Determination of the structures and

abundances of alkanes and olefins in

Fischer-Tropsch products using 13C and ‘H
n.m.r. methods

David J. Cookson and Brian E. Smith

The Broken Hill Proprietary Co. Ltd, Melbourne Research Laboratories,
245 Wellington Road, Mulgrave, Victoria, 37 70, Australia
(Received 26 September 7988; revised 28 November 1988)

13C and ‘H n.m.r. methods are shown to provide detailed and quantitative information regarding the
structures and abundances of major and minor hydrocarbon components in products from Fischer-Tropsch
synthesis. A test jet fuel product (boiling range 19&230”(Z) has been studied together with a test diesel
fuel product (23&32O”C). The jet fuel consists primarily of n-alkanes (87 wt%), while branched alkanes
with, on average, 12 carbon atoms per molecule and one CH, branch per molecule, represent another
9wt% of the sample. All possible monomethyl branched alkane isomers are shown to occur with
approximately equal probability. Olefins represent nearly 4 wt% of the sample. 13C n.m.r. distinguishes
l-ene (0.6 wt%), cis-2-ene (0.8 wt%), trans-2-ene (1.4 wt%), 2-methyl-2-ene (0.4 wt%) and trans-3-ene
(0.5 wt%) structures. ‘H n.m.r. distinguishes structures in the form - CH = CH, (0.7 wt%), - CH = CH -
(2.5 wt%), >C=CH, (0.1 wt%) and >C=CH- (0.5 wt%). Grossly similar results have been found for
the diesel fuel, in which the average number of carbon atoms per molecule is approximately 15.

(Keywords: carbon-13 n.m.r.; n.m.r.; structural analysis)

The Fischer-Tropsch process ‘-’ involves reaction of EXPERIMENTAL

synthesis gas (CO/H2 mixtures) over a catalyst that
Sample origins
usually incorporates iron, cobalt, nickel or ruthenium.
To date, large scale Fischer-Tropsch plants operate only Jet and diesel fuels were distilled from a Fischer-
in South Africa, where coal is used for synthesis gas Tropsch syncrude produced in a fixed-bed reactor using
production. In Australia there are large reserves of a Co/ZrO,/SiO, catalyst. The feed gas consisted of a 2:l
uncommitted natural gas. One way in which these mixture of H,:CO pumped at a gas hourly space velocity
reserves could be utilized involves conversion to synthesis of 200 h- ’ at nearly atmospheric pressure and approxi-
gas, followed by production of hydrocarbon transport mately 210°C. Samples studied include the whole fuels
fuels (gasoline, jet and diesel fuels) via a Fischer-Tropsch as well as residual fractions obtained after urea
route. adduction7 and removal of the n-alkane rich adduct.
The chemical and physical characteristics of Fischer-
Tropsch products are greatly influenced by process Instrumental procedures
conditions, catalyst formulation and feed gas com- G.c. analyses were performed by the BHP Petroleum
position. Typically, hydrocarbon products are domi- Laboratory on a Hewlett-Packard model 5880 chromato-
nated by straight methylene chain structures with carbon graph with flame ionisation detection. Full details of the
numbers described by the Schulz-Flory distribution’ 4. procedure have been described elsewhere5g6. All n.m.r.
Hydrocarbons incorporating polymethylene structures spectra were recorded at 30°C on a modified Bruker
include normal alkanes, branched alkanes and alkenes. WP200 spectrometer operating at 50.3 MHz for 13C and
In the jet fuel and automotive diesel fuel boiling range 200.13 MHz for ‘H. Chemical shifts are reported relative
(190-320°C in the present work), n-alkanes are easily to internal TMS. For ‘H n.m.r. measurements, a 90”
identified by g.c. The purpose of this work is to investigate pulse flip angle (9.5 ps) and a 20 s relaxation delay period
the utility of n.m.r. techniques, previously applied to were employed’. Solution concentrations of < 5 wt% in
petroleum fuels and synfuels5,6, for the identification of deuterochloroform were used.
the structures present in two test Fischer-Tropsch fuels. Conventional inverse-gated 13C n.m.r. experiments
One of these boils in the jet fuel range (19&23o”C) and were carried out using previously established principles’.
the other boils in the automotive diesel fuel range Samples of the whole fuels were 4&50wt% in
(23G32o”C). Components other than the n-alkanes are deuterochloroform, while the minor saturates fractions
of prime concern in this study. Quantitative analysis is were approximately 10 wt% in deuterochloroform.
sought. All solutions contained 0.7 wt% of the relaxation
00162361/89/06077W6$3.00
9’. 1989 Butterworth & Co. (Publishers) Ltd

