J Energy 2021 120877
J Energy 2021 120877
Energy
journal homepage: www.elsevier.com/locate/energy
a r t i c l e i n f o a b s t r a c t
Article history: Empty fruit bunches (EFB), after fruits-bunches separation, are highly potential as an energy source.
Received 27 December 2020 However, due to some inherited inferior fuel properties such as bulk density, hydrophobicity, moisture
Received in revised form content and calorific value (CV), pretreatment is required before EFB can be utilized as a solid fuel.
30 April 2021
Torrefaction is promising to upgrade EFB fuel properties, and thorough products characterization is
Accepted 4 May 2021
Available online 7 May 2021
essential to reveal process mechanism and potential application. In this study, the effects of temperature
(225e300 C) and retention time (20e60 min) on EFB torrefaction were investigated in a fixed bed
reactor. While an increased temperature and retention time had led to a lower energy and mass of solid
Keywords:
Torrefaction
yields, the liquid and gaseous yields showed otherwise. The oppositely reduced O (37%) and increased C
Oil palm biomass (45%) contents of torrefied EFB with increasing temperature, and the resulting decrease in O/C ratio, and
Solid fuel H and O contents, had greatly improved the CV of the torrefied solid up to 46% compared to its raw form.
Liquid fuel In addition, the chlorine content was significantly reduced (21e54%) after torrefaction. Both the raw and
Gaseous fuel torrefied solid EFB showed a strong thermal degradation of hemicellulose with lower decomposition rate
Deoxygenation of the latter due to mass depletion during torrefaction. Oxygen removal from EFB during torrefaction
migrated to the liquid and gaseous products, most dominant as water and CO2, followed by CO and
oxygen-containing organic compounds. The liquid product consisted of condensable components; water
(main) and functionalised compounds, among which were acids (as relatively higher concentration),
furans, ketones, alcohols and phenols. The gaseous product was dominated by CO2 and CO in all torre-
faction conditions. The findings contribute partly as biomass resource database for future bioenergy
upscaling and deployment.
© 2021 Elsevier Ltd. All rights reserved.
https://doi.org/10.1016/j.energy.2021.120877
0360-5442/© 2021 Elsevier Ltd. All rights reserved.
M.A. Sukiran, W.M.A. Wan Daud, F. Abnisa et al. Energy 230 (2021) 120877
these, converting EFB into fuel as energy resource is the most time and reactor heating rate on torrefaction of EFB in a laboratory-
popular research area that has received a lot of attention [10]. scale standard retort. The optimum trade-offs for higher torrefied
The use of EFB as a fuel, however, is limited mainly due to high EFB yields but lower heating values occurred at heating rate of 10 K/
moisture and oxygen contents, low energy density, hygroscopic min with retention time of 30 min and temperature of 272 C. In
nature and low calorific value (CV). These inferior properties have microwave-assisted torrefaction of EFB with waste oil (used engine
caused difficulty for direct fuel application and subsequent low oil and cooking oil), Lam et al. [35] found that the co-torrefied
utilization and combustion efficiency [3]. A pretreatment is thus product has a much improved CV of up to 28 MJ/kg from its raw
needed as means to improve its fuel properties, and the most pellet form of 17.9 MJ/kg. Besides, Nyakuma et al. [36] investigated
appropriate and widely accepted option by far is through thermo- the non-oxidative torrefaction of EFB pellets at different tempera-
chemical process i.e. torrefaction. Torrefaction operated at tem- ture in a horizontal fixed bed tubular reactor. The torrefied EFB
perature ranging from 200 C to 300 C in the absence of oxygen or pellets were highly grindable, hydrophobic and thermally stable,
air [11,12] has ability to decompose most of the hemicelluloses in a and also high in CV (22.83e25.81 MJ/kg) and carbon but low in
biomass into light volatile condensable products of water, acids and oxygen contents.
lightweight chemicals [13]. Such release of volatiles is able to These previous studies give remarkable insights into under-
reduce moisture and oxygen contents and lead to an improved CV standing the torrefaction characteristics of EFB under different
[14]. Torrefaction is influenced by many operating parameters, conditions. Most studies focus on solid products in terms of yields
including temperature, retention time, particle size, type of and properties as a potential solid fuel. However, it is noteworthy
biomass and moisture content [15]. As agreed by most of the re- that the torrefaction of EFB contains not only solid main product
searchers, temperature has the greatest influence on torrefaction but also other co-products i.e. liquid and gases. To the best of the
process compared to other parameters [16,17]. authors’ knowledge, products distribution of EFB torrefaction and
A lot of studies on biomass torrefaction have been published in characterization of the resulting liquid and gas products were not
the past 15 years, describing the process using various feedstock investigated in details yet. The quantitative and qualitative data
e.g. sawdust [18], rice husk [19,20], pine wood [21], bamboo [22,23] from analysis of these co-products together with the well-
and sorghum [11]. Most of these studies focused on understanding established solid product play a crucial role in examining mass
the conversion process via manipulating certain parameters for and energy flows of EFB torrefaction. According to Yue et al. [11],
better product yield. In examining torrefaction of rice husk, Chen the liquid and gas products from torrefaction are derived from both
et al. [20] found a damaged and disrupted fiber structure with volatile components in hemicellulose and cellulose and partial
increasing temperature leading to lower atomic oxygen ratio, and degradation of lignin. The liquid product consists mainly of water
higher atomic carbon ratio and energy density. Similarly, Li et al. and functionalised compounds such as acids, furans, ketones, al-
[23] showed that temperature and retention time could affect the dehydes and phenols [19]. Meanwhile, the gas product is made up
structure and properties of the solid product of bamboo. The of mainly CO2 and CO with trace amounts of CH4 and H2 [28]. We
resulting mass and energy yields gradually decreased with an in- also found information gap in the literature on thermal behaviour
crease of the investigated parameters. In 2010, a laboratory-scale of torrefaction product of EFB. As agreed by most of the researchers,
horizontal tubular reactor was first deployed and operated by understanding thermal behaviour of torrefied biomass is an
Uemura et al. [24] to treat EFB at different temperatures (220, 250 important first-hand experience before embarking on other ther-
and 300 C) for a fixed retention time of 30 min. The relationship mal processes such as pyrolysis, gasification and combustion [37]. It
between CV and elemental compositions of the torrefied product provides essential information to motivate researchers in deep-
was established. Subsequently, over 30 papers on torrefaction of oil ening R&D in this area for design, optimisation and modification of
palm biomass were published with some recently highlighted for industrial-scale thermal conversion systems in the future.
