ww.sciencedirect.com

b i om a s s an d b i o e n e r g y 4 9 ( 2 0 1 3 ) 2 1 4e2 2 1

Available online at w

ht tp: / /www.elsevier .com/locate/biombioe

Pelletizing properties of torrefied wheat straw

Wolfgang Stelte a,*, Niels Peter K. Nielsen a, Hans Ove Hansen a, Jonas Dahl a, Lei Shang b,Anand R. Sanadi c

aRenewable Energy and Transport, Energy and Climate Division, Danish Technological Institute, Gregersensvej 1, 2630 Taastrup, DenmarkbDepartment of Chemical and Biochemical Engineering, Technical University of Denmark, 2800 Kgs. Lyngby, DenmarkcBiomass and Ecosystem Science, Faculty of Life Sciences, University of Copenhagen, 1958 Frederiksberg, Denmark

a r t i c l e i n f o

Article history:

Received 16 June 2012

Received in revised form

16 November 2012

Accepted 10 December 2012

Available online 25 January 2013

Keywords:

Biomass

Torrefaction

Wheat straw

Pelletization

Pellet

* Corresponding author. Tel.: þ45 7220 1072.E-mail addresses: wst@teknologisk.dk, st

0961-9534/$ e see front matter ª 2012 Elsevhttp://dx.doi.org/10.1016/j.biombioe.2012.12.0

a b s t r a c t

Combined torrefaction and pelletization are used to increase the fuel value of biomass by

increasing its energy density and improving its handling and combustion properties.

However, pelletization of torrefied biomass can be challenging and in this study the tor-

refaction and pelletizing properties of wheat straw have been analyzed. Laboratory

equipment has been used to investigate the pelletizing properties of wheat straw torrefied

at temperatures between 150 and 300 �C. IR spectroscopy and chemical analyses have

shown that high torrefaction temperatures change the chemical properties of the wheat

straw significantly, and the pelletizing analyses have shown that these changes correlate

to changes in the pelletizing properties. Torrefaction increase the friction in the press

channel and pellet strength and density decrease with an increase in torrefaction

temperature.

ª 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Biomass from wood and agricultural residues is increasingly

important for sustainable heat and power production at

industrial scale [1]. Political decision makers have set ambi-

tious goals to push forward renewable energies, and power

producers have identified biomass utilization as a potential

way to increase their share of renewable energy relatively fast

and cost efficient when using it in their existing coal power

plants [2]. However the physical properties of biomass have

been identified as a challenge for power producers due to

its poor handling properties and combustion related

problems. Biomass as a material is inhomogeneous in struc-

ture and composition and its bulk and energy density is

low when compared to conventional fossil fuels. Furthermore

is it necessary to bridge long distances between biomass

elte@gmail.com (W. Steltier Ltd. All rights reserve25

production sites and power plants and that require expensive

logistics [3]. There are different technologies to improve the

physical properties of solid biomass. Mechanical treatment

such as pelletization and briquetting results in a homogeneous

biomass product of high density [4]. Thermal treatment i.e.

torrefaction removesmoisture, volatiles and degrades parts of

its carbohydrate fraction, resulting in a product with higher

energy density and better moisture resistance [5]. Both

processes have been studied extensively and recent efforts

have been made to combine both processes where biomass is

thermally treated (torrefied) and subsequently pelletized [6e8]

The resulting product is a “coal like” product that is expected to

be used as replacement and/or supplement for coal in existing

heat and power plants [8]. The properties of pellets made from

torrefied biomass are inmanyways similar to coal as shown in

Table 1.

e).d.

Table 1 e Benchmark of torrefied pellets with selected bio- and fossil-fuels [8e13].