776 FUEL, 1989, Vol 68, June

 Examination of Fischer- Tropsch products using n.m.r.: 0. J. Cookson and B. E. Smith

in reasonable accord, given that the per cent of n-alkyl

carbon should marginally exceed the wt% n-alkanes13
and given the integration limitations in both n.m.r. and
g.c.. Allowing that IX,B, y and 6 n-alkyl resonances shown
in Figure 1b each correspond to two carbon atoms per
molecule, the average n-alkane chain length is calculated
as N,= 15.0. This value is also in close accord with the
average chain length calculated from g.c. (14.9).

A
x 1200

4 The major ‘H n.m.r. intensity for the diesel fuel occurs

at 1.27 ppm (CH, groups) and 0.88 ppm (CH, groups).
For a sample consisting solely of n-alkanes, of course,
the average chainlength can be readily calculated from
CH, and CH2 resonance intensities. For this sample,
however, such a calculation yields an erroneously low
value of N,= 13.4, because branched alkane CH, and
CH, group resonances are coincident with the n-alkane
resonances. The major n.m.r. spectral features of the
Fischer-Tropsch jet fuel are very similar to those shown
in Figure 1 for the diesel fuel. The jet fuel is composed
predominantly of n-alkanes (C, = 90%, n-alkane content
by g.c.=84 wt%) with an average chainlength of 11.9
(11.6 by g.c., a low estimate of 10.3 by ‘H n.m.r.). Olefinic
sp’ carbon is also of very low abundance (0.6%).

