Adlard E.R. (ed.) Chromatography in the Petroleum Industry
Подождите немного. Документ загружается.
272
Chapter
10
SF-Sotuer
Program
Temp
:
68.48
"C
Density
in
gdat
I
2.5
t4
Naae
TenpOC
1
Naphtha1
55.88
-
Analytes
-
D
E
N
S
I
I
T
Y
-
Comands
=
(T
)ah
late
Print iny
. .
.
1
(Ehit
Solubi
1
it9
R
.133
Fig.
10.1.
Plot
of
variation
of
so1ubilit)l
of
naphthalene
with
density in carbon dioxide
at
55°C.
(Reproduced
from
SF-Solver Software package courtesy
ISCO
Inc.
USA.)
critical fluid, the solvent power can be adjusted over a wide range by varying the
pressure and therefore
the
density as illustrated
in
Fig.
10.1.
This means that the
solvent power and hence the selectivity of the solvent can be fine tuned to target
the analytes of interest. Table
10.2
lists
data where the density of carbon dioxide
has been calculated for various pressures at
35°C
and
55°C.
The calculated
density has then been allocated a value for Hildebrand solubility parameter and a
solvent with a comparable value selected. Although this
is
a
somewhat simplistic
approach, it does illustrate that the solvent power
of
a supercritical fluid can be
easily mmipulated by changing the applied pressure.
Furthermore, the combination of pressure and temperature can also be em-
ployed to obtain selectivity as illustrated in Fig.
10.2
where the variation of
solubility with pressure
for
naphthalene in carbon dioxide is shown at
35°C
and
55°C.
At
a pressure
of
90
atm, an increase in temperature from
35°C
to
55°C
results in a decrease in solubility and naphthalene would be precipitated from
solution.
In
contrast, however, at a pressure
of
200
atm, the reverse is true and an
Supercritical
fluid
extraction
273
-
Analytes
--
U
Hare
TerpOC
1
Naphthal
35.88
2
Naphthal
55.88
rn
Connands
-
(Tlabulate
Printiny
.
. .
IElxit
TABLE
10.2
PRESSURE, DENSITY, HILDEBRAND SOLUBILITY PARAMETER, AND EQUIVALENT
USA)
SOLVENT DATA AS PRODUCED BY THE SF-SOLVER COMPUTER PACKAGE, (ISCO,
~~ ~
Pressure
Pressure Density
Hildebrand Equivalent
at
35°C
(bar)
at
55°C
(bar)
(g/ml)
solubility solvent
parameter
78
100
0.329 2.81
Hydrogen
133 23 5
0.800
6.82
Methane
155 265
0.828 7.07
n-Pentane
180
300
0.860 7.33
n-Hexane
220 350
0.888
7.57
n-Octane
360 520
0.936 8.21
Cyclohexane
480 >660
1.009
8.60
Carbon tetrachloride
Pressure
in
ATM
SF-So
1
uer
Program
Isco
,
Inc
.
Teap
68.48
"C
I
I
284.8
P
R
E
s
S
0
R
E
a1.1
8.881
Solub
i
1
ity
El. 133
Fig.
10.2.
Plots
of
variation
of
solubility
of
naphthalene with pressure in carbon dioxide at
35°C
and
55°C.
(Reproduced
from
SF-Solver Software package courtesy ISCO Inc USA.)
References
pp.
301-303
274
Chapter
I0
increase in temperature causes a dramatic increase in solubility. This tunability
of solvent power means that one supercritical fluid can be used to extract a wide
range of analytes. For more complete discussions
on
the factors affecting solu-
bility in supercritical fluids, references
[24]
are recommended.
10.2.
I.
3
Volatility
This may seem an unusual benefit, but in almost all cases the analyst wishes
to perform some further analysis on the extracted species and it is useful to be
able to isolate the extract from the solvent. This can be particularly true for trace
analysis when it
is
useful to be able to get the analytes in as concentrated a form
as possible in order that they can be measured with sufficient sensitivity by the
chosen analytical technique.
