Adlard E.R. (ed.) Chromatography in the Petroleum Industry
Подождите немного. Документ загружается.
332
Chapter
I1
ins
~
'\
4
Backflush
Aromatics
L
olefins was not as good as that achieved with the C02 and the silver-modified
column
[65],
the results obtained correlated reasonably well with
FIA
data.
A
pressure and temperature ramp
(2.30-340
atm, 50-150°C) had to be used after
elution of the saturates and olefins in order to elute the aromatics as
a
compact
band of peaks. Campbell
[67]
using identical separation conditions, except
for
a
larger surface area silica, achieved
a
much superior separation between saturates
and
olefins (see Fig.
11.8).
Campbell
[67]
also reported the use of
two
different
separation columns which effected
a
partial but different hydrocarbon type sepa-
ration on each column. Using 10%
C02
in
SF6
as the mobile phase and a silica
column, the unresolved saturates
+
olefins elute first and the aromatics are back-
flushed off (Fig.
1
1.9(a)). With
C02
as
the mobile phase and a silver-loaded
sul-
Supercritical
jluid
chromatography
333
phonic acid silica column, the saturates only are eluted first and unresolved ole-
fins
+
aromatics are backflushed off the column (Fig. 11.9(b)). Peak areas were
normalized and the olefin content determined by difference. In practice, this
system worked well although it is not suitable for low olefinic samples because
this hydrocarbon group is determined by difference.
This idea was then developed further into a multidimensional
SFC
system
enabling saturates, olefins and aromatics to be quantified in a single chroma-
tographic run of around
50
min (Fig. 11.10). The sample was injected onto the
silica column where the saturates
+
olefins were separated
from
the aromatics.
The aromatic fraction was isolated by valve switching while the saturates and
olefins were separated on the silver-loaded column. After backflushing the
ole-
fins off the silver-loaded column, the aromatics were switched back into the
system and eluted
[68].
The multidimensional method is applicable to higher
boiling samples including diesels but suffers from
two
practical disadvantages:
relatively complex and expensive instrumentation and the use of
SF6
which pro-
duces the toxic and corrosive
HF
in the FID. Moreover, the widespread use of
oxygenates and polar additives in fuel will have a detrimental effect on the sta-
bility of the silver-loaded columns.
Anderson et al.
[69]
reported a similar multidimensional
SFC
method which
used a different elution order (aromatics were eluted first while the saturates and
olefins were ‘parked’ on the silver-modified cation exchange column) for the
hydrocarbon types. This system uses
COz
as the mobile phase and therefore
overcomes objections regarding the use of SF6. By using micropacked capillar-
FID
Saturates
I
I
I
1
4.0
5.0
6.0
7.0
8.0
9.0 10.0 11.0
12.0
13.0
14.0
Mlnutes
Fig.
1
1.1 1
Dual
SFC
chromatograms
of
standard
gasoline
RR2.
References pp.
343-345
334
Chapter
11
ies, the analysis times are drastically reduced to around 6 min and the aromatic
fraction is separated into mono- and di-aromatics.
Schulz and Genowitz
[70]
proposed a much simplified SFC-FID procedure
using a single silica column,
COz
mobile phase and dual UVD/FID detection.
Samples are injected neat and analysed at constant density using forward elution
only. On this system, the aromatics are completely resolved whereas the satu-
rates and olefins are only partially resolved from each other (see Fig.
1
1.1
1).