Diesel fuel branched alkanes

In addition to the intense resonances due to n-alkanes,
several low intensity resonances, also mainly associated
with saturated species, are observed in the i3C n.m.r.
, I
spectrum of the Fischer-Tropsch diesel fuel (FigureZa).
IL0 120 loo 80 60 CO 20 wm
Since the major n-alkane peaks mask intensity due to
Figure 1 a, ‘H n.m.r. spectra; b, ‘jC n.m.r. spectra of the the minor saturates in Figure2a, urea adduction7 has
Fischer-Tropsch diesel fuel. Assignment of ‘%Z resonance intensity to been used to extract n-alkanes, leaving a sample enriched
carbon groups in n-alkanes is shown in the minor saturates. This sample contains just 3 wt%
residual n-alkanes, as shown in Figure3, and its 13C
n.m.r. spectrum is illustrated in Figure2b.
agent tris(acetylacetonato)chromium(III), Cr(acac),. A GASPE selected multiplet n.m.r. spectra for the minor
90” pulse flip angle (16 ps) and a relaxation delay period saturates fraction of the diesel fuel are shown in Figure 4.
of 10 s were used. Gated spin-echo (GASPE) experiments Individual CH and CH, group spectra have been
were carried out using the sequence described previously’“. calculated but are not shown. In Figure4a. all intensity
Solutions contained 0.4 wt% Cr(acac), and a relaxation to low field of 30ppm is due to CH groups, while all
delay of 20s was employed. The spectral combinations intensity to high field of 30ppm is due to CH, groups.
used to derive selected multiplet spectra were similar to Non-protonated carbon groups were not detected in this
those described in Ref. 11 rather than those described in sample. Carbon group distributions (expressed as
Ref.10. Artefacts remaining in subspectra were removed fractions) are given in Table I. Calculations using these
using procedures described elsewhere’ 2. CH, group distributions (Table 1) show that for the minor
saturates fraction, the average number of branches per
RESULTS AND DISCUSSION molecule (NB) is 1.1 and the average number of rings per
molecule is near zero (N, = 0.06). This fraction therefore
N.m.r. spectra of whole fuels-identification of compound consists primarily of singly-branched alkanes.
classes A more detailed assignment of the i3C n.m.r. spectrum
The ‘H and 13C n.m.r. spectra of the Fischer-Tropsch in Figure2b can be gained by considering an average
diesel fuel are shown in Figurel. In the ‘H n.m.r. branched alkane molecule which consists of approxi-
spectrum, resonances of very low intensity due to olefinic mately 15 carbon atoms (by analogy with N, determined
proton groups are detected between 4.5 and 6.0 ppm. The for the n-alkanes and in keeping with Figure3) and
corresponding olelinic carbon group resonances are includes just one methyl branch per molecule (as
observed between 114 and 140ppm in the 13C n.m.r. determined by selected multiplet n.m.r.). For Cl5 there
spectrum (Figurelb). The percentage of 13C n.m.r. are just six such possible isomers. In Figure2 we have
intensity due to olefinic carbon is measured as just 0.2%. used the notation 2M through 7M to distinguish these
The five most intense resonances in the 13C n.m.r. isomers in terms of the branch position. Consideration of
spectrum (Figurelb), at 32.3, 30.1, 29.8, 23.0 and literature chemical shift datai4, expected relative
14.2 ppm, arise mainly from n-alkanes which are clearly resonance intensities for a given isomer, and use of the
the major component of the diesel fuel. Integration of CH, type identified via GASPE subspectra, yields the
these five peaks yields the percentage of 13C n.m.r. assignments shown in Figure2b in terms of isomer (2M
intensity due to n-alkyl carbon, namely C, = 90%. G.c. to 7M) and carbon number (1 to 15) in each isomer. A
suggests 83 wt% of n-alkanes. These two figures are number of resolved peaks can be uniquely associated

FUEL, 1989, Vol 68, June 777

Examination of Fischer- Tropsch products using n.m.r.: 0. J. Cookson and B. E. Smith

I
co 3.5 30 25 20 15 10 wm
fraction, namely 0.3 (2M): 0.7 (3M): 1.0 (4M): 2.8
(5M +6M + 7M). This is most obvious in the 39 ppm
region of Figures2a and 2b. As might be expected, urea
adduction has selectively removed branched alkane
isomers with long straight chain segments (i.e. 2M and
3M) along with the n-alkanes. Thus for the whole diesel
fuel, unlike the enriched fraction, there is a near equal
abundance of all possible monomethyl branched isomers.
So far, we have only been concerned with relative
abundances of branched alkanes. For these samples there
is little doubt that within a reasonable margin (e.g. a
10% margin would be fully adequate), branched alkane
isomers have the same molecular weight as other species
present and all species have grossly similar stoichi-
ometries. In this circumstance, the per cent 13C n.m.r.

1
b15

2M A------
i
‘
3M
I
b16

I
LM
1‘
5M
1‘

1‘
7M
110 15 0 19 0 230
1
co 35 30 25 20 15 10 __ ELUTION T/ME immsl -
wm

Figure 3 Gas chromatogram of the branched alkane fraction of the

Fischer-Tropsch diesel fuel. Peaks due to branched alkanes (b) and
normal alkanes (n) are shown together with the corresponding carbon
number

