Purge and Trap Analysis Using
Fused Silica Capillary Column Gas Chromatography/
Mass Spectrometry
BY
F. A. Dreisch and T. 0. Munson*
Central Regional Laboratory, Region 3
U.S. Environmental Protection Agency
839 Bestgate Road
Annapolis, Maryland 21401
U. S. Environmental Protection Agency
Environmental Science Center
701 Mapes Road
Ft. Meade, MD 20755-5350
-------
*Current Address:
Chemistry Research Unit
Forensic Science Research and Training Center
F.B.I. Academy
Quantico, Virginia 22135
-------
BRIEF ABSTRACT
Purge-and-trap GC/MS analysis using a fused-si1ica capillary column was
compared to the EPA Method 624 which uses packed columns. The capillary
column method was found comparable for spiked water samples and far superior
for complex environmental samples.
-------
ABSTRACT
Purge-and-trap analysis using fused-silica capillary column gas
chromatography/mass spectrometry (GC/MS) was evaluated for analysis of
priority pollutant organics compared to the EPA Method 624 which uses
packed columns. Using a packed column, our purge-and-trap GC/MS system
generated precision data very similar to that reported by others (5.8%
compared to 5.1%). Preliminary tests of a 30 m SE54 capillary column
system, using the EPA Volatile Organic Performance Evaluation Samples
WP006 and WP007 unknown test materials, indicates satisfactory precision
(8.5%) and excellent accuracy (98.5 and 102 percent recoveries). Not only
was the capillary column much faster than the packed column method (13
min. compared to 33 min.), it also overcame difficulties normally
encountered with samples of excessive complexity, or very large
concentration spread among components, or containing high concentrations
of high boiling components.
-------
INTRODUCTION
Purge-and-trap gas chromatography/mass spectrometry (GC/MS) is a
widely used technique for the measurement of volatile organic compounds in
aqueous samples—particularly for the screening of environmental samples
for the U.S. Environmental Protection Agency's (EPA's) list of volatile
priority pollutants (1,2). We have analyzed several hundred such samples
employing the technique essentially as described in the EPA recommended
Method 624 (2), and while for most samples the method was quite
satisfactory some samples presented serious problems. These problems can
be classified as three types: 1.) excessive complexity--the eluting
compounds overlap extensively making identification difficult;
2.) excessive concentration spread—with the lowest-level components just
above the detection limit, some of the highest-level components cause very
broad peaks obscuring many of the lowest-level components; and 3.) high
concentrations of high boiling components—the gas chromatographic runs
must be lengthened to allow the late peaks to elute, greatly extending the
required analysis time.
Others have reported measuring the EPA volatile priority pollutants by
purge-and-trap GC/MS using high resolution capillary column chromatography
with cryogenic trapping of the desorbed compounds onto the head of the
capillary column (3,4). We have established this method in our
laboratory.* We will describe our version of this technique in this
report, the precision and accuracy of the technique when used for
determining some of the EPA priority pollutants, and how the technique
does overcome the difficulties we encountered when using Method 624 on
certain samples.
* A portion of this work was presented at the 16th Middle Atlantic
Regional Meeting of the American Chemical Society, Newark, Delaware,
April 21-23, 1982
-------
EXPERIMENTAL
Materi als
The GC/MS reference compound 4-bromofluorobenzene was obtained as the
pure compound from PCR Research Chemicals, Inc., Gainesville, Fla. All of
the other compounds were obtained as prepared standard solutions from
Supelco, Inc., Bellefonte, Pa.:
Purgeable A (0.2 g/L in Methanol) Purgeable B (0.2 q/L in Methanol)
methylene chloride
trichlorof1uoromethane
1,1-dichloroethene
(E)-l,2-dichloroethene
1,1-dichloroethane
1,2-dichloroethane
chloroform
1,1,1-trichloroethane
carbon tetrachloride
bromodichloromethane
1,2-dichloropropane
(E)-l,3-dichloropropene
trichloroethene
(Z)-l,3-dichloropropene
1,1,2-trichloroethane
benzene
dibromochloromethane
bromoform
tetrachloroethene
1,1,2,2-tetrachloroethane
chlorobenzene
toluene
ethylbenzene
Purgeable C (0.2 q/L in Methanol) Internal Standards (20 g/L in Methanol)
chloromethane bromochloromethane
dichlorodifluoromethane 1,4-dichlorobutane
bromomethane 2-bromo-l-chloropropane
vinyl chloride
chloroethane
2-chloroethyl vinyl ether (0.2 g/L in Methanol)
Acrolein/acylonitrile (1 g/L in Water)
-------
Other materials were obtained as follows: Tenax GC, 60/80 mesh
(Applied Science Laboratories, Inc, State College, Pa.); activated coconut
charcoal, 40/60 mesh (Analabs, Inc, North Haven, Ct.); and 1% SP1000 on
60/80 mesh Carbopack B (Supelco, Inc).
Organic-free water was generated in the laboratory by passing water
from a Super-Q System (Mi 11ipore Corp., Bedford, Massachusetts) through a
cannister (all glass and metal) containing about 750 gms of activated
charcoal (NUCHAR WV-L, 8/30 mesh, Westvaco, Covington, Va.).
