Professional Documents
Culture Documents
Presented at the Waste Testing and Quality Assurance Symposium 2001, August 13-16, 2001, Arlington,
A V
ABSTRACT the analyte. The MDLs for these analytes range from
16 ppb for carbon tetrachloride to 157 ppb for
Dexsil Corporation has recently developed a total
methylene chloride with the majority around 25
halogen method for the analysis of water samples for
ppb.
chlorinated organics. The new procedure is an addi-
tion to the existing methods for the L2000DX
The method development data indicate that the
Chloride Analyzer. The L2000DX has been in use in
L2000DX water method should be a useful tool for
the field for 12 years and can be used for the analysis
monitoring ground water in areas contaminated with
of transformer oil, surface wipes and soils. The soil
chlorinated solvents. Now, one instrument can be
method is the basis for SW-846 Method 9078
used to track contamination in both soil and water.
"Screening Test Method for Polychlorinated
This type of screening can be used for checking the
Biphenyls in Soil." Using the same reaction and
progress of a treatment process, defining the limits of
quantification steps as the established L2000DX
a contamination plume, as well as, monitoring
methods, the new method uses a liquid-liquid
ground water over long periods of time. All of these
extraction step to achieve method detection limits
applications do not require expensive laboratory
(MDLs) in the 15-30 ppb range for most chlorinated
analysis and the L2000DX can provide the informa-
solvents. This new method will allow nearly real
tion in the field in 10 minutes
time data to be collected from ground water moni-
toring wells at a fraction of the cost of laboratory
INTRODUCTION
analysis.
The L2000DX is a screening tool for chlorinated
This paper describes the new method, the develop- organic compounds and has been used in the field
ment work in establishing the feasibility of the for the analysis of transformer oil for PCB contami-
method and the comparison with SW-846 laboratory nation since 1989. Shortly after its introduction, a
methods for spiked samples. Recoveries and MDLs method was developed to use the L2000DX for the
have been determined for the following analytes: screening of soils for PCBs. This method eventually
Carbon Tetrachloride, 1,2-Dichloroethane (DCEA), became SW-846 Method 90781and has been evalu-
cis-1,2-Dichloroethylene (DCEE), trans-1,2- ated in two different forms, first, through the SITE
Dichloroethylene (DCEE), 1,2-Dichloroethylene program and, most recently, under the ETV pro-
(cis/trans mix), Methylene Chloride, gram2 both run by the USEPA. The procedure for
Pentachlorophenol (PCP), Tetrachloroethylene (PCE), PCBs in transformer oil has just recently been vali-
1,1,1-Trichloroethane (TCEA), Trichloroethylene dated through the ETV program and a method has
(TCEE), Vinyl Chloride. As expected, the extraction also been developed for use on wipe samples.3
recoveries vary inversely with the solubility of the
analyte in water which, in turn, affects the MDL for
In the late 90s, with the rise of chlorinated solvents matrix preparation steps and the conversion to the
and pesticides as contaminants in the environment, appropriate units. The fourth correction necessary to
the demand grew for the development of L2000DX obtain a useful result is derived from empirical data
based methods for other chlorinated compounds in on the overall efficiency of the analytical steps,
soil. Laboratory experiments and field trials have including the extraction efficiency from the original
since demonstrated that the L2000DX, following matrix, the reaction efficiency for the conversion to
suitable matrix preparation, is suited to the analysis chloride, the recovery of the chloride ions in the final
of basically any extractable chlorinated organic in buffer solution, etc. The theoretical conversion fac-
soil.4 tors can be calculated for all analytes and tabulated
for each matrix, but the correction for recoveries
As chlorinated solvents leach into ground water and must be experimentally determined for each matrix
interest in cleaning up old dry cleaning and manufac- and each analyte.