The
UVD is operated at
190
nm where it functions as a quantitative olefin-selective
detector by virtue of its relatively uniform response to olefins. The UVD is cali-
brated against a standard olefin or a sample of known olefin content and the FID
acts as a general mass-sensitive detector for all three hydrocarbon types. Olefin
contents down to 0.02% can be measured and the results correlate closely with
FIA data. Pressure programming can be utilized to sharpen up the aromatic peaks
and
so
improve detection limits
[71].
ii.3.4.2
Kerosenes
and
naphthas
SFC methods developed for gasoline analysis are usually suitable for kerosenes
and naphthas. Aviation kerosenes must typically meet specifications on olefin,
total aromatic and naphthalene content. Olefins and total aromatics are deter-
mined by FIA while naphthalenes are measured by a spectrophotometric method
(ASTM D1840). A modified version of ASTM D5 186 (aromatic content in die-
sel fuel by SFC) has been proposed for the determination of aromatic types in
aviation fuel. An additional performance check, using tetralin and
2-
methylnaphthalene, has been included
in
the kerosene method to ensure suffi-
cient resolution between mono- and di-aromatics. Since the di-aromatic hydro-
carbon group contains several different compound types, but predominantly
naphthalenes and biphenyls, it may not be possible to correlate the SFC results
with those of ASTM
D1840.
Experimental design optimization studies for the hydrocarbon type analysis of
aviation fuels have indicated that the optimum resolution for saturates and mono-
aromatics has to be compromised
in
order to attain the resolution demanded for
the mono-/di-aromatic separation
[72].
Fortunately SFC conditions (55OC and
150 atm) were identified which met the resolution requirements and produced a
good separation of all three hydrocarbon types in a kerosene (Fig.
1
1.12).
Di Sanzo and Yoder [63] undertook a wide ranging validation procedure of
their SFC-FID method for jet fuels. The conditions used by these workers (silica
column, C02 at 30°C and 115 atm) favoured the separation between saturates
and mono-aromatics at the expense of the aromatic type separation,
in
agreement
with the optimization studies of Fraile and Sanchez
[72].
Jet fuel blends of
known saturate/aromatic content were prepared fiom pre-separated fractions and
analysed by
SFC.
A small positive bias (1-2% RSD) in the total aromatic con-
Supercritical
fluid
chromatography
335
li
AROMATICS
-1
I-
Fig.
1
1.12.
Separation
of
aviation fuel under optimized conditions.
tent by SFC was observed, possibly indicating slight differences in
FID
response
to
saturates and aromatics. The quantification of naphthalenes was evaluated by
spiking
known
amounts of a dimethylnaphthalene mixture into jet fuel saturates
before analysing by SFC. The agreement between actual and measured values
was excellent.
LC methods use amino-bonded or polar amino/cyano-bonded stationary
phases to minimize the effects of alkylation on the retention characteristics of
aromatic compounds. This aspect of stationary phase selectivity has not been
explored to any great extent in SFC. Figure
1
1.13
illustrates how these charge
transfer type columns. when operated under high density(1iquid-like) conditions,
reduce the influence of aromatic alkylation to give more compact peaks which
are easier to integrate and give lower detection limits
[73].
Olefin contents are very low in aviation fuels and are difficult to quantify ac-
curately. The dual
UVD/FID
method of Schulz and Genowitz
[70]
may have
sufficient sensitivity as Fig.
11.14
shows.
References pp.
343-345
336
Chapter
1
I
I
I
I*
11
e
8%
*I"#'
10.0
12,s
15.0
17.5
2.5
5.0
7.5
Il..
I.l""I..l
.o
.o
1
Fig.
11.13.
Partisil PAC
column
at
240
bar/50°C
or
125
bad40"C;
250
X
2
mm
ID,
COz
FID,
aviation
fuel.
11.3.2.3
Diesel
fuels
The first standard method
for
hydrocarbon type analysis
of
diesel fuels, based on
HPLC
with refractive index detection. was published in
1990
[74].
At the time
Supercritical fluid chromatography
337
One-Ring
Aromatics
i
4.0
5.0
6.0
7.0
8.0
9.0
10.0 11.0 12.0 13.0 14.0
Mlnules
Fig.
I
1.14.
UV
and
FID
chromatograms of
a
jet fuel (jet fuel
1
190-208
ASTM).
the HPLC method was undergoing development and validation in Europe, SFC-
FID methods were also under study in the
US.
