Adlard E.R. (ed.) Chromatography in the Petroleum Industry
Подождите немного. Документ загружается.
52
Chapter
2
slices rather than peak areas
[36].
A
computer program then compared data from
the fresh oil with that from the used oil, and from the difference in areas the per-
centage of fuel
in
the lube was calculated. The computer program was very
so-
phisticated, being able to compensate for differences in retention time and in
sample size.
Cartellieri and Tritthart used
GCD
to analyse the oil soluble fraction of par-
ticulates emitted by diesel engines
[37]
so
that they could estimate the propor-
tions attributable to the fuel and lube. They did this by constructing
a
calibration
graph of the
50%
points of a range of fuel/lube mixtures against composition.
The
50%
points for samples were determined and the proportion
of
fuel and lube
obtained from the graph. It was necessary to determine
a
calibration graph for
each fuel and lube being used (nine standards), but by using an autosampler the
authors were able to analyse
a
large number of samples.
2.7.
CONCLUSIONS
It
is
now over
30
years since the first report
of
GC distillation
[2].
The tech-
nique has been very successful in that it is now in widespread use and is used for
quality control
in
hundreds
of
locations. It
is
usually the method of choice for
unknown samples such as spillages, since it can be run quickly and cheaply.
However, the technique has not replaced older methods of physical distillation in
specifications, and
in
that respect has not attained its full potential. The major
reason for this
is
the lack of good reproducibility data discussed earlier, and it is
to be hoped that
this
problem will be solved. Another major challenge is the de-
velopment
of
a viable method for crudes, which would have large economic
benefits
in
many laboratories. The introduction of
SFC
has extended the scope of
GCD,
and provided that good, easy to use instrumentation becomes available, it
should have a bright future.
2.8
REFERENCES
t
2
3
ASTM
methods can be found in the current edition
of
ASTM standards.
F.T.
Eggerston,
S.
Groenings and
J.J.
Holst, Anal. Chem., 32 (1960) 904.
K.D.
Butlcr. in
K.11.
Altgelt and T.H.
Gouw
(Eds), Chromatography in Petroleum Analysis,
Marcel Dekker. New York (1979).
Institute
of
Petroleum sub-panel ST-G-6B, unpublished work.
D.J.
Abbott,
J.
Chromatogr. Sci., 21 (1983) 425.
I-I.E.
Schwartz,
R.G.
Brownlee, M. Boduszynski and
F.
Su,
Anal. Chem., 59 (1987) 1393.
1.G. Southern, A. Iachelli, D. Cuthiell and
M.L.
Selucky,
Anal.
Chern., 57 (1985) 303.
I..
Luke and
J.E.
Ray,
J.
High Res. Chromatogr.,
8
(1985) 193.
S.
Trestianu,
G.
Zilioli, A. Sironi,
C.
Saravalle,
F
.Munari,
M.
Galli,
G.
Gaspar,
J.M.
Colin
and
J.L.
Jovelin,
J.
High Res. Chromatogr.,
8
(1985)
771.
Advances in simulated distillation
53
10
A.D. Bashall,
Am.
Lab.,
19
(1987) 56.
11
A. Rastogi, J. High Res. Chromatogr.,
10 (1987) 479.
12
N.R. Warren and B.M. Lawrence, Lab. Pract.,
36 (1 987) 80.
13
F. Noel, J. High Res. Chromatogr.,
11 (1988) 837.
14
R.L. Firor,
Am.
Lab.
21(2A) (1989) 32.
15
R.L. Firor and R.J. Philips,
J.
High Res. Chromatogr.,
12, (1 989) 181.
16
P.S. Spock, M.V. Robillard and R.E. Long, Adv. Inst. Cont.,
45 (1990) 593.
17
M. Dorbon,
S.
Lamaison and A. Chevalier,
J.
Chromatogr.,
557 (1991) 155.
18
M.D. Kiser and D.P. Malone,
Am.
Chem. SOC. Div. Fuel Chem.,
27 (1982) 114.
19
R.B. Pannell and
A.
Sood, J. Chromatogr. Sci.,
20 (1982) 433.
20
N. Petroff, A. Hoscheitt and J.P. Durand,
J.
Chromatogr.,
395 (1987) 241.
21
J.L. Buteyn and
J.J.
Kosman,
J.
Chromatogr. Sci.,
28 (1990) 19.
22
K. Zuber and P. Bartl, Fuel,
68 (1989) 659.
