Combustion and Flame 194 (2018) 72–84

Contents lists available at ScienceDirect

Combustion and Flame

journal homepage: www.elsevier.com/locate/combustflame

On the particle sizing of torrefied biomass for co-firing with

pulverized coal
Aidin Panahi a, Mahmut Tarakcioglu a, Martin Schiemann b, Michael Delichatsios a,
Yiannis A. Levendis a,∗
a
Mechanical and Industrial Engineering Department, Northeastern University, Boston, MA, USA
b
Ruhr University Bochum, Bochum, Germany

a r t i c l e i n f o a b s t r a c t

Article history: In biomass harvesting and fuel preparation processes, grinding causes a prominent energy consumption
Received 8 January 2018 penalty, which results in an analogous cost impact. This is due to the fibrous and tenacious nature of
Revised 13 February 2018
biomass. Torrefaction of biomass makes it brittle, as it diminishes its fibrous nature and, hence, it en-
Accepted 12 April 2018
hances its grindability. Nevertheless, grinding costs are still important and increase with decreasing tar-
geted particle size. Therefore, this study introduces a methodology for assessing the torrefied biomass
Keywords: grind size that is suitable for firing or co-firing with coal in existing pulverized fuel boilers. It examines
Combustion combustion of biomass of different origins, herbaceous, woody, or crop-related. Biomass was torrefied for
Torrefied biomass 30 min at 275 °C in nitrogen. It was subsequently ground and sieved to various size cuts, which reflect
Size
the mean widths rather than the lengths of these typically elongated particles. Subsequently, the particles
Burn-out time
were burned, one at a time, in a drop tube furnace (DTF) under high temperature and high heating rate
Temperature
Co-firing conditions. Luminous burnout times were observed pyrometrically and cinematographically for a num-
ber of single particles from various size cuts. Such burnout times were then contrasted with those of
individual coal particles in the size range of 75–90 μm, i.e., at the upper end of particle sizes burned in
coal-fired boilers. Based on this comparison, the nominal sieve size of the examined torrefied biomass
particles whose overall observed burnout times matched those of the 75–90 μm coal particles was deter-
mined to be 212–300 μm. Hence, to minimize the grinding cost of co-firing such torrefied biomass with
coal in existing boilers, its finer pulverization may not be necessary.
© 2018 The Combustion Institute. Published by Elsevier Inc. All rights reserved.

1. Introduction process, which can ameliorate the biomass utilization character-

Approximately 30% of the total power produced in the USA in
2016/2017 was harvested from coal according to the Energy Infor-
mation Administration and the American Coalition for Clean Coal
Electricity [1,2]. Similarly to the benefits of co-firing of different
coal types, co-firing of coal and biomass can be used to reduce the
emission of a number of environmental pollutants [3–5]. In addi-
tion, co-firing coal with renewable biomass can greatly reduce the
net emissions of carbon to the atmosphere [6–8].
Since biomass is plentiful and naturally grown, it is categorized
as a renewable energy source [9,10]. However, its low calorific
value, high moisture content, hygroscopic nature and soot emis-
sions during combustion diminish the appeal of biomass as a fuel.
Torrefaction is a practical method for improving the properties of
biomass [11–17]. Torrefaction is a thermochemical pretreatment
∗
Corresponding author.
E-mail addresses: y.levendis@neu.edu, yal@coe.neu.edu (Y.A. Levendis).
istics, including calorific value and resistance to decay by remov-
ing moisture and volatiles [17–24]. Therefore, torrefied biomass can
be a suitable alternative fuel in existing large-scale pulverized coal
boilers because torrefaction brings the properties of biomass closer
to those of coal [12,25–28]. Also, due to their fibrous nature, raw
biomass particles cannot be readily pulverized down to the same
particle sizes as coal [29–33]. Thus, their grindability is an issue.
Torrefaction improves grindability [18,34].
To consume biomass in utility boilers burning pulverized coal,
grinding of the solid fuel is a necessary and key step [9,35]. The
resulting biomass particle size, and even particle shape are impor-
tant physical properties, which influence the fuel fluidization and
combustion parameters of the particles [36–41]. Marinelli et al.
[42] reported that grinding to very fine particle sizes increases
the cohesive strength of the powder, which impedes fluidization.
Grinding to fine sizes is also expensive as it requires a lot of en-
ergy. Whereas grinding of torrefied biomass consumes a fraction of
the energy requirement for grinding raw biomass [34,43–45], the

