H++--=**:++ 0016-2361(95)00030-5

F,e/ Vol 74 No. 7, pp. 1011-1017,1995

Copy¡isht O 1995Elsevi€rScienceLrd
P¡intedin Great B¡itain.All rishls reserved
0016-2361/95/S10.00 + 0.00

Comparison of oils and asphaltenes from

extraction of bituminous coal and lignite
with hydrogenated anthracene oil

Roberto Rosal, Fernando V. Díez, Herminio Sastre and Golin E. Snape*

Depattamento de lngenieríaQuímica,Universidadde Oviedo,33071Oviedo,Spain
. Departmentof Pureand Applied Chemistty,Univers¡tyof Strcthclyde,GlasgowGl 1XL,
UK
(Received13J une 1994)

The applicationol'H n.m.r. spectroscopyand g.c. m.s. to the analysisof aromaticfractionsobtainedin

the liquefactionof two coalsoldiffe¡ent ¡ank is desc¡ibed.The conversionand the yield ofasphalteneshave
been¡elatedto the p¡oportion of aliphatic hydrogenin the ext¡actioú oils afte¡ liquefaction.The n.m.r.
bands for asphaltenesare simila¡ to those of recoveredanthraceneoils, but massspect¡ashow that a
substa¡ltialamount of high-molecular-weight compoundsis producedduring the extractionola bituminous
coal,compoundsnot encounter€din lignite liquelaction.

(Keywords:
coal;liquefaction;
asphslt€nes)

Large-scale direct liquefaction processesare being
developedto replacec¡ude oil by high-quality liquids
that could be processedin current refining facilitiesl.
However,it is not likely that coal-derivedliquids will
competewith petroleumin the near futu¡e. The needs
for refiningto meetenvironm€ntalregulationsmay delay
the implementationof these processesuntil dwindling
petroleumresourceshaveto be supplemented. None the
less,there is considerableinterest in the production of
fractions as precursorsfor binder pitch for anodes,
advancedcarbon materialsetc.2-s.Coal tar oitch from
coke ovens is currenrl¡ the major raw máterial [or
coal-basedca¡bon fibres and other ca¡bon materials.
However,the amountand quality ofcoal tar are declining
steadily while the demand for this product and most
aromaticchemicalsis rapidly increasing.
Subbituminousand low-rank bituminous coals have
beenrecognizedas major taw materialsfor liquefaction
processes.However, some interest also remains in
lower-rank coals. Lignite, for example, constitutes a
major resourcein the USA and Russiaand a considerable
lraction of the total reservesof fossil fuels in some
Europeancountdessuch as Germanyand Spain.
Theanalysisolthe liquid fractionsfrom coalliquefaction
is normally performedby chromatographicand spectro-
scopic methods.Most of the componentsof the light
fractions are identified and quantified by gas chroma-
tographyand massspectrometry.The analysisof heavier
fractions is usually based on structu¡al parameters
derivedfrom n.m.r.spectroscopy, providing information
on bulk compositional properties of these extremely
complex mixtures6-8.The determination of individual
componentsis hinderedby their lack of volatility and is
generallylimited to moleculesofmolecularweight ¡ 300,
which in the caseof aromaticsystemsrelersto compounds
of up to five to six ringse.
This paper describesthe applicationof 1H n.m.r. and
g,c.-m.s.usingelectronimpact ionization to the analysis
ofthe oils and asphaltenes obtainedfrom the liquefaction
of a lignite and a bituminous coal.
EXPERIMENTAL
Materiqls
A low-volatile bituminous coal (True Energy, Olga)
and a lignite (Meirama, La Coruña) were used.Their
proximateanalysisare givenin Tablel. Beforereaction,
the coals were ground to three particle size ranges:
0.03G{.100,0.10G{.400and0.40G-1.500 mm. The samples
weredried at 110'C to constantwsight beforereaction.
The anthraceneoil (90% recoveredbetween215 and
400"C) was hydrogenated over a sulfided nickel-
molybdenum catalyst (BASF, M8-24) to obtain two
fractions (denoted 1HAO and 2HAO) with different
concentratiónsof hydroaromaticcompounds.The hydro-
genationprocedurehas beendescribedpreviouslylo.
M¡üoautoclqueextraction
Batch hydrogen transferextÍactionswere conducted
in 15cm3 stainlesssteelbomb reactorsshakenvertically
in a fluidized bath of alumina sand at 600 cyclesper
minute and with an oscillation amolitude of 25 mm.
Reactionconditionswe¡e monitoreá with a oressure
rransducer and a thermocouple.The devicewai similar
to that describedby Neavelrr.The coal (3g) and the
solvent(9 g) were accuratelyweighed,thoroughly mixed
and transferredto the bomb. Reactionswerecar¡iedout

