TB85-158145
Thermally Modulated Electron Affinity
Detector for Priority Pollutant Analysis
Radian Corp., Austin, TX
Prepared for
Environmental Monitoring and Support Lab.
Cincinnati, Oil
Jan 85
-------
PB85-1561H5
-EPA/600/4-.S5/009
-January 1985
THERMALLY MODULATED ELECTRON
AFFINITY DETECTOR FOR
PRIORITY POLLUTANT ANALYSIS
by
R.C. Hanisch
L.D. Ogle
A.E. Jones
R.C. Hall
Radian Corporation
Austin, Texas 78756
Contract No. 68-03-2965
Project Officer
Stephen Billets
U.S. Environmental Protection Agency
26 West St. Clair Street
Cincinnati, Ohio 4.W68
\ENVIRONMENTAL MONITORING AND SUPPORT LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
TECHNICAL REPORT DATA
(Please read tnslruclions on iht itvrne before completing)
I. REPORT NO.
EPA/6007 A-85/009
12.
3. RECIPIENT'S ACCESSION NO.
PBS 5 1 5 8 i A 5 /AS
4. TIT LE AND SUBTITLE
Thermally Modulated Electron Affinity Detector for
Priority Pollutant Analysis
5. REPORT DATE
January 1985
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S) " '
R.C. Hanisch, L.D. Ogle. A.E. Jones, R.C. Hall
8. PERFORMING ORGANIZATION REPORT NO,
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Radiai Corporation, Austin.Texas 78766
10 PROGRAM ELEMENT NO
11, CONTRACT /GRANT NO.
EPA 68-03-2965
12. SPONSORING AOENCY NAME AND ADDRESS
Environmental Monitoring and Support Laboratory
26 W. St. Clair Street
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final
14 SPONSORING AGENCV CODE
EPA 600/06
IS. SUPPLEMENTARY NOTES
16 ABSTRACT
In the area of environmental monitoring, a need exists for a rapid, sensitive,
and selective method to analyze for chlorinaced organic compounds such as
pesticides, PCB, PCDD, and PCDF at trace lovels in complex samples.
In response to this neeu, a program w*s conducted to determine the feasibility
of using a new detector concept in the gas chromatographic analysis of certain
priority pollutants. The concept is based on tne thermal alteration of a
compound's electron affinity in a flow-through reactor, which can be used to
modify the selectivity and sensitivity of the ECD to certain compounds. The TM
ECO consists of two ECDs connected by a temperature-controlled reactor. Different
classes of organic compounds respond to the reactor conditions in different ways:
some compounds exhibit an enhanced ECD response after passing through the reactor;
others a diminished signal; and still others no change in the magnitude of the
signal. The ratio of a co^uurid's response from the post-reactor ECD to that
obtained from the prereactor LCD appears to be a property characteristic of each
compound. This peak area ratio can be used in conjunction with its retention time
to increase the confidence level of the identity of a given compound while still
taking advantage of the excellent sensitivity characteristics of the ECD.
17.
WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
l> IDENTIFIERS'OPEN ENDED TERMS C. COSATI field/Croup
B. DISTRIBUTION STATEMENT
'-.Distribute to Public
IB. SECURITY CLASS (ThisReport 1
Unclassified
21. NO. OF PAGES
56
S6COHITY ClfiSS/rhi
Unclassified
22 PRICE
EPA Fwa 2220-1 (R»». 4-77) PREVIOUS EDITION it OBIOLCTE
-------
DISCLAIMER
This report has been reviewed by the Environmental Monitoring and
Support Laboratory-Cincinnati, U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents
necessarily reflect the views and policies of the U.S. Environmental
Protection Agency, nor does mention of trade names or commercial products
constitute endorsement or recommendations for use.
ii
-------
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, conducts research to:
• Develop and evaluate methods to measure the presence and
concentration of physical, chemical, anu radiological
pollutants in water, wastewater, bottom sediments, and solid
waste.
• 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.
• Develop and operate an Agency-wide quality assurance program
to assure standardization and quality control of systems for
monitoring water and wastewater.
• Develop and operate a computerized system foe Instrument
automation leading to improved data collection, analysis, and
quality control.
Thiu report describes the development and evaluation of an analytical
system designed to measure selected priority pollutants in various
environmental media. The ef£e. -.3 of varying caCalvst and tennerature
to modify compound response for an electron capture detector were
evaluated.
Robert L. Booth
Director
Environmental Monitoring and
Support Laboratory-Cincinnati
ill
-------
ABSTRACT
In Che area of environmental monitoring, a need exists for a rapid,
sensitive, and peleccive method to analyze for chlorinated organic compounds
such as pesticides, PCfli PCDD, and PCDF at trace levels in complex samples.
Current methodologies typically employ analysis by GC-ECD, GC-HECD, or GC-MS
for identification and quantification of the compounds of Interest. Each of
these approaches has its advantages and limitations.1»2 The ECD usually
provides the highest sensitivity for these types of analyses but this detec-
tor responds to any compound capable of capturing electrons In an ionized
field including ubiquitous contaminants such as phthalate esters.^* The
HECD is very selective for chlorinated organic compounds when operated In the
halogen-specific mode but it sometimes lacks the required sensitivity. Mass
spectrometry provides unequivocal Identification of compounds but also has
problems with respect to adequate sensitivity and cost of analysis.
In response to this need, a program was conducted to determine the
feasibility of using a new detector concept in the gas chrosatographic
analysis of certain priority pollutants. The concept is based on the thermal
alteration of a compound's electron affinity in a flow-through reactor, which
can be used to modify the selectivity and sensitivity of the ECD to certain
compounds. "The TM ECO consists of two ECDs connected by a temperature-
controlled reactor. Different classes of organic compounds respond to the
reactor conditions in different ways: some compounds exhibit an enhanced ECD
response after passing through the reactor; others a diminished signal; and
still others no change in the magnitude of the signal. The ratio of a com-
pound's response from the post-reactor ECD to that obtained from the pre-
reactor ECD appears to be a property characteristic of each compound. This
peak area ratio can be used in conjunction with its retention time to
Increase the confidence level of the identity of a given compound vhlle still
taking advantage of the excellent sensitivity charcteristics of the ECD.
This report was submitted in fulfillment of Contract No. 68-03-2965 by
Radian Corporation under the sponsorship of the U.S. Environmental Protection
Agency. This report covers the period from September 1, 1980 to November 9,
1961, and work was completed as of June IS, 1984.
iv
-------
CONTENTS
Disclaime r
Foreword[[[
Abstract lv
Figures *i vi
Tables vil
Abbreviations and Symbols viii
1. Introduction 1
Background 1
Scope of Work 1
2. Conclusions.» 3
3. Recommendations 4
4. Experimental. 5
Hardware Development and Preliminary Evaluation 5
Method Validation Study.^.^....,. ; 7
5. Results and Discussion..-;.;. U
— -Response Ratios Obtained with Argon/Methane Carrier
and a Gold Reactir'ii Tube 11
Response Ratios Obf-alr.ed with Nitrogen Carrier and
a Gold Reaction Tube 15
Response Ratio's Obtained with Hydrogen/Helium and a
Gold Reaction Tube..' 15
Response Ratios Obtainei^wich Hydrogen/Helium and a
Nickel ReactioTf'Tube 21
Response Radios Obtained with Nitrogen Carrier and a
Nickel Reaction Tube 2i
Response Rftios Obtained with Argon/Methane Carrier
and a Nickel Reaction Tube....' 24
-------
FIGURES
Nunber
Page
1 Block diagram of detector system ................................ 5
2 Chromatcgrams of chlorinated pesticides with Ar/CH4 carrier
and gold reaction tube ........................................ 13
3 Chromatograms of Coxaphene with Ar/CH^ carrier and gold
reaction tube
.\
4 • 'Chromatograms of parathions with Ar/Cfy carrier and gold
reaction tube ................................. ..... ...... ..... 16
5 Chromatograms of phthalates with Ar/CH* carrier and gold
reaction tube ................................................ . 17
6 Chroma tog rams of Aroclor 1254 with* nitrogen carrier and gold
reaction tube ................................................. 19
7 Linearity plot - commercial ECD ......................... •„ ....... ->6
8 Linearity plot - TM ECD ......................................... 27
9 Accuracy estimates: Phases I ar.J II ............................ 32
10 Precision estimates: Phases 1 and II ........................... 33
11 Percent recovery plot - commercial ECD...* ....................... 34
12 Percent recovery plot - TM ECD .................................. 35
13 Percent recovery plot - p re-cleanup ..................... ^s ..... 3g
14 .Percent recovery plot - post-cleanup ......... .j^r. .............. 37
*
15 Differential chroma tog ram of a toxapherte-contaminated
industrial waste ............ ^.-r;".'. ... .......................... 3e
16 Differential chroma tcgJMn of Sample X-2 prior to cleanup ........ $p
17 Differential chroma tog raa of Sample X-2 after cleanup ....... .... 41
18 Precision estimates for toxaphene determinations in waste
effluent s ........ . ............. . ................ .. ............ ..
