Adlard E.R. (ed.) Chromatography in the Petroleum Industry
Подождите немного. Документ загружается.
3
02
Chapter
I0
16
17
18
19
20
21
22
23
24
25
26
27
28
79
30
31
32
33
34
35
36
37
38
39
40
41
42
D.R. Gere, R. Board and D. McGinigill, Anal. Chem., 54 (1982) 736.
T. Greibokk,
AL.
Blillie, E.J. Johansen and E. Lundanes, Anal. Chem., 56 (1984) 2968.
J.M. Levy, Advances in Analytical SFE, American Laboratory, August (1991).
K.H. Pickel? Proceedings
of
the 2nd International Symposium on Supercritical Fluids, Boston
(1991) p. 457.
M. Saito and Y. Yamauchi, Supercritical Fluid Extraction and Supercritical Fluid Chromatog-
raphy for Preparative Separation, Jasco Report Publisher, Japan (1990).
B.W. Wright, C.W. Wright, R.W. Gale and R.D. Smith, Anal. Chem., 59 (1987) 38.
H.M. McNair and
J.O.
Frazier, American Laboratory (1991) 24d.
L.J. Mulcahey, J. Hedrick and L.T. Taylor, Anal. Chem., 63 (1991) 2225.
M.
Ashraf-Khorassani, R.K. Houck and J.M. Levy, J. Chromatogr. Sci., 30 (1992) 361.
N.L. Porter, A.F. Ryanski, E.R. Campbell, M. Saunders, B.E. Richter, J.T. Swanson, R.B.
Nielsen and B.J. Murphy, J. Chromatogr. Sci., 30 (1992) 367.
J.L.
Hedrick and L.T. Taylor, Anal. Chem.,
61
(1989) 1986.
J.L. Hedrick,
L.
J. Mulcahey and L.T. Taylor, Supercritical Fluid Technology: Theoretical and
Applied Approaches in Analytical Chemistry, Chapter 15 (F.V. Bright and M.E.P. McNally,
Eds.), ACS Symposium Series
488
(1992) ISBN 0-8412-2220-7.
Supercritical Fluid Technology Synopsis SFE Application: PCBs in Oil. SFE-82 (1991),
Suprex Corporation. Pittsburgh,
PA.
D.C. White, D.E. Nivens, D. Ringelberg and D. Hedrick, Proceedings of the International
Symposium on Supercritical Fluid Chromatography and Extraction, Park City,
UT
(1991)
p.
43.
S.B.
Hawthorne, D.J. Miller, D.E. Nivens and D.C. White, Anal. Chem., 64 (1992) 405.
D.J. Miller,
S.B.
Hawthorne,
J.J.
Lagenfield and D.C. White, Proceedings
of
the International
Symposium on Supercritical Fluid Chromatography and Extraction, Park City, UT (1991).
p.
155.
J.J.
Hills, H.H. Hill and T. Maeda, Proceedings
of
the International Symposium on Supercriti-
cal Fluid Chromatography and Extraction, Park City, UT (1991) p. 113.
J.J.
Hills. H.H. Hill and T. Maeda, Anal. Chem., 63 (1991) 2152.
C.P. Eppig,
B.M.
Putnam and R.P. de Fillippi, US Patent application no. 255,037, filed April
17th (1981).
Supercritical Fluid Technology Synopsis, SFE-87, Suprex Corporation, Pittsburgh, PA
(1991).
Off-line supercritical fluid extraction with the Dionex SFE-703, Dionex Corporation, Salt
Lake Division.
B. Subramaniam and
S.
Saim, J. Supercritical Fluids, 3 (1990) 214.
S.
Saim
,
D.M. Ginosar and B. Subramaniam, Proceedings of the 2nd International Sympo-
sium on Supercritical Fluids, Boston, MA (1991) p. 137.