https://doi.org/10.1016/j.combustflame.2018.04.014
0010-2180/© 2018 The Combustion Institute. Published by Elsevier Inc. All rights reserved.
 A. Panahi et al. / Combustion and Flame 194 (2018) 72–84
energy for grinding any solid fuel still increases with decreasing
targeted particle size. Hence, the targeted torrefied biomass parti-
cle size should be as large as combustion considerations in boil-
ers would allow. Additionally, it is advantageous that the targeted
size of biomass particles in pulverized fuel combustion be larger
than that of coal particles because of their typically lower bulk
density, faster devolatilization rates and higher volatiles to fixed
carbon ratios [46–49]. On the other hand, the larger particle sizes
of biomass can induce lengthier ignition delays than coal and can
generate larger quantities of unburned residues [35,50–54]. Hence,
finding the right biomass particle size to co-fire with coal is im-
portant in achieving high combustion efficiency, while also min-
imizing the grinding costs. The particle size selection has been
addressed theoretically by Saastamoinen et al. [51] and experimen-
tally by Mock et al. [55]. The former authors concluded that single
biomass in the size range of 500 μm burns with the same time du-
ration as a 200 μm coal single particle. The latter authors burned
pulverized biomass and sewage sludge and found that particles of
355–425 μm and smaller, when exposed to a hot upwards moving
oxidizer gas stream (heated to 1340 K), moved in the direction of
the stream and burned completely. To the contrary, larger particles
fell to the bottom of the furnace and failed to burn completely.
This research examines the maximum particle size of torrefied
biomass for co-firing with coal based on their relative burnout
times under identical conditions. Torrefied biomass is chosen since
this fuel is more suitable for co-firing with coal in existing boil-
ers [45,56]. Besides examining overall particle burnout times, this
work also examines the contributions of the volatile and char com-
bustion phases to the cumulative burnout times. It is also reporting
on pyrometrically-measured particle combustion temperatures and
ignition delays. Different types of pulverized torrefied biomass fu-
els in the size cuts of (75–90) μm, (180–212) μm, (212–300) μm,
(300–350) μm and (350–500) μm were burned in a DTF operated
at a wall temperature of 1400 K which resulted to an axial gas tem-
perature of ∼1350 K, i.e, the same as that used in the prior coal
combustion studies [55,57]. Based on combustion observations, a
recommendation for the size of torrefied biomass particles that ig-
nite and burn in time-frames that are comparable to those of coal
particles in the range of 75–90 μm is given. This coal particle size
was selected based on the industry standard of 65–70% of coal par-
ticles passing 75 μm (200 mesh) [58].
2. Methods
2.1. Preparation of samples
Pulverized corn straw and rice husk were harvested in Harbin
province of China and were provided by Harbin Institute of Tech-
nology. Miscanthus originated from an agricultural farm in Ger-
many (Sieverdingbeck-Agrar). Beechwood was from trees grown
in the Netherlands. Pulverized miscanthus and beechwood were
provided by Ruhr-University Bochum, Germany [59]. Sugarcane
bagasse was obtained from a bio-ethanol production plant in
Brazil. DDGS (Distiller’s Dried Grains with Soluble) was provided
by a North American ethanol-producing company. Torrefaction of
all samples was carried out in a laboratory-scale muffle furnace
in nitrogen. The furnace was charged with small amounts (a few
grams) of millimeter-size particles of biomass and, subsequently,
they were heated to 275 °C with heating rates in the order of
10 °C/min. Upon reaching the final temperature, each sample was
treated at constant conditions for 30 min. All torrefied biomass fu-
els were air-dried, chopped in a household blender, and size clas-
sified by sieving to obtain size cuts of (75–90) μm, (180–212) μm,
(212–30 0) μm, (30 0–350) μm and (350–500) μm. Optical micro-
scope photographs and scanning electron microscope photographs
of each torrefied biomass are shown in Figs. 1 and 2, respectively.
73
Therein, it can be observed that most types of biomass, except
DDGS and rice husk, are elongated and, thus, have high length-
to-diameter aspect ratios. Therefore, for those biomass types the
aforesaid particle size ranges, are nominally based on size classi-
fication by sieving. Whereas the mean widths of biomass particles
were in the ranges of mesh sizes, their mean lengths in most cases
exceeded the mesh sizes. This shows that the traditional sieve clas-
sification method, which is intended for spherical or spheroidal
particles, does not represent all of the dimensions of all biomass
particles because of their elongated shapes. In the previous work
in this laboratory, it was observed that the torrefaction process re-
duced the particle aspect ratios [60] and produced more uniform
particle sizes. All coal samples were procured from the Penn-State
Coal Bank and burned in this laboratory as described in a prior
publication [3]. The proximate analysis and the ultimate analysis
of the biomass fuels on a dry basis, are given in Table 1, whereas
those of the coals are given in Ref. [53,61]. The proximate and ul-
timate analysis, as well as the determination of the heating value
of the torrefied biomass fuels, were performed at Harbin Institute
of Technology, according to GB/T 212-2008, GB/T 30733-2014, GB/T
30733-2014 and to GB/T 213-2008 Chinese standards, see Ref. [60].
2.2. Experimental apparatus
The combustion of free-falling fuel particles took place in a
laminar flow, vertical drop tube furnace, operated at a wall temper-
ature, Twall , of 1400 K. The resulting axial gas temperature profile
was measured to be approximately constant at Tgas = 1350 K [62].
The radiation cavity of this furnace (an ATS unit) was 25 cm long
and it was electrically heated by hanging MoSi2 elements. A verti-
cal 7 cm i.d. transparent quartz tube was fitted in this furnace. Air
was introduced into the tube through a water-cooled stainless steel
injector, and, also, through a flow straightener placed coaxially to
the furnace injector, see Fig. 3. The air flow through the injector
was set at 0.5 l/min. To enable single particle combustion, the fuel
particles were introduced through a port at the top of the injec-
tor by first placing them on the tip of a beveled needle syringe.
Gentle taps on the needle allowed single particles to enter the in-
jector and, subsequently, the furnace. Pyrometric observations of
those particles were conducted from the top of the furnace injec-
tor, viewing downwards the central axis of the furnace, Fig. 3, i.e.,
along with a particle’s path line. Details of the pyrometer optics,
electronics, calibration, and performance were given by Levendis
et al. [63]. The voltage signals generated by the three detectors
were amplified and then were processed by a microcomputer us-
ing the LabView software.
A high-speed, high-resolution camera was located at one side
of the furnace and viewed through slotted side quartz windows to
record the particle combustion histories, against a backlight frosted
glass position at the diametrically-opposite side of the furnace. An
Edgertronic self-contained digital high-speed broadband video cam-
era was used, at speeds of 10 0 0 frames per second. The camera
was fitted with an Olympus-Infinity model K2 long-distance micro-
scope lens to provide high-resolution images of the combustion
events. Scanning Electron Microscopy, with a Hitachi S-4800 instru-
ment, was used to observe the initial biomass particles. The instru-
ment was operated with a 3 kV of accelerating voltage, 10 μA of
beam current and 8.5 mm working distance.
3. Results and discussion
3.1. Cinematographic observations
High-speed, high-resolution cinematographic sequences of sin-
gle torrefied biomass particles burning in air are shown in Fig. 4.
The combustion behaviors of the various types of torrefied biomass
74 A. Panahi et al. / Combustion and Flame 194 (2018) 72–84

Fig. 1. Optical microscope photographs of various torrefied biomass fuel particles, each at two different magnifications. (a) Corn straw. (b) Miscanthus. (c) Sugar-cane bagasse.
(d) DDGS. (e) Rice husk. (f) Beechwood. In each case, the left side frames show particles in the size cut of 180-212 μm (based on sieving), whereas the right-side frames
show particles in the size cut of 350–500 μm. A size bar is included at the bottom right of this figure.

Fig. 2. Scanning electron microscope images of torrefied biomass fuels in this study, taken at two different magnifications: (a) Corn straw. (b) Miscanthus. (c) Bagasse. (d)
DDGS, (e) Rice husk. (f) Beechwood. The size bars included at the bottom of the figure, corresponds to all photographs in each column.
 A. Panahi et al. / Combustion and Flame 194 (2018) 72–84 75

Table 1
Chemical compositions and energy contents of six types of torrefied biomass.

Rank/fuel source Biomass (Torrefied at 275 °C for 30 min)

Herbaceous Crop-derived Woody

Corn straw Miscanthus Sugarcane bagasse Corn DDGS Rice husk Beechwood
Proximate analysis (dry basis)

Volatile matter (%) 67.55 75.00 73.74 71.46 55.62 72.20

Fixed carbon (%) 24.56 22.30 23.49 21.09 21.58 27.30
Ash (%) 7.89 2.70 2.71 7.45 22.80 0.50
Ultimate analysis (dry basis)
Carbon (%) 52.77 52.80 55.82 58.22 44.16 55.40
Hydrogen (%) 5.32 5.70 5.45 6.32 4.41 5.30
Oxygen (%) (by difference) 32.46 38.6 34.94 22.98 26.73 38.88
Nitrogen (%) 1.50 0.18 1.00 4.00 1.22 0.18
Sulfur (%) 0.07 0.46 0.03 1.01 0.03 0.31
Calcium (%) 0.56 0.55 0.44 0.15 0.35 0.19
Sodium (%) 0.02 0.04 0.03 0.80 0.02 0.01
Potassium (%) 1.19 0.97 0.06 1.50 0.73 0.19
Magnesium (%) 0.29 0.14 0.06 0.32 0.08 0.07
Chlorine (%) 0.18 0.03 0.01 0.12 0.04 0.02
Heating value (MJ/kg) 19.4 20.1 20.3 23.7 16.1 20.9

Fig. 3. Schematic illustration of the drop tube furnace, particle introduction and sequential photographs of a single particle combustion event (not to scale).
76 A. Panahi et al. / Combustion and Flame 194 (2018) 72–84

Fig. 4. Images from high-speed high-resolution cinematography of single torrefied biomass particles burning in air in a DTF operated at Twall = 1400 K. The displayed numbers
in each frame are in milliseconds, where zero does not mark the beginning of ignition, instead, it merely represents the beginning of the depicted sequences. Nominal initial
particle sizes were in the range of 212–300 μm.