Torrefied pellets Wood pellets Charcoal Lignite coal Bituminous coal Anthracite coal

Moisture [g kg-1] 10e50 70e100 10e50 65e400 20e160 20e160

Fixed carbon [g kg-1] 280e350 200e250 850e870 314e410 340e780 800e860

Bulk density [kg m3] 750e850 550e750 ca. 200 560e865 670e910 800e930

Energy density [GJ m3] 15e18.5 7.5e10.4 6e6.4 14e19 19e30 18e24

Hydroscopic properties Moderate hydrophobic Hydrophilic Hydrophobic Hydrophobic Hydrophobic Hydrophobic

Biological degradation No Yes No No No No

Milling requirements Classic Special Classic Classic Classic Classic

Fig. 1 e Process stages of biomass pellet production.

b i om a s s a n d b i o e n e r g y 4 9 ( 2 0 1 3 ) 2 1 4e2 2 1 215

Major advantages are their high energy density hydro-

phobic nature and good grindability, and it is expected that

torrefied pellets can be handled similar as coal. Anothermajor

advantage of torrefied pellets pointed out by the pellet

industry is that they are an ideal fuel to be used in coal fired

power plants since only few changes have to be made to the

existing power plant design. A recent study has shown that it

is possible to fire 100% torrefied biomass in a pulverized coal

boiler without decrease of the boiler efficiency and fluctuation

of boiler load [14].

Torrefaction is amild pyrolysis process at about 200e300 �Cin the absence of oxygen that changes the biomass properties

significantly [5,8]. Due to thermal depolymerization and

degradation a large fraction of the biomass carbohydrate

fraction is removed from the biomass, especially the hemi-

celluloses that account for about 25e35% dry matter of ligno-

cellulose type biomass. Its natural function in biomass is to

crosslink the cellulose fibers and the ligninmatrix and thereby

providing the biomasswith its tenacious and flexible structure

which is crucial for the mechanical stability of each plant.

It has been shown that hemicelluloses start to degrade at

low torrefaction temperatures of about 230 �C and that the

rate of degradation increases strongly with an increasing

torrefaction temperature [15]. Cellulose and lignin have been

shown to be more resistant to thermal decomposition and

significant degradation occurs at higher temperature than for

hemicellulose. Chen and Kuo [15] found that at a torrefaction

temperature of 260 �C almost 40% of the hemicellulose has

been removed from the biomass compared to less than 5% of

the cellulose and about 3% of the lignin.

The degradation of hemicelluloses results in the loss of the

biomass’s tenacious and flexible properties resulting in

a brittle material with poor mechanical properties. The heat-

ing value increases from about 10 to 17 MJ kg�1 to about

12e24.5 MJ kg�1 depending on species, torrefaction tempera-

ture and residence time [5,8,16]. Biomass torrefaction

processes have been reviewed in great detail recently by van

der Stelt et al. [5] and others [17e19].

Biomass densification has been an established process to

improve the handling properties of biomass for several

decades [4]. The raw materials are mainly softwood but also

hardwoods are being used and recently agricultural residues

i.e. straws and husks from various crops and pulps are gaining

interest. Straw is available in large quantities and a low value

by product of the agricultural industry. It is already used in

large scale heat and power plants [20] but the handling and

transportation of straw have remained an issue. Therefore

several studies have been made to investigate the pelletizing

properties of agricultural residues [21e23].

The pelletizing process consists out of multiple steps that

are shown in Fig. 1.

The working principle of a pelletizing process has been

explained in detail in one of our earlier studies [24]. In general

the process can be subdivided into different steps: compres-

sion, flow and friction component as shown in Fig. 2.

Every time the roller approaches the surface of the die, the

biomass is compressed and forms a temporary layer on the die

surface (compression component). The flow component

represents the energy required to force the compressed layer

into the press channels and the friction component stands for

the energy required to press the compressed saw dust in the

channels. The work required for the three components can be

determined experimentally for different raw materials and

process parameters. The resulting data provides and estima-

tion for the energy consumption of the pelletizing process that

can be used for process design and optimization [24].

Turning torrefied biomass into pellets has so far shown to

be a challenging step in completing the full production chain

for torrefied biomass pellets [6,25]. The problems are two-fold

in the sense that both good pellet quality and acceptable pellet

production capacities have been difficult to achieve. In terms

of quality, the pellet durability and density are often low in

comparison to conventional natural wood pellets, and high

energy consumption in the pellet mill causes lower produc-

tion capacities and high costs.

Fig. 2 e Different components of a biomass pelletization

process. Figure reproduced from Ref. [24] with permission

from Society of Wood Science and Technology.

b i om a s s an d b i o e n e r g y 4 9 ( 2 0 1 3 ) 2 1 4e2 2 1216

Using a standardized single pellet pressmethod has shown

that the pelletization of torrefied biomass requires more

energy due to the high friction between the torrefied biomass

and themetal surface of the press channel within a pelletmill.

