A review of torrefaction for bioenergy feedstock production

Correspondence to: Daniel Ciolkosz, The Pennsylvania State University Department of Agricultural and Biological Engineering, 249 Ag Engineering Building,

University Park, PA 16802, USA. E-mail: dec109@psu.edu

Review

317

A review of torrefaction
for bioenergy feedstock
production

Daniel Ciolkosz, The Pennsylvania State University, University Park, PA, USA

Robert Wallace, Booz Allen Hamilton, Pittsburgh, PA

Received October 4, 2010; revised version received December 2, 2010; accepted December 3, 2010

View online January 28, 2011 at Wiley Online Library (wileyonlinelibrary.com); DOI: 10.1002/bbb.275;

Biofuels, Bioprod. Bioref. 5:317–329 (2011)

Abstract: The torrefaction of biomass is a thermochemical decomposition process in which hemicellulose degrada-

tion is the dominant reaction, with the cellulose and lignin fractions largely unaffected. The primary product is a solid

material that retains 75–95% of the original energy content. Properties of the torrefi ed solid include improved grind-

ability, hydrophobicity, and energy density. Torrefi ed biomass has been processed successfully in batch-mode and

continuous process devices; net thermal effi ciencies of the process as high as 90% have been reported. Torrefi ed

biomass has been proposed as a feedstock for coal co-combustion, as well as for gasifi cation-combustion and

Fischer-Tropsch fuel production. Analyses of supply chain impacts indicate that, in some scenarios, torrefaction can

be the lowest cost and most energy effi cient option for supplying fuel, especially when combined with pelletization of

the material.

Signifi cant gaps still exist in our understanding of torrefaction; there is need to further study this important process

for its potential benefi ts to bioenergy production. Some of the more pressing needs include characterization of

chemical pathways of the torrefaction reaction, investigation of equipment performance and equipment-related

Wiley & Sons, Ltd

Keywords: torrefaction; thermochemical conversion; biomass; bioenergy; pyrolysis

Introduction

he thermochemical processing of biomass is the act

of exposing the material to elevated temperatures in

an oxygen-constrained environment, which leads

to thermally-activated breakdown of lignocellulosic mate-

rial without the oxidation that occurs during combustion.

e predominant product of the process, a gas, liquid, or

solid, can be selected by controlling the processing condi-

tions: pressure, temperature, and residence time. Th

e most

severe and complete form of thermochemical conversion

is known as gasiﬁ cation, and results primarily in gaseous

products (mostly carbon monoxide, hydrogen, water vapor,

and methane). Th

e least severe form of thermochemical

318

D Cielkosz, R Wallace

Review: Torrefaction for bioenergy feedstock production

processing, known as torrefaction, consists of processing

biomass in a relatively low temperature, inert gas environ-

ment (200–300°C) for a duration of generally 1 h or less at

or close to atmospheric pressure. Th

e torrefaction process

is oft en referred to as a ‘roasting process’ (‘torréfaction’,

translated from French, means ‘roasting’), and the product

is comparable to a low-grade charcoal with diﬀ erent proper-

ties from the raw biomass, including increased stability and

reduced susceptibility to microbial degradation, improved

hydrophobic properties, and a higher carbon density than

raw biomass.

Torreﬁ ed biomass has been proposed as a suitable feed-

stock for coal co-combustion, gasiﬁ cation and thermochem-

ical fuel production (including Fischer-Tropsch processes)

due to its high energy content, grindability, and hydrophobic

properties,

4,5

and may be suitable as a feedstock for bio-

chemical processing as well.

e earliest known studies of

torrefaction were carried out in France in the 1930s, for the

generation of syngas, and again in the 1980s as studies inves-

tigated its potential for metallurgical processes.

1, reported in 2

Recently, torrefaction has come under increased scrutiny

as a possible means of improving the suitability of biomass

as a feedstock for power plants or bioprocessing facilities.

As a result, many studies have been undertaken recently to

analyze and assess torrefaction mechanisms and properties.

Most studies have consisted of analysis of the products of

torrefaction for diﬀ erent feedstocks and process conditions.

Laboratory-scale reactors are the most common device uti-

lized, and some Th

ermogravimetric Analysis (TGA) has also

been carried out to assess kinetics of the reaction.

In addi-

tion, chemical modeling has also been reported as a comple-

ment to the experimental studies. Relatively little work has

been done on the life cycle cost and logistics of the process

or its impacts on the entire bioenergy supply chain, although

initial indications have been promising.

Physical processes and equipment

e torrefaction processing of biomass involves raising its

temperature to the desired level for a speciﬁ ed residence

time. Th

is is usually accomplished using convective heating

within a sealed chamber ﬁ lled with inert gas. Pre-treatment

of biomass prior to torrefaction usually consists of grinding

(grinder and/or hammermill) and/or drying. Post-treatment

of the biomass can include cooling and/or densiﬁ cation.

Oft en, densiﬁ cation (pelletizing or briquetting) is used to

improve the handling and transportation characteristics of

the material.

Torrefaction equipment can be designed for either batch

processing or continuous processing, and both approaches

have been used for laboratory-scale investigations. Steam

is oft en used as the heat transfer medium for the reaction

vessel, although dry roasting is also utilized, sometimes

using hot combustion gases as the heat transfer medium.

Pilot-scale devices have included screw reactors and tray

ovens,

8–10

(Fig. 1), but relatively little commercial-scale

production has occurred to date. Careful characterization

of scale-related eﬀ ects and commercial torrefaction system

performance is a signiﬁ cant need as the industry develops.

Some studies have suggested utilizing microwave radiation

to heat large chunks of wet biomass – allowing for uniform

processing without the need to grind the feedstock prior to

roasting.

11,12

Alternately, ‘wet processing’ of biomass under

high pressure conditions allows for torrefaction without ﬁ rst

drying the feedstock.

