Environmental Monitoring Series
Isolating Organic Water Pollutants:
XAD Resins, Urethane Foams,
Solvent Extraction
I
55
O
\.
SE
ul
CD
National Environmental Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Corvallis, Oregon 97330
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development,
U.S. Environmental Protection Agency, have been grouped into
five series. These five broad categories were established to
facilitate further development and application of environmental
technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in
related fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL MONITORING STUDIES
series. This series describes research conducted to develop new
or improved methods and instrumentation for the identification and
quantification of environmental pollutants at the lowest conceivably
significant concentrations. It also includes studies to determine
the ambient concentrations of pollutants in the environment and/or
the variance of pollutants as a function of time or meteorological
factors.
EPA REVIEW NOTICE
This report has been reviewed by the National Environmental
Research Center—Corvallis, and approved for publication. Mention
of trade names or commercial products does not constitute endorsement
or recommendation for use.
-------
EPA-660/4-75-003
June 1975
ISOLATING ORGANIC WATER POLLUTANTS:
XAD RESINS, URETHANE FOAMS, SOLVENT
EXTRACTION
By
Ronald G. Webb
Southeast Environmental Research Laboratory
National Environmental Research Center
Athens, Georgia 30601
ROAP 16ADN Task 37
Program Element 1BA027
NATIONAL ENVIRONMENTAL RESEARCH CENTER
OFFICE OF RESEARCH AND DEVELOPMENT
U. S. ENVIRONMENTAL PROTECTION AGENCY
CORVALLIS, OREGON 97330
For Bale by the Superintendent of Documents, U.S. Government
Printing Office, Washington, D.C. 20402
-------
ABSTRACT
Isolation, separation, and concentration into an organic
solvent are generally required prior to identification and
quantitation of ~6rganic pollutants in water by gas
chromatography or mass spectrometry. These operations can
be simplified or improved by the use of XAD-resins
(macroreticular resins) and by changes in solvent extraction
procedures. XAD-2,4,7, and 8 and mixtures of these resins
effectively extracted a broad range of individual industrial
pollutants and mixtures typical of paper mill wastewaters,
dissolved fuel oil, and textile dyes. Resin recovery
efficiencies were typically 65-15% for individual compounds;
direct chloroform extraction efficiency was 80%. Polyure-
thane foams were not effective for extracting these
compounds. Chloroform is generally recommended over diethyl
ether as an extraction solvent. Drying of chloroform
extracts before evaporation was shown to be unnecessary.
For typical industrial effluents, extract concentration to
10 ml with a Kuderna-Danish evaporator and to as low as 0.3
ml with a micro-Snyder column is the most quantitative
procedure. Extraction with tetralin sometimes allows
detection of nonpolar low-boiling pollutants that are
usually obscured in gas chromatographic analysis by the
solvent peak.
This report was submitted in partial fulfillment of ROAP
16ADN Task 37 at the Southeast Environmental Research
Laboratory, Athens, Georgia. Work was completed by July
1974.
-------
CONTENTS
Sections Page
I Conclusions 1
II Introduction /' 2
/
III Macroreticular Resins 3
Industrial Effluents 3
Fuel Oil 6
Textile Dyes 7
Operational Notes 7
IV Polyurethane Foams 9
Paper Mill Wastewater Components 9
Fuel Oil 10
Textile Dyes 10
V Solvent Extraction 11
Solvents Used 11
Drying Chloroform Extracts 11
Concentration of Organic Extracts 13
First Stage Evaporation 13
Final Evaporation 15
Block Tube Heater 16
Micro-Snyder Column 16
High-boiling Solvents 17
VI References 19
111
-------
SECTION I
CONCLUSIONS
XAD resins extract a broad range of organic compounds from
water with slightly lower efficiency than direct chloroform
extraction.
The resins may be useful for long-term or composite
sampling.
Polyurethane foams are not useful adsorbents for many
compounds found in industrial effluents although they
function well for PCB's.
Either chloroform or methylene chloride is generally
preferable to diethyl ether or hexane as an extraction
solvent.
Drying chloroform extracts with sodium sulfate or glass wool
results in lower recoveries than direct concentration
without dryin g.
The recoveries of dissolved organics are greater when
extracts are concentrated in two stages using Kuderna-Danish
and micro-Kuderna-Danish equipment rather than rotating
evaporators and airstream-waterbath methods.
Extraction with tetralin sometimes allows detection of low-
boiling materials that are usually obscured in gas
chromatographic analysis by the solvent peak.