Furthermore it was shown that the resulting pellets are less

stable compared to conventional wood pellets [6,7]. The lower

strength has both advantages and disadvantages. On the one

hand lower pellets strength increases the risk for fines and

dust formation during handling and transportation but on the

other hand it has been shown that torrefied pellets are easier

and less energy intense to grind into a fine powder (e.g. prior to

co-firing in a coal boiler) [26].

It has been shown that pelletization of torrefied biomass

requires higher energy and results in lower crushing strength

anddensityof thepellets compared tountreatedbiomass.These

problematic properties of torrefied material have been verified

using a production scale (500 kg h�1) pellet mill [25]. The small

scale trials using a single pellet press offer a number of possi-

bilities to test different parameters, which affect the torrefied

material’s behavior in the pelletizing process. The effect of die

temperature, pressure,moisture content, degree of torrefaction,

andalso the effect of additives can therefore bequickly screened

before scaling up the trials into medium and large scale.

Thepress cannot simulate all aspects of a “real life” pelletmill

but can give an indication about how different process param-

eters affect the product quality and friction (pressure) built up in

the press channel. In a production size pellet mill, some process

parameters are connected and dependent on each other; for

example, friction and temperature depend on the press channel

length. Other parameters such as the particle size or moisture

content are easier to adjust. The results obtained with a single

pellet press tool will become useful when setting up a pilot scale

test, i.e. allowingamoreefficient testing, sincemanyparameters

have been tested already in advance on a small scale.

The present study investigates the pelletizing properties of

torrefied wheat straw (Triticum aestivum L.) using a single

pellet press unit.

Wheat is among the most grown crop species in the world

with an estimated annual production of about 685 Mt per year

[27]. It is therefore considered to be a potential resource for

heat and power production. Nevertheless, the utilization of

straw is limited due to poor handling properties related to low

bulk densities, in the range of 40 kg m�3 for loose straw [28]

and about 120 kg m�3 for bales [29]. It is likely that the fuel

properties of straw can be improved by torrefaction and

subsequent pelletization. The present study investigates the

pelletizing properties of torrefied wheat straw using a single

pellet press test system.

2. Methods

2.1. Materials

Wheat straw (T. aestivum L.) was harvested in the summer

season 2011 on the island of Sjælland, Denmark and collected

after it has been dried on the field. The wheat straw was

chopped into particles <5 mm in diameter using a hammer

mill (Model 55, Jensen and Sommer Aps, Denmark). The

material was packed in paper bags permeable to air and

moisture and stored in a dry storage until use.

2.2. Torrefaction

Torrefaction was carried out in a bench scale batch reactor

that was designed and built similar as suggested by Pimchuai

et al. [30]. A metal box with a volume of about 2 L and two

openings (5 mm in diameter) for nitrogen inlet and gas outlet

was used as reactor. The box was installed in a programmable

muffle furnace (S90, Lyngbyoven, Denmark) and connected to

a nitrogen cylinder with pressure and flow regulator, water

seal valve and fittings and pipes for gas inlet, outlet and

temperature sensors. A temperature sensor (iron-constantan

thermocouple) was installed in the center of themetal reactor

and connected to a thermometer (52 kJ, John Fluke, USA) and

a computer system controlling the heating of the oven. The

material was torrefied at 150, 200, 250 and 300 �C for 2 h.

2.3. Fiber analysis

Fiber analysis method was based on the standard method used

by Sluiter et al. [31]. The moisture content of the raw materials

was determined using a moisture analyzer (MA 30, Sartorius,

Germany) at 105 �C. The samples were dried in vacuum at 40 �Cfor 2 days, and milled to <1 mm particle size. The composition

of the raw materials was analyzed using a two-step strong

acid hydrolysis. Monosaccharides (D-glucose, D-xylose and

L-arabinose) were quantified using high-performance liquid

chromatography (HPLC) (Summit system, Dionex, USA) equip-

ped with an RI-detector (Shimadzu, Japan). Klason lignin con-

tent was determined based on the filter cake, subtracting the

ash content after incinerating the residues from the strong acid

hydrolysis at 550 �C for 3h. For extractives analysis, about 6 gper

sample were soxhlet extracted in acetone for 3 h, and

Fig. 3 e Setup of a single pellet testing system.

b i om a s s a n d b i o e n e r g y 4 9 ( 2 0 1 3 ) 2 1 4e2 2 1 217

subsequent calculation of the dry matter loss. The moisture

absorption of the fibers at 25 �C and 65% relative humidity was

determined as described in one of our earlier studies [6]. The

higher heating value was analyzed in a bomb calorimeter (6800,

Parr InstrumentCompany,USA)asdescribedbyShangetal. [26].