13–16

Torrefaction chemistry

Chemical analyses of torreﬁ ed biomass suggest that the tor-

refaction process is dominated by the thermal activation

and depolymerization of hemicellulose molecules within

the biomass.

Hemicellulose soft ens at temperatures of

150–200°C and undergoes dehydration, deacetylization, and

depolymerization reactions at processing temperatures in

the 200–300°C range.

18,19

Xylan (Fig. 2) is the predominant

Figure 1. Diagram of pilot-scale torrefaction device 10. Dashed

arrows denote feedstock fl ow.

319

Review: Torrefaction for bioenergy feedstock production

D Cielkosz, R Wallace

form of hemicellulose in deciduous wood and herbaceous

ﬁ eld crop residues, and is a pentose polysaccharide consist-

ing of D-Xylose units with 1-b -4 linkage. acetoxy- and meth-

oxy- groups are attached at intervals to the xylose units.

Glucomannan, the predominant form of hemicellulose in

conifer wood, is a polysaccharide of glucose and mannose

units in a ratio of 1.0G : 1.6M (Fig. 3). Th

e typical repeating

unit of the polymer is MMGGMGGMMGMMM. Acetate

groups occur every 9 to 19 units and monosaccharide side

chains occur every 50–60 units.

Xylan tends to break down more quickly and at lower

temperatures than glucomannan, and consequently biomass

samples that are higher in xylan content tend to break down

more rapidly.

21,22

Major reaction pathways that have been

proposed for torrefaction include dehydration reactions to

form water and solid ‘torreﬁ ed biomass’, deacetylization, and

depolymerization, leading to the formation of levoglucosan.

e products of lignocellulosic biomass torrefaction are

approximately 70–90% solids, 6–35% liquid, and 1–10% gas

(on a mass basis). Th

e solid fraction, known as ‘char’, is usu-

ally the quantity of interest. Increasing the severity of the

reaction (which can be thought of roughly as the product of

temperature and duration) increases the relative yield of gas

and liquid.

Several studies suggest that a small degree of cellulose

and/or lignin degradation also occurs during torrefaction.

Torrefaction at higher temperature conditions (>~270°C) is

reported to initiate a greater degree of cellulose breakdown.

Lignin, which soft ens at temperatures as low as 80–90°C, has

not been found to undergo signiﬁ cant chemical alteration

during torrefaction.

TGA of biomass during the torrefaction process suggests

a dominant two-step reaction. One possible explanation of

the two steps would be diﬀ ering reaction kinetics of the two

main types of hemicellulose in biomass.

Others have sug-

gested that the reaction involves the generation of intermedi-

ate compounds.

Lipinsky et al. suggest that hemicellulose

breaks down to ‘reactive hemicellulose’, from which point

it decomposes and recombines to form a variety of sub-

stances.

e composition and quantity of these substances

have not been carefully examined as of yet. It is interesting

to note that several studies indicate that the absolute amount

of ﬁ xed carbon in a sample (non-volatilized carbon at

950°C) increases as the torrefaction processing temperature

increases.

2,27

is suggests that higher temperature torrefac-

tion processing transforms hemicelluloses into compounds

that have greater thermal stability.

Several mathematical models have been utilized to simu-

late torrefaction, with generally good success.

28–31

However,

simpler models appear to be suitable and more easily imple-

mented at this state.

Our analysis mass yield data from

multiple studies indicates that the percent mass yield of

Figure 2. Xylan molecule (Haworth), showing D-xylose units with 1-

b-4 linkage.

Figure 3. Section (GMMG) of glucomannan molecule (Haworth) with 1-

b-4 linkage

(side chains of polysaccharides or acetate not shown).

320

D Cielkosz, R Wallace

Review: Torrefaction for bioenergy feedstock production

recovered char solids (on a dry, ash-free basis) generally ﬁ ts

an Arrhenius-type relationship (Fig. 4) as follows:

100 * Mt

= K1 + (100-K1) e-

(K2t)/(RT)

Where Mt = mass of solids at time t (g dry solids, dry, ash free)

Mo = original mass of solids (g dry solids, dry, ash free)

t = processing time (minutes)

T = processing temperature (Kelvin)

R = universal gas constant, 8.314

−3

(J K

−1

mol

−1

)

K1, K2 = reaction coeﬃ

cients (percent, J mol

−1

min

−1

)

Practically speaking, the coeﬃ

cient K1 corresponds to

the mass yield of solid material (char) at an inﬁ nitely long

processing time (as a percentage of the initial mass). Our

analysis of data from four separate studies suggests that the

value of K1 may be related to the processing temperature,

possibly in a linear fashion (Fig. 5).

e coeﬃ

cient K2 corresponds to the activation energy

of the overall reaction, and as such can be expected to cor-

respond to hemicellulose composition in the sample, with

lower values of K2 for samples with a higher xylan concen-

tration. At least one study did ﬁ nd a trend corresponding to

feedstock xylan concentration.

However, our analysis of 13

feedstocks from 5 separate studies

32–34

found values of K2

ranging from 1.3

× 10

−3

to 1.4

× 10

−2

, with no obvious trend

associated with feedstock type. It may be that results are

impacted by the type of experimental setup used in the vari-

ous studies – perhaps caused by diﬀ ering heat transfer rates

Figure 4. Example measured and modeled dry mass of torrefi ed sugarcane bagasse –

data from Pach et al.

Figure 5. Coeffi cient K1 as a function of processing temperature.

321

Review: Torrefaction for bioenergy feedstock production

D Cielkosz, R Wallace

from varying heating equipment and diﬀ erent particle sizes,

causing changes to reaction kinetics.

e solid fraction resulting from torrefaction presum-

ably consists of the unreacted cellulose, unreacted lignin,

and non-volatile byproducts of hemicellulose degradation.