turing sites grows, so has the need for an L2000DX
method for water analysis. The objective of this After some experimentation, it was determined that a
work was to develop and document a water extrac- simple liquid/liquid extraction would be the simplest
tion technique for concentrating organo-chlorine and also effective. Of the commonly used solvents,
contaminants into an organic solvent suitable for e.g., hexane, heptane, cyclohexane, it was determined
introduction into the L2000DX system. that 2,2,4-trimethylpentane (isooctane) worked well
and had the advantage of being less volatile. A liq-
The L2000DX is a total chlorine analyzer. A require- uid/liquid extraction will accomplish two things: by
ment for the system is that all chlorine must be choosing the appropriate solvent, it eliminates inter-
chemically converted to inorganic chloride for quan- ferences due to inorganic chloride and by choosing
tification. By changing the sample preparation steps the correct solvent to sample ratio, a concentration of
either total chlorine or total organic chlorine will be 1:100 can easily be accomplished.
measured. There are three elements to the analysis:
sample preparation/extraction, converting the organic The effectiveness of a liquid/liquid extraction is deter-
chlorine present into inorganic chloride using metal- mined primarily by the extracted compound and is,
lic sodium and the quantification of the resulting therefore, not constant over all possible contami-
chloride using a chloride ion selective electrode nants. In general, compounds with a high water sol-
(ISE). Over the 12 years of its use in the field, the ubility will tend to have low recoveries in the solvent
L2000DX has proven to be a very reliable instrument layer. Extraction efficiency can also be affected by
and the ISE based system has been shown to be accu- water quality parameters, e.g., pH, salinity, ionic
rate and relatively free of interferences. (NOTE: All strength, etc. In this work we set out to determine
organic chlorine is quantified as the target analyte, the extraction efficiency for a liquid/liquid extraction
however, this is the nature of a TOC measurement combined with the L2000DX sample preparation as
and is not an interference, per se. Inorganic chloride well as the MDL for the quantification of environ-
can be an interference, but can be removed in the mentally significant organo-chlorine contaminants.
matrix cleanup step and, therefore, would not inter-
fere.)
difluorobenzene, chlorobenzene-d5, and 1,4- samples since these samples were diluted by a factor
dichlorobenzene at a fixed concentration reading of of five. Similarly for GCMS samples, the dilution fac-
50.00 mg/L. In the internal standard mixture, the tor was the size multiplier. Once these theoretical
surrogate analytes present were dibromofluo- corrections had been made, a regression analysis of
romethane, toluene-d8, and 4-bromofluorobenzene at the corrected instrument response versus the theoret-
concentrations of approximately 100.0 mg/L. A cali- ical concentration for each analyte indicates the lin-
bration curve for PERC was made of concentrations earity of the response. Table 1 shows the comparison
ranging from 5 ppb to 200 ppb. of the coefficient of regression for both the L2000DX
and the GC/GCMS response curves. Figure 1 show-
RESULTS AND DISCUSSION ing the measured concentration versus the theoretical
concentration for PCE, illustrates a typical response
The methylene chloride experiments indicate that
curve for both methods. The slopes of the regression
varying water quality parameters have little effect on
line can be taken as the average percent recoveries for
the recovery of chlorinated solvents. While this
each analyte, as shown in Table 2.
means that the recovery of methylene chloride cannot
be easily improved, it also indicates that variations in
A plot of the average recovery versus the analyte sol-
ground water should not significantly affect recover-
ubility in water for all analytes tested, indicates that
ies in the field. As expected, pH did affect the recov-
the recovery is inversely proportional to an analyte’s
ery of PCP . The recovery improved from 54% to
solubility (See Figure 2). Using published data, it
60% when the pH was reduced from neutral to less
can be seen in Figure 3 that a reasonable approxima-
than 2.
tion of the recovery can also be made based on an
analytes octanol/water partition coefficient.7
The one water parameter that did have an affect on
recovery of chlorinated solvents was temperature.
