ROLL PRESS BRIQUETTING AND PELLETING

OF CORN STOVER AND SWITCHGRASS

N. Kaliyan, R. V. Morey, M. D. White, A. Doering

ABSTRACT. Corn stover and switchgrass, potential biomass feedstocks for bioenergy and bioproducts industries, are often
harvested during a limited harvest season and stored as bales with bulk densities of about 100 to 200 kg m-3. Because of low
bulk density, corn stover and switchgrass are difficult to handle, transport, store, and use in their natural forms. One of the
solutions to reduce these problems and the associated costs is to densify these biomass feedstocks into pellets or briquettes.
In this study, roll press briquetting characteristics of corn stover and switchgrass were studied using a pilot‐scale roll press
briquetting machine. Almond‐shaped briquettes 28.7 to 31.3 mm in length were made. Results showed that high‐durability
corn stover and switchgrass briquettes with bulk densities of 480 to 530 kg m-3 could be produced. This corresponds to about
a three‐ to five‐fold increase in bulk densities compared to those of bales. Briquettes produced with the roll press briquetting
machine had bulk densities (351 to 527 kg m-3), durabilities (39% to 90%), and crushing strengths (28 to 277 N) that were
somewhat less than, but in a range comparable to, the pellets (9.6 to 9.8 mm diameter) produced with a conventional ring‐die
pelleting machine. The bulk density, durability, and hardness of the pellets ranged from 528 to 610 kg m-3, 75% to 95%, and
148 to 224 N, respectively. Micro‐structural studies (chemical composition analyses, scanning electron microscopy imaging,
and UV auto‐fluorescence imaging) on grinds, briquettes, and pellets confirmed that highly dense, strong, and durable
briquettes and pellets from corn stover and switchgrass could be produced without adding chemical binders (i.e., additives)
by activating (softening) the natural binders such as water‐soluble carbohydrates, lignin, protein, starch, and fat in the
biomass materials by providing moisture and temperature in the range of glass transition of the biomass materials. No steam
conditioning was necessary to produce good‐quality corn stover briquettes. For switchgrass, a grind temperature of about
75°C obtained by steam conditioning was necessary to produce good‐quality briquettes. Roll press briquetting appears to
be a promising low‐cost, low‐energy, high‐capacity densification approach for commercial production of biomass briquettes.
Keywords. Biomass, Briquettes, Corn stover, Densification, Pelleting, Pellets, Roll press briquetting, Switchgrass.

T
o reduce dependence on fossil fuels, there is tre‐ inal form. One solution to these problems is densification of
mendous interest in using biomass materials such biomass materials into pellets or briquettes. Densification
as corn stover and switchgrass in the U.S. for pro‐ can increase the bulk density of biomass materials from an
ducing liquid transportation fuels (ethanol), com‐ initial bulk density (including baled density) of 40 to 200 kg
bined heat and power, chemicals, and bio‐products (DOE, m-3 to a final bulk density of 600 to 800 kg m-3 (Holley, 1983;
2005). In addition to numerous advantages, use of biomass Colley et al., 2006). Densification of biomass materials could
materials in place of fossil fuels would result in low emissions reduce the costs of transportation, handling, and storage. In
of greenhouse and acid gases (DOE, 2005). In order to make addition, because of uniform shape and size, densified prod‐
biomass materials available for a variety of applications, ucts can be more easily handled using existing handling and
challenges with the use of biomass materials in their original storage equipment, and can be easily adopted in direct‐
forms must be resolved. Because of high moisture content, combustion or co‐firing with coal, gasification, pyrolysis,
irregular shape and size, and low bulk density, biomass is and in other biomass‐based conversions.
very difficult to handle, transport, store, and utilize in its orig‐ In the U.S., pelleting and roll press briquetting/compac‐
tion technologies are widely used for densifying particulate
materials for various end uses (Pietsch, 2002). Pelleting
technology is commonly used for producing animal feed (Sit‐
Submitted for review in October 2008 as manuscript number FPE 7757;
approved for publication by the Food & Process Engineering Institute
kei, 1986), whereas the roll press is used for densifying coals,
Division of ASABE in March 2009. fertilizers, minerals, and metals (Pietsch, 2002). These two
The authors are Nalladurai Kaliyan, ASABE Member Engineer, densification approaches (pelleting and roll press briquetting
Research Associate, Department of Bioproducts and Biosystems with the configurations shown in fig. 1) would be readily
Engineering, University of Minnesota, St. Paul, Minnesota; R. Vance adapted to densifying biomass materials to use for renewable
Morey, ASABE Fellow, Professor, Department of Bioproducts and
Biosystems Engineering, University of Minnesota, St. Paul, Minnesota; energy applications. In pelleting equipment, the feed materi‐
Michael D. White, Director of Marketing, Bepex International LLC, al is pressed through open‐ended cylindrical holes (dies)
Minneapolis, Minnesota; and Alan Doering, Associate Scientist, made in the periphery of a ring (fig. 1a). One to three small
Agricultural Utilization Research Institute (AURI), Waseca, Minnesota. rotating rolls push the feed material into the die holes from
Corresponding author: R. Vance Morey, Department of Bioproducts and
Biosystems Engineering, University of Minnesota, 1390 Eckles Ave., St.
inside of the ring towards the outside of the ring. The skin
Paul, MN 55108; phone: 612‐625‐8775; fax: 612‐624‐3005; e‐mail: friction between the feed particles and the wall of the die re‐
[email protected]. sists the free flow of feed, and thus the particles are com‐