However, the composition of torreﬁ ed biomass has not been

conclusively characterized as yet. Solids from torrefaction

are composed primarily of C and O, with smaller amounts

of N and H (Table 1). As the severity of the treatment

increases, the relative amount of oxygen in the solid frac-

tion decreases. Measurements of pH indicate that, at higher

processing temperatures, the pH of the torreﬁ ed biomass

increases.

e liquid fraction, which is condensed from exhaust gases

during torrefaction, consists primarily of water, acids, meth-

anol, furfural, hydroxyacetone, and phenol. Water vapor

forms as a result of a dehydration reaction that breaks down

the hemicellulose. Th

e presence of xylan in the feedstock is

believed to lead to acetic acid formation, while glucomannan

leads to formic acid production. Th

e presence of water dur-

ing the reaction is believed to be beneﬁ cial for the depolym-

erization reactions, and the presence of acids can increase

the degree of cellulose degradation.

e gaseous fraction contains primarily CO

and CO,

with traces of O

and C

. As the severity of the proc-

ess increases, the total amount of gas production increases

and the relative amount of CO increases. Th

e production

of CO

is believed to be a byproduct of decarboxylation of

acid groups in the wood. Th

e source of CO is not as read-

ily apparent, but may be due to a secondary ‘water gas shift

Table 1. Typical yield of various torrefied biomass samples, 60 min process t ime.

Material

Temp

(°C)

Gas %

Liq %

Solid

Solids composition

Energy

retained

Ref.

Pine

230

0.6

92.4

49.7

5.9

44.3

96.5

Pine

250

10.8

88.2

50.9

5.8

43.2

94.4

Pine

280

2.1

19.8

78.1

56.4

5.5

38.0

93.9

Bagasse

230

2.6

9.9

87.5

48.6

5.6

45.5

96.4

Bagasse

250

10.4

10.7

78.9

50.6

5.6

43.5

92.0

Bagasse

280

12.9

18.5

68.6

52.8

5.3

41.5

82.9

Birch

250

1.7

12.8

85.5

51.5

5.8

42.5

97.9

Birch

230

0.8

93.2

48.2

5.9

45.7

93.8

Birch

250

1.2

10.8

49.5

5.7

44.7

90.0

Birch

280

51.3

5.6

43.0

84.3

Salix

230

45.6

5.9

48.2

94.4

Salix

250

1.5

85.5

45.8

5.8

48.1

88.4

Salix

280

46.3

5.6

47.7

81.8

Miscanthus

230

44.4

6.1

48.7

87.7

Miscanthus

250

47.4

5.8

46.1

87.7

Miscanthus

280

51.3

5.7

42.4

80.0

Straw pellets

230

0.1

47.8

6.3

45.2

95.1

Straw pellets

250

0.3

9.8

49.0

6.1

44.1

91.6

Straw pellets

280

19.1

79.9

52.8

6.1

40.3

89.8

Wood pellets

230

0.06

3.5

96.5

49.8

6.3

43.8

97.5

Wood pellets

250

0.15

5.5

94.4

50.7

6.2

43.0

96.9

Wood pellets

280

0.6

89.4

52.5

6.2

41.3

96.0

Notes: %E = energy content (higher heat value), as a % of original feedstock value. Solids composition is on a dry ash-free basis.

322

D Cielkosz, R Wallace

Review: Torrefaction for bioenergy feedstock production

reaction’ occurring during torrefaction, or to other reactions

that are catalyzed by the presence of transition metals found

in biomass ash.

Particle size is believed to be a factor in torrefaction reac-

tions, due to varying conditions experienced by reaction

intermediates when larger particles of biomass are proc-

essed. One study

found that particle sizes of less than

5 mm did not aﬀ ect the reaction, and most studies to date

have dealt with torrefaction of ground material and have

not examined the complicating factors introduced by the

torrefaction of larger biomass particles. Most studies of

torrefaction have also utilized dried (<10% MC) samples.

However, the torrefaction processing of higher moisture-

content biomass has been found to increase mass loss of the

sample, perhaps due to fracturing of the biomass by escap-

ing steam.

Torrefaction is an endothermic process below about 270°C,

and exothermic above – possibly due to exothermic break-

down of sugars at higher temperatures.

erefore, care must

be taken if processing at higher temperatures to prevent a

runaway reaction.

However, the magnitude of the heat of

reaction is reported to be relatively small.

Claims have been made as to the uniformity of the biomass

being improved by torrefaction.

is may be true of vari-

ation of a single feedstock type, but does not appear to be

the case when comparing several feedstocks. Our analysis of

eight feedstocks, all processed at the same temperature

(250°C) and duration (60 min) does not show a large

decrease in variability of elemental composition or energy

content, although ash content is slightly more uniform

(Table 2). A more detailed study of this issue is needed,

including an assessment of the molecular form of the bio-

mass pre- and post-treatment.

Physical properties of torrefi ed biomass

Torreﬁ ed biomass tends to be brown to black in color, with

an appearance that is otherwise similar to that of the origi-

nal feedstock. However, individual particles of feedstock

tend to be somewhat rounded relative to their original

shape,

suggesting that torreﬁ ed biomass may have better

ﬂ owability characteristics than unprocessed material. Other

physical properties of importance include density, compress-

ibility, grindability, and hydrophobicity.

e density of torreﬁ ed biomass is measured in terms of its

bulk density as well as its energy density, where energy den-

sity is the energy content per unit mass.

Bulk density of biomass does not appear to change appreci-

ably during torrefaction.

However, our analysis indicates

that the energy density does increase noticeably – resulting

in energy densities generally ranging from 102% to 120% of

the original (Fig. 6). Th

is is presumably due to the decrease

in the oxygen content relative to the mass of carbon. Th

increase is greatest for more severe processing conditions –

high temperatures and/or residence times.

