Journal of Analytical and Applied Pyrolysis

65 (2002) 111136
www.elsevier.com/locate/jaap

Vacuum pyrolysis of sugarcane bagasse

Manuel Garc`a-Pe`rez a, Abdelkader Chaala b,
Christian Roy a,b,*
a
Departement de genie chimique, Uni6ersite La6al, Ste-Foy, Que., Canada G1K 7P4
b
Institut Pyro6ac Inc., 333 rue Franquet, Ste-Foy, P. Que., Canada G1P 4C7

Received 24 March 2001; accepted 28 September 2001

Abstract

The objective of this study was to provide background data on sugarcane bagasse vacuum
pyrolysis. Product yields and properties were investigated. Vacuum pyrolysis tests were
performed at bench and pilot plant scales. The bagasse finest particles with a diameter
smaller than 450 mm were removed in order to overcome difficulties caused by their low
density and high ash content. In comparison with the pyrolysis test carried out in the pilot
reactor, the pyrolysis tests performed at the laboratory scale yielded more oil (34.4 vs. 30.1
wt.%) and less charcoal (19.4 vs. 25.7 wt.%). This bio-oil was found to be a potential
valuable liquid fuel: it has a low ash content (0.05 wt.%), a relatively low viscosity (4.1 cSt
at 90 C), a high calorific value (22.4 MJ/kg) and a low content of methanol insoluble
materials (0.4 wt.%). Accelerated ageing tests performed at 80 C indicated that the bio-oil
has a thermal susceptibility similar to oils obtained from other biomass materials. The
charcoal, gas and aqueous phase were also characterized.
2002 Elsevier Science B.V. All rights reserved.

Keywords: Vacuum pyrolysis; Sugarcane bagasse; Charcoal; Bio-oil; Ageing

1. Introduction

Over the last two decades, a special attention has been paid to the conversion of
residual biomass and renewable materials into biofuels. Use of waste agricultural
biomass such as bagasse from the sugarcane industry is an attractive solution to

* Corresponding author. Tel.: + 1-418-656-7406; fax: + 1-418-656-2091.

E-mail address: [email protected] (C. Roy).

0165-2370/02/$ - see front matter 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 1 6 5 - 2 3 7 0 ( 0 1 ) 0 0 1 8 4 - X
112 M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136

mitigate the Third World energy needs. It can also contribute to the global CO2
reduction. Typically, one ton of crude sugarcane can produce 100 kg of sugar, 35
kg of molasses and approximately 270 kg of dry bagasse [1]. Oven-dry sugarcane
bagasse is a solid fuel with a heating value of approximately 18 MJ kg 1 [2,3].
Generating power by direct combustion of sugarcane bagasse in boilers has a
maximum efficiency of 26%. Combustion systems with low efficiency are tradition-
ally used in sugarcane plants. In populated areas, bagasse-fired boilers can be one
of the major health hazards due to airborne fly ash. Beaton et al. [4] and Brito et
al. [5] recently disclosed an improved method to increase the efficiency and capacity
of bagasse boilers. Use of more efficient technologies could create new opportuni-
ties to export power [6 8]. Bagasse transformation into high-density renewable
fuels, like charcoal and bio-oil, can significantly increase the profitability of
sugarcane plantations [1]. Such solid and liquid fuels can be easily transported to
distant consumers.
Two families of biomass conversion technologies have been investigated over the
last two decades. Gasification and pyrolysis can advantageously achieve higher
energy conversion than combustion. During gasification, simple molecules (H2, CO)
are produced as biomass is converted to gases. Souza-Santos [9] wrote a paper on
the feasibility of sugarcane bagasse gasification. Biomass Integral Gasifier/Gas
Turbine at high pressure (BIG/GT) provides efficiencies at least 10% higher than at
atmospheric pressure. These results were used by Ponce et al. [6] to demonstrate the
economical and environmental advantages of sugarcane gasification in a typical
Cuban sugarcane mill. The other method is pyrolysis which offers a rapid route to
crack the biomass polymeric structure to high yields of liquid products that are the
source of valuable products [1,2,10 12].
Carry-over of small bagasse particles in bagasse fluidized reactors is troublesome.
The finest particles may find their way in the gaseous product stream. These
problems markedly affect the gas product quality and require the installation of
expensive gas cleaning systems. Pelletization of sugarcane bagasse and separation of
bagasse raw feedstock into individual fractions have been studied as alternatives to
avoid elutriation problems during fluidization [13,14]. On the other hand, lower
turbulence inside moving and stirred bed pyrolysis reactors such as the one
developed by Pyrovac [15,16] reduces the carry-over of particles along with the
condensable products.
Drummond et al. [1] studied pyrolysis of ground, dried and sieved sugarcane
bagasse (100150 mm) in a 7 mg wire mesh reactor. The role of important pyrolysis
parameters like peak temperature and heating rate on the tar yield was investigated.
The authors obtained 54.6% by wt. of total condensable liquids on a bagasse
anhydrous basis. Zandersons et al. [2] carried out experiments in a laboratory scale
fixed bed reactor using 500 g of bagasse briquettes. They demonstrated that
pyrolysis of bagasse produces 23 28 wt.% charcoal on a bagasse anhydrous basis.
Xia et al. [17] performed experiments with bagasse to produce activated carbon by
steam and chemical activation. According to Mobarak et al. [11] high amounts of
tar are produced from bagasse during rapid pyrolysis. Ether extraction produces a
fraction rich in phenolic compounds with yields of 9.4% by wt. (bagasse anhydrous
 M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136 113

basis). Brossard et al. [12] pyrolyzed bagasse in a 50 kg h 1 continuous slow

pyrolysis reactor. The authors obtained approximately 30% charcoal, 45% pyrolig-
neous acid phase, 5% tar and 20% gas (wt.%, bagasse anhydrous basis). Bagasse
vacuum pyrolysis has been first reported by Pakdel et al. in 1996 [10].
Several biomass-derived oil combustion experiments in conventional boilers have
been conducted [18 20]. Burning bio-oils in a gas turbine has been discussed by
Elliot in 1994 [21]. Andrew et al. [22] reported on the combustion of approximately
13 000 l of bio-oil in a 2.5 MW gas turbine pilot unit. The authors concluded that
the technical limitations observed can be overcome by optimizing the bio-oil
properties in terms of alkali, ash content, viscosity, heating value and by modifying
the gas turbine engines.
The main objective of the present work was to assess the yields and characteristics
of the products obtained by vacuum pyrolysis of bagasse.