vi
-------
TABLES
Number Page
1 Compounds Used for the Electron Affinity Study .................. '8
2 Gas Chromatographic Parameters for the Method Validation Study.. 10
3 Response Ratios at Various Temperatures with Ar/CH^ Carrier
and a Cold Reaction Tube ...................................... 12
4 Response Ratios at Various Temperatures with Nitrogen Carrier
and a Cold Reaction Tube ...................................... 18
S Response Ratios of Chlorinated Pesticides with Helium/Hydrogen
Carrier and a Cold Reaction Tube at 350"C ..................... 20
6 Response Ratios Obtained with Helium/Hydrogen Carrier and a
Gold Reaction Tube at Four Temperatures. ....... . .............. 22
7 Response Ratios at Various Temperatures with a Nitrogen Carrier
and Nickel Reaction Tube.. ..................... •. .............. 23
8 Response Ratios at Various Temperatures with Ar/CH^ Carrier
and a Nickel Reaction Tube ................... . ................ 25
9 Method Validation Study: Toxaphene Spiking and Recovery Data
for Phases 1 and II ........................................... 29
10 Statistical Summary: Accuracy and Precision Estlmaces of
Percent Recovery for Toxaphene Spikes in Reagent Water ........ 31
11 Toxaphene Levels in Industrial Waste Effluent
12 Statistical Summary: Precision Estimates of Recovery for
Toxaphene In Waste Effluent Samples. ........ ...... . ........... 63
vii
-------
LIST OF ABBREVIATIONS AND SYMBOLS
ABBREVIATION'S
C ~ centigrade
ECD — electron capture detector
CC-ECD — gas chromatography-electron capture detector
GC-HECD — gas chromatography-Hail electrolytic conductivity
detector
CC-MS — gas chromatography-iuss spe;.trometry
L — liter
Mln. ~ minute
mL — mllllllter
on ~ millimeter
MS —mass spect rone try
ng — nanogram (10~9 gram)
PCS — polychlorinated blphenyl
PCDD — polychlorinated dlbenzo-p-dloxins
PCDF — polychlorinated dlbenzofurans
pg — plcogram (10~!2 gram)
TM ECO — thermally-modulated electron capture detector
ug — microgram (10~b gran)
uL — mlcrollter (10~6 liter) V
US EPA — U.S. Environmental Protection Agency
SYMBOLS
Ar — argon
— methane
vlil
-------
SECTION' 1
INTRODUCTION
BACKGROUND
r« 3?e.P°t f°r enhancln* ehe selectivity and sensitivity of the ECD
R c ?*}" "BP°"nd8wa8 demonstrated In preliminary 8tudle8 conducted by
R.C. Hall at Purdue University In 1973. Thftse studies utilized a detector
system comprised of two ECDs and a flow-through reactor. The detectors "
"
cmber i 8eile"' bUt "•» "P^ted by a reaon
design. reaction tube and nitrogen carrier gas wire used In this
It was found chat at moderate reactor temperatures (up to 800eC) the
i KhS 8eC°wd detector «•« ""««<* relative to the first for
« f ^"f^bon Pesticides; approximately the same for PCBs; and
S5I ?V Pht»'al«e e««». At high reactor temperatures (950'C) PCBs and
phthalates continued to produce a significant response in the second
detector; most chlorinated hydrocarbon pesticides exhibited Uttle or no
response. The response ratio of the. detector" was also foum* to be
compound-specific and useful for confirmation of compound identity.
These phenomena were used to enhance detector selectivity by directly
Ht fn&^ response ot certain components and by differentially summing
SS-"lTp2 T 8" !S° ellnlnate the re8F°«e of stable components. §
These techniques were used to enhance detector selectivity to chlorinated
hydrocarbon pesticides in the presence of PCBs and phthalJte..
'I!"6 P™1^"8^ results were d^couraging. this approach was
L 3nd dld n°C "Ver a U>lde Varlety of compounds. The
«lllv f compouns. e
stability of the system ua, uut verlfieil ^ ..tuai samples wire not investl-
af It r?U8r Uentl,y>, fUndlng W" sollcite" to ^ly investigate th« utility
0f the technique and determine the feasibility of constructing a commercial
detector system based on this principle.
SCOPE OF WORK
The proposed program consisted of four phases. Phase I was designed to
demonstrate the feasibility of the concept using existing off-the-shelf
components. Phase II would involve the construction and evaluation of a
Uow°in »£.!eiM' ^"^i011 and teSCln* °f 8 P'O'OW detector would
^ ,!5 II t detailed evaluation of the prototype detector would
conducted » Phase IV. Initial funding was sought and granted for only
-------
Che £irst phase, during which a number of specific objectives were required
Co be accomplished. They were:
• construction of a detection system;
• deCerolnatlon of response characteristics Cor selected
coopounds as a function of reactor temperature using at lease
two dltferenc materials as reactor cubes;
• investigation of response characteristics in different
reaction gas coaposlclons Including nitrogen, argon/methane,
and helluo/hydrogen;
• determination of Che most promising reaction Woe/resetion gas
conbination; '
• syeCetn evaluation with a Halted nutnVer of model compounds;
• determination of Che relation of-'detector response as •
function of key operational parameters to reproducibility and
signal aCabillcy for each rjmpound studied;
• evaluation of detector specificity; and
• determination of &n absolute detection level for each compound
studied.
-------
SECTION 2
' CONCLUSIONS
;—.'.__ In this Initial phase of work, ItJbaS^been demonstrated' cfTat & TM-ECD
svsteffcad. have an enhanced select ivirjT fjr Che determination jjt;ss»€e levels
i>i--chloriu.ifv;els of
rlTitBTferences, .-Th^se interfereng>tf"are compounds that eRhi.Ki^ifr'StD'-''
response, sneci'ilcally PCBs ar.u phthalato esters,-andiier^s^U^ifjaiati.'sresent at
significant levels in_sonie_&tutt-goi.m»iugarwastewater rcdtrices. The output
_/.rrm n 'I'M .XLi can De^ii'ferentially amplified to^pull the signals cue to PCBs
and phthalste estB»'-&and-thereby Isolate the signal due only, to chlorinated
pesticides. TVre oevelopcicnt. of tht^'iri ECO could eliminate the necessity for
sample-extract cleanup in the screening by GC of environmental samples for
the -priority pollutant pe.sL~ici«es. -
The opcicuo selectivity for chlorinated hydrocarbon pesticides was
achieved when the TM ;ECD was operated with argon/uiethane reaction gas and'a
gold reaction-tub? ,sfa5iearhed ^it "350*C. The TM BCD provides a response ratio
which is gene-r3tre. vmics of ECDy/2 by the
refipor.se in . rSa units of* ECD/fl. The ntsgultude of this ratio appears to be a
function of.-cKr.iiSca! cla^«. ^-Phthalate esters exhibit a larger ECD response
'after passing'through the reactor and yield a response ratio greater than
-one. -jhider the conditions listed above, PCBs display less than 10 percent
reduction in^jxsgpbnse. As a result, their response ratio is approximately
op.
-------
SECTION 3
-RECCMMESDATIONS-
- The evaluation of the detector prototype has demonstrated the validity
of the TM ECD concept. A final determination of the deteqtor's potential for
widespread application in the area of environmeucol •ecr.itruring is not
possible without additional work. The following areas mexit continued
Investigation:
• determination of the reaction products responsible for the
post-reactor signals; - ' '
• determination of molecular positional effects on the response
factors of various isoners;' "
"• utilva<.viroii of capillary columns in conjunction with the TM ECD;
• definition of the limits of matrix effects on TM ECD response
ratios; and
• optimization of TM ECD selectivity and sensitivity for PCDD and
FCDF.