Casual observation suggests that torreﬁ ed biomass is much

more prone to aerial dispersion than untreated biomass. Th

may pose either a respiratory or combustion hazard, requir-

ing appropriate safety precautions. A solution to this prob-

lem is to densify the biomass aft er torrefaction. Densifying

also renders the product more suitable for handling and

Figure 6. Survey of average energy density (% relative to untreated) of

torrefi ed biomass as a function of processing time and temperature

(standard deviation varies from 1.5 to 10.2%)

2,19,32,34,49,57

Table 2. Coefficient of variation of biomass
properties before and after torrefaction.

Property

CV Pre-

treatment

CV Post-

treatment

C yield (mass %)

0.038

0.040

H yield (mass %)

0.044

0.035

O yield (mass %)

0.038

0.042

Ash content (mass %)

1.453

1.029

Energy content (MJ kG

−1

)

0.067

0.058

323

Review: Torrefaction for bioenergy feedstock production

D Cielkosz, R Wallace

for transport. Studies of densiﬁ cation of torreﬁ ed biomass

suggest both positive and negative impacts from the proc-

ess. Pelletization of torreﬁ ed biomass utilizes less energy

than raw biomass pelletization, with energy requirements

reduced by as much as 50%; pellet strength is also notice-

ably reduced.

40,41

Both eﬀ ects can be attributed to the loss

of structural integrity from the breakdown of the hemicel-

lulose. Simultaneous torrefaction and pelletization has been

proposed as a possible approach that could reduce overall

energy use for generating a densiﬁ ed, torreﬁ ed product.

8,42

Since hemicellulose creates structural linkages within ligno-

cellulosic material, it follows that its breakdown during torre-

faction would yield a material with lower strength and easier

grindability. Reported measurements of grindability have

been either in terms of the Hardgrove Grindability Test,

4,43

or measurements of the energy use for grinding the torre-

ﬁ ed material with a laboratory mill.

24,44,45

Most experiments

report that the energy requirement for grinding torreﬁ ed bio-

mass is between 10 and 20% of the amount required for raw

biomass – a very large reduction. However, laboratory-scale

grinding equipment can be very diﬀ erent from commercial

devices, and there is a need to assess the grinding perform-

ance of torreﬁ ed biomass at pilot or commercial scales.

Researchers have noted an increase in the hydrophobic

nature of the material, as indicated by lower equilibrium

moisture levels in ambient conditions. Th

is phenom-

enon is due presumably to a combination of the following

factors:

5,35,46,47

1) Th

e breakdown of hemicellulose unbinds the cellulose

and lignin, allowing the last water molecules not stored

at the cell level to be released.

2) Th

e deconstruction of hemicellulose leads to a greater

brittleness for cellulose and lignin, also lending to its

hydrophobic nature.

3) Th

e removal of OH groups from the hemicellulose

reduces the feedstocks’ ability to form hydrogen bonds

with water.

4) Th

e non-polar molecules that result from the breakdown

of hemicellulose tend to be hydrophobic – this inciden-

tally aids in the resistance to biodegradation.

Quantiﬁ cation of this property, especially under condi-

tions of repeated wetting over extended periods of time,

should be a priority for future study given its relevance for

outdoor storage and transport.

e ideal torreﬁ ed product would have minimal energy

loss while exhibiting improved grindability and energy den-

sity. Generally, conditions that favor improved energy den-

sity result in greater mass loss and therefore reduced energy

retention. An ideal torreﬁ ed fuel will need to optimize

these two competing characteristics. Bergman et al. sug-

gests that the optimum conditions for torrefaction consist

of high temperatures and low processing times – resulting

in material with good grindability, low processing cost, and

high energy content.

However, not all studies agree in this

respect – especially in terms of energy yield of the torreﬁ ed

biomass. Figure 7 illustrates energy retention as a function

of processing time and temperature from six studies we

reviewed. It may be that the optimum processing conditions

vary according to the type of equipment or the feedstock.

Additional pilot- or full-scale studies could provide impor-

tant information in this regard.

Torrefaction energy balance

e energy balance of torrefaction is dependent on the char-

acteristics and performance of both the equipment and the

feedstock. Since torrefaction is a nascent industry, actual

energy balances for industrial-scale operations are not gen-

erally available. However, some general approximations and

Figure 7. Survey of energy content retained (%) vs processing

time and temperature (standard deviation varies from

2.9 to 26.3%)

2,19,32,34,48,56

324

D Cielkosz, R Wallace

Review: Torrefaction for bioenergy feedstock production

conclusions can be drawn. Energy needs for torrefaction can

be generally categorized as pre-treatment energy, process

energy, or parasitic loads (Fig. 8).

Pre-treatment usually consists of chipping and or grind-

ing, and drying. Energy requirements for these steps can be

highly variable – some typical values are given in Table 3.

Process energy is the energy provided to drive the torrefac-

tion process, equal to the enthalpy of the products leaving the

device minus the enthalpy of all materials entering the device

(i.e. feedstock, moisture, inert gases) plus any additional

heat needed to compensate for losses in the system. Process

energy can be provided internally (from combustion of tor-

refaction gases) or externally (from fossil fuels or other heat

sources). Parasitic loads are those energy inputs required to

operate the torrefaction equipment (fans, pumps, etc.), and

can be a signiﬁ cant contributor to the overall energy balance.

ermal losses are the heat losses (primarily convection, con-

duction and latent) from the torrefaction device. At present,

values for the energy balance are generally based on models

or laboratory-scale experiments, and very little is known

about the commercial-scale energy balance of the torrefac-

tion process.

e energy eﬃ

ciency of the torrefaction process is reported

in several ways by diﬀ erent researchers, which can lead to

confusion when comparing results. Probably the most com-

mon measure in use is the net thermal process eﬃ

ciency

– the ratio between the energy yield in the product and the

total energy (feedstock plus process) input:

= 100 * Q yield/ (Q process ext + Q pre-treatment + Q

parasitic + Q feedstock)

Where η

= net thermal eﬃ

ciency (%)

Q yield = energy content of torreﬁ ed biomass solids (kJ/kg of

dry feedstock)

Q process ext = process energy provided from external

sources (kJ/kg of dry feedstock)

Q pre-treatment = energy required for pre-treatment of feed-

stock (kJ/kg of dry feedstock)

Q parasitic = parasitic load of torrefaction process (kJ/kg of

dry feedstock)

Q feedstock = gross energy content of raw feedstock (kJ/kg

of dry feedstock)

Note that all energy values are expressed ‘per dry kg of

feedstock’. It is important to keep this basis consistent for

the calculation. Also, some researchers utilize the Higher

Heating Value (HHV) of the feedstock and product, whereas

some use the Lower Heating Value (LHV). Th

e HHV is

probably the more appropriate value to use for this calcula-

tion, since it gives the total (gross) heat available from the

material.