In 1992,
ASTM
D5186 was pub-
lished as the first standard SFC method. Both the HPLC and the SFC methods
were developed to meet
an
urgent analytical requirement, viz the determination
of
aromatic content of diesel fuel. Aromatic compounds in diesel fuel are gen-
erally considered responsible for the emissions from diesel engines. Legislation
to control and specify the maximum aromatic content allowable in diesel fuel is
already
in
force in come countries. Olefins may cause coating of injection noz-
zles and valves leading to impaired combustion and more toxic emissions.
Many laboratories use a simple column and follow a method similar to or
identical to that prescribed in
ASTM
D5 186. Lee et al. [62] reported that a single
silica column with C02 as the mobile phase gave repeatable results which corre-
lated with FIA data. The aromatics elute over a relatively broad range of reten-
tion times which can
make
integration difficult particularly for low aromatic die-
sels. Poor precision can often be traced to difficulties in the integration step
when the hydrocarbon
group
is spread out along the baseline. Di Sanzo and
Yo-
der [63], using a similar SFC set-up, also reported problems determining accu-
rate aromatic contents when diesel fuels contained less than
10%
mass aromat-
ics; errors were attributed to incomplete separation of saturates and aromatics.
These workers used specially prepared blends of diesel saturates and aromatics
to confirm the excellent accuracy and precision of the SFC method over the
range
10-50%
mass aromatic content. Chen and colleagues
[75]
compared SFC
References pp. 343-345
338
Chapter
1
I
Analysis
Name
:
[RSXI
1
DIESELS,
3.
1.
OIESELf
.
Amount
:
1.000.
L
~~"""'i"""'i"'
2.0
3.0
4.0
5.0
Tim
hinutea)
.o
Instrument
:
Method
:
SFC
Channel Title
:
Calibration
:
NONAHE
Lims
ID
:o
Run
Sequence
:
SFC
Acquired on
13-FIAR-1988
at
16:31
Reported on
19fEB-igW
€it
17:38
Fig.
1
1.15.
SFC-FID
analysis
of
diesel
fuels
(100
mm
X
4.6
mm
ID
3
pm SiO, column) with col-
umn backflush to elute the aromatics.
aromatic ring type distribution with D2425
MS
data obtained on the aromatic
fractions isolated by
LC.
The
SFC
results showed good agreement with the
gra-
vimetric aromatic contents and with mono- and di-aromatic data from the
MS
analysis; the relationship with tri-aromatic contents was fair.
It may be preferable to analyse low aromatic fuels using pressure program-
ming to compress the aromatics peak
[71];
the effect
of
the increased
C02
flow
on
FID
response should, however, be checked out. Alternatively aromatics can
be backflushed off the column as a sharp band for easy integration and quantifi-
cation (Fig.
1
1.15). Large diameter columns (4-5 mm
ID)
operated close to or
above their optimum linear velocities yield rapid efficient separations (Fig.
1
1.16) and are a practical alternative to the more extensively used microbore
columns. Dual
UV/FID
detection can be employed because only a small portion
of
the effluent
is
sent to the FID; the
UV
detector is a useful adjunct
as
it pro-
vides a fingerprint of the aromatic distribution and helps to define the
satu-
rate/aromatic cut point
in
low aromatic content samples. The analysis of sulphur
compound types
in
diesel fiiels using a silica/C02 separation with dual
SCDKJVD
has been described [42c].
Supercritical
jluid
chromatography
339
Ti
lmin)
Fig.
11.16.
Aromatic content in diesel fuel. SFC with dual detection. Conditions: column,
4
x
250
mm
silica
(5
pm); mobile phase,
C02;
flow rate,
5
ml/min; column temperature, 30°C; outlet
pressure,
150
bar.