Transactions of the ASABE

Vol. 52(2): 543-555 E 2009 American Society of Agricultural and Biological Engineers ISSN 0001-2351 543
pressed against each other inside the die to form pellets. One
or two adjustable knives placed outside the ring cut the pellets
into desired lengths. Typical diameter of the pellets may Ring-die
range from 4.8 to 19.0 mm, and the length of the pellets may
Feed
range from 12.7 to 25.4 mm.
In a roll press briquetter/compactor, material is densified
by compression between two counter‐rotating rolls (fig. 1b). Roll
Initial densification of the material may occur through com‐
pressing the material with a tapered auger in the feed mecha‐ Die
nism. The primary purpose of this initial densification phase (hole)
is to remove air from low bulk density material. The final
compaction occurs because of high pressures created as the Knife
material flows between the two rolls. The roll surfaces have Pellets
pockets to form briquettes of desired size and shape when the
material passes between the rolls. The briquettes are easily (a)
separated and handled after leaving the machine. The densi‐
fied products are mostly pillow shaped with a size of 10 to Screw
40mm or larger. feeder
Pelleting would produce densified biomass (i.e., pellets)
suitable for home, business, and institutional heating where
high bulk density and high quality (i.e., high strength and du‐
rability) of densified biomass are required. On the other hand, Roll with
roll press briquetting would produce densified biomass (i.e., pockets
briquettes) suitable for applications requiring a large amount
of biomass materials such as cellulosic ethanol production or
combustion/gasification at ethanol plants where moderate
bulk density and moderate quality of densified biomass are
sufficient to facilitate the transportation, handling, and stor‐
Briquettes
age of biomass with minimum costs. Based on this rationale,
densification behaviors of the biomass materials (i.e., corn
stover and switchgrass) in pilot‐scale roll press briquetting
(b)
and conventional ring‐die pelleting machines were studied in
this research. In addition, in this research, more effort was Figure 1. Schematics of (a) conventional ring‐die pelleting and (b) roll
placed on roll press briquetting than on pelleting because press briquetting machines.
very limited fundamental studies on roll press briquetting of
biomass materials are available in the literature (e.g., Köser types of commercial densification machines (i.e., roll press
et al., 1982; Holley, 1983; Dec, 2002). Conversely, several briquetting and pelleting machines) was the overall objective
studies on the performance of conventional pelleting of bio‐ of this study. The specific objectives of this study were to:
mass materials have been conducted in the past (e.g., Hill and S Determine the roll press briquetting characteristics of
Pulkinen, 1988; Tabil and Sokhansanj, 1996; Samson et al., corn stover and switchgrass.
2000; Jannasch et al., 2004; Colley et al., 2006). S Determine optimum roll press briquetting conditions
(machine variables and biomass variables) for densify‐
OBJECTIVES ing corn stover and switchgrass.
Kaliyan and Morey (2006) studied the densification char‐ S Compare performances of the roll press briquetting
acteristics of corn stover and switchgrass using a lab‐scale machine and a conventional pelleting machine at se‐
piston‐cylinder uniaxial compression apparatus at a forming lected densification conditions for corn stover and
pressure of 150 MPa. Compression pressure of 150 MPa or switchgrass.
higher can be achieved in commercial roll press briquetting S Assess the binding mechanisms of corn stover and
and pelleting machines (Dec, 2002; Kaliyan and Morey, switchgrass by micro‐structural analyses.
2009). Kaliyan and Morey (2006) concluded that the particle The densification of biomass always involves size reduc‐
size obtained from a hammer mill screen size of 3.0 mm (geo‐ tion (i.e., grinding) and may require drying of biomass before
metric mean particle diameter of 0.56 to 0.66 mm), grind compressing of biomass in densification machines. In this re‐
moisture content of 15% (w.b.) for corn stover and 10% (w.b.) search, only the densification process was studied in detail.
for switchgrass, and preheating temperature of 75°C or
above would produce strong and high‐durability briquettes
from corn stover and switchgrass. They also concluded that
when the moisture content of the corn stover grind was 15%
MATERIALS AND METHODS
BIOMASS FEEDSTOCKS
(w.b.), there was no need to preheat the corn stover grind in
Corn stover was purchased from Mat Ag‐Fiber LLC
order to produce high‐quality corn stover briquettes. For
(Floodwood, Minn.) in October 2006 as coarsely ground
switchgrass, preheating to at least 75°C with a grind moisture
(i.e.,hammer milled) corn stover. Two particle sizes of corn
content of 10% (w.b.) was required to produce good‐quality
stover were obtained by grinding the coarse corn stover in an
switchgrass briquettes. Evaluation of these optimum densi‐
18.7 kW (25.0 hp) hammer mill (Jacobson Quality Machin‐
fication conditions for corn stover and switchgrass in two