” L1, CH+CH3

Figure 2 r3C n.m.r. spectra of: a, the Fischer-Tropsch diesel fuel; and
b, the branched alkane fraction of the diesel fuel. Assignments are given
in terms of six possible isomeric structures (2M to 7M) wiith specific
carbon groups designated 1 to 15
b

with individual isomers. For example, 39.2, 28.6 and

28.3 ppm with isomer 2M; 34.5, 19.3 and 11.4ppm with
3M; 39.5,26.6 and 20.2ppm with 4M; and 23.1 ppm with
5M. Peaks at 31.2 ppm and 32.8 ppm are associated with
a superposition of resonances from 5M, 6M and 7M.
From the intensities of these resolved peaks, the relative C

abundances of isomers occurring in the minor saturates

fraction can readily be calculated as 0.3 (2M): 0.7 (3M):
1.0 (4M): 1.0 (5M): 1.8 (6M +7M).
The assignments for the minor saturates fraction in
Figure2b can be directly translated to the whole diesel
fuel spectrum in Figure2a. Inspection of resonance LO 35 30 25 20 15 10 pm

intensities shows that the relative abundances of Figure 4 GASPE selected multiplet spectra for the branched alkane
monomethyl branched alkane isomers in the whole fuel, fraction of Fischer-Tropsch diesel fuel: a, CH +CH, groups; b, CH,
namely 1.O(2M): 1.O(3M): 1.O(4M): 2.8 (5M + 6M + 7M), groups; and c, CH, (n = 1-3) groups. Quaternary C groups are of
differs from that observed for the minor saturates negligible abundance

778 FUEL, 1989, Vol 68, June

 Examination of Fischer-Tropsch products using n.m.r.: D. J. Cookson and B. E. Smith

Table 1 Carbon group abundances and average structure parameters 2-methyl and 3-methyl isomers in this fraction.
for the minor saturates fractions of the Fischer-Tropsch diesel and
Resonance assignments follow directly from those
kerosene fuels
discussed for the diesel fuel. In the whole jet fuel,
Parameter” Diesel fuel Jet fuelb resonance intensities imply isomer abundances in the
ratio 1.0 (2M): 1.0 (3M): 1.0 (4M): 1.0 (5M): 0.5 (6M).
f (CH,) 0.202 0.216 (0.227) For an average Cl2 structure, this is consistent with
f KHz) 0.722 0.713 (0.686)
monomethyl branching at all possible sites along a Cl 1
fW) 0.076 0.071 (0.087)
1.1 1.w backbone with near equal probability. A procedure
N,
N, 0.06 0.16“ analogous to that used for the diesel fuel, yields the
absolute wt% abundances of branched alkane isomers
of 13C n.m.r. intensity attributable to CH,
“flCH.) = fraction which are listed in Table2. The whole jet fuel contains
(n=(r3) groups; N,=average number of branches per molecule
(= N,[2f(C)+flCH)], where Nc is assumed to be 11.9 for jet fuel
9 wt% branched alkanes.
branched alkanes and 15.0 for diesel branched alkanes); N, = average
number of rings per molecule (0.5 Nc[2flC)+flCH)-f(CH,)] + 1). See OleJins
Ref. 5 for full details
13C n.m.r. Low intensity resonances attributable to
*Values in brackets adjusted for n-alkane content of branched alkane
fraction, namely 18.1 wt% olelinic groups are observed in the ‘H and 13C n.m.r.
e Parameters calculated using adjusted figures for flCH.) spectra of the Fischer-Tropsch jet and diesel fuels. Since
d N, value is sensitive to variations in ACH) and ACH,). As with the the olelins are more abundant in the jet fuel, we will focus
diesel fuel, cyclic molecules are not considered to be present to any on this fuel first. Olehnic carbon resonance intensity
significant extent in the jet fuel
accounts for 0.6% of the total i3C n.m.r. intensity.
Consideration of relative intensity and chemical shift
data15 enables assignment of resonances in the olefmic
Table 2 Absolute wt% abundances ofdiesel and jet fuel components” carbon region to five principal structures (Figure 5),
which differ in terms of the position of the olefin bond
Wt% in whole Wt% in whole
relative to the carbon chain terminus, and in terms of
diesel fuel jet fuel
branching at olelinic carbon sites. The structures and
Branched alkanes resonance positions are as follows: I-ene (139.1 and
2-methyl 2.1 2.0 114.2 ppm); trans-2-ene (13 1.8 and 124.6 ppm); cis-2-ene
3-methyl 2.1 2.0 (131.0 and 123.6ppm); 2-methyl-2-ene (131.0 and
4-methyl 2.1 2.0
S-methyl 2.1 2.0
125.2ppm); and trans-3-ene (131.9 and 129.5ppm).
> 6-methyl 3.8 1.0 Taking the relative abundance of the major isomeric form
as unity, the relative abundances of isomers is, 0.42
Oletins (“C n.m.r. classification)
I-ene nc 0.6
(1-ene): 0.58 (cis-2-ene): 1.O(trans-2-ene): 0.26 (2-methyl-
cis-2-ene nc 0.8 2-ene): 0.32 (trans-3-ene).
trans-2-ene nc 1.4
2-methyl-2-ene nc 0.36
trans-3-ene nc 0.45