Some studies have suggested that net thermal process

eﬃ

ciencies of over 90% can be obtained commercially, but

this is probably only possible for dry feedstocks that require

minimal pre-processing. More likely scenarios for torrefac-

tion would have a process eﬃ

ciency of 80% or lower – addi-

tional studies are needed to address this issue.

ermal process eﬃ

ciency can be increased by increasing

the use of torrefaction gases and liquids as an energy source

for process heat, or by selecting processing conditions that

maximize the energy yield of the torreﬁ ed material. Gaseous

and liquid products of torrefaction, containing 10–30%

Figure 8. Energy balance of torrefaction process, assuming

isenthalpic reaction.

Table 3. Typical pre-processing energy
requirements.

Property

Range of
values

Units

References

Chipping of
wood

180–2360

kJ per kg of
wood

Grinding

270–450

kJ per kg of
feedstock

Drying of
green wood

3000–9000

kJ per kg water
removed

50–53

325

Review: Torrefaction for bioenergy feedstock production

D Cielkosz, R Wallace

of the original energy value of the feedstock, can be used

to provide process heat to control reaction temperatures

or to dry incoming feedstock. About 2% of the biomass

feedstock’s energy is needed to heat ‘dry’ (5% moisture,

wet basis) biomass to a processing temperature of 250°C.

Feedstock at 20% moisture (wet basis) would require about

4% of the feedstock’s energy, and 50% moisture (wet basis)

– typical for green wood – would require about 15% of the

feedstock’s energy for heating and drying. Pre-drying the

feedstock before it enters the torrefaction chamber can

reduce overall energy requirements, as can the recovery of

heat from the processed feedstock. Maximizing the energy

yield, by selecting processing conditions that maximize the

amount of energy retained in the feedstock, is important

both for improving process eﬃ

ciency and for minimizing

the amount of biomass feedstock needed for the process.

Parasitic loads for torrefaction can be minimized through

careful engineering design of equipment and processes.

ese loads are oft en neglected in analyses, although their

magnitude can be a large portion of the overall energy

requirements.

Torrefaction emissions

Emissions from the torrefaction process consist of the gase-

ous and volatile products of the process. If the gaseous and

liquid products of the process are captured and combusted,

the remaining emissions proﬁ le can be expected to consist

primarily of CO

, H

O and particulates. NOx emissions

should be negligible due to the low processing temperature,

and SOx emissions should be negligible due to the extremely

low levels of sulfur in most lignocellulosic biomass.

However, study of this topic is still needed, and it is uncer-

tain what magnitude and nature of particulate emissions can

be expected from this process.

Torrefaction economics

While torrefaction does increase the speciﬁ c energy density

of the solid and improve its hydrophobic properties, there

is some question as to whether or not torrefaction provides

a net beneﬁ t to the bioenergy value chain. Unprocessed

biomass, depending on the fuel type and application, may

perform as well as torreﬁ ed biomass in some applications,

without the added processing cost. In order to assess the net

beneﬁ t of torrefaction, the impact of the process on all stages

of the value chain must be carefully accounted. In essence,

the question to be asked is whether or not the cost of tor-

refaction can be compensated by reduced costs or increased

performance in other portions of the supply chain.

e segments of the supply chain that are most likely to

beneﬁ t from torrefaction are transport, storage, and conver-

sion or utilization, whereas the torrefaction process (and

associated densiﬁ cation) will add to overall costs. Th

e trans-

portation step beneﬁ ts from the higher energy density of the

fuel, allowing for reduced costs per joule of fuel transported.

However, beneﬁ ts will probably be realized only if the torre-

ﬁ ed biomass is densiﬁ ed into an easily handled pellet or bri-

quette. Otherwise, the torreﬁ ed biomass is likely to require

specialized handling and transportation equipment that will

add to the transportation costs.

While it has not been conclusively demonstrated, it is

widely claimed that torreﬁ ed biomass, by its virtue of being

hydrophobic, can be successfully stored outdoors, thus

obviating the need for an enclosed storage bin or building.

However, it should be noted that, in dry climates, wood

chips have been successfully stored in large outdoor piles.

e relative fuel losses (shrinkage) during storage are not

well known, but can be expected to be higher for outdoor

storage. Comparisons of shrinkage losses of torreﬁ ed vs

raw biomass are needed for diﬀ erent storage conditions and

climates.

Utilization beneﬁ ts are related to the higher energy con-

tent, lower oxygen content, and (probable) lower moisture

content, relative to unprocessed biomass. Torreﬁ ed biomass

is expected to perform as well or better than raw biomass

for many bioenergy applications, including combustion,

gasiﬁ cation, and fuel production applications.

is may

not be true for biochemical processing, where the loss of

hemicellulose and increased hydrophobicity may reduce

conversion rates. Enhanced conversion and utilization,

when compared to the other steps in the supply chain,

probably provide the most signiﬁ cant opportunity for cost

savings (followed by transport costs). Torreﬁ ed biomass is

believed to be a superior solid fuel for combustion, espe-

cially when co-ﬁ red with coal due to its higher energy

density and coal-like handling properties.