-------
SECTION II
INTRODUCTION
During the last five years, great advances have been made in
techniques for identification of small amounts of organic
compounds in water. Progress, however, has not kept pace in
methods to isolate, separate, and quantitate these
materials. Two areas in which research is needed are the
development of alternative methods for carbon adsorption for
long-term sampling, and improvement in recoveries in solvent
extraction. To this end, macroreticular resins and urethane
foams were investigated as adsorbents for the isolation of
organic pollutants from water. Also, the procedures and
reagents for solvent extraction were examined for
inefficiencies and changes are proposed in the usual
procedures for drying and evaporation.
-------
SECTION III
MACRORETICULAR RESINS
Solvent extraction and carbon filters both have disad-
vantages for extracting large amounts of water containing
small amounts of organic pollutants. For solvent
extraction, the labor required for large samples is too
great; and carbon filters give incomplete recovery.
Recently, macroreticular resins have been proposed for this
application.*
Macroreticular refers to the relatively large, controlled
pore size of the resin beads. Each grain of resin is formed
from many microbeads cemented together during the
polymerization process. The pores are the spaces left
between the cemented-together microspheres. The most
popular macroreticular resins are the XAD series made by the
Rohm and Haas Co. They are hard insoluble beads of 20-50
mesh, varying from white to light brown in color. Four
materials are available, XAD-2,4,7,and 8. XAD-2 and 4 are
styrene-divinyl benzene copolymers. XAD-2 has an average
area of 330 m2/g and a nominal pore size of 90 angstroms;
XAD-4 has more surface area (750 m2/g) and smaller pores (50
angstroms).2 XAD-7 and 8 are chemically similar, both being
acrylate esters, but have different physical charac-
teristics; XAD-7 has a 750 m2/g area and 80 angstrom pores
and XAD-8 has an area of 140 m2/g and a pore size of 250
angstroms.
Adsorption on the surface of the resin is the basis for the
separations; no ion exchange mechanisms are involved. In
practice, the water sample is passed through a column of XAD
resin and the pollutants are adsorbed. Then the organics
are desorbed from the resin by elution with a small amount
of organic solvent, which is analyzed by gas chromatography
or GC-MS.
INDUSTRIAL EFFLUENTS
To be of general utility in water pollution analysis, the
resins must be capable of extracting a wide variety of
contaminants. A mixture containing 13 materials previously
identified in industrial effluents3 was used to test the
resins' extraction ability. To one liter of distilled water
was added 1 ml of acetone that contained all the test
compounds, each at a concentration of 50 micrograms/ml. The
final concentration of each pollutant in the water was
therefore 50 micrograms/1 (50 ppb). After thorough mixing,
the solution was passed through a 1.5 cm x 3 cm column
-------
(i.e., 5 ml of resin) at a flow rate of 20 ml per min (4 bed
volumes per min).* The pollutants were eluted from the
resin with acetone followed by chloroform. The combined
eluate was separated from a small amount of water,
concentrated in a Kuderna-Danish evaporator and analyzed by
GC using a computer-assisted quantitation system. All
experiments were done at least in duplicate. The percent
recovery of each compound is shown in Table 1.
For this investigation, the best resin was defined as the
one that gave the largest value when the recoveries of the
individual test compounds were averaged. Eight of the nine
resins gave average recoveries of 65-76%, the best being
from an equal mixture of XAD-U and XAD-8. XAD-7 was
significantly poorer with an average of only 51%. These
recoveries should be compared with an average recovery of
80% for direct chloroform extraction of a similar sample.
The difference in recoveries for each compound in duplicate
resin extractions was generally less than 10%.
Solvent extraction of the aqueous effluents from the columns
revealed measurable amounts of eight of the thirteen test
compounds; the phenols, hexadecane, alpha-terpineol and
dibenzofuran were in highest concentration. However,
although significant amounts of phenols were found to pass
through the column, the efficiency of some resins for
extraction of phenols was greater than for direct solvent
extraction. Paraffins, on the other hand, are more poorly
extracted by the resins than by direct solvent extraction,
as confirmed by later experiments with fuel oil.
In one experiment, XAD-48 was allowed to completely air-dry
in the column for three days before elution. The recoveries
from the dry column were only about 12% lower than those
from immediate extraction of the wet resin. This suggests
that class separations and clean-up of complex mixtures
could be made by elution with solvents of varying
polarities.
In addition to the classes of compounds in Table 1,
alcohols, ketones, acids, and esters are frequently found in
industrial effluents. Limited experiments indicate that the
resins are reasonably effective in extracting these
materials, as shown in Table 2.
