United States
Environmental Protection
Agency
Environmental Monitoring
Systems Laboratory
Research Triangle Park NC 2771
Research and Development
EPA/600/S4-87/040 Jan. 1988
Project  Summary
Supercritical  Fluid  Extraction  of
Paniculate and Adsorbent
Materials:  Part  II
Bob W. Wright and Richard D. Smith
  Supercritical fluids have  solvent
strengths similar to those of liquids, but
have higher diffusion coefficients,
lower viscosities and an extended
temperature range which provides the
potential for more rapid and efficient
extraction than can  be achieved with
liquids. In this project, the applicability
and efficiency of  analytical supercrit-
ical fluid extraction and related  meth-
odologies were evaluated. These stu-
dies  included  development  of
quantitative off-line  supercritical fluid
extraction methodology and a compar-
ison to traditional Soxhlet extraction,
the development and evaluation  of on-
line supercritical  fluid extraction-gas
chromatography for combined sample
preparation  and analysis, and  direct
supercritical  fluid  extraction-mass
spectrometry  for the monitoring of
specific extraction profiles as a func-
tion of time. The  sample  matrices
included an air particulate  sample,
XAD-2 resin, polyurethanefoam (PUF),
and Spherocarb®* adsorbent,  which
were spiked with  various model com-
pounds. Carbon  dioxide, isobutane,
and methanol-modified (18.2 mole %)
carbon dioxide were used as supercrit-
ical fluid systems.  The quantitative
analysis capability of a  fluorescence
detection supercritical fluid  chroma-
tography method  and solute focusing
methods for capillary supercritical fluid
chromatography were also developed
and evaluated.
 •Mention of trade names or commercial products
 does not constitute endorsement or recommenda-
 tion for use.
  This Project Summary  was devel-
oped by EPA's Environmental Monitor-
ing Systems Laboratory, 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
  Numerous methods and combinations
of methods with varying  degrees  of
complexity are currently employed  to
prepare samples for analysis. Less
complex, more rapid, and more sensitive
preparation procedures would be desir-
able in many cases, particularly for small
samples or low concentrations of ana-
lytes. Supercritical fluid extraction (SFE)
has been successfully used for several
large-scale chemical processing applica-
tions and is currently attracting renewed
interest. However, only limited attention
has been  given  to SFE methods  for
analytical  applications. A  number  of
advantages including more  rapid extrac-
tion rates, more efficient  extractions,
increased selectivity, combined analyte
fractionation in conjunction with extrac-
tion, and  direct continuous on-line
analysis, such as with  mass spectrome-
try or  periodic gas chromatographic
analysis, are possible with SFE compared
to  conventional  liquid   extraction
methods. The goals of this study were
to determine the  potential utility of SFE
for analytical sample preparation  and to
develop  improved methods of analysis
utilizing these and other  supercritical
fluid analytical methodologies.
  The potential advantages of SFE
accrue from the properties of a solvent

-------
at temperatures and pressures above its
critical point. At elevated pressure this
single phase will have properties that are
intermediate between those of the gas
and liquid phases and depend on the fluid
composition, pressure, and temperature.
The compressibility of supercritical fluids
is large just above the critical temper-
ature,  and small  changes  in pressure
result in large changes in density of the
fluid. The density of a  supercritical fluid
is typically 102 to 103 times greater than
that of the gas. Consequently, molecular
interactions increase due to shorter
mtermolecular distances. However, the
diffusion coefficients and viscosity of the
fluid,  although  density-dependent,
remain similar  to those of a gas  The
"liquid-like" behavior  of a supercritical
fluid  results  in  enhanced  solubilizmg
capabilities compared  to the subcritical
gas, and m higher diffusion  coefficients,
lower viscosity, and an extended temper-
ature range compared to the correspond-
ing  liquid. These properties allow solvent
strengths similar to those of liquids but
with  greatly improved mass  transfer
characteristics, which  provide the poten-
tial  for both more rapid extraction rates
and more  efficient extraction  due to
better penetration of the matrix.