2. Experimental

2.1. Materials

2.1.1. Sugarcane bagasse

The sugarcane bagasse was provided by United States Sugar Corporation, Clewis-
ton, Florida. Bagasse consists of three main fractions: pith, fibre, and ring.
Typically, the proportion of pith is approximately 5% by wt., fibres represent 73
wt.% and ring 22 wt.% [23]. The bagasse received in the laboratory contained 46
wt.% moisture. The moisture content was lowered to 68 wt.% after three days of
natural drying in ambient atmosphere. The bagasse sample was sieved in a Hoskin
Scientific sifting machine during 10 min. Each fraction was subjected to proximate
analysis (see Table 1).
The ash content of the smallest particles is much higher. Additionally, the finest
particles exhibit a very low bulk density (220 kg m 3) in comparison to the larger
particles (550 kg m 3). Thus, it was desirable to remove the finest particles
(diameter B450 mm) in order to reduce the entrainment of fines in the process gas,
otherwise such fines might find their way into the condensates resulting in an
increased methanol insoluble material (MIM) content of the oil. The discarded

Table 1
Proximate analysis of bagasse fractions (wt.% bagasse anhydrous basis)

Dimensions (mm) Fraction Volatile matter Ash Fixed carbon

d\4.75 34 82.3 1.3 16.4

0.85BdB4.75 44 82.1 1.6 16.3
0.45BdB0.85 15 81.8 2.3 15.9
0.25BdB0.45 4 72.5 13.0 14.5
dB0.25 3 60.0 27.7 12.3
114 M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136

fraction of fines can be burnt and hence supply a portion of the energy required for
the drying and the pyrolysis. Simultaneous drying and sieving of bagasse is a
feasible operation [24].
Proximate and elemental analyses of the sieved bagasse were performed using a
MAC-400 instrument and a CHN-600 apparatus from LECO, respectively. The
results are shown in Table 2. The hemicellulose, cellulose and lignin content of the
bagasse was determined using a thermogravimetric technique described by Garc`a-
Pe`rez et al. [25].

2.2. Vacuum pyrolysis tests

Samples of 80 g of sugarcane bagasse (6 wt.% moisture) were pyrolyzed (runs c

A-238, A-239, A-240 and A-241) in a cylindrical stainless steel retort. Some tests
were performed twice. The retort was placed in a Lindberg three zone vertical
electric furnace. The heating rate was 12 C min 1 until a temperature of 500 C
was reached. The reactor total pressure was kept below 8 kPa throughout the test.
Vapours were rapidly pumped out of the reactor across the bed of biomass (250
mm in height). The organic vapours and steam were condensed in three traps
connected in series and maintained at 30, 78, and 78 C, respectively (Fig.
1). The non-condensable gases were collected and stored in an empty vessel at room
temperature. The solid residue was left behind in the reactor. At the end of each
test, the reactor was cooled down to room temperature under a flow of nitrogen. A
detailed description of the equipment used can be found elsewhere [26].
A large scale pyrolysis test was performed in a pilot plant reactor (run c H-57)
in order to collect enough oil for the characterization tests. A 20 kg sample of
air-dry bagasse (8% by wt. moisture) was placed in four pans and introduced into
a cylindrical horizontal reactor 3 m long and 0.6 m in diameter. The bagasse bed
height was 140 mm. The sample was heated by an indirect electrical heating system
during 200 min until the feed material reached a temperature of 530 C. The pilot

Fig. 1. Vacuum pyrolysis laboratory scale installation.

Table 2
Elemental composition, proximate and chemical analyses of the sieved bagasse feedstock

a
C H N S O Volatile matter Ash Fixed carbon Hemicellulose and extractives Cellulose Lignin
Ash-free basis (wt.%) Bagasse anhydrous basis (wt.%) Water and ash-free basis (wt.%)

49.6 6.0 0.5 B0.1 43.8 82.1 1.6 16.3 35.8 43.1 21.1

a
By difference.
M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136
115
116 M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136

reactor, which was operated in a batch mode, was held at 530 C for 1 h. The solid
material was cooled down to room temperature under a flow of nitrogen. The
average total pressure during the test was approximately 12 kPa. Four traps
maintained at 25, 0, 30 and 80 C, were used as condensation train. A
detailed description of the equipment used can be found elsewhere [15,27].
The condensed liquids collected in the traps were mixed and then evaporated at
45 C during half an hour in a rotatory evaporator (Bu chi, RE 111). This process
was carried out in order to remove the lightest compounds and a portion of the
water. The objective of this operation was to obtain an oil sample as close as
possible to the one produced using a large condensation system installed in an
industrial vacuum pyrolysis plant such as the one operated by Pyrodev Inc. (3.0 t/h)
in Jonquie`re, Que bec, Canada [28]. The evaporation residue is referred to as
bio-oil in this paper and the distillate consisting of light organic fraction and water
is called aqueous phase.

2.3. Characterization methods

2.3.1. Gas analysis

The non-condensable gas phase was analyzed in two Varian model 3700 gas
chromatographs. Each GC was equipped with two stainless steel 1.83 m 3.2 mm
o.d. columns connected in series and packed with 150180 mm size Porapak Q and
180250 mm size molecular sieves 5 A, , respectively. A detailed description of the
method used can be found elsewhere [29].

2.3.2. Condensate analysis

2.3.2.1. Physico-chemical characterization. The methanol insoluble material (MIM)

was determined following the method described by Oasmaa et al. [30]. The
Conradson Carbon Residue (CCR) was measured according to ASTM D-189
method. The water content was determined by Karl Fischer titration (ASTM
D-1744). Kinematic viscosity was measured with Cannon-Fenske upflow viscome-
ters (ASTM D-445-88). Flash point was evaluated using a Tag Closed Tester
(Fischer-Tag) (ASTM D 3828-93). Density was measured using a 10 ml picnometer
(ASTM D-369). Gross calorific value was determined in a Parr calorimetric bomb
(ASTM D-4809). The ash content was determined according to ASTM D-482
method. The metal content was measured by Inductively Coupled Plasma-Atomic
Emission Spectroscopy (ICP/AES) on an ash sample previously dissolved in
mineral acids (mixture of nitric and hydrochloric acids at a 2:1 ratio).
Microscopic analysis of the bio-oil was performed with an optic microscope using
plane and cross polarized light. The bio-oil sample was pressed between glass slides
and then a microphotography was performed with photo camera.

2.3.2.2. Distillation. Samples of 400 g of oil (run c H-57) were distilled both under
atmospheric and vacuum (15 kPa abs.) conditions in the temperature range of
25 140 C. The distillates were subjected to water analysis. The equipment con-
sisted of a flask, a heater, a condenser and a vacuum pump.
 M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136 117

2.3.2.3. Separation of bio-oil fractions. A 10 g bio-oil sample was successively

extracted during 20 min in a separating funnel with 100 ml of the following
solvents: pentane, benzene, dichloromethane, ethyl acetate, methanol. The solvents
were removed in a rotatory evaporator and the recovered bio-oil fractions were
weighed and analyzed by GC MS.