-------
Table 1. TABLE OF RECOVERIES
Percent Recovery from 50 pg/1 Samples
COMPOUND
bis-Chloro isopropyl ether
sym-Tetrachloroethane
n-Hexadecane
alpha-Terpineol
Naphthalene
o-Nitrotoluene
2 -Methyl naphthalene
1-Methyl naphthalene
Benzothiazole
Phenol
p-Cresol
Acenaphthene
Dibenzofuran
Averages excluding
n-Hexadecane
XAD Resin
7
c
35
3
36
64
53
63
64
40
19
33
72
73
51
24a
74
58
c
76
66
75
61
62
80
30
58
68
70
65
8
77
59
c
62
78
77
77
80
53
29
47
20
95
69
27a
76
66
8
77
77
79
72
75
75
32
50
81
83
70
2
76
61
3
81
79
82
75
76
74
14
44
99
93
71
28a
77
68
18
75
81
81
80
82
73
33
49
85
86
72
2478a
71
68
14
75
82
81
81
80
77
41
60
84
85
74
4
80
72
c
80
80
83
77
77
82
38
69
81
82
75
48a
77
72
11
8.0
80
83
77
79
82
46
68
81
84
76
CHCl3b
92
82
36
92
87
91
86
86
96
19
50
91
92
80
a - Mixture of equal dry weights of each resin.
b - Sample directly extracted with two 50-ml portions.
c - Peak unsuitable for accurate quantitation.
-------
Table 2. POLAR COMPOUNDS ADSORBED ON XAD RESINS
Percent Recovered from 100 yg/1 Sample
XAD Resin
Compound 2 4 7 8
2-Ethylhexanol
Isophorone
Pentachlorophenol
Palmitic acid
Dehydroabietic acid
2-Ethylhexyl phthalate
85
76
84
67
94
33
91
86
84
79
86
11
74
46
83
12
31
22
79
47
77
16
85
13
Average 73 73 45 53
Results were not as reproducible as the earlier studies
because of losses in the extra analytical step required
(esterification of the eluted acids with diazomethane before
GC) and problems with the GC column. The consistently
better recovery of dehydroabietic acid, a resin acid, over
palmitic acid, a fatty acid, is possibly due to the affinity
of the resins for molecules having an aromatic ring. The
adsorption of palmitic acid probably would have been more
efficient if the test solution had been acidified before
adsorption as Junk5 suggests.
FUEL OIL
The resins were found to be unacceptable for sampling
natural waters for oil residues. They sorb the polar and
aromatic portions of the oil but they do not remove the
aliphatic hydrocarbons that constitute the bulk of most
oils. Solvent extraction with carbon tetrachloride removes
all the oil components from water and is the preferred
method for quantitative analysis.
When No. 2 Fuel Oil (home heating oil) is dispersed in water
and the mixture is extracted with carbon tetrachloride, the
gas chromatogram of the extract is identical to that of the
original oil. In particular the regularly spaced peaks of
the straight-chain hydrocarbons are obvious. When a similar
mixture is passed through the XAD resins and processed in
the usual manner, the chromatographic peaks of the normal
hydrocarbons are very small. The peaks corresponding to
polar and aromatic materials are of similar intensity as in
the original oil. There is no question that the normal
hydrocarbons pass through the column because they can be
recovered from the effluent by solvent extraction.
-------
TEXTILE DYES
The author of an EPA sponsored study6 on photodegradation of
textile dyes in waste streams pointed out that many of these
dyes are difficult to detect when mixed with other colored
wastes because they are not extracted from water with the
usual organic solvents. Six of his test dyes that do not
extract with ether, chloroform, or hexane were tested for
extraction by XAD-2 and XAD-7. Dyes tested were Vat Blue 6,
Acid Red 18, Direct Red 83, Acid Red 37, Basic Blue 9, and
pararosaniline. In general, a few grams of XAD-2 resin
would remove all the visible color from 200 ml of 1-10 mg/1
solutions of the dyes. Adjustment of the solutions to acid
pH was required for the sulfonic acid dyes. The dyes could
be eluted with ethanol or aqueous base. XAD-7 was less
effective for all dyes, particularly in the elution step.
Only Vat Blue 6, an organic base, was not extracted by the
resins.
OPERATIONAL NOTES
• The commercial XAD resins require extensive pre-
extraction to remove interfering impurities. The
best technique seems to be washing large batches in
a column with acetone, methanol, and either
chloroform or methylene chloride. Batch extraction
in shake flasks was not as efficient.