Procedure
  Studies  were conducted  to  evaluate
the applicability and efficiency of analyt-
ical  SFE  and related methodologies.
These  studies included off-line SFE of
adsorbents and particles, on-line  super-
critical fluid extraction-capillary gas
chromatography (SFE-GC), and  direct
supercritical fluid extraction-mass spec-
trometry (SFE-MS).  The  quantitative
analysis capability of  the fluorescence
detection supercritical fluid chromatog-
raphy (SFC) system that was previously
developed in this project was  also
evaluated. Preliminary studies aimed at
developing solute focusing  methods for
capillary SFC similar to cryogenic trap-
ping in gas chromatography were also
conducted.
  In the off-line SFE studies, various fluid
systems were evaluated and compared
to Soxhlet extraction for the recovery of
low concentrations  (ppm) of  model
compounds  from selected  adsorbents.
Higher molecular weight (228-450 dal-
tons) polycyclic aromatic hydrocarbons
(PAH) were extracted from XAD-2 resin,
polyurethane foam,  and Spherocarb®
adsorbents. Air paniculate  matter was
also  used as  a  complex matrix  for
evaluation  of  the SFE methods. In
addition to carbon dioxide, isobutane and
methanol-modified (18.2  mole percent)
carbon  dioxide were used for SFE
solvents. Instrumentation was designed
to allow extraction of gram quantities of
the matrix  at pressures up to 400 bar
and  temperatures  up  to  235°C, with
collection of  the effluent in a  sealed
liquid-nitrogen-cooled flask.
  A logical extension of SFE is to com-
bine the process with a chromatographic
analysis  method. The variable solvating
power of a supercritical fluid furnishes
the mechanisms for the selective extrac-
tion of the  components of interest from
the sample matrix and provides the basis
for an automated method  where sample
preparation and analysis  can be  instru-
mentally linked.  Instrumentation  and
methodology  were  developed for the
automated  SFE-GC system. The  super-
critical fluid extracts were  decompressed
through a restrictor  to  deposit  and
concentrate the analytes  at the inlet of
a standard capillary gas chromatography
column  for subsequent analyses. This
methodology  allows  several  modes  of
operation including quantitative extrac-
tion of all analytes from a sample matrix,
quantitative extraction and  concentra-
tion of trace analytes,  selective extrac-
tions  at various solvating  powers  to
obtain specific  fractions,  and multiple-
step extractions at various pressures for
qualitative characterizations.  These
modes of operation were developed, and
the potential utility of this methodology
for sample extraction and  selective
fractionation was demonstrated using a
standard polycyclic aromatic hydrocar-
bon  mixture and two  complex sample
matrices.
  Direct  SFE-MS was  used  to monitor
the individual extraction  of each model
compound as it was extracted from XAD-
2 resin,  PDF, and Spherocarb® adsorb-
ents with carbon dioxide,  isobutane, and
methanol-modified (18.2 mole %) carbon
dioxide fluid systems. Individual extrac-
tion profiles  of each of the  spiked
compounds as  a function of time were
obtained, and  the different  extraction
behaviors (threshold pressures, extrac-
tion rate, solubility, decay curves,  etc.)
of the model compounds from the various
adsorbents were observed.
  A fluorescence detection capillary SFC
system  was developed to obtain  high
sensitivity detection  when using fluids
that are not  compatible with flame
mnization detection, and its quantitative
analysis capability was evaluated. These
procedures included  evaluation of a
multi-level detector calibration of several
compounds, documentation of the repro-
ducibility of multiple determinations of
specific mixtures, and comparison of the
quantitative values obtained from paral-
lel determinations using this methodol-
ogy and capillary gas chromatography.
  Preliminary studies were conducted to
develop solute  focusing  methods  for
capillary SFC similar to cryogenic trap-
ping  in  gas  chromatography.  One
approach that proved  moderately suc-
cessful was to create density gradients
within a short length of the column by
altering the fluid temperature.