544 TRANSACTIONS OF THE ASABE

ery, Minneapolis, Minn.) with two different screens with 2.4 fines that leaked through the rolls) to room temperature
and 4.0 mm (3/32 and 5/32 in.) openings. Switchgrass was (about 20°C), and screening the product using a vibratory
harvested as 1.2 × 1.2 × 2.4 m (4 × 4 × 8 ft) square bales screen with 5.2 mm openings to separate the fines and bri‐
in September 2006 from a field in Owatonna, Minnesota. quettes. The cleaned briquettes were used for measuring their
Size reduction of switchgrass was done in two steps. First, properties. The moisture content of the biomass grind was ad‐
switchgrass bales were ground in a 5.6 kW (7.5 hp) chopper justed by mixing a predetermined quantity of distilled water
(Agri‐Val, DC Atlas Co., Loyal, Wisc.) to a length of 101.6 with the grind in a mechanical mixer. Before briquetting, the
to 127.0 mm (4 to 5 in.). Then, fine grinding of the chopped biomass grinds were left in plastic containers for about 2 to
switchgrass was done using the hammer mill with two differ‐ 3 h for moisture equilibration.
ent screens with 2.4 and 4.0 mm (3/32 and 5/32 in.) openings. For the briquetting conditions where the temperature of
Moisture content of the biomass grinds was determined the grind should be higher than room temperature, the grind
using the procedure given in ASABE Standard S358.2 was steam conditioned using a Turbulizer (Bepex Interna‐
(ASABE Standards, 2006a). The moisture content values re‐ tional LLC, Minneapolis, Minn.) steam conditioner. The
ported in this article are on a wet basis (w.b.). Bulk density Turbulizer was a steam‐jacketed cylindrical container with
of the biomass grinds (wet basis) was calculated from the mixing paddles at its center. A feeder metered the grind to an
mass of grind that occupied a 250 mL glass container with airlock. From the airlock, the grind was fed to the inlet of the
80mm diameter. While measuring the bulk density of the Turbulizer. In the Turbulizer, paddles moved the grind from
grinds, the glass container was manually filled by slowly dis‐ the inlet to the outlet while increasing the temperature of the
charging the samples to the container from a height of about grind to the required levels (within 30 s). Steam at 0.35 to
100 mm, and the container with the sample was tapped gently 0.83 MPag (50 to 120 psig) and 150°C to 178°C was input to
five times on a lab bench to remove large voids inside the the steam‐jacket of the Turbulizer to indirectly raise the tem‐
sample as well as to reduce the sample filling errors. There‐ perature of the grind. The paddles in the Turbulizer were op‐
fore, the bulk density measurements resulted in tapped bulk erated at 1200 rpm to obtain a grind temperature of 75°C or
density of the grinds. Particle size and particle size distribu‐ at 1650 to 1800 rpm to obtain 100°C or above. In addition,
tion of the biomass grinds were determined based on ASABE some amount of steam at 0.10 to 0.14 MPag (15 to 25 psig)
Standard S319.3 (ASABE Standards, 2006b). Corn stover and was also sprayed directly inside the Turbulizer to avoid any
switchgrass grind samples were sent to a forage analysis labo‐ possible moisture loss from the grind during steam condition‐
ratory (Dairy One, Ithaca, N.Y.; www.dairyone.com) to de‐ ing. More details on the moisture conditioning, steam condi‐
termine the compositions of these two biomass materials tioning, and roll press briquetting procedures can be found in
using near‐infrared reflectance (NIR) spectroscopy. The NIR Kaliyan and Morey (2007a). Figure 2 shows examples of corn
calibration standards developed by Dairy One (Ithaca, N.Y.) stover and switchgrass briquettes produced.
based on the Foss model 6500 spectrophotometer (Foss NIR‐ To estimate the throughput (kg h-1) of the roll press bri‐
Systems, Inc., Laurel, Md.) with Win ISI II v1.5‐AOAC quetting machine, the mass of briquettes collected for about
989.03 procedures for corn stover and grass hay, respectively, 4 to 10 min was weighed. The experimental setup of the pilot‐
were used to determine the compositions of corn stover and scale roll press briquetting machine did not allow accurate
switchgrass. Glass transition temperatures of the corn stover measurement of specific energy consumption.
and switchgrass grinds were determined using a differential
scanning calorimeter (DSC; Pyris‐1, PerkinElmer Life and PELLETING
Analytical Sciences, Shelton, Conn.) (Kaliyan and Morey, Pelleting experiments were conducted during February
2006). 2007 at the Agricultural Utilization Research Institute
(AURI), Waseca, Minnesota, using a pilot‐scale convention‐
ROLL PRESS BRIQUETTING al ring‐die pelleting machine (CPM Master model 818806,
Roll press briquetting experiments were conducted during California Pellet Mill (CPM) Co., San Francisco, Cal.). A
January 2007 at Bepex International LLC (Minneapolis, 29.8 kW (40.0 hp) motor powered the two corrugated rolls
Minn.) using a pilot‐scale roll press briquetting machine and the ring‐die in the pelleting chamber, and a 3.7 kW
(model CS‐25 compactor/briquetter, Bepex International (5.0hp) motor powered a screw conveyor below the feed
LLC, Minneapolis, Minn.). In the roll press briquetting ma‐ hopper, which directed the biomass grind to the pelleting
chine, the rolls were powered by a 3.7 kW (5.0 hp) motor, and chamber. The ring‐die was operated at a constant speed of
the screw feeder was powered by a 2.2 kW (3.0 hp) motor. 232 rpm. Before pelleting corn stover and switchgrass, the
The diameter and width of the rolls were 228.6 mm (9.0 in.) ring‐die was cleaned to remove rust and any previous plug‐
and 38.1 mm (1.5 in.), respectively. A constant force of ging in the holes of the ring‐die. To achieve effective cleaning
5.1kN mm-1 width of rolls (i.e., hydraulic system pressure of the ring‐die, a mixture of oat grain, fine sand, and vegeta‐
of 2600 psig) was applied to the rolls during briquetting of ble oil was extruded through the ring‐die for about 2 to 3 min.
corn stover and switchgrass. The pockets on the rolls pro‐ This start‐up procedure also preheated the ring‐die. In addi‐
duced almond‐shaped briquettes. The dimension of one tion, the same mixture was used to clean the ring‐die whenev‐
pocket was 31.75 mm (1.25 in.) length × 20.32 mm (0.8 in.) er there was a change in the pelleting condition, such as
width × 5.08 mm (0.2 in.) depth. The gap between the rolls pelleting switchgrass after pelleting corn stover.
was kept constant at about 1.0 mm (0.04 in.). Therefore, the The pelleting process involves moisture conditioning the
approximate dimension of an almond‐shaped briquette was biomass grind, and then feeding the moisture‐conditioned
31.75 mm length × 20.32 mm width × 11.16 mm depth. grind to the pelleting mill to make pellets with a die diameter
The densification process involves moisture/steam condi‐ of 9.5 mm (3/8 in.). The moisture content of the biomass
tioning the biomass grind, and then roll press briquetting the grinds was adjusted by mixing predetermined quantities of
grind, followed by cooling of the product (briquettes plus water with the biomass grinds in a mechanical mixer. No