APPARATUS
A Finnigan 3300 GC/MS (Finnigan Instruments, Sunnyvale, Ca.) coupled
to an RDS Dual Computer Data System (RDS Nermag, Santa Clara, Ca.) was
used for this work. The mass spectrometer was operated in the electron
impact ionization mode (70 eV) using instrument settings similar to those
suggested by Budde and Eichelberger (5).
The GC/MS unit was modified by replacing the Finnigan 9500 GC with a
Hewlett-Packard 5840A GC equipped with a capillary injector system
(Hewlett-Packard Co., Avondale, Pa.). The packed column data were
generated using a 2.4 m x 2.1 mm I.D. stainless steel column packed with
1% SP1000 on 60/80 mesh Carbopack B, and the capillary column data were
generated using a 30 m x 0.25 mm I.D. SE-54 fused-silica capillary column
(J&W Scientific, Rancho Cordoba, Ca.). When a packed column was used, the
outlet of the column was connected ,to the mass spectrometer by a glass
-------
jet separator enrichment device and a glass-lined stainless steel transfer
line. The separator oven and the transfer line were kept at about
225°C. In order to connect the capillary column to the mass
spectrometer, the glass jet separator was removed and the end of the
capillary column passed from the GC oven into the separator oven, down
the transfer line into the vacuum manifold with the end of the column
positioned about 1 mm back from the plane of the opening into the ion
volume.
The purge-and-trap procedures and the necessary equipment have been
described in great detail elsewhere (2,6), and only specific operating
parameters and modifications will be included in this report. An
automatic Hewlett-Packard 7675A Purge-and-Trap Sampler was mounted on t-he
HP5840A GC with the heated, rigid portion of the sample line assembly
extending through the top of the GC oven. When a packed GC column was
used, the effluent from the sampler was routed directly to the GC column
by connecting the end of the flexible, 1.6 mm (1/16 in) O.D. stainless
steel tubing portion of the sample line assembly to the top of the GC
column. The sampler trap needed to be desorbed with 15 to 30 ml/min gas
flow. Since the capillary column needed a flow of 0.5 to 1.0 ml/min, the
effluent from the sampler was connected to the capillary column via the
capillary injector, with the injector adjusted to yield about a 20:1
split, when the capillary column was used. The flexible portion of the
sample line assembly was routed back through the top of the GC oven and
connected to a 23 cm length of 3.2 mm (1/8 in) O.D. teflon tubing fixed to
a syringe needle which pierced the septum on the capillary injector. The
trap on the sampler (4.5 mm I.D.) was packed with 4.5 cm of Tenax at the
inlet end followed by 4.5 cm of activated charcoal, which has been found
suitable for trapping all of the EPA priority pollutant volatiles (7).
-------
Cryogenic trapping was necessary to achieve narrow peak widths on the
early eluting compounds, particularly those compounds more volatile than
methylene chloride. The portion of the fused-silica capillary just
downstream of the injector was shaped into a loop by being taped to a rod
extending down from the top of the GC oven. The configuration was such
that, when an insulated 100 ml beaker full of liquid nitrogen was
positioned around the loop, about 14 cm of the capillary column was
immersed in the liquid nitrogen.
For all of the purge-and-trap runs described in this report, a 5.0 ml
sample volume was purged in the large (25 mm x 150 mm) sample tube using
the "deep vortex" technique (8). An egg-shaped magnetic spin bar (20 mm
long) in the bottom of the sample tube just below the terminus of the
purge gas delivery tube was caused to rotate with sufficient speed that a
deep vortex formed all the way to the spin bar. When a sample was purged,
the purge gas passed over the surface of the sample but not through it,
thus eliminating the problem of sample foaming.
For the samples which were heated during the purging process, a small
heated water bath was constructed out of a clear plastic food storage
container (18.5 cm x 14.5 cm x 10.5 cm) with a tight-fitting lid (18.5 cm
x 14.5 cm x 1.5 cm). A Braun Thermomix 1419 heating-stirring unit (SGA
Scientific, Inc, Silver Spring, Md.) was inserted into one end of the
water bath through a hole cut in the lid; another hole was cut through the
lid at the other end to allow the bath to be raised up under the sample
tube on the purge-and-trap device. , With the sample tube inserted
completely to the bottom of the wat'er-bath, the magnetic mixer could be
placed under the water bath to turn the stir bar in the sample tube.
-------
The HP7675A Purge-and-Trap Sampler was designed with a requirement for
fairly high-volume 70 psi air to rotate two pressure-activated valves and
to cool the trap. Because our laboratory lacked centrally-supplied
high-pressure air, the air requirement was initially satisfied using a
pure air generator. After several inconvenient and costly failures of the
air generating system, the air flow system on the 7675A Purge-and-Trap
Sampler was restructured so that the pressure-activated valves were
supplied 70 psi air from a compressed air cylinder, and the trap cooling
system was supplied 15 psi air from the laboratory air system. Although,
thus configured, the trap seemed to be cooled adequately between runs,
additional cooling air was supplied to the exterior of the trap heating
assembly using 6.4 mm (1/4 in) I.D. tubing to supply 15 psi laboratory air
during the trap cooling cycle.
