vvEPA
United States
Environmental Protection
Agency
Health Effects
Research Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S1-84-028 Jan. 1985
Project Summary
Isolation and Concentration of
Organic Substances from Water
— An Evaluation of Supercritical
Fluid Extraction
Daniel J. Ehntholt, Christopher P. Eppig, and Kathleen E. Thrun
This study describes the use of
supercritical fluid carbon dioxide (SCF
062) as an extraction solvent for the
isolation and concentration of 23
specified organic solutes in water at
trace levels. Direct extraction using a
non-toxic, non-hazardous solvent such
as carbon dioxide has not previously
been applied to the isolation and
concentration of trace levels of organic
compounds from water. Most of the
recovery studies performed on the
model compounds in this research were
conducted on 400 mL aqueous samples
in a stainless steel extractor operated at
2,500 psi and 45°C.
The ability of SCF CO2 system to
extract and subsequently trap model
solutes with widely varying chemical
and physical properties was generally
found to be lacking. Recovery values of
greater than 40 percent were demon-
strated for only four of the model
solutes, 2,4-dichlorophenol, isopho-
rone, phenanthrene and stearic acid.
The low recoveries were attributed to
the inability of SCF CO2 to extract
highly water soluble or alkaline solutes
such as glucose, glycine, trimesic acid,
quinaldic acid, humic acid, caffeine, 5-
chlorouracil and quinoline. Mass balance
studies also indicated losses resulting
from an ineffective trap system for
volatile solutes (chloroform, furfural
and methylisobutyl ketone) and adsorp-
tion of hydrophobic compounds (bi-
phenyl, 1-chlorododecane, 2,4'-dichlo-
robiphenyl and 2,2'5,5'-tetrachlorobi-
phenyl) to the extraction system.
This Project Summary was developed
by EPA's Health Effects Research Labo-
ratory, Research Triangle Park, NC, to
announce key findings of the research
project that is fully documented in a
separate report of the same title (see
Project Report ordering information at
back).
Introduction
One means of understanding and
evaluating the possible lexicological
effects of organic substances in drinking
water is through biological tests Many of
these tests however, require significantly
higher concentrations of organic com-
pounds than those normally found in
drinking water because exisitmg test
systems are not sufficiently sensitive to
contaminants at trace levels. In addition,
although hundreds of organic compounds
have been identified and quantified m
samples of natural waters, much of the
organic matter present cannot readily be
characterized using currently available
analytical protocols Without such char-
acterization the substances cannot be
purchased or synthesized for use in
preparing of the concentrated solutions
required for health effects testing
Therefore, direct concentration/isolation
of organic contaminants in aqueous
samples for biological testing offers a
potential solution.
The Health Effects Research Laboratory
(HERL) U S EPA ha.s funded several
independent studies in an effort to
determine the effectiveness of different
isolation/concentration techniques
-------
Systems or techniques investigated
include reverse osmosis, vacuum distilla-
tion, solid adsorbents, and supercritical
fluid carbon dioxide (SCF COa) extraction
For purposes of comparison, a mixture of
23 model compounds was chosen by the
HERL—U S EPA to evaluate each system.
While solubility data in supercritical
fluids were reported as early as the late
1800s, commercial applications of super-
critical fluids (e.g , hops extraction and
the decaffemation of coffee) did not come
on stream until the 1970s The renewed
interest in superficial fluid extraction
was largely spurred by the increased
scrutiny of industrial solvents because of
health and safety considerations and in-
creasing costs associated with energy-in-
tensive separation processes such as
evaporation and distillation. Use of a non-
toxic, non-hazardous, volatile solvent
such as carbon dioxide offers several dis-
tinct advantages in the extraction of or-
ganic substances from water for biologi-
cal testing.