S.A. Schmidt, R.E. Yoder and J.V. Smith, Supercritical Fluid Extraction
of
Refinery Bottom
Sludges and the Analysis for Polynuclear Aromatic Hydrocarbons, Application Note 228-1 19.
November
1990,
Hewlett-Packard Company.
G.W.
Kelly, K.D. Bartle and A.A. Clifford, Proceedings
of
the 4th International Symposium
on
Supercritical Fluid Chromatography and Extraction, Cincinnati,
OH
(1992) p. 139.
R. Flaherty and
J.V.
Smith, Hewlett-Packard application note 228-133, Supercritical Fluid
Extraction of Filter Media. April 1991.
A.
Lopez-Avilla,
J.
Benedicto. N.S. Dodhiwala and R. Young,
J.
Chromatogr. Sci., 30 (1992)
335.
Supercritical fluid extraction
3 03
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
Determination
of
Total Petroleum Hydrocarbons (TPH)
in
Soil, Application Note Ref RC-5,
Suprex Corporation, Pittsburgh,
PA
(1991).
M. Richards and R.M. Campbell, LC-GC Int., 4 (1991) 33.
N.J. Cotton, K.D. Bartle, A.A. Clifford,
S.
Ashraf, R. Moulder and C.J. Dowle, J. High Reso-
lut. Chromatogr., 14 (1991) 164.
H. Daimon and
Y.
Hirata, Chromatographia, 32 (1991) 549.
M. Ashraf-Khorassani, D.S. Boyer and
J.M.
Levy,
J.
Chromatogr. Sci., 29 (1991) 517.
R.W. Wright, S.R. Frye, D.G. McMinn, R.D. Smith, Anal. Chem., 59 (1987) 640.
J.C. Monin, D, Barth, M. Perrut, M. Espitalie and
B.
Durand, Report Ref 35604, Institute
Francais du Petrole (1987).
S.B. Hawthorne and D.J. Miller, J. Chromatogr. Sci., 403 (1987) 63.
S.B. Hawthorne, D.J. Miller and M.S. Kreiger,
J.
Chromatogr. Sci., 27 (1989) 347.
S.B. Hawthorne, D.J. Miller and M.S. Kreiger, Anal. Chem., 60 (1988) 472.
S.B. Hawthorne, D.J. Miller, M.S. Kreiger, Fresenius
Z.
Anal. Chem., 330 (1988) 21
1.
J.M. Levy,
J.P.
Guzowski, W.E. Huhak,
J.
High Resolut. Chromatogr., Chromatogr. Com-
mun., 10 (1987) 337.
J.M. Levy and J.P. Guzowski, Fresenius
Z.
Anal. Chem., 330 (1986) 207.
J.M. Levy and A.C. Roselli, Chromatographia, 28 (1986) 613.
T.P.
Lynch, R.E.A. Escott, P.G. McDowell,
I.
Roberts and M.J. Carrott, Proceedings
of
the
4th
International Symposium
of
Supercritical Fluid Chromatography and Extraction, Cincin-
nati, OH (1992) p. 179.
T.P. Lynch, Applications
of
Supercritical Fluids in Industrial Analysis (J.R. Dean, Ed.),
Blackie, Glasgow (1993) Chapter
8,
p. 208, ISBN 0-7514-057-2.
This Page Intentionally Left Blank
E.R.
Adlard
(Ed.),
Chromatography in the Petroleum Industry
Journal
of
Chromatography
Libmy Series, Vol.
56
0
1995
Elsevier Science B.V. All rights reserved
305
CHAPTER
11
Supercritical
fluid
chromatography
I.
Roberts
Analytical
&
Applied Science Division,
BP
Research
&
Engineering Centre, Sunbury-on-Thames, Middlesex,
TWl6
7LN
UK
11.1
INTRODUCTION
The term supercritical fluid chromatography (SFC) was introduced by Sie and
Rijnders in 1967 [I], some 9 years after Lovelock had first proposed using su-
percritical carrier gases for the analysis of non-volatile compounds by gas chro-
matography (GC). In 1962 the first published ‘SFC’ application appeared de-
scribing the separation of porphyrins on a packed GC column using supercritical
chlorofluoroalkanes as the mobile phase [2].