5, 19, 55

Torreﬁ ed

326

D Cielkosz, R Wallace

Review: Torrefaction for bioenergy feedstock production

Figure 9. Delivery costs of pelletized biomass.

Tor = torrefaction, Pel = pelletization,

TOP = integrated torrefaction and pelletization. Numbers indicate nominal capacity of

system (dry kilotonnes of raw biomass feedstock per year).

biomass is also expected to provide advantages as a fuel for

thermochemical processing, due to the removal of acids and

oxygen.

Gasiﬁ cation using torreﬁ ed biomass allows for

improved ﬂ ow properties of the feedstock, increased levels

of H

and CO in the resulting syngas, and improved overall

process eﬃ

ciencies.

54,57

Techno-economic analysis (TEA) couples a physical model

of a device or process, with an economic analysis to yield

understanding of material ﬂ ow, eﬃ

ciency, and cost. As such,

it is the most appropriate method for assessing the overall

impact of these issues. However, few studies have been pub-

lished to date that examine the techno-economic feasibility

of torrefaction. Th

e most thoroughly analyzed supply chain

scenario involves the importation of biomass from South

America to Europe, and concludes that torrefaction com-

bined with pelletization provides a lower cost fuel for power

or fuel production when compared to pelletizing alone, with

cost savings ranging from 4% to 16%, depending on the end

use of the biomass.

Figure 9 shows supply chain costs for

several scales and processing options for biomass, indicat-

ing that pelletizing of torreﬁ ed biomass signiﬁ cantly reduces

costs, that larger-scale operations are more cost eﬃ

cient,

and that integrated torrefaction and pelletizing is less costly

than pelletizing alone. Zwart et al. conclude that, while tor-

refaction is one of the most cost-eﬀ ective options for supply

of overseas biomass, modiﬁ cations to the supply chain, such

as the centralized processing of raw feedstock, can result in

similar reductions in overall costs.

Magalhaes et al. compare pre-treatment options for liquid

fuel production, and conclude that pre-treatment via tor-

refaction is more cost eﬀ ective and ecologically sound than

options that utilize raw biomass.

However, these studies are

hampered by the lack of actual commercial-scale perform-

ance data, requiring many assumptions to be made as to

system performance. Additional assessment and commercial-

scale analysis is needed to determine the degree to which tor-

refaction can provide overall beneﬁ ts to the many conﬁ gura-

tions possible in the bioenergy supply chain.

Conclusions

Interest in biomass torrefaction has grown signiﬁ cantly in

recent years, as has knowledge of its processes and proper-

ties. Th

e process of torrefaction is dominated by the ther-

mal breakdown of hemicellulose to a combination of gases,

condensable liquids, and solid components that, together

with the feedstock’s cellulose and lignin, comprise a ‘char’

product suitable for downstream utilization as heat, elec-

tricity, fuels, or chemicals. Th

e torrefaction reaction is

believed to be dominated by a two-step process, which is

yet to be fully characterized. Mass and energy yields of the

solid product are generally in the 75–95% range, although

process conditions can greatly inﬂ uence the results.

Torreﬁ ed biomass also exhibits improved grindability and

resistance to moisture uptake, which are perhaps the most

valuable properties of the material, when compared to raw

327

Review: Torrefaction for bioenergy feedstock production

D Cielkosz, R Wallace

biomass. Torreﬁ ed biomass most likely requires densiﬁ ca-

tion if it is to be handled successfully in a bioenergy sup-

ply chain. However, the durability of densiﬁ ed torreﬁ ed

biomass appears to be a signiﬁ cant technological challenge

at this point. Th

e overall inﬂ uence of torrefaction on sup-

ply chain eﬃ

ciencies is expected to be positive, via reduced

transportation costs and improved end-use utilization.

Modeled predictions of the cost beneﬁ ts from torrefaction

tend to be modest.

Several gaps still exist in our understanding of torrefac-

tion, and there is need for continued work to characterize

and optimize this promising option for bioenergy feedstock

processing. While a general framework understanding of

the chemical reaction pathways has been developed, suc-

cinct understanding of the reaction network has yet to be

established, in part due to the complex chemical nature of

the feedstock. Greater understanding is also needed of the

chemical composition of the solid ‘char’ that is produced. In

addition, the eﬀ ects of large-particle processing have yet to be

suﬃ

ciently investigated. Supply chain impacts of torrefaction

are in need of signiﬁ cant further assessment, including the

real testing of commercial scale torrefaction supply chains,

and the identiﬁ cation of those scenarios and industries that

stand to beneﬁ t the most from torrefaction. Health and safety

issues related to torrefaction also require careful considera-

tion if torreﬁ ed biomass is to be successfully implemented as

a component of the global energy economy.

References

1. Armines, Procede de transformation de matiere ligneuse d’origine veg-

etale et matiere d’origine vegetale ligneuse transformee par torrefac-

tion. French Patent 81-16493, European Patent 82-40-25762, US Patent

410–636 (1981).

2. Bourgois J and Guyonnet R, Characterization and analysis of torrefi ed

wood. Wood Sci Technol 22:143–155 (1988).

3. Chen W and Kuo P, A study on torrefaction of various biomass materials

and its impact on lignocellulosic structure simulated by a thermogravim-

etry. Energy 35:2580–2586 (2010).

4. Bridgeman T, Jones J, Williams A and Waldron D, An investigation of the

grindability of two torrefi ed energy crops. Fuel 89:3911–3918 (2010).

5. Bergman P, Boersma A, Zwart R and Kiel J, Torrefaction for biomass co-

fi ring in existing coal-fi red power stations ‘BIOCOAL’. ECN Project Report

ECN-C-05-013. ECN, Petten, the Netherlands (2005).