2.4. Infrared spectroscopy

ATR-FTIR spectra of the fracture surface and raw material

surface were recorded as described earlier [6] using a Fourier

transform infrared spectrometer (Nicolet 6700 FT-IR, Thermo

Electron Corporation, USA), equipped with a temperature

adjustable ATR accessory (Smart Golden Gate diamond ATR

accessory, Thermo Electron Corporation, USA), and equili-

brated at 30 �C. Five measurements per sample were per-

formed. To ensure good contact with the diamond surface, all

hard, solid samples were pressed against it, using a metal rod

and consistent mechanical pressure. All spectra were ob-

tained with 200 scans for the background (air), 100 scans for

the sample and with a resolution of 4 cm�1.

2.5. Single pellet press

The biomass was pelletized using a single pellet press man-

ufactured and designed at the Danish Technological Institute,

Denmark. The units consisted of a cylindrical die made of

hardened steel, lagged with heating elements and thermal

insulation. The temperature was controllable between 20 and

180 �C, using a thermocouple (HSS braid coil, HT 30 regulator,

Horst GmbH) connected to a control unit. The end of the die

could be closed using a backstop. Force is applied to the press

using a universal testing system (AGX, Shimadzu, Japan) and

detected using a 200 kN loadcell connected to a detection

software (TrapeziumX, Shimadzu, Japan). The setup is shown

in Fig. 3 and its working principle is explained in great detail in

an earlier work by Nielsen et al. [24].

Data recorded where the compression force and friction

(force that is requited to extrude the pellet from the press

channel).

2.5.1. CompressionAn 8 mm press channel die was used at 125 �C temperature.

750 mg sample material was compressed at a rate of

100 mm min�1 until a maximum pressure of 300 MPa was

reached. The pressure was held for 10 s.

2.5.2. FrictionSubsequently to compression, the bottom piston was

removed from the die, and the pellet was pressed out of the

channel at a rate of 100 mm min�1. A value for the static

friction was determined from the force required to initiate the

pellet motion. The average value of at least 5 measurements

was determined and a 95% confidence interval was calculated.

2.6. Characterization of pellets

2.6.1. Strength and densityThe produced pellets were cylindrical with 8 mm diameter

and about 10e14 mm in length. The pellets were left to cool

overnight, and the pellet strength was made by crushing the

pellet perpendicular to the length of the pellet. The pellet was

laid down on the side and increasing force was applied until

crushing. Prior to this, the length, diameter, and mass were

measured for density calculation. Twelve replicates were

made for each sample. The error bars in the plots refer to the

95% confidence interval of the mean. The TukeyeKramer test

[32] was used to determine whether the observed differences

were significant.

3. Results and discussion

The fiber analysis (Table 2) indicates a chemical change during

torrefaction with increasing treatment temperature. The

cellulose content of the torrefied wheat straw samples

Table 2 e Fiber analysis of torrefied wheat straw (Data in g kgL1).

Sample Cellulose Hemicellulose Insoluble fraction Ash Yield [%DM]

Wheat straw 37.2 � 0.3 27.3 � 0.4 19.2 � 0.5 4.9 � 0.2 100

T150 36.4 � 0.4 27.1 � 0.5 19.1 � 0.6 5.5 � 0.3 99.7

T200 36.7 � 0.4 26.1 � 0.2 20.7 � 0.5 5.4 � 0.2 96.8

T250 37.5 � 0.2 5.8 � 0.1 47.2 � 0.3 6.7 � 0.4 74.7

T300 1.4 � 0.3 4.6 � 0.3 >80 10.7 � 0.4 45.9

b i om a s s an d b i o e n e r g y 4 9 ( 2 0 1 3 ) 2 1 4e2 2 1218

remain stable up to 250 �C but decreases dramatically when

increasing the torrefaction temperature to 300 �C. These

results match with earlier studies on wood torrefaction indi-

cating severe degradation of cellulose at high temperatures

between 250 and 300 �C [6,15,16,33e36]. Hemicelluloses are

more sensitive to thermal degradation than cellulose and

degradation starts at temperatures >200 �C. Di Blasi and

Lanzetta [35] have studied the thermal degradation of xylan,

one building block of hemicelluloses, and suggested a two-

step thermal degradation process starting at about 200 �C.The applied standard method for fiber analysis defines the