Vol. 52(2): 543-555 545

 (a)

(a)

(b)

Figure 2. (a) Corn stover and (b) switchgrass briquettes produced.

steam conditioning was done either for corn stover or switch‐

grass grind for any of the pelleting tests. Hot pellets were
(b)
spread in a thin layer on a screen with 6.4 mm (1/4 in.) diame‐
ter round holes, which is less than the pellet mill die diameter, Figure 3. (a) Corn stover and (b) switchgrass pellets produced.
to naturally cool the pellets to room temperature (about
12.8°C). After cooling, the screen was mechanically shaken For roll press briquetting experiments, the effects of two
to collect fines less than 6.4 mm, and the cleaned pellets were particle sizes (grinds obtained from hammer mill screen sizes
used for measuring their properties. Kaliyan and Morey of 2.4 and 4.0 mm), two moisture levels (10% and 15% w.b.),
(2007a) describe the procedure involved for the pelleting of and three preheating (i.e., steam conditioning) temperatures
corn stover and switchgrass. Figure 3 shows examples of the (room temperature, 75°C, and 100°C) on briquetting of both
corn stover and switchgrass pellets produced. corn stover and switchgrass were studied. Although we
To calculate the throughput (kg h-1) of the pelleting ma‐ planned to study the effect of two moisture contents (10% and
chine, the mass of pellets collected for about 1.5 to 2.0 min 15% w.b.), the moisture contents of the moisture‐conditioned
was measured. During pelleting, the current input to the over‐ or steam‐conditioned grinds deviated from these two values
all operation of the pelleting machine was recorded using an and ranged from 7% to 20% (w.b.). In addition, it was diffi‐
ammeter to estimate the specific energy consumption for the cult to achieve the other experimental variables at exact pre‐
pelleting process. The specific energy consumption per unit set levels. During briquetting, the roll speed and screw feeder
mass of pellets produced (MJ t-1) was estimated by subtract‐ speed had to be adjusted simultaneously to match (i.e., syn‐
ing the energy consumption for running the empty pelleting chronize) them for each briquetting condition to produce
machine (i.e., no load condition) from the energy consump‐ consistent briquettes. If the speed of the screw feeder was too
tion during biomass pelleting (Kaliyan and Morey, 2007a). fast or the speed of the rolls was too slow, the biomass grinds
plugged the hopper just above the rolls, and thus no briquettes
MACHINE AND BIOMASS FEEDSTOCK VARIABLES STUDIED were formed. We then needed to stop the roll press and re‐
The ranges of biomass variables, such as particle size, move the plugged materials before further operation. If the
moisture content, and preheating temperature, that could pro‐ speed of the rolls was too fast, biomass grinds came out
duce good‐quality densified products in the pilot‐scale roll through the rolls without forming briquettes. For the roll
press briquetting and pelleting machines were selected based press briquetting machine used in this study, the lowest me‐
on the results from the laboratory‐scale densification studies chanically achievable roll speed was 4 rpm. Therefore, the
on corn stover and switchgrass (Kaliyan and Morey, 2006). speed of the rolls was controlled using a frequency controller
to achieve roll speeds of less than 4 rpm. The speeds of the