particles appear to be rather similar, despite differences in their
physical and chemical properties, such as mass, shape, aspect ra-
tios, surface structure and roughness, porosity, volatile content
and chemical composition. All torrefied biomass particles in these
size ranges appeared to ignite homogeneously and to burn in two
phases: volatile matter combustion and char combustion. Volatiles
burned in fairly luminous envelope flames, surrounding individ-
ual particles. The shape of the envelope flames varied from spher-
ical to ellipsoidal, depending on the shape of the devolatilizing
particles. Upon extinction of the volatile flames, char combustion
took place. This behavior has also been observed in previous work
in this laboratory [64].
3.2. Initial particle aspect ratio
Aspect ratios of the torrefied biomass particles were obtained
by accessing images of as many as 500 particles, photographed
under an optical microscope. An overall cylindrical shape was
 A. Panahi et al. / Combustion and Flame 194 (2018) 72–84
Fig. 5. Mean initial aspect ratio of torrefied biomass fuels in this study for five
different size cuts. Standard deviation in the data is shown.
assumed for all particles, as suggested by visual examination under
the microscope. Results are shown in Fig. 5, where the aspect ratio,
denoted as average, is the ratio of the mean length to the mean
width of the particles. The mean widths of the particles were all
in the range of the mesh aperture sizes of the sieves; however, the
mean lengths of all biomass particles exceeded the mesh because
many types of biomass particles are elongated. This is in agree-
ment with previous research in this laboratory which showed that
the torrefaction process reduces the axial dimension of needle-like
particles of their raw biomass precursors [60,64]. Furthermore, it is
noticeable that for all the biomass samples, with the exceptions of
DDGS and rice husk, the particle aspect ratios increase with in-
creasing average particle width. This is illustrated in the optical
microscopy and SEM images displayed in Figs. 1 and 2. As it was
also reported in a prior publication [53], torrefied biomass particles
have much higher aspect ratios than coal.
3.3. Ignition of torrefied biomass particles
Single torrefied biomass particles burning in air ignited homo-
geneously forming envelope flames, with some exemptions for the
largest size particles. This is in agreement with prior observations
on biomass samples both raw and torrefied that were also burned
in the same facility [64], where similar homogeneous ignition be-
havior was observed. While there have been some other efforts to
study the ignition behavior of single particle biomass fuels [65],
studies focusing on the ignition delay of different types of tor-
refied biomass particles are scarce. Herein, the ignition delay was
defined based on still frames of back-lit cinematography. It has
been taken as the elapsed time from the moment when a parti-
cle was recorded to exit the water-cooled DTF injector to the mo-
ment when particle-related luminosity became visually apparent,
see for example Fig. 4. Particle ignition delays derived with this
method are displayed in Fig. 6 for all the torrefied biomass fuels of
this study. Therein it is shown that the ignition delay of torrefied
biomass particles increases with increasing size, is in agreement
with the results of Austin et al. [65] for ignition delays of raw cel-
lulosic dust particles.
There have been some additional reports [66–68] on the ig-
nition behavior of raw biomass, however, studies on the ignition
of torrefied biomass fuels are scarce. Whereas the fact that igni-
tion delay increases with particle size is rather predictable, since
small particles heat up faster, the separation of ignition delays ac-
cording to the types of torrefied biomass, examined herein, is no-
table. In ascending order, the ignition delays were as follows: Sug-
arcane bagasse, corn straw, miscanthus, beechwood, DDGS and rice
husk. While it is likely that many physical and chemical factors
77
Fig. 6. Mean torrefied biomass particle ignition delay times, in air, for various par-
ticle size cuts in the DTF operated at Twall = 1400 K (Tgas = ∼1350 K). Standard devi-
ation in the data is shown.
Fig. 7. Ignition in torrefied corn straw and miscanthus particles, in air, for large
particles injected in the DTF, operated at Twall = 1400 K (Tgas = ∼1350 K).
contribute to these biomass-type-dependent particle ignition de-
lays, it is remarkable to notice that the observed trend of igni-
tion delays is the opposite of the trend of particle aspect ratios of
these six torrefied biomass types in each size fraction (see Fig. 5).
It appears that the particles with the highest aspect ratios ignite
more readily than the rest. Past work by one of the authors in-
vestigated preferential ignition of irregularly-shaped carbonaceous
particles (coal chars) at localized “hot-spot” sites [69], where heat
generation to heat loss ratios can be favorable. Such phenomena
may be at play herein in the case of needle-shaped particles with
large aspect ratios, such as bagasse, corn straw and miscanthus and
to a lesser extent beechwood. Particles of those biomass types are
more likely to exhibit this behavior than the rest. Some evidence
for this phenomenon was obtained herein, see Fig. 7, but further
studies are needed. Another influential factor in the ignition delay
trend, albeit not as consistent as the particle aspect ratio, is the
volatile matter content of the torrefied biomass.
3.4. Combustion of torrefied biomass particles
The combustion behavior of single torrefied biomass particles
in all nominal size ranges of this investigation was recorded with
high-speed cinematography, as illustrated in Fig. 4, and with opti-
cal pyrometry, as was described in Section 2.2. Pyrometric inten-
sity profiles of a single torrefied particle of miscanthus biomass
in the size cut of 212–300 μm and a single bituminous coal (Pitts-
burgh #8) in the size cut of 75–90 μm are shown superimposed
in Fig. 8. Therein the sequential phases of volatile mater com-
bustion and residual char combustion of both particles are ev-
ident. It can be seen that in these two cases the overall lu-
minous combustion durations of the two particles are compara-
ble. The volatile combustion phase of the torrefied miscanthus
78 A. Panahi et al. / Combustion and Flame 194 (2018) 72–84
Fig. 8. Pyrometric monochromatic (at 998 nm) intensity signals of a single parti-
cle of torrefied miscanthus in the size range of 212–300 μm and a bituminous coal
(Pittsburgh #8) in the size range of 75–90 μm, both burning in air in the DTF oper-
ated at Twall = 1400 K (Tgas = ∼1350 K).
particle is longer, as it has a higher volatile mass fraction (75%)
than the coal (35%), but its intensity peak is lower, as it also ap-
pears to be less bright than coal.
This work is concerned with the overall burnout times of in-
dividual torrefied biomass particles; however, a breakdown of the
volatile matter burnout times and the residual burnout times is
also included. Moreover, for completion, the volatile matter flame
temperatures and the char surface temperatures, as measured by
optical pyrometry, are also presented in this manuscript.
3.5. Combustion burn-out times
Burnout times of volatiles and chars of the torrefied biomass
particles with nominal initial sizes in the ranges of (75-90) μm,
(180–212) μm and (212–30 0) μm, (30 0–350) μm and (350–500) μm
are shown in Fig. 9a and b. In each range, the combustion his-
tories of over 40 particles were recorded and average times were
calculated. All burnout times were measured by optical pyrometry
and were verified by observations with high-speed cinematogra-
phy. The two techniques were found to be in reasonable agree-
ment. Char burnout times were longer, by factors of 2–3, than
volatile matter burnout times for all six torrefied biomass types.
Predictably, both the volatile matter and the char burnout times
were seen to increase with particle size.
Based on experimental observations, particles of (75–90) μm,
(180–212) μm and (212–300) μm size cuts experienced complete
burnouts in the particular DTF, used in this study, without leaving
any black color residues at its exit, where a white filter was placed.
It should be emphasized that in the case of elongated particles
these nominal size ranges are based on the mesh apertures of the
sieves and, thus, correspond to the mean widths of the particles,
not their mean lengths. Collected residues of larger particles, (300–
500) μm appeared to occasionally contain black carbon. Cumulative
(total) volatile and char luminous burnout times for these particles
are plotted in Fig. 10; they span the time-frame of 60–300 ms. Ac-
cording to these results, it appears that torrefied biomass particles
of all the sizes experimented with herein, i.e. up to a nominal size
of 500 μm (based on size classification by sieves, i.e., by the aper-
ture of the mesh), experience burnout times lasting up to 300 ms.
The time frames recorded in this work are those of the luminous
combustion of the char particles. Smoldering or otherwise low-
luminosity combustion may have been neither recorded by the py-
rometer nor visually detected in the cinematographic frames. Pre-
vious studies in this laboratory burned single 75–90 μm particles
from five different ranks of coals [53,61,70] in the same DTF, in air,
also at Tgas = 1350 K. The aforesaid burnout times of these torrefied
biomass (tbournout < 300 ms) appeared to be in the overall range of
those of coal particles from all ranks (tbournout < 300 ms), including
the slower-burning anthracite. However, if torrefied biomass is to
be fired in bituminous or lignite coal boilers, then it would be rec-
ommended to limit the nominal size (based on size classification
by sieves) of these biomass particles to 300 μm. In the case of elon-
gated particles (cylindrical, elipsoid, needle-like, etc.) this dimen-
sion corresponds to their mean widths, not their lengths. Thereby,
their burnout times will be consistent with those of 75–90 μm coal
particles, that range between 100 and 300 ms. The above recom-
mendation is in line with results reported by Mock et al. [55], who
used a different criterion (the upwards motion of the fuel particles)
to recommend a maximum nominal torrefied biomass particle size
of 425 μm under similar gas temperature conditions. Such recom-
mendations for torrefied biomass particle size selection (300 μm)
can bring about potential savings in the power consumption for
grinding. For instance, the grinding energy requirement for bitumi-
nous coal and torrefied biomass particles in the size range of 75–
90 μm has been reported to be ∼20 kWe/MWth, whereas that of a
number of torrefied biomass types of a nominal size of 300 μm has
been reported to be ∼5 kWe/MWth [45]. Thus, a factor of four in
energy savings can be achieved by avoiding to grind the torrefied
biomass down to the size of the coal.
Close examination of Figs. 9 and 10 reveals differences in
burnout times among the different torrefied biomass types. Such
differences, in volatile and char combustion burnout times can be
attributed to physical and chemical variations of the pyrolyzate
gases and the residual chars. They include dissimilarities in molec-
ular weight, volatility and auto-ignition temperature of the de-
volatilizates. They also include dissimilarities in size, shape, aspect
ratio, density of the chars, as well as catalytic activity of the miner-
als therein, as previously reported [71]. For instance, in the case of
rice husk, it should be pointed out that whereas its volatile content
is lower than those of the other torrefied biomass types, its ash
content is exceedingly high, at almost 23%. As a result, the amount
of its combustible char component is the lowest, see Table 1. These
are likely reasons why its volatile matter combustion times are
among the briefest shown in Fig. 9a , also, its char combustion
times are among the briefest shown in Fig. 9b.
3.5.1. Flame temperatures
Particle envelope flame temperatures, presented in Fig. 11, cor-
respond to the volatile combustion phase, the occurrence of which
was illustrated in Fig. 4. Such temperatures were obtained from
pyrometric recordings of the soot radiation emanating from such
flames, assuming that the soot emissivities therein were indepen-
dent of wavelength, in the region of the pyrometric wavelengths
employed herein, i.e., 650–10 0 0 nm [70]. This assumption has been
shown to be acceptable in the case of optically-thin envelope
flames wherein temperature gradients are steep in the radial di-
rection [72]. This assumption is likely to be valid for the volatile
matter envelope flames of this work, but confirmation needs to
be obtained in future work. Temperatures obtained with this as-
sumption are expected to be in the vicinity to the highest tem-
perature of the soot in the flame at any instant of time [72].
Moreover, the presented values for the volatile flame temperatures
herein are the peak temperatures measured in single particle pyro-
metric temperature-time histories, averaged over the 40 individual
biomass particles that were successfully detected by the pyrome-
ter in each case. Standard deviation bars are shown in all cases.
The radiation intensity signals captured by the pyrometer during
the combustion of the torrefied biomass particles were compara-
bly strong to those captured during the volatile phase combustion
of smaller coal particles, under identical experimental conditions
[61], as illustrated in Fig. 8.
 A. Panahi et al. / Combustion and Flame 194 (2018) 72–84 79