2.3.2.4. GC MS analysis. The fractionated and extracted bio-oil samples were

analyzed in a Hewlett Packard model 5890 gas chromatograph. The separation
was made using a HP5-MS fused silica capillary column, 30 m 0.25 mm i.d. with
0.25 mm film thickness. The GC oven temperature was held at 50 C for 2 min,
then programmed to 290 C at 5 C min 1. The injector temperature was 290 C
with split mode. Helium was the carrier gas. The end of the column was directly
introduced into the ion source of a HewlettPackard model 5970 series mass
selective detector (MSD) operated in an electron impact ionization mode. Typical
mass spectrometer operating conditions were: transfer line 270 C, ion source
250 C, electron energy 70 eV. Data acquisition was carried out with a PC base
G1034C Chemstation software and NBS mass spectra laboratory data base.
Computerized matches were manually evaluated by matching mass spectra and
retention time with standard compounds to ensure identification quality.

2.3.2.5. Fatty acids. One gram of pyrolytic oil sample was derivatized using a
freshly prepared diazomethane solution. The methylated oil sample was separated
with 15 g of dicloromethane. The dicloromethane soluble fraction was analyzed in
detail by GC MS. Standard decanoic, hexadecanoic, eicosanoic and docosanoic
acids were methylated and their response factors were measured with respect to
anthracene, which was used as an internal standard during the course of this work
[31].

2.3.2.6. Low molecular weight acids. The C1 C7 carboxylic acids in the condensed
total liquid (bio-oil+aqueous phase) were derivatized into their benzyl esters and
analyzed by gas chromatography [32,33].

2.3.2.7. Sugars. A 400 mg sample of the total liquid (bio-oil+ aqueous phase) was
dried using benzene in a rotary evaporator. Five millitres of a 1 g l 1 solution of
sorbitole in pyridine was added to the oil. A half millilitre of the solution was
derivatized using a half millilitre of 1-trimethylsilylimidazole solution. The deriva-
tized solution was subjected to GC MS analysis. A standard solution of glucose,
levoglucosane, mannose, galactose, arabinose, xylose, ribose was prepared and their
respective response factors were determined with respect to sorbitole used as the
internal standard. The procedure described above was carried out to determine the
content of monomeric sugars in the oil. The polysaccharide content was determined
after hydrolysis and derivatization. A 1 g sample of the total collected liquid was
hydrolyzed using 1 N sulphuric acid solution during 4 h. The hydrolyzed solution
was than neutralized to pH 5 using a saturated solution of barium hydroxide in
water. The neutralized solution was evaporated to dryness and derivatized as
explained above.
118 M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136

2.3.2.8. Stability test. Bio-oil samples of approximately 80 g were placed in tightly

closed 80 ml volume bottles and heated in an oven for 168 h at 80 C. The bio-oil
thermal stability was evaluated by measuring the viscosity, solid content, molecular
weight distribution and water content before and after heating [30,34,35].

2.3.2.9. Molecular weight distribution. Molecular weight distribution was determined

by gel permeation chromatography (GPC) using a Waters 510 pump with 410
refractive index detector. Separation was performed using two columns in series:
Styragel HR 4E and HR 1 with 5 mm particle diameter. Both columns were 300
mm long with a diameter of 7.8 mm. Tetrahydrofuran (THF) was used as the eluent
at a flow rate of 1 ml min 1. Bio-oil samples were dissolved in THF at a
concentration of 2% by wt.

2.3.3. Charcoal characterization

A MAC-400 instrument from LECO, St. Joseph, MI, USA, was used to perform
the proximate analysis. Carbon, hydrogen and nitrogen content were determined
using a LECO CHN-600 apparatus according to ASTM D 5291-92 method. This
technique was also used for the elemental analysis of the bio-oil.

3. Results and discussion

3.1. Vacuum pyrolysis

Mass balances and product yields for the two different scale pyrolysis tests are
presented in Table 3. The pyrolysis tests at the laboratory scale yielded more
condensable matter, less charcoal and less gas than the pyrolysis in the pilot plant.
Bagasse decomposition reactions generated condensed liquids composed of 18.8
and 20.3 wt.% of pyrolytic water and 43.2 and 31.0 wt.% of organics (on bagasse
anhydrous basis) at the laboratory and pilot plant scales, respectively. The con-
densed liquids were formed of two distinct phases that could be easily separated by
decantation (a tarry phase and a water-rich phase). The former is rich in organic
compounds of relatively low polarity. These compounds are mainly derived from

Table 3
Product yields for the vacuum pyrolysis of sugarcane bagasse at various scales (wt.%, bagasse
anhydrous basis)

Laboratory scale Pilot plant scale

Liquid organic compounds 43.29 2.9 31.0

Pyrolytic water 18.89 2.5 20.3
Gas 17.6 9 2.2 22.0
Charcoal 19.4 9 1.2 25.6
Losses 1.09 0.5 1.1
Total 100 100
 M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136 119

lignin and amount to 5 9 wt.% on a bagasse anhydrous basis [2,10,12]. This phase
is rich in phenolic compounds. The water-rich phase obtained after decantation is
mainly composed of acids, sugars and other highly polar organics compounds. The
liquids collected at the laboratory and pilot plant scales were subjected to an
evaporation process. The evaporation residue called bio-oil, accounted for 34.2 and
30.1 wt.% (on bagasse anhydrous basis) of the products for the laboratory and pilot
plant conditions, respectively. The distillate that was removed and further con-
densed was called aqueous phase and represented 27.8 and 21.2 wt.% (on bagasse
anhydrous basis) under the laboratory and pilot plant tests, respectively.
When the biomass thermal decomposes, a thermal scission of the chemical bonds
occurs. This stage is referred to as the primary pyrolysis reaction step. The species
formed during this initial step in the bulk of the decomposing material may
undergo additional bond-breaking reactions in a condensed phase to form more
volatile compounds or may experience polycondensation/polymerization reactions
to form char. The volatile species may undergo further heterogeneous reactions
with the residual solids and/or homogeneous reactions in the gas phase [36]. The
vapour phase cracking homogeneous reactions convert some tar products into gases
and a small amount of char [36]. All these reactions are minimized under vacuum.
The increase in char production when increasing the initial bed depth and the
reactor configuration reflects an increase in the effect of heterogeneous extra-parti-
cle secondary reactions. According to Boronson et al. [37], the char yield increases
sharply at first and then approaches an apparent asymptote upon increasing of the
bed depth. The authors observed the formation of extra-charcoal up to 14 wt.%
(biomass anhydrous basis). In the present work the extra-charcoal formed in the
pilot plant was 6.2 wt.% (Table 3) which indicates that bed height alone is not
enough to explain the reaction mechanisms. The overall reactor configuration, the
vapour outlet localization inside the chamber and the exact position of the heat
transfer sources also play a significant role. The scale effect under various vacuum
pyrolysis conditions for softwood bark has been reported earlier by Yang et al. [38]
and Pakdel et al. [39]. Yields obtained in a continuous feed pilot plant reactor are
more reliable than data obtained batchwise [28].