546 TRANSACTIONS OF THE ASABE

rolls and screw feeder were variables, and these speeds were chine (model 4206, Instron Corp., Canton, Mass.) at a com‐
recorded for each briquetting condition. pression speed of 2.54 mm min-1. Pellet hardness (along
To compare the densification performance of the roll press diameter) was measured using a hand‐operated Dillon Quan‐
briquetting machine with the conventional ring‐die pelleting trol basic force gauge (model BFG 500 N, Itin Scale Co., Inc.,
machine, pelleting experiments were conducted for selected Brooklyn, N.Y.). The crushing strength test for briquettes is
conditions. For pelleting experiments, the effects of particle equivalent to the hardness test for pellets because compres‐
size (grinds obtained from 2.4 and 4.0 mm hammer mill sive stress is applied to fail the briquettes and the pellets
screen sizes for corn stover, and grinds obtained only from (Raghavan and Conkle, 1991).
2.4mm hammer mill screen size for switchgrass), and two Angle of repose of briquettes and pellets (about 250 g sam‐
die (i.e., hole in the ring‐die) length to diameter (L/D) ratios ple) on a galvanized steel surface was measured using a
(5.3 and 6.0) were studied. The diameter of the die was kept tilted‐plane angle of repose apparatus (Mohsenin, 1986).
constant at 9.5 mm (3/8 in.) for all of the pelleting tests. Pre‐ Scanning electron microscopy (SEM) (Hitachi S3500N)
liminary pelleting trials showed that at a moisture content of images were taken for corn stover and switchgrass grinds,
15% (w.b.) or below, consistent pellets were not formed; pel‐ and for cross‐sections (i.e., fractured surfaces) of the bri‐
lets appeared to be soft with weak particle‐particle bonding. quettes and pellets. The SEM images were analyzed to under‐
Therefore, a constant moisture content of about 20% (w.b.) stand the binding mechanisms of corn stover and switchgrass.
was used for both corn stover and switchgrass for all of the Ultraviolet (UV) auto‐fluorescence images (Olympus IX70
pelleting tests. In addition, no steam conditioning was used inverted microscope; UV excitation at 330‐385 nm, dichroic
for any of the pelleting tests. mirror at 410 nm, emission at 420‐700 nm) of corn stover and
No external binding agents (i.e., additives) were used for switchgrass grinds, and of cross‐sections (i.e., fractured sur‐
any of the briquetting or pelleting experiments. faces) of the briquettes and pellets, were obtained to identify
the natural binders. According to Rost (1995), the color inter‐
PROPERTIES OF BRIQUETTES AND PELLETS pretation of UV auto‐fluorescence is: deep red for chloro‐
Properties of briquettes and pellets were measured imme‐ plasts; green or yellow‐green for protein compounds;
diately after forming, and after one week of storage at room brilliant blue or bluish‐white for lignin, cutin, suberin, or phe‐
temperature (about 23°C) in zip‐lock plastic bags to allow nolic acids such as ferulic acid; and whitish fluorescence for
more curing of particle‐particle bonds (Raghavan and Con‐ cutin (cuticle). In addition, pure carbohydrates (cellulose, he‐
kle, 1991). The properties measured were bulk density, dura‐ micellulose, and starch) and lipid/fat molecules do not fluo‐
bility, crushing strength or hardness, individual briquette and resce (Rost, 1995).
pellet dimensions, and moisture content of the briquettes and
pellets. The bulk density and durability are properties of a STATISTICAL ANALYSES
bulk sample of briquettes or pellets, whereas the crushing All statistical analyses were performed using SPSS 16.0
strength or hardness is a property of individual briquettes or for Windows (SPSS, Inc., Chicago, Ill.) at 5% significance
pellets. level. We first performed paired‐samples t‐tests on data for
Bulk density of the briquettes and pellets (wet basis) was bulk density or durability of briquettes/pellets taken immedi‐
measured from the mass of briquettes or pellets occupying a ately after forming and after one week of storage at room tem‐
1.0 L glass container with 125 mm diameter. While measur‐ perature to determine if there were significant differences. If
ing the bulk density of briquettes or pellets, the glass contain‐ there were no significant differences, then we pooled the data
er was manually filled by slowly discharging the samples to from the results at these two times in order to increase the
the container from a height of about 100 mm, and the contain‐ sample size for additional statistical analyses.
er with the sample was tapped gently five times on a lab The results of the paired‐samples t‐tests for briquettes
bench to remove large voids inside the sample as well as to showed that the bulk density and durability of the briquettes
reduce the sample filling errors. Therefore, the bulk density measured after one week of storage were significantly higher
measurements resulted in tapped bulk density of the bri‐ than the values measured immediately after forming (P <
quettes or pellets. The moisture content of the briquettes and 0.05). Therefore, pooling of the data measured at two differ‐
pellets was quantified based on ASABE Standard S269.4 ent times was not appropriate for briquettes, and hence fur‐
(ASABE Standards, 2006c). Individual briquette or pellet di‐ ther statistical analyses could not be done for the briquettes.
mensions were measured using a digital Vernier caliper. The briquetting results measured after one week of storage
Durability of briquettes and pellets was measured using are discussed based on the trends observed for the effects of
the tumbling can method given in ASABE Standard S269.4 particle size, moisture content, and steam conditioning tem‐
(ASABE Standards, 2006c). About 500 g of briquettes or pel‐ perature.
lets were tumbled in a tumbler at 50 rpm for 10 min. Then, The results of the paired‐samples t‐tests for pellets showed
the percentage of original weight of briquettes or pellets re‐ that the bulk density and durability of the pellets measured
tained on a 3.2 mm (1/8 in.) screen was calculated as the dura‐ after one week of storage were not significantly different
bility. The screen size of 3.2 mm, used for removing the fines from the values measured immediately after forming (P >
from the “good” quality briquettes or pellets after tumbling, 0.05). This indicates that the change in pellet mass and/or
was based on the recommendation by the U.S. Pellet Fuels In‐ volume and the effect of curing occurred during the one‐week
stitute (PFI, Arlington, Va.; www.pelletheat.or g). The U.S. storage period were negligible. Therefore, the data on bulk
Pellet Fuels Institute (Arlington, Va.) classifies particles that density or durability of pellets measured immediately after
are less than 3.2 mm as fines in packaged biomass pellets forming and after one week of storage at room temperature
used for heating purposes. were pooled to conduct independent‐samples t‐tests to inves‐
Crushing strength of the briquettes (along length and tigate the effect of die L/D ratio (5.3 and 6.0) for corn stover
width) was measured using an Instron universal testing ma‐ and switchgrass, and the effect of particle size (grinds ob‐