Quality Control
The performance of the GC/MS system was monitored through the daily
use of a GC/MS reference compound (9). The instrument was tuned so that
the purge-and-trap analysis of 50 ng of 4-bromof1uorobenzene (BFB)
provided a spectrum for that compound which satisfied recommended ion
abundance ratio criteria (10). This compound was used as one of the
surrogate compounds which were spiked into each sample analyzed. In this
manner, the tune of the instrument could be verified for each sample run.
The BFB criteria were used as a "benchmark" reference for tuning the GC/MS
system and monitoring its performance, rather than as a performance
requirement which had to be slavishly satisfied prior to sample analysis.
-------
Every purge-and-trap run (method blank, field blank, standards run, or
sample) was spiked with a four-component standard spike:
bromochloromethane, 1,4-dichlorobutane, 2-bromo-l-chloropropane, and
4-bromof1uorobenzene, to give concentrations of 20, 20, 20, and 10 ug/L
respectively. The ion current for a particular ion for each compound
(128, 77, 55, and 176, respectively) was measured for each purge-and-trap
run. On the one hand, these data were used as a diagnostic tool to
monitor the performance of the entire purge-and-trap GC/MS system in terms
of instrument response, assuming that the data from these surrogate
compounds would be applicable to all of the compounds of interest. On the
other hand, one of the compounds (2-bromo-l-chloropropane) was used as the
internal standard for quantitation of the other three compounds (and all
other compounds detected) when compared to their values in the standards
run. The calculated mean concentrations recovered for a particular sample
set, and the relative standard deviations from the mean, provided
precision and accuracy information for the entire analysis (from filling
the syringe through processing the data) for those three surrogate
compounds, and by inference, for all of the compounds of interest. The
mean concentrations and the relative standard deviations were reported in
the Quality Control Summary provided with each analytical report.
Reagent blanks (organic-free water) were run both at the beginning of
each day to demonstrate that the purge-and-trap GC/MS system was
sufficiently free of internal contamination, and whenever it was necessary
to measure possible carry-over of compounds from one run to the next
(e.g., after a sample containing very high concentrations, or after a
standards run). A field blank consisting of organic-free water was
transported to the sampling site, taken back to the laboratory with the
samples, and stored with the samples until the time of analysis. It was
then analyzed with each set of samples.
-------
One of the samples from each sample set was analyzed in duplicate and
the values reported as part of the Quality Control Summary. Occasionally
sample duplicates (5 to 10% of the samples) were spiked with all of the
priority pollutant purgeable standards, and the recoveries reported in the
Quality Control Summary.
QUANTITATION
The only compounds which were routinely quantitated against standards
run on our instrument were the EPA priority pollutant volatile organic
compounds. Each compound was quantitated using the ion current for a
specific ion (usually the ion recommended in Method 624) compared to a<-
single standards run (one concentration) using the internal standards
method as described in Method 624, with 2-bromo-l-chloropropane as the
internal standard. All of the data in this report were generated using
the RDS Automatic Search and Quantitation software program.
In order to identify non-priority pollutant compounds, background
corrected spectra were processed with the Library Search Routine contained
in the RDS SADR Data Reduction Program. A library disk supplied by RDS
Nermag (NBS Library Disk, containing the 35,000 spectra EPA/NIH data base
in abbreviated form) was used as the on-line data base. Using the list of
the ten best-fit compounds found in this manner as a guide, the background
corrected spectrum was compared manually to the spectra displayed in the
EPA/NIH Mass Spectra Data Base volumes (11). A compound was considered
presumptively identified if all of the ions (including those with ion
abundances less than 5%) in the published spectrum were present in the
sample spectrum, and any additional ions in the sample spectrum could be
accounted for.
-------
The concentrations of identified compounds for which standards had not
been run were grossly estimated using the average of the responses of the
compounds in the four component spike.
PROCEDURE
Aqueous samples were collected as specified in Method 624 (2) and
stored refrigerated together with the appropriate field blanks until the
time of analysis. Four replicates of each sample were collected routinely
in case replicate analysis was necessary at some later date. At the time
of analysis, the sample was poured into the barrel of a 5 ml gas-tight
syringe equipped with a two-way valve (in the open position) and a
15 cm 20 gauge needle. With the barrel of the syringe filled to
overflowing, the plunger was inserted, depressed to the 5.0 ml mark and
the valve closed. If a laboratory duplicate of the sample were run, at
this point a second syringe assembly would have been loaded with the same
sample. The appropriate standard solution was added to each sample by
removing the needle from the sample syringe assembly and passing the
needle of the spiking syringe through the two-way valve into the sample
contained in the sample syringe assembly. Every sample was spiked with
5.0 ul of the four component standard spiking solution (an organic-free
water solution containing 20 mg/L of three components and 10 mg/L of
4-bromofluorobenzene). The needle was replaced on the syringe assembly
and the sample transferred to the HP7675A sample tube containing the
magnetic stir bar, care being taken not to atomize the sample. The loaded
sample tube was then mounted on the HP7675A, the deep vortex established,
and the purge-and-trap cycle started.