Fig.. 9. Experimental data on combustion times of (a) volatile flames (b) chars of torrefied biomass particles in three ranges within 75–350 μm burning in air in the DTF
operated at Twall = 1400 K (Tgas = ∼1350 K). Columns (left to right): (75–90) μm, (180–212) μm, (212–300) μm, (300–350) μm and (350–500) μm.

Fig. 10. Comparison of cumulative combustion times of single particles of coal and torrefied biomass burning in air in the DTF operated at Twall = 1400 K (Tgas = ∼1350 K).
Columns (left to right): (75–90) μm, (180–212) μm, (212–300) μm, (300–350) μm and (350–500) μm. The solid line (———) represents the anthracite, the dotted line
(•••••••••) represents the semi-anthracite, the dashed line (- - - - -) represents the bituminous and the dashed-dotted line (− • − • − ) represents both the lignite and
the sub-bituminous coals. Bituminous, lignite, sub-bituminous are in the nominal range size of 75–90 μm and anthracite and semi-anthracite are in the nominal size range
of 75–150 μm.

Fig. 11. Average peak flame temperatures for all biomass fuels of this study burning with preheated air inside the DTF operated at Twall = 1400 K (Tgas = ∼1350 K). Columns
(left to right): (75–90) μm, (180–212) μm, (212–300) m, (300–350) μm and (350–500) μm.
80 A. Panahi et al. / Combustion and Flame 194 (2018) 72–84
Fig. 12. Average measured char temperatures for all biomass fuels of this
study burning with preheated air inside the DTF operated at Twall = 1400 K
(Tgas = ∼1350 K). Columns (left to right): (75–90) μm, (180–212) μm, (212–300) μm,
(300–350) μm and (350–500) μm.
The measured volatile matter envelope flame temperatures
were in the range of 210 0–230 0 K. Such temperatures significantly
exceed the surrounding gas temperature, because of heat release
during exothermic combustion of the devolatilized hydrocarbons
and hydrogen with oxygen. Particle size appears to have a consis-
tent effect on volatile flame temperature as bigger particles were
hotter than smaller ones, albeit with values within experimental
error. This may be explained by reasoning that the energy gener-
ated increases with the flame diameter cubed, whereas heat losses
from the flame increase with the flame diameter square.
Differences in the volatile temperatures among the various
types of torrefied biomass are mild (within 100 K for comparable
sizes), and they may be related to volatile mass released or chem-
ical compositions of pyrolyzates gases. The cellulose and hemicel-
lulose components of biomass decompose to small molecules in
the form of volatile hydrocarbons, carbon oxides, tars and pyrolytic
water [73,74]. In fact, the volatile flames of raw biomass fuels were
previously observed in this laboratory to be significantly less bright
(and less sooty [75]) than those of high-rank coals. Oxygen con-
tents of biomass fuels are much higher than those of coals. Hence,
the volatile pyrolyzates contain CO2 and CO, in addition to hydro-
gen and light hydrocarbons [76]. The volatile envelope flames of
the torrefied biomass fuels burned herein are considerably brighter
than those of their raw biomass precursors [64].
3.5.2. Char temperatures
Pyrometric char temperatures for all biomass types were mea-
sured to be in the 170 0–190 0 K range, as shown in Fig. 12.
Char temperatures initially increased with increasing particle size,
reached a maximum and thereafter decreased with further increas-
ing particle size. The latter behavior (decreasing trend) is typically-
attributed to diffusion-controlled burning of solid fuel particles as
the mass transfer reaction rate coefficient is inversely proportional
to the particle diameter [77,78]. This is defined as Regime III of
combustion [79], which is likely to be experienced by particles
of large size and/or particles burning at high temperatures. The
former behavior (i.e., the increasing trend) may be manifested in
combined kinetic and pore diffusion controlled burning (Regime II
of combustion [79]). Therein the variation of char particle tem-
perature with size depends upon the ambient conditions (Tgas ,
PO2 ) and the heat generation versus the heat loss of the particle
[77]. This behavior was investigated theoretically using Field’s sim-
plified single film model for combustion of a solid carbonaceous
particle [80]. With this model the variation of char surface tem-
peratures with particle size were calculated based on a number
of assumptions. The brief description of the model, the obtained
results and a parametric analysis are presented in the Appendix.
Remarkably, both the char particle temperature values and the
temperature trend with char particle size were represented with
good fidelity. In general, biomass chars burned with somewhat
higher (by 50 K) average temperatures than bituminous char par-
ticles previously burned in this laboratory [57].
4. Conclusions
The determination of the maximum particle size for complete
combustion of pulverized torrefied biomass in coal-fired boilers is
essential for maintaining high efficiency and minimizing the un-
burned carbon in the ash. A method for expedient determination
of the nominal size (based on sieving) of torrefied biomass par-
ticles for firing or co-firing with coal in existing boilers is pre-
sented herein. It is based on matching the burnout times of sin-
gle biomass particles to those of coal particles that are typically
burned in a boiler. Six types of biomass from different origins were
torrefied at 275 °C for 30 min and, subsequently, they were burned
with air in a DTF, heated to Tg = 1350 K. It was observed that par-
ticles of torrefied biomass with nominal sizes smaller than 300 μm
(50 mesh) burned completely with luminous burnout times match-
ing those of 75–90 μm pulverized coal particles, whose size-cut
selected based on industry standards. As most types of torrefied
biomass particles have elongated shapes, i.e., they are cylindrical,
ellipsoidal, cuboids or polygonal prisms, their size classification
by sieves corresponds to their widths (minor axes), not to their
lengths (major axes). Hence, with this method the recommended
maximum nominal size (width) for grinding the torrefied biomass
types examined in this study was 300 μm. Grinding to smaller par-
ticle sizes may not be necessary as it results in higher grinding
costs. It should be noted that this recommendation may vary for
other biomass types, or even for the same biomass types had they
been torrefied under different conditions. Moreover, it is further
recommended that additional confirmation of these results is ob-
tained based on combustion observations of streams of blended
coal/biomass powders, coupled with examination of their combus-
tion residues for unburned carbon content.
Additional conclusions from this work that may be noteworthy
in the selection of torrefied biomass size are: (a) the ignition de-
lay of particles increased with the particle size and decreased with
the particle aspect ratios; and (b) the burnout times of the de-
volatilizates were half as long as those of the char residues and
their temperatures were significantly hotter.
Acknowledgment
Aidin Panahi was supported by the Department of Mechani-
cal and Industrial Engineering at Northeastern University by an
