o < o o o
vo $ vo o% o. £3 s
- co .5 co CM s 03
§ R w SB 83
co
in
i-. ^V
RJ
CM t>~ ^1 CO
0) _, 0»
^ S'
s-r
CO
b
CM
|
t5
8.
i
1
O)
5
t
6 6|
'&'&&
re lo > i
iJHi^
ll
64
-------
i
a
£
t3
I
S
UJ
o
en
LO
co
2
S'
CJS
(0
t
s
O
O
o
*E
+\
%
CO
co c_>
o
"fi
+1
co
3 CO
o5cn
oo
CO
3 O
CO
a> en
o
1
CM
CM
3 O
£cn
a> en
s
§
a>
CM co to «r ^i- ID ro
OOOOi-ioCOCMI^.
+1 +1 +1 +1 +1 +1 +| +1 +1
n ip
S
00
d '
co
00 * 00
+1 +1 +1 +1 +1 +1 -H +1 +1
IO O CM 00 IO '
m to 10 CM co
SOOCMOOlACMCMi 1
CO ^H Cn
CM
- ro r-«. rt CM
SCOOOO^-IOCMVO
CM ^ ^- I 1
^- in CM en en o co
CM m * en
.CM CO i-I i-i O O
H +1.0 +| +i +i +i +i
oj i oo 00 CM en r^
CM c ii
to r-. CM t
o a>
+^ O Q>
i-i 0»r
*->» 4-> L. 4-> 10 O 10
C 0) >>- .M - X C X
i- S- C -i- C O (0 0)
Q»f OOCO '-+>.£
i
i
O>
co <
S- V
5S
I
o *o -a
^5 .^- Q)
|fl
3 a> co
O- C i Q)
r f0 -r +J
i i. tO CO
10 c -> u a>
^K ^T ^^ ^S
f-8°g-.^
co a> n> s 2.
3*" o> a) i- a>
e o s-
.^ fl> (JI .c
a> +>
_ i a) o
? «0 « g cjl
*J C
C (0 10 0) C
(0 -O -CO
w ^ *v v^
? Q. - O T3
O g CO - i-
w 05 0) 4") *0
I- g +J > <0
O 10 O) 4->
co o i- a to
o
a oo
65
-------
TABU; 10. BREAKTHROUGH CHARACTERISTICS OF ACRDIECN AND ACETCNITRILE ON
NOVEL TRAPPING MATERIAIS
Trap Type (a)
Analyte
Temperatures
Tested, °C
25°C
Breakthrough
Volume(b), L
Specific Breakthrough
Volume, L/gffl
at 25 °C
DADS/HOAfal
Acrolein 70, 72, 91, 121
Acetonitrile 91, 100, 123, 150
DAEM/IMDAfa)
Acrolein 46, 66, 92
Acetonitrne 46, 66, 92, 120
27
36
3.3
10
76
99
14
44
(a) See text for chemical description of trapping material. The amounts of
trapping materials used were: DADS/PMDA, 0.36 gm; DAHVPMDA, 0.23 gm.
(b) Fran extrapolation of the plot of log (retention time) versus
1/T("K) to 25'C
66
-------
TABI£ 11. BREAKTHROUGH CHARACTERISTICS OF ACRDIEEN AND ACETONTTRIIE ON
TENAX
Analyte
Acrolein
Temp, *C
Tested Extrapolated(a)
Elution Volume, mL
Onset(fc) Peak Top(c)
Aoetonitrile
29.9
30.0
25.0
24.8
20.8
18.0
15. CT
360
348
504
512
720
860 (a)
1,070 (a)
412
400
592
584
852
996 (a)
1,240 (a)
30.0
30.1
24.9
25.1
20.6
18.0
15.0
352
352
552
504
808
988 (a)
1,250(a)
400
396
648
584
912
1,130(a)
1,430(a)
(a) Ocnputed from an extrapolation of the other data for which a plot of
log elution volume vs. 1/T(°K) is a straight line.
(b) Less than 0.1% of the analyte had eluted at the onset volume.
(c) Peak asymmetry factors at 10% peak height (front:back) :acetonitrile,
1:3.7 and acrolein, 1:7.2.
67
-------
TABLE 12. ANALYTE CALIBRATION RESPONSE FACTORS FOR GC-MS QUANTIFICATION
Response Factor and Percent Fit(b)
for the Given Calibration Range
Analyte
Quantification
ion,(a)
m/e
Analyte
Level (e)
of Lowest
Calibration
Standard
mg/L
100- Fold Range
30-Fold Range
Average
Response
Factor(c)
Percent
Fit(b)
Average
Response
Factor(c)
Percent
Fit(b)
Acrolein
Methyl ethyl ketone
Methacrylonitrile
Acrylonitrile
Acetonitrile
Propionitrile
1,4-Dioxane
Isobutanol
56
43
41
53
41
54
88
43
20
1.5
5
5
10
5
15
15
0.098
(d)
0.24
0.28
0.18
0.20
0.09
0.24
8
(d)
11
23
14
13
16
15
0.102
(d)
0.22
0.25
0.17
0.19
0.09
0.22
4
(d)
7
22
9
8
16
7
(a) In all cases, the quantification ion was the base peak over the mass range scanned.
(b) The percent fit value is provided by the data system software as an average measure of
how near the calibration points fall to the best straight line calibration model.
(c) Relative to the molecular ion, m/e 84, of the quantification internal standard,
benzene-Ds
(d) Due to the background level of methyl ethyl ketone in the reagent water used, only the
three highest level standards were useable, i.e., a 10-fold range, for which the
response factor was 0.73 and the percent fit 17.
(e) The other four calibration levels were 3-, 10-, 30-, and 100- fold higher than that
for the lowest calibration standards.