significant improvement. Among these, Sabil et al. [25,26], Sulai- In the present study, EFB was torrefied under nitrogen atmo-
man et al. [27], Gallego et al. [28], Poudel et al. [29], Uemura et al. sphere in a fixed bed reactor to investigate the effects of temper-
[30,31] and Harun et al. [32] extended the scope of previous works ature (225e300 C) and retention time (20e60 min) on distribution
to optimize the process parameters (temperature, retention time, and characteristics of the torrefaction products. The structural
input gas type, and particle size) in enhancing quality of torrefied (surface morphology, functional property), thermal and fuel prop-
products in terms of CV, energy density or mass and energy yields. erties (CV, proximate and ultimate compositions, chlorine content)
Besides, torrefaction kinetics and process modeling were also re- of the torrefied EFB were comprehensively characterized via mul-
ported. For instance, Chew et al. [33] determined the isothermal tiple techniques such as scanning electron microscopy (SEM),
torrefaction kinetic of oil palm biomass. The kinetic prediction of Fourier transform infrared (FTIR), x-ray diffraction (XRD) and
torrefaction performance of three types of oil palm biomass using a thermogravimetric analysis (TGA). The liquid product was charac-
two-stage decomposition mechanism was found to closely match terized for chemical compositions via gas chromatography-mass
the experimental data. Besides, Sukiran et al. [12] estimated the spectrometry (GC-MS), CV, water content, pH and elemental
mass and energy yields from the torrefaction of EFB at different compositions while the gas product for gas types and distribution.
moisture levels using the least squares method on the basis of In addition, thermal behaviour of torrefaction products in terms of
quadratic equations. The empirical equations generated by the oxygen migration and energy distribution were discussed in-depth
developed model were found accurate, with a satisfactory fit of the to better understand what lies behind torrefaction process of EFB.
coefficient of determination values higher than 0.96 to the exper-
imental data. From literature, although palm kernel shell is the 2. Materials and methods
most popular biomass used in torrefaction, followed by EFB and
mesocarp fiber [4], efforts to torrefy EFB has been intensified due to 2.1. Sample preparation
its vast availability, versatility and less-competitive use.
Recently, Uemura et al. [3] conducted torrefaction of EFB under EFB was collected from a palm oil mill located in Negeri Sem-
flue gases atmosphere to replace the commonly used nitrogen, and bilan, Malaysia. It was milled and sieved using a test sieve shaker
found higher torrefaction severity in the former. A simulated flue (Endecotts EFL 2000) for particle separation to obtain test fractions
gases from biomass combustion in palm oil mills was used. Besides, of desired size ranged 106e250 mm. The resulting EFB particles
Talero et al. [34] evaluated the effects of temperature, retention were dried and weighed repeatedly at 103 C for 24 h until constant
2
M.A. Sukiran, W.M.A. Wan Daud, F. Abnisa et al. Energy 230 (2021) 120877
weight in order to remove any remaining water in the sample. The using Eq. (1).
dried EFB sample, denoted as raw EFB throughout the manuscript,
was stored in an air-tight container prior to analysis and torre- Mass of torrefied EFB=liquid product
Mass yield ð%Þ ¼ 100%
faction experiment. The physicochemical properties of the EFB Mass of raw EFB
sample used in torrefaction are: volatile matter, 77.42 wt%; fixed (1)
carbon, 13.84 wt%; moisture, 4.55 wt%; ash content, 4.19 wt%; ox-
The yield of the gases was calculated by subtracting the weight
ygen content, 50.57 wt%; carbon content, 42.82 wt% and CV,
percentages of the torrefied solid and liquid yields from the total of
17.57 MJ/kg.
100%.
The schematic diagram of the torrefaction system used for this TGA is well recognized as a proven method to determine the
study is illustrated in Fig. 1. It comprises a fixed bed reactor (in- thermal decomposition of both the raw and torrefied products. In
ternal diameter: 40 mm; length: 135 mm), an electrically heated this study, the TGA of the raw and torrefied EFB products was
furnace with temperature controller, a thermocouple, a gas supply performed using a thermogravimetric analyzer (Perkin Elmer-Pyris
system and a condenser unit. Approximately 10 g of EFB sample 6). Approximately 10 mg of the sample was pyrolyzed from room
with particle size of 106e250 mm was used for each experiment. temperature to 750 C at a constant heating rate of 20 C/min.
After loading the EFB sample, the reactor was closed tightly to Nitrogen was used as a carrier gas and its flow rate was controlled
ensure it was connected to the gas supply and condenser unit. The at 0.1 L/min for all experiments. The respective degradation pattern
reactor placed in the center of furnace was continuously purged was determined from the thermogravimetry (TG) and derivative
with nitrogen gas at a higher rate of 2.0 L/min for 15 min to elim- thermogravimetry (DTG) curves.
inate oxygen and any volatiles present so as to maintain an inert
atmosphere in the reactor prior to torrefaction process. The nitro- 2.4. Solid product characterization
gen gas flow rate was then reduced to about 1.0 L/min during tor-
refaction experiments. A lower flow rate was introduced during The proximate analysis of the raw and torrefied EFB (solid
torrefaction to avoid issues on sample being carried over and product) was carried out using a thermogravimetric analyzer (LECO
cooling of lower operating temperature. The EFB sample was TGA-701) to determine moisture, volatile matter, fixed carbon and
heated up using an electric furnace at a heating rate of 10 C/min to ash contents. The carbon, hydrogen and nitrogen contents of the
4 final temperatures (225, 250, 275 and 300 C); each designated raw and torrefied EFB were performed using a CHNS analyzer (LECO
for 3 different retention times (20, 40 and 60 min). An ice-bath CHNS-628). Meanwhile, the oxygen content was calculated from
condenser was installed to condense the released evaporated the difference between 100% and the total percentage of carbon,
gases from the process. The condensed liquid products were hydrogen and nitrogen. The CV of the raw and torrefied EFB was
majorly collected from the liquid-receiving bottle collector with determined using an isoperobolic system bomb calorimeter (LECO
traces from its connectors (Fig. 1). Due to the well-performed AC-600). For microscopic observation, the raw and torrefied EFB
condensation system installed in our study, the liquid products were subjected to a SEM (Hitachi Se3400 N model, Horiba EMAX)
that remained in the reactor were too small and contributed <1% to using backscattered electrons mode at 5.0 and 10.0 kV. The func-
the total liquid products; hence they were not collected. Mean- tional groups of the raw and torrefied EFB were detected using a
while, the non-condensable gases released from the reactor were FTIR spectrometer (Nicolet Magna-IR550, Madison). A small
measured and recorded using a gas analyzer (Quintox KM 9106). amount of the sample was mounted on a potassium bromide (KBr)
After the process was completed, the remaining solid inside the disc that was previously scanned as a background. The X-ray dif-
reactor was collected and then weighed. The final weight per- fractograms of the raw and torrefied EFB were collected via XRD
centages of the torrefied solid and liquid yields were calculated (Bruker AXS). These analyses were conducted using a CuKa
Fig. 1. Schematic diagram of experimental setup for torrefaction of empty fruit bunches.
3
M.A. Sukiran, W.M.A. Wan Daud, F. Abnisa et al. Energy 230 (2021) 120877
monochromatized radiation which operated at 40 kV and 40 mA at (Quintox KM 9106) equipped with electrochemical sensors. The
ambient temperature in the scanning range of 10e60 . The crys- non-condensable gases released from each experiment were
tallinity index (CrI) was then calculated according to Eq. (2). sampled for a period of 40 min at every 5-min interval.