remaining acid insoluble fraction as Klasson lignin [31]. The

insolublemass fraction remains constant up to 200 �C at about

20% and the jumps to 80% and more at 300 �C. This can be

explained with different effects. First of all does the volatili-

zation of the carbohydrates results in a relative increase of the

remaining components i.e. insolubles and ash. Secondly

carbohydrates undergo scission reactions at elevated

temperatures resulting in acid insoluble solids and thus an

increase of the insoluble fraction [35].

The data from fiber analysis was supplemented with

infrared spectra obtained from an ATR-FTIR spectrometer

(Fig. 4). The spectra show the composition of the first few mm

of the straw particle surface. The OH-stretching vibration

region at 3600e3200 cm�1 gives an indication about the

amount of hydroxyl groups and as such potential bonding

sites for water. The intensity of the OH stretching vibration

decreases with increasing torrefaction temperatures which

could be an indication that the concentration of OH groups

decreases with cellulose and hemicelluloses dehydration at

increasing temperatures. The double peak at about 2850 and

2920 cm�1 has earlier been assigned to CH-stretching vibra-

tions of plant waxes [37]. Against expectations these waxes

remain relatively stable to thermal degradation and are still

present at elevated temperatures. Major spectral changes

Fig. 4 e ATR-FTIR spectra of straw surface before and after

torrefaction at different temperatures.

occur between 1750 and 1600 cm�1 (carbonyl region) and this

has earlier been observed by Gonzalez-Pena and Hale [38] who

studied torrefied wood. Spectral changes were explained with

a general increase of absorption in the carbonyl region due to

lignin condensation, carboxylation of polysaccharides and

opening of aromatic rings.

The thermal degradation of hemicelluloses and cellulose

has a great effect on the ability to absorb water. An increasing

torrefaction temperature results in decreased moisture

absorption of the wheat straw samples (Fig. 5).

This is likely due to the thermal degradation of hemi-

celluloses and cellulose and thus a smaller amount of avail-

able bonding sites for water molecules [33]. However Hakkou

et al. [39] have studied the wettability of torrefied beech wood

and studied its thermal degradation by means of nuclear

magnetic resonance spectroscopy and FTIR. They concluded

that the decreased moisture absorption rose from conforma-

tional changes of the wood during torrefaction and lignin

plasticization. The degradation and volatilization of the

biomass constituents result in an increasing energy density of

the remaining biomass, and as such an increase of the heating

value as shown in Fig. 6. The corresponding mass yield is

shown in Table 2. The energy density increases by about 40%

when comparing the untreated biomass with the straw tor-

refied at 300 �C.To increase the density of the biomass further they are

pelletized using a single pellet press and characterized

according to their density and mechanical stability. All

materials, apart from the wheat straw torrefied at 300 �C,result in compact and homogeneous pellets as shown in Fig. 7.

The material torrefied at 300 �C has significantly different

pelletizing properties which is shown in Figs. 8e10. The fric-

tion (Fig. 8) is increasing moderately when increasing the

Fig. 5 e Moisture absorption of torrefied wheat straw.

Fig. 6 e Heating value of torrefied wheat straw.

Fig. 8 e Friction generated in the press channel.

