Adlard E.R. (ed.) Chromatography in the Petroleum Industry
Подождите немного. Документ загружается.
192
Chapter
7
r-'--7-
I
---
T-----7-----7----.
.
.
7-
0
10
20
30
40
Tfme
<min.>
Fig.
7.16.
C,
H
and
S
chromatograms
for
SimDist
analyses
of
a
long
residue.
gramming. When this constant-flow mode
is
used without any further considera-
tions. the electronics will programme the pressure assuming that the column
head pressure equals the column pressure drop. This will result
in
a
non-linear
flow
programme during analysis. The solution here is to calculate the necessary
pressure programme to provide constant flow and to introduce the results
of
these calculations via the keyboard into the GC electronics.
Generally, samples on which a SimDist analysis is performed have
a
wide
boiling range and on-column injection is the only mode of injection to be used
here. As mentioned previously, solvent vent should be used to prevent the huge
amount of solvent entering the AED. Depending on the column starting tempera-
ture, co-eluting, low-boiling parts of the sample will be vented as well. With
a
typical wide-bore column
of
6
m length and a starting temperature
of
3OoC,
the
initial boiling point
of
the sample must not be below
150°C.
Programming the
column up
to
430°C
will enable compounds with normal boiling points up to
700°C
to be eluted. Repeated experiments demonstrated that the temperature
of
the transfer line and cavity should not
be
more than
30°C
below the maximum
oven temperature, otherwise high-boiling compounds will not be detected com-
pletely.
7.7.
1.
1
Multi-element SimDist
software
Hewlett Packard provides
a
software package specifically written to perform
SimDist analyses with the AED. It consists of
a
set
of
macros that execute the
Microwave
plasma
detectors
193
II
I
w
w
Fig.
7.17.
Boiling point calibration with the
AED
using normal alkanes.
various program files used. These include boiling point calibrations, calibrations
for the correction of the carbon response on other element lines (back-amounts),
response calibrations as well as blank run determinations. The package enables
the user to set up their
own
element set, to choose which elements should be de-
termined. Reports can be produced that give the mass of an element recovered,
as well as the hydrogen-to-carbon ratio, as a function of the boiling range, pre-
sented in a table as well as graphically,
in
either degrees Celsius or degrees
Fahrenheit. It automatically determines and controls all analytical runs that have
to be performed. These include all calibration runs as well as runs necessary to
detect the specified elements.
Calibrations to determine the back-amounts and responses of specific ele-
ments have to be performed by injection of synthetic mixtures including one or
more compounds containing these elements. In a future release of the software,
the possibility of calibration with real oil samples will be incorporated.
As
in
general most hetero compounds
are
distributed throughout the boiling range, this
kind of calibration can be expected to be more accurate than single-compound
calibration.
Some publications suggest that the
AED
exhibits a compound or structure de-
pendency of response
[30].
We have determined the response of the
AED
to-
wards carbon and hydrogen for a set of different structured organic compounds.
The results are given in Table
7.8.
Clearly, dependency on structure is not de-
References
p.
200
194
Chapter
7
'TABLE
7.8
STRUCTURE INDEPENDENCY
ON
RESPONSE
FOR
THE HP
AED
Compound Formula Percentage
of
hydrogen (C
+
H
=
100%)
Theor. DCPDSa
r1-Cl2~
Naphthalenea
2-Methyl-hexadecane
2-Methyl-
1
-dodecene
1
-Methyl-decalin
I,
1
-Dimethyl-tetralin
Tetramethy I-benzene
Fluoranthene
1,4-Dithiane
Quinoline
Dimethyl-sulfolane
15.0
14.3
13.2
10.1
10.4
4.9
14.4
6.1
14.4
15.1
14.2
13.3
10.2
10.4
4.9
14.4
6.0
14.4
15.2
14.2
13.3
10.1
10.4
4.9
14.5
5.9
14.4
~~~~~ ~~
~
"DCPDS, calibrated with dicyclopentyl dodecyl sulphide; n-C
12,
calibrated with n-dodecane;
naphthalene, calibrated with naphthalene.
monstrable. These results have
all
been obtained using a strictly constant column
flow
(by means
of
a proper pressure programme; see also Section
7.7.1)
and
taking the non-linearity
of
the hydrogen response into account. More about this
hydrogen non-linearity is described in the next paragraph.
1.7.1.2
Linear dynamic ranges
Quimby and Sullivan [23] present an extensive description
of
the
HP
AED
design, including a table with selectivities, sensitivities
and
linear dynamic
ranges
(LDR).
