Analysis of Trihalomethanes
by Micro Liquid-Liquid
Extraction and Capillary
Gas Chromatography with
the HP 6890 Micro-ECD
Gas Chromatography
April 1997
Application Note 228-379
Authors Series GC system with micro- niques are time-consuming and
F. David ECD will allow detection of require more complex instrumen-
Research Institute for trihalomethanes at levels as tation than direct injection into a
Chromatography, low as 0.1 ppb. gas chromatograph (GC)—even
Kennedypark 20, B-8500 with a liquid-liquid extraction step.
Kortrijk, Belgium Introduction
Although direct aqueous injection
P. Sandra Trihalomethanes, including chlo-
gives reliable results while not
University of Gent, roform, dichlorobromomethane,
requiring sample preparation, the
Krijgslaan 281, S4, B-9000 dibromochloromethane, and bro-
use of retention gaps that have to
Gent, Belgium, Eindhoven moform, are commonly present
be replaced at regular intervals3
University of Technology, in tap water at low (less than
is required. Micro liquid-liquid
The Netherlands 25 µg/L) concentrations. These
extraction with pentane followed
Matthew S. Klee compounds are formed from or-
by GC-electron capture detection
Hewlett-Packard Company ganic material during chlorination
analysis is a desirable alternative.
2850 Centerville Road of drinking water. According to
Wilmington, DE 19808-1610 World Health Organization (WHO)
In this application note, automatic
USA regulations,1 the sum of the four
micro liquid-liquid extraction
compounds should not exceed
using an HP 7686 Series Prep-
100 µg/L. There is, therefore, a
Abstract Station, followed by capillary gas
need to monitor the presence of
chromatography separation of an
Trihalomethanes were mea- these compounds in drinking wa-
aliquot from the upper (pentane)
sured in tap water samples by ter using reliable methods. In
layer, is shown to be a rapid and
capillary gas chromatography addition, the use of fast analysis
reliable method for trihalometh-
(CGC) with a new electron and automated methods save
ane analysis. The new HP 6890
capture detector (ECD) after laboratories time and expense.
Micro-ECD was used for sensitive
automated micro liquid-liquid
detection of the trihalomethanes.
extraction. The percent rela- Trihalomethanes can be analyzed
tive standard deviation (% in drinking water samples using
RSD) for the entire process— different techniques, including Experimental
from sample preparation static headspace, purge and trap, Calibration solutions of the four
through analysis—was less liquid-liquid extraction, and direct target solutes were prepared in
than 10 percent for the target aqueous injection.2, 3 Static pentane. Water samples were
compounds. Total sample turn- headspace and purge and trap are taken directly from a tap. Then,
around time was 10 minutes. very sensitive techniques; concen- 1-mL aliquots were pipetted into
Extrapolation of the data indi- trations below 1 ppb can easily 2-mL automatic sampler vials
cates that the new HP 6890 be detected. However, both tech- (this manual operation could
2
also have been done by the Extraction
HP PrepStation). The remainder solvent
of the analytical procedure was ((( )))
automated.
As shown in figure 1, the HP 7686
PrepStation was used to dispense
0.5 mL of pentane into each 1 mL Water Dispense 0.50 mL pentane Withdraw sample
sample vial. The compounds were and mix and inject
extracted by vortex mixing for
1 minute at medium speed on the
Figure 1. Variable Sampling Depth and PrepStation for Automated Liquid-Liquid
HP G1296A bar code reader. The Micro-Extraction
two solvent layers were allowed
to sit for 1 minute to separate.
Then the sample vial was trans-
ferred to the HP 7673 automatic Table 1. Instrumental Configuration and Analytical Conditions
sampler for analysis. Chromatographic System
Gas chromatograph HP 6890 Series
Tap water varies in pH and ionic
Inlet Split/splitless
strength. A more complete
Detector HP 6890 Micro-ECD
method validation might show
that adjustment of pH and ionic Automatic sampler HP 7673 Series
strength prior to extraction would Liner Single taper deactivated (HP part number 5181-3316)
increase yields and reproducibility. Data handling ChemStation (DOS Series)
Column 30 m x 0.53 mm id x 2.65 µm HP-1
For the experiments reported in (HP part number 19095Z-123)
this application note, 100 percent Experimental Conditions
recovery of the trihalomethanes Inlet temperature 250 °C
was assumed for the pentane Injection volume 1 µL
extraction step. The specific Injection mode Splitless
recovery would have to be known Sampling depth 10.0 mm
to report the most accurate Purge time 0.50 min
results. However, if the recovery
Purge flow 50 mL/min
was less than 100 percent, the
Carrier gas Hydrogen
reported results would be less
than the actual, so any result Head pressure 12 kPa at 35 °C
indicating higher than allowable Carrier gas mode Constant flow
halocarbon concentrations would Flow, velocity 5 mL/min, 38 cm/s
still be valid. Oven temperature 35 °C, 1 min initial, 10 °C/min to 125 °C
Detector temperature 340 °C
Analytical Conditions Detector gases Argon/5% methane: 30 mL/min
The analyses were performed on
an HP 6890 Series GC. Automated
splitless injection was done using
an HP 7673 automatic sampler
with enhanced parameters, which
allow injection from different
heights in a sample vial.3 The
instrumental configuration and
analytical conditions used are
summarized in table 1.