Oletins (‘H n.m.r. classification)

-CH=CH2 nc 0.7
-CH=CH- nc 2.5
>C=CH, nc 0.1
>C=CH- nc 0.5

a Based on C 15 for diesel fuel and Cl 2 for jet fuel; nc = not calculated,
total oletins 2 1.5 wt% in diesel fuel

intensity due to all carbons in a given isomer is a good

approximation of the absolute wt% abundance of the
isomer. For the whole diesel fuel, a substantial proportion
(at least 60%) of branched alkane intensity is masked by
the dominant n-alkane peaks, and thus cannot be
integrated directly. Nevertheless, as a result of the
spectral assignments indicated in Figure 2, it is sufficient
to integrate resolved isomer intensity and multiply by li0 7j5 130 135 120 li5 ppm
the appropriate factor to calculate total intensity. Results
are listed in Table 2. The whole diesel fuel sample contains P

x120
12 wt% of branched alkanes. i

Jet fuel branched alkanes

The Fischer-Tropsch jet fuel saturates are close
analogues of the diesel saturates, except for the shorter
_I--- 1LO 120 160 Bo 60 io 2’0 ppm
chain length (12 versus 15). The fraction remaining after
Figure 5 ‘T n m.r. spectrum of the Fischer-Tropsch jet fuel.
urea adduction is enriched in monomethyl branched Resonance assignments in the olefmic carbon region (114-140 ppm) are
alkanes (N, = 1.O in Table I ) and, as a result of the urea shown together with the corresponding olefm structures: 1, 1-ene; c2,
adduction procedure, there is a relative reduction of cis 2-ene; t2, trans 2-ene; 2m2. 2-methyl 2-ene; and t3. trans 3-ene

FUEL, 1989, Vol 68, June 779

 Examination of Fischer- Tropsch products using n.m.r.: 0. J. Cookson and B. E. Smith

Table 3 Assignment of oletinic ‘H n.m.r. intensity (4.5-6.0ppm) for

the Fischer-Tropsch jet fuel and comparison of calculated carbon
intensities with those measured directly by 13C n.m.r.

Oletinic carbon intensity

Measured Calculated
‘H chemical ‘H n.m.r. from from
shift range intensity 13C n.m.r. ‘H n.m.r.
(ppm) Assignment W) (%) (X)

6.s5.6 CH* =CH, 0.027 0.10” 0.12

5.054.8 CH = CH2* 0.054 1
I I
60 55 50 45 wm 5.6-5.3 CH* =CH* 0.21 0.43b 0.42

5.3-5.05 C=CH* 0.019 0.07’ 0.08

4.84.5 C=CH,* 0.010 ndd 0.02

a Structure 1 (see FigureS)