CrI ¼ ðI002 Iam Þ = I002 (2) 2.7. Energy distribution of torrefaction products
where I002 (at 2q ¼ ~22 ) is the intensity at the maximum 2q peak Assumption was made to calculate the energy distribution of
and Iam is the intensity at the minimum 2q between ~15.5 and torrefaction products. Since the total energy yields of the torre-
~18.7 [38]. faction products (solid, liquid and gas) was 100%, the energy yield
The chlorine content of the raw and torrefied EFB was deter- of gaseous product was estimated from the difference between
mined via oxidative pyrolytic combustion followed by ion chro- 100% and the total percentage of energy yields of solid and liquid
matography (IC) using an automatic combustion unit (AQF-2100H, products.
Mitsubishi Chemical Analytech). The sample was combusted and
decomposed firstly in a pyrolysis tube under a flow of argon (carrier 3. Results and discussion
gas), followed by oxygen (oxidative pyrolysis) at about 1000 C. The
released hydrogen chloride (HCl) gas passed through an aqueous 3.1. Products distribution
absorbing solution (hydrogen peroxide) and then was injected
directly into the IC instrument, thereby eliminating sample Fig. 2 shows products distribution of EFB torrefaction at
matrices and any other associated interferences. Chlorine content 225e300 C for 20e60 min. Three types of products were pro-
was determined from the respective IC chromatogram based on duced, namely solid (torrefied EFB), liquid and gas. As shown, the
standard solution containing 1 mg/L of NaCl. yields of torrefied EFB (38.0e90.4 wt%) decreased with increasing
Based on the obtained mass yield and CV of the torrefied EFB, torrefaction temperature due to release of moisture and volatiles
energy yield (Ey) was calculated according to Eq. (3). [40]. At the lowest temperature of 225 C, the gas and liquid yields
were relatively very small compared to solid product which
CV of torrefied EFB
Ey ¼ Mass yield torrefied EFB (3) showed negligible mass reduction. However, as torrefaction tem-
CV of raw EFB perature increased from 250 C to 300 C, high solid mass yields
reduction occurred approaching a maximum of z40 wt% at 300 C
for 60 min while the gas and liquid yields were on an uptrend,
2.5. Characterization of liquid product reaching 44.7 wt% and 17.3 wt%, respectively. At this point, com-
plete decomposition of the amorphous hemicellulose took place
The liquid product was performed via GC-MS using Shimadzu's together with partial decomposition of the more crystalline cellu-
gas chromatography (QP2010 Ultra) equipped with a RTX-5 column lose and lignin that were subsequently released and transformed
to separate various organic compounds that were formed during into the corresponding components with lower molecular weight
the torrefaction process. The starting temperature of the oven was i.e. gas and liquid products [41]. The solid yield was higher than
40 C. It was held for 2 min before the temperature was increased to those of gas and liquid products, except at temperature of 300 C.
250 C at a rate of 10 C/min and held at this temperature for Hence, the solid yield was inversely proportional to the gas and
20 min. The injector port temperature and the detector tempera- liquid yields as shown by the torrefaction product distribution
ture were set at 280 C. The spectra obtained were identified and trend in Fig. 2. The trend of mass reduction was consistent with
characterized via the MS search libraries (NIST11). The water con- other studies concerning torrefaction of biomass other than EFB
tent of the liquid product was obtained via Karl Fischer titration [21,42]. On the other hand, the yield of torrefied EFB decreased with
method. The pH of the liquid product was measured by a pH meter an increase in retention time, while gas and liquid yields increased.
(Eutech Instruments, pH Tutor). The functional groups of the liquid However, this effect was less significant than that of torrefaction
product were detected using a FTIR spectrometer (Nicolet Magna- temperature for all the experiments conducted.
IR550, Madison). A small amount of the liquid product was
mounted on a KBr disc that was previously scanned as a back- 3.2. Characterization of torrefied product
ground. The density and viscosity of the liquid product were
determined using a digital density meter (DE 40) and a viscometer Fig. 3 shows the volatile matter, moisture, fixed carbon and ash
(Haake, VT 550) at 50 C, respectively. The carbon, hydrogen and contents of the raw and torrefied EFB. Fixed carbon increased from
nitrogen contents of the liquid product were determined using a 13.4 wt% to 40.9 wt%, while ash content from 3.7 wt% to 10.9 wt%,
CHNS analyzer (LECO CHNS-628). Meanwhile, the oxygen content which corresponded to an increase change in temperature and
was calculated from the difference between 100% and the total retention time from 200 C, 20 min to 300 C, 60 min, respectively.
percentage of carbon, hydrogen and nitrogen. The obtained The study on torrefaction of wood chip and oil palm frond by
elemental contents were used to calculate the CV of the liquid Homdoung et al. [43] showed similar results. A much larger per-
product according to Eq. (4) on the basis of >15 wt% of oxygen centage increase in fixed carbon indicates that the torrefied product
content obtained [39]. is good for combustion, as it can extend burning time and release
fairly high amount of heat. The ash content of the torrefified EFB
CVðMJ =kgÞ¼0:336 C þ1:418 H ð0:1530:00072 OÞ O þ0:0941S ranged from 3.7 wt% (lowest at 225 C) to 10.9 wt% (highest at
300 C). The breakdown of CeH bonds during torrefaction had
(4)
resulted in volatile loss and further concentrated the ash content in
the torrefied product. Meanwhile, the volatile matters of the tor-
refied EFB decreased when torrefaction temperature increased. The
2.6. Characterization of gas product volatile matters present originally in the raw EFB (80.1 wt%) had
significantly been reduced to 44.7 wt% at 300 C for a 60-min tor-
The compositions of non-condensable gases at the outlet of the refaction period. The release of volatile matters was accelerated due
torrefaction system were analyzed using a portable gas analyzer to catalytic effect of inorganic minerals present in the biomass
4
M.A. Sukiran, W.M.A. Wan Daud, F. Abnisa et al. Energy 230 (2021) 120877
Fig. 2. Torrefaction product distribution of empty fruit bunches at different temperatures and retention times.
Fig. 3. Proximate composition of raw and torrefied empty fruit bunches at different temperatures and retention times.
during torrefaction [44]. The moisture content of the torrefied EFB variation was consistent with other torrefaction studies using
ranged from 1.18 wt% to 4.86 wt%. The findings indicated that different biomass feedstock [47,48]. Using the ultimate results, the
retention time was less impactful on the proximate composition atomic H/C and O/C relations were calculated, plotted and refer-
compared to temperature for all the experiments conducted. enced in the Van Krevelen diagram (Fig. 4) together with some coal
Table 1 shows the elemental composition (via ultimate analysis), types (lignite and bituminous). As shown, torrefaction changed the
CV and energy yield of the torrefied EFB at different temperatures chemical compositions of EFB significantly and shifted the fuel
and retention times. As shown, the elemental composition of the properties of EFB downwards to be closer to coal. Under more se-
torrefied EFB changed as a function of torrefaction temperature and vere torrefaction conditions (300 C, 60 min) as investigated in this
retention time. However, the changes were more dominant for study, the chemical compositions of torrefied EFB had been much
temperature than retention time, especially on carbon and oxygen improved and were closer to those of lignite and bituminous,
contents. As the torrefaction temperature increased from 225 C to showing a significant increased fuel properties for combustion.