3.2. Product characterization

3.2.1. Gas
The gas composition, heating value and average molecular weight are shown in
Table 4. The gas produced is mainly composed of CO2, CO, CH4, H2 and C3H6.
These compounds represent between 87.5 and 92.4% by vol. of the total non-con-
densable product. The high CO2 content is due to the high amount of oxygen in the
feedstock (43.9 wt.%). The resulting process gas has a low heating value. The gas
composition was found to be similar at the laboratory and pilot plant scales and
corresponds very well to the values reported for other biomasses subjected to
similar pyrolysis conditions [31,33].
120 M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136

Table 4
Range of gas composition (% vol.) (runs c A-238-241, H-57)

Gas Range

H2 1.12.4
CH4 2.26.7
CO 28.631.8
CO2 48.256.0
C 2H 4 0.30.6
C 2H 6 0.60.9
C 3H 6 1.11.6
C 3H 8 0.50.8
CH3OH 0.81.3
C4H10 0.40.9
C5H12 0.10.6
Others 4.28.0
Average molecular weight (g mol1) 38.138.5
Gross calorific value (MJ kg1) 6.49.8

3.2.2. Bio-oil
The amount of pyrolytic oil collected at the laboratory scale was not large
enough to enable a complete characterization. This is the reason why the bio-oil
properties presented are related only to the oil obtained in the pilot plant equip-
ment (run c H-57).

3.2.2.1. Physico-chemical characterization. To be used as a liquid fuel, bagasse-

derived bio-oil must meet several specifications in terms of viscosity, water content,
Conradson carbon residue (CCR), methanol insoluble materials (MIM) content
and calorific value. Some of these properties are shown in Table 5.
The sugarcane bagasse bio-oil obtained has a low moisture content (13.8 wt.%).
The viscosity observed (26.7 cSt at 40 C) is relatively low compared to values
reported for other bio-oils (35 53 cSt at 40 C) [40,41]. For gas turbine applica-
tions, the viscosity should be around 3 4 cSt at 90 C [40]. Bagasse-derived bio-oil
meets this specification (4.1 cSt at 90 C). The bio-oil density at 20 C was
approximately 1.2 times greater than distilled water. The oil density decreased
almost linearly with increasing temperature (thermal expansion coefficient =0.8
10 3 g ml 1 C). This value is very similar to the one reported for decanted oil
from bark (0.8210 3 g ml 1 C) [40,41]. Bagasse bio-oil is an acid oil with a pH
of 2.7. The acidity is due to the presence of low molecular weight carboxylic acids.
The bio-oil gross calorific value was found to be 22.4 MJ kg 1 (on a bio-oil
as-received basis), and is lower than that of petroleum fuels used typically in gas
turbines (approximately 40 MJ kg 1). Thus, 1.88 kg or 1.57 l of sugarcane bagasse
oil are required to provide the same energy density as 1 kg of petroleum fuel. The
MIM of the bio-oil was 0.38 wt.%. It lies in the lower range of bio-oil solid contents
reported in the literature (Table 5).
 M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136 121

The presence of ash in the bio-oil can cause erosion, corrosion and incrustation
problems. Ash content is problematic for gas turbine applications and the limit is
0.1 wt.% [40]. The main source of ash in pyrolysis oils is the solid char particles carried
over by the pyrolysis vapours. Elliott [21] reported that filtering the bio-oil, after
dilution with methanol through a 2.5 mm filter, removed only 2050% of the inorganic
content of the oil. This implies that the remaining inorganic elements are suspended
char particles less than 2.5 mm in diameter, or they are in ionic form in the bio-oil.
The ash content of the bio-oil was about 0.05 wt.%. It meets the specifications of
gas turbine fuels. The metal content of the charcoal, bio-oil and pyrolytic water is
presented in Table 6.
The metal upper limits for a gas turbine fuel are 0.51 ppm sodium+ potassium,
0.5 ppm calcium and 0.5 ppm vanadium [40 42]. The bagasse-derived bio-oil
vanadium content was below the maximum allowed for gas turbine fuels (0.02 ppm).
However, the alkali metals and calcium content was 21.5 ppm Na, 5.3 ppm K, 33.8

Table 5
Bagasse bio-oil physico-chemical properties

Properties Units Bio-oil Other bio-oils

(run c H-57) [14,19,25,29,45]

Moisture wt.% 13.8 10.235

Density at 20 C kg m3 1211 12081238
at 40 C 1195 11901211
at 60C 1180 11711197
at 80 C 1160
Kinematic viscosity at 20 C cSt 116.5 50672
at 40 C 26.7 3553
at 50 C 16.4 9137
at 60 C 11.2
at 70 C 8.2
at 80 C 5.4 11
at 90 C 4.1
Flash point C \90 64 to \106
Gross calorific value MJ kg1 22.4 1524.3
(as-received basis)
Elemental composition wt.%
(anhydrous basis)
C 54.60 48.063.5
H 6.45 5.27.2
N 0.73 0.070.39
S B0.10 0.000.05
O (by difference) 38.07 3246
Ash 0.05 0.030.3
Solid content (as wt.% 0.38 0.171.14
methanol insoluble
material)
Conradson carbon wt.% 18.6 1823
residue
Acidity (g NaOH/100 g oil) 8.2
pH 2.7 2.03.8
122 M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136

Table 6
Metal content (mg kg1) of the charcoal, bio-oil and aqueous phase (run c H-57)

Metals Charcoal Bio-oil Aqueous phase

Na 448.8 21.5 1.2

Ni 2.6 1.1 B0.01
P 1324.3 3.3 0.2
Pb 2.1 4.9 0.03
Sb 10.8 7.7 B0.05
Se 6.5 0.2 0.9
Ti 12.5 0.2 B0.005
U 12.9 0.08 0.03
V 0.4 0.02 B0,002
Zn 27.6 43.7 0.04
Al 281.4 10.1 0.2
Ca 5384.4 33.8 1.7
Co 1.5 0.08 B0.003
Cr 2.4 0.6 B0.003
Cu 15.3 2.3 B0.01
Fe 950.1 105.8 0.2
K 5536.7 5.3 0.07
Mg 2257.1 2.2 0.09
Mn 73.5 1.1 0.009

ppm Ca. These values are in agreement with the range of values reported in the
literature for other bio-oils (8 83 ppm Na, 4 440 ppm K, 15 540 ppm Ca) [37,38].
Here, the gas turbine usual fuel standards are not met.
Fig. 2 shows bio-oil microphotographs using plane and cross polarized light. It is
obvious that the bio-oil is not an homogeneous system. Some colloidal associates
formed by large molecules are not fully solubilized at room temperature (perhaps
pyrolytic lignin compounds or some methanol insoluble materials). Some of these
insoluble materials can be observed clearly using cross-polarized light. According to
Radlein [43] the bio-oil must be described as a heterogeneous system. The author
suggested that the pyrolytic lignin is suspended in a micellar or microemulsion phase
by the monomeric phenols, pyrans and furans acting as bridging agents between the
high molecular weight pyrolytic lignin and the continuous and more polar phase
formed by the water and its soluble components.