In
conventional solvent extraction, this normally means removal
of
the solvent
by vacuum distillation using
a
rotary evaporator but this can introduce problems
with the
loss
of analytes by both evaporation and thermal lability. It can also be a
problem in that trace impurities in the solvent are concentrated along with the
analytes and can interfere with their subsequent determination. In the case of
SFE,
the most common solvents are gases at room temperature and below and as
such,
on
depressurization from the extraction system, are easily removed from
even a cryogenically cooled analyte collection vessel without the need for distil-
lation
or
the application of heat. Furthermore, as gases they are generally easier
to obtain
in
a
high degree
of
purity and therefore problems with solvent impuri-
ties can be reduced. This ability to remove the solvent easily
is
particularly im-
portant when
SFE
is
coupled on-line to chromatographic techniques such as gas,
liquid,
or
supercritical fluid chromatography.
10.2.
I.
4
Mntrix penetration
In the majority of cases where an extraction technique
is
used, the sample
is
a
solid and if the chosen solvent has sufficient power to dissolve the analyte(s) of
interest, then the rate determining step
in
the extraction process will be governed
by how well the solvent interacts with the matrix and removes the analyte from
it.
This
is
where the gas-like physicochemical properties of a supercritical fluid
such as diffusivity, viscosity and surface tension begin to show benefit. The dif-
fusivity of supercritical fluids
is
typically
1-2
orders of magnitude greater than
that of solutes in liquids and this gives a more rapid transfer of material from a
solid surface to the extract phase compared to that with liquids. In addition, the
viscosity of the supercritical phase is typically an order of magnitude lower than
that of a liquid and this combined with a zero surface tension allows easy pene-
tration of microporous matrices. These factors, when combined, typically give
extraction times which are an order
of
magnitude less than an equivalent solvent
extraction. The use of temperature can also be influential on the release of ana-
Supercritical fluid extraction
275
lyte from the sample matrix, particularly when volatile compounds such as petro-
leum-based components are involved. Temperature can also be employed to ef-
fect a structural change in the sample matrix, e.g. melting or converting crystal-
line polymers into amorphous ones, thereby allowing penetration of the matrix
by the solvent. The use of “solvent modifiers” as mentioned in the section on
solubility can also have an important part to play in matrix interaction where
they can act to displace molecules of analyte which are adsorbed on active sites
on the sample matrix.
10.2.2 Environmental and safety advantages
Carbon dioxide is by far the most popular analytical supercritical fluid and
probably the biggest driving force for the application
of
analytical supercritical
fluid extraction has been as a result of the benign nature
of
this fluid. It is cheap,
non-flammable, non-toxic, and is not considered as environmentally unfriendly.
As
a result there has been an initiative in the United States by the Environmental
Protection Agency
(EPA)
to employ it to replace the environmentally unfriendly
and often toxic solvents used in analytical methods which employ solvent ex-
traction procedures. This is becoming more and more important as the evidence
for the toxicity of once common solvents such as chloroform and more recently
hexane has been collected. In addition, the removal of the solvent as gaseous
carbon dioxide means that the costly problems associated with the storage and
disposal of conventional solvents are avoided.
So
we can conclude that the main reasons for using supercritical fluid extrac-
tion can be summarized as:
-
reduces sample preparation time
-
selective extractions are possible
-
reduces chemical exposure
-
minimizes solvent purchase and disposal costs
-
compatible with coupling on-line to chromatography
Let us now
look
at the equipment and techniques employed to perform SFE
and some petroleum related applications, with particular reference to the use of
SFE
as a pre-separation technique for chromatography.
10.3 SFE EQUIPMENT
The basic requirements for equipment for analytical scale SFE are as follows:
References
pp.
301-303
216
Chapter
10
--
~I
-
FEED FLUID
Fig.