Vol. 52(2): 543-555 547

tained from 2.4 and 4.0 mm hammer mill screen sizes) for ROLL PRESS BRIQUETTING
corn stover. The paired‐samples t‐tests for briquettes showed that the
bulk density and durability of the briquettes measured after
one week of storage were significantly higher (by 31.1 kg m-3
in bulk density and 3.0% points in durability) than the values
RESULTS AND DISCUSSION measured immediately after forming (P < 0.05). This may be
CHARACTERISTICS OF BIOMASS FEEDSTOCKS
due to the change in the briquette mass and/or volume (possi‐
Currently, corn stover and switchgrass are collected from bly due to the drying and shrinkage of briquettes) and curing
the field as bales with bulk density of 100 to 200 kg m-3
of bonds during the one‐week storage period. Thus, the bri‐
(Shinners et al., 2003, 2006). Compared to the bulk density
quette properties measured after one week of storage were
of the bales, the increase in the bulk density of the biomass
used to evaluate the effects of particle size, moisture content,
feedstocks due to the size reduction (i.e., chopping and then
and steam conditioning temperature (table 3). The briquette
hammer milling) was small (table 1). Table 2 gives the com‐
properties measured immediately after forming the bri‐
positions of the corn stover and switchgrass grinds used for
quettes can be found in Kaliyan and Morey (2007a).
this study. In addition, in table 2, the composition values pro‐
vided by Mani et al. (2006) and DOE (2007) are compared. Effect of Roll Press Briquetting Machine Variables
The constituents such as lignin, crude protein, starch, crude The biomass briquettes produced were of almond shape with
fat, and water‐soluble carbohydrates, are “natural binders” in length of 28.7 to 31.3 mm. The combination of screw feeder
the biomass materials. These natural binders can be activated speed and roll speed that created consistent briquettes was dif‐
(softened or melted locally) either by high moisture or heat ferent for each briquetting condition; however, the differences
or steam to use their binding functionality (Kaliyan and in the respective speeds were small (table 3). The roll speed and
Morey, 2009). Lignin and hemicellulose were found to be screw feeder speed obtained from this study could serve as start‐
amorphous thermoplastic materials that would undergo plas‐ ing values for making briquettes in commercial‐scale roll press
tic deformation at temperatures in the range of their glass briquetting machines, although these speeds can be adjusted on
transition (i.e., softening temperatures) (Back and Salmen, the go. For briquetting cases without steam conditioning, the
1982). The glass transition occurs in corn stover and switch‐ temperature of the corn stover and switchgrass briquettes ranged
grass from 50°C to 113°C (Kaliyan and Morey, 2006). The from 51°C to 63°C due to the frictional heating of the grinds
mean glass transition temperature for both corn stover and during briquetting (table3). Because this temperature range is
switchgrass is 75°C (Kaliyan and Morey, 2006). The ash con‐ well within the range of glass transition temperature of corn
tents (i.e., mineral contents) in corn stover (11.2% d.b.) and stover and switchgrass (i.e., 50°C to 113°C), the natural binders
switchgrass (5.0% d.b.) show their relative abrasiveness to in the biomass materials would have been activated to some ex‐
densification equipment when there is high friction/shear tent to help produce durable particle‐particle bonding.
during densification, such as in pelleting. The higher the min‐ Throughput of the roll press briquetting machine ranged
eral content, the higher the abrasion (Kaliyan and Morey, from 16.8 to 32.6 kg of cleaned briquettes per hour for corn
2009). stover, and 9.6 to 30.6 kg of cleaned briquettes per hour for

Table 1. Properties of corn stover and switchgrass grinds used for the roll press briquetting and pelleting experiments.
Initial Grind Particle Size Bulk Density
Hammer Mill Screen Moisture Content[b] of Grinds[c] of Grinds[b]
Biomass Feedstock Used for Grinding[a] (% w.b.) (n = 3) (mm) (n = 3) (kg m‐3) (n = 3)
Corn Stover 2.4 mm (3/32 in.) 7.3 ±0.1 0.34 ±0.29 160.5 ±6.8
4.0 mm (5/32 in.) 8.5 ±0.04 0.36 ±0.35 139.1 ±7.8
Switchgrass 2.4 mm (3/32 in.) 9.2 ±0.01 0.49 ±0.38 219.9 ±2.7
4.0 mm (5/32 in.) 10.4 ±0.1 0.59 ±0.43 184.2 ±1.2
[a] Diameter of the holes in the screen.
[b] Mean ± standard deviation.
[c] Geometric mean particle diameter by mass (mm) ± geometric standard deviation of particle diameter by mass (mm).

Table 2. Compositions of corn stover and switchgrass grinds.