3.2.2.2. Distillation. Distillation tests in the temperature range of 25140 C under

atmospheric and under vacuum (15 kPa abs.) conditions were performed to determine
the content of the bio-oil light fraction. The distillates were collected and their water
content was determined (Fig. 3). The first drop of distillate corresponding to the initial
boiling point (IBP) appeared at a lower temperature under vacuum (60 C) than at
atmospheric pressure (90 C). However, the amount collected from IBP to 140 C
was approximately the same (30 wt.%) in both distillations. The small differences
observed in the yield of organics and water at 140 C are mainly due to experimental
limitations associated with the condensation of vapors under vacuum.
Fig. 2. Microphotograph of bio-oil obtained from bagasse at laboratory scale: (a) plane polarized light; and (b) cross polarized light.
M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136
123
124 M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136

Table 7
Successive oil solubility in different solvents-run c H-57 (wt.% total bio-oil basis)

Solvent Pentane Benzene CH2Cl2 Ethyl acetate MeOH

wt.% 2.7 15.0 23.6 37.5 21.2

The distillate contained considerable amount of water, thus invalidating its use as
fuel. The distillation residue became practically insoluble in methanol and ethyl
acetate, indicating an appreciable polymerization during the distillation process. The
bio-oil thermal instability limits the recourse to upgrading alternatives involving a
high temperature environment.
3.2.2.3. Separation of bio-oil fractions and GC MS analysis. Bio-oil was extracted
successively using various solvents with different polarities. The percentage corre-
sponding to each soluble fraction is reported in Table 7.
As can be observed, the bio-oil is less soluble in low molecular weight aliphatic
hydrocarbons (2.7 wt.%) than in aromatic hydrocarbons such as benzene (15 wt.%).
This difference is due to the high polarity of the bio-oil main compounds. The
recovered fractions were further analyzed by GCMS (Fig. 4). The list of the most
abundant compounds in the oil fractions is presented in Table 8.
3.2.2.4. Fatty acids, organic acids and sugars. The content of fatty acids in the oil is
very low (less than 0.01 wt.%) which is a typical characteristic of hardwood-derived
pyrolysis oils. The main fatty acids were identified and quantified in Table 9. The
low molecular weight carboxylic acids are abundant in the condensed phase (4.1 wt.%
of formic acid, 9.2 wt.% of acetic acid, 0.8 wt.% of propanoic acids, 0.4 wt.% of
butanoic acid). These values are in agreement with those reported by Sipila [44] and
Pakdel [32] for other pyrolytic oils. The monosaccharide sugar content was approx-
imately 1.6 wt.% of the oil (0.5 wt.% on a bagasse anhydrous basis). The most common
monosaccharides found were: levoglucosane (77 wt.%), xylose (13 wt.%) and
arabinose (9 wt.%). The total sugar content determined after bio-oil hydrolysis was
approximately 3.1 wt.% on a bagasse anhydrous basis, or approximately 10.3 wt.%
on an oil basis. The major monomeric sugars found after oil hydrolysis were: glucose
(68.9 wt.%), xylose (15.3 wt.%), arabinose (6.5 wt.%) and ribose (5.9 wt.%).
3.2.2.5. Stability test. Bio-oil is not a thermodynamically stable product during the
pyrolysis process. It is produced over a short lapse of time (at high temperature)
followed by a rapid cooling operation. In the cooling train, the vapor compounds
are condensed and many undergo to a certain extent chemical reactions even after
having been condensed. This phenomenon is observed during the bio-oil storage and
can be critical during bio-oil pre-heating for combustion applications. Accelerated
oil stability tests were performed in order to identify the bio-oil thermal susceptibility.
Viscosity measurements of bio-oil samples after storage at 80 C are shown in Fig.
5. Experimental results indicate that the viscosity continuously increased with storage
time. Over 1 week (168 h) at 80 C, the bio-oil viscosity measured at 40 C did
increase by a factor of 2.3. The ageing process was more evident when viscosity
measurements were conducted at low temperature.
Fig. 3. Bio-oil distillation curves: (a) atmospheric pressure; and (b) vacuum (15 kPa abs.).
M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136
125
126 M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136

Fig. 4. GC MS chromatograms of the aqueous phase and oil fractions.

M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136 127

Fig. 4. Continued.
Table 8

128
Sugarcane bagasse pyrolytic oil chemical composition (run c H-57)

No. Compounds Aqueous C5H12 Benzene CH2Cl2 Ethyl acetate MeOH

phase soluble soluble soluble soluble soluble

1 Acetic acid X n.a. n.a n.a n.a n.a

2 1-Hydroxy-2-propanone X n.a n.a n.a n.a n.a
3 Propanoic acid X n.a n.a n.a n.a n.a
4 1-Hydroxy-2-butanone X n.a n.a n.a n.a n.a
5 Cyclopentanone X n.a n.a n.a n.a n.a
6 Methoxyethene X. X X X
7 Butyl acetate X
8 2-Furancarboxaldehyde X X X X
9 N,N,2,2-Tetramethyl-N-oxide-propylamine X
10 1-Acetyloxy 2-propanone X X X
11 1,2-Ethanediol X
12 Methylcyclopenten-1-one X X
13 Ethanone-1-(2-furanyl) X X
14 2-Furanone X X X X
15 2-Butenal X
16 2-Hydroxycyclopent-2-en-1-one X X X X X X
17 3-Methylene-2, 5-furandionedihydro X X
18 5-Methyl-2-furancarboxaldehyde X X X
19 (2-Methyl-1-propenyl)hydrazine X X
20 3-Methyl-2-furanone X
21 Phenol X X X X
22 3-Methyl-2,4-imidazolidinedione X X X X
23 3-Methyl-1,2-cyclopentanedione X X X X
M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136

24 1,3,5-Trioxane X
25 2-Methylphenol X X X
26 4-Methylphenol X X X
27 Methylester-3-furancarboxilic acid X X
28 2-Methoxyphenol X X X
29 Cyclopropylcarbinol X X X X
Table 8 (Continued)