• A water flow rate of U bed volumes per minute
results in greater extraction efficiency than a rate
of 12 bed volumes per minute.
• Direct elution with chloroform or carbon
tetrachloride was not efficient. Treatment of the
column with a solvent such as acetone or methanol,
which is soluble in both water and water-immiscible
solvents, elutes the water from the interior of the
resin and makes the resin wettable by the chloroform
or carbon tetrachloride. On the other hand. Junk et
al.s recommended the direct addition of diethyl
ether to the wet column followed by a 10-minute
penetration period before drawing off the solvent.
He reports excellent recoveries by this procedure.
• Elution by acetone followed by chloroform was more
efficient than elution by acetone followed by carbon
tetrachloride.
-------
Two bed volumes of acetone followed by eight of
chloroform was as efficient as five of acetone
followed by five of chloroform.
Acetone contains trace amounts of diacetone alcohol
and mesityl oxide that are found in the final
extract.
After a resin has been eluted with chloroform it can
be washed with acetone and then with water and then
used to extract another water sample. Recoveries
from re-used columns were not significantly
different from the original values. No irreversible
adsorption or saturation effect therefore took place
during the first cycle.
-------
SECTION IV
POLYURETHANE FOAMS
Polyurethane foam has many everyday uses including seat
padding for automobiles and furniture. In the biological
laboratory it sometimes replaces cotton balls as closures
for culture flasks. In a recent paper7, Gesser et. al.
report that the porous foam can be used to extract
polychlorinated biphenyls (PCB's) from water. If the foams
are coated with various GC liquid phases such as DC-200,
they also extract the DDT class pesticides with high
efficiency.8
The technique is very similar to that for the XAD-resins. A
plug of foam is placed in a column chromatography column,
the water sample is passed through, and later the column is
stripped of absorbed components with a solvent. In the
experiments outlined below, a gram or two of foam was used,
usually with a plug about 20 mm in diameter. Two brands of
foam were first tested for extraction of PCB's to be sure
they behaved as the materials reported7 and then experiments
were done with paper mill wastewater components, fuel oil,
and textile dyes. Foam stoppers from Gaymar Industries
extracted PCB's from water at 20 micrograms/liter with the
same efficiency (84%) as direct extraction with hexane.
Foam plugs sold by Analabs, Inc. for PCB extraction also
worked well.
PAPER MILL WASTEWATER COMPONENTS
One-liter volumes of test mixtures containing 5 mg/1 each of
typical materials found in paper mill effluents (camphor,
fenchone, fenchyl alcohol, alpha-terpineol, guaiacol,
phenol, and para-cresol) were extracted with the foam. In
general, no more than 10% of the material was extracted by
the foam plug since 90% of the original material was found
on analysis of the effluent.
To determine whether the 5 mg of each of the materials had
overloaded the column, a one-liter sample containing only 20
micrograms of each solute was extracted. Still no more than
10% of any of the materials was extracted.
The effect of increased contact time was tested by shaking
the plug in a jar of the stock eresol solution for several
hours. Extraction efficiency was increased only to 20%.
The inefficiency of extraction by the urethane foam
therefore is probably not a result of insufficient contact
time.
-------
Foam plugs were coated with 3% DC-200, DECS, or Carbowax 20M
and a one-liter sample containing 20 micrograms of each
solute was extracted. The extraction removed only 10 to 40%
of each test compound.
In contrast, the macroreticular resin XAD-2 removed nearly
100% of this test mixture at 0.2 mg/1 level and 65% to 80%
of each compound was recovered on elution of the resin with
acetone and carbon tetrachloride. XAD-7 resin was less
satisfactory, with 30% to 45% of the material being
extracted and 13% to 40% recovered from the resin.
FUEL OIL
Although urethane foams act as sponges or wicks for oil
floating on water, our tests show that for dispersed oil, as
opposed to oil slicks, the foams alone or coated- with DC-200
are largely ineffective as extraction media. Only 9% of the
oil was extracted when a liter of water with 1.6 mg/1 No. 2
fuel oil dispersed in it was passed through a 0.7-g foam
plug. In a similar experiment using 2.2 g of 4% DC-200
coated foam, 80% of the oil was removed from the water, but
only 34% was recovered from the plug by elution with carbon
tetrachloride.