-------
SECTION A
- EXPERIMENTAL
HARDWARE DEVELOPMENT AND PRELIMINARY EVALUATION
The chroaatcgraphlc system u sed in t hi a Etli'4v~consi s t ed of a "Radinn—11 OB
Gas Chromatograph modifled to accept an experimental detector systeja which IB
shown schematically in Figure 1. It consists of two modified Tracor. ECDs
with a flow-through reactor bei-^een L'ue two detectors. >
The detectors were modified by replacing the 1/4-inch inlet fitting with
.a 1/16-inch fitting and reworking the top detector biscuit to reposition, the
exit port. The modification Involved:
• removing th«.. I/4-inch in-let—a-e- the- base •of~rfie'"de"tect"or"^hd ~
replacing it with a 1/16-inch bored-through male tube fitting;
• removing the 1/8-inch exit tube from the side of the top detector
biscuit and plugging the resulting hole; and
• machining a female tube connection (10-32 thread with 39 degree
'ferrule seat) in the top face of the detector biscuit.
These modifications resulted in a straight flow-through carrier path of
minimum volume. It also enabled the reaction tube and sample transfe.r inlet
tube to be positioned flush to the detector cavity, thus eliminating
stainless steel surfaces from the flow path.
The reactor used for this study consisted.of &• two-hole ceramic tube
wrapped with resistance heating wire and enclosed in ceramic insultation.
The reaction tube was Inserted through one hole of the ceramic tube and a
thermocouple in the other hole. Temperatures of the reactor were controlled
+2°C from 350°C to 900°C.'
A 1/8-inch o.d. glass-lined stainless steel column (1.8 mm l.d. x 168 cm
long) was used for all separations. The exit of the column was interfaced to
the first detector via a short piece of l/16*-inch o.d. (0.035-inch l.d.) gold
tubing. Gold was used for this interface to prevent catalytic decomposition
of the sample prior to detection.
All temperatures of the chromatographic system were controlled by the
DART computer which is an integral part of the chromatograph. An inlet
temperature of 226°C and detector temperatures of 340°C were used. The
" cofOmn temperature was varied to achieve the desired separation. Reactor
-------
GC Column
ECD #1
Reactor
ECD #2
Signal B
DART II
Computer
T | - •
i i .
• 1 i
Report For
Signal A
(area percents)
Reoort For
Signals A & B
(response ratios)
Figure 1. Block Diagram of Detector System
-------
temperatures were va"Yled from 350°C to 900°C. Detector outputs were routed
through the DART II computer which integrated all peaks. The Integration and
chronacographic information were printed on a Teletype printer. A strip-
chart recorder was also employed in order to obtain a record of the
separation.
Seventeen model compounds were selected for use in the determination of
the TM ECD's response characteristics. The basis for their selection was the
fact chat they represent electron-capturing analytes from a variety of com-
pound classes '.ncluding chlorinated hydrocarbon pesticides, PCBs, phthalate
esters, organophosphate pesticides, chloroaromatics, nltroaromatics, and
chlorophenols. The test compounds were grouped into mixtures each of which
contained similar compounds that could be resolved chromatographieally under
the analytical conditions employed. The mixtures were Introduced to the
chromatograph in 5 uL injections. The total amount injected for each test
compound and its absolute detection limit are presented in Table 1.
METHOD VALIDATION STUDY
The target compounds selected for use in the method validation study included
phthalate esters and toxaphene. The validation study itself was based oa^the
analytical procedures employed In US EPA's Method 608 for organochlorine
pesticides and PCBs. The study was conducted based on the assumption that
the phthalate esters were to act as interferants in the analysis of water
samples for toxaphene.
The 200°C i so thermal column temperature required 'for US EPA Method 608
precluded utilization of all six priority pollutant phthalate esters. At
200°C the 1.5% SP-2250/1.952 SP-2A01 mixed phase column employed in the
analysis of organochlorine pesticides would yield a c>rooatogram in which
dimethyl phthalate and diethyl phthalate'would be merged with the solvent
front ar.d di-n,-octyl phthalate would,-elute well after the ast toxaphene
component. As a result, the following phthalate esters were selected for use
in the method validation: di-n-butyl phthalate, benzyl butyl phthalate, and
bls(2-ethylhexyl) phthalate.
System Linearity
Prior to initiating the method validation, the linear range of both the
TM ECD and a commercially available linearized Ni63 gcD was established.
This was accomplished by generating a five-point linearity check using
toxaphene standards prepared at different concentration levels. - Triplicate
injections of each concentration level were made in order to ensure
statistically valid data.
Preparation of Youden Pairs
The toxaphene spiking solutions used in the validation study were
prepared as ten Youden pairs. Acetone was used as the solvent in order to
enhance the water solubility of the concentrates. The spiking solutions were
prepared at toxaphene levels that would yield sample concentrations ranging
from 9.8 Ug/L to 115 ug/L when 1.0 mL of the concentrate was used to fortify
a 1 L water sample. The phthalate ester concentrations were prepared at a
single level that would yield a sample concentration of 50 yg/L of each of
-------
TABLE 1. COMPOUNDS USED FOR THE ELECTRON AFFIMjrY STUDY
Amount (ng) Injecte^' Absolute
Compound Into Chromatograph Detection Limit (ng]
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
Lindane
Hcptachlor
Heptachlor Epoxlde
p,p'-DDE
p.p'-DDT
Aroclor 1254
Aroclor 1016
Toxaphene
Chlordane
Diethyl Phthalate
Dibutyl Phthalate
bis-(2-Ethxlhexyl^ Phthalate
Methyl Parathfott^J
Ethyl ParatSj?*5
1,2,4-Trichrorbb^^-
Nitrobenzene
2, 4-Dichloro phenol • ""
0.55
0.51
0.48
0.50
0.50
~5.2
-------
the esters when a 1 L water sample was fortified with 1.0 tnL of the concen-
trate. The toxaphene concentration in each of the spiking concentrates was
verified oy duplicate analysis against a toxaphene quality control check
solution obtained from the Quality Assurance Branch of US EPA's Environmental
Monitoring and Support Laboratory in Cincinnati, Ohio.
Phase I
The method validation study was conducted in three phases. During Phase
I, twenty 1 L samples of reagent water were fortified with a 1.0 n»L aliquot
of spiking solution containing appropriate concentrations of tcxaphene and
the three selected phthalate esters. After spiking, the 20 water samples
were extracted according to the protocol detailed in US EPA Method 608.
Briefly, this Involved extraction of the sample with three 60 mL portions of
nethylene chloride, drying of the combined extracts oa an anhydrous sodium
sulfate column, and concentration of the dried extracts followed by solvent
exchange into hexane. A method blank was extracted with each set of samples
using the protocol described above.
After extraction, the samples were analyzed on both a conventional ECD
(ECDll in Figure 1) and TH ECD. The gas chroaatographic parameters utilized
in the method validation study are presented in Table 2. Quantitation of the
samples was achieved by peak height comparisons against external standards.
Phase II
In Phase II, the extracts analyzed in Phase I were cleaned up on
-F3ro?£-silffiLcolumns using the procedure recommended in US EPA Method 608. The
6£_diethyl ether in hexane cluate fraction was analyzed under the same
chromatographic conditions as were used for Phase I. Prior to beginning this
phase of the study, the Florisii<& eluticn pattern tor toxaphene was
established using standard solutions.
Phase"IIT ~
During Phase III, I L industrial wastewater samples known to contain
toxaphene as a contaminant were extracted and analyzed according to US EPA
"Method 608 procedures. The analyses were conducted prior to and after
Flori7ltS-Qplu:nn cleanup using both detection systems.
A total of fo;Tr^£'\xaphene-containir.g wastewater samples were obtained
for use in the validation^Studjr. One sample was divided into three 1 L
aliquots.
Two of the-ailquots were extracted aritT-aaalyzed as previously described.
The third aliquot* was fortified with a toxaphene-spiking solution as a method
recovery"check. -The three remaining samples were each- treated as single
determinations.
-------
TABLE 2. GAS CHROMATOGRAFHIC PARAMETERS FOR THE
METHOD VALIDATION STUDY
Coliimn - 168 cm X 1.8 mm I.d. glass-lined stainless
steel packed with 1.5% SP 2250/1.95% SP 2401
on 100/120 mesh Supelcoport.
Temperature Program - Isothermal at 2003C.
Carrier Gas - 5% methane/952 Argon at 35 uL/mln.
Purge Gas - 5Z methane/95% Argon at 25 mL/min.
Detector Temperature - ECD - 340°C.
TMEDC - 340°C.
Reactor Temperature - 850°C.
Inlet Temperature - 220°C.