No. Compounds Aqueous C5H12 Benzene CH2Cl2 Ethyl acetate MeOH

phase soluble soluble soluble soluble soluble

30 Maltol X
31 3-Ethyl-2-hydroxy-2-cyclopenten-1-one X X
32 2,4-Dimethylphenol X
33 O-Methylisoureahydrogensulfate X
34 4-Ethylphenol X X
35 2-Methoxy-4-methylphenol X X
36 1,2-Benzenediol X X X
37 1,4: 3,6-Dianhydro-a-D-glucopyranose X
38 2,3-Dihydrobenzofuran X X X X X
39 5-Hydroxymethyl-2-furancarboxaldehyde X X
40 2,3-Dimethoxytoluene X
41 3-Methyl-1,2-benzenediol X X
42 4-Ethyl-2-methoxyphenol X
43 3,4-Diethylphenol X
44 2,6-Dimethoxyphenol X X X
45 Eugenol X
46 4-Hydroxybenzaldehyde X X
47 Vanillin X X
48 2,5-Dimethylbenzo(b)thiophene X
49 1,2,4-Trimethoxybenzene X X
50 2-Methoxy-4-(1-propenyl)phenol X
51 Ethanone-1-(4-hydroxy-3-methoxyphenol) X
52 1,6-Anhydro-b-glucopyranose X X
53 Hydroquinone X X
54 6,7-Dihydro-3,6-dimethyl-4-benzofuranone X X
55 2,6-Dimethoxy-4-(2-propenyl)phenol X X
M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136

56 4-Hydroxy-3,5-dimethoxybenzaldehyde X X
57 1,2-Dimethoxy-4-(2-propenyl)benzene X X
58 3-(4-Hydroxyphenyl)-2-propenoic acid X
59 1-(2,4,6-Trihydroxyphenyl)-2-pentanone X
60 Hexadecanoic acid X
61 1,2,4-Trimethoxy-5-(1-propenyl)benzene X X
129

62 9,12-Octadecadienoic acid X

n.a: not analyzed.

130 M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136

Table 9
Main fatty acids present in the bio-oil (run c H57)

Fatty acids wt.%

Hexadecanoic acid 45.4

Octadecanoic acid 10.8
8-Octadecenoic acid 6.0
Eicosanoic acid 5.2
Tetracosanoic acid 4.5
Octacosanoic acid 3.1
Tricosanoic acid 3.0
Pentanedioic acid 2.9
Docosanoic acid 2.1
Pentadecanoic acid 1.9
Octanoic acid 1.7
Benzenepropanoic acid, 3,4 dimethoxy 1.7
Hexacosanoic acid 1.3
Dodecosanoic acid 1.2
Others 9.2
Total 100

Viscous flow activation energy was calculated using viscosity values obtained at
different temperatures and under different ageing times. Viscous flow energy is a
characteristic of the fluid. It indicates the amount of energy required for a fluid to
flow. It is determined by the slope of the curve obtained on a semi-logarithmic
graph of the dynamic viscosity versus the inverse of viscosity measurement temper-
ature (Fig. 6). The equation used for the calculation is:
v = Ave (Ew/RT) (1)
where, v, dynamic viscosity (mPa s), Av, intercept with Y-axis (equivalent to the
pre-exponential factor) (mPa s), T, viscosity measurement temperature (K), Ev,
viscous flow activation energy (J mol 1), and R= 8.314 J mol 1 K 1.
The value of the flow activation energy obtained for different ageing times is
almost constant at 37.8 kJ mol 1. The intercept with the Y-axis (equivalent to the
pre-exponential factor) increases with ageing time (Fig. 7). This intercept point
represents the lowest viscosity which could be obtained upon increasing the
temperature, supposing that the chemical nature of the oil was maintained [45].
Viscosity increments are due to the formation of new high molecular weight
compounds resulting from polymerization reactions. In complex mixtures, average
molecular weight indicators such as Mw and Mn can be used to describe the
molecular weight distribution (Table 10) [34]. Mw is the average molecular weight
of a (fictive) homogeneous compound which, with regard to light scattering or
sedimentation, would have the same properties as the real mixture. This average
molecular weight is calculated on basis of the weight concentration of each fraction.
The number average molecular weight (Mn) could be obtained applying colligative
methods and is an average of molecular weight calculated on the basis of molar
concentrations. The polydispersity index (Pd=Mw/Mw) is thus a measure for the
non-homogeneity of the mixture [46].
 M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136 131

Fig. 5. Bio-oil viscosity during ageing process at different measurement temperatures.

These parameters can be calculated automatically by the GPC software MILLEN-

NIUM version 3.05 from Waters. The change in the concentration of several bio-oil
fractions of different molecular weights during ageing is shown in Fig. 8. As can be
observed, the variation of the concentration of low molecular weight compounds
(100200 u.m.a.) was found to be high indicating that those are the most reactive
compounds. Furthermore, large amounts of high molecular weight compounds
were also obtained as the ageing process progressed. It is important to note from
Fig. 7 and Table 10 that a correlation between the pre-exponential factor (Av) and
the average molecular weight (Mw) can be drawn. The pre-exponential factor
increases with the increase of the average molecular weight (Mw) following the
equation:
Av =4 10 8Mw 5 10 6 (2)

Fig. 6. Bio-oil viscosity behaviour at different ageing times.

132 M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136

Fig. 7. Viscous pre-exponential factor at different ageing times.

The ageing process is mainly due to the presence of a large number of oxygenated
organic compounds which are very reactive. High temperature, presence of charcoal
particles as well as the acidity of the medium accelerate bio-oil ageing. During
ageing, etherification and/or esterification reactions occur between hydroxyl, car-
bonyl and carboxyl group components [34,47 49]. The presence of condensation
reactions during bio-oil ageing is confirmed by the increase of the water content in
the bio-oil.
The MIM content of the aged bagasse bio-oil is similar to that of the fresh bio-oil
(Table 10). This may be due to the solubility increase of fatty acids and other
methanol insoluble matter in the medium during storage. However, increase in solid
content during ageing of other kinds of bio-oils has been reported [34,41].

3.2.3. Aqueous phase

The aqueous phase contains 16 wt.% organic compounds. These compounds are
responsible for the low pH of 2 and high acidity (1.37 g NaOH/100 ml at pH 7).
A chromatogram of the aqueous phase is presented in Fig. 4. This phase is very rich
in acids and low molecular weight compounds which are very reactive. Boucher
[40,41] proved that the addition of portions of the pyrolytic aqueous phase to the

Table 10
Bio-oil ageing tests (run c H57)

Heating time Solid content (% wt/wt) Water (wt.%) Mn (a.m.u.) Mw (a.m.u.) Pd = Mw/Mn
(h)

0 0.38 13.84 251 434 1.73

12 0.28 15.05 322 587 1.82
24 0.28 15.18 338 624 1.84
48 0.18 15.86 383 702 1.83
168 0.14 16.00 414 839 2.03
 M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136 133

Fig. 8. Molecular weight distribution vs. ageing time: (a) low molecular weight compounds; and (b) high
molecular weight compounds.

bio-oil increases its rate of ageing and consequently reduces its stability. The
presence of acids in the aqueous phase accelerates ageing of the oil. It is well known
that polymerization and polycondensation reactions are favoured by the acidity of
the medium. The acid effect on the polymerization reactions is remarkable in spite
of the relatively low dissociation factor of the formic and acetic acid compared to
the dissociation factor of mineral acids.