Akira Yamamura Fellowship. Martin Schiemann was supported by
the German Research Foundation (DFG) within the project SCHI
1272/2-1. The authors would like to thank Mr. Xiaohan Ren for his
assistance in conducting the proximate and ultimate analysis of the
fuels.
Appendix: Calculation of the char temperature as a function of
char particle size
In this section, Field’s [80] single film model was used to calcu-
late the char temperatures of single particles of torrefied biomass,
the experimentally-determined values of which are shown in
Fig. 12. This model assumed that torrefied biomass particles are
spherical carbon particles surrounded by a boundary layer of stag-
nant gas several particle diameters thick. Oxygen diffuses through
this gas from the free stream and reacts at the particle surface to
form combustion products, which diffuse outwards and pass into
 A. Panahi et al. / Combustion and Flame 194 (2018) 72–84
the free stream [81]. The reason for making the spherical char par-
ticle assumption, besides the obvious computational simplicity af-
forded therefrom, was that upon high-temperature pyrolysis most
types of biomass char particles were observed to undergo vari-
ous degrees of spherodization, as documented in Refs. [59,64]. The
gaseous combustion products of chars may be a mixture of car-
bon monoxide or carbon dioxide, but several studies [82–85] have
documented that at elevated temperatures the formation of CO is
more favored over that of CO2 .
Combustion of a carbon particle, of diameter x, in thermal equi-
librium with its surroundings has been considered before [86,87],
considering the rate of heat generation and the rate of heat loss to
be equal. Hence,
Hg = Hc + Hr
where Hg denotes the rate of heat generated by chemical reactions,
Hc is rate of heat loss by conduction, and Hr is rate of heat loss by
radiation. In order to find the equilibrium particle temperature, it
is necessary to express each of the three terms in Eq. (1) as a func-
tion of particle temperature, assuming that there is no temperature
gradient within the particle. The rate of heat loss by conduction is
given by:
 0.75
Ts + Tg (Ts − Tg )
Hc = 2λ0 × (2)
2Tg x
Whereas the rate of heat loss by radiation is given by:
 
Hr = σ Ts4 − Tw4 ,
where Tw is effective wall temperature of the surfaces which sur-
round the particle.
Finally, the rate of heat generation is given by:
Hg = qQ
where Q is heat release at surface per unit mass of carbon burnt
and q is the overall particle burning rate per unit external surface
area. The parameter q can be expressed in terms of diffusional and
surface reaction rate coefficients, kdiff and ks as follows:
pg
q=  
1 1
+
kdi f f ks
The surface reaction rate, ks , is given by an Arrhenius-type ex-
pression that represents the dependence of the rate constant of a
chemical reaction on the absolute temperature, the activation en-
ergy, and a pre-exponential frequency factor [88]:
 
E ii. The effects of changing the particle diameter, x, are shown in
ks = A1 exp −
RT
The mass diffusion across the boundary layer surrounding the
burning particle yields the following expression for the diffusional
reaction rate coefficient:
48 D
kdi f f =
xRTm
The set of the fuel’s physical and chemical properties used in
this calculation are summarized in Table A1.
Experimentally-measured and theoretically-predicted particle
char temperatures are superimposed in Fig. A1, as functions of the
initial torrefied particle size cut.
The experimental char temperatures are taken from Fig. 11 and
the predicted char temperatures were calculated using the model
described above. Both the experimental and the calculated results
support the notion that the char temperatures increase mildly
with the char particle size and then, upon reaching a maximum
value, they decrease mildly. Moreover, as the initial torrefied parti-
cle size increases, both heat loses and heat generation decrease.
81
(1) Fig. A1. Experimentally-measured and theoretically-predicted torrefied biomass
char temperatures versus initial biomass particle size. Char combustion took place
with air inside the DTF operated at Twall = 1400 K (Tgas = ∼1350 K).
Under the conditions of these experiments, the maximum tem-
perature corresponds to a biomass char size of 175 μm, which, in
turn, corresponds to an initial torrefied particle size of 180–212 μm.
This initial particle to char particle size correspondence was de-
termined from separate experiments in the DTF, where torrefied
biomass particles were pyrolyzed at high-temperature (1350 K) at
high-heating rates (104 K/s) in nitrogen, in the manner of Ref. [64].
The minimum temperature was calculated for a biomass char size
of 52 μm, which, in turn, was derived from initial torrefied parti-
cles with nominal sizes in the range of 75–90 μm.
(3)
Since there is uncertainty in some key property values used in
these calculations, a parametric analysis was conducted. Values of
the following parameters: particle emissivity, particle diameter, ac-
tivation energy, and pre-exponential factor were perturbed accord-
ing to prior data in the literature. Results from such perturbations
(4) on the calculated char particle temperatures are shown below:
i. The effect of the particle emissivity, ε , is shown Fig. A2a. Only
the heat radiation equation was affected by the emissivity of
particle. The emissivity of particle that was used in the mod-
eling results shown in Fig. A1, was 0.9. However, as emissiv-
ity values for carbonaceous particles have been reported in the
(5)
literature to span the range of 0.8–1 [77], the model was per-
turbed to reflect the results of the char particle temperatures
when ε took values in this range; results are shown in Fig. A2a.
When ε increased to 1, the particle temperature decreased by
32 K, whereas when ε decreased to 0.8, the particle tempera-
ture increased by also 35 K. This is because higher emissivity
increases the heat loss form a particle.
(6)
Fig. A2b. Both the diffusion term in the heat generation equa-
tion and the conduction term in the heat loss equation are af-
fected by a change in the particle diameter, whereas the chem-
ical term and the radiation term in these equations remain
unaffected. For example, torrefied Miscanthus particles in the
(7) size cut of 180–212 μm, upon pyrolysis generated char particles
with estimated diameters of 175(± 10) μm [64]. Char sizes for
this and other torrefied biomass particle size cuts are shown
in Table A1. The model was perturbed to reflect the results of
the char particle temperatures when the particle diameter, x,
took values in the range of 175(± 10), and results are shown in
Fig. A2b. The calculated effect of reducing the particle diame-
ter is to reduce the particle temperature for the 75–90 μm size
but increase the particle temperature for other particle sizes.
The effect of increasing the particle diameter is to increase the
calculated particle temperature for 75–90 μm size, but decrease
the particle temperatures for other particle sizes.
iii. The effects of perturbing the assumed activation energy, E,
in the kinetic expression of Eq. (6) are shown in Fig. A2c.
82 A. Panahi et al. / Combustion and Flame 194 (2018) 72–84