10
-------
SECTION 5
RESULTS AND DISCUSSUN.-
Thermal stabilities of the 17 model compounds were determined with tha
detector configuration phown in Figure 1. The primary objective of this work
was to find the most useful combination of temperature, carrier S&B, and
reaction tube material which can be used to differentiae between chlorinated
pesticides, rCBs, phthalates, and chlorinated hydrocarbon^
RESPONSE RATIOS OBTAINED WITH ARGON/METHANE CARRIER AND A GOLD REACTION TUBE
The response ratios determined experimentally at various temperatures
with argon/ire thane carrier and a gold reaction tube are presented in Table 3. .
The ratios ?>hovn are the average of three determinations. The standard devi-
ations for vhe determinations are also shown. These ratios are based on the
assumption that the response of ECDtfl and ECDfr'2 are equal for all compounds
with an ambient reactor temperature. An ambient temperature for the reactor,
however, was impossible to attain due to the temperature of the detectors
(340°C each). Therefore, equal response was assumed and a reactor temperd^
ture of 3M)°C was used as the minimum temperate-re.
At 350°C, the response ratios were not equal to 1.0 for all compounds.
This indicates that some rearrangement takes place either, in the first
detector or in the reaction tube at 350°C. The compounds altered the most
were toxaphene, chlordane (both chlorinated hydrocarbons), and the
phthalat.es.
As the temperature of the reactor was increased, the chlorinated
pesticides were degraded tc species less responsive to the electron capture
detector. This resulted in response ratios less than 1.0. The response of
ECDfll and ECDfr'2 for the chlorinated pesticides at a reactor temperature of
900°C is shown in Figure 2. The most thermally stable chlorinated pesticides
were p,p'-DDE and p,p'-DDT. Toxaphene and chlordane, both aliphatic chlori-
nated hydrocarbons, were found to be very unstable as shown by the data in
.Table 3 and the chromatograms reproduced in Figure 3.
PCBs.. nitrobenzene, 1,2,4-trichlorcbenzene and 2,4-dichlorophenol
exhibitec1. high thenaal stability. The thermal stability of PCBs is well
documented. Therefore, the response ratios were expected to be close to 1.0
and show little change as a function of temperature. The stability of
nitrobenzene, 1,2,4-trlchlorober.zene and 2,4-dichlorophenol was surprising.
These compounds, particularly nitrobenzene and 2,4-dichlorophenol, are
chemically reactive and were therefore expected to exhibit thermal
instability. -
11
-------
TABLE 3. RESPONSE RATIOS1 AT VARIOUS TEMPERATURES KITH Ar/CHu CARRIER
AND A COLD REACTION TUBE
1.
2.
3.
4.
5.
6.'
?:
8.
9.
10.
11.
12.
13.
•I/..
15.
16.
17.
Compound
Llndane
llvptdclilor
; lleptaclilor Epoxldc
I>,P'-DDI£
'. p,p'-l)DT
.Aroclor 1016 *
Aroclor 1254
Toxaplicne
Clilordane
Dlothyl I'll thai ate
Ulbutyl Piithalate
l>ls(2-EU,ylhexyl)
Phthalate
Methyl Parothioi)
Ltliyl Parathlon
Nitrobenzene
1,2,4-Trlclilorobunzene
2,4-Dlclilovoplienol
Reactor Tumpurattire (°C)
350
0.9310
0.8210
0.8210
0.8UO
0. 7810
0.9810
0.9410
0.6110
0.7210
1.5410
1.6110
1.8410
0.9010
0.9210
l.OCUO
0.8210
1.0710
.01
.01
.01
.01
.01
.02
.06
.06
.01
.30
.24
.39
.01
.01
.02
.01
.01
0.
0.
0.
0.
0.
1.
1.
0.
0.
1.
1.
1.
0.
0.
0.
I.
1.
600
9J10.01
77i0.01
8210.01
8310.01
7810.05
001 0. 01
OU0.01
6410.02
6910.02
5810.10
5510.04
8010.05
8910.01
9110.01
8410.01
0210.02
0410.01
700
0.9010.01
0.4110.06
0.7710.02
0.8210.01
0.5410.03
0.9810.02
1.0110.02
0.0510.03
0.2710.10
1.8210.01
1.5710.04
1.8510.07
0.3110.01
0.2610.01
0.9010.03
0.9710.01
0.9010.05
800
0.1110.
0.05
0.1710.
0.5910.
0.2110.
0.9210.
0/.9210.
<0.01
0.(,)3
3.4810.
1.81.10.
2.04^0.
0.0910.
0.09tO.
0.841,0.
0.9210.
0.8510.
859
01
01
02
03
03
03
04
05
OS
01
01
02
01
03
0.05
0.02
0.0810
0.4810
0.2510
0.9210
0.8910
<0.01
0.01
6.5510
2.5210
2.4410
0.0810
0.0910
0.8510
0.9310
0.7510
.01
.01
.03
.01
.03
.04
.02
.08
.01
.01
.01
.01
.01
900
0.04
9.02
0.05
0.40*0.01
O.lSlv-.Ol
C.('.S10.01
0.7710.04
<0.01
<0.01
8.0910.15
3.1610.02
3.0310.12
0.0710.01
0.0810.01
0.9010.02
0.9110.01
0.7710.03
1Based on equal response of ECD tfl and ECD
-------
-------
-------
Phthalates and the parathions were found to be thermally Bistable. The
organophosphate pesticides, methyl and ethyl parathion, lere thermally
degraded to products having very little electron affinity. As a result, the
response ratios were quite small (see Figure 4). The phthalates were also
very unstable. However, the products formed during the thermal degradation
of the phthalates had a greater electron affinity than the parent compounds.
Figure 5 Illustrates the threefold enhancement of bls(2-ethylhexyl) phthalate
and the eightfold enlianceoent of the diethyl phthalate at a reaction tempera-
ture of 900er.
RESPONSE RATIOS OBTAINED WITH NITROGEN CARRIER AND A COLD REACTION TUBE
The response ratios obtained at six reactor temperatures using nitrogen
carrier and a gold reaction tube are shown in Table 4. In general, the
results with nitrogen as the carrier were similar to those obtained with
argon/methane. The chlorinated pesticides, the parathions, and the
phthalates were again found to be thermally unstable. The PCBs, nitro-
benzene, trichlorobenzsne, and dichlorophenol, were found to be more
thermally stable.
Chlorinated pesticides appear to be less stable with nitrogen carrier
than with au argon/methane carrier. Significant degradation was observed at
700°C with nitrogen, whereas argon/methane required a temperature of 8CO°C
before significant decomposition was observed. PCBs also appear less stable
with a nitrogen carrier. For example, Aroclor 1254 has a response ratio of
0.77 at 900°C with argon/methane and a ratio of 0.29 at 900°C with nitrogen.
This is illustrated by the chroma tog rams presented l:i Figure 6.
Comparison of the response ratios of phthalates with nitrogen and
argon/methane carriers show contradictory results. The response ratio of
diethyl phthalatr. is greater in nitrogen than a.r&on/methane at 900°C (14.86
versus 8.09, respectively), but bis(2-ethylhexyl) phthalate has a greater
response ratio at 90f:°C in argon/methane (3.03 versus 1.98 for nitrogen).
The remaining compounds, the parathions, nitrobenzene, 1,2,4-trichloro-
benzene, and 2,4-dichlorophenol, displayed similar response ratios in
nitrogen and argon/methane. It is interesting to note that the response
ratio of nitrobenzene, gradually decreases as the temperature approaches
850°C, but then increase at a reactor temperature of 900°C for both nitrogen
and argon/methane carriers.
RESPONSE RATIOS OBTAINED WITH HYDROGEN/HELIUM AND A GOLD REACTION TUBE
The response ratios for the helium/hydrogen gas composition were
obtained by using a carrier of 3CmL/min helium £nd a make-up gas of 30 mL/oin
hydrogen. The make-up gas was adcisd to the column effluent immediately prior
to entering the first detector.
Response ratios obtained with a helium/hydrogen carrier gas did not
yield reproducible results as indicated by the data presented in Table 5.
This table shows the response ratios obtained for the chlorinated pesticides
at 350°C with different equilibration times. The phthalates also exhibited
poor reproducibiJiricc.
15
-------
B
I
V
OJ
C
O
o.
CO
&
FIGURE 4.
Chromatograca of ParatHons with Ar/CH* Carrier and Gold
Reaction Tube.
Order of elution: Methyl Parathior. and Ethyl Parathion
Chronatogram A, ECt) 11, 500 pg each. Attenuation X10.
Chromatogran B, ECU 92, 500 pg each after passing through a gold reactor
at 900°C, Attenuation X2.
16
-------
a.
§
'CU
ca
O-
JC.1
o
V)
f •/
B
Time
FIGURE 5. Chromategrams of ?hthalates with A.r/CH* Carrier and Gold
Reaction Tube.