300 C, the torrefied EFB lost some of its hydrogen and oxygen el- The CV of the torrefied EFB increased when temperature and
ements due to the release of hydrogen and oxygen-rich volatiles retention time were increased (Table 1). It increased up to 25.7 MJ/
such as water, CO2 [45], acetic acid, methanol, aldehydes and ke- kg at torrefaction temperature of 300 C and retention time of
tones [46]. These losses had resulted in a relatively higher carbon 60 min which was about 46% higher compared to CV of the raw EFB.
content, which led to an increased CV. For example, the carbon Such a similar trend has been reported previously for other torre-
content of the torrefied EFB increased from 45.8 wt% (225 C, fied biomasses [42]. The increased carbon coupled with both the
20 min) to 62.2 wt% (300 C, 60 min), whereas the oxygen content decreased oxygen and hydrogen contents of the torrefied EFB had
decreased from 47.5 wt% to 31.8 wt%, respectively. This range of resulted in an increased CV because of a much higher energy
5
M.A. Sukiran, W.M.A. Wan Daud, F. Abnisa et al. Energy 230 (2021) 120877
Table 1
Elemental composition, calorific value (CV) and energy yield of raw and torrefied empty fruit bunches (EFB) at different temperatures and retention times.
6
M.A. Sukiran, W.M.A. Wan Daud, F. Abnisa et al. Energy 230 (2021) 120877
Fig. 5. Chlorine content and reduction of chlorine content of raw and torrefied empty fruit bunches (EFB) at different temperatures for 40 min residence time.
Fig. 6. FTIR spectra of raw and torrefied empty fruit bunches (EFB) at different temperatures for 40 min residence time.
difractograms and the calculated CrI of the raw and torrefied EFB at release of some volatile molecules such as H2O and CO2 [55].
different temperatures for 40 min. As shown, the two height of the Reduction in crystallinity of the torrefied celluloses might also be
peaks of each sample at 2q of 16 and 22 represented the typical translated to better susceptibility of any intended downstream
crystalline cellulose Ia (triclinic) and cellulose Ib (monoclinic), processes for a more efficient enzymatic digestibility, e.g. in a
respectively [55]. The height of the peak at 2q ¼ 22 , as attributed to biorefinery.
the crystalline region of cellulose, shrank a lot for EFB torrefied at The morphological and structural changes of the raw and tor-
300 C. The untorrefied EFB with the highest CrI (75.36%) experi- refied EFB, as examined via SEM (Fig. 8), allow us to understand
enced around 30% reduction in crystallinity with increasing torre- how thermal treatment affects the lignocellulosic structure of EFB,
faction temperature. At 225 C, the CrI of the torrefied EFB gradually and whether there is correlation between temperature and the
decreased to 70.21%, and finally to 52.38% at a much severe tem- anticipated devolatilization along the torrefaction process path.
perature of 300 C. This trend was in accordance with Chen et al. Prior to torrefaction, the raw EFB had well-defined tubular surfaces
[55] and Ma et al. [59]. It is important to note that while only little (Fig. 8). There were noticeable changes on the surfaces of the tor-
CrI variation was observed at lower torrefaction temperature refied EFB at 225 C and 250 C probably due to decomposition of
(225e250 C), huge reduction in crystallinity occurred at 300 C. hemicelluloses and losses of moisture and some light organic vol-
This indicated that the crystalline region of cellulose was less atiles [60]. However, no significant structural change was observed
reactive at lower torrefaction temperature, while a much induced except for the appearance of very small pores on the surfaces due to
thermal degradation reactivity occurred at higher temperature release of gases from inner part during torrefaction [38]. As tem-
causing the crystalline region to decompose and break. The crys- perature increased further to 275 C, inter-particle pores in the cell
talline region of cellulose was disturbed and changed involving wall of the torrefied EFB increased and some cracks were formed on
breakage of inter- and intramolecular hydrogen bonds, leading to the surfaces. Such increase was found to have been caused by the
7
M.A. Sukiran, W.M.A. Wan Daud, F. Abnisa et al. Energy 230 (2021) 120877
Fig. 7. XRD difractograms of raw and torrefied empty fruit bunches (EFB) at different temperatures for 40 min residence time.
trapping of volatile gases during the devolatilization stage. As the 0.8%, since almost all the hemicelluloses and celluloses had been
gases were released, decomposition of hemicellulose took place at removed at this point. The curve was then characterized by a
the same time, restricting some of these gases from moving well continuous slight devolatilization zone where the remaining lignin
through and thus trapped [60]. The entrapped air pockets caused would be decomposed and char formed at temperature between
uneven pressure on the surfaces leading to cracks. This phenome- 480 C and 700 C. At this stage, the mass loss decreased slowly as
non was more serious at 300 C with more cracking and voids shown in Fig. 9 (a) and (b).
visible on the surfaces. At this stage, celluloses from EFB had been
completely decomposed, thus microstructural deformation 3.4. Characterization of liquid products
occurred causing surface reorientation, as intense degradation
continued and micro holes formed. In addition, lignin and hemi- A retention time of 40 min at different torrefaction tempera-
cellulose which had far better thermal degradation tolerance tures was selected to represent the characteristics of the liquid
initially had now exceeded their glass transition temperatures, products. Fig. 10 shows the relative contents (%) of organic com-
leading to further cracks and crevice [61]. Similar phenomenon has pounds in the liquid products obtained at torrefaction temperature
been observed during torrefaction of pruned olive tree under ni- of 250, 275 and 300 C for 40 min. The relative content of each
trogen atmosphere [47]. chemical compound was calculated based on peak area % from the
GC-MS results. The total of peak area % of each chemical compound
3.3. Thermal behaviour of torrefied EFB at different temperature represents the relative content of chemical
species present. As determined via GC-MS, the liquid products were
The thermal degradation behaviour of the raw and torrefied EFB mainly composed of organic acids (33.21e41.12%), phenol
at different temperatures was represented in Fig. 9 (a) and (b). The (13.04e23.26%) and phenol derivatives (5.54e8.49%), furans
curves, showing increasing trend in thermal degradation of torre- (1.62e3.56%), ketones (2.20e13.21%) and alcohols (3.26e10.37%).