Fraile and Sanchez [72], using experimental design optimization, have defined
the conditions of pressure and temperature required to meet the D5 186 resolu-
tion requirement for the separation between docosane and toluene. Maximum
resolution was achieved at 190 atm and 32°C although a resolution
>4
was at-
tained if pressures >120 atm and temperatures <40"C were used. These limits
show good agreement with empirically derived conditions for diesel analysis. A
recent study has noted that silica stationary phases with surface areas >350 m2/g
are more likely to meet the D5186 resolution requirement than low surface area
silicas [75].
Multidimensional approaches to diesel fuel hydrocarbon type speciation are
also documented. Fuhr and co-workers [64] reported the use of a dual column
system (silica and amino-bonded silica stationary phases) for the analysis of
aromatic types by SFC-FID (C02/4000 pd35"C). Diluted fuel samples were in-
jected onto the silica column to effect the saturates/aromatics separation; after
elution of the saturates to the FID, the aromatics were diverted onto the amino-
bonded column to give an improved aromatic-type separation. The validity of the
aromatic-type separation was confirmed by SFC analysis of both model com-
pounds and diesel mono-, di- and poly-aromatic fractions isolated by preparative
LC (Fig. 11.17). Analysis times were roughly three times longer than those
on
a
single silica column and the additional band broadening of the PAH fraction
could give rise to integration problems. Approximate FID response factors of
1.00,
1.05 and 0.87 were found for the mono-, di- and polyaromatics, respec-
tively; the authors considered these to be sufficiently close to allow aromatic
References pp.
343-345
340
Chapter
I I
r
I
I
I
I
0
5
10
15
20
lime
(min)
Figure
11.17.
SFC
chromatograms
of
(a) a whole middle distillate and
(b)
fraction
1,
(c) fraction
2,
and (d) fraction
3
obtained
from
the whole middle distillate. Conditions are described in the text.
type contents to be determined using area normalization. Repeatabilities for all
aromatic types were around
1%
mass which
is,
not unexpectedly, poorer than
that achieved for total aromatics
(0.4%
mass) using a single silica column
[62].
The
SFC
results compared favourably with aromatic-type data determined by
a
MS
matrix method.
Anderson et
al.
[61,69,76]
have described a number
of
multidimensional
SFC-
FID systems using micropacked capillaries and carbon dioxide mobile phase.
The first instrument used
two
columns (silica and silver-modified cation
ex-
change) to separate saturates, olefins and aromatic-types in under
10
min
[53].
Separations were performed isobaricaily
(300
atm) and isothermally
(65°C).
A
higher column temperature was required to reduce the strength
of
the olefin/Ag+
interaction and
so
allow the olefin group to be eluted
as
a narrow band.
To
com-
pensate for the higher column temperatures, higher column pressures were re-
quired to maintain densities of
>0.82
g/ml. The success of this method is criti-
cally dependent on the availability
of
well-packed capillary columns which are
stable under conditions
of
repeated backflushing. The next generation instrument
[76]
introduced an additional cyano-bonded silica column to isolate and quantify
polar compounds in diesel. A schematic of the valve and column arrangement is
shown in Fig.
11.18.
Higher column temperatures and pressure
(75"C,
375
atm,
density
0.81
g/ml) were used to ensure rapid elution of the olefins from the sil-
ver-loaded column.
SFC-MS
studies were carried out using
a
micropacked silica
50
pm
capillary column, which generated only 50pVmin of gaseous
C02
even at
Supercritical fluid chromatography
341
t
8i
w-
Poe.
c.
J
Pca.
D.
Si
Fig.
1
1.18.
Schematic diagram
of
the column-switching system. (A) Separation
on
the
CN
and mSi
columns, polar compounds retained
on
the
CN
column, saturates and alkenes transferred to the Ag
column;
(B)
aromatics transferred from Si column to flame ionization detector;
(C)
polar com-
pounds backflushed from the
CN
column;
(A)
saturates eluted from the Ag column;
(D)
alkenes
backflushed from the Ag column.
References
pp.
343-345