Fig. 9 e Compression strength of pellets.

b i om a s s a n d b i o e n e r g y 4 9 ( 2 0 1 3 ) 2 1 4e2 2 1 219

torrefaction temperature up to 250 �C. There is a big difference

between material torrefied at 250 and 300 �C. The friction

increases by factor 10e30 and the recorded data from repeated

trials showed a wide spread, indicated by the error bar (95%

confidence interval). The wide spread is a sign for an instable

pelletizing behavior. This observation is also reflected by the

compression strength data (Fig. 9) and pellet unit density

(Fig. 10). The material that has been torrefied at 300 �C, doesnot form stable pellets and disintegrates when pressed out of

the pellet press, resulting in a product of poor mechanical

stability and low density. In general it can be said that the

pellet compression strength tends to decrease with increasing

torrefaction temperature, as already observed forwood pellets

in earlier studies [6,7]. The lower compression strength is

beneficial when it comes to crushing and milling of pellets

prior to firing, since conventional coal mills might be utilized,

but certain strength is needed to prevent dust and fines

formation during handling and transportation. Earlier studies

have shown that torrefied biomass as such has favorable

grinding characteristics [40,41] and this is also valid for when

wood pellets are used as feedstock for torrefaction processes

[26]. The strong decrease of pellet unit density for samples

Fig. 7 e Pellets made from torrefied wheat straw.

torrefied at 300 �C (Fig. 10) can be explained with the poor

inter-particle bonding between the single biomass particles as

seen in Fig. 7, and that result in a phenomenon also known as

spring back effect [42e44]. This phenomenon has also been

observed for pellets made from torrefied spruce [6].

Looking closer at the pellet surface after it has been frac-

tured (Fig. 11) reveals some important differences between

pellets made from wheat straw and torrefied wheat straw.

Fig. 11a shows a flat surface, most likely the smooth cuticle of

Fig. 10 e Unit density of pellets.

Fig. 11 e Fracture surface of pellets made from a) wheat straw and b) wheat straw torrefied at 300 �C.

b i om a s s an d b i o e n e r g y 4 9 ( 2 0 1 3 ) 2 1 4e2 2 1220

a wheat straw leaf while Fig. 11b shows a flaking, brittle

surface The harsh torrefaction conditions have made the

material brittle and less flexible which makes it crack and

disintegrate into smaller particles during pelletization. The

relatively flat surface of both samples is indicating weak

interaction forces between the individual biomass particles

and explains also why the compression strength of both,

straw pellets and torrefied straw pellets, is much lower as it

has been observed for wood pellets produced and tested under

the same conditions [42].

4. Conclusion

Fiber analysis has shown that wheat straws polymers are

degrading during torrefaction. The thermal degradation of

hemicelluloses starts at about 200 �C and at about 250 �C for

cellulose. Those results were confirmed by infrared spectros-

copy that also indicated structural changes and degradation

for lignin at high torrefaction temperature. The thermally

induced chemical and physical changes of the wheat straw

result in a greater hydrophobicity that is reflected in reduced

moisture absorption of the biomass after torrefaction. The

mass specific calorific value of the wheat straw increases with

torrefaction due to the degradation and evaporation of low

molecular weight compounds. Nevertheless the mass loss at

high torrefaction temperature and 2 h torrefaction time is

unacceptable high. Good pelletizing properties are obtained

for torrefied biomass heated up to 250 �C. Above that

temperature the mechanical properties of the pressed pellets

are poor which is also reflected in a low unit density. The

severe polymer degradation at high temperatures result also

in a high friction between biomass and press channel surface

resulting in high pelletizing pressures that will most likely

result in high energy uptake when up-scaling the process to a

pellet mill. Up to 250 �C torrefaction temperature the pellet-

izing process results in mechanically strong pellets with

increased properties i.e. higher heating value and reduced

moisture adsorption. The study focused on a broad tempera-

ture range between 150 and 300 �C and from the obtained data

it can be seen that most changes occur in the range between

250 and 300 �C. Future studies should therefore focus on the

temperature range between 250 and 300 �C to identify

optimum conditions for the torrefaction and pelletization of

wheat straw.

r e f e r e n c e s

[1] Tolon-Becerra A, Lastra-Bravo X, Bienvenido-Barcena F.Proposal for territorial distribution of the EU 2020 politicalrenewable energy goal. Renew Energ 2011;36(8):2067e77.

[2] Neville A. Biomass cofiring: a promising new generationoption. Power 2011;155(4):52e6.

[3] Rentizelas AA, Tolis AJ, Tatsiopoulos IP. Logistics issues ofbiomass: the storage problem and the multi-biomass supplychain. Renew Sust Energ Rev 2009;13(4):887e94.

[4] Tumuluru JS, Wright CT, Hess JR, Kenney KL. A review ofbiomass densification systems to develop uniform feedstockcommodities for bioenergy application. Biofuels BioprodBiorefin 2011;5(6):683e707.