*Includes contributions from structures t2, c2 and t3 (see Figure 5)
I 1
‘Structure 2m2
7 6 5 L 3 2 I pm dCorresponding structure not observable in t3C n.m.r. spectrum,
Figure 6 ‘H n.m.r. spectrum of the Fischer-Tropsch jet fuel. probably due to signal/noise limitations
Resonance assignments in the oletinic proton region (6.0-4.5 ppm) are
shown together with the corresponding olefinic proton groups
The spectral signal/noise for the diesel fuel olelinic ‘H
It is apparent that most olefins are internal (z 84%) and ’ 3C resonances is substantially poorer than observed
rather than terminal (z 16%). Also, trans isomers are for the jet fuel. There is little point, therefore, in pursuing
more abundant than cis isomers as is expected on detailed integrations and calculations. Nevertheless,
thermodynamic grounds 16*17 Since the average number comparison of Figure la with Figure 6, as well as
of carbon atoms per olelin molecule for this 19s230°C Figure I b with Figure 5, clearly indicates that the spectral
fraction must be approximately equal to that observed patterns observed for the diesel fuel are closely
for n-alkanes (Nc= 12), the 0.6% of i3C n.m.r. intensity comparable with those found for the jet fuel. Relative
observed as due to olefinic sp’ carbon, implies that the abundances of isomers/structures are therefore similar
sample contains approximately 3.6 wt% olefin molecules. for both diesel and jet fuel samples. The percentage of
Thus, the absolute wt% abundances of olefin isomers is 13C n.m.r. intensity due to olelinic carbon is approxi-
readily calculated (Table 2). mately 0.2% for the diesel fuel, which implies a total
abundance of olelins of 1.5 wt%.
‘H n.m.~. Since the 13C n.m.r. data implies that we are
dealing with simple olelins, the ‘H n.m.r. data may be
CONCLUSIONS
interpreted in a straightforward manner18-20 to yield
quantitative structural information in a distinctive but In the diesel fuel, a total of 12.2 wt% branched alkanes
complementary fashion to that from 13C n.m.r. Olelinic and 1.5 wt% olelins have been found. The remainder
proton groups give rise to intensity in distinctive (86 wt%) is almost wholly n-alkanes. This value is in good
regions (Figure 6), namely - CH* = CH, (6.c5.6 ppm); agreement with the g.c. estimate of n-alkane content
CH = CH2* (5.05-4.8 ppm); CH* = CH* (5.6-5.3 ppm); (83 wt%) given the inherent limitations in integration for
C = CH* (5.3-5.05 ppm) and C = CH,* (4.884.5 ppm). both g.c. and n.m.r.. Similarly, the jet fuel has been found
Integrated intensities report on the abundances of these to consist of 9 wt% branched alkanes and 3.6 wt%
olefinic ‘H group types (see Table3). olelins, leaving 87 wt% of n-alkanes. The latter value is
Just 0.32% of total ‘H n.m.r. intensity is attributed to also in good agreement with that estimated by gc.
olefinic ‘H resonances. Just 0.6% of 13C n.m.r. intensity (84 wt%).
has been attributed to olelinic carbons. In both cases In the examples chosen for the present study, a very
spectral features have been assigned to specific structures, detailed approach has been taken (including urea
and relative intensities have been measured. It is adduction, selected multiplet ’ 3C n.m.r. spectra, ‘H-l 3C
instructive to compare ‘H and 13C n.m.r. measurements compatibility checks, etc.). This detail has been pursued
as a check on the self-consistency of proposed to establish the credibility of structure assignments and
assignments and on the accuracy of intensity measure- interpretations of intensities. For subsequent related
ments. Consideration of stoichiometries makes it a simple samples, a much simpler approach based on conventional
matter to convert the observed per cent ‘H n.m.r. 13C n.m.r. and/or ‘H n.m.r. would be sufficient.
intensities reported in Table 3 to per cent 13C n.m.r.
intensities. Table 3 shows that these calculated 13C n.m.r. ACKNOWLEDGEMENT
intensities obtained from ‘H n.m.r. compare very well
with the intensities measured directly from 13C n.m.r. The authors thank the Broken Hill Proprietary Co. Ltd.
given the limitations of signal to noise. Thus, for the for support and for permission to publish this work, and
jet fuel, ‘H and 13C n m.r. observations provide a Dr P. J. Jackson for supplying the samples. Partial
fully consistent and complementary view of olelin support was also provided under the National Energy
composition. Absolute wt% abundances of olefin species Research Development and Demonstration Program,
containing the structural types identified via ‘H n.m.r. administered by the Australian Commonwealth Depart-
are given in Table2. ment of Primary Industries and Energy.