-------
Fish samples were frozen with solid CO^ in the field as they were
collected and stored frozen until they were prepared for analysis. Each
fish was cut into cubes with a band saw and the cubes coarsely ground with
a sausage grinder. The resulting icy slush was stirred and replicate
subsamples (about 5 gm each) stored frozen in 40 ml VOA vials (Pierce
No.13075, Pierce Chemical Co., Rockford, 111.). At the time of analysis,
ice-cold organic-free water (20 ml) was added to the vial of frozen ground
fish and the vial shaken vigorously until the granules of fish were
suspended in the water. The slurry was then sonicated as described by
Easley, et al. (12), with 20 1 of the 4 component standard spike added
prior to the sonication. A 5 ml portion of the sonicated slurry was
measured with an ice-cold 10 ml graduated cylinder, poured into a HP7576A
sample tube and analyzed according to the procedure for aqueous samples.
For all of the data presented in this report, the
purge-and-trap/desorb sequence was the same: a 1 min pre-purge period
during which cooling air was passed over the trap; an 11 min
purge-and-trap period during which a 40 ml/min flow of helium was passed
sequentially through the purge vessel and trap; a 4 min desorb period
during which the trap was heated to 180°C and backflushed with a 15
ml/min stream of helium (20 ml/min for packed column runs); and a 9 min
vent period during which the trap was baked at 230°C while being
backflushed with 40 ml/min helium.
When the packed column (1% SP1000) was used, the column temperature
was 45°C during the 4 min desorb period while the compounds were being
stripped from the trap and carried to the GC column; then the temperature
was increased to 225°C at 10°C/min,' and maintained at 225°C for 10
min. The temperature was further increased at 20°c/min to 245°C and
maintained for 4 min to sweep late-eluting compounds from the column,
resulting in a total GC run time of 37 min.
-------
When a capillary column was used, the gas stream carrying the desorbed
compounds was split approximately 20:1 in the capillary injection port.
The cryogenic trap was placed around the column about 1 min prior to the
start of the desorb period and removed exactly at the end of the desorb
period. Two GC temperature programs were used with the capillary
column—a short program when late-eluting peaks were not expected to
occur, and a longer program when they were. For the short program, the
column was held at 30°C for 5.5 min (measured from the start of trap
desorb), heated at 2°C/min for 4.5 min, and heated at 10°C/min to a
final temperature of 110°C. The total GC run time was 17.5 min. For
the longer program, the temperatures were programmed exactly the same for
the first 15 min, then the temperature was increased to 150°C at
20°C/min and held for 5 min giving a total run time of 22.5 min.
For the packed column purge-and-trap runs, the data collection by the
GC/MS data system was initiated coincidentally with the start of the
desorb phase. A split mass range (20-27; 33-260 amu) was scanned at 9
msec/amu with a resultant scan rate of 2.4 sec/scan. For the capillary
runs, data collection was initiated 3.0 min after the start of the desorb
phase. The same split mass range was scanned at 0.86 sec/scan for the
initial capillary runs---the fastest scan rate attainable by the GC/MS
data system interface at that time. Subsequently, the interface was
modified by RDS to decrease the settling time, and a scan rate of 0.63
sec/scan was used.
-------
RESULTS AND DISCUSSION
Table 1 shows the results of a precision study for our purge-and-trap
GC/MS system when operated using a packed column. The mean relative
standard deviation (RSD) of 5.8 percent compares favorably with the values
reported by Olynyk, et al. (10) of 5.2, 5.4, and 5.0 percent for selected
groups of these compounds. Their tests did not include six of the most
volatile priority pollutant compounds (chloromethane, bromomethane, vinyl
chloride, dichlorodifluoromethane, chloroethane, and
trichlorofluoromethane), and two compounds of quite poor purging
efficiencies (acrolein and acrylonitrile). Calculating a mean RSD from
the data in Table 1 excluding these compounds yields 5.1 percent, which is
even closer to the reported values.
Figure 1 shows the extracted ion current profile (EICP) from a
fused-silica capillary column purge-and-trap GC/MS run of all of the
priority pollutant purgeable organics compared to the EICP obtained when a
packed column was used. The tremendous resolving power of the capillary
column is obvious--the capillary peaks are so narrow that during
two-thirds of the run time, no peaks were eluting. The potential exists
for adequately resolving mixtures much more complex than this 35 component
mixture. The increased resolution is achieved using a much shorter run
time for the capillary run, only about 13 mins to elute the last compound
compared to about 33 mins for the packed column (the time axis on the EICP
for the packed column includes the 12 min purge time). Better separation
of the individual components in the mixture can be achieved using a longer
capillary run time (and flatter temperature ramps), but because one can
distinguish between overlapping compounds using the specific ions recorded
by the mass spectrometer, we chose to trade increased separation for
shorter run times.
-------
Table 2 presents data which allow an estimation of the precision of
the capillary purge-and-trap method for some of the priority pollutant
purgeable organic compounds. The mean RSD of 8.5 percent which can be
calculated from these data represents only a rough estimate of the
precision of this method for the priority pollutant purgeables because
fewer than one third of the compounds in the mix were represented in the
test, and only three runs were performed. Table 3 presents precision data
from the duplicate analysis by capillary purge-and-trap GC/MS of a "real
world" sample from a steel mill treatment lagoon. In addition to the
compounds shown in Table 3, this sample contained 15 non-priority
pollutant volatile organic compounds in the 1-100 ug/L range.