In order to be able to perform Multi-Element SimDist of the ele-
ments H,
C,
S
and
N,
we determined the ability of the
AED
to determine these
elements. Although the minimal detectable level for nitrogen given by Quimby
and Sullivan is
7.0
pg/s, this low level
is
hard
to
reach in high-temperature
GC
use. In normal practice
of
high-temperature
GC,
it is almost impossible to ex-
clude minor leaks in injection port (septa) and column couplings,
so
that a con-
stant relatively high background of nitrogen
is
always present. Since nitrogen-
containing compounds in oil samples generally are distributed along the boiling
range, nitrogen is present in thousands
of
compounds. The maximum concentra-
tion of nitrogen rarely exceeds a few thousand parts per million. Thus, the car-
bon-to-nitrogen ratio
of
eluting material is in general over
10
000,
and equals or
even exceeds the selectivity of
6
X
lo3
given in
1231.
For these reasons, the
AED
is
not the detector
of
choice for SimDist of the element nitrogen in high-boiling
oil samples.
Hydrogen exhibits a deviant behaviour in terms of linearity, as is depicted in
Fig.
7.18.
We have observed this phenomenon
on
different plasma detectors (see
Microwave plasma detectors
195
1.5
1.4
1.3
1.2
1.1
P
$
l.o
5;
$
0.7
0.6
0.5
0.4
u
h
1 3
2
1
5
10
5
10
5
10
5
10
5
10
5
Fig.
7.18.
Non-linearity
of
hydrogen.
also Section 7.4.3), but have not yet been able to produce a conclusive explana-
tion. The shape of the “linearity curve” is indeed highly reproducible and can
therefore be determined exactly. The Multi-Element SimDist software includes a
method
to measure this non-linearity and calculate response factors accordingly.
It should be realized, however, that these response factors vary with the abun-
dance of the hydrogen line intensity, and therefore vary along the hydrogen peak.
In normal peak integration mode, as was the case with the measurements of Fig.
7.18, these factors cannot be incorporated and thus hydrogen will not be meas-
ured correctly. Implementation of these factors is automatically performed by the
SimDist software, since it does not integrate separate peaks, but rather integrates
and treats the chromatogram in narrow area slices. It generates a LDR for
H486
of about
10
000.
Together with the LDR of
C496
and
S181
of
over
10
000,
it pro-
vides an excellent detector for SimDist analysis for these elements.
7.7.1.3
Simdist results
In “normal” SimDist analyses with the flame ionization detector, where only
carbon is measured, no response factors are used. For samples containing a high
concentration of polynuclear aromatics, too high a recovery will therefore be
reported, whereas very highly paraffinic samples will exhibit too low a recovery.
Apart from these errors, hetero elements (such as sulphur) present in the sample
References
p.
200
196
Chapter
7
TABLE
7.9
COMPARISON
OF
ELEMENT
RATIOS
WITH THE HP AED
YOC
%H
%S
Comb.
AED
Comb.
AED
Comb.
AED
Sample
A
12.1
12.8/12.9
85.5 84.8B4.8 2.40 2.36/2.35
Sample
B
12.0
13.0112.8
85.0 83.9184.2 3.03
3.1
U2.98
Sample
C
11.5
12.6/12.4 87.7 86.3B6.5 1.03
1.06/1.09
SampleD 13.8
14.3113.8 86.2 85.6/86.
I
<0.
1
0.06/0.04
SampleE 13.9
1
3.9/
1
3.6
86.1
86.1/86.4
<0.01
0.015/0.015
will not be detected at all. When using the
AED,
none of these errors will occur.
These
two
types of SimDist results should therefore be compared with this con-
sideration in mind. Since the carbodhydrogen ratio determined with the
AED
is
a
good yardstick for the estimation of the amount of paraffinic and (po1y)-
aromatic material in a sample, we were able to select a number of samples for
this comparison. The conclusion is that differences
in
boiling points reported for
1.64-
1.60-
m
W
1.55-
9
.
1.45-
1.4@
_,
-
---*
Y
!i
8
t
In
6
CIS
I1
Nrn
00
Ii
1
I
I
0.00
5.00
10.00
15.00
Retention time
In
minutes
Fig.
7.19. Chlorinated components
in
the
EO process.
Microwave plasma detectors
197
carbon and hydrogen, will in general not be larger than
3°C.
The AED results of
the totals of the different element distributions and the comparison with the re-
sults obtained by combustion analyses of total samples are presented in Table
7.9.
7.7.2
Determination
of
chlorine-containing compounds in the ethylene oxide
process
In the ethylene oxide (EO) process, ethylene is converted to ethylene oxide
over a catalyst. Traces
(1
ppm)
of
chlorinated components are added to tune this
catalyst. Figure
7.19
is
a
typical analysis
of
the reactor outlet using the C1 chan-
nel of the home-built apparatus.