23
0.
Glinzer, Erdoel und Kohle,
38 (1985) 213.
24
J.
Curvers and P. van den Engel, J. High Res. Chromatogr.,
12 (1989) 16.
25
Chrompack, Middelburg, The Netherlands.
26
H.E.
Schwartz, J.W. Higgins and R.G. Brownlee, LC-GC
4 (1986) 639.
27
H.E. Schwartz,
J.
Chromatogr. Sci.,
26 (1988) 275.
28
J.S. Thomson and A.F. Rynaski,
J.
High Res. Chromatogr.,
15 (1992) 227.
29
F.P.
DiSanzo, C.J .Lowther, R.E. Yoder and J.L. Lane,
J.
Chromatogr. Sci.,
26 (1988) 39.
30
D.J. Abbott, J. High Res. Chromatogr.,
7 (1984) 577.
3 1
D.R. Deans, Chromatographia
1
(1968) 18.
32
J.B. Maynard and W.A. Michalik,
J.
Chromatogr. Sci.,
26 (1988) 290.
33 S.
Coulombe and G. Duquette,
Am.
Chem.
SOC.
Div. Fuel Chem.,
33 (1988) 920.
34
G.B. Levy,Am. Lab.,
(1986) 12.
35
D.J.
Abbott, J. High Res. Chromatogr.,
9 (1986) 598.
36
E.R. Adlard and R.E. Davies, Petroanalysis
81,
Wiley, New York
(1982).
37
W.
Cartellieri and P. Tritthart,
SAE
technical paper
84041 8 (1 984).
This Page Intentionally Left Blank
E.R.
Adlard
(Ed.),
Chromatography
in
the
Petroleum
Industry
Journal
of
Chromatography
Library
Series,
Vol.
56
0
1995
Elsevier Science
B.V.
All
rights
reserved
55
CHAPTER
3
The chromatographic analysis
of
refined
and synthetic waxes
Arthur
Barker
Dussek
Campbell Ltd,
Thames
Road, Crayford, Dartjord,
Kent
DAl4QJ,
UK
3.1
INTRODUCTION
During the petroleum refining process, various residues containing wax are
produced: tank bottoms, lube stock residues and distillate residues. These oily
wax materials, or slackwaxes, can have the oil removed to produce microcrys-
talline, intermediate and paraffin waxes. They are complex materials containing
straight chain and branched chain alkanes, cycloalkanes and very small amounts
of aromatics. The ratio of straight chain
to branched chain alkanes varies accord-
ing to the type of wax. Microcrystalline waxes have the highest proportion
of
branched chain alkanes (see Table
3.1).
Also the carbon number distribution of
waxes varies according to the crude oil used and the degree of refining.
Materials with wax type properties can be synthesized from several routes.
The Fischer-Tropsch process is used to produce a unique series
of
high melting
waxes containing mainly straight chain alkanes. Polyethylene wax materials,
produced from the polymerization of ethylene, have a higher molecular weight
distribution than most petroleum waxes and consist almost wholly
of
even-
numbered straight chain alkanes. Originally polyethylene waxes were obtained
as
a
by-product in the manufacture
of
high density polyethylene. However, over
the past
5
years there has been a rapid growth in specialized processes to pro-
duce polyethylenes of narrow molecular weight distribution. There are also new
processes to manufacture synthetic low molecular weight waxes.
Waxes are usually sold according to average physical properties (such as
melting point, viscosity,
or mean molecular weight), as the actuaI properties
of
each batch will vary depending on the refinery process (typical values are shown
References
pp.
90-93
56
Chapter
3
TABLE
3.1
PERCENTAGE STRAIGHT, BRANCHED AND CYCLO
ALKANES
IN
TYPICAL PETRO-
LEUM WAXES
Type
of
alkane
Type
of
way
Paraffin
wax
Intermediate wax Microcrystalline wax
%
Straight chain alkanes
65-95 -60 20-40
%
Cyclo-alkanes
3-1
5
-25
-3
5
%
Branched chain alkanes
3-20
-15
15-40
Carbon
number
range
1840 20-60
20-80
in
Table
3.2).
This is particularly true
of
petroleum waxes which also vary ac-
cording
to
the crude oil source. Most refined and synthetic waxes are not nor-
mally used in industry as single substances, but are blended together, with pos-
sibly other materials added. This gives the specific physical and chemical prop-
erties
for
the end use. Wax blends are important as adhesives and coatings in
packaging, as additives to
inks
and rubber components and in many other prod-
uct areas
[
13.