Results
  Optimized off-line  analytical  SFE
instrumentation and methods  were
developed that  allowed high molecular
weight polycyclic aromatic compounds to
be successfully extracted and recovered
from the adsorbent  matrices. A recipro-
cating pump provided  the potential for
use of unlimited volumes  of fluid  in an
extraction,  and  collection of the extrac-
tion effluent in a sealed liquid-nitrogen-
cooled flask prevented analyte losses due
to aerosol formation during fluid decom-
pression.  When an open  collection
method was  used  under conditions
where aerosol formation was  expected
to be  inhibited (low temperature expan-
sion where a two-phase  product was
expected or utilization  of a polar liquid-
modified  fluid), significantly higher
extraction recoveries were  obtained.  The
model compounds were  more easily
extracted from  PDF than  from XAD-2
resin. Higher  levels of  background
components were also extracted from
PUF, and the PUF itself appeared to be
slightly soluble in  methanol-modified
carbon  dioxide. Except for  PUF,  the
various adsorbents did not appear to be
affected by  SFE.  Supercritical  fluid
extraction  of  the  highly adsorptive
Spherocarb® matrix with isobutane  and
with methanol-modified carbon dioxide
fluids provided more efficient extraction
recoveries than those achieved by Soxh-
let extraction. However, only low levels
of the less polar analytes were recovered.
Supercritical fluid extraction also pro-
vided more rapid and improved extraction
efficiency of PAH from Urban Dust than
did Soxhlet extraction. The  concentra-
tions  of selected compounds  obtained
from  these  analyses exhibited surpris-
ingly  good  agreement with the  NBS
certified values.
  On-line SFE-GC provides potential for
combined sample preparation and anal-

-------
ysis. In addition to completely automated
operation,  rapid  analyses and high
sensitivities can be achieved with this
methodology. The  selectivity obtainable
with the wide range of solvent powers
available with SFE  provides the potential
for fractionation of complex samples and
isolation of specific  analytes  from a
matrix. The quantitative extraction and
transfer of PAH analytes from  an XAD-
2  resin  to the gas  chromatographic
column for analysis was demonstrated.
Quantitative extraction and  transfer is
dependent  on  solute  solubility which
requires using an adequate volume of the
supercritical fluid.
  Direct  SFE-MS provides a means of
simultaneously monitoring the extrac-
tion  profiles of the  individual  model
compounds from the adsorbents. The
extraction behavior of the model com-
pounds from the various adsorbents with
the different fluid systems was consist-
ent with the off-line  SFE studies. The
compounds were more soluble and better
extractability was achieved with isobu-
tane than  carbon  dioxide. More rapid
exhaustion of the compounds was also
obtained with isobutane. The methanol-
modified carbon  dioxide  mixture was
used under  subcritical temperature
conditions which resulted in a two-phase
liquid-gas mixture  rather than  a super-
critical fluid. Consequently, the  polar
compounds capable of hydrogen bonding
with  liquid methanol were extracted
more rapidly and more efficiently than
the  neutral  compounds. The  model
compounds were  all  extracted  more
easily from PUF than from XAD-2 resin
with all the fluid systems studied. The
extraction profile intensities from PUF
correspond to  the expected solubility
orders of the  compounds. The  model
compounds were so strongly adsorbed by
Spherocarb® that SFE-MS  extraction
profiles could not be obtained.
  Linear  response of  the  neutral PAH
model compounds was achieved over the
narrow concentration range studied, and
sensitivities in the low picogram range
were obtained using fluorescence detec-
tion SFC. Calibration plots with less than
an 8% standard error  were constructed
from replicate analyses at four concen-
tration levels. Slightly better  calibration
was  obtained from peak heights than
from peak areas. Replicate analyses of
standard mixtures and authentic sample
extracts agreed to within approximately
15%. Quantitative determinations of
authentic sample extracts from  parallel
analyses using capillary gas chromatog-
 raphy  and_ fluorescence  detection SFC
 were generally consistent within a 20%
 range.  A significant portion of the
 quantitative variance  can be attributed
 to problems associated with the injection
 method.
  A moderately successful solute focus-
 ing technique for  capillary SFC was
 developed. The technique used a  local
 zone of lowered fluid density created  by
 heating a short length of the column
 inlet.  Preliminary evaluations of  other
 approaches using a  retention gap  and a
 pressure gradient were also conducted.