780 FUEL, 1989, Vol 68, June

Examination of Fischer-Tropsch products using n.m.r.: D. J. Cookson and B. E. Smith

REFERENCES 8 Cookson, D. J. and Smith, B. E. Anal. Chem. 1982,54, 2591

9 Cookson, D. J. and Smith, B. E. 1. Magn. Reson. 1984,57,355
10 Cookson, D. J. and Smith, B. E. Org. Magn. Resort. 1981,16,1 I I
Anderson, R. B. in ‘Catalysis Vol. IV, Hydrocarbon Synthesis, 11 Cookson, D. J. and Smith, B. E. Fuel 1983,62,986
Hydrogenation and Cyclization’ (Ed. P. H. Emmett), Reinhold 12 Cookson, D. J. and Smith, B. E. Energy & Fuels 1987, 1, 111
Publishing Company, New York, USA, 1956 13 Cookson, D. J., Smith, B. E. and Shaw, I:M. Fuel 1987,66,758
Thvrion. F. C. in ‘Svnthetic Fuels from Coal’ (Eds. I. Romev. 14 Netzel, D. A., McKav._ D. R., Heonner. R. A. et al. Fuel 1981.
.I

P. F. M.‘Paul and G:Imarisio), Graham and Trotman, London, 60, 307

UK, 1987 15 Couperous, P. A., Clague, A. D. H. and van Dongen, J. P. C.
Dry, M. E. in ‘Catalysis, Science and Technology’ (Eds. J. R. M. Org. Magn. Reson. 1976,8, 426
Anderson and M. Boudart), Springer-Verlag,Berlin, FRG, 1981, 16 Rossini, F. D., Pitzer, K. S., Amett, R. L., Braun, R. M. and
Vol. I, Ch. 4 Pimentel, G. C. ‘Selected Values of Physical and Thermo-
Gray, D., Lytton, M., Neuworth, M. and Tomlinson, G. ‘The dynamic Properties of Hydrocarbons and Related Compounds’,
Impact of Developing Technology on Indirect Liquefaction’, Carnegie Press, Pittsburgh, USA, 1953
Report No. MTR-80W326, Mitre Corporation, VA, USA, 1980 17 Stull, D. R., Westrum, Jr., E. F. and Sinke, G. C. ‘The Chemical
Cookson, D. J. and Smith, B. E. Anal. Chem. 1985,57, 864 Thermodynamics of Organic Compounds’, John Wiley and
Cookson, D. J., Rolls, C. L. and Smith, B. E. Fuel 1989, 68, Sons, New York, USA, 1969
788 18 Stehling, F. C. and Bartz, K. W. Anal. Chem. 1966, 38, 1467
Weiss, F. T., ‘Chemical Analysis’; ‘Determination of Organic 19 Pouchert, C. J. and Campbell, J. R. ‘The Aldrich Library of
Compounds: Methods and Procedures’, (Eds. P. J. Ewing and NMR Spectra’, Aldrich Chemical, Milwaukee, WI, USA, 1974
I. M. Kolthoff), Wiley-Interscience, New York, USA, 1970, Vol. 20 Wilson, M. A., Lambert, D. E. and Collin, P. J. Fuel 1985, 64,
32, Chapter 1, p. 14 1647

FUEL, 1989, Vol 68, June 781