Refined and synthetic waxes are complex multicomponent materials
of
great
variety. Their structural properties, i.e. molecular weight distribution
(or
carbon
number distribution) and straighthranched chain nature, have
a
direct bearing
on
physical properties such as melting point, crystallinity, hardness, adhesion,
flexibility and viscosity
[2-51.
Since the early
1980s
there has been an expansion
in
demand
for
more complete analysis
of
these materials (and blends
of
them)
for
both quality control and product development. This has included gas liquid
Tr\BI,E 3
2
TYPICAL
PHYSICAL PROPERTIES
OF
WAXES
AND
POLYWAXES
Type
of
material Typical physical properties
Melting point
Viscosity Penetration
(in "C)
(in CPS)~
(dmm125"C)
Paraffin
wax
4646 2.0-3.4 9-20
Intermediate
wax
6674 3 3-7.9
-20
Microc
60-93
7.9-21 2-3
0
Sasol
c1
-100
8
2
Poly~ax
500 86 38 7
Polywm
1000 113 74
1
Polywax
2000
125
270
0.5
~~ ~~ ~~ ~
"\'iscosity
of petroleum waxes is measured at
99T,
Sasol at
120°C
and Polywaxes
at
149°C.
The chromatographic analysis
of
refined and synthetic waxes
57
chromatography
(GLC)
for carbon number distribution analysis and size exclu-
sion chromatography
(SEC)
for synthetic wax analysis. Wax additive separation
and food contact analysis can be carried out using high performance liquid
chromatography
(HPLC).
There is a growing interest in using supercritical fluid
chromatography
(SFC)
for wax analysis, but this type
of
chromatography has yet
to find a niche.
3.2
GAS
LIQUID
CHROMATOGRAPHY
3.2.1
Establishing present technology
The
GLC
analysis of waxes can separate each of the straight chain alkanes (n-
alkanes) from the branched and cyclo-alkanes (iso-alkanes) of the same molecu-
lar mass (see Fig.
3.1).
Using the best possible capillary
GLC
technique, the al-
kanes can be separated up to approximately
C80.
At this point, the branched and
straight chain molecules tend to merge,
so
that all the alkanes
of
each carbon
number are detected
up
to approximately
C
120
(approx. molecular weight of
1900).
This
precise form of separation of wax materials is
known
as the carbon
number distribution.
Waxes are mainly non-polar and thermally stable; therefore absorption onto
GLC
columns
is
not
a
problem and relatively inert stationary phases can be used.
The complex and extensive carbon number distribution analysis of waxes pro-
duces a high demand on the separation efficiency and thermal stability
of
any
GLC
column used, particularly when it is taken above
350°C.
Because of the
STRAIGHT
CHAIN ALKANES
----
>
A
10
C25
I
C30
f
20
I
c35
30
4ominrtes
Fig.
3.1.
Packed column
GLC
analysis
of
a typical rubber
wax
blend. 1.P method
372/85
(modified),
150-375°C
at
5°C
mid.
References pp.
90-93
58
Chapter
3
wide boiling point range of wax components, temperature programming is used.
For
each wax sample, the repeatability of alkane retention data depends on accu-
rate oven temperature regulation and carrier gas control. In quantitative analysis
the design of both the injector and detector is important, as well as the mode
of
data acquisition and calculation.
The basic techniques for the GC separation of all petroleum waxes were
es-
tablished
in
the 1960s [6,7]. In 1970, the stable stationary phase Dexil 300 was
introduced
[S],
to be used up to 400°C to separate natural waxes [9]. Also in
1970,
the
first short capillary column was used to rapidly separate a microcrys-
talline wax up to carbon number C58
[lo].
In the
1970s,
the basic principles of
capillary
GLC
and bonded stationary phases were established. However, al-
though microprocessor controlled data stations were used
for
data handling, GC
instrumentation was still based on transistors and resistors. The
ASTM
put for-
ward a proposed general method for the packed column GLC analysis of waxes
[
1
11,
but temporarily abandoned it for standardization. It was left
to
the IP Wax
Panel
ST-G-GA
to
carry out round robin testing in 1979 and produce
a
packed
column method
for
carbon number distribution of paraffin wax by gas chroma-
tography
[
121.