Order of elution: ^Diethyl, Dibutyl, and bis(2-Ethylhexyl) Phthalate.
ChromatosjrJm A, ECD #1, 5 ng of each. Attenuation X10.
Ghromatogram B, ECD #2, 5 ng of each after passing through a gold reactor
at 900°C, Attenuation X20.
17
-------
TABLE 4. ' RESPONSE RATIOS1 AT VARIOUS TEMPERATURES WlTH^NITROGEN CARRIER AND A'
C'0(,D REACTION TUBE
•a K c: «w«s
L.
2.
'3.
4.
i.
h.
7.
8.
9.
io.
11.
12.
13.
14.
15.
16.
17.
a. »j=»^j^aa- «a^j-j-^=a=a..,-yia.-»._.j.i^ j j A
Compound
L Ind Jim
llcptaclilor '
lleptaclilor Cpoxtde
p.p'-DDli
p.p'-DDT
Aroclor 1016
Aroclor 1254
Toxaphune
(Milordane
Uieihyl Plulialate
Dibutyl Phtlmlate
bis(2-Ethylhoxyl)
I'lilhalate
Methyl Parathlon
Ethyl ParatliJon
Nitrobenzene
1 , 2 , 4-Tr Icli lorobenzene
2,4-DJchlorophcnol
Keaetiir Tempera cure (°C)
350 .
0*9310.01
0.4710.03
0.7610.01
0.7510.01
0.5110.01
0.9710.01
0.95A0.04
0.6110.01
0.7310.01
2.0910.09
1.6410.08
s2. 0710. 12
.
0.8610^.01
0.8810.01
0.7810.01
0.9810.01
l.U.'»10.04
600
0.9110.01
0.3810.01
0.7310.01
0.7310. 01 «
0.4410.04
0.9510.01
0.8910.03
0.5110.02
0.7110.01
2.0910.07
1.7810.06
1.9310.07
0.6810.02
0.00 10. 02
0.8410.01
0.99*0.01
0.9810.03
700
0.12^0.01
0.041
0.1610.01
0.4910.01
0.1810.0^'
o.83to. 01 ;.
0.7310.02 '
0.0910.01
0.03
2.8510.23
1.8510.14
2.2010.02
0.1710.01
0.1810.01
0.9210.03
0.9810.02
0.6910.03
'800
0.03
0.02
0.05
0.1710.01
0.0610.02
0.7210.02
0.4410.02
0.04
0.01 '
5.7810.74
1.9410.17
1.8810.32
0.1010.01
0.1210.01
0.8810.02
0.9210.03
0.5710.03
850
0.02
0.01
0.02
0. 1010. 01
0.0510.01
0.6310.01
0.34±0.03
0.0?
<0.01
6.9610.15
2.1710.28
1.8810.20
0.0710.01
0.09:4.01
0.8810.01
0.9010.01
0.6010.02
1
900
0.02
0.01
0.02
0.0910.01
0.03
0.6010.01
0.2910.01
i
0.02
<0.01
14.8610.72
3.0210.57
1.9810.14
0.0710.01
0.0810.01
0.9710.02
0.9010.01
0.7210.02
on equal response of BCD ffl and LCD
-------
0*
CO
C
o
a*
V)
CJ
PS
B
I
o
CO
Time
FIGURE 6. Chromatograms of Aroclor 1254 with Nitrogen Carrier and Gold
Reaction Tube.
Clircmatogram A, ECD #1, 5.2 ng of Aroclor 1254, Attenuation X10.
Chromatograin B, ECS #2, 5.2 ng of Aroclor 1254 after passing through a gold
reactor at 900°C, Attenuation X10.
19
-------
-'TABLE 5. RESPONSE RATIOS OF CHLORINATED PESTICIDES WITH HELIUM/
HYDROGEN CARRIER AND A GOLD REACTION TUBE AT 350°C
RESPONSE RATIOS (ECD 02/ECD
Comoound
Llndane
Heptachlor
Heptachlor Epoxide
p.p'-DDE
p,?'-DDT
Held at 350°C
for 1 Day
0.14
0.30
0.39
0.80
0.23
Held at 350°C
for 2 Davs
0.90
0.43
0.88
C.81
0.45
Reactor Heated to 900°C
and Cooled to 350 9C
0.01
0.08
0.09
0.31
0.07
-------
Reproducible results were obtained for the PCBs, 2,4-dichlorophenol, and
l»2,4-trichlorobenzene. Toxaphene, chlordane, nitrobenzene and the para-
thions had very low response ratios at all temperatures, as indicated in
Table 6, and were not studied in great detail.
Heating the reactor to temperatures above 800°C caused an apparent
"activation" of the reaction tube. The reactor would require several days to
return to its original level of activation once the temperature had been
reduced. This "activation" caased very poor reproducibility, especially for
the chlorinated pesticides and phthalates. The reactor "activation" was
postulated to be due to a temperature dependent reaction between the hydrogen
and some substance coating either the inside of the ECDs or the reaction
tube. Removal of the material would then leave an active surface which could
degrade compounds until a new "coating" was built up at lower temperatures by
column bleed or reaction products. Regardless of the-cause, this problem
makes helium/hydrogen an unlikely choice for the carrier gas except for the
thermally stable species.
Detector linearity with a helium/hydrogen gas composition was determined
with the geld reaction tube at 600°C using standards spanning three orders of
magnitude. The observed detector response was not linear over the complete
range of concentrations. Similar results were obtained with dibutyl
phthalate and p,p'-DDE. In addition, the response ratios determined were not
consistent over three orders of magnitude for these compounds. The non-
linear response observed is not surprising since the detector's relative
pulss width was not adjusted for this gas composition.
/
RESPONSE RATIOS OBTAINED WITH HYDROGEN/HELIUM AND A NICKEL REACTION TUBE
Replacement of the gold reaction tube with a nickel tube did not
alleviate the problems observed with gold. All compounds, except diethyl
phthalate, were completely destroyed at a reactor temperature of 900CC. The
"activation" of the reactor was again observed at high -emperatures. After
reducing the reactor temperature, several days were required for restoration
of the original activity level.
The temperature at which the reaction tube was "activated" was deter-
mined to be approximately 860°C. Below 860°C, most of the compounds
responded on ECD#2. Above 860°C, all of the analytes except nitrobenzene and
diethyl phthalate were destroyed. The poor reproducibility and lack of
discrimination in destroying the various compounds makes helium/hydrogen a
poor gas composition for use with a nickel reaction tube.
RESPONSE RATIOS OBTAINED WITH NITROGEN CARRIER AND A NICKEL REACTION TUBE
Response ratios determined with a nitrogen carrier and a nickel reaction
tube at six different temperatures are shown in Table 7. When compared to
the results of both argon/methane and nitrogen with a gold reaction tube, the
response ratios of the nickel tube with a nitrogen carrier were found to be
lower at 350°C for every compound except the phthalates and nitrobenzene.
The phthalates were observed earlier to have larger response ratios due to
the formation of a degradation product with a greater electron affinity. The
increased degradation of all compounds at 350°C with-nitrogen/nickel sug-
gested that catalytic reactions were occurring (i.e., the gold tube is more
21
-------
TABLE 6. RESPONSE RATIOS OBTAINED WITH HELIUM/HYDROGEN CARRIER
AMD A COLD REACTION TIIBF AT FOUR TEMPERATURES
Teroeracure (°C)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
Compound
Lindane
Heptachlor
Ueptachlor Epoxide
p.p'-DDE
p,p'-DDT
Aroelor 1254
Aroclor 1016
Toxaphene
Chiordane
Diech/1 Phthalate
Dibutyl; Phchalate
bis<2-Ethylhexyl) Phthalate
Methyl Parathion
Ethyl Parathion
Nitrobenzene
1,2, 4-Trichloroben zane
2,4-Dichlorophenol
- 350'
0.14
0.30
0.3S
0.80
0.23
0..75
0.76
0.12
0.33
3.76
7.01
2.26
0
0
0.08
0.92
0.76
600
<0.01
0.02
0.02
0.18
0.09
"0™. >6 .
0.75
0.04
0.02
7.03
2.7,4
2.05
0
0
0
0.96
0.75
800
<0.01
<0.01
0.01
0.08
0.07
0.57
0.76"
<0.01
<0.01
_J.2.86"
2.84
2.18
0
0
0
0.91
0.61
9UO
0
0.02
0.01
0.09
0.07
0.77
0.77
0
0
20.29
3.93
2.36
0
0
0
1.00
0.64
1Reactor held at 350°C for one day prior to analyses.