7.7.3
Determination
of
oxygenates in gasoline
Modem non-leaded gasolines contain oxygenates as octane boosters. Figure
7.20
represents the
0
and
C
chromatograms
of
such a gasoline using the
Hp
atomic emission detector. See also the work of Wylie
et
al.
[31].
7.7.4
Determination
of
noble gases in natural gas
In oil-producing reservoirs, tracers
are
sometimes introduced via the steam
injector and samples are taken
from
the producing wells in order to identify
preferential gas or steam paths. The noble gases are preferred because they are
chemically inert and non-radioactive. Figure
7.2
1
represents a neon chroma-
togram of a synthetic mixture using the home-built apparatus.
7.7.5
Determination
of
metal porphyrins in crude oils
For the detection of nickel, vanadium and iron porphyrins in crude oils, the
reader
is
referred to the work
of
Quimby
et
al.
[32] and Zeng
ei
al.
[33].
7.8
CONCLUSION
At the time of writing, the only commercially available plasma detector is the
Hewlett Packard 5921A Atomic Emission Detector. This detector has been de-
References
p.
200
198
1400-
Chapter
7
5
000
4
000
3Q0Q
Z@OQ
1
El00
..*
2
4
Time
(min.)
jl.;
6
?
R
o
777
of
DRTR:CMDPSTRC.D
Q
2
.1
6
Time
(rnin.)
8
Fig.
7.20.
Carbon
and oxygen
chromatogram
of
gasoline.
signed such that it
is
compatible with capillary columns up to temperatures
of
400°C.
It is a very good compromise between performance, ease of operation
and price. The choice
of
a moveable diode array as spectrometric detection de-
vice
enables the determination
of
a few elements simultaneously, if the emission
lines are within the range
of the array. For elements with emission lines more
apart, the automatic
GC
sampler enables multiple analyses in an accurate man-
ner.
Microwave plasma detectors
199
5.40-
5.36-
5.31-
5.27-
5.23-
.01
Fig.
7.21.
Natural
gas,
90
ppb
neon added.
To
decrease erosion of the discharge tube, which contains the helium plasma,
the system is water cooled. The construction of the cooling of the cavity makes it
very sensitive towards water leakages at temperatures above 300°C. Apart from
this source of trouble, the detector in combination with the Chemstation software
is a powerful and easy-to-use device. It provides the user with menu’s in which
the gas chromatograph as well as the detector can be controlled automatically,
calibrated and used for analysis.
The detector enables measurement of both the elemental composition of
eluting compounds as well as element distributions of complex samples.
Ex-
periments pointed out that the response of the detector for the various elements
is structure independent.
In our laboratory, the detector has been used successfully for the determina-
tion of
a
wide variety of elements: viz. metals, such as lead, tin and mercury, the
noble gases as neon, krypton, argon and xenon; oxygen-containing compounds,
where it successfully replaces the
OFID;
halogen-containing compounds, where
it replaces the
EC
detector; sulphur-containing compounds, replacing the
FPD;
and last but not least, hydrogen, for which no other selective detector is avail-
able.
In
most of these cases, the
MDL
of the
AED
is the same or even better than
that of the other selective detectors. Furthermore, the
AED
in general is imme-
diately ready for use for detection of a new element, whereas installing a new
detector needs quite an investment in stabilization time and calibration.
Although the detector has only been available since
1990,
a number of inter-
esting applications have been published already and it is to be expected that
many more will follow.
200
Chapter
7
7.9
REFERENCES
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
H.P. Broida and G.H. Morgan, Anal. Chem. 24 (1952) 799.
A.J. Mc Cormack, S.C. Tong and W.D. Cooke, Anal. Chem. 37 (1 965) 1470.
C.A. Bache and D.J. Lisk, Anal. Chem. 37 (1965) 1477.
C.A. Bache and D.J. Lisk, Anal. Chem. 38 (1966) 783.
C.A. Bache and D.J. Lisk, Anal. Chem. 38 (1966) 1757.
C.A. Bache and D.J. Lisk, Anal. Chem. 39 (1967) 786.
H.A. Moye, Anal. Chem. 39 (1967) 1441.
R.S. Braman and
A.
Dynako, Anal. Chem, 40 (1968) 95-106.
R.M. Dagnall, S.J. Pratt, T.S. West and D.R. Deans, Talanta 16 (1969) 797.
W. Braun, N.C. Peterson, A.M. Bass and M.J. Kurylo, J. Chromatogr. 55 (1971) 237.
W.R. McLean, D.L. Stanton and G.E. Penketh, Analyst 98 (1973) 432.