Corn Stover Switchgrass
This Study Mani et al. DOE (2007) This Study Mani et al. DOE (2007)
(% of dry matter) (2006) (range, (% of dry matter) (2006) (range,
Component (n = 1) (% of dry matter) % of mass) (n = 1) (% of dry matter) % of mass)
Cellulose 49.4 31.3 30.6‐38.1 43.8 44.3 27.8‐37.1
Hemicellulose 26.2 21.1 19.1‐25.3 28.8 30.0 22.4‐28.6
Lignin 8.8[a] 3.1[a] 17.1‐21.3 [a] 9.2[a] 7.4[a] 13.2‐22.5 [a]
Crude protein 3.6 8.7 NA[b] 3.9 1.6 NA
Starch 0.4 NA NA 1.0 NA NA
Crude fat 0.7 1.3 NA 0.9 1.9 NA
Water‐soluble carbohydrates 7.9 NA NA 2.2 NA NA
Ash 11.2 7.5 9.8‐13.5 5.0 5.5 2.5‐7.6
[a] Lignin values measured for the biomass materials used in this study and in Mani et al. (2006) were acid‐insoluble lignin contents, whereas the lignin
contents obtained from DOE (2007) were total lignin in the biomass materials.
[b] NA = data not available.

548 TRANSACTIONS OF THE ASABE

 Table 3. Properties of corn stover and switchgrass briquettes measured after one week of storage at room temperature (about 23°C).
Roll Press Briquetting Conditions (mean ±SD, if given) Properties of Briquettes (mean ±SD, if given)
Moisture Fines Throughput Temp.
Content Temperature Screw Leakage of Cleaned of Bri‐ Moisture Bulk Crushing Strength
of Grind of Grind Roll Feeder through Briquettes quettes Content Density Durability (N) (n = 3)
(% w.b.) (°C) Speed Speed Rolls[a] (kg h‐1) (°C)[b] (% w.b.) (kg m‐3) (%) Along Along
(n = 3) (n = 3) (rpm) (rpm) (n = 1) (n = 1) (n = 1) (n = 3) (n = 1) (n = 1) Length Width
Corn stover grind with particle size of 0.34 mm ±0.29 (hammer mill screen size = 2.4 mm (3/32 in.))
7.3 ±0.1 22.0 ±0.0 1.5 42.0 20.7% 19.2 57.0 7.1 ±0.1 480.8 66.7 174.7 ±61.9 122.6 ±65.1
15.0 ±0.0[c] 20.3 ±0.0 1.6 41.0 7.9% 20.0 51.0 13.9 ±0.1 422.4 86.5 88.1±17.9 138.5 ±35.5
9.9 ±0.2 77.3 ±5.9[d] 2.3 52.0 4.2% 32.6 >70.0 9.0 ±0.1 472.3 89.8 250.3 ±27.0 276.7 ±22.8
Corn stover grind with particle size of 0.36 mm ±0.35 (hammer mill screen size = 4.0 mm (5/32 in.))
16.6 ±0.6[c] 18.0 ±0.0 1.5 44.0 7.1% 16.8 54.0 12.3 ±0.1 452.5 87.5 128.2 ±28.3 169.3 ±40.1
14.8 ±0.3 75.4 ±0.6[d] 1.5 49.0 4.6% 20.5 >52.0 9.5 ±0.03 478.4 87.7 179.2 ±35.6 227.0 ±25.6
Switchgrass grind with particle size of 0.49 mm ±0.38 (hammer mill screen size = 2.4 mm (3/32 in.))
9.2 ±0.01 19.0 ±0.0 1.3 46.0 47.1% 9.6 58.0 8.0 ±0.04 467.3 39.5 NA[e] NA
19.8 ±0.5[c] 19.0 ±0.0 1.3 42.0 24.8% 21.2 63.0 12.0 ±0.1 351.1 50.2 27.5 ±7.3 28.5 ±7.3
11.7 ±0.2 68.3 ±7.6[d] 2.35 41.0 10.7% 30.6 75.0 9.8 ±0.1 419.7 70.0 58.9 ±4.2 170.8 ±58.2
10.1 ±0.1 109.4 ±10.7[d] 2.5 48.0 18.9% 26.9 >85.0 5.5 ±0.02 526.7 61.6 95.6 ±14.9 142.3 ±62.5
Switchgrass grind with particle size of 0.59 mm ±0.43 (hammer mill screen size = 4.0 mm (5/32 in.))
10.4 ±0.1 19.0 ±0.0 1.75 43.0 51.2% NA 54.0 7.6 ±0.1 433.1 39.1 NA NA
10.0 ±0.0 85.0 ±0.0[d] 1.92 41.0 23.2% NA 75.0 7.3 ±0.1 409.6 58.1 52.9 ±3.6 80.1 ±6.3
[a] Fines leakage through rolls (%) = (kg of fines × 100) / (kg of fines + kg of cleaned briquettes). The size of fines is <5.2 mm.
[b] Temperature of the briquettes measured immediately after forming. For cases with steam conditioning, temperature of the briquettes was less than the
temperature of the grinds because of faster cooling of briquettes exiting the roll press briquetting machine.
[c] The grind was moisture conditioned by adding distilled water to increase the moisture content of the grind before briquetting (no steam conditioning was
involved).
[d] The grind was steam conditioned to increase the temperature of the grind before briquetting. The corresponding moisture content of the grind was
measured after steam conditioning (but before briquetting) to include the effect of steam conditioning on the initial moisture content of the grind.
[e] NA = data not available. For these briquetting conditions, consistent briquette shape was not formed although the switchgrass particles agglomerated.