22
-------
TABLE 7. RESPONSE RATIOS AT VARIOUS TEMPERATURES'WITH A NITROGEN CARRIER AND
M
NICKEL REACTION TUBE
Ruatitor Tcnruraturu (°C)
tni\ i onn "
Compound 3r>0 'fi'lO 70?) 800 850 9(10.
" * ' •—•• f -i
1. Llndanu 0.8510.01 0.84*0.01 0.66+0.05', 0.35+0.04 0.03 0
2. llupcuclilor 0.6110.01 0.^310.01 0.1710.02 0.09*0.01 0.02 0;
I. lluptarhlor Exoxlde 0.6110.01 0.5BJ0.01 0.3310.03 \ 0.1610.01 0.03 , 0'
4. p.p'-DDE 0.5410.UI ;0.i3iO.()l 0.4510.01 0.36i0.01 0.17+0.01 0.02 •
5. p,|>"-l)l)T 0.5Ji0.01 0.4810.02 0.2410.01 0.2010.02 O.lOlO.Ol 0
6. Aroclor 1016 0.7:»iO.Ol/ 0.7410.01 0.7110.01 Q.63-«0.01 0.05 0.0.1
7. Aroclor 1254 0.7iiO.O? 0.6510.00 0.60i0.02 O'.44i0.02 0.1210.01 0.01
8. Toxaphone 0.47J0.02 0.4710.02 0.3510.02 0.1210.04 0.0) 0
9. Chlorddiic O.n3'0l0l O.VJ10.01 0.49+0.01 0.1810.02 0.01 0
10. Dicthyl Phtlialate 2.05i6.03 2.4610.12 1.8410.29 10.0011.64 0.3510.01 0.1210.01
11. Dlbucyl I-liLliulale 2.1bJ0.02 2.6710.03 2.95+0.10 3.52+0.07 0.1U0.01 0.03
12. bis(2-Erhylliexyl) 3.5910.15 4.3610.17 4.38t0.17 4.1810.19 2.0310.29 0.4010.05-
riithuluic
13. Methyl Purathion '0.7210.01 0.69J0.01 0.57i0.03 0.02'. 0 0
14. Ethyl Paratliion 0.74i0.01 0.69JG.OI 0.4510.01 <0.01 ' 0 0
15. Nitrobenzene • 1.4910.02 1.4910.02 1.3110.01 000
lb. 1.2.4-Trlclilcrobtinzcno 0.7510.01 0.7310.01 0.7110.01 0.05+0.01 0.01 0
17. 2,4-Uicliloropliuiml 0.7110.02 0.65i0.04 0.5)10.04 0.04 0.02 0.01
-------
inert Chan the nickel Li be). The nickel tube displayed another difference
from the gold tube with nitrogen. At temperatures >900°C all compounds were
degraded including the normally thexnally stable PCBs, chloroaromatlcs and
nitroaromatics. In ad lit ton, the response ratios of the phthalates were
smaller suggesting that the intermediate species formed by the thermal
degradation of tiie phLhalates were degraded further to species which have
small electron ifiinities. Due to the.se difficulties, this system was not
investigated further.
RESPONSE RATIOS OBTAINED WITH ARGON/METHANC CARRIER AND A NICKEL REACTION
TUBE
Response ratios obtained using an ar^on/cethane carrier and a nickel
reaction tube at six different temperatures aru shown in Table 8. At 3SO°C,
Che response characteristics for the argon/methane carrier are very similar
to those observed for a nitrogen carrier when a nickel tube is employed as
the reactor. The reactivity of the nickel tube appears to increase more
rapidly with a corresponding Increase in temperature when argon/ex* thane is
used in place of nitrogen as the carrier.
This trend is evident when the values at a reaction temperature of 8tiOcC"
are compared for the two systems. The response ratios of all the compounds -
except the phthalates are extremely small for the argon/methane system at
800°C whereas ~he response ratios of the chlorinated compounds and the
phthalates resulting from the nitrogen system are significantly greater at
the same temp'srature. The response ratios for the organophosphate pesti-
cides, the chloroaromatics, and the nitroaromatics are very similar for both
systems at this temperature.
In the case of the chlorinated pesticides, these data cay be indicative
of the occurrence of free radical formation in the argon/methane atmosphere
within the aickel reactor followed by subsequent recombination reactions in
which-species are-formed that have lower electron affinities. The nickel
tube appeals to act as a catalyst under these conditions since the test
compounds exhibit much greater thermal stability when argon/methane is used
in conjunction with a gold reaction tube. Because of these characteristics,
this syst-*m was not selected for use in the validation study.
METHOD VALIDATION STUDY
System Linearity Check
In order to determine whether the prediction equation fits the
five-point calibration data, measures of the adequacy of the fit were
examined. In particular, residual plots were constructed and used as
graphical aids in visually inspecting the fit of the prediction equati'on.
The plots of amount (ng) versus the response (mm) for both the commer-
cial ECB and the TK ECD presented in Figures 7 and 8 respectively, indicate
that a linear a<>£o?iatlon between the two variables Is the dominant
relationship. Because of the minimum nvnber of points, it is difficult to
say whether other components, peihcps a quadratic term, should be added to
the prediction equation.
-------
TABLE 8. RESPONSE RATIOS AT VARIOUS TEMPERATURES WITH Ar/CH., CARRIER AND
A NICKEL REACTION TUBE
10
Reactor Temperature °C
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
Compound
Llndane
Haptachlor
HepCnchlor Epoxide
p,p'-DDE
p,p'-DDT
Arocloi 1016
Aroclor 1254
Toxaphene
Chlordane
Diethyl Phthalate
Dlbutyl Phthalate
bis(2-Ethylhexyl)
Phthalate
Methyl Parathion
Ethyl Parathion
Nitrobenzene
1,2,4-Trichlorobenzene
2 , 4-Dichlorophenol
350
0.84+0.01
0.65;:0.01
0.62+0.01
0.5 /tO. 01
0.53+0.02
0.79+0.01
0.85+0.02
o.4:»+o.n
0.6(1+0.02
2.29+0.09
2.15+0.03
3.49+0.27
0.80+0.01
0.85+0.03
1.64+O.OJ.
0.79+0.0?.
0.79+0.0-4
600
0.82+0.01
0.52+0.01
0.59+0.01
0.57+0.01
0.46+0.02
0.74+0.01
0.67+0.08
0.34+0.0*
0.49+0.0)
2.67+0.41
2. 35 H), 08
3.54+0.60
O.;8f_0.04
0.51+0.03
1.2040.01
'0.76^0.01
0.74+0.01
700
0.73+0.01
0.34+0.01
0.48+0.01
0.50+0.01
0.37+0.01
0.. 32+0. 05
i ^
0.43+0.10
0.06+0.03
(),25K).02
1 -6.73+P.38
', 2. 78+0, 14
.3.85+0.29.
A. 08+0. 01
10.11+0.02
1 0.02
0.46+0.03
p.d6_H).oi
800
0.01
<0.01
0.01
0.01
0.03
0.06+0.01
0.04+0.01
<0.01
0.04+0.02
2.56+1.90
1.06+1. OR
2.37+0.78
0.02
0
0
0.14+0.02
0.06+0.02
850
0
0
0
0.01
0
0.07+0.01
0.05+0.01
0
0
0.97+0.10
0.13J.0.01
0.09+0.02
0
0
C
0.12+0.02
Q. 05
900
0
0
0
0
0
0.02
0.01
0
0
0.31+0.20
0.03
0
0
0
0
0.07+0.01
0.02
-------
10 M
o^ ™
I
1250 *
I
1
i:co *
i
R I
E I
. S 1
' P 750 *
'" I
t) I
S. I
r i
500 *
M I
I
I
250 * X
I
* '
. I V
I X
0 *
-4*--4«— .4-— .4---4 — 4 — 4---4— «4— «-4 — -* — 4 — 4— »4-
2P ?« ?0 IF 40 *5 5C 55 60 6F. 7f 7F PC «"5
Figure; 7. Linearity Plot - Commercitl\ BCD
•, \
-------
I
1250 *
I
I
I
I
1000 »
1
R *
r l
s '
P ^750 +
0 f
N I
S I
r t I
50J •»
•» I
M IN
, I
?fO •»
I
r x
i . *
o *
0 5 1," IF ?0 2? 30 3F ^0 45 5fl 5S 69 6F 7S 75 flO
AMOUNT (NO
Figure 8. Linearity Plot - TM BCD
-------
In Investigating the possible need to include a quadratic terra or some
other component in the prediction equation, residuals of the' linear fit were
plotted against the amount (ng) for each of the two ECOs. The results did
not indicate a need for some transformation of the predictor variable as the
residuals appear to be randomly centered around the value zero.