 Conclusions and
 Recommendations
  Essentially quantitative recoveries  of
 all the  model compounds  from the XAD-
 2 and PUF adsorbents were achieved  by
 Soxhlet extraction. In general, the SFE
 methods provided extraction  efficiencies
 comparable   to  traditional  Soxhlet
 methods, but with  over an  order  of
 magnitude increase in extraction speed.
 Supercritical fluid extractions  were
 accomplished in  30-45  minutes  com-
 pared to 16 hours for Soxhlet extraction.
 The more polar and  higher molecular
 weight analytes were more difficult  to
 extract  from  the adsorbents, and the
 methanol-modified carbon dioxide fluid
 and higher critical  temperature isobu-
 tane systems provided higher extraction
 efficiencies  than did carbon  dioxide
 alone.  Limited solubility  of  the higher
 molecular weight materials  in carbon
 dioxide limited their recoveries with the
 relatively  small  volumes of fluid that
 were used.
  The utililty  of on-line SFE-GC, and  its
 qualitative operation for sample extrac-
 tion from adsorbent matrices,  and  its
 potential for fractionation and isolation
 of specific components  from complex
 matrices was demonstrated.
  Direct SFE-MS allowed the individual
 extraction profile of each model  com-
 pound to be monitored as it was extracted
 from XAD-2  and  PUF adsorbents  with
 carbon dioxide, isobutane, or methanol-
 modified (18.2 mole %) carbon dioxide.
 Extraction profiles for Spherocarb® could
 not be  constructed since only very low
 signal levels were detected.
  The fluorescence  detection SFC sys-
tem was able to provide  reliable quan-
 titative determinations with a combined
 uncertainty of approximatley  15%.
  Several recommendations  for off-line
SFE studies  can  be made.  Since the
overall  SFE rates appear to be limited by
the solubility of  the analytes in the
 extraction fluid rather than by desorption
 or mass transfer rates, more rapid and
 more  efficient  extractions should  be
 feasible by  use  of larger fluid volumes
 at higher rates.  The use of higher fluid
 volumes may be  best obtained by recom-
 pression and recycling of the  fluid.
 Additional studies should be conducted
 to develop collection methods that can
 be  routinely applied. A  more rigorous
 study of the chemical composition of the
 supercritical fluid extracts of NBS Urban
 Dust should  be conducted.  Detailed
 fractionations  and analyses of the
 extracts would  not only provide addi-
 tional chemical composition information,
 but would provide additional insights  on
 the extraction process.
  The  on-line  SFE-GC  and SFE-MS
extraction studies should be expanded.
More rigorous quantitative studies using
multiple analyses  and  a  multi-level
calibration  with a  variety of model
compounds should be conducted to verify
the performance of the on-line SFE-GC
methodology. Methods of increasing the
extraction fluid flow rate to allow extrac-
tion of  larger sample matrices in a short
time-frame should be investigated. The
methanol-modified carbon dioxide SFE-
MS studies should be repeated using an
extraction temperature above the critical
point and the results compared to those
obtained in this study using a subcritical
two-phase   mixture.   Appropriate
response factors and quantitative SFE-
MS  extraction   profiles  should  be
obtained. The SFE-MS approach should
be expandedto include a chromatograph-
ic separation step to allow concentration
of trace  components in the extraction
effluent and to provide separation of the
components prior to MS analysis.
  The  linearity of  the fluorescence
 detector response should be evaluated
 over a much wider range,  e.g.,  >103.
 Improved splitless injection  methods
 should be  developed  to improve the
 accuracy and reproducibility of quantita-
 tive SFC analyses. Such improvements
 could include high-speed valve switching
 to effect lower volume injections and use
 of long retention gaps (>10 m) to aid
 solute focusing on the analytical column.
The use of valves with 1732" zero-dead-
volume  outlet fittings  should also  be
 investigated.
  The development of on-column solute
focusing methods for capillary  SFC
should  be continued. This development
should include use of long retention gaps
coupled with thermal focusing and use
of on-column  solute  concentration

-------
  methods employing  extraction-injection
  and pressure focusing techniques
                                            Bob W.  Wright and Richard D.  Smith are with Battelle,  Pacific Northwest
                                              Laboratory, Richland, WA 99352.
                                            Nancy K. Wilson is the EPA Project Officer (see below).
                                            The complete report, entitled "Supercritical Fluid Extraction of Paniculate and
                                              Adsorbent Materials: Part II," (Order No. PB 88-133 699/AS; Cost: $14.95,
                                              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:
                                                    Environmental Monitoring Systems Laboratory
                                                    U.S. Environmental Protection Agency
                                                    Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
                                                                                                 US.GfriCiALWIA
'ENAL1Y
OR
        U.S.POSTAG1
 Ril'ATF
JS.I- S300i ~ ft  9 *?
»    * i — U ,L L
Official Business
Penalty for Private Use $300

EPA/600/S4-87/040
          0QQC329   PS
          0  S  €*IVIt  PROTECTION  *Gg«CY

-------