Experimental Procedures
Preparation of Model
Compound Test Solutions
Test solutions of model compounds
used in the small scale extractor studies
(400 mL) and the 10 liter extractor studies
were prepared by simply diluting the
required volume(s) of stock solution with
organic free water containing an inorganic
salt matrix. The salt matrix consisted of
70 ppm NaHCO3, 120 ppm CaSO4 and 47
ppm CaCI2-2H2O Table 1 lists the final
concentration at which each model
compound was tested
Small Scale Extraction
Recovery studies were conducted on
400 mL aqueous samples in a stainless
steel extractor (extractor capacity was
approximately 600 ml) operated at 2500
psi and 45°C. Supercritical conditions are
achieved for CO2 at pressures >1,070 psi
and temperatures >31.1°C. Approxi-
mately 300 standard liters of COa were
passed through the aqueous solutions
into the trapping system via a pressure
reduction valve. While various systems
were evaluated, the trapping system used
for the recovery studies consisted of a set
of three sequential glass U-tubes in
series, maintained at -76°C by a dry ice-
acetone bath. Operation at this tempera-
ture prevented clogging of solid COa.
To enhance the COa/aqueous phase
interfacial area and facilitate contact by
dispersion of the CO2 as fine bubbles, a
plug of silanized glass wool was placed in
the bottom of the extraction vessel. After
the vessel was charged with 400 mL of
aqueous feedstock solution, it was slowly
pressurized to the extraction pressure
and simultaneously heated to the desired
temperature. Carbon dioxide was then
passed through the aqueous phase at a
velocity of slightly more than 10 cm/mm
(about 10 standard liters/min at 1 atm.,
70°F). After the pre-determmed amount
of carbon dioxide (300 standard liters)
flowed through the sample, the system
was depressurized and the extracted
aqueous raffmate was drained into a col-
lection vessel. Analyses of the extracted
aqueous raffinate and the residue in the
trapping system were used for mass bal-
ance determinations.
Ten Liter Extractor
An original objective of this effort was
the extraction of a single five-hundred
liter sample at the conclusion of the
program. During the course of the project,
however, it was decided that a smaller-
scale run combined with additional
trapping experiments would yield more
useful results. A final series of ten liter
extractions was therefore carried out on
solutions containing all of the organic
compounds of interest and the three
inorganic compounds specified. The
extraction apparatus used in these
studies was similar to that used for the
small-scale work, but it had an internal
Table 1. Summary of Small Scale Extraction Study
Compound
Anthraqumone
Biphenyl
Bis(2-ethylhexyl)-phthalate
Caffeine
Chloroform
1 -Chlorododecane
5-Chlorouracil
Crotonaldehyde
2,6-Di-t-butyl-4-
methylphenol
2,4'-Dichlorobiphenyl
2,4-Dichlorophenol
Furfural
Glucose
Glycme
Hum ic Acid
Isophorone
Methyl Isobutyl Ketone
Phenanthrene
Qu maid ic Acid
Quinoline
Steanc Acid
*2,2'.5,5'-Tetrach/oro-
biphenyl
Tnmesic Acid
* ~ not detected.
- = not analyzed.
Concentration
(ng/U
50
50
50
50
50
5
50
50
50
50
50
50
50
50
2000
50
50
1
50
50
50
5
50
Number of
Determinations
2
1
3
2
*
/
1
1
1
3
1
*
*
2
1
1
1
1
2
1
3
1
Trap
Mean Recovery
21 4
38
15.4
*
—
20.7
*
7.8
327
20.3
454
223
—
—
*
40.4
173
97.0
*
34
47.5
187
*
% Recovery
Raffinate
Mean Recovery
846
23.4
11 3
81 4
—
*
56.0
31 0
•*-
5.5
28.0
10.8
—
—
42.0
24.5
11.4
*
85.0
46 1
22.0
12.0
91.0
Mass
Balance
106.0
27.2
267
81.4
—
207
96.0
38.8
327
28.8
734
33.1
—
—
420
64.9
28.7
97.0
85.0
49.5
69.5
307
91.0
-------
volume of approximately fifteen liters
The traps used were stainless steel
impmgers with a volume capacity of
approximately one liter
Results and Discussion
General
A series of SCF CC>2 extractions of
model solutes was conducted. In all
instances, the organic free water used to
prepare the model compound test solutions
contained a salt matrix (70 ppm NaHCOa,
120 ppm CaSCU and 47 ppm CaCI2-2H20)
to simulate the salt content of drinking
water Analysis of the trapping system
after extraction indicated that these salts
as well as lead nitrate were not extracted
by SCF CDs. Experiments that were
conducted to determine whether artifacts
were produced by the presence of a
chlorine residual (2 ppm) also showed
that no new compounds were formed
Small Scale Extraction
Table 1 detailsthe experimental results
obtained for the SCF CO2 extraction of the
model compounds The compounds
selected for investigation, the nominal
spiking levels, and the number of
experiments performed are provided in
the first three columns The mean trap
recoveries representing the sum of the
three U-tube trap in series are then
presented along with the mean raffmate
recovery (SCF C02 extracted feedstock)
and the mass balance (mean trap recovery
plus mean raffinate recovery). While
values for mass balance determinations
exceeded 40 percent for 11 of the model
compounds, only 2,4-dichlorophenol,
isophorone, phenanthrene and stearic
acid could be extracted and recovered
from the trapping system at levels >40
percent.