3.2.4. Charcoal
Table 11 presents the proximate and elemental analysis of the charcoal obtained
at the laboratory and pilot plant scales. The charcoal average volatile matter (V.M.)
content of 17.2 wt.% indicates that a good bagasse conversion has been reached.
The results obtained correspond very well to the ones reported by Zandersons et al.
134 M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136

Table 11
Charcoal proximate and elemental analyses (wt.%)

Laboratory scale (run c A-238) Pilot plant scale (run c H-57)

Proximate analysis
Volatile matter 18.9 15.4
Fixed carbon 74.4 79.1
Ash 6.7 5.5
Elemental analysis
Carbon 85.6 81.5
Hydrogen 2.9 3.1
Nitrogen 1.3 0.8
Sulphur B0.1 B0.1
O+ash(by difference) 10.2 14.6

[2]. The charcoal from the pilot plant was produced at a temperature 30 C above
the laboratory test in the retort. Hence the lower residual volatile matter content for
the charcoal obtained in the pilot unit.
The metals identified in the charcoal are also present in the bio-oil. The content
of metals in the charcoal produced in the pilot reactor is presented in Table 6. The
metal content of the charcoal is considerably larger than that of the bio-oil. The
metals detected in the oil find their origin in the selective leaching of some metals
present in the charcoal. The charcoal obtained in the laboratory reactor has a
heating value of 36 MJ kg 1 and exhibits a high surface area (529 m2 g 1) of
microporous nature. After steam activation, the activated carbon obtained had a
surface area of 1947 m2 g 1 (Dubinin-Radushkevich (DR) method) with an iodine
number of 1140 mg g 1. The surface characterization of the bagasse charcoal and
the steam activation procedure have been reported by Darmstadt et al. [50].

4. Conclusions

Between 30 and 34 wt.% of bio-oil (bagasse anhydrous basis) can be recovered by

vacuum pyrolysis of sugarcane bagasse. The bio-oil potentially represents a valu-
able liquid fuel for boilers. For gas turbine applications, Na, Ca and K content of
the oil must be lowered. Bagasse finest particles, which exhibit a very high ash
content, must be removed from the reactor feedstock in order to reduce the MIM
content of the pyrolysis oil obtained. The chemical composition suggests that the
bio-oil is a challenging matrix for chemical, nutritional and pharmaceutical pur-
poses. A series of ageing tests indicates that the bagasse-derived bio-oil has a
normal bio-oil thermal stability. The yield of charcoal is comprised between 20 and
26 wt.% (bagasse anhydrous basis). This solid fuel could be used in boilers where
bagasse is presently burned. It also represents a good feedstock for the production
of activated carbon.
 M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136 135

Acknowledgements
The authors are thankful to the United States Sugar Corporation (Clewiston,
Florida) for having gratefully provided the sugarcane bagasse. The authors are very
thankful to Dr Hooshang Pakdel for some of the analyses (GC MS and GPC).
The authors are also grateful to Dr Annette Schwerdtfeger for her technical review
of the paper and to Mrs Micheline Gingras and Mrs Joanne Lagace who have
conducted the laboratory tests.

References
[1] A.R. Drummond, I.W. Drummond, Ind. Eng. Chem. Res. 35 (1996) 1263.
[2] J. Zandersons, J. Gravitis, A. Kokorevics, A. Zhurinsh, M. Bikovens, A. Tardenaka, B. Spince,
Biomass Bioenergy 17 (1999) 209.
[3] I. Kilicaslan, H.I Sarac, E. Ozdemir, K. Ermis, Energy Convers. Manage. 40 (1999) 1.
[4] P.A. Beaton, A.L. Brito, J. Ballester, C. Dopazo, Retrofitting and optimization of bagasse boilers,
in: R.P. Overend, E. Chornet (Eds.), Proceedings of the Fourth Biomass Conference of the Americas,
Oakland, August 29 September 2, Elsevier Science Ltd., UK, 1999.
[5] A.L. Brito, P.A. Beaton, J. Ballester, C. Dopazo, Novel approach for the analysis of heat transfer
in bagasse-fired furnaces, in: R.P. Overend, E. Chornet (Eds.), Proceedings of the Fourth Biomass
Conference of the Americas, Oakland, August 29 September 2, Elsevier Science Ltd., UK, 1999.
[6] F. Ponce, A. Walter, Potential assessment of BIG-GT technology in Cuban sugarcane mills, in: R.P.
Overend, E. Chornet (Eds.), Proceedings of the Fourth Bio-mass Conference of the Americas,
Oakland, August 29 September 2, Elsevier Science Ltd., UK, 1999.
[7] M.S. Scott, G.E. Staats, S.N. Rao, R. Goldman, D. Hess, Fuel Process. Techol. 54 (1998) 227.
[8] R. Tamaro, C. Echeverr`a, Use of combustion turbines to produce electricity from sugarcane bagasse,
in: R.P. Overend, E. Chornet (Eds.), Proceedings of the Fourth Biomass Conference of the Americas,
Oakland, August 29 September 2, Elsevier Science Ltd., UK, 1999.
[9] M.L. Souza-Santos, Fuel 78 (1999) 529.
[10] H. Pakdel, C. Roy, A. Chaala, Production of useful products by vacuum pyrolysis of biomass, Proc.
XV Chemical Conference of the Oriente University, Santiago de Cuba, November 27 29, Editorial
Oriente, 1996.
[11] F. Mobarak, Y. Fahmy, Wood Sci. Technol. 16 (1982) 59.
[12] L.E. Brossard, L.A.B. Cortez, Biomass Bioenergy 12 (5) (1997) 363.
[13] E. Esperanza, J. Arauzo, G. Gea, Fluidised bed air gasification of sugarcane bagasse. Bed temperature
and air factor influence, in: R.P. Overend, E. Chornet (Eds.), Proc. of the Fourth Biomass Conference
of the Americas, Oakland, August 29 September 2, Elsevier Science Ltd., UK, 1999.
[14] D.L. Larson, E. Olivares, Energy losses due to elutration in fluidised bed gasification of sugarcane
bagasse, in: R.P. Overend, E. Chornet (Eds.), Proc. of the Fourth Biomass Conference of the
Americas, Oakland, August 29 September 2, Elsevier Science Ltd., UK, 1999.
[15] C. Roy, J. Yang, D. Blanchette, L. Korving, B. De Caumia, Development of a novel vacuum pyrolysis
reactor with improved heat transfer potential, in: A.V. Bridgwater, D.G.B. Boocock (Eds.), Proc.
of Developments in Thermochemical Biomass Conversion, London, October 15 19, 1996, Blackie
Academic and Professional, UK, 1997.
[16] C. Roy, D. Blanchette, B. De Caumia, Horizontal moving and stirred bed reactor. Canadian patent
claim number 2 196 841 January 30, 1998. US Patent number 8,811,172, March 1997. International
application number PCT/IB98/00224.
[17] J. Xia, K. Noda, S. Kagawa, N. Wakao, J. Chem. Eng. Jap. 6 (1998) 987.
[18] W. Van der Kamp, J. Smart, Evaluation of the combustion characteristics of pyrolytic oils derived
from biomass, in: G. Grassi, A. Collina, H. Zibetta (Eds.), Proc. of Biomass and Energy, Industry
and the Environment, 6th E.C. Conference, Athens, Greece, April 22 26, 1991, Elsevier Applied
Science, UK, 1992.
136 M. Garc`a-Pe`rez et al. / J. Anal. Appl. Pyrolysis 65 (2002) 111136