TEXTILE DYES
The six textile dyes mentioned earlier that will not extract
directly from water with the usual solvents were tested for
extraction by polyurethane foam. In general, only part of
the color was removed from 200 ml of 1-10 mg/1 solutions of
the dyes by a gram of foam. The dye that was removed could
be only partially recovered from the foam by elution with
ethanol even though ethanol dissolves the pure dyes.
Alcoholic sodium hydroxide removed the dyes but also
disintegrated the foam.
These results indicate that the foams are very limited in
their extraction ability. The optimism expressed in a
recent paper9 for their promise as an alternative to carbon
adsorption was premature.
10
-------
SECTION V
SOLVENT EXTRACTION
The most common method for isolating organic water
pollutants is direct extraction with a solvent. With the
exception of polymers and complex biological molecules, most
organic chemicals can be extracted from water by a judicious
choice of conditions.
SOLVENTS USED
With EPA analysts, the two most popular solvents for
extraction are methylene chloride and chloroform. Diethyl
ether, hexane, or mixtures of these solvents10 are also used
frequently. Methylene chloride and chloroform are
efficient solvents for a broad range of compounds. They are
reasonably insoluble in water so the volumes recovered are
nearly equal to those added; they do not require drying with
inorganic salts to remove water; their density, which is
greater than that of water, makes multiple extractions
possible with a minimum of manipulation; they evaporate
easily but can be stored without further loss for reasonable
periods of time; and they give acceptably narrow solvent
peaks on flame detector GC's.
Ether and hexane are poorer in all of these respects. Ether
was compared in efficiency with chloroform for extraction of
short (C^-C,) fatty acids, phenol, Q-cresol, and 2-ethyl-1-
hexanol. With the exception of phenol, chloroform was the
better extractant for each compound. The high solubility of
ether in water may be partly responsible for the relatively
low efficiency of the ether extraction. When one liter of
water was extracted with two 50-ml portions of chloroform,
U2 ml and 49 ml of the chloroform were recovered for each
fraction, for a total solvent loss of only 9%. In contrast,
100 ml of ether must be added to a liter of water before any
separable layer develops. Then when two additional 50-ml
portions of ether are used for the extraction, a total
volume of only 98 ml is recovered. The "aqueous" layer is
then 10% ether by volume and must retain some of the
organics.
DRYING CHLOROFORM EXTRACTS
After a water sample is extracted with an organic solvent,
the solvent is usually dried to remove dissolved water and
then evaporated to a smaller volume before analysis. As
shown below, for quantitative recovery of many common
industrial pollutants, there is no advantage in drying a
11
-------
chloroform extract. Fear of losses due to a steam
distillation effect in undried extracts is unfounded.
Identical liter samples of water containing 33 micrograms of
each of 13 industrial pollutants were extracted with two 50-
ml portions of chloroform by the usual separatory funnel
technique. The extracts were dried by passing them through
a short column of anhydrous sodium sulfate or a short column
of glass wool pre-wet with solvent. The extracts were
evaporated to about 6 ml in a Kuderna-Danish apparatus and
then to 1 ml by blowing a stream of nitrogen over the
sample. An undried extract and a reference sample of 100 ml
of chloroform spiked with 33 micrograms of each compound
were concentrated similarly. The experiments were done in
duplicate and each sample was analyzed by GC and quantitated
by the computer-assisted data system. The results,
tabulated as percent recovery of the amount added to the
water, are shown below (listed in order of eluti-on from a
Carbowax 20M-TPA column):
Na2SO4 Glass Wool Undried Evap. Ref.
bis-Chloro isopropyl 87 82 92 91
ether
sym-Tetrachloroethane 76 70 83 89
n-Hexadecane 29 23 36 93
alpha-Terpineol 86 81 92 91
Naphthalene 82 76 87 91
o-Nitrotoluene 84 81 91 93
2-Methyl naphthalene 80 75 86 93
1-Methyl naphthalene 81 77 86 93
Benzothiazole 85 83 96 94
Phenol 19 17 19 91
p-Cresol 46 44 50 83
Acenaphthene 85 82 91 93
Dibenzofuran 84 83 92 92
The last column is the maximum amount that could be
recovered if extraction were 100% efficient and indicates
the magnitude of losses on evaporation to one ml. Low
recoveries of hexadecane, phenol, and p_-cresol are due to
the inefficiency of extraction rather than the effect of
drying. The undried sample gave the best overall results.