F u e l1 9 9 5 V o l u m e7 4 N u m b e r7 1013
Comparisonof o¡ls and asphaltenes:R. Rosal et al.
Table I P¡ope¡tiesof the coalsused
Meirama Olga
Moisture {wt% as ¡eceived) 51.4 7.8
Proximateanalysis(wt% db)
Ash 8.3 6.0
Volatile matter ) t.l 19.6
Sulfur (wt% db) 1.3
0.6
under nitrogen (3 MPa at toom temperature)to ensure
the absenceof hydrogen transfer from the gas phase.
Liquefactionwasperformedat temperatures between375
and 425"Cand for reactiontimesbetween20 and 60 min.
Additional detailsa¡e given elsewherel2.
Anaf,ses
After the reaction,the bomb wasquenchedwith water,
gaswasventedandthe restofthe productswereseparated
by sequentialext¡action into tetrahydrofuran(THF)-
insolubles,
preasphaltenes (THF-soluble,toluene-insoluble),
asphaltenes (toluene-soluble,
r-hexane-insoluble)and oils
(n-hexane soluble)13.The weightofthe insolublefractions
wasrecordedafter2 h at 120"C.The amountofinsoluble
organic residue was calculated assuming a constant
weight of mineral matter l¡om the coal.
Anthracene oil and the n-hexane-solublefractions
were analysed by gas chromatography with a 50 m,
0.25mm i.d. capillary column and OV-101 as stationary
phase. Peaks were assignedby g.c. m.s. Calibration
factors were determinedwith pure standardswhenever
available;lor other compoundsthe calibration factors
wereassumedto be unity, with acceptableaccuracytn.
rH n.m.r.spectraofthe solublefractionswererecorded
on a Bruker 300MHz instrument.Deute¡ochloroform
was used as solvent and tet¡amethylsilaneas internal
standard.Chemicalshiftswere assignedas follows:
0.5-1.5ppmCH, and CH in naphthenicrings and
further than B from an aromatic ring;
non-cr-CH3 groups(Hp.)
1.5-2.0ppm CH, and CH B to an aromaticring and in
naphthenicrings (H")
2.0-3.5ppm CH3, CH2 and CH a to an aromatic dng
(H,)
3.5-4.5ppm ring-joining methylene(H,2)
6.9-9.1ppm aromatic protons (Hu.)
The hydrogen aromaticity (fJ is defined as the
fraction of total hydrogen bonded to aromatic carbon.
For anthraceneoil beforehydrogenation,¿H:0.755,and
for the hydrogenatedfractions used in the extraction
Ieactions,¿H:0.689 (lHAO) andÁH:0.600 (2HAO).
Analysesofoils wereperformedby gaschromatography
on a capillary OV-101 column. Electronimpact spectra
were obtainedat 70 eV.
RESULTS AND DISCUSSION
The yields of preasphaltenes, asphaltenesand oils were
calculatedrelative to the initial weight of dry ash-free
coal. The tetrahydrofuran-solubleconversion(THFSC)
was calculatedas follows:
(initial m.a.f.coal)-(THF ash-freeinsoluble)
THFSC:
(initial m.a.l coal)
1 0 1 4 F u e l1 9 9 5 V o l u m e7 4 N u m b e r7
Extractions were performedunder conditions that did
not favour the formation of light products.The yieldsof
oils and gaseswas calculatedby differencebecausethe
light fractions contained a substantial amount of
volatile compounds,someof which wereinevitably lost
during solventevaporation.The calculationignoredthe
possibility of formation of asphaltenesfrom the solvent
and considered only the occurrence of progressive
reactionsl2.On this assumption,the amount of oils that
could be attributedto coal-derivedmaterialwas < 5 and
< 10wt% for Olga coal and Meiramaligniterespectively.
Most of the r-hexane-solublefraction ol the reaction
products was a depletedanthraceneoil.
Figure1 showsthat coalconversion(THFSC)is closely
related to the amount of aliphatic hydrogen in the
extractionoils.The hydrogenaromaticityis considerably
higherin the oils from Olga bituminouscoal (fr>0.56)
than in thosefrom lignite exttaction.This is due to the
differencein compositionof the n-hexane-soluble fraction
obtainedfrom the coalin the two casesand not in amount
of oils relative to solvent.It was previously reportedl2
that und€¡ the reactionconditionsusedin this work, the
rate constantat 400'C for the reactionof asphaltenes to
oils for Olga bituminouscoal is more than 30 timesthat
for Meirama lignite liquefaction.As a consequence, the
diflerencein reactivity betweenthe two coals ¡esultsin
a much greateryield of asphaltenes in lignite extraction,
and a lower yield of liquefactionoils. Aliphatic chains
make a signiflcant contribution to overall hydrogen
distribution. The disparity betweenthe two setsof data
should be attributed to the more aliphatic characterof
the reaction products from lignite. For both coals,but
more clea y for the bituminouscoal, a lower hydrogen
aromaticity is observedin the oils from runs using the
more hydrogenated solvent. The higher content of
hydroaromaticsresultsin a higheryield oflight products
from extraction and subsequentlya higher content of
aliphatic protons in the oils.
In the runs performed with Olga coal, different
reactivitieswe¡eobservedfor differentoarticlesizes.The
inilial rate of coal dissolutionfoi small oarticles
{0.0100.100mm) was more rhan twice rhat foi 0.400
1.500mmparticles.This led to differencesin conversion
which, however,resultedin a similar quantity of oils
o
n
!q
é,
I
Alipharichydrogen(%)
Figure 1 Co4version of coal to tetrahydrofu¡an-solublefraction
(THFSC) as a function of aliphatichyd¡ogencontentof oils obtained
in liquefaction.Meirama lignite (O); Olga bituminousext.actedwith
1HAO (tr) and 2HAO (t)
 Comparison of o¡ls and asphaltenes:B. Rosal et al.