The low recoveries were largely
attributed to the inefficiency of SCF CO2
as an extraction solvent for highly water
soluble or alkaline solutes such as
glucose, glycine, trimesic acid, quinaldic
acid, humic acid, caffeine, andqumolme.
Poor extraction efficiency was also
demonstrated for anthraquinone and 5-
chlorouracil as indicated by the high
recoveries for these substances in the
raffinate. Mass balance determinations
suggested losses resulting from an
ineffective trap system for volatile
compounds (chloroform, furfural and
methylisobutyl ketone) and adsorption to
he extraction system for hydrophobic
solutes (biphenyl, 1-chlorododecane,
2,4'-dichlorobiphenyl and 2,2',5,5'-
etrachlorobiphenyl
Ten Liter Extraction
Based on scale-up considerations from
the 400 mL runs, each sample extraction
with SCF CO2 was conducted at 1950 ±
50 psi and 37-45°C, and involved passing
approximately 11,200 standard liters of
carbon dioxide through the aqueous
solution in about 110 minutes Since
pressure/flow rate fluctuations might
occur in the large scale apparatus that
could lead to the rupture of the glass
traps, a series of three stainless steel
impingers maintained at -76°C were
used to collect the organics present in the
effluent carbon dioxide stream.
Because the small scale extraction
experiments had shown that quantitative
removal of organics from the traps was a
problem, a trap rinse sequence was
designed to assure the dissolution of all of
the organic compounds from the traps.
The solvents used were compatible with
any denvatization/sample preparation
steps necessary before analysis. Thus, at
the conclusion of each experiment, the
three traps were rinsed sequentially with
methylene chloride, methylene chloride/
base (5N NH4OH added dropwise to each
trap), and Milli-Q water. The first methyl-
ene chloride trap rinse yielded some
aqueous phase extract (approximately
twenty milliliters) which was added to the
Milli-Q rinse. Aliquots of the trap rinses,
raffinate, and feedstock were analyzed
according to methods previously devel-
oped specifically for this project. Table 2
summarizes the results obtained from
these runs. In general, the types of com-
pounds which were extracted and trapped
were the same as those found in the
small scale experiments. In particular,
the hydrocarbons and phenols were col-
lected in the traps, whereas the more
water soluble compounds were not de-
tected in the trapping system. The mass
balances for some types of materials (e g.,
5-chlorouracil and the humic acid) were
poorer m the 10 liter extraction; however,
these runs contained all 23 compounds at
the same time and the extractions were
also conducted for a longer period of time
It is possible that the interactions be-
tween compounds under the acidic ex-
traction conditions accounts for the low
total recoveries in certain of these cases.
For example, the absence of humic acid in
the raffinate and the observation of a
brown organic material upon cleaning
the extractor suggested that this material
was precipitated.
Conclusions
This study demonstrated the utility of
supercritical fluid carbon dioxide for the
isolation and concentration of selected
compounds present in water at low
concentrations Compounds exhibiting
greater solubility in water (e.g., trimesic
acid, glucose, and glycine) do not show
evidence of extraction; in addition, those
materials which tend to precipitate
(humic acid) or form more soluble species
(caffeine) under acidic conditions were
not extracted.