Table A1
Biomass particle properties and furnace operating parameters input to the model.

pg Partial pressure of oxygen, atm 0.21

A Pre-exponential factor, g/cm2 s atm 18 [88]
E Activation energy, cal/mole 190 0 0 [88]
R Gas constant, cal/mole K 1.986
Tg Gas temperature, K 1350 [54]
Ts Char surface temperature, K From Fig. 8
H Heat release at surface per unit mass of carbon burnt, cal/g 2340 [89]
D Diffusion coefficient of oxygen in the gas, cm2 /s Varies with Ts [90]
ks Surface reaction rate coefficient, g/cm2 s atm Eq. (2)
kdiff Diffusion reaction rate coefficient, g/cm2 s atm Varies
x Char particle diameter, cm Varies with size [64]
σ Stefan-Boltzmann constant, cal/cm2 s K4 1.36 × 10−12 10
 Emissivity 0.9 [64]
λ0 λ0 is the thermal conductivity at gas temperature, cal/cm s K 0.0 0 02
Q Heat release at surface per unit mass of carbon burnt, cal/g 2340 [89]

Fig. A2. Effects of the char particle: (a) emissivity, (b) assumed size, (c) assumed chemical reaction activation energy and (d) assumed chemical reaction pre-exponential fac-
tor on char particle temperatures versus initial torrefied biomass particle size. Char combustion took place with air inside the DTF operated at Twall = 1400 K (Tgas = ∼1350 K).