Table 4 presents data generated during tests of the accuracy of the
capillary purge-and-trap GC/MS method using the EPA Volatile Organic
Performance Evaluation Samples WP006 and WP007 as unknown test materials-
As might be expected from the good mean recoveries achieved (98.5 and 102
percent), all of the values fell within the reported 95% confidence
intervals (13).
Table 5 shows the use of the surrogate standards to provide both
precision and accuracy information for a data set consisting of 16
capillary purge-and-trap GC/MS runs accumulated during the analysis of 6
steel mill samples. The mean precision and accuracy values were quite
satisfactory considering the nature of the samples.
A closer examination of the surrogate data shows reduced recoveries in
2 of the 6 steel mill samples. Table 6A shows the concentrations of the
spiked surrogate standards which were calculated using the internal
standard method, and Table 68 shows the raw data which were used to
generate the values shown in Table 5A. The internal standard method
yielded very acceptable quantitation of the surrogate compounds in all 6
-------
of the samples even though the raw data indicate very poor recovery of the
surrogate compounds from sample 6, and somewhat diminished recovery from
sample 4. These poor recoveries represent "matrix effects" as discussed
by Olynyk, et al. (10). Because both of these samples contained more than
50 compounds, many at quite high concentrations (samples 6 and 4 are
estimated by this purge-and-trap analysis to contain 0.2-2% and 0.01-0.03%
total organics by weight, respectively), the surrogate compounds were
probably more soluble in these sample matrices than in the less
contaminated samples or the organic-free water of the blanks and spikes,
and therefore, purged less efficiently. One might guess from these
results that the internal standard method would adequately correct for
this type of solubility matrix effect for a broad range of organic
compounds which are sparingly soluble in water.
Purge-and-trap analysis of compounds less volatile than
4-bromofluorobenzene is not generally considered practical using the
packed columns suggested in Method 624 (2) due to the excessively long
elution times involved. That such compounds do purge-and-trap is clearly
evident from the late-eluting peaks which frequently occur as lumps in
subsequent runs. The use of capillary purge-and-trap greatly extends the
range of compounds that can effectively be analyzed by the purge-and-trap
technique. Table 7 presents a list of the compounds which we detected
using fused-silica capillary purge-and-trap GC/MS.
-------
Figure 2 presents an EICP (from the purge-and-trap analysis of a fish
tissue sample) which illustrates the power of the method described in this
report. This sample run displays: 1.) the excessive complexity which
could not have been adequately handled by a packed column; 2.) the broad
range of concentrations which would have lead to a loss of valuable
information on a packed column (the small peak just after scan 800
represents about 100 ng); and the high concentration of late eluting peaks
(everything after scan 1050) which would have necessitated long delays
between packed column runs.
-------
Table 1
. Packed
Column Method Precision
Compound3 %
RSDb
Compound9 %
RSDb
chloromethane
5.9
1,2-dichloropropane
4.2
bromoethane
8.8
(E)-1,3-dichloropropene
2.7
vinyl chloride
8.2
trichloroethene
5.5
di chlorodif1uoromethane
8.2
benzene
6.5
chloroethane
7.5
(Z)-l,3-dichloropropene
3.3
methylene chloride
14
1,1,2-trichloroethane
4.3
acrolei n
8.2
dibromochloromethane
4.8
acryloni trile
5.4
2-chloroethyl vinyl ether
5.8
trich1 orof 1uoromethane
11
2-bromo-l-chloropropane (I.S.)
c
1,1-dichloroethene
7.6
bromoform
5-4
bromochloromethane
6.4
1,1,2,2-tetrachloroethane
1.3
1,1-dichloroethane
6.1
tetrachloroethene
4.2
(E)-l,2-dichloroethene
8.4
1,4-dichlorobutane
1.2
chloroform
6.2
toluene
4.7
1,2-dichloroethane
5.5
chlorobenzene
4.9
1,1,1-tri chloroethane
3.0
ethylbenzene
4.7
carbon tetrachloride
4.7
4-bromof1uorobenzene
3.1
bromod i ch1oromethane
4.8
mean RSD = 5.8%
a. All of the compounds were at 20 ug/L except for the following:
chloromethane, bromomethane, vinyl chloride, dichlorodifluoromethane,
chloroethane at 40 ug/L; acrolein and acrylonitrile at 400 ug/L; and
2-chloroethyl vinyl ether at 100 ug/L.
b. Percent relative standard deviation calculated using the data from 5
standard runs
c. This compound was used as the internal standard for the calculations.
-------
Table 2. Precision Study: WP007
Compound
chloroform
1,2-dichloroethane
],1,1-trichloroethane
carbon tetrachloride
tri chloroethene
bromodichloromethane
dibromochloromethane
tetrachloroethene
chlorobenzene
bromoform
methylene chloride
mean RSD = 8.5 + 5.5
Concentration Measured (ug/L)
Sample # .