becauseof the damping originating from the fact that

oils are mainly a secondaryreactionproduct and the rate
of coal dissolutiondoesnot affectth€ rate constantsfor
preasphalteneand asphaltenedegradation.The points ts
located above the line in Figure 1 represent runs
performed with particles <0.100mm; the remainder
correspond to sizesof0.100{.400 and 0.400-1.500mm.
The distribution ol points for the bituminous coal
indicatesthat thereis no significantcorrelationwith the
prima¡y factors: particle size and aliphatic content
of the HAO. For lignite reactions no differencewas
0.1
obtained\ryithdifrerentparticle sizesand thero was no
E
dispersionsuch as that encounteredin the caseof Olga
bituminouscoal.The most probableexplanationfor this
behaviouris that th€ lignite underwentrelativelyrapid
disintegration,masking any effect of different initial (HdHú)oir
particlesizer5,16.
Figure4 Non-d-aliphatic hydrogen as a function ol H,/H", in
The relative yields of asphaltenesare plotted in liquefactionoils. Symbolsas in ]tigrle J
Fígure2 asa finction ofthe amountofaliphatic hydrogen
in the oils. The differencein slopefor the two coalsis a
Table2 Proton distributionin selectedsamples,
consequence of the diffe¡encein global rate of formation
of asphaltenes. For Meirama lignite,the specificrate for H,"/H^, HJH., H./H".
the reaction of asphaltenesis two orders of magnitude
lower than that of formation, and therefore they Fresh anthraceneoil 0.058 0.116 0.008 0.083
lHAO 0.124 0.253 0.023 0.051
2HAO 0.067 0.336 0.0'72 0.192
Aslphaltenes'
Olga 0.037 0.195 0.033 0.0'77
Meirama 0.041 0.213 0.036 0.065
Liquefactionoils
"/ Olga,1HAO 0.049 0.21s 0.083 0.117
/ Olga, 2HAO 0.052 0.246 0.103 0.213
0,8 Meirama, IHAO o.062 0.412 0.145 o.379
E- Meirama,2HAO
3 0.051 0.611 0.219 0.516
b ""/,,
'Liquefaction
conditions:
425"C.
ó0min
"/" 'Using2HAO