The work carried out by the
IP
Wax Panel optimized the conditions for the
quantitative analysis of paraffin waxes up to
350°C
maximum separation tem-
perature. Dual columns containing an inert packing coated with
a
silicone gum
rubber stationary phase were used. However, the method did not address the
need
for
accurate quantitative analysis of microcrystalline waxes
or
wax blends,
which was then becoming important for quality control in the rubber industry.
There are several problems associated with the packed column GLC methods
of
analysis
of
waxes
[
131:
1.
From the production of waxes
in
the refinery to the end-user, different
GLC
methods were used
for
carbon number distribution analysis (with
different column sizes, packing and temperature programs). This led to
different degrees of separation of branched and straight chain alkanes.
2.
The flame ionization detector
(FID)
for a packed column GLC does not
have a linear response factor for alkanes above C32. Unless the instru-
ment is calibrated for response then inaccurate quantitative results are
achieved.
3.
Packed column analysis
of
waxes
is
inaccurate because the straight and
branched chain alkanes cannot be resolved sufficiently to measure the
peak areas with any certainty.
4.
Different methods of data handling, including peak area measurement and
carbon number distribution calculation, are used in different laboratories.
Harmonization between laboratories analysing wax blends and microcrys-
talline waxes is very difficult.
The chromatographic analysis of refined and synthetic waxes
59
A
typical packed column chromatogram of a rubber wax blend is shown in
Fig. 3.1. Although most petroleum GC analysis has now changed to using capil-
lary columns, it is only recently that the wax industry has abandoned packed
column methods for capillary column analysis.
3.2.2
The
1980s
revolution
Many research ideas that were initiated in the late 1970s became commer-
cially practical in the early 1980s. The use of high temperature gas liquid chro-
matography (HTGLC) for the accurate analysis of waxes would not be possible
without the fused silica column containing
a
bonded phase, the modern micro-
processor controlled GC and the rapid growth in sophisticated data handling
techniques.
Glass capillary columns used in the 1970s were expensive to make, very frag-
ile and had an active inner surface that did not readily take stationary phases for
HTGLC .The introduction of the fused silica capillary column [14] with its inert
inner surface (providing a stable base for bonded phases) and outer polyimide
coating (giving the column strength and flexibility), made the detailed separation
of waxes possible. This coincided with the development of stable bonded phases
capable of being used to 350°C without excessive bleeding from the column
[15,16]. Although before 1980 packed column GLC had been used to separate
waxes up to an oven temperature
of
400°C, it was not until 1985 that fused silica
columns were reported to be used in this manner. However, all the initial devel-
opments in capillary HTGLC were made in simulated distillation research for
petroleum refining. This technique is not required to give the resolution neces-
sary for the detailed separation
of
waxes. It was not until the late 1980s that pa-
pers on the high resolution, HTGLC separation of waxes were published.
The GC equipment of the 1970s, was inadequate for the high resolution sepa-
ration
of
waxes. The early 1980s saw a rapid development in GC design. Ovens
were redesigned with thin metal walls to reduce thermal lag and allow accurate,
reproducible temperature programming. Accurate diffusion-free pressure regula-
tors and low flow rate, mass flow controllers became available allowing accurate
control of the carrier gas at high oven temperatures. Microprocessor controlled
equipment enabled more precise and flexible regulation of the column oven and
detector temperatures to be obtained. The commercialization of on-column and
programmed temperature vaporizer
(PTV)
injectors revolutionized sample intro-
duction, minimized sample discrimination and allowing ,accurate quantitative
analysis above C30.
Microprocessors also revolutionized data handling and quantitative analysis.
The demands for more detailed analysis (including further calculations such as
carbon number distribution), data storage and comparison of chromatograms
References pp.
90-93
6
0
Chapter
3
have now led to the common use
of
PC technology and specialized sofhvare for
wax
analysis.
3.2.2.
I
Sample introduction
Rcfore a sample can be introduced into a capillary system it has
to
be dis-
solved
in
a suitable high purity solvent. The high molecular weight alkanes (i.e.
above C40)
in
microcrystalline and synthetic waxes tend to be insoluble
in
most
common solvents at room temperature. When an injector is used in the split
mode then the sample concentration has to be high (at least
I%),
and the solvent
and syringe needle have to be warm to ensure that the higher molecular weight
components do not condense. This means that a rapid, skilful injection tech-
nique, which is difficult to automate, has to be used to produce accurate results.
With on-column injection (including splitless PTV injection), a very dilute
sample of
wax
(1-3
mg ml-l) has to be used
so
as
not to overload the thin-film
columns.