Rep. 1
Rep. 2
Rep. 3
% RSD
1
7.8
6.4
7.0
9.9
2
54
49
52
4.9
1
3.5
2.8
3.6
13
2
68
64
66
3.0
1
35
30
30
9.1
2
5.4
4.9
4.9
5.7
1
39
32
31
13
2
9.6
8.8
8.8
5.1
1
7.7
6.0
6.0
15
2
25
22
14
28
1
2.1
1.9
1.8
7.9
2
16
16
16
0
1
5.1
4.5
4.9
6.3
2
22
19
19
8.7
1
17
15
14
10
2
36
32
31
8.0
1
5.3
4.8
5.6
7.7
2
26
23
25
6.2
1
32
27
30
8.5
2
2.7
2.6
2.3
8.2
1
16
14
15
6.7
2
3.6
3.8
3.8
3.1
-------
Table 3. Precision in the Analysis of a Sample From
a Steel Mill Treatment Lagoon
Concentration in ug/L
Compound
Sample
Duplicate
% RSD
methylene chloride
18
18
0
1,1-dichloroethene
0.75
1.1
27
chloroform
7.4
7.5
0.95
1,1,1-trichloroethane
34
34
0
1,2-dichloroethane
7.4
7.5
0.95
benzene
1.4
1.8
18
trichloroethene
7.0
6.4
6.3
bromodichloromethane
1.2
1.3
5.7
toluene
28
28
0
tetrachloroethene
3.7
4.0
5.5
chlorobenzene
1.8
1.9
3.8
ethyl benzene
5.1
4.7
5.8
mean RSD = 6.2 + 8.3
-------
Table 4. Accuracy Study
A. WP006 Single Analysis
Compound
Sample No.
Result3
True Value9
% Recovery
1,2-dichloroethane
1
44.8
47.8
93.7
2
13.9
12.8
109
chloroform
1
67.7
74.6
90.8
2
11.0
11.2
98.2
1,1,1-trichloroethane
1
21.4
24.1
88.8
2
3.30
3.21
103
trichloroethene
1
28.0
28.3
98.9
2
3.20
3.33
96.1
carbon tetrachloride
1
46.2
60.2
76.7
2
2.90
3.01
96.3
bromodichloromethane
1
12.5
12.0
104
2
36.1
34.2
106
dibromochloromethane
1
48.1
42.8
112
2
8.10
8.56
94.6
bromoform
1
20.5
18.2
113
2
4.90
4.05
121
methylene chloride
1
6.7
8.37
80.0
2
32.2
33.4
96.4
chlorobenzene
1
37.0
41.4
89.4
2
13.0
13.8
94.2
tetrachloroethene
1
9.60
9.12
105
2
61.0
60.8
100
mean recovery = 98.5 + 10.6
aAll concentration values in ug/L
-------
Table 4. Accuracy Study (con't)
B. WP007 Multiple Analysis (3 runs)
Compound
Sample No.
Result9
True Value3
% Recovery
1,2-dichloroethane
1
3.30
3.19
103
2
66.3
63.8
104
chloroform
1
7.07
7.46
94.8
2
51.6
55.9
92.3
1,1,1-trichloroethane
1
32.1
32.1
100
2
5.07
4.82
105
trichloroethene
1
6.56
6.66
98.5
2
23.2
25.0
92.8
carbon tetrachloride
1
34.1
36.1
94.5
2
9.05
9.04
100
bromodichloromethane
1
1 .94
1.71
113"
2
15.8
15.4
103
dibromochloromethane
1
4.84
4.28
113
2
20.0
19.3
104
bromoform
1
29.9
24.3
123
2
2.54
2.03
125
methylene chloride
1
14.9
16.7
89.2
2
3.74
3.34
112
chlorobenzene
1
5.24
5.52
94.9
2
24.3
27.6
88.0
tetrachloroethene
1
15.4
15.2
101
2
32.8
33.4
98.2
mean recovery = 102 + 9.69
aAll concentration values in ug/L
-------
Table 5. Surrogate Standard Precision and Accuracy*
Concentration Mean Recovered
Compound Spiked (ug/L) Concentration (ug/L) % Recovered % RSD
bromochloromethane 20.0 20.9 104 5.42
1,4-dichlorobutane 20.0 19.9 99.5 4.50
4-bromofluorobenzene 10.0 10.1 101 7.55
*These data are drawn from 16 runs spread over a two-day period; 8 of the runs
were samples (6 sample runs plus 2 duplicate sample runs) from a steel mill,
the remainder were the associated blanks and standards runs.
-------
Table 6. Recovery of Surrogate Standards3
A. Expressed as Concentration (
Sample No. BCM
1 21.6
2 20.7
3 20.5
4 21.1
5 21.9
6 22.9
9/L)
BCP DCB BFB
(20) 20.2 9.64
(20) 20.4 10.0
(20) 19.0 9.94
(20) 20.7 9.34
(20) 19.0 11.1
(20) 20.4 11.7
mean 21.4 (20) 20.0 10.3
% RSD 4.13 --- 3.77 8.88
% Recovery 107 — 100 103
B. Expressed as Area Counts
Sample No. BCM
1 340.0
2 348.2
3 335.6
4 303.1
5 348.2
6 147.2
BCP DCP BFB
696.3 735.2 278.5
704.5 823.3 299.0
686.1 747.5 290.8
606.2 696.3 274.4
667.6 729.1 315.4
270.3 315.4 135.2
meanb 335.0 672.1 746.3 291.6
% RSDb 5.57 5.85 6.30 5.66
a. BCM = bromochloromethane, 128 amu; BCP = 2-bromo-l-chloropropane, 77 amu;
DCB = 1,4-dichlorobutane, 55 amu; BFB = 4-bromofluorobenzene, 176 amu.
b. The values for sample 6 were not included in the calculations of the mean
values or the % RSD values.