The petroleum industry immediately recognised the possibilities offered by
this new form of chromatography. Sie and Rijinders carried out much of the
early pioneering work at the Shell Laboratories in Amsterdam
[3,4]
while, at
Chevron, Jentoft and Gouw demonstrated the use of polar modifiers and pressure
programming as a means of controlling retention in SFC [5,6]. Although work
on SFC continued during the 1970s, particularly at Chevron [7], the oil industry
lost interest in SFC as it focussed its attention on the emerging technique
of
high
performance liquid chromatography (LC).
In
an
earlier review, Gouw and Jentoft described the status of
SFC
in petro-
leum analysis as it stood in 1978
[8].
Separations were performed on packed
columns containing large particle-sized stationary phases, detection was pre-
dominantly by
UV
and typical mobile phases were alkanes (e.g. pentane, hex-
ane) and oxygenated solvents (e.g. propan-2-01, diethyl ether). Although carbon
dioxide was in use as a mobile phase, it did not occupy the dominant position it
does today. Open tubular capillary columns had not yet arrived on the scene and
instrumentation was still built in-house. The GC origins
of
SFC were clearly
evident in that the main benefits of SFC were perceived to be the separation of
References
pp.
343-345
306
Chapter
I
I
components with ‘low GC migration rates and/or low thermal stabilities’. Inter-
estingly, this review did not mention supercritical fluid extraction (SFE) but did
give an example of SFC directly coupled to mass spectrometry (SFC-MS).
The beginning of the 1980s saw renewed interest in SFC spurred on by the
development of open tubular capillary columns [9] and the introduction of the
first commercial SFC instrument, a modified Hewlett-Packard liquid chroma-
tograph, in 1982.
In
1986 the first commercially-built capillary supercritical fluid
chromatograph appeared. This period of rapid growth, up to the present day, has
seen many improvements
in
instrumentation, particularly injector and detector
technology, and on-line coupling with SFE, GC, LC,
MS
and infrared
(IR).
Capillary and packed columns are now more robust and more efficient while
developments in the design and synthesis of stationary phases has produced op-
timised SFC phases with a range selectivities. The theoretical basis of SFC is
now well established
[
101.
In
spite
of
the technical success of SFC, there has been a reticence within the
general scientific community to accept it as an equal partner with GC and LC.
Some of the blame for this must be attached to exaggerated claims, made in the
early 1980s, that SFC would make GC and LC redundant. Perhaps there was too
much emphasis on the theoretical aspects, during these early development years,
at the expense of demonstrating the problem-solving powers of SFC (cf. the de-
velopment of GC and LC). Certainly a large number
of
the published SFC appli-
cations dealt with analytical solutions which were more readily achievable by
GC or LC, and encouraged the sceptics to dismiss SFC as ‘a technique looking
for an application’. Nowadays SFC is rightly regarded as a complementary, if
still minor, partner
in
the GC/SFC/LC triad of column chromatography.
The modern revival in petroleum SFC may be traced to a paper by Rawdon,
published early in 1984, describing a modified flame ionization detector (FID)
for packed columns
[I I].
The significance
of
this paper was that petroleum ana-
lysts could now perform quantitative hydrocarbon-type analyses which had pre-
viously proved elusive using LC techniques. The example Rawdon gave in his
paper was the separation of diesel fuel into saturates and aromatics; some 9 years
later the first standard SFC test method (ASTM D5186-91: Determination
of
Aromatic Content of Diesel Fuels by Supercritical Fluid Chromatography) was
published
[
121.