An extraction conducted on an aqueous
solution containing a two part-per-
million chlorine residual did not indicate
the generation of any new species m the
extract. All of the tests in this program
were conducted on aqueous solutions
containing IMaHCOa, CaSCU, and CaCI2
added at concentration levels typical of
drinking water Experiments were also
conducted to determine whether or not
these inorganic materials or PbNOs
(added to several solutions as a surrogate
for possible toxic metal concentration)
were extracted. Results indicated that the
inorganics were not isolated or concen-
trated.
The extraction conditions used in the
study were determined based on approx-
imately seventy percent extraction of
phenolic compounds in early runs. While
additional treatment with supercritical
fluid carbon dioxide might increase the
extraction efficiency of the process,
additional trapping (recovery) problems
may occur
Although the aqueous extraction
sampling and analysis procedures were
well developed for the study, the trap
systems and trap rinse procedures for the
small scale extractions (0.4 L) and the 10 L
extractions were different. Therefore, the
results obtained for trap recoveries are
not directly comparable, but the raffinate
analysis results are
The overall conclusion from this study
was that the supercritical fluid carbon
dioxide extraction of drinking water
represents an alternative path for selected
organic compounds which are not highly
soluble in water It can be used in lieu of
organic solvents or membrane techniques
when those interfere with biological tests.
Recommendations
Since the concept should be adaptable
to large scale extraction of certain types
of organic compounds from water,
further study of the supercritical fluid
extraction concept is recommended.
Specifically, the efficiencies of alternative
trapping systems should be defined For
example, complete trapping of all effluent
carbon dioxide in a vessel from which
fractional distillation of C02 can take
place is likely to yield higher recoveries of
-------
Table 2. Summary of Ten Liter Extraction (Avg. 3 Runs)
Compound
Anthraqutnone
Biphenyl
Bis(2-ethylhexyl)-
phthalate
Caffeine
Chloroform
1 -Chlorododecane
5-Clorourac/l
Crotonaldehyde
2.6-Di-t-butyl-4-
methylphenol
2. 4'-Dichlorobiphenyl
2.4 -Dich/orophenol
Furfural
Glucose
Gtycme
Hum tc A cid
Isophorone
Methyl Isobutyl Ketone
Phenanthrene
Ctuinaldtc Acid
Qumoltne
S tear ic Acid
2,2. '5,5'-Tetrach/oro-
biphenyl
T rime sic Acid
Concentration
(M9/LJ
50
50
50
50
50
5
50
50
50
50
50
50
50
50
2000
50
50
1
50
50
50
5
50
Trap
Mean Recovery
32
15
30
6
NA
25
9
3
31
45
26
3
»
*
;
28
5
14
»
4
«
30
*
% Recovery
Raffinate
Mean Recovery
31
*
»
71
*
»
13
2
*
*
*
*
»
*
+
*
4
*
89
31
27
*
84
Mass
Balance
63
15
30
77
0
25
22
5
31
45
26
3
0
0
1
28
9
14
89
35
27
30
84
* = not detected.
NA = not analyzed
+ = none detected, brown precipitate was recovered from the extractor
organics. In addition, the use of "closed
systems" in which the effluent CO2
stream is recycled through the aqueous
stream after removal of some portion of
the dissolved organic compounds may
permit more efficient collection of those
organics. If these studies are conducted
on a small scale, particular attention
should be paid to irreversible adsorption
to the traps and inefficient removal of the
organics from the effluent COa stream as
likely causes of low organic compound
recovery
DanielJ. Ehntholt, Christopher P. Eppig, and Kathleen E. Thrunare withArthurD.
Little, Inc., Cambridge, MA 02140.
H. Paul Ringhand is the EPA Project Officer (see below).
The complete report, entitled "Isolation and Concentration of Organic Substances
from Water—An Evaluation of Supercritical Fluid Extraction," (Order No. PB
85-138 899; Cost: $10.00, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
BULK RATE
POSTAGE & FEES PA
EPA
PERMIT No G-35
Official Business
Penalty for Private Use $300
* U S GOVERNMENT PRINTING OFFICE, 1985 - 559-016/789
------- |