C.I.M. Beenakker, Spectrochim. Acta 31B (1976) 483.
C.I.M. Beenakker, Spectrochim. Acta 32B (1977) 173.
B.D. Quimby, P.C. Uden, R.M. Barnes, Anal. Chem. 50 (1978) 21 12.
R.J. Lloyd, R.M. Barnes, P.C. Uden and W.G. Elliot, Anal. Chem. 50 (1978) 2025.
M.
Moisan,
R.
Pantel, J. Microwave Power 14 (1975) 57.
M.H.
Abdallah and
S.
Coulombe, Spectrochim. Acta. 37B (1982) 583.
S.A. Estes, P.C. Uden and R.M. Barnes, Anal. Chem. 53 (1981) 1829.
Estes, P.C. Uden and R.M.
Barnes,
J. Chromatogr. 239 (1982)
181.
S.R.
Koirtyohann, E.D.
Glass,
D.A. Yates, E.J. Hindenberger
and
F.E.
Lichte, Anal. Chem. 49
(1977) 1121.
J.T.
McCaffrey and R.G. Michel, Anal. Chem. 55 (1983) 488.
A. de Wit, Spectrochim. Acta 38B (1983) 369.
B.D. Quimby and J.J. Sullivan, Anal. Chem. 62 (1990) 1027.
J.J. Sullivan and B.D. Quimby, Anal. Chem. 62 (1990) 1034.
K. Tanabe, Spectrochim. Acta 36B (1981) 633.
P.
Brassem, F.J.M.J. Maessen and
L.
de Galan, Spectrochim. Acta31B (1976) 537.
Proceedings
of
the Tenth International Symposium on Capillary Chromatography
ume
11,
1025 pp.
Advanced Product Release, Carlo Erba Instruments, Rodano, Italy, May 1990.
J.A. Stewart and A. Scheeline, Anal. Chem.
56
(1984) 2995.
989,
‘ol-
,,
30
S.
Pedersen-Bjergaard,
T.N.
Asp and T. Greibrokk, Anal. Chem. 15 (1992)
89.
31 B.D. Quimby, P.C. Dryden, and J.J. Sullivan, High Resolut. Chromatogr. Chromatogr. Com-
mun. 14 (1991)
110.
32
P.L. Wylie and B.D. Quimby, High Resolut. Chromatogr. Chromatogr. Commun. 12 (1989)
813.
33
Y.
Zeng, J.A. Seeley, T.M. Dowling, P.C. Uden and M.Y. Khuhawar, High
Resolut.
Chroma-
togr. Chromatogr. Commun. 15 (1992) 669.
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
20
1
CHAPTER
8
The sulfur chemiluminescence detector
Richard
S.
Hutte
Sievers Instruments
Inc.,
2500
Central Avenue, Suite
HI,
Boulder,
CO 80301,
USA
8.1
INTRODUCTION
All petroleum samples from crude oil to refined products contain varying
concentrations of sulfur-containing compounds. These compounds include gases
(H2S, COS,
SO2
and CS2), aliphatic, and aromatic thiols, sulfides and polysul-
fides, thiophenic and other sulfur-heterocyclic compounds. The geochemistry of
sulfur has been recently reviewed
[
13
and an extensive list of the types and con-
centrations of sulfur compounds found in petroleum systems is presented. The
concentration of sulfur compounds in crude oils varies widely
from
<0.05%
to
>14%
sulfur by weight
[l].
Most of the sulfur is in the form
of
organic com-
pounds, with only relatively small amount of dissolved H2S and elemental sulfur
present. Most of the organo-sulfur compounds are higher molecular weight com-
pounds with boiling points greater than
300°C
[l].
Hydrotreating and other proc-
esses to remove sulfur compounds from crude oil and other feedstocks result in
the conversion of the higher molecular weight materials to
H2S
and lower mo-
lecular weight thiols and sulfides. Thus, a wide range of different types of sulfur
compounds is usually present in petroleum products.
Sulfur compounds are also intentionally added to petroleum products. Low
molecular weight thiols, sulfides and tetrahydrothiophene are used as odorants
for natural gas and LPG and certain additives for lubricants and other products
contain sulfur.
The presence
of
even parts per billion levels
of
sulfur can cause numerous
problems in the processing of petroleum including poisoning of the expensive
catalysts used in the refining process, corrosion of reactors and pipelines and the
presence of sulfur-containing compounds imparts undesirable odors to petroleum
products. Combustion of petroleum products containing sulfur compounds is a
References
pp.
227-229