fied EFB over temperature, were similar to other torrefied bio- Table 2 shows the distribution of chemical components in the liquid
masses such as corncobs [62], wood [63] and pruned olive tree [47] products at different torrefaction temperatures. Acid was found
when heated up to 700 C. The whole degradation process for the most dominant, followed by phenol and ketone. Acetic acid was the
raw and torrefied EFB was divided into three different stages. The major acid present, accounted for about 70% of the total pool of
first stage prior to torrefaction, from 30 C to 200 C, took place acids in the torrefied liquid products. Formation of acetic acid
mainly due to evaporation of moisture and decomposition of some during thermal degradation of hemicellulose can occur via two
light organic compounds, at <5e10% mass losses [38,62]. After that, pathways [56]: firstly, by elimination of O-acetyl groups linked to
the weight loss gradually increased with increasing initial degra- xylan side chains at the C2 position and secondly, by ring-opening
dation temperature (torrefaction severity). For example, the weight reaction of 4-O-methylglucuronic acid unit after breakage of the
loss at 225 C and 275 C were 8.11 wt% and 8.78 wt%, respectively. carbonyl and O-methyl groups [56]. The torrefied liquid products
The main mass loss occurred at 200e480 C. At this stage, the also contained small amounts of other acids such as propanoic acid,
cellulose and hemicellulose in the EFB had completely been formic acid, dodecanoic acid and oleic acid. The amount of these
depolymerized [18]. The DTG curves showed an upward shift of the acids was reduced when torrefaction temperature increased from
maximum degradation temperature from 331 C, 337 Ce344 C for 250 C to 300 C, probably due to secondary cracking of the long-
the EFB torrefied at 225, 250 and 275 C, respectively. This upward chain fatty acids as intermediates. The maximum yield of acid
trend might be due to the presence of mostly cellulose and lignin compounds in liquid products was 41% at torrefaction temperature
[63] while hemicellulose had been substantially decomposed. of 250 C. Similar finding was reported by Chen et al. [19].
Meanwhile, the corresponding blunt peak at 380 C for the EFB Besides, no significant difference was observed in distribution of
torrefied at 300 C was associated with a much lower mass loss of chemical compounds in the liquid products at 250 C and 275 C
8
M.A. Sukiran, W.M.A. Wan Daud, F. Abnisa et al. Energy 230 (2021) 120877
Fig. 8. SEM images of raw and torrefied empty fruit bunches (EFB) at different temperatures for 40 min residence time.
(Table 2). However, remarkable cross reactions had occurred be- with rising temperature, the content of phenol derivatives in tor-
tween holocellulose and the decomposed lignin typically from refied liquid product will increase [64]. The relative content of each
275 C to 300 C, as was evidenced by a diverse distribution profile individual chemical compound, as analyzed via GC-MS, was accu-
of phenols derivatives and ketones. The relative contents of ketones rate at time of measurement. Further degradation leading to any
and alcohols decreased when torrefaction temperature increased chemical compositional changes and their relative quantity might
from 250 C to 300 C, while those of phenol and phenol derivatives occur subjected to period of storage, storage condition and tem-
increased. Such increase could be due to cleavages of thermally perature. Overall, when torrefaction temperature increased, the
unstable ether bonds including b-O-4 linkages of the lignin struc- chemical compositions of liquid products of EFB became more
ture in the torrefied EFB. As bond cleavages are normally enhanced complex (Table 2).
9
M.A. Sukiran, W.M.A. Wan Daud, F. Abnisa et al. Energy 230 (2021) 120877
Fig. 9. TG and DTG curves for raw and torrefied empty fruit bunches (EFB) at different temperatures for 40 min residence time.
10
M.A. Sukiran, W.M.A. Wan Daud, F. Abnisa et al. Energy 230 (2021) 120877
Table 2
Distribution of chemical components in liquid products of empty fruit bunches at different torrefaction temperatures for 40 min residence time.
Acids Acetic acid (34.97) Acetic acid (32.21) Acetic acid (29.29)
Propanoic acid (5.96) Propanoic acid (1.11) Propanoic acid (1.01)
Oleic acid (0.19) Formic acid (3.79) Dodecanoic acid (0.36)
n-Hexadecanoic acid n-Hexadecanoic acid (1.21)
(0.34) Oleic acid (1.05)
Tetradecanoic acid (0.29)
Ketones 1-Hydroxy-2-butanone (7.61) 1-Hydroxy-2-butanone (1.25) 2-Butanone, 1-(acetyloxy)- (0.42)
1,2-Cyclopentanedione (4.01) 1,2-Cyclopentanedione (1.93) Cyclohexanone (0.88)
3-Pentanone (1.59) 3-Pentanone (2.01) 2-Cyclopenten-1-one, 2-hydroxy-3-methyl- (0.97)
Ethanone, 1-oxiranyl- 4-Hexen-3-one, 4,5-dimethyl- (0.54)
(6.92) 2-Cyclopenten-1-one, 3-ethyl-2-hydroxy- (1.43)
1-Hydroxy-2-butanone 4-Hydroxy-3-methylacetophenone (0.50)
(0.76) 2-Propanone, 1-(4-hydroxy-3-methoxyphenyl)- (0.67)
Alcohols 3-Butyn-2-ol (0.97) Cyclopentanol (0.57) Catechol (0.63)
1-Propanol (3.84) Catechol (1.21) 1,2,3-Propanetriol, 1-acetate (0.42)
1,2-Benzenediol (5.56) 1,2-Benzenediol, 3-methoxy (1.10) 1-Octen-1-ol, acetate (1.07)
Glycidol (0.47) 4-Octen-3-ol, 2,2-dimethyl- (1.14)
Cyclopropyl carbinol
(1.16)
Phenols Phenol (13.04) Phenol (18.65) Phenol (23.26)
Phenol, 2-methoxy (3.68) Phenol, 2-methoxy (2.51) Phenol, 2-methoxy- (1.49)
Phenol, 3,4-dimethoxy (1.86) Phenol, 3,4-dimethoxy- Phenol, 2,6-dimethoxy (3.41)
(0.72) Phenol, 2-methoxy-5-(1-propenyl)-, (E)- (1.34)
Phenol, 2,6-dimethoxy-4-(2-propenyl) (1.75) Phenol, 2,6-dimethoxy-4-(2-propenyl)- (2.25)
Furans Furfural (1.05) Furfural (1.22) Furfural (1.24)
2- Furancarboxaldehyde (3H)-Furanone, dihydro-4-hydroxy- (0.46) 2-Furanmethanol (0.96)
(0.57) 2-Hexanoylfuran (0.70)
2-Furanmethanol (1.18)
Table 3
Elemental composition, calorific value (CV), water content and pH of liquid products of empty fruit bunches at different torrefaction temperatures for 40 min residence time.
Sample Ultimate analysis (wt%) CV (MJ/kg) Water content (wt%) pH Density (g/cm3)
a
C H N O S
225 21.37 6.94 0.26 71.38 0.05 6.16 64.2 2.51 1.12
250 24.38 6.83 0.51 68.27 0.01 7.48 58.7 2.62 1.17
275 27.74 6.75 0.50 64.95 0.06 9.01 52.4 2.77 1.18
300 32.54 6.48 0.64 60.30 0.01 10.94 46.1 2.81 1.17
a
By difference.