10.3.
Schematic representation
of
a simple
SFE
system.
-
a
supply of fluid at the required pressure
-
a pressure-safe sample vessel
-
a means of controlling the extraction temperature
-
a depressurization device
-
a means of collecting the extract on depressurization
The coupling
of
these requirements
is
shown schematically in Fig.
10.3.
In
or-
der
to
fully describe all the variants of each component
which
are available and
their relative merits and disadvantages would require a chapter on its own,
so
we
will only briefly cover the basics here. Readers who wish to find out more are
advised to consult the specialist texts recommended earlier
or
to contact all the
manufacturers and decide for themselves which product best suits their require-
ments.
10.3.1
Pumping
systems
The pumping system is the heart of an
SFE
system. It must be able to supply
a
continuous stream
of
fluid at the desired pressure and preferably at
a
user-
selectable flow rate. There are
two
main types
of
pumps in
use
at present,
namely syringe and reciprocating piston pumps. Syringe pumps were primarily
developed
for
SFC
where the main requirement
is
for
a
pump which can be pres-
sure programmed but also provide a pulseless controlled
flow
of fluid. However,
they
do
have limitations
for
use in
SFE,
namely they are relatively expensive,
Supercritical fluid extraction
277
and have a limited volume (typically between
150
and
300
ml) and unless two
are used in tandem, the extraction may have to be stopped to refill the pump. In
addition, they are difficult to employ with solvent modifiers.
Reciprocating piston pumps
of
the type commonly used in HPLC can easily
be modified to pump supercritical fluids such as carbon dioxide
[
16,171. They
can be easily coupled together to provide binary or even ternary solvent systems
as in HPLC, and will provide a continuous flow of fluid at high flow rates. They
do not give a pulseless flow like a syringe pump but this is not as important in
SFE as it is in both HPLC and
SFC.
One drawback of employing HPLC pumps
to pump carbon dioxide is that
it
is usually necessary to chill the pump head and
the feed carbon dioxide to ensure that it remains as a liquid. However, recently a
new syringe type pump has appeared
[
181
which is specifically designed for use
with analytical
SFE
and this does not require the pump head to be chilled. An-
other type of piston pump which is air driven has also found uses where flam-
mable or corrosive solvents have been used. This type of pump has a high ca-
pacity condenser on the inlet which can liquefy gases such as propane, ethylene,
and ammonia
[
191.
10.3.2
Sample vessels and temperature control
The main requirement for sample vessels is that they must be able to safely
withstand the pressures and temperatures used in extraction. In addition, they
should be easy to open, load and seal, and to connect into the extraction system.
Ideally they should have a low thermal mass
so
that they reach thermal equilib-
rium quickly. Early cells were empty HPLC columns but these were normally
small diameter and it was difficult to fill and remove the sample. More recently,
custom designed cells with removable end frits and finger tight connections have
been developed which are available in sizes ranging from a few hundred microli-
tres to about
50
ml. It is also necessary to be able to maintain the temperature
of
the sample cell at the required value and this is most commonly done using a
fan-assisted oven to facilitate rapid heating and cooling of the extraction cell. In
the early days of SFE, one of the main claimed advantages of the technique was
that it worked best at low temperatures (i.e. high density) and was therefore
considered useful for thermally labile substances. As a result, many of the ovens
used had a maximum temperature of
100°C.
However, recent work has shown
that temperatures in the region of
200°C
are often needed to obtain maximum
analyte recovery of
PAHs
and this should be considered when looking at sys-
tems. In addition, if chilled pump heads are being used, the temperature of the
fluid may be low and it is worth having some mechanism to pre-heat the fluid to
the desired temperature before it enters the extraction vessel.
References pp.