accumulateas the reaction proceeds.For Olga coal,

\.
" however,the rate constantfor the degradationreaction
¡mdr 12.
\ is higherthanthatoftheirformationfrompreasphaltenes
Figures3 and.4 relatethe proton aromaticitiesand the
02040ó080 proportions of non-a-aliphatic hydrogen of several
Aliphatichydrogen
(7o) samplesof oils and asphaltenesto the proportion of H,.
Fig¡rre2 Yield of asphaltenesrelative to the initial daf coal as a The two samplesof asphaltenes for which hydrogendata
fúnctionof aliphatichyd¡ogencont€ntof oils obtainedin liquefaction. are reportedwereobtainedf¡om the liquefactionof Olga
Symbolsas in ¡brl¿ J coal and Meirama lignite at 425'C for 60min with the
more hydrogenatedsolvent (2HAO). There is little or
no differencebetween the hydrogen distributions for
asphaltenesand oils from runs in which conversionwas
low. In suchcasesJ the extractionoil was predominantlv
a spent anthraceneoil and the compoiition of botÍr
systemsis basedon condensedaromatics.Full hvdrosen
'o o,7
distributions[or selectedsamplesare present.d-in
Table 2. Figure 5 shows an example of a lH n.m.r.
spectrumof an asphaltenesampl€from the processingof
É Olga coal at 425'C for 60min.
É The mass spectra of the asphaltenesfrom coal
liquefactionrevealedtracesof the heaviercomponentsof
the solvent.appearingin lhe r-bexane-insolubie fracrion.
Pyrene, fluoranthene, benzopyrenesand some other
compoundsthat werepreviouslydetectedin the solventla
0 0.2 0.4 0.6 0.E appearedin the EI mass spectra together with other
(Hcq/Ha)o¡ compoundsthat presumablyarose.frgm anthraceneoil,
-iroton
Figuib:
suchasdibenzopyrenes and coronene,The reasonis that
drbmati¡ity as a function ot H"/H". in liquefaction
oils: a, asphaltenes from extraction at 425"C; ¡, oils f¡om Olga no specialprecautionswere taken to remove distillable
bituminous coal; O, oils from Meirama lignit€ material occludedin the asphaltenes. It should be also

F u e l1 9 9 5 V o l u m e7 4 N u m b e r7 1015
Comparison of oils and asphaltenes;R. Rosal el al.

Figure 5 1H n.m.r.spectrumo{ an asphaltenesampleobtainedfrom Olga bituminouscoal

302328 42?
354376400

Ít/z

Figur€ 6 EI m.s. spectra of asphaltenesamples from (A) Olga bituminouscoal a¡d (B) MeiramaIignite