11.2
INSTRUMENTATION
The supercritical fluid chromatograph is
a
hybrid instrument assembled from GC
and LC components. The basic elements of the SFC instrument are illustrated
schematically in Fig.
1
1.1.
Mobile phase, in the form of a liquid or pressurized
Supercritical
fluid
chromatography
307
r-----
co2
detector
1
I
-_-
0
restrictor
OVEN
X)
detector
LC)
-----a
_-__
___
__
restrictor
i-
Fig.
ll.
1.
Schemata
of
SFC
chromatography.
gas, is supplied to the LC pump where the pressure of the fluid is raised above
its critical pressure. Samples are introduced into the fluid stream via an LC in-
jection valve and separated on a column placed in a GC oven thermostatted
above the critical temperature of the mobile phase. A post-column restrictor en-
sures the fluid is maintained above its critical pressure throughout the separation
process. Analytes eluting from the column are monitored by detectors positioned
either before or after the post-column restrictor.
Prior to the mid-l980s, laboratories designed and assembled home-made SF
chromatographs using readily available LC (pumps, injectors, columns and de-
tectors) and GC (ovens, columns, detectors) components. The modifications re-
quired to build a working SFC instrument are relatively minor
so
that this ap-
proach can still provide
a
cheap introduction to SFC by using redundant LC and
GC
parts. For industrial laboratories, it is more cost-effective to invest in pur-
pose-built instrumentation which is now available from a number of suppliers
and manufacturers. Commercial SF chromatographs offer sophisticated micro-
processor-controlled units (allowing the simultaneous programming of fluid
pressure, temperature, flow and composition), a range of detector options, auto-
mated sample injection and on-line coupling with SFE. Of equal importance
is
the back-up (maintenance, spare parts, consumables and practical advice) pro-
vided by the manufacturers.
11.2.1
Mobile phase pumps
The specifications for an SFC pumping system are similar to those for LC, i.e. it
should provide a stable pulseless flow over the operating pressure range
References pp. 343-345
308
Chapter
11
(typically
70-400
bar). Pumps designed to operate at higher pressures (e.g. up to
650
bar) are available but their use is often restricted by pressure limitations on
other system components such as injection valves. Since flow rate requirements
vary from a few
pl
min-I in capillary SFC to several ml min-l in packed column
SFC,
two
different pump types are used to cover this range.
11.2.1.1
Reciprocatingpumps
Generally the cheapest option for SFC, these conventional LC pumps supply
stable flow rates over the range
0.5-5
ml
mid, making them ideal for use with
packed columns having internal diameters
>2
mm. For efficient pumping, the
pump heads must be cooled to 0°C or less by means of a clamp-on heat ex-
changer and a recirculating chiller. Stream-splitting may be used to reduce the
flow rate but this
is
not a reliable means for producing the very low flow rates
required for capillary SFC unless flow and pressure can be precisely and accu-
rately maintained through a programmable, electronically-controlled variable
restrictor
[
131.
A
major advantage of reciprocating pumps is the ease with which liquid
modifiers can be added to the main fluid using an additional pump or pumps (the
modifier pump does not need to chilled). Standard high pressure LC mixers en-
sure rapid and reproducible mixing of the mobile phase streams
so
that compo-
sitional gradients can be used to enhance selectivity or elute components of
widely varying polarity.
The check valves in the chilled pump head need to be replaced more fre-
quently than their LC counterparts but, otherwise, reciprocating pumps offer low
maintenance and trouble-free operation.
11.2.1.2
Syringepumps
The ability of a syringe pump
to
provide a virtually pulse-free flow in the
1-
10
pl
min-' range makes them ideal for capillary SFC. They are also suitable for
use with the larger diameter packed columns provided the capacity of the syringe
(typically
20-500
ml) is adequate.
Syringe pumps are frequently used with packed columns up to
1
mm
ID;
pis-
ton seal wear can become significant if flow rates of several ml min-' are used
routinely.