Recoveries from duplicate experiments differed by an average
of only two percentage points for the undried sample. For
the sodium sulfate dried sample, the average difference was
6% and for the glass wool samples, it was 16%. The
evaporation technique was very reproducible with an average
12
-------
difference in results of 2%. Even when undried extracts
were evaporated to volumes as small as 0.1 ml, no water
separated out and recoveries did not differ significantly
from those of samples that had not been in contact with
water.
A separate experiment showed that undried extracts using
hexane or 15X methylene chloride in hexane as the solvent
can be analyzed by electron-capture GC without any
discernible change in the detector's performance. Samples
obtained by extraction of a liter of water with two 50-ml
portions of solvent were concentrated with a Kuderna-Danish
evaporator to 10 ml and further reduced to 1 ml by the
airstream method. Analysis with a Ni-63 detector on a
pesticide column showed immediate return to the baseline
after elution of the solvent for both undried and sodium
sulfate dried samples. Even concentrates with a visible
layer of water chromatographed in a normal fashion when the
organic layer was sampled.
CONCENTRATION OF ORGANIC EXTRACTS
The common methods for evaporation of organic extracts were
also examined. The usual procedure is to concentrate from
the original volume (100 ml or more) to 5 or 10 ml in one
step and to further reduce the 10-ml volume to about 1 ml in
a second step.
First Stage. Evaporation
Three methods of evaporation were tested with 100-ml
portions of chloroform containing 33 micrograms of each of
13 known pollutants. The beaker method was tested by
warming the sample in a 250-ml beaker on a steam bath while
directing a gentle stream of air on the surface of the
sample. The Rotavapor samples were evaporated from a 250-ml
round-bottom flask at a waterbath temperature of 30°C and a
pressure of 450 mm mercury. The K-D samples were processed
in a 24/25 size evaporator (Figure 1). Each sample was
evaporated to 5-10 ml and then transferred to a K-D tube and
reduced to 1 ml by the airstream-waterbath method. The
experiments were done in duplicate (K-D in quadruplicate).
13
-------
J
Figure 1. Kuderna-Danish
Evaporator
Figure 2. Micro-Snyder
Column Evaporator
14
-------
The average percent recovery of each compound is shown
below:
Evaporation Method
Compounds Beaker Rotavapor K-D
bis-Chloro isopropyl ether — 84 91
sym-Tetrachloroethane 76 73 87
n-Hexadecane 88 89 92
alpha-Terpineol 82 85 90
Naphthalene 85 86 91
o_-Nitrotoluene 86 86 91
2-Methyl naphthalene 92 90 92
1-Methyl naphthalene 84 86 91
Benzothiazole 82 88 91
Phenol 83 85 90
p-Cresol 78 82 85
A~cenaphthene 86 89 91
Dibenzofuran 87 88 92
Each of the methods was reproducible, with duplicate runs
showing differences of only about 5 % in recovery for
individual compounds. The average recovery of all 13
compounds was 84% for the beaker method, 85% for the RotoVap
and 9035 for the K->D. Although the K-D method is best, the
RotoVap and beaker methods are acceptable. In particular,
the beaker method is not nearly as poor as expected. The K-
D method is actually nearly 100% efficient since, as shown
next, losses of about 10% occur in the final evaporation to
1 ml.
Zisai Evaporation^ Airstream-Waterbath Method
Three methods were tested for the final evaporation step,
i.e., evaporation to 1 ml or less. The most used technique
is the airstream-waterbath method. Each of eleven people
evaporated a 5-ml sample and a 10-ml sample of chloroform,
each containing 33 micrograms of the compounds shown above.
The samples were all in standard K-D tubes and were all
evaporated to 1 ml. The average recovery for all compounds
in the 10-ml samples was 90%; in the 5-ml samples it was
92%. The longer exposure time to the air stream was
expected to result in significantly lower recoveries for the
10-ml sample, but this was not the case. For 10 out of the
11 people the recovery for each compound was 85-95%. (The
other person had recoveries about 12% lower.) The compounds
for which more than 10% was lost were tetrachloroethane,
13%; o-nitrotoluene, 13%; benzothiazole, 14%; and p-cresol,
17%. "with this set of test compounds, there was no
15
-------
difference in the recoveries when air was substituted for
nitrogen as the evaporating gas.
Block Tube Heater
A variation of the airstream evaporation that is sometimes
used in pesticide analysis is a specially designed aluminum
block heater.11/12 The concentrator tube extends through the
bottom of the heating block and when the liquid level drops
below the heated zone, evaporation immediately slows down.