6. Tedeschi S, Zimbardi F, Braccio G, Chiaramonti D and Rizzo A, Coupling

of torrefaction and grinding as biomass pre-treatment for ethanol pro-

duction. [Online]. 32nd Symposium on Biotechnology for Fuels and

Chemicals. Society for Industrial Microbiology, Clearwater Beach, FL,

April 20 (2010). Available at: http://sim.confex.com/sim/32nd/

webprogram/Session1799.html [December, 2010].

7. Uslu A, Faaij A and Bergman P, Pre-treatment technologies, and their

effect on international bioenergy supply chain logistics. Techno-

economic evaluation of torrefaction, fast pyrolysis and pelletisation.

Energy 33:1206 –1223 (2008).

8. Bergman PC and Kiel J, Torrefaction for Biomass Upgrading. In

Proceedings of 14th European Biomass Conference and Exhibition. Paris,

October 17–21 (2005).

9. Sridhar G, Subbukrishna D, Sridhar H, Dasappa S, Paul P and Mukunda

H, Torrefaction of Bamboo. [Online]. 15th European Biomass Conference

and Exhibition. Berlin, Germany, May 7–11 (2007). Available at: http://

cgpl.iisc.ernet.in/site/Portals/0/Publications/NationalConf/torrefaction

ofbamboo.pdf [December, 2010].

10. Anonymous, Wyssmont Turbo Drier Product Information. [Online].

Available at: www.wyssmont.com [July, 2010].

11. Huang Y, Kuan W, Lo S and Lin C, Total recovery of resources and energy

from rice straw using microwave-induced pyrolysis. Bioresource Technol

99:8252–8258 (2008).

12. Budarin V, Milkowski K, Shuttleworth P, Lanigan B, Clark J, MacQuarrie

et al., Microwave torrefaction of biomass. International Patent C10L 5/44

(2006.01), C10B 53/02 (2006.01) (2009).

13. Yan W, Hastings J, Acharjee T, Coronella C and Vasquez V, Mass and

energy balances of wet torrefaction of lignocellulosic biomass. Energ

Fuel 24(9):4738–4742 (2010).

14. Allen S, Kam L, Zemann A and Antal M, Fractionation of sugar cane with

hot, compressed, liquid water. Ind Eng Chem Res 35:2709–2715 (1996).

15. Petchpradab P, Yoshida T, Charinpanitkul T and Matsumura Y,

Hydrothermal pre-treatment of rubber wood for the saccharifi cation proc-

ess. Ind Eng Chem Res 48:4587–4591 (2009).

16. Mochidzuki K, Sakoda A and Suzuki M, Measurement of the hydrother-

mal reaction rate of cellulose using novel liquid-phase thermogravimetry.

Thermochim Acta 348:69–76 (2000).

17. Lipinsky RS, Arcate JR and Reed T, Enhanced wood fuels via torrefaction.

American Chemical Society. Division of Fuel Chemistry 47:408–410 (2002).

18. Fisher T, Hajaligol M, Waymack B and Kellogg D, Pyrolysis behavior and

kinetics of biomass derived materials. J Anal Appl Pyrol 62:331–349

(2002).

19. Deng J, Wang G, Kuang J, Zhang U and Luo Y, Pre-treatment of agri-

cultural residues for co-gasifi cation via torrefaction. J Anal Appl Pyrol

86:331–337 (2009).

20. Theander O, Cellulose, hemicellulose, and extractives, in Fundamentals

of Thermochemical Biomass Conversion, ed by Overend RP. Elsevier

Applied Science, London (1988).

21. Ramiah MV, Thermogravimetric and differential thermal analysis of cel-

lulose, hemicellulose, and lignin. J Appl Polym Sci 14:1323–1337 (1970).

22. Prins MJ, Ptasinski KJ and Janssen F, Torrefaction of wood part 1: weight

loss kinetics. J Anal Appl Pyrol 77:28–34 (2006).

23. Nimlos M, Brooking E, Looker M and Evans R, Biomass torrefaction

studies with a molecular beam mass spectrometer. American Chemical

Society. Division of Fuel Chemistry 48:590–591 (2003).

24. Arias B, Pevida C, Fermoso J, Plaza MG, Rubiera F and Pis J, Infl uence

of torrefaction on the grindability and reactivity of woody biomass. Fuel

Process Technol 89:169–175 (2008).

328

D Cielkosz, R Wallace

Review: Torrefaction for bioenergy feedstock production

25. Prins MJ, Ptasinski KJ and Janssen F, More effi cient gasifi cation via tor-

refaction. Energy 31:3458–3470 (2006).

26. ASTM. Standard Test Method for Volatile Matter in the Analysis of

Particulate Wood Fuels. Standard E872-82. ASTM International. West

Conshohocken, PA (2006).

27. Ferro DT, Vigouroux V, Grimm A and Zanzi R, Torrefaction of Agricultural

and Forest Residues. Cubasolar, Guantánamo, Cuba (2004).

28. Prins MJ, Ptasinski KJ and Janssen F, Torrefaction of wood part 2: analy-

sis of products. J Anal Appl Pyrol 77:35–40 (2006).

29. Repellin V, Govin A, Rolland M and Guyonnet R, Modelling annhydrous

weight loss of wood chips during torrefaction in a pilot kiln. Biomass

Bioenerg 34:602–609 (2010).

30. Felfl i F, Luengo C and Soler P, Mathematical modelling of wood and bri-

quettes torrefaction. In Proceedings of the 5th Encontro de Energia no

Meio Rural, Campinas, Spain, October 19–21 (2004).

31. Almeida G, Brito J and Perre P, Alterations in energy properties of euca-

lyptus wood and bark subjected to torrefaction: the potential of mass loss

as a synthetic indicator. Bioresource Technol 101:9778–9784 (2010).

32. Zanzi R, Ferro D, Torres A, Soler PB, and Bjornbom, E.biomass torrefac-

tion, in Second World Conference and Technology Exhibition on Biomass

for Energy, Industry and Climate Protection, Rome, Italy, pp. 859–862

(2004).