Efficient filling of the pump is achieved either by cooling the pump cylinder
to
around
5°C
or by using gas cylinders in which the mobile phase
is
overpres-
surized with helium at
100-125
bar. The former method is preferred since the
presence
of
even small amounts of helium in the mobile phase can have a detri-
mental effect on chromatographic reproducibility in capillary
SFC
[
141.
The fundamental design of syringe pumps (large cylinder volumes and no
check valves) combined with high fluid compressibilities gives rise to a major
Supercritical fluid chromatography
3
09
drawback, the inability to generate mobile phase composition gradients by add-
ing a liquid (low compressibility) modifier to the highly compressible fluid [15].
Additionally, the least compressible liquid may force its way back into the tub-
ingkylinder of the highly compressible fluid giving rise to cross contamination
problems and inaccurate or variable solvent compositions.
A
single syringe pump can be used to deliver
a
mixed solvent, under isocratic
conditions, either by using cylinder bottles pre-doped with modifier (e.g. 10%
m/m methanol in carbon dioxide) or by pre-mixing the liquid modifier and fluid
in the pump cylinder [16,17]. The practical difficulties in generating mixed
mobile phases with syringe pumps are only a minor inconvenience for the petro-
leum analyst where multicomponent fluids are the exception rather than the rule.
Modern syringe pumps, although more expensive, are extremely reliable and
easy to maintain. The piston drive mechanism requires regular lubrication to
prevent corrosion problems arising from water condensation (due to cylinder
cooling) and the high pressure piston seal needs to be replaced approximately
every 6 months.
11.2.2 Ovens/temperature control
The role of the oven in SFC is to raise and maintain the temperature of the mo-
bile phase, in a controlled and repeatable manner, above its critical temperature.
Standard GC ovens meet all the requirements for temperature management in
SFC by providing accurate, precise and uniform (no localized hot or cold spots)
control of temperatures, including temperature programming, from ambient to
300°C. Most commercial SFC instruments are supplied with GC-type ovens
so
that they can also operate in the GC mode. Many home-made
SF
chroma-
tographs utilize redundant GC ovens, not only for their temperature control
function, but because they have additional in-built features such as detectors,
heated injectioddetection zones and a large oven cavity. The temperature pro-
gramming facility, although not widely used in petroleum SFC applications, is
an area worthy of further investigation.
A
number of cheaper options are available for packed column SFC carried out
under isothermal conditions (the most common requirement for petroleum SFC
applications). Circulating air ovens and block column heaters designed for LC
provide excellent temperature control and even a simple oil or water bath is ade-
quate.
11.2.3 Injectors
Sample introduction is often the most critical part of any chromatographic sys-
tem. The injection system aims to transfer a concentrated (minimum band
References pp. 343-345
310
Chapter
11
spreading), representative (i.e. no sample discrimination) and repeatable (i.e.
quantitative) aliquot
of
the sample to the head of the column without disturbing
the chromatographic equilibrium. In practice there is no ideal injection mecha-
nism for SFC and the injection mode that works best for one sample does not
necessarily
work
best for others.
The standard LC injection valve has monopolized the introduction of liquid
samples (including solutions) onto both capillary and packed columns, even
though the amount of sample injected onto a packed column (e.g.
1Opl)
may be
three orders
of magnitude greater than that injected onto an open tubular capil-
lary column (e.g.
10
nl). Commercial LC injectors offer fixed volume injections
from 60nl upwards, allowing the
hi1
loop volume to be injected onto packed
columns
(>OS
mm
ID) with excellent repeatability
(4%
RSD)
[15].
Sample
injection onto packed columns presents no real problems provided the analyst
follows standard chromatographic practices.
For
capillary columns, injection volumes smaller than 60 nl are preferred and
various methods have been devised to achieve this. The dynamic split injection
method uses a two-way splitter, positioned between the injector and column, to
send most of the sample (e.g.