-------
Table 7. Compounds Identified Using Purge-and-Trap
Capillary Column GC/MSa
an Number Compound Identified
274 sulfur dioxide
277 hydrogen sulfide
278 chloromethane
281 vinyl chloride
286 butane
287 bromomethane
290 methanethiol
290 chloroethane
298 ethanol
303 trichlorofluoromethane
303 acetone
305 diethyl ether
313 ethanethiol
317 1,1-dichloroethene
317 acrylonitrile
318 methylene chloride
321 thiobismethane
322 1,1,2-trichloro-l,2,2-trifluoroethane
328 carbon disulfide
338 (E)-1,2-dichloroethene
347 2-metboxy-2-methylpropane
353 thirane
354 1,1-dichloroethane
367 2-methyl-2-propanethiol
367 2-butanone
-------
Table 7. Compounds Identified Using Purge-and-Trap
Capillary Column GC/MSa(con1t)
an Number Compound Identified
367
diisopropyl ether
368
pentane
373
2-butanol
379
(Z)-l ,2-dichloroethene
380
ethyl acetate
383
1-propanethiol
383
bromochloromethane
383
chloroform
388
(methylthio)ethane
401
tetrahydrofuran
415
methylthirane
420
2-methyl-1-propanol
420
1,1,1-tri chloroethane
427
1,2-dichloroethane
432
hydrochloric acid
433
1-bromo-1-chloroethane
438
cyclohexane
443
carbon tetrachloride
444
benzene
458
2-(methylthio)propane
458
2-butanethiol
494
2-pentanone
508
3-pentanone
509
1,2-dichloropropane
511
heptane
-------
Table 7. Compounds Identified Using Purge-and-Trap
Capillary Column GC/MSa(con't)
Scan Number Compound Identified
513 dibromomethane
515 trichloroethene
524 bromodichloromethane
525 1 , 1 '-thiobisethane
531 1,4-dioxane
552 methylcyclohexane
585 2-chloroethyl vinyl ether
601 (Z)-1,3-dichloropropene
605 4-methyl-2-pentanone
620 dimethyl disulfide
661 toluene
661 (E)-l,3-dichloropropene
675 1,1,2-trichloroethane
688 2-bromo-l-chloropropane
691 5-hexen-2-one
712 dimethylcyclohexane (isomer)
713 3-hexanone
730 dibromochloromethane
745 dimethylcyclohexane (isomer)
765 dimethylcyclohexane (isomer)
769 tetrahydrothiophene (T)*3
773 tetra'chloroethene
790 butyl acetate
835 ethylcyclohexane (isomer)
844 ethyldimethylcyclohexane (isomer)
-------
Table 7. Compounds Identified Using Purge-and-Trap
Capillary Column GC/MSa(con1t)
Scan Number Compound Identified
864 chlorobenzene
877 2-methyl tetrahydrothiophene (7)13
895 ethylbenzene
915 xylene (isomer)
950 bromoform
958 2,5-dimethyl tetrahydrothiophene (isomer)
968 2,5-dimethyl tetrahydrothiophene (isomer)-
971 xylene (isomer)
975 1 ,3,5,7-cyclooctatetraene (7)13
990 nonane
1007 1,4-dichlorobutane
1020 1,1,2,2-tetrachloroethane
1033 1,2,3-trichloropropane
1041 4-bromofluorobenzene
1061 bromobenzene
1106 chloromethylbenzene (isomer)
1112 propyl benzene
1142 trimethylbenzene (isomer)
1148 dimethyltrisulfide
1171 benzonitrile
1182 6-methyl-5-hepten-2-one
1189 trimethylbenzene (isomer)
1199 decane
1211 butylbenzene
-------
Table 7. Compounds Identified Using Purge-and-Trap
Capillary Column GC/MSa(con't)
Scan Number Compound Identified
1215 1,3-dichlorobenzene
1232 1 ,4-dichlorobenzene
1237 trimethylbenzene (isomer)
1258 2,3-dihydroindene
1271 lH-indene
1273 1-propynylbenzene
1282 1,2-dichlorobenzene
1289 diethylbenzene (isomer)
1294 tetramethylbenzene (isomer)
1304 methylphenol (isomer)
1307 methylpropylbenzene (isomer)
1324 dimethylethylbenzene (isomer)
1334 dimethylethylbenzene (isomer)
1344 methylphenol (isomer)
1345 nitrobenzene
1359 methylbenzoate
1379 tetramethylbenzene (isomer)
1386 tetramethylbenzene (isomer)
1408 3,5-dimethyl-1,2,4-trithiolane (isomer)
1418 3,5-d'imethyl-1,2,4-trithiolane (isomer)
1446 1,2,3,4-tetrahydronaphthalene
1478 naphthalene
1479 trichlorobenzene (isomer)
-------
Table 7. Compounds Identified Using Purge-and-Trap
Capillary Column SC/HSa{cori't)
Scan Humber Compound Identified
1523 trichlorobenzene (isomer}
1S95 1,2,314-tetrahydromett)ylnaphthaleri6 (isomer)
1644 methyl naphthalene ("isomer)
l£?3 methylncphthalene {isomer)
1718 tetrachlorobenzene (isomer)
1 SO? tetrachlorobenzene (isomer)
a. Tae non-priority pollutant compounds otter than the cooipouids
listed fn Table 1) were identified with a computer-based matching
system using the EFA/lflH 'library of spectra-
fa. The parenthetical symbol indicates that the reference spectrum had
some minor fragments not found in the sample spectruw.