The ultimate sensitivity of the L2000DX method is
Spiking water of various temperatures at 1 ppm with
determined by the sensitivity of the ISE. The practi-
PCE, resulted in recoveries ranging from 35% at 2EC
cal limit for chloride detection, with some low-end
to approximately 60-65% at 12EC. Above 12EC the
corrections, is 1 ppm in the final extract solution
recoveries remained more or less constant in the 60-
used for quantification. Since all of the preparation
70% range, up to the highest temperature tested at
steps are the same for each analyte, and the final
45 EC.
extraction multiplier is determined by the extraction
efficiency, the achievable MDL for each analyte will,
In order to determine the recoveries, theoretical cor-
therefore, be a function of the recovery. The results
rections were made to the instrumental readings to
from the MDL determinations, listed in Table 2, con-
account for percent chloride (on L2000DX readings
firm this general trend with the lowest MDL of 16
only) and dilution/concentration factors. While the
ppb achieved for carbon tetrachloride ( 58% recov-
GC experiments analyzed the VOC analyte concen-
ery) and the highest of 157 ppb for methylene chlo-
trations directly, the L2000DX uses a chloride-specif-
ride (9% recovery).
ic electrode, which measures free chloride concentra-
tion. The L2000DX readings were converted from
chloride concentrations to analyte concentrations by
dividing the measurement by the percent chlorine in
the analyte. The extraction of the analyte from water
to isooctane produced a size multiplier for the
L2000DX samples of 965/10 = 96.5. Although there
was no extraction involved in the GC samples, a size
multiplier of 5 was used for the most concentrated
SUMMARY REFERENCES
The results from the response range experiments indi- 1. US-EPA (1998) Screening Test Method for
cate that the L2000DX liquid/liquid extraction Polychlorinated Biphenyls in Soil, Test Method for
method for water is has a linear range up to 10 ppm Evaluating Solid Waste, Physical/Chemical Methods
for all of the analytes tested. The coefficient of regres- (SW-846), Update IVA, U.S. Government Printing
sion was greater than 0.96 for all analytes including Office, Washington, D.C.
methylene chloride, the hardest to extract. The
recovery for each of the analytes has been demon- 2. Dindal, A.B., Billetts, S, (1998) Electrochemical
strated to be consistent and reproducible enough to Technique/Ion Specific Electrode Dexsil Corporation
make this method a suitable method for field use. L2000 PCB/Chloride Analyzer, Environmental
Typical MDLs are in the 25 ppb range for most of the Technology Verification Report, US-EPA, Office of
analytes of interest which should make the L2000DX Research and Development, Washington, D.C.,
water method a useful tool for monitoring ground EPA/600/R-98/109, August
water in areas contaminated with chlorinated sol-
vents. Now, one instrument can be used to track con- 3. Dindal, A.B., Bayne, C.K., Koglin, E.N., (2000)
tamination in both soil and water. This type of PCB Detection Technology Dexsil Corporation
screening can be used for checking the progress of a L2000DX Analyzer, Draft Environmental Technology
treatment process, defining the limits of a contamina- Verification Report, US-EPA Office of Research and
tion plume, as well as, monitoring ground water over Development, Washington, D.C. Report (in Press)
long periods of time. All of these applications do not
require expensive laboratory analysis and the 4. Lynn, T.B., Kneece, J.C., Meyer, B.J., Lynn, A.C.
L2000DX can provide the information in the field in (1997) Determination of Chlorinated Hydrocarbon
10 minutes. Concentrations in Soil Using a Total Organic Halogen
Method, Presented at the 13th Annual Waste Testing
and Quality Assurance Symposium, July 6-9,
Arlington, VA
Table 1: Comparison of Recoveries and Regression Coefficients for GC and L2000DX Response Curves
14
DX
GCMS
12
Measured Concentration (ppm)
y = 1.186x - 0.0149
10 R2 = 0.9902
4 y = x - 0.0223
R2 = 0.9692
2
0
0 2 4 6 8 10 12
Theoretical Concentration (ppm)
0.7
0.6
y = -0.366x + 0.5067
0.5 R2 = 0.7628
Recoveries
0.4
0.3
0.2
0.1
0.0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
Solubility in Water (wt%)
0.7
0.6
0.5
Recoveries
0.4
0.3
0.2
y = 0.2614x - 0.1793
0.1 R2 = 0.8781
0.0
1 1.5 2 2.5 3
Log Kow