99%) to waste and only a small aliquot onto the
column. Timed-split injection relies on the ability to switch the injection valve
from load to inject and back to load over very short time periods (tens of milli-
seconds)
so
that only a portion of the loop volume is pushed on to the column by
the mobile phase. Both methods are available on commercial instruments and are
capable of good repeatability, especially if internal standards are used. Overall,
timed split injection has a slight edge over dynamic split by virtue of its better
repeatability, lower sample discrimination, reduced solvent tailing and more
flexible adjustment of injection volume.
A
disadvantage of the split injection techniques is that complex samples, of
which petroleum is a prime example, must be injected as concentrated solutions
because of the very small volumes transferred to the column. Preparing concen-
trated solutions may require the use of either
a
‘strong’ solvent, which may not
be compatible with the mobile or stationary phase, or heating the solution to im-
prove solubility. In the latter case, the injection valve must also be heated to en-
sure the sample does not precipitate in the valve and cause problems with sample
discrimination and carry-over.
A
number of solvent elimination
(or
venting)
techniques have been developed to allow the injection of more dilute sample
solutions.
A
fuller discussion of these techniques, which bring trace analysis into
the realm
of
capillary SFC, may be found elsewhere
[
15,181.
11.2.4
Detectors
One of the major attractions of
SFC
is the ability to ‘mix and match’ detectors
Supercritical jluid chromatography
311
which hitherto were only available exclusively to GC or LC. Most GC and LC
detectors can be used with little or no modification on both packed and capillary
column systems. LC detectors normally require a high pressure flow cell for use
with packed columns and a degree of miniaturization for capillary SFC. GC de-
tectors can be used without modification on capillary systems but some allow-
ance for the higher flow rates must be made for packed column SFC applica-
tions. In column chromatography there are
two
basic detector designs: (i) liquid
phase detectors where the analyte remains in solution, and (ii) gas phase detec-
tors where the analyte is vaporized. In SFC the former (usually LC detectors) are
located before the restrictor and the latter (mainly GC detectors) are positioned
after the restrictor.
I
1.2.4.1 Universal detectors
Simple detectors whose primary function is to ‘sense’ all analytes eluting from
the chromatographic column are termed universal detectors. They provide mini-
mal chemical or structural information on the eluate but are routinely used for
quantification. More sophisticated instruments, such as mass spectrometers and
Fourier transform infrared spectrophotometers, can also function as universal
detectors although, for reasons of cost, this is not their primary role.
(i)
Flame ionization detector
(FID).
The FID
is
the most popular detector in SFC
due, at least in part, to its widespread use for petroleum analysis
[19].
As
the
first-choice detector for quantitative analysis of petroleum hydrocarbons it has
played a major role in the development of SFC within the oil industry. Apart
from optimizing the flame gas conditions
[20],
the standard FID can be used
without modification on both packed and capillary columns. Large bore packed
columns
(>2
mm
ID)
require post-column flow splitting to avoid extinguishing
the flame
[11,21].
With carbon dioxide as the mobile phase, FID sensitivity and
linearity are comparable to those achieved in GC.
Nitrous oxide, sulphur hexafluoride and xenon are also compatible with the
FID. Nitrous oxide produces a much higher background signal than carbon diox-
ide, giving a sensitivity reduction of several orders of magnitude. Nickel- or
gold-plated collectors must be used with sulphur hexafluoride to counteract the
corrosive action of hydrogen fluoride formed in the FID flame. Xenon is used
exclusively with flow cell
IR
detectors, for reasons of cost, but may subse-
quently be directed to an
FID
for quantification.
The sensitive and universal response of FIDs to organic compounds effec-
tively precludes the use of organic modifiers with carbon dioxide or other FID-
compatible fluids. Formic acid and water, which are sparingly soluble
(4%)
in
carbon dioxide, are the only
two
polar modifiers which have been used success-
References pp. 343-345