68
-------
85
;£
is
CO
en
00
CO
ID
CO T3
to en
to
ooro ofoo 80 en otco
oj oo csj in roo
in cO cn«d
ovo
m
co vo in ^ o in CM o>
00 »-" VO I-H 00
s s
oo co
as
oil/) com
CO 00
cO cvj
in in
I^-CM CSJCM 00"*
S2 S
en o >i ^f oo oo
in r-
ffl'lS
i-N CO
{82
oo en
8S
o PO
i
un <£>
S co co *
oq
S
00
S S S 8 8 8
^i
X
0)
o £
c o
;5
Si
a> o >- <£-
£ -««
1o iz'wS'
_^ ^^^. ^ j
-------
cu ^*
> oo
8
1 c
0
f-H
ex
02
cn
Q.
<§!
00
Q.
0)
*
^
f
\
a vo
.8-
-get:
§
u_ in
'ai cu
>o;
a>
J
a.
ai
ec
CO
f
CVJ
8-
C£.
"
Q.
Q)
ac
f
J* "to 1
'5. Q) as
00 _l 3
m ^f in in in in cnvo
3"
8vo o vo in ^ csj m
oo cn vo cO oo cn oo
i i
cn
Si < VO VO CO « I CO CO
in vo cvj csi vo co oo
CVJ i-H r 1 CO " ' * f
^
gen co m CM CM in co
00 «* VO CSJ CSJ VO O r-»
cvi i 1 1 i co i i m ^-
in
£3 5> co co vo oo csj in
co f i i <*>< vo m
CO
'* ^ o CM in i»» in i
en in cn «* in co vo ^t-
cvj ?-« . i co i in in
cn
CM o r*. r»~ CO m coco
cn in r-- csj csj vo TT cn
cvj i 1 t i co < in *
CO
SCSI O CO O VO O CVJ
in cn co co »»» cvj cvj
co i 1 1 i co i in in
CO
* co co vo co -! cn in
csj m coco vo co ocvj
co i i i co -^ vo m
vo
co < in o vo 03 vo co
o in r^.co covo ii i
co f« i i co -> vo in
cn
2S S ^ & ob1 in 2
co i i co I-H in m
"
CO 00 CM < < O VO CO
CO'* i>- CM co i»» cn K
CM r-* < i l CO > I ^" ^"
88 88 88 88
CO CM CM «* CSJ VO VO
0)
O CU (A
i-> r -a
a> -r- C
4-> a> a> -a
*>> c -r- !- o> re
£ oi- "-s- ci +J
+j i *J i-*J reo oo
c cu >»»- -»J - x c
- t? c -r- c o re cu
c/i > flS f~ O *^ C3 3 *0
,_ y g» ni u . co
* *
Sen o «-i
i co m
^- CO
cn o
vo cn en in
CO CM
vo o O in
tf CO
^
§*
1
CD CO CP
i fZj *?
^^ A) i ^
^=: SS
Mi- re i
0) 4-> X V|_
">, o 5 §
JC +> i S-
*-> a> * co
- o.
.
i
i5
o
2
VO
oo
re
cn
re
£n
-Q
1
^:
. i
1^2
re ->
r
(^ °
"s"i
re a>
C r
Ti-o
oo §
cu £
+J (J
425
a: cu
j=
c
?«£
a>
a. -i->
n u
° o
o; u
a«=>
re .a
70
-------
TKELE 15. CCMPARISON OF CHRCMftTOGRAFHIC PEAK WIDIHS FOR H-PTD
PROCESSED AND SPIZT SEPTUM INJECTED SAMPIES
Peak Width(a> at Half Height, sec, for
the Given Sairolincr Condition
Split Septum
H-PTP(b) In-iection(c)
Analyte Rep 1 Rep 2 Rep 1 Rep 2
Acrolein 6732
Methyl ethyl ketone 10 9 4 4
Methaciylonitrile 12 12 6 6
Acrylonitrne 9956
Acetonitrile 8855
Propicnitrile 10 9 5 6
1,4-Dicxane 20 20 7 7
Isobutanol 13 14 15 16
(a) For the extracted ion current profile of the quantification ion.
(b) The sample was the next to highest level H-PTD calibration
standard.
(c) The injector split ratio was 10:1 and the 2.0 uL injected sairple
contained 10 ug/uL of all analytes in water. The internal and
surrogate standards were excluded.
71
-------
Reproduced by NTIS
0 = 0 (3J
£83
OSJ.EO
'§§£
WSIS5
ri-ET
ZS.S2.E
National Technical Information Service
Springfield, VA 22161
This report was printed specifically for your order
from nearly 3 million titles available in our collection.
For economy and efficiency, NTIS does not maintain stock of its vast
collection of technical reports. Rather, most documents are printed for
each order. Documents that are not in electronic format are reproduced
from master archival copies and are the best possible reproductions
available. If you have any questions concerning this document or any
order you have placed with NTIS, please call our Customer Service
Department at (703) 605-6050.
About NTIS
NTIS collects scientific, technical, engineering, and business related
information then organizes, maintains, and disseminates that
information in a variety of formats from microfiche to online services.
The NTIS collection of nearly 3 million titles includes reports describing
research conducted or sponsored by federal agencies and their
contractors; statistical and business information; U.S. military
publications; multimedia/training products; computer software and
electronic databases developed by federal agencies; training tools; and
technical reports prepared by research organizations worldwide.
Approximately 100,000 new titles are added and indexed into the NTIS
collection annually.
For more information about NTIS products and services, call NTIS
at 1-800-553-NTIS (6847) or (703) 605-6000 and request the free
NTIS Products Catalog, PR-827LPG, or visit the NTIS Web site
http://www.ntis.gov.
NTIS
Your indispensable resource for government-sponsored
informationU.S. and worldwide
------- |