[5] van der Stelt MJC, Gerhauser H, Kiel JHA, Ptasinski KJ.Biomass upgrading by torrefaction for the production ofbiofuels: a review. Biomass Bioenerg 2011;35(9):3748e62.

[6] Stelte W, Clemons C, Holm JK, Sanadi RA, Shang L,Ahrenfeldt J, et al. Pelletizing properties of torrefied spruce.Biomass Bioenerg 2011;35(11):4690e8.

[7] Li H, Liu X, Legros R, Bi XT, Jim Lim C, Sokhansanj S.Pelletization of torrefied sawdust and properties of torrefiedpellets. Appl Energ 2012;93:680e5.

[8] Bergman PCA. Combined torrefaction and pelletizationetheTOP process. Petten, The Netherlands: Energy Centre of TheNetherlands; July 2005. p. 29 Report No.: ECN-C-05e073.

[9] Prins MJ, Ptasinski KJ, Janssen FJJG. From coal to biomassgasification: comparison of thermodynamic efficiency.Energy 2007;37:1248e59.

[10] Senior CL, Bool-Ill LE, Srinivasachar S, Pease BR, Porle K. Pilotscale study of trace element vaporization and condensationduring combustion of a pulverized sub-butuminous coal.Fuel Process Technol 2000;63:149e65.

[11] Atimtay A, Topal H. Co-combustion of olice cake with lignitecoal in a circulating fluidized bed. Fuel 2000;83:859e67.

[12] Lohrer C, Schmidt M, Krause U. A study on the influence ofliquid water and water vapour on the self-ignition of lignitecoal-experiments and numerical simulations. J Loss PreventProcess Ind 2005;18(3):167e77.

[13] Miller BG. Introduction to coal. In: Dorf RC, editor. Coalenergy systems. Sustainable world series. Waltham (MA):Academic Press; 2005. p. 1e26.

b i om a s s a n d b i o e n e r g y 4 9 ( 2 0 1 3 ) 2 1 4e2 2 1 221

[14] Li J, Brzdekiewicz A, Yang W, Blasiak W. Co-firing based onbiomass torrefaction in a pulverized coal boiler with aim of100% fuel switching. Appl Energ 2012;99:344e54.

[15] Chen W-H, Kuo P-H. Torrefaction and co-torrefactioncharacterization of hemicellulose, cellulose and lignin aswell as torrefaction of some basic constituents in biomass.Energy 2011;36(2):803e11.

[16] Rousset P, Macedo L, Commandre J-M, Moreira A. Biomasstorrefaction under different oxygen concentrations and itseffects on the composition of the solid by-product. J AnalAppl Pyrol 2012;96:86e91.

[17] Quing C, Zhou JS, Liu BJ, Mei QF, Luo ZY. Influence oftorrerfaction pretreatment on biomass gasificationtechgnology. Chin Sci Bull 2011;56(14):1449e56.

[18] Kim YH, Lee SM, Lee HW, Lee JW. Physical and chemicalcharacteristics of products from the torrefaction of yellowpoplar (Liriodendron tulipifera). Bioresour Technol 2012;116:120e5.

[19] Wang GJ, Luo YH, Deng J, Kuang JH, Zhang YL.Pretreatment of biomass by torrefaction. Chin Sci Bull2011;56(14):1442e8.

[20] Anonymous. Energistatistik 2011. Copenhagen: DanishEnergy Agency; September 2012. p. 59.

[21] Serrano C, Monedero E, Lapuerta M, Portero H. Effect ofmoisture content, particle size and pine addition on qualityparameters of barley straw pellets. Fuel Process Technol2011;92(3):699e706.

[22] Stelte W, Clemons C, Holm JK, Sanadi RA, Shang L,Ahrenfeldt, et al. Fuel pellets from wheat straw: the effect oflignin glass transition and surface waxes on pelletizingproperties. Bioenerg Res 2012;5(2):450e8.

[23] Theerarattananoon K, Xu F, Wilson J, Ballard R, McKinney L,Staggenborg S, et al. Physical properties of pellets made fromsorghum stalk, corn stover, wheat straw, and big bluestem.Ind Crop Prod 2011;33(2):325e32.