33. Couhert C, Salvador S and Commandre J, Impact of torrefaction on syn-

gas production from wood. Fuel 88:2286–2290 (2009).

34. Pach M, Zanzi R and Bjornbom E, Torrefi ed Biomass as a Substitute for

Wood. [Online]. 6th Asia Pacifi c International Symposium on Combustion

and Energy Utilization, Kuala Lumpur, May 20–22 (2002). Available at:

http://hem.fyristorg.com/zanzi/paper/zanzi_apisceuVI.pdf

35. Sadaka S and Negi S, Improvements of biomass physical and ther-

mochemical characteristics via torrefaction process. Environmental

Progress & Sustainable Energy 28(3):427–434 (2009).

36. Encinar J, Beltrán F, Bernalte A, Ramiro A and González J, Pyrolysis of

two agricultural residues: olive and grape bagasse. Infl uence of particle

size and temperature. Biomass Bioenerg 11:397–409 (1996).

37. Medic D, Darr M, Potter B and Shah A, Effect of Torrefaction Process

Parameters on Biomass Feedstock Upgrading. Paper 1009316. Annual

International Meeting of the ASABE, Pittsburgh, PA, June 20–23

(2010).

38. Knezevic D, van Swaaij W and Kersten S, Hydrothermal conversion of

biomass: I, glucose conversion in hot compressed water. Ind Eng Chem

Res 48:4731–4743 (2009).

39. Rodriguez T and Rousset P, Effects of torrefaction on energy properties

of eucalyptus grandis wood. Cerne Lavras 15:446–452 (2009).

40. Gi lbert P, Tyu C, Sharifi V and Swithenbank J, Effect of process param-

eters on pelletisation of herbaceous crops. Fuel 88:1491–1497 (2009).

41. Bergman P, Combined Torrefaction and Pelletization: The TOP Process.

ECN Project Report ECN-C-05-073. ECN, Petten, the Netherlands (2005).

42. Kiel J, Verhoeff F, Gerhauser H and Meuleman B. BO

Technology for

Biomass Upgrading Into Solid Fuel – Pilot Scale Testing and Market

Implementation. ECN Project Report ECN-M-08-036. ECN, Petten, the

Netherlands (2008).

43. ASTM. Standard Test Method for Grindability of Coal by the Hardgrove-

Machine Method. Standard D409-09a. ASTM International, West

Conshohocken, PA (2010).

44. Bergman PC, Boersma AR, Kiel JH, Prins MJ, Krzyzstof JP and Janssen

F, Torrefaction for Entrained Flow Gasifi cation of Biomass. [Online]. 2nd

World Conference and Technology Exhibition on Biomass for Energy,

Industry, and Climate Protection, Rome (2004). Available at: http://www.

ecn.nl/docs/library/report/2004/rx04046.pdf [December, 2010].

45. Repellin V, Govin A, Rolland M and Guyonnet R. Energy requirement for

fi ne grinding of torrefi ed wood. Biomas Bioenerg 34:923–930 (2010).

46. Felfl i FF, Luengo CA, Suarez JA and Beaton P , Wood briquette torrefac-

tion. Energy for Sustainable Development 9:19–22 (2005).

47. Wooten J, Crosby B and Hajaligol M, Evaluati on of cellulose char struc-

ture monitored by 13C CPMAS NMR. American Chemical Society.

Division of Fuel Chemistry Preprints 46:191–193 (2000).

48. Bridgeman TG, Jones JM, Shield I and William s P, Torrefaction of reed

canary grass, wheat straw, and willow to enhance solid fuel qualities and

combustion qualities. Fuel 87:844–856 (2008).

49. Holtzapple M, Humphrey A and Taylor J, Energ y requirements for the

size reduction of poplar and Aspen wood. Biotechnol Bioeng 33:207–210

(1989).

50. Comstock GL, Energy requirements for drying of wood products, in

Proceedings of the Forest Products Research Society Energy Workshop.

WI: Forest Products Research Society, Denver, CO. pp. 8–12 (1975).

51. Simpson WT and Tschernitz J, Time, costs, an d energy consumption for

drying red oak lumber as affected by thickness and thickness variation.

Forest Prod J 30:23–27 (1980).

52. Nicholls D, Zerbe J, Bergman R and Crimp P, Use of Wood Energy for

Lumber Drying and Community Heating in Southeast Alaska. General

Technical Report FPL−GTR−152. USDA Forest Products Laboratory,

Madison, WI (2004).

53. Gruber T, Drying of wood chips with optimized energy consumption and

emission levels, combined with production of valuable substances. In

Proceedings of the 3rd European COST E15 Workshop on Wood Drying.

Helsinki, Finland, June 5 (2001).

54. Svoboda K, Pohorely M, Hartman M and Martinec J, Pre-treatment and

feeding of biomass for pressurized entrained fl ow gasifi cation. Fuel

Process Technol 90:629–635 (2009).

55. Pimchuai A, Animesh Dutta A and Basu P, Torref action of agriculture resi-

due to enhance combustible properties. Energ Fuel 24:4638–4645 (2010).

56. Mani S, Integrating biomass torrefaction with thermo-chemical conver-

sion processes. Paper 160229. In Proceedings of the 2009 AIChE Annual

Meeting, Nashville, TN, November 8–13 (2009).

57. Hakansson K, Torrefaction and Gasifi cation of Hydrolysis Residue. MSc

Thesis, Umea Institute of Technology (2007).

58. Zwart R, Boerrigter H and Van der Drift A, The impact of biomass pre-

treatment on the feasibility of overseas biomass conversion to Fischer-

Tropsch products. Energ Fuels 20:2192–2197 (2006).

59. Magalhaes A, Petrovic D, Rodriguez A, Putra Z and Thielemans G, techno-

economic assessment of biomass pre-conversion processes as a part of

biomass-to-liquids line-up. Biofuels, Bioprod Bioref 3:584–600 (2009).