Samples rarely go to dryness unless left unattended for a
long time. The one disadvantage mentioned for this system
was a tendency for bumping to occur. Boiling chips will not
prevent bumping in the block heater because they are not in
the heated portion of the solution. The original method for
initiating boiling was a small ebullator, but bumping was
still a hazard. The introduction of a fine stream of
nitrogen bubbles near the bottom of the heated area via a
steel capillary inserted from the top of the concentrator
tube was recommended as an improvement.l*
The commercial version of this device gave recoveries
ranging from 35-95% in different tests with the industrial
pollution test compounds. Poor recoveries always resulted
when the solution was allowed to evaporate to less than 1
ml. It was very difficult to know when the volume of
solution was approaching the desired point because most of
the tube is out of sight. Each evaporation took about three
times as long as with the airstream method. For relatively
volatile materials (compared to pesticides) there is little
to be recommended in this method.
Micro-Snyder Column
A third method, using a two chamber micro-Snyder column
(Figure 2) attached directly to a K-D tube containing a few
boiling chips and heated by a steam bath, gave better
recoveries with less variation in results. Again 5- and 10-
ml tests were made. Two sizes of condensers were used, a
commercial unit (available from Kontes Glass Co., Vineland,
N.J., 08360) with a 19/22 size joint and a locally made unit
with a 24/25 size joint. There was no difference in
efficiency between the 5- and 10-ml samples with either size
condenser. The average recovery for the pollution mixture
with a 19/22 condenser was 97% and with the 24/25 condenser,
94%. A commercial 2-chamber 24/25 condenser tested later
gave an average recovery of 99.9% in the 10-ml test.
Results from all micro-Synder column tests were better than
those for the airstream method. In addition, the range of
recoveries for tetrachloroethane, o-nitrotoluene,
benzothiazole, and g-cresol varied by only 10% or less in
16
-------
all these experiments as opposed to the 13-17% reported
earlier for the airstream-waterbath method.
«
Sometimes it is necessary to evaporate an extract to less
than 1 ml to attain adequate concentrations for analysis.
Recovery experiments were done with one-mi samples of the 13
compound pollution mixture that were evaporated to 0.7, 0.5,
0.3, 0.2, and 0.1 ml by the micro-Snyder column and the
airstream-»waterbath methods. Tests were also made by both
methods on spiked water samples that were extracted and then
evaporated to these levels.
The micro-K-D was found to be the best way to evaporate a
sample to 0.3 ml. Recovery a-t this volume is 90-95% in
comparison to an airstream recovery of 81-87%. Below 0.3 ml
the analyst has more control over the final volume with the
mild conditions of the airstream-waterbath method. The
results become highly variable from sample to sample at
these volumes. The airstream method at 0.2 ml gave
recoveries of 55-80% for the same compounds in duplicate
tests, and 60-80% at 0.1 ml. The micro-Snyder column gave
at 0.2 ml, 60-80% recovery and at 0.1 ml, 50-60%.
There are three keys to good micro-Snyder column recoveries.
To avoid bumping, the tube must be gently agitated for a few
seconds over the steam bath until the boiling chips start
bubbling. The condenser must be constructed so that the
glass bubbles move freely in their chambers and always
maintain a layer of liquid solvent that scrubs the escaping
gases. The K-D receiver tube must not be allowed to boil
dry. When the receiver tube has almost boiled dry, the
entire unit must be removed from the steam bath and allowed
to cool and drain. The final volume is 0.3-O.tt ml.
Based on the recovery of dissolved organics, the Kuderna-
Danish concentrator is the most efficient method for
concentration of large solvent volumes. The use of a
rotating vacuum evaporator or open evaporation in a beaker
results in lower recoveries. For the final concentration,
directly distilling the solvent through a micro-Snyder
column is more efficient, more reproducible, and faster than
the airstream-waterbath method.
HIGH-BOILING SOLVENTS
Industrial solvents and volatile taste and odor causing
compounds are sometimes overlooked in GC analysis because
they are obscured by the solvent peak. One proposed
solution is to use a high-boiling (and therefore late-
eluting) solvent for these applications. The difficulty
17
-------
lies in finding materials that are pure and are good
extractants. The ideal solvent contains no impurities above
the microgram/1 level. Also, the solvent peak on the GC
elutes about ten minutes after injection and returns to
baseline quickly, without tailing.
Several solvents have been tested for these characteristics
under GC conditions typically used to examine industrial
effluent extracts. The best solvent found was tetralin
(tetrahydronaphthalene), first recommended by B. F.