3
Results and Discussion Peaks
A typical chromatogram obtained 1. Chloroform
for a calibration mixture containing 2. Dichlorobromomethane
5 ppb of each trihalomethane in Hz 3. Dibromochloromethane
pentane is shown in figure 2. All 4. Bromoform
2 3
compounds are well separated
400 1
within 8 minutes. In this chro-
matogram, it appears that GC run 4
time can be shortened, but when 300
extracted samples are analyzed,
the presence of possible interfer- 200
ing peaks indicates the necessity
for the longer run time. 100
The linearity of the detector was
0
measured in the range of 1 to
100 ppb with a 1-µL injection 2 3 4 5 6 7 8 min
volume corresponding to the Figure 2. Analysis of 5 ppb Standard Mixture of Trihalomethanes in Pentane
range that is usually monitored (Experimental conditions are listed in table 1.)
in tap water samples. The micro-
ECD gave a linear response for Area
all four compounds, as shown Dichlorobromomethane
18,000
by the calibration curve for Area = 163.720137*Amt + 71.211287
dichlorobromomethane in figure 3. 16,000 Correlation: 0.99933
5
Table 2 shows the correlation
14,000
coefficients for the four analytes.
12,000
A sample of tap water was
10,000 6
analyzed five times using these
calibration curves (five sample 8,000
vials, each prepared by the
PrepStation and analyzed once 6,000
4
with the HP 6890 Micro-ECD). A
4,000
typical chromatogram for the tap 3
water sample is shown in figure 4. 2,000 2
1
The four trihalomethanes were
easily detected. Other unidenti- 0
fied compounds were also present 0 50 100
in the tap water. Amount [ppb]
The quantitative data for the five Figure 3. Calibration Curve for Dichlorobromomethane (Experimental
replicate analyses are summarized conditions are listed in table 1.)
in table 3. The concentration of
chloroform (an estimation extrapo- GC-ECD analysis were found to Table 2. Correlation Coefficients for
Trihalomethanes
lated from the calibration curve) be highly repeatable with a relative
greatly surpassed the World standard deviation of less than Peak Correlation
Health Organization limit. The 10 percent for all compounds. Coefficient
results of the entire analytical The total sample turnaround time Chloroform 0.99048
procedure, including pippeting, was less than 10 minutes. Dichlorobromomethane 0.99933
PrepStation extraction, and Dibromochloromethane 0.99893
Bromoform 0.99791
Hz
1
Peaks
1. Chloroform
800 2
2. Dichlorobromomethane
3. Dibromochloromethane
600 4. Bromoform
400 3
4
200
2 3 4 5 6 7 8 min
Figure 4. Analysis of Trihalomethanes in Tap Water Extract (Experimental
conditions are listed in table 1.)
Table 3. Reproducibility of Trihalomethane Determination
CHC13 (ppb) CHBrCl2 (ppb) CHBr2Cl (ppb) CHBr3 (ppb)
Mean 4605* 11.09 3.86 1.12
RSD (Percent) 5.23 5.41 6.43 8.68
* Estimated by extrapolation of calibration curve
Conclusion References
The new HP 6890 Micro-ECD 1. “Guidelines for Drinking Water
allows very sensitive detection of Quality,” Vol. 1, Recommenda-
trihalomethanes. Detection limits tions, World Health Organization,
below 1 picogram (1 µL of a 1-ppb Geneva, 1984.
sample) were obtained, and the 2. R. Soniassy, P. Sandra, and
detector provided linear response C. Schlett, “Water Analysis,”
in the concentration range of Hewlett-Packard Company
importance. Trihalomethanes Publication (23) 5962-6216E,
in tap water samples can be 1994.
monitored in a fast, simple, 3. I. Temmerman, F. David,
and reproducible way using P. Sandra, and R. Soniassy,
automated micro liquid-liquid Hewlett-Packard Company Hewlett-Packard shall not be liable
extraction. Application Note 228-135, for errors contained herein or for
incidental or consequential damages
Publication (23) 5071-1675E, in connection with the furnishing,
1991. performance, or use of this material.
4. F. David, P. Sandra, and Information, descriptions, and specifica-
S. Stafford, Hewlett-Packard tions in this publication are subject to
Company Application Note change without notice.
228-361, Publication Copyright © 1997
(23) 5965-1611E, May 1996. Hewlett-Packard Company
Printed in USA 4/97
(23) 5965-8014E