Only a single measurement was done due to limited sample vol- be due to the presence of water as impurity. The CeH stretching
ume. The high content of water might have caused dilution of the vibrations between 2800 and 3000 cm1 and CeH deformation and
liquid product for this low viscosity value. Meanwhile, the liquid bending vibrations between 1200 and 1400 cm1 indicated the
products contained the highest oxygen element (60.3e71.4 wt%) presence of alkanes. The presence of significant peaks between
followed by carbon (21.4e32.5 wt%). The oxygen content in the 1640 and 1700 cm1 could be due to C]O stretching vibrations
liquid products was proportional to the amount of water content, indicating compounds such as ketones, phenols, carboxylic acids or
mainly in the form of oxygen-containing organic compounds [19]. esters. C]C stretching vibrations indicative of alkenes and aro-
As shown in Table 3, the carbon content of the liquid products in all matics were found in the range of 1450 and 1550 cm1. The peaks
cases increased and vice versa for oxygen content. The increasing between 1300 and 950 cm1 showing the CeO and OeH stretching
carbon content of liquid products was due to formation of some vibrations may be due to the presence of primary, secondary and
phenols and aromatic compounds as a result of lignin decomposi- tertiary alcohols, phenols, ethers and esters.
tion (Fig. 10) from 250 C to 300 C. Similar trend was observed by
Chen et al. [19] previously. Based on the results, the CV of liquid 3.5. Characterization of gas products
products (Table 2) was calculated using Eq. (4). The CV was low, in
the range of 6.16 MJ/kg (lowest, at 225 C) to 10.94 MJ/kg (highest, The main gas components based on relative content (volume
at 300 C), due to high water content. In general, the liquid products concentration, vol%) during a 40-min EFB torrefaction at temper-
are considered not a good fuel for used in boiler or engine due to ature of 250, 275 and 300 C are shown in Fig. 12. The gas products
the low CV and high moisture content. However, according to Chen of torrefied EFB in decreasing order were CO2 (main), CO and CH4
et al. [65], dewatering can greatly improve CV of liquid products by (traces). It was found that the volume concentration of CO and CH4
>100% from 8.9 to 12.7 MJ/kg to 22.4e24.8 MJ/kg. The dewatered increased with increasing torrefaction temperature from 250 C to
liquid products with such a high CV can then be used as a fuel. In 300 C, while that of CO2 otherwise. According to Zhang et al. [66],
Fig. 11, the OeH stretching vibrations between 3200 and 3600 cm1 as torrefaction severity increases, biomass degradation through
indicated the presence of phenols from lignin and alcohols from decarbonylation reaction is more favorable involving loss of CO
cellulose. However, the broad OeH stretching vibrations could also than decarboxylation reaction. The latter, exothermic in nature,
11
M.A. Sukiran, W.M.A. Wan Daud, F. Abnisa et al. Energy 230 (2021) 120877
Fig. 11. FTIR spectra of liquid products of empty fruit bunches at different temperatures for 40 min residence time.
Fig. 13. Gas distribution profile during a 40-min torrefaction of empty fruit bunches at different temperatures.
(raw EFB) to 31.81e47.50 wt%, which corresponded to some 6e37% oxygen migration in the form of CO2, CO, water and oxygen-
oxygen removal efficiency. The removed oxygen has been shown to containing organic compounds is evident during torrefaction of
have migrated to the gaseous and liquid products in the form of EFB as fairly good degree of deoxygenation has taken place resulted
CO2, CO, water and oxygen-containing organic compounds [19]. in a promising torrefied product.
According to Ma et al. [56], migration of oxygen from water in- Oxygen migration during torrefaction of EFB had significantly
creases with increasing torrefaction temperature. In this study, free imparted the energy distribution of the 40-min torrefaction prod-
water was evaporated from the raw EFB at 100 C during initial ucts (solid, liquid and gas) at different temperatures. Fig. 14 shows
heating of torrefaction. As torrefaction temperature increased from that the energy yields of gaseous and liquid products increased
100 C to 300 C, more bound water molecules were released gradually as the torrefaction temperature increased, while that of
through hydroxyl bond breakages linking the basic structural units torrefied EFB solid product decreased. Nevertheless, the energy
(glucan, glycosyl and phenylpropane) present in the cellulose, yield of torrefied EFB was not much affected as it was still able to
hemicellulose and lignin, respectively [40]. As shown in Fig. 12, CO2 retain most of the energy containing in its original raw form. The
was the most dominant oxygen carrier for migration to the torre- energy yield in the liquid products was the lowest due to the
fied gaseous products at all temperatures. For the liquid products, resulted low mass yield (Fig. 2), carbon content and CV (Table 3).
the employed torrefaction temperatures had demonstrated signif-
icant potential routes for oxygen transfer. At lower torrefaction 3.7. Social impact of torrefaction of oil palm biomass
temperature, oxygen was transferred to acids, ketones and alcohols.
With an increase in torrefaction temperature, the relative content Torrefaction of oil palm biomass can affect the surrounding
of phenol and its many derivatives increased (Fig. 10) due to an communities and stakeholders through various ways. Broadly, it
accelerated decomposition of lignin, showing higher degree of can range from improving the environment and quality of life with
oxygen migration to phenols in the liquid products. Therefore, reduction of GHG emissions to creation of new jobs and business
13
M.A. Sukiran, W.M.A. Wan Daud, F. Abnisa et al. Energy 230 (2021) 120877
conditions, and last but not least, the level of education will be
enhanced too. Ultimately, it is hoped that biomass valorisation into
value-added production for energy and non-energy applications
can help building a more responsible society and a circular
economy.
4. Conclusion
Appendix A. Supplementary data torrefaction. Waste and Biomass Valorization; 2018. p. 1e12.
[29] Poudel J, Ohm T-I, Gu JH, Shin MC, Oh SC. Comparative study of torrefaction of
empty fruit bunches and palm kernel shell. J Mater Cycles Waste Manag
Supplementary data to this article can be found online at 2017;19:917e27.
https://doi.org/10.1016/j.energy.2021.120877. [30] Uemura Y, Omar W, Othman NA, Yusup S, Tsutsui T. Torrefaction of oil palm
EFB in the presence of oxygen. Fuel 2013;103:156e60.
[31] Uemura Y, Omar WN, Tsutsui T, Yusup SB. Torrefaction of oil palm wastes.
References Fuel 2011;90:2585e91.
[32] Harun NHHM, Samad NAFA, Saleh S. Development of kinetics model for tor-
[1] Mohd Faizal H, Shamsuddin HS, Heiree MHM, Muhammad Ariff Hanaffi MF, refaction of empty fruit bunch from palm oil waste. Energy Procedia
Abdul Rahman MR, Rahman MM, et al. Torrefaction of densified mesocarp 2017;105:744e9.
fibre and palm kernel shell. Renew Energy 2018;122:419e28. [33] Chew J-J, Doshi V, Yong S-T, Bhattacharya S. Kinetic study of torrefaction of oil
[2] Petinrin JO, Shaaban M. Renewable energy for continuous energy sustain- palm shell, mesocarp and empty fruit bunch. J Therm Anal Calorim 2016;126:
ability in Malaysia. Renew Sustain Energy Rev 2015;50:967e81. 709e15.