"As a further check on the adequacy of the linear '^t, a quadratic tetm
was assumed as part of the true underlying codel:
Response • a •+• bi (amount) + b2 (amount}2 + e
where a, bi, and 02 are the unknown model parameters and e is the error
tetm.
The least-squares coefficient estimates exhibited below indicate that
amount is the most influential determinant of the response for both
detectors.
Intercept Coefficient Estimates
Detector a f>\ €2
Commercial ECD -27.5 12.8 0.01
\
TH ECD -34.5 13.3 0.005
The hypothesis that 02 « 0 was tested at 0.25 significance level (a
very conservative test) with an individual t-test for each of the ECDs.
-Since the significance probabilities were greater than 0.25, the hypothesis
of a zero 'regression coefficient (02) was not rejected.
It appears from the above analyses that the linear fit is appropriate
for both the commercial ECD and the TH ECD. In addition, there does not
appear to be a statistically significant difference in the linear reRpor.se
characteristics of the two detector? for toxaphene over the range of 5 ng to
80 ng.
Phases I and II
A phased approach was utilized in the method validation study in order
to determine the effects of Florisll® column cleanup and sample matrix on the
precision and accuracy %>f the analytical methodologies. The sample fortifi-
cation and recovery data for Phase I of the method validation study are
presented in Table 9. It should be noted that multiple data entries at a
given concentration level are indicative of replicate analyses.
In addition, the data listed under the "ECD" category In Table 9 are the
result of the chromatograms generated from the response of ECDtfl (Figure 1).
The data tabulated under the "TMECD" heading were obtained from the chromato-
grams generated by the differential amplifier. This device subtracts the
signal produced oy ECD02 from the signal generated by ECDffl. With this con-
figuration, compounds with low response ratios «1.0) yield positive peaks on
the chromatogram while compounds with high response ratios (>1.0) produce
28
-------
TABLE 9. METHOD VALIDATION STUDY:
PHASES I AND II
TOXAPHSNE SPIKING AND RECOVERY DATA FOR
Analysed Concentration (UR/L)
Phase I (Pre-Cleanup)
Youden Pair
1
2
3
4
5
b
7
8
9
10
11
12
13
14
15
16
17 v
18
19
20
True Value (ug/L)
9.8
12.5
i5.5
17.3
19.6
25.0
24.5
30.2
33.0
38.0
50.3
57.3
66.0
76.0
76.1
86.3
80.6
99.5
100.6
114.6
ECD
7.8, 7.5-
11.1
11.5
14.8
17.3
16.0
2J.O
30.0
33.6. 33.0
38.7
44.5
52.8
63.6
74.0
75.9
85.6
68. C
87.0
85.0
95.0, 100.
97.7
1MECD
7.8, 8.0
11.0
11.2
14.9
17.9
16.2
22.6
31.9
31.7, 31.0
35.6
47.8
57.6
67.1
77.8
82.0
90.0
74.0
93.7
87.8
103. 106.
103
Phase II (Post-Cleanup)
ECD
7.6
9.6
11.6
13.0
11.3
15.6
20.0
26.0
27.2
3-'.. 5
40.0
50.2
55.7
64.5
70.5
84.9
65.0, 65.0
84.0
80.0
91.0, 91.9
TMECD /
3
7.5
9.4
11.0
12.7
11.4
15.0
21.4
26.0
27.0
33.3
40.0
51.6
60.0
67.0
73.0
83.5
76.0, 73.0
SI. 7
84.8
98.0, 96.7
-------
negative chroaatographic peaks. Under the conditions used In this study,
toxaphene pvoducec positive peaks on the differential chronatogram while the
phthalate esters exhibited a negative response.
For the recovery Jata presented in Table 9, the standard deviation
appeared to depend on the toxaphene concentrarlon, with constant proportion
over ell mean con-rx-nt rat ions. Therefore, the percent relative standard
deviation was judged to be the nost useful measure of variation.
A statistical sunnary of percent recovery of toxaphene spiked in reagent
water is presented in Table 10 for both the commercial FCD and the TMECD.
The intralaboratory percent relative standard deviation (precision estimate)
which indicates che repeatability of an Individual laboratory or analyst when
using the method i.s computed by:,-''
Sr
RSD x 100
seat percent recovery
where the single analyst standard deviation, Sr, was calculated from the
Youden pair data and deflr.cc by Youden6 as
-D)2
\/ -
S,
/E(Di -
Jr -V 2(n -
where n •" cumber of paired observations,
D£ • the difference In percent recovery between observations for a
sample palr,~-
D - the average value Cor Dj.
A comparison of the two detectors in terms of accuracy and precision for
both the pre- and post-cleanup aethods can he made using the statistical
suanary given In Table 10. A graphical presentation of the accuracy and
precision estimates are presented In Figures 9 and 10 respectively. It Is
apparent that the commercial ECD and the TM ECD work equally well for the'"
analysis of toxaphene in fortified reagent water samples. This is.trfie for
both pre--aod post-cleanup methods.
Additionally, it is observed that while the variability of recoveries
using the post-cleanup method is significantly, smaller than that for the pre-
cleanup method, recoveries using the post-cleanup method are consistently
lower.
Figures 11 through 14 siaphically summarize the recoveries for the two
types of detectors and-cleanup cethods used. The concentration level does
not appear_tp af/cct the recovery of toxaphene.
A paired comparisons t-test~-*a«-._run to statistically test for dif-
ferences between the two detectors and the pre-_and post-cleanup recoveries.
The results of this test were<"*as foiiO«*s:
30
-------
TABLE 10. STATISTICAL SUMMARY: ACCURACY AND PRECISION ESTIMATES OF:PERCENT
RECOVERY .FOR TOXAPHENfe SPIKES IN REAGENT WATlvR
Method
Pre-cleanup
Commercial ECD
TM ECD
Post-cleanup
Commercial ECD
x TM ECD
Mean
Estimate
89.1
92.2
80. '8
82.7
Pc-.rclpnt Recovery
Approximate
95% Confidence Interval
\
i •
(83, 95)
(86, 93)
i(79, 83)
(80, 85)
% Relative
Estimate
fe!e
3.0
,3.5
Standard Deviation (RSD)
Approximate
95% Ccu fide rice 'Interval
(5.7, 15)
(6.1, 16)
\
(2.0, 5.4)
(2.4, 6.4)
!„ \ _i Measured concentration ., . j,
Percent Becovery •= —~ —: X-10(
r i ' True concentrations. '
Measured concentration
""'\
: sr _.-.._ , . ••
rrN—sr-r X 100 where S,. is the intralaboratoiry standard deviation
Mean >i Recovery r . . n
•I . : . '
-------
U)
1 VV
P 95
E
R
C
F 90
L-
N
T
85
R
E
C
0 .88
V
5 75
Y
-._
•
i™"
i 1
1
_
e
N*
_
'• «
i
i
\
1
\
\
i
.
_i
( ,
i • i
.
-
EC
D ' TM
ECD
f V L_^
:
>
i
1 •"*' '
ECD TM ECD
1
Figure 9. Accuracy Estimates: Phases I and II (Estimate <> 95% Confide
-------
w
w
1 /
E 15
E
L
A 13
T
I .,
V H
E
9
S
T
D 7
E 5
V
3
r
f-
-
m
i_
•
-i
i
i_
"1
i
1
_i
_i
El
-
:D TM E
:CD
•
i
'
1 1
i
ECD TM ECD
Figure 10. Precision Estimates: ^.,'phape I and II (Estimate
35% Confidence Interval)
-------
\100
1' 'JO
t
c
r
N fiO
T
h
r.