[19] R.G. Graham, A. Barry, D.R. Huffman, Application of rapid thermal processing of biomass, in: A.V.
Bridgwater (Ed.), Proc. of Advances in Thermochemical Biomass Conversion, Interlaken, Switzer-
land, May 11 15, 1992, Blackie academic and professional, UK, 1994.
[20] S. Gust, O. Neste, Flash pyrolysis fuel oil, Proc. of the Symposium: Biomass Pyrolysis Oils-Properties
and Combustion Meeting. Estes Park, Colorado, September 26 28, 1994.
[21] D.C. Elliot, Biomass Bioenergy 7 (1 6) (1994) 179.
[22] R. Andrews, P.C. Patnaik, Q. Liu, Firing fast pyrolysis oil in turbine, Proc. of the Symposium: Biomass
Pyrolysis Oils-properties and combustion meeting, Ester Park, Colorado, September 26 28, 1994.
[23] M.G. Rasul, V. Rudolph, M. Carsky, Fuel 78 (1999) 905.
[24] G.A. Roca, J.M. Boizan, Int. Sugar J. 95 (1993) 177.
[25] M. Garc`a-Pe`rez, A. Chaala, C. Roy, Co-Pyrolysis of Sugarcane Bagasse with Petroleum Residue.
Part I: Thermogravimetric Analysis, Fuel 80 (2001) 1245.
[26] H. Pakdel, G. Couture, C. Roy, Tappi J. 77 (7) (1994) 205.
[27] C. Roy, D. Morin, F. Dube , The Biomass Pyrocycling process, in: M. Kaltschmitt, A.V. Bridgwater
(Eds.), Proc. of Biomass Gasification and Pyrolysis, Stuttgard, April 9 11, 1997, CPL press, UK,
1997.
[28] C. Roy, D. Blanchette, B. de Caumia, F. Dube , J. Pinault, E. Be langer, Industrial scale demonstration
of the Pyrocycling process for the conversion of biomass to bio-fuel and chemicals, Proc. of the
First World Conference and Exhibition on Biomass for Energy and Industry, Sevilla, Spain, June
5 9, 2000, p. 1032.
[29] H. Pakdel, C. Roy, W. Kalkreuth, Fuel 78 (1999) 365.
[30] A. Oasmaa, E. Leppamaki, P. Koponen, J. Levander, E. Tapola, Physical Characterisation of
Biomass-based Pyrolysis Liquids: Application of Standard Fuel Oil Analysis, VVT Energy Publica-
tion, 1997, p. 306.
[31] H. Pakdel, H.G. Zhang, C. Roy, Bioresour. Technol. 47 (1994) 45.
[32] H. Pakdel, C. Roy, Biomass 13 (1987) 155.
[33] C. Roy, H. Pakdel, D. Brouillard, J. Appl. Polym. Sci. 41 (1990) 337.
[34] S. Czernik, D.K. Johnson, S. Black, Biomass Bioenergy 7 (1 6) (1994) 187.
[35] L. Fagernas, Chemical and physical characterisation of biomass-based pyrolysis oils-literature review,
VTT Research notes, Finland, 1995, p. 1706.
[36] C. Roy, B. de Caumia, P. Plante, H. Me nard, Production of liquids from biomass by vacuum
pyrolysis Development of data base for continuous Process, Proc. of Energy from Biomass and
Wastes VII, January 24 28, Lake Buena Vista, Florida, 1983, pp. 1147-1170.
[37] M.L. Boroson, J.B. Haward, J.P. Longwell, W.A. Peters, Energy Fuels 3 (1989) 735.
[38] J.D. Yang, D. Blanchette, B. de Caumia, C Roy, Modelling, scale up and demonstration of vacuum
pyrolysis reactor, Proc. of Progress in Thermochemical Biomass Conversion, Tyrol, Austria, 17 22
September, 2000.
[39] H. Pakdel, C. Amen-Chen, C. Roy, Can. J. Chem. Eng. 75 (1997) 121.
[40] M.E. Boucher, A. Chaala, C. Roy, Biomass Bioenergy 19 (5) (2000) 337.
[41] M.E. Boucher, A. Chaala, H. Pakdel, C. Roy, Biomass Bioenergy 19 (5) (2000) 351.
[42] J. Odgers, D. Kretschmer, Gas turbine fuels and their influence on combustion, Energy and
Engineering Sciences Series, Abacus Press, Massachusetts, 1986, p. 185.
[43] D. Radlein, in: A. Bridgwater, et al. (Eds.), Fast Pyrolysis of Biomass. A Handbook, CPL press, UK,
1999, p. 164 (Chap. 13).
[44] K. Sipila, E. Kuoppala, L. Fagernas, A. Oasmaa, Biomass Bioenergy 14 (2) (1998) 103.
[45] J. Leblanc, Rhe ologie des e lastome res et mise en uvre des polyme`re, E ditions ARTEL, Paris, 1996,
p. 1999.
[46] H. Determann, Gel Chromatography. Gel Filtration, Gel Permeation, Molecular Sieves. A Labora-
tory Handbook, Springer, New York, 1968, p. 120.
[47] J.P. Diebold, S. Czernik, Energy Fuels 11 (1997) 1081.
[48] R.K. Sharma, N.N. Bakhshi, Upgrading of biomass-derived pyrolytic oils over HZSM-5 catalyst,
Report of Contract File No. 058SZ-23283-8-6116. Bio-energy Development Program, Energy, Mines
and Resources, Canada, 1989, p. 79.
[49] J.D. Adjaye, R.K. Sharma, K. Ramesh, N. Bakhshi, N. Narenda, Fuel Process. Technol. 31 (1992)
241.
[50] H. Darmstadt, M. Garcia-Pe rez, A. Chaala, C. Roy, Carbon 39 (2001) 815.