[3] Uemura Y, Sellappah V, Trinh TH, Hassan S, Tanoue K-i. Torrefaction of empty [34] Talero G, Rinco n S, Go
mez A. Biomass torrefaction in a standard retort: a study
fruit bunches under biomass combustion gas atmosphere. Bioresour Technol on oil palm solid residues. Fuel 2019;244:366e78.
2017;243:107e17. [35] Lam SS, Tsang YF, Yek PNY, Liew RK, Osman MS, Peng W, et al. Co-processing
[4] Sukiran MA, Abnisa F, Wan Daud WMA, Abu Bakar N, Loh SK. A review of of oil palm waste and waste oil via microwave co-torrefaction: a waste
torrefaction of oil palm solid wastes for biofuel production. Energy Convers reduction approach for producing solid fuel product with improved proper-
Manag 2017;149:101e20. ties. Process Saf Environ Protect 2019;128:30e5.
[5] Ghulam Kadir AP, Elina H, Soh Kheang L, Meilina OA, Kamalrudin MS, Mohd [36] Nyakuma BB, Oladokun O, Wong SL, Abdullah TAT. Torrefaction of oil palm
Noor Izuddin ZB, et al. Oil palm economic performance in Malaysia and r&d empty fruit bunch pellets: product yield, distribution and fuel characterisation
progress in 2019. J. Oil Palm Res. 2020;32:159e90. for enhanced energy recovery. Biomass Conversion and Biorefinery; 2021.
[6] Derman E, Abdulla R, Marbawi H, Sabullah MK. Oil palm empty fruit bunches [37] He Q, Ding L, Gong Y, Li W, Wei J, Yu G. Effect of torrefaction on pinewood
as a promising feedstock for bioethanol production in Malaysia. Renew Energy pyrolysis kinetics and thermal behavior using thermogravimetric analysis.
2018;129:285e98. Bioresour Technol 2019;280:104e11.
[7] Yiin CL, Ho S, Yusup S, Quitain AT, Chan YH, Loy ACM, et al. Recovery of cel- [38] Li S-X, Zou J-Y, Li M-F, Wu X-F, Bian J, Xue Z-M. Structural and thermal
lulose fibers from oil palm empty fruit bunch for pulp and paper using green properties of Populus tomentosa during carbon dioxide torrefaction. Energy
delignification approach. Bioresour Technol 2019;290:121797. 2017;124:321e9.
[8] Siddiquee S, Shafawati SN, Naher L. Effective composting of empty fruit [39] Buckley TJ. Calculation of higher heating values of biomass materials and
bunches using potential Trichoderma strains. Biotechnol Rep 2017;13:1e7. waste components from elemental analyses. Resour Conserv Recycl 1991;5:
[9] Idris R, Chong WWF, Ali A, Idris S, Hasan MF, Ani FN, et al. Phenol-rich bio-oil 329e41.
derivation via microwave-induced fast pyrolysis of oil palm empty fruit bunch [40] Chen D, Gao A, Cen K, Zhang J, Cao X, Ma Z. Investigation of biomass torre-
with activated carbon. Environ Technol Innov 2021;21:101291. faction based on three major components: hemicellulose, cellulose, and
[10] Chen W-H, Kuo P-C. Torrefaction and co-torrefaction characterization of lignin. Energy Convers Manag 2018;169:228e37.
hemicellulose, cellulose and lignin as well as torrefaction of some basic con- [41] Li S-X, Chen C-Z, Li M-F, Xiao X. Torrefaction of corncob to produce charcoal
stituents in biomass. Energy 2011;36:803e11. under nitrogen and carbon dioxide atmospheres. Bioresour Technol
[11] Yue Y, Singh H, Singh B, Mani S. Torrefaction of sorghum biomass to improve 2018;249:348e53.
fuel properties. Bioresour Technol 2017;232:372e9. [42] Wang L, Barta-Rajnai E, Skreiberg Ø, Khalil R, Czege
ny Z, Jakab E, et al. Effect of
[12] Sukiran MA, Abnisa F, Syafiie S, Wan Daud WMA, Nasrin AB, Abdul Aziz A, torrefaction on physiochemical characteristics and grindability of stem wood,
et al. Experimental and modelling study of the torrefaction of empty fruit stump and bark. Appl Energy 2018;227:137e48.
bunches as a potential fuel for palm oil mill boilers. Biomass Bioenergy [43] Homdoung N, Sasujit K, Uttharuan J, Wongsiriamnuay T, Tippayawong N.
2020;136:105530. Influence of torrefaction temperature and time on the yields and properties of
[13] Nam H, Capareda S. Experimental investigation of torrefaction of two agri- torrefied biomass. Eng Appl Sci Res 2019;46:170e5.
cultural wastes of different composition using RSM (response surface meth- [44] Almeida G, Brito J, Perre P. Alterations in energy properties of eucalyptus
odology). Energy 2015;91:507e16. wood and bark subjected to torrefaction: the potential of mass loss as a
[14] Sarvaramini A, Gravel O, Larachi F. Torrefaction of ionic-liquid impregnated synthetic indicator. Bioresour Technol 2010;101:9778e84.
lignocellulosic biomass and its comparison to dry torrefaction. Fuel 2013;103: [45] Niu Y, Lv Y, Lei Y, Liu S, Liang Y, Wang D, et al. Biomass torrefaction: prop-
814e26. erties, applications, challenges, and economy. Renew Sustain Energy Rev
[15] Nhuchhen DR, Basu P, Acharya B. A comprehensive review on biomass tor- 2019;115:109395.
refaction. Int J Renew Energy Biofuels 2014:1e56. 2014. [46] Chiou B-S, Valenzuela-Medina D, Bilbao-Sainz C, Klamczynski AP, Avena-
[16] Bajcar M, Zaguła G, Saletnik B, Tarapatskyy M, Puchalski C. Relationship be- Bustillos RJ, Milczarek RR, et al. Torrefaction of almond shells: effects of tor-
tween torrefaction parameters and physicochemical properties of torrefied refaction conditions on properties of solid and condensate products. Ind Crop
products obtained from selected plant biomass. Energies 2018;11:2919. Prod 2016;86:40e8.
[17] Chin KL, H’ng PS, Go WZ, Wong WZ, Lim TW, Maminski M, et al. Optimization [47] Martín-Lara MA, Ronda A, Zamora MC, Calero M. Torrefaction of olive tree
of torrefaction conditions for high energy density solid biofuel from oil palm pruning: effect of operating conditions on solid product properties. Fuel
biomass and fast growing species available in Malaysia. Ind Crop Prod 2017;202:109e17.
2013;49:768e74. [48] Chiou B-S, Valenzuela-Medina D, Bilbao-Sainz C, Klamczynski AK, Avena-
[18] Li H, Liu X, Legros R, Bi XT, Lim CJ, Sokhansanj S. Torrefaction of sawdust in a Bustillos RJ, Milczarek RR, et al. Torrefaction of pomaces and nut shells. Bio-
fluidized bed reactor. Bioresour Technol 2012;103:453e8. resour Technol 2015;177:58e65.