[24] Nielsen NPK, Gardner DJ, Poulsen T, Felby C. Importance oftemperature, moisture content, and species for theconversion process of wood residues into fuel pellets. WoodFiber Sci 2009;41(4):414e25.

[25] Larsson SHR M, Nordwaeger M, Olofson S. Effects of moisturecontent, torrefaction temperature, and die temperature inpilot scale pelletizing of torrefied Norway spruce. Appl Energ2013;102:827e32.

[26] Shang L, Nielsen NPK, Dahl J, Stelte W, Ahrenfeldt J, Holm JK,et al. Quality effects caused by torrefaction of pellets madefrom Scots pine. Fuel Process Technol 2012;101:23e8.

[27] FAOSTAT [online database]. Rome, Italy: Food AgriculturalOrganization of the United Nations [cited 2012 October 8]Available online from: http://faostat.fao.org; 2012. Filesupdated annually.

[28] Sokhansanj S, Mani S, Zaini P. Binderless pelletization ofbiomass. ASAE paper No 05e6061. St Joseph: AmericanSociety of Agricultural and Biological Engineers; 2005.

[29] Smith IE, Probert SD, Stokes RE, Hansford RJ. Briquetting ofwheat straw. J Agr Eng Res 1977;22(2):105e11.

[30] Pimchuai A, Dutta A, Basu P. Torrefaction of agricultureresidue to enhance combustible properties. Energ Fuel 2010;24:4638e45.

[31] Sluiter A, Hames B, Ruiz R, Scarlata C, Sluiter J, Templeton D,et al. Determination of structural carbohydrates and lignin inbiomass. Golden, Corlorado: National Renewable EnergyLaboratory; 2010 July. p. 17, Report No. TP-510e42618.

[32] Kramer CY. Extensions of multiple range tests to groupmeans with unequal numbers of replication. Biometrics1956;12(3):307e10.

[33] Yan W, Acharjee TC, Coronella CJ, Vasquez VR. Thermalpretreatment of lignocellulosic biomass. Environ ProgSustain Energ 2009;28(3):435e40.

[34] Prins MJ, Ptasinski KJ, Janssen FJJG. Torrefaction of wood e

part 1. Weight loss kinetics. J Anal Appl Pyrol 2006;77(1):28e34.

[35] Di Blasi C, Lanzetta M. Intrinsic kinetics of isothermal xylandegradation in inert atmosphere. J Anal Appl Pyrol 1997;40(1):287e303.

[36] Agrawal RK. Kinetics of reactions involved in pyrolysis ofcellulose. 1. The 3-reaction model. Can J Chem Eng 1988;66(3):403e12.

[37] Merk S, Blume A, Riederer M. Phase behaviour andcrystallinity of plant cuticular waxes studied by Fouriertransform infrared spectroscopy. Planta 1998;204(1):44e53.

[38] Gonzalez-Pena MM, Hale MDC. Colour in thermally modifiedwood of beech, Norway spruce and Scots pine. Part 1: colourevolution and colour changes. Holzforschung 2009;63(4):385e93.

[39] Hakkou M, Petrissans M, Zoulalian A, Gerardin P.Investigation of wood wettability changes during heattreatment on the basis of chemical analysis. Polym DegradStab 2005;89(1):1e5.

[40] Shang L, Ahrenfeldt J, Holm JK, Sanadi AR, Barsberg S,Thomsen T, et al. Changes of chemical and mechanicalbehavior of torrefied wheat straw. Biomass Bioenerg 2012;40:63e70.

[41] Arias B, Pevida C, Fermoso J, Plaza MG, Rubiera F, Pis JJ.Influence of torrefaction on the grindability and reactivity ofwoody biomass. Fuel Process Technol 2008;89(2):169e75.

[42] Mani S, Tabil LG, Sokhansanj S. Effects of compressive force,particle size and moisture content on mechanical propertiesof biomass pellets from grasses. Biomass Bioenerg 2006;30(7):648e54.

[43] Stelte W, Holm JK, Sanadi AR, Barsberg S, Ahrenfeldt J,Henriksen UB. A study of bonding and failure mechanisms infuel pellets from different biomass resources. BiomassBioenerg 2011;35(2):910e8.

[44] Kaliyan N, Morey RV. Factors affecting strength anddurability of densified biomass products. Biomass Bioenerg2009;33(3):337e59.