C 7C
^J _
^ L
V
t
K
y 60
i
' 50
\
\ i; L«
*••*****•***••**•«»** [<*«»*»***«***«***«»**<«|, »***>]*••***••••••*•»**•*•••
\ I' A
\ H
\
\ h A
•• s • A
\ I M ti A U
\ n T. A
\ \ . A A H A
\ > A A 11
* U N A A A A
t
\ A A '
\ A
\
*
\
^ B » Pre-cleanup
\ n
^ * x A « Post-cleanup
* "
\ A
•J- v
V l N
\
•1 . \
\ \
\ \
I' .10 2u 30 <<0 b(l t.O 7U t!0N
-------
ion
•
!•
F '•
i. mj
c
r
M
T
HO
B
t
C
w °
u- V 70
L
l«
Y
60
. ,0
* , ''
\ ('
\ "
V
\ H H / A
\ 1- A L
\ h I' !• A
+ A A o
\ ,- A A A L1
\ h A A
\ A
\ *
* b A
\
\ AA
\ A
'A • Pre-cleanup
••A •
* B « Post ^cleanup
\
\ »
\
+ A
\ A
\
\
\ '
10 ?o 30 <«n "so i-o 7o . fcii -
-------
\
ico «* *...»•,•*..•***••**•••••*••••••••«•»••••••••"•*•••••••••••••••••*•*•••
\ • L i
X T T
T r '
? 90 ; T i i i
P N L LCI
C \ L C » ,
r N c L' ' , r r
r; so * * L t, C
T \ t
\ U C
K \ CT
r \
C 70 * T
C V c . Comaercial ECD
V \ '
C N T - TM ECD
K N E
Y. (>0 « T
\ \ »-
\
\
\ ^
•JO «
\
N ' .,.«.._.»,
" *,"""lo"""iO il ''0 5iO tO 70 I-C SO IOC 110 1*0
UUL LONCF.MKA1 ION (i'l !•)
Figure 13. Percent Recovery Plot - Pre-Cleanup
-------
\
\
\
V
[
P
C
f
K.
T
P
I
c
f
V
C
K
V
\
9'J *
\
V
. \
\
HO +
\
\
v
\
70 *
\
\
\
\
60 *
\
\
\
\
,
C
E - Conunercial ECD
T • TM ECD
lc 2-0 JO 10
SU
(.6
7.'p » C
^0
IOC- 110 i:-0
V-l'L COUCCMRA1 lOJ- It Phi
Figure 14. Percent Recovery Plot -\Post-Cleanup
-------
• [The commercial ECD and the TM ECD are significantly different
(at the 0.05 significance level) with respect Co percent
recovery for the precleanup method—TM ECD recoveries-are
closer to the true value; and
• The pre-cleanup and post-cleanup methods are significantly
different (at the 0,05 significance level) with respect to
percent recovery for* both ECD-detectors—pre-cleanup recovery
results are closer to true value.
Phase III
The purpose of this phase of the method validation study was ro
determine the effect of sample matrix on the precision and accuracy of the
analytical methodologies. Four Industrial waste effluents known to contain
Coxaphene were ajialyzed using both detection systems,.prior to and after
cleanup. One 'sample was of sufficient volume to be-divided into three
aliquots for a check on analytical precision. One of these allquots was
fortified with a toxaphene spiking solution at a level which would yield an
equivalent sample concentration of 86..3 pg/L- The quantitation of the
samples was achieved by peak heigr.t comparison against external standards.
For all samples except X-2, a minimum of three peaks were used to quantitate
the toxaphene concentration. ' The level of phthalate interference prior to
cleanup was so great for sample X-2 that the quantitation was based on a
single peak in the chromatogram. Figures 15 and 16 are illustrative of the
differential chromatograms obtained for a sample containing only toxaphene
and sample X-2 containing both toxaphene and phthalates respectively. The
differential chromatogram of cample X-2 after cleanup is presented in Figure
17. It snould be noted that the same peaks v.-ere used in the quantitation of
the toxaphene concentrations for both detection systems. The analytical data
for the industrial waste samples are summarized in Table 11.
A statistical summary of recovery precision for toxaphene in effluent
samples is given in Table 12 for the two detectors. As was illustrated for
clean samples, the TM ECD appears to be equivalent to the commercial ECD in
this particular application. Figure 18 Is a graphical presentation of the
precision estimates for toxaphene in waste effluents. The method detection
limit for this type of sample is estimated to be 5 pg/T" The results of the
method recovery check listed In Table 13 indicate quantitative recovery of
toxaphene from the fortified industrial waste sample.
Phthalic Anhydride Experiment
An additional series of analyses using the TM ECD vere performed in an
attempt to identify the mechanism responsible for the enhanced phthalats
response. It was hypothesized that phthalate esters were being converted to
phthalic anhydride within the heated reactor and that the enhanced signal was
the result of the relatively greater electron affinity of the anhydride. To
test this hypothesis, a 5 ug/pL phthalic anhydride in acetone solution was
analyzed on the 110 GC equipped with the TM ECD. The analysis was performed
at two different reactor temperatures: 650°C and 850°C. The response ratio
for this compound was calculated and found to be 0.8? for both reactor
temperatures. This indicates that the phthalic anhydride is thermally stable
•under the conditions employed. These conditions included a gold reactor tube
38
-------
01
§ 3
-
OS
Time
Figure 15. Differential Chromatogram of a Toxaphene-
Contaminated Industrial Waste (128 pg/L)
39
-------
a,
c
Time
I
Figure 16. Differential Chromatogram of Sample X-2 Prior to
Cleanup (43.2 pg/L)
40
-------
I
.1
I
Time
Figure 17. Differential Chromatogram of Sample X-2 After
Cleanup (AC.7 ug/L)
41
-------
TABLE 11. TU.XATHESE LEVELS IN INDUSTRIAL WASTE EFFLUENT „
Concentration (ug/L)
Sample ID
X-1A
X-1B
X-1C
X-2
X-3
X-4
Spiking Pre-Cleanup
Level (Ug/L) ECD TKECD
123
112
126
86.3 235
43.2
158
35.0
33.4
108
105
123
195
63.2
112
38.3
37.3
Post-Cleanup
ECD THE CD
115
121
194
199
39.9
147
150
39.0
111
106 "
184
182
40.7
130
127
38.4
-------
TABLE 12. STATISTICAL SUMMARY: PRECISION ESTIMATES OF RECOVERY
I FOR TOXAPHENE IN WASTE EFFLUENT SAMPLES
Intralaboratory
% Relative Standard Deviation
Approximate
Method Estimate 95% Confidence Limit
Pre-Cleanup
Commercial ECO . 6.8 (3.7, 26)
TM BCD 12.7 (7.2, 47)
Post-Cleanup
Commercial ECO 1.8 (1.0. 6.8)
TM ECD 1.8 (1.0, 6.8)
-------
50
R 45
E
L 40
A
T 35
I
V 30
s25
D15
V 10
% 5
I
ECD
I
TM ECD
ECD
TM ECD
^ Prc-Cleanup
Post-Cleanup - - - - -
Figure 18. Precision Estimates for Taxaphene
1 Determinations in Waste Effluents (Estimate
952 Confidence Interval)
-------
TABLE 13. METHOD RECOVERY
Soriceritration (
" _-• - - => Pro-Cleanup
Parameter
. Post-Cleanup
'"TMECD
Average toxaphene cone. (Samples
-X1-A and Xl-B)
Toxaphene spiking level
Predicted toxaphene
(yg/L) -"" .
Analyzed toxaphene concentration
Olg/L)
Z Toxaphene recovery
_86.
235 195
ItS (8
i9.e-- 183
.95 94
-------
and Ar/CH4 carrier gas.' An attempt was made to compare the pose-reactor
response of an eqjji'valent amount of di-n-butyl phthalate to that of the
phthallc anhydride.
The phthalic anhydride chromatographic peak tailed so badly, however,
that it was impossible to get an accurate comparison. Consequently, It was
not possible to conclusively identify the formation of phthalic anhydride as
the'underlying mechanism for the phthalate ester response characteristics.
Nevertheless, it was found that phthalic anhydride was more sensitive than
-the phthalate ester. The difference between the sensitivities of the
anhydride and the ester was approximately the same as the elevated response
observed for the"^srer after passing through the reactor at S50°C. Thus, the
formation of phthalic anhydride is a plausible" .reaction mechanism.
-------
REFERENCES
1. Aue, V.A. , Detectors for Use ic GC Analysis of Pesticides. J.
Chromatogr. Sci. , 13: 329-333, 1975.
2. Budde, W.L. and J.W. Eichelberger, The Mass Spe&nrog^ter as a
Substance-Selective Detector in Chroma tog raphy. J. Chroma cog r. 134:
147-158, 1977.
*3r
3. Aue, V.A. , and S. Kapila, The Electron Capture Detector— Controversies ,
Comments, and Chromatograms. J. Chromatogr. Sci. 11: 255-263, 1973.
4. -David, D.J. , Gals Chroma tog raphic Detectors. John Wiley & Sons, Inc..,
NY, 1974. 295 pp. """" — —-^
Register. Vol. 44, No. 233, Monday, December 3, 1979, p. 69501.
n, W.J. , "Statistical Technique for Collaborative Tests,"
elation of Official Analytical Chemists, Washington, DC, 1967.
"47
------- |