[19] Chen D, Gao A, Ma Z, Fei D, Chang Y, Shen C. In-depth study of rice husk [49] Mamvura TA, Pahla G, Muzenda E. Torrefaction of waste biomass for appli-
torrefaction: characterization of solid, liquid and gaseous products, oxygen cation in energy production in South Africa. S Afr J Chem Eng 2018;25:1e12.
migration and energy yield. Bioresour Technol 2018;253:148e53. [50] Ren X, Sun R, Chi H-H, Meng X, Li Y, Levendis YA. Hydrogen chloride emis-
[20] Chen D, Zhou J, Zhang Q, Zhu X, Lu Q. Upgrading of rice husk by torrefaction sions from combustion of raw and torrefied biomass. Fuel 2017;200:37e46.
and its influence on the fuel properties. BioResources 2014;9:5893e905. [51] Saleh SB, Flensborg JP, Shoulaifar TK, Sarossy Z, Hansen BB, Egsgaard H, et al.
[21] Ben H, Ragauskas AJ. Torrefaction of loblolly pine. Green Chem 2012;14:72e6. Release of chlorine and sulfur during biomass torrefaction and pyrolysis.
[22] Li M-F, Li X, Bian J, Chen C-Z, Yu Y-T, Sun R-C. Effect of temperature and Energy Fuel 2014;28:3738e46.
holding time on bamboo torrefaction. Biomass Bioenergy 2015;83:366e72. [52] Keipi T, Tolvanen H, Kokko L, Raiko R. The effect of torrefaction on the chlorine
[23] Li M-F, Chen C-Z, Li X, Shen Y, Bian J, Sun R-C. Torrefaction of bamboo under content and heating value of eight woody biomass samples. Biomass Bio-
nitrogen atmosphere: influence of temperature and time on the structure and energy 2014;66:232e9.
properties of the solid product. Fuel 2015;161:193e6. [53] Chen H, Chen X, Qiao Z, Liu H. Release and transformation characteristics of K
[24] Uemura Y, Omar W, Tsutsui T, Subbarao D, Yusup S. Relationship between and Cl during straw torrefaction and mild pyrolysis. Fuel 2016;167:31e9.
calorific value and elementary composition of torrefied lignocellulosic [54] Duranay ND, Akkuş G. Solid fuel production with torrefaction from vineyard
biomass. J Appl Sci 2010;10:3250e6. pruning waste. Biomass Conversion and Biorefinery; 2019.
[25] Sabil KM, Aziz MA, Lal B, Uemura Y. Effects of torrefaction on the physi- [55] Chen D, Cen K, Cao X, Li Y, Zhang Y, Ma H. Restudy on torrefaction of corn stalk
ochemical properties of oil palm empty fruit bunches, mesocarp fiber and from the point of view of deoxygenation and decarbonization. J Anal Appl
kernel shell. Biomass Bioenergy 2013;56:351e60. Pyrol 2018;135:85e93.
[26] Sabil KM, Aziz MA, Lal B, Uemura Y. Synthetic indicator on the severity of [56] Ma Z, Zhang Y, Shen Y, Wang J, Yang Y, Zhang W, et al. Oxygen migration
torrefaction of oil palm biomass residues through mass loss measurement. characteristics during bamboo torrefaction process based on the properties of
Appl Energy 2013;111:821e6. torrefied solid, gaseous, and liquid products. Biomass Bioenergy 2019;128:
[27] Sulaiman MH, Uemura Y, Azizan MT. Torrefaction of empty fruit bunches in 105300.
inert condition at various temperature and time. Procedia Eng 2016;148: [57] Tong S, Sun Y, Li X, Hu Z, Dacres OD, Worasuwannarak N, et al. Gas-pres-
573e9. surized torrefaction of biomass wastes: roles of pressure and secondary re-
[28] Gallego LJ, Cardona S, Martínez E, Rios LA. Valorization of palm-oil residues: actions. Bioresour Technol 2020;313:123640.
integrated production of a good quality bio-coal and electricity via [58] Karimi K, Taherzadeh MJ. A critical review of analytical methods in
15
M.A. Sukiran, W.M.A. Wan Daud, F. Abnisa et al. Energy 230 (2021) 120877
pretreatment of lignocelluloses: composition, imaging, and crystallinity. Bio- [65] Chen W-H, Liu S-H, Juang T-T, Tsai C-M, Zhuang Y-Q. Characterization of solid
resour Technol 2016;200:1008e18. and liquid products from bamboo torrefaction. Appl Energy 2015;160:
[59] Ma Z, Wang J, Li C, Yang Y, Liu X, Zhao C, et al. New sight on the lignin tor- 829e35.
refaction pretreatment: relevance between the evolution of chemical struc- [66] Zhang S, Su Y, Ding K, Zhu S, Zhang H, Liu X, et al. Effect of inorganic species on
ture and the properties of torrefied gaseous, liquid, and solid products. torrefaction process and product properties of rice husk. Bioresour Technol
Bioresour Technol 2019;288:121528. 2018;265:450e5.
[60] Daud D, Rahman AA, Shamsuddin A. Devolatilization studies of oil palm [67] Singh RK, Chakraborty JP, Sarkar A. Optimizing the torrefaction of pigeon pea
biomass for torrefaction process through scanning electron microscopy. In: stalk (cajanus cajan) using response surface methodology (RSM) and char-
IOP conference series: earth and environmental science. IOP Publishing; 2016, acterization of solid, liquid and gaseous products. Renew Energy 2020;155:
012072. 677e90.
[61] van der Stelt MJC, Gerhauser H, Kiel JHA, Ptasinski KJ. Biomass upgrading by [68] Loh SK. The potential of the Malaysian oil palm biomass as a renewable en-
torrefaction for the production of biofuels: a review. Biomass Bioenergy ergy source. Energy Convers Manag 2017;141:285e98.
2011;35:3748e62. [69] Du S-W, Chen W-H, Lucas JA. Pretreatment of biomass by torrefaction and
[62] Tian X, Dai L, Wang Y, Zeng Z, Zhang S, Jiang L, et al. Influence of torrefaction carbonization for coal blend used in pulverized coal injection. Bioresour
pretreatment on corncobs: a study on fundamental characteristics, thermal Technol 2014;161:333e9.
behavior, and kinetic. Bioresour Technol 2020;297:122490. [70] Nunes LJ, Matias JC. Biomass torrefaction as a key driver for the sustainable
[63] Singh S, Chakraborty JP, Mondal MK. Torrefaction of woody biomass (Acacia development and decarbonization of energy production. Sustainability
nilotica): investigation of fuel and flow properties to study its suitability as a 2020;12:922.
good quality solid fuel. Renew Energy 2020;153:711e24. [71] Sacchelli S. Social, economic, and environmental impacts of biomass and
[64] Chang S, Zhao Z, Zheng A, He F, Huang Z, Li H. Characterization of products biofuel supply chains. Biomass supply chains for bioenergy and biorefining.;
from torrefaction of sprucewood and bagasse in an auger reactor. Energy Fuel 2016. p. 191e213.
2012;26:7009e17.
16