301-303
278
Chapter
10
10.3.3 Depressurization
There are a number of devices which allow the depressurization of the solvent
from the extraction pressure to atmospheric pressure and therefore cause precipi-
tation of the extract
in
the collection device. Until recently, the most popular
device was one which provided a restriction to the flow of the fluid by passing it
through a narrow tube
or
orifice, therefore building a back pressure. The most
common way of achieving this used
a
fused silica capillary to restrict the escape
of the gas and therefore maintained pressure in the extraction cell. The pressure
which was maintained was dependent on the length and internal diameter of the
restrictor. The main problem with these restrictors was that the flow could not be
controlled independently of pressure. In addition, they were fragile and prone to
blockage with extract. Other variants employed crimped metal tubing and frits in
the end of tubing but they suffered from the same problems.
It
is
amazing to think that a technique which relies on pressure and pressure
control for
its
success employed such primitive technology. The second type
of
device which has been employed
is
the back pressure regulator. These were
originally based on pressure relief type valves where the force exerted by
a
spring was used to force a needle into a seat. When the system pressure is suffi-
cient to lift the valve it was released. The pressure could be set by adjusting the
spring force via a screw thread. These systems gave independent control of flow
and pressure but were still prone to blockage and sticking. Saito
et
al.
[20]
made
a significant breakthrough by designing an electronically controlled high speed
switching valve. The configuration of the valve was such that precipitated ex-
tract and solid carbon dioxide was cleared from the valve by the vibration and
punching action of the valve needle. This and other electronically controlled
valves which have been developed since have brought analytical SFE to the
stage where analysts in different laboratories can repeat exactly the same ex-
periment
in
terms
of
pressure, temperature and flow rate and expect to get the
same result.
10.3.4 Extract collection
In
the early days
of
analytical SFE development this was one of the most ne-
glected areas. Indeed,
in
much of the early work on SFE which reported low ex-
traction efficiencies,
it
was the trapping on depressurization which was the
problem and not incomplete extraction. This is not surprising when one consid-
ers that the flow of solvent gas on depressurization
is
often
in
the litre/minute
range and that the precipitated extract
is
in
the form of a fine aerosol
in
the gas
phase. Since these early days, many techniques have been developed to ensure
Supercritical fluid extraction
279
complete collection of the extract and some of the available options are listed
below:
-
in an empty glass vial
-
onto a solid surface
-
on an inert solid, e.g. glass balls
-
on an active solid sorbent, e.g.
C18
stationary phase
-
bubbling into solvent
The efficiency
of
all of these methods can often be improved when they are
combined with each other, e.g. glass balls in a solvent to maximize surface area,
or combined with cryogenic cooling of the trapping material. There have been
many studies made on the efficiencies of the different trapping methods
[21-251
and there are a number of factors to consider when selecting the best for a spe-
cific application. The chemical nature of the analyte is important as is the ana-
lytical method which will be used to analyse the extract. One disadvantage of the
solid-based traps is that most subsequent analytical methods (e.g.
GC
or
HPLC)
require a solution and therefore it is necessary to then carry out a solvent extrac-
tion, albeit on a smaller scale, to get the analytes from the trap into solution.
Also if a solvent modifier
is
used, it may be impossible to use solid type traps as
the precipitated co-solvent washes the analyte from the trap and losses can oc-
cur. In such cases, it is probably best to use the modifier solvent as the trapping
solvent.
One way of studying trapping efficiency is the use
of
“spiked samples”. In
these, known amounts of pure components, usually in solution, are added either
to an inert support
or
to a “clean” sample of the matrix of interest. These Sam-
ples are then analysed using the proposed method and the recoveries of the
spiked analytes calculated. These type of results are often shown in publications
as a proof
of
complete extraction. This is not valid as “spiked samples” cannot
mimic the way in which real analytes interact with the matrix, particularly when
the samples have been aged and weathered as in many environmental samples
such as soil. However, they can be used to check the integrity of the extraction
and trapping systems and as such are very useful tools. In fact it is recommended
that this
is
done on a regular basis as a quality control check for routine analyti-
cal SFE methods.