-------
REFERENCES
1. Sampling and Analysis Procedures for Screening of Industrial Effluents
for Priority Pollutants. U.S. Environmental Protection Agency,
Environmental Monitoring and Support Laboratory, Cincinnati, Ohio,
March, 1977, revised April, 1977, pp. 1-15.
2. Guidelines Establishing Test Procedures for Analysis of Pollutants,
Proposed Regulations; Purgeables-Method 624. Fed. Reg. 44:
69532-69539 (1979).
3. F. DeWalle, and E. Chian. Presence of Priority Pollutants in Sewage
and Their Removal in Sewage Treatment Plants: Grant 806102, S.
Hannah, Project Officer. U.S. Environmental Protection Agency,
Municipal Environmental Research Laboratory, Cincinnati, Ohio (Draft).
4. A. R. Trussel, J. G. Moncur, F-Y Lieu and L. Y. C. Leong.
Simultaneous Analysis of All Five Organic Priority Pollutant
Fractions. J.. High Res. Chrom. Chrom. Commun. 4: 156-163 (1981 ).
5. W. L. Budde and J. W. Eichelberger. Organics Analysis Using Gas
Chromatography/Mass Spectrometry. Ann Arbor Science Publishers, Inc.,
Ann Arbor, Michigan, 1979, p. 12.
6. Operating and Service Manual: Purge and Trap Sampler 7675A. Hewlett
Packard Co., Avondale, Pa., (1979).
-------
7. D. F. Bishop. GC/MS Methodology for Priority Organics in Municipal
Waste Water Treatment, EPA 600/2-80-196, U. S. Environmental
Protection Agency, Municipal Environmental Research Laboratory,
Cincinnati, Ohio (1980).
8. J. Poole and W. D. Snyder. A Technique for the Reduction of Foaming
in Purge and Trap Samples, Application Note AN 228-13, Hewlett Packard
Co., Avondale, Pa.
9. J. W. Eichelberger, L. E. Harris and W. L. Budde. Reference Compound
to Calibrate Ion Abundance Measurements in Gas Chromatography-Mass
Spectrometry Systems, Anal. Chem. 47: 995-1000 (1975).
10. P. Olymyh, W. L. Budde and J. W. Eichelberger. Simultaneous
Qualitative and Quantitative Analyses. I. Precision Study of
Compounds Amenable to the Inert Gas-Purge-and-Trap Method. J. Chrom.
Sci. 19: 377-382 (1981).
11. S. R. Heller, G. W. A. Milne, editors. EPA/NIH Mass Spectral Data
Base, Volumes 1-4, 1978; Supplement 1, 1980. U.S. Superintendent of
Documents, Government Printing Office, Washington, D.C.
12. D. M. Easley, R. D. Kleopfer, and A. M. Carasea. The Analysis of
Volatile Organic Compounds in Fish. J. Assoc. Off. Anal. Chem. 64:
653-656 (1981).
13. P. Britton, Environmental Monitoring and Support Laboratory, U. S.
Environmental Protection Agency, Cincinnati, Ohio, Personal
Communication, 1981.
-------
Acknowledgement
The authors wish to thank Janet Roberson for diligently assembl
this manuscript.
-------
Tt^ocr: . JL
TJIL£= PURCEMLE SIANDARDS ON 3011 SE-54 FS CAPILLARY COLUMN
fJl£: OB .DA ACQUISITION IAfil£= CAPTD3 .AT DATE: 1/2171981
100Z= 3260416 XI
50.00 - 260.00
TIME- " 2^04 3=03 4=14
6:24 1:29 8:34 9=39
TITLE: PURGEAfiLE STANDARDS ON PACKED COLUMN SP-1000/CARB0PACK 0
FILE: 29 .DA ACQUISITION TABLE: PURGE .AT DATE: 4/3/1981
50.00 - 260.00
TIME
-------
TJIi£ - nsfi SAflPlt 6Y PURSL AND TRAP
FILL- 35 ACGUJSIIiOJt TABL£ HAfl£= CAPTM mi- S/30/138J
r>/?£/t,c i'
^ *• .
/ V ^
1007. 3G")llSSf x:
SO.00 - 260.0)
SCAf(
1000
-------
Purge and trap analysis
using fused silica
capillary column gas
EJDM OD79 C45J60
------- |