The calculated heat generation is affected by the value of
the activation energy. The activation energy used to obtain
the modeling results shown in Fig. A1 was 19,0 0 0 cal/mole
(or 79.5 kJ/mole mole
kJ
) [88]. This value was mildly perturbed to
18,0 0 0 cal/mole and to 20,0 0 0 cal/mole. When E increased to
20,0 0 0 cal/mole, the particle temperature decreased by 88 K,
whereas when E decreased to 18,0 0 0 cal/mole, the particle
temperature increased by 83 K. Lowering the activation en-
ergy barrier promotes facilitates the combustion reactions and,
thereby, increases the heat generation.
iv. The effect of perturbing the assumed pre-exponential factor,
A, in the kinetic expression of Eq. (6) is shown in Fig. A2d.
Only the heat generation equation term is affected by the pre-
exponential factor. The pre-exponential factor used to achieve
the modeling results shown in Fig. A1 was 18 g/cm2 s atm.
However, as pre-exponential factor values for biomass have
been reported to be in the range of 13–20 g/cm2 s atm [88], the
model was perturbed to reflect the results of the char parti-
cle temperatures when A took values in this range, as shown
in Fig. A2d. When the factor A increased to 20 g/cm2 s atm, the
particle temperature increased by 29 K, while when A decreased
to 13 g/cm2 s atm, the particle temperature decreased by 84 K.
References
[1] EIA, EIA total energy data, (2017).
[2] R. Lueken, K. Klima, W.M. Griffin, The climate and health effects of a USA
switch from coal to gas electricity generation, J. Appl. Energy 109 (2016)
1160–1166.
[3] E. Rokni, A. Panahi, X. Ren, Y.A. Levendis, Reduction of sulfur dioxide emissions
by burning coal blends, J. Energy Res. Technol. 138 (2016) 032204-1–032204-8.
[4] E. Rokni, A. Panahi, X. Ren, Y.A. Levendis, Curtailing the generation of sul-
fur dioxide and nitrogen oxide emissions by blending and oxy-combustion of
coals, Fuel 181 (2016) 772–784.
[5] D.A. Tillman, D.N. Duong, N.S. Harding, Solid fuel blending: principles, prac-
tices, and problems, Elsevier, Waltham, MA, 2012.
[6] A. Toptas, Y. Yildirim, G. Duman, J. Yanik, Combustion behavior of different
kinds of torrefied biomass and their blends with lignite, Bioresour. Technol.
177 (2015) 328–336.
 A. Panahi et al. / Combustion and Flame 194 (2018) 72–84
[7] J. Nadziakiewicz, M. Kozioł, Co-combustion of sludge with coal, Appl. Energy
75 (2003) 239–248.
[8] E. Rokni, X. Ren, A. Panahi, Y.A. Levendis, Emissions of SO2 , NOx , CO2 , and HCl
from co-firing of coals with raw and torrefied biomass fuels, Fuel 211 (2018)
363–374.
[9] A. Demirbas, Potential applications of renewable energy sources, biomass com-
bustion problems in boiler power systems and combustion related environ-
mental issues, Prog. Energy Combust. Sci. 31 (2005) 171–192.
[10] P. McKendry, Energy production from biomass (part 1): overview of biomass,
Bioresour. Technol. 83 (2002) 37–46.
[11] T. Bridgeman, J. Jones, I. Shield, P. Williams, Torrefaction of reed canary grass,
wheat straw and willow to enhance solid fuel qualities and combustion prop-
erties, Fuel 87 (2008) 844–856.
[12] B. Arias, C. Pevida, J. Fermoso, M.G. Plaza, F. Rubiera, J. Pis, Influence of tor-
refaction on the grindability and reactivity of woody biomass, Fuel Process.
Technol. 89 (2008) 169–175.
[13] J. Chew, V. Doshi, Recent advances in biomass pretreatment–torrefaction
fundamentals and technology, Renew. Sustain. Energy Rev. 15 (2011)
4212–4222.
[14] D.R. Nhuchhen, P. Basu, B. Acharya, A comprehensive review on biomass tor-
refaction, Int. J. Renew. Energy Biofuels 2014 (2014) 1–56.
[15] W.-H. Chen, J. Peng, X.T. Bi, A state-of-the-art review of biomass torrefac-
tion, densification and applications, Renew. Sustain. Energy Rev. 44 (2015)
847–866.
[16] S. Sadaka, S. Negi, Improvements of biomass physical and thermochemi-
cal characteristics via torrefaction process, Environ. Prog. Sustain. Energy 28
(2009) 427–434.
[17] Q.-V. Bach, Ø. Skreiberg, Upgrading biomass fuels via wet torrefaction: a re-
view and comparison with dry torrefaction, Renew. Sustain. Energy Rev. 54
(2016) 665–677.
[18] M. Van der Stelt, H. Gerhauser, J. Kiel, K. Ptasinski, Biomass upgrading by tor-
refaction for the production of biofuels: a review, Biomass Bioenergy 35 (2011)
3748–3762.
[19] D. Agar, M. Wihersaari, Bio-coal, torrefied lignocellulosic resources–key prop-
erties for its use in co-firing with fossil coal–their status, Biomass Bioenergy
44 (2012) 107–111.
[20] R.H. Ibrahim, L.I. Darvell, J.M. Jones, A. Williams, Physicochemical characterisa-
tion of torrefied biomass, J. Anal. Appl. Pyrolysis 103 (2013) 21–30.
[21] B. Colin, J.-L. Dirion, P. Arlabosse, S. Salvador, Quantification of the torrefac-
tion effects on the grindability and the hygroscopicity of wood chips, Fuel 197
(2017) 232–239.
[22] J. Peng, H. Bi, S. Sokhansanj, J. Lim, A study of particle size effect on biomass
torrefaction and densification, Energy Fuels 26 (2012) 3826–3839.
[23] R.B. Bates, A.F. Ghoniem, Biomass torrefaction: modeling of reaction thermo-
chemistry, Bioresour. Technol. 134 (2013) 331–340.
[24] M. Phanphanich, S. Mani, Impact of torrefaction on the grindability and fuel
characteristics of forest biomass, Bioresour. Technol. 102 (2011) 1246–1253.
[25] V. Repellin, A. Govin, M. Rolland, R. Guyonnet, Energy requirement for fine
grinding of torrefied wood, Biomass Bioenergy 34 (2010) 923–930.
[26] J. Jones, T. Bridgeman, L. Darvell, B. Gudka, A. Saddawi, A. Williams, Combus-
tion properties of torrefied willow compared with bituminous coals, Fuel Pro-
cess. Technol. 101 (2012) 1–9.
[27] P.C. Bergman, A. Boersma, R. Zwart, J Kiel, Torrefaction for biomass co-firing in
existing coal-fired power stations, Energy Centre of Netherlands, 2005 Report
No. ECN-C-05-013.
[28] J. Li, A. Brzdekiewicz, W. Yang, W. Blasiak, Co-firing based on biomass torrefac-
tion in a pulverized coal boiler with aim of 100% fuel switching, Appl. Energy
99 (2012) 344–354.
[29] A. Bharadwaj, L.L. Baxter, A.L. Robinson, Effects of intraparticle heat and mass
transfer on biomass devolatilization: experimental results and model predic-
tions, Energy Fuels 18 (2004) 1021–1031.
[30] M. Sami, K. Annamalai, M. Wooldridge, Co-firing of coal and biomass fuel
blends, Prog. Energy Combust. Sci. 27 (2001) 171–214.
[31] T. Grotkjær, K. Dam-Johansen, A.D. Jensen, P. Glarborg, An experimental study
of biomass ignition, Fuel 82 (2003) 825–833.
[32] P.T. Williams, A.R. Reed, High grade activated carbon matting derived from the
chemical activation and pyrolysis of natural fibre textile waste, J. Anal. Appl.
Pyrolysis 71 (2004) 971–986.
[33] A.R. Reed, P.T. Williams, Thermal processing of biomass natural fibre wastes by
pyrolysis, Int. J. Energy Res. 28 (2004) 131–145.
[34] J. Shankar Tumuluru, S. Sokhansanj, J.R. Hess, C.T. Wright, R.D. Boardman, A
review on biomass torrefaction process and product properties for energy ap-
plications, Ind. Biotechnol. 7 (2011) 384–401.
[35] D.A. Tillman, Biomass cofiring: the technology, the experience, the combustion
consequences, Biomass Bioenergy 19 (20 0 0) 365–384.
[36] M. de Souza-Santos, A feasibility study of an alternative power generation
system based on biomass gasification/gas turbine concept, Fuel 78 (1999)
529–538.
[37] X.L. Yin, C.Z. Wu, S.P. Zheng, Y. Chen, Design and operation of a CFB gasifica-
tion and power generation system for rice husk, Biomass Bioenergy 23 (2002)
181–187.
[38] A. Demirbas, Combustion characteristics of different biomass fuels, Prog. En-
ergy Combust. Sci. 30 (2004) 219–230.
[39] S. Mani, L.G. Tabil, S. Sokhansanj, Effects of compressive force, particle size and
moisture content on mechanical properties of biomass pellets from grasses,
Biomass Bioenergy 30 (2006) 648–654.
83
[40] V. Ganesan, K.A. Rosentrater, K. Muthukumarappan, Flowability and handling
characteristics of bulk solids and powders–a review with implications for
DDGS, Biosystems Eng. 101 (2008) 425–435.
[41] Y.D. Kim, C.W. Yang, B.J. Kim, K.S. Kim, J.W. Lee, J.H. Moon, W. Yang, U.Y. Tae,
U. Do Lee, Air-blown gasification of woody biomass in a bubbling fluidized bed
gasifier, Appl. Energy 112 (2013) 414–420.
[42] J. Marinelli, J.W. Carson, Solve solids flow problems in bins, hoppers, and feed-