1 0 16 F u e l 1 9 9 5V o lu m e 7 4 N u m b e r7

pointed out that the mass spectraare only qualitative,
giving rise to fewer parent ions than softer techniques,
particularly Fl m.s. Figure ól corresponds to an
asphaltenefrom Olga coal whereas-B shows EI mass
spectraofan asphaltenefraction obtainedfrom Meirama
lignite. Hydrogen deficiency compared with alkanes,
conventionally denoted by a number z defined as
C,Hrn*", ranged from -20 to -40 for these highly
condensedand unbranched compounds.The spectra
showedthe presenceof longer alkyl chain homologues
in the region of higher-molecular-weightsubstances,
which could be attributed to coal fragm€nts.
In the extractionswith Olga bituminous coal, there
werepeaksfor a signiñcantnumberof massesrepresenting
> 4lyo of the total ion current (TIC) with massesup to
550. Lignite asphaltenesappeared to have a much
narrower range,typically <450. A possibleexplanation
for this behaviour is that lignite asphaltenesare more
polar, so they are included in the asphaltenesolubility
classeven though they have a lower molecularweight.
The ¡esults indicated that the amount of asphaltenes
generatedduring the extractionoflignite washigherthan
that obtainedfrom Olga coal.The hydrogendistribution
from n.m.¡. analyses,however,was similar. Apparently,
themolecularweightoffragmentsfrorn the coal of higher
rank wasgreaterbut the structuresofthe lignite products
werenot greatlydifferentin alkyl substitution.Another
effectthat could partly explainth€ similarity betweenthe
n.m.r. bands is masking by the asphalteniccompounds
f¡om the anthraceneoil. The fraction of asohaltenes that
couldbe atl.ributed to the solventis in the iange2O-40%
lor the lignite and up to 40% fo¡ the bituminous coal,
dependingon the extractionconditions.
ACKNOWLEDGEMENTS
This researchwassupportedby FICYT (Foundationfor
the Support of Research in Applied Science and
Technology, Asturias, Spain) and British Council
Comparison of oils and asphaltenes;R. Rosal et al.
Ministry of Education of Spain (CooperativeResea¡ch
Projects).The authorsare also gratefulto the University
of SwanseaSERC massspectrometryservice,wherethe
m.s.analyseswereperformed.
REFERENCES
1 Motuñí,E. C. Aü. Chem.Soc.Diu. Fuel Chem.Preptints1991,
2 Schultz,H. D.I¡'Coal Liquefaction P¡oducts:NMRSpectroscopic
Cha¡¿cteízationand ProductionProcesses'lEd. H. D. Schuttz).
Wiley, New York, 1982
3 M¿rsh. H. (Ed.)'Ne$ and traditional carbon mate als f¡om
p€troleum and coal sources',Paper presentedat Information
T¡ansfer Session,Cooperative Program in Coal Research,
PennsylvaniaStateUniversity,November1991
4 Donnet, J. B. and Ba¡sal, R. P.'Carbon Fibers',Dekker, New
Yo¡k. 1984
5 Walker, P. L. Catbon 1990,28,261
6 Snape,C. 8., Ladnet W. R. and Bartle,K.D. Fael1985,64,1394
7 Snape, C. E. In'Analytical NMR' (Eds L. D. Field and
S. Stemh€]I),Wiley, New York, 1989
8 Snape, C. E., Ladne¡, W. R. and Bartle, K. D. In 'Coal
LiquefactioúProducts:NMR SpectroscopicCha¡acterization
ard ProductionProcesses'(Ed. H. D. Schultz),Wiley,New York,
1982
9 Burchill, P., Hc¡od, A. A. and James,R. G, In 'Carcillogenesis',
Vol. 3, 'PolynuclearAromatic Hydrocarbons'(Eds P. W. Jones
and R. I. Freude¡thal),Rav€nP¡ess,New Yo¡k, 19?8,p. 35
10 Roasl,R., Díez,F. V. and Sastre,H.Ird. Eng. Chem.Res,1992,
31, 1007
11 Neavel,R. C. Fuel 1976,55,237
12 Rosal,R., Díez,F. V. and Sashe,H.Ind, Eng, Chem.Res.1992,
31,2441
13 Mima, M. J., Schulz,H. and McKinstry, W. E.'Method lo. the
determinationof benzeneinsolubles.asDhaltenes and oils iú
coal derivedo¡l liquids. eenCTnt Ze¡0.US EnergyR&D
Administration,1976
14 Rosal,R., Diez, F. V. and Sast¡e,H. ¡re|1992,11,761
15 Whitehurst, D. D., Mitchell, T. O. and Farcasiu,M.'Coal
Liquefaction: the Chemistry and Techúology of The¡mal
Processes', AcademicPress,New York, 1980,pp. 108 112
16 Pradham,V. R., Holder, G. D., Wender,L and Tiemey, J. W.
Ind. Eng. Chem.Res.1992,31,2051
F u e l1 9 9 5 V o l u m e7 4 N u m b e r7 1 0 ' 1 7