For
alkane standards, the concentration of each alkane
is
reduced to
0.1
mg
ml-I.
This
still presents a problem as to what solvent can be used, particu-
larly
if the method
is
to be automated.
Cyclohexane
[
171
and carbon disulphide
[
181,
both on their own and as
a
1
:
1
mixture, have been used as solvents for polywaxes. Mixtures
of
1
:
1
decane and
carbon disulphide,
or
xylenes and carbon disulphide, or iso-octane and carbon
disulphide
[
191
have been used for crude oil/wax mixtures and polywaxes. Both
carbon disulphide and xylenes are now considered as undesirable solvents
for
normal use
in
a laboratory. Cyclohexane and heptane are practical solvents for
the analysis
of
refined waxes, but require heating to solubilize synthetic waxes.
The methods of sample introduction into capillary GLC systems have been
very ably covered in a book edited by Sandra [20]. Here, it is intended to em-
phasize and expand on the principles that affect HTGLC. The mode of injection
is
of
prime importance
in
the accurate quantitative analysis of complex, high
boiling point alkane mixtures.
A
limited amount
(0.1-1
pl)
of
sample must be
injected rapidly into a small volume
so
that it can transfer to the capillary col-
umn as
a
narrow band and be separated into its components without decomposi-
tion, peak broadening,
or
tailing
[21].
The design of the injector is very impor-
tant
in
achieving an unbiased response for each component over the whole car-
bon number range.
The
two
most common injection techniques used in HTGLC are the on-
column injector invented by Schomburg and Husmann [22] and the Program-
mable Temperature Vaporizer (PTV) introduced by Vogt
et
al.
[23]. There has
been continuous development of both these methods of injection over the past
I0
years.
On-column injectors are now designed
so
that a dilute solution
of
the sample
is injected through a septumless valve directly into a capillary column. The in-
The chromatographic analysis
of
rejined and synthetic
waxes
61
jection point is externally cooled below the oven temperature to focus the sample
as a narrow band at the entrance
of
the column. This secondary cooling is
switched off before initiating temperature programming. Normally, a retention
gap
(70
cm to
5
m of uncoated, deactivated fused silica capillary) is placed be-
tween the injector and the separation column. This focuses the less volatile com-
ponents and minimizes band broadening of alkane peaks eluting during tempera-
ture programming
[24].
If this mode of injection is used without a retention gap
to separate waxes, then there
is
a
risk of peak broadening and splitting.
This injection technique is very useful in the analysis of refined and synthetic
waxes in that it requires dilute sample solutions (minimizing the solubility
problems). Also the technique has proved to give a reasonably linear response
for alkanes up to
C78 [25]
and has been used to separate polywaxes up to
C120
chain length. It can be used to quantitatively analyse waxes (with components of
wide boiling range), giving accurate and precise carbon number distributions.
However, there are several problems inherent in using this technique:
1.
2.
3.
-
Due to limitations in minimum needle size available, the smallest column
that the sample can be injected into has to be at least
0.32
mm internal di-
ameter. If a retention gap is used, then the separating column can be of
smaller diameter. However, for high temperature work, the columns have
to be joined by a "Graphpack" type connector (which can be a source of
problems).
The thin silica syringe needles used are difficult to handle and keep clean.
It is possible that a small amount of sample will move back along the
sy-
ringe by capillarity, or be transferred from the outer syringe wall and
smeared along the inlet wall of the column. To guard against this, it is
recommended that there should be a minute air and solvent gap between
the sample and the syringe plunger, as well as an air gap at the syringe tip.
This method
of
sample introduction is most accurate when a stainless
steel needle is used in the automated mode. Precision engineering and mi-
croprocessor control have now made automatic on-column injection a re-
liable technique.
Residues can build up in the injection zone leading to contamination of
later samples, particularly if the column is used at a higher temperature
than usual. This part of the system needs regularly removing, which is
easier if a retention gap
is
used.
The programmable temperature vaporizer (PTV) injector was developed by
Poy
[26]
and Schomburg
[27]
into a practical method that could be used in the
split, splitless (on-column) or solvent flush mode
of
injection. In a modern PTV
injector, a sample solution is injected, via a septum, into a narrow glass tube
containing glass- or quartz-wool, connected to the capillary column. The sample
condenses onto the wool insert, which is initially kept at ambient temperature.
References
pp.
90-93