One other approach, which is particularly useful for samples which will be
analysed by chromatography, is to employ internal standards which are added to
the sample
in
the cell and then used both as a means
of
quantification
of
the
analyte and as a means of checking the efficiency of the system plumbing and
trapping system. For example, in looking at the concentration of low molecular
weight oligomers in the range
C8
to
C20
in polyethylene, it is possible to employ
References
pp.
301-303
280
Chapter
10
an internal standard which contains three normal hydrocarbons
as
internal stan-
dards. This is possible because the oligomers are all even carbon numbers and
therefore all the odd carbon numbers are available
for
use as internal standards.
So
a
solution of known concentration of three normal hydrocarbons
of
odd car-
bon number which cover
the
range of interest, e.g.
nC9,
nC13, and nC19 is
added to the sample in the extraction cell and the extraction carried out. The re-
sulting extract
is
analysed and the response of the intermediate standard checked
against a direct analysis of the internal standard solution. This standard is then
used to calculate the concentration of the other components in the extract. The
efficiency
of
analyte trapping can then be assessed over the whole range of in-
terest by checking the recoveries of the other
two
internal standards.
10.4
SFE
TECHNIQUES
We
now consider some of the techniques employed to actually carry out the
extractions.
10.4.1
Static extraction
In static extraction, the sample cell
is
pressurized with fluid and the sample is
allowed to steep
in
the solvent until equilibrium is reached. An aliquot of the
extract is then removed,
or
the system is then depressurized and the analyte re-
covered. The main benefit
of
this method
is
that the fluid has time to penetrate
the matrix
.
It is most applicable when the analyte has a high affinity for the
sol-
vent and a low affinity for the matrix and also when the solubility limit of the
analyte in the fluid
is
much higher than the actual level reached during the ex-
traction.
10.4.2
Dynamic extraction
This
is
the most common mode
of
operation for analytical
SFE
and
a
typical
flow manifold for operation in this way
is
shown in Fig. 10.3. New fluid
is
con-
tinuously flushed through the sample and out through the back pressure device
into the trap. This
is
a particularly useful mode when the concentration of the
solute in the sample is high relative to its solubility in the fluid
or
when the sol-
ute has a high affinity for the sample matrix. Many modem instruments can op-
erate in both dynamic and static modes and often the best approach is to combine
both in the same method with a static step followed by a dynamic step.
Supercritical fluid extraction
FEED
PUMP
-'
J
FEED
FLUID
OVEN
EXTRACTION
CELL
3
OVEN
EXTRACTION
CELL
YV
SAMPLING
L
VALVE
Ak
1r
EXTRACTION
MONITOR
RECIRCULATING,-
PUMP
VALVE
EXTRACTION
MONITOR
28
1
-
EXTRACT
Fig.
10.4.
Schematic representation
of
a recirculating extraction manifold.
10.4.3 Recirculating extraction
This combines some of the features of both static and dynamic extraction. In
this mode, the sample cell
is
part of a loop which is pressurized with the super-
critical fluid. The loop also contains a recirculating pump which can pump the
fluid around the loop and over the sample as in dynamic extraction.
A
flow dia-
gram of a typical recirculating system is shown in Fig.
10.4.
Alternatively the sample can be allowed
to
steep in the fluid as in static ex-
traction. These steps are often alternated several times until the system reaches
equilibrium. The extract can then be collected by depressurization of the loop or
an aliquot of known volume can be removed via a sampling valve for further
analysis. This can be a useful means of studying the effect of temperature and
pressure on solubility as only a small volume is removed in relation to the total
loop volume. Therefore experiments can be carried out and the solubility calcu-
lated at several different temperatures and pressures on a single sample. This
method, however, is not generally available on commercial instruments and has
not found many published analytical applications.
10.4.4 Extraction
of
liquids
Conventional SFE extraction cells are often not suited for the direct extraction
References
pp.
301-303