Dudenbostel of Region II, EPA. This material contained only
one major impurity and very few minor materials, leaving
considerable "open space" in the baseline. Ortho-
dichlorobenzene and one lot of tetradecane were also
acceptable.
Several solvents, including the purest commercially
available materials (99+%) , contained over a dozen
impurities at 5-mg/l concentrations. These make the solvent
unacceptable. In this class were decane, tridecane, 1,3-
cyclooctadiene, chlorobenzene, and 2-ethylnaphthalene.
Diethyl terephthalate contained only small impurity peaks
but the parent compound and the higher-boiling impurities
required an excessively long time to elute.
Tetralin is relatively nonpolar and does not extract highly
polar molecules. For example, extraction tests with ethyl-
mercaptan, suspected as the odor source in a water supply,
were completely unsuccessful. On the other hand, extraction
of a liter of landfill leachate with one ml of tetralin
revealed the presence of acetone, benzene, chloroform,
ethyl-* benzene, n-hexane and toluene.
When 2 ml of tetralin was used to extract 1 liter of water
spiked with 1 microgram each of chloroform, benzene, and
bis-chloroethyl ether, about 1.5 ml of tetralin was
recovered. Injection of 1 microliter of this extract on a
GC using a 10% SE-^30 column produced peaks for each of the
three compounds. Enough material was also extracted for
interpretable mass spectra using a GC-MS system with a
similar GC column and one-microliter injection.
Quantitative recovery studies were not made.
18
-------
SECTION VI
REFERENCES
1. Burham, A. K.r G. U. Calder, J. S. Fritz, G. A. Junk,
H. J. Svec, and R. Willis. Identification and
Estimation of Neutral Organic Contaminants in Potable
Water. Anal. Chem. 44: 139-142, January 1972.
2. Summary Bulletin Amberlite Polymeric Adsorbents. Rohm
and Haas Co. Independence Hall West, Philadelphia,
Pennsylvania. IE-172-200, April 1970. 9 p.
3. Webb, R. G., A. W. Garrison, L. H. Keith, and J. M.
McGuire. Current Practice in GC-MS Analysis of
Organics in Water. Environmental Protection Agency,,
Washington, D.C. EPA-R2-73-277. August 1973. 91 p.
4. Harvey, George R. Adsorption of Chlorinated
Hydrocarbons from Seawater by a Crosslinked Polymer.
Woods Hole Oceanographic Institute. Washington, D.C.
EPA-R2-73-177. Environmental Protection Agency. March
1973. 26 p.
5. Junk, G. A., J. J. Richard, M. D. Grieser, D. Wiltiak,
J. L. Witiak, M. D. Argeullo, R. Vick, H. J. Svec, J.
S. Fritz, and G. V. Calder. The Use of Macroreticular
Resins in the Analysis of Water for Trace Organic
Contaminants. J. Chromatogr. 99:745-762, 1974.
6. Porter, J. J. A Study of the Photodegradation of
Commercial Dyes. Clemson University. Washington, D.
C. EPA-R2-73-058. Environmental Protection Agency.
March 1973. 94 p.
7. Gesser, H. D., A. Chow, F. C. Davis, J. F. Uthe, and J.
Reinke. The Extraction and Recovery of Polychlorinated
Biphenyls (PCB) Using Porous Polyurethane Foam. Anal.
Lett. 4:883-886, December 1971.
8. Uthe, J. F., J. Reinke, and H. Gesser. Extraction of
Organochlorine Pesticides from Water by Porous
Polyurethane Coated with Selective Absorbent. Environ.
Lett. 3:117-135, February 1972.
9. Gesser, H. D., A. B. Sparling, A. Chow, and C. W.
Turner. The Monitoring of Organic Matter With
Polyurethane Foam. J. Amer. Waterworks Assn. £5:220,
April 1973.
19
-------
10. Meiggs, T. O.r Workshop on Sample Preparation
Techniques for Organic Pollutant Analysis.
Environmental Protection Agency. Office of
Enforcement, National Field Investigations Center,
Denver, Colorado. November 1973. 35 p.
11. Beroza, M., and M. C. Bowman. Device and Procedure for
Concentrating Solutions to a Small Volume with Minimum
Attention. Anal. Chem. 39 (10):1200-1203, August 1967.
12. Beroza, M., M. C. Bowman, and B. A. Bierl. Improved
Ebullator for Solution Concentrator. Anal. Chem.
44(14) :2411-2413, December 1972.
20
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-660/4-75-003
3. RECIPIENT'S ACCESSIO(*NO.
4. TITLE
------- |