ers, Chem. Eng. Prog. 88 (1992) 22–28.
[43] N. Anderson, W. Chung, D. Loeffler, J.G. Jones, A productivity and cost compar-
ison of two systems for producing biomass fuel from roadside forest treatment
residues, Forest Products J. 62 (2012) 222–233.
[44] G. Marrs, R. Zamora-Cristales, J. Sessions, Forest biomass feedstock cost sen-
sitivity to grinding parameters for bio-jet fuel production, Renew. Energy 99
(2016) 1082–1091.
[45] P.C. Bergman, J.H. Kiel, Torrefaction for biomass upgrading, 14th European
Biomass Conference Paris, France (2005), pp. 17–21.
[46] I. Adeyemi, I. Janajreh, T. Arink, C. Ghenai, Gasification behavior of coal and
woody biomass: validation and parametrical study, Appl. Energy 185 (2017)
1007–1018.
[47] J. Riaza, J. Gibbins, H. Chalmers, Ignition and combustion of single particles of
coal and biomass, Fuel 202 (2017) 650–655.
[48] M. Saha, B.B. Dally, P.R. Medwell, A. Chinnici, Effect of particle size on the
MILD combustion characteristics of pulverised brown coal, Fuel Process. Tech-
nol. 155 (2017) 74–87.
[49] A. Demirbaş, Sustainable cofiring of biomass with coal, Energy Convers. Manag.
44 (2003) 1465–1479.
[50] D. Gera, M. Mathur, M. Freeman, A. Robinson, Effect of large aspect ratio of
biomass particles on carbon burnout in a utility boiler, Energy Fuels 16 (2002)
1523–1532.
[51] J. Saastamoinen, M. Aho, A. Moilanen, L.H. Sørensen, S. Clausen, M. Berg,
Burnout of pulverized biomass particles in large scale boiler – single particle
model approach, Biomass Bioenergy 34 (2010) 728–736.
[52] J. Riaza, R. Khatami, Y.A. Levendis, L. Álvarez, M.V. Gil, C. Pevida, F. Rubiera,
J.J. Pis, Combustion of single biomass particles in air and in oxy-fuel condi-
tions, Biomass Bioenergy 64 (2014) 162–174.
[53] J. Riaza, R. Khatami, Y.A. Levendis, L. Álvarez, M.V. Gil, C. Pevida, F. Rubiera,
J.J. Pis, Single particle ignition and combustion of anthracite, semi-anthracite
and bituminous coals in air and simulated oxy-fuel conditions, Combust. Flame
161 (2014) 1096–1108.
[54] A. Panahi, Y.A. Levendis, N. Vorobiev, M. Schiemann, V. Scherer, Combustion
behavior of herbaceous and woody biomass, International Technical Confer-
ence on Clean Coal & Fuel Systems, Clearwater, Florida (2016).
[55] C. Mock, H. Lee, S. Choi, V. Manovic, Combustion behavior of relatively large
pulverized biomass particles at rapid heating rates, Energy Fuels 30 (2016)
10809–10822.
[56] M. Pach, R. Zanzi, E. Björnbom, Torrefied biomass a substitute for wood and
charcoal, 6th Asia-Pacific International Symposium on Combustion and Energy
Utilization (2002).
[57] R. Khatami, Y.A. Levendis, An overview of coal rank influence on ignition
and combustion phenomena at the particle level, Combust. Flame 164 (2016)
22–34.
[58] R. Ganguli, S. Bandopadhyay, Relationship between particle size distribution
of low-rank pulverized coal and power plant performance, J. Combust. 2012
(2012) Article ID 786920, 7 pages, doi:10.1155/2012/786920.
[59] N. Vorobiev, A. Becker, H. Kruggel-Emden, A. Panahi, Y.A. Levendis, M. Schie-
mann, Particle shape and Stefan flow effects on the burning rate of torrefied
biomass, Fuel 210 (2017) 107–120.
[60] X. Ren, R. Sun, X. Meng, N. Vorobiev, M. Schiemann, Y.A. Levendis, Carbon,
sulfur and nitrogen oxide emissions from combustion of pulverized raw and
torrefied biomass, Fuel 188 (2017) 310–323.
[61] Y.A. Levendis, K. Joshi, R. Khatami, A.F. Sarofim, Combustion behavior in air of
single particles from three different coal ranks and from sugarcane bagasse,
Combust. Flame 158 (2011) 452–465.
[62] Y.A. Levendis, A. Panahi, E. Rokni, X. Ren, Emissions from cofiring
coals, in: B.A. Sakkestad (Ed.), 40th International Technical Conference
on Clean Coal & Fuel Systems, Clearwater, Florida, USA (2015), pp.
533–544.
[63] Y.A. Levendis, K.R. Estrada, H.C. Hottel, Development of multicolor pyrometers
to monitor the transient response of burning carbonaceous particles, Rev. Sci.
Instrum. 63 (1992) 3608–3622.
[64] A. Panahi, Y.A. Levendis, N. Vorobiev, M. Schiemann, Direct observations on the
combustion characteristics of Miscanthus and Beechwood biomass including
fusion and spherodization, Fuel Process. Technol. 166 (2017) 41–49.
[65] P. Austin, C. Kauffman, M. Sichel, Ignition and volatile combustion of cellulosic
dust particles, Combust. Sci. Technol. 112 (1996) 187–198.
[66] A. Carvalho, M. Rabaçal, M. Costa, M. Alzueta, M. Abián, Effects of potassium
and calcium on the early stages of combustion of single biomass particles, Fuel
(2017).
[67] G. Simões, D. Magalhães, M. Rabaçal, M. Costa, Effect of gas temperature and
oxygen concentration on single particle ignition behavior of biomass fuels,
Proc. Combust. Inst. 36 (2017) 2235–2242.
[68] D. Magalhães, F. Kazanç, A. Ferreira, M. Rabaçal, M. Costa, Ignition behavior
of Turkish biomass and lignite fuels at low and high heating rates, Fuel 207
(2017) 154–164.
[69] Y.A. Levendis, R. Sahu, R.C. Flagan, G.R. Gavalas, Post-ignition transients in the
combustion of single char particles, Fuel 68 (1989) 849–855.
84 A. Panahi et al. / Combustion and Flame 194 (2018) 72–84
[70] P.A. Bejarano, Y.A. Levendis, Single-coal-particle combustion in O2 /N2 and
O2 /CO2 environments, Combust. Flame 153 (2008) 270–287.
[71] Y. Levendis, S. Nam, M. Lowenberg, R. Flagan, G. Gavalas, Catalysis of the com-
bustion of synthetic char particles by various forms of calcium additives, En-
ergy Fuels 3 (1989) 28–37.
[72] T. Panagiotou, Y. Levendis, M. Delichatsios, Measurements of particle flame
temperatures using three-color optical pyrometry, Combust. Flame 104 (1996)
272–287.
[73] J. Lédé, H.Z. Li, J. Villermaux, H. Martin, Fusion-like behaviour of wood pyroly-
sis, J. Anal. Appl. Pyrolysis 10 (1987) 291–308.
[74] R. Narayan, M.J. Antal, Thermal lag, fusion, and the compensation effect during
biomass pyrolysis, Ind. Eng. Chem. Res. 35 (1996) 1711–1721.
[75] R. Khatami, Y.A. Levendis, M.A. Delichatsios, Soot loading, temperature and size
of single coal particle envelope flames in conventional-and oxy-combustion
conditions (O2 /N2 and O2 /CO2 ), Combust. Flame 162 (2015) 2508–2517.
[76] D. Neves, H. Thunman, A. Matos, L. Tarelho, A. Gómez-Barea, Characterization
and prediction of biomass pyrolysis products, Prog. Energy Combust. Sci. 37
(2011) 611–630.
[77] R.E. Mitchell, W.J. McLean, Temperature measurements of single pulverized–
fuel particles by two-color pyrometry. II: particle-burning behavior, Sandia Na-
tional Labs, Livermore, CA. USA, 1983.
[78] R.E. Mitchell, Experimentally determined overall burning rates of coal chars,
Combust. Sci. Technol. 53 (1987) 165–186.
[79] I. Smith, The combustion rates of coal chars: a review, Symp. (Int.) Combust.
19 (1982) 1045–1065.
[80] M. Field, D. Gill, B. Morgan, P. Hawksley, Combustion of pulverized coal, British
Coal Utilization Research Association, Leatherland, UK, 1967.
[81] M.A. Field, Rate of combustion of size-graded fractions of char from a low-rank
coal between 1200 K and 2000 K, Combust. Flame 13 (1969) 237–252.
[82] Y.A. Levendis, A. Atal, J.B. Carlson, On the correlation of CO and PAH emissions
from the combustion of pulverized coal and waste tires, Environ. Sci. Technol.
32 (1998) 3767–3777.
[83] B.J. Waters, R.G. Squires, N.M. Laurendeau, R.E. Mitchell, Evidence for formation
of CO2 in the vicinity of burning pulverized carbon particles, Combust. Flame
74 (1988) 91–106.
[84] T. Zeng, W.B. Fu, The ratio CO/CO2 of oxidation on a burning carbon surface,
Combust. Flame 107 (1996) 197–210.
[85] C.R. Shaddix, F. Holzleithner, M. Geier, B.S. Haynes, Numerical assessment
of Tognotti determination of CO2 /CO production ratio during char oxidation,
Combust. Flame 160 (2013) 1827–1834.
[86] A.F. Kamenetskii, Diffusion and heat exchange in chemical kinetics, Princeton
University Press, 1955.
[87] L.A. Vulis, Thermal regimes of combustion, McGraw-Hill, 1961.
[88] N. Vorobiev, M. Geier, M. Schiemann, V. Scherer, Experimentation for char
combustion kinetics measurements: bias from char preparation, Fuel Process.
Technol. 151 (2016) 155–165.
[89] N.V. Lavrov, The thermodynamics of gasification and gas-synthesis reactions,
Pergamon Press, 1963.
[90] W. Bartok, A.F. Sarofim, Fossil fuel combustion: a source book, Wiley, 1991.