switchgrass (table 3). The throughput was mostly influenced increased the durability of briquettes (table 3). The lower
by the roll speed rather than the screw feeder speed. In gener‐ bulk density at the higher moisture content of the grind may
al, the faster the roll speed, the higher the throughput of bri‐ have been due to loss of mass due to the evaporation of water
quettes. For corn stover, 4.2% to 20.7% of the product from from the briquettes, and due to the possible expansion of the
the roll press briquetting machine was fines (<5.2 mm) with briquettes after leaving the roll press briquetting machine.
bulk densities of 231 to 357 kg m-3 (table 3). For switchgrass, The higher briquette durability at the higher moisture content
10.7% to 51.2% of the product was fines (<5.2 mm) with bulk may have been due to the activation of the water‐soluble car‐
densities of 252 to 392 kg m-3 (table 3). It was found that bohydrates in the biomass materials. Furthermore, when bri‐
when the percentage of fines through the rolls was lower, the quetted without steam conditioning, at the higher moisture
durability and crushing strength of the briquettes were higher. content, corn stover resulted in higher‐durability briquettes
During commercial‐scale production of briquettes, the fines (durability of 87% at a moisture content of about 15% w.b.)
can be recycled to produce good‐quality briquettes. than switchgrass (durability of 50% at a moisture content of
about 20% w.b.). This may be due to the presence of larger
Effect of Biomass Feedstock Variables
amounts of water‐soluble carbohydrates (i.e., natural binding
No previous studies on roll press briquetting of corn stover
components) in corn stover (7.9% d.b.) than in switchgrass
and switchgrass were found in the literature. For roll press
(2.2% d.b.).
briquetting of corn stover without steam conditioning, in‐
For the roll press briquetting of corn stover with a particle
creasing the particle size from 0.34 mm to 0.36 mm increased
size of 0.34 mm or 0.36 mm, increasing the temperature of
the bulk density, durability, and crushing strength of the bri‐
the grind from room temperature (i.e., 18°-22°C) to
quettes; however, the differences between the briquette prop‐
75°-77°C by steam conditioning increased the durability
erties at these two particle sizes were small because of the
and crushing strength of the briquettes, but decreased the
small difference between the two particle sizes tested
bulk density of the briquettes (table 3). In addition to activat‐
(table3). For briquetting of switchgrass, with and without
ing the water‐soluble carbohydrates, steam conditioning of
steam conditioning, increasing the particle size from
corn stover grind appeared to activate the binding functional‐
0.49mm to 0.59 mm decreased the bulk density and durabili‐
ities of the additional natural binding components such as lig‐
ty of the briquettes (table 3). Therefore, for large‐scale roll
nin, protein, starch, and fat, which need a grind temperature
press briquetting of corn stover and switchgrass, particle
close to the glass transition temperature (i.e., 75°C) for com‐
sizes with geometric mean particle diameters of 0.36 mm
plete activation. This improved binding effect due to the ele‐
(i.e., hammer mill screen size of 4.0 mm) and 0.49 mm
vated temperature of the grind was reflected in increased
(i.e.,hammer mill screen size of 2.4 mm), respectively, could
durability and crushing strength values of the corn stover bri‐
be used to produce good‐quality briquettes. quettes (table 3). However, the results suggest that if the corn
For roll press briquetting without steam conditioning of
stover grind moisture content is 15% (w.b.) or above, then
corn stover and switchgrass, increasing the moisture content
steam conditioning to increase the temperature of the grind
of the grinds decreased the bulk density of the briquettes, but

Vol. 52(2): 543-555 549

to about 75°C is not necessary in order to produce strong and PELLETING
durable corn stover briquettes. The paired‐samples t‐tests for pellets showed that the bulk
With no steam conditioning (i.e., briquetting at room tem‐ density and durability of the pellets measured after one week
perature), a consistent briquette shape was not formed during of storage were not significantly different from the values
roll press briquetting of switchgrass. However, the switch‐ measured immediately after forming (P > 0.05). This sug‐
grass grind was agglomerated, and the agglomerated product gests that the change in pellet mass and/or volume and the ef‐
had durability of 39% to 40% (table 3). Steam conditioning fect of curing that occurred during the one‐week storage
the switchgrass grind to a temperature of 68°C or 85°C re‐ period were negligible. Therefore, the data for the bulk densi‐
sulted in higher values of bulk density (410 to 420 kg m-3), ty and durability of the pellets measured immediately after
durability (58% to 70%), and crushing strength (53 to 171 N) forming and after one week of storage were pooled for statis‐
of briquettes compared to cases without steam conditioning tical analysis (table 4).
for both particle sizes. The improved strength and durability
Effect of Pelleting Machine Variables
of the switchgrass briquettes for the steam conditioning cases
A review by Kaliyan and Morey (2009) found that the fol‐
was probably due to the activation of the natural binding
lowing pellet machine variables influence the strength and
components in the switchgrass. Steam conditioning of
durability of the pellets: length to diameter (L/D) ratio of the
switchgrass grind to 109°C resulted in lower durability of bri‐
die, speed of the ring‐die, gap between the rolls and ring‐die,
quettes than those values measured at 68°C (table 3). This
the specific energy input to the pellet mill, and steam condi‐
may be due to the loss of moisture from the switchgrass when
tioning/high shear processing before pelleting, such as feed
heating the grind to more than 100°C. Because there was not
conditioning in an expander. In this study, only the effect of
enough moisture (at least 10% w.b.) available to soften the
the L/D ratio of the die (5.3 and 6.0) was tested, and other pel‐
natural binding components that require water for activation
let machine variables were kept constant. In addition, in this
(softening), the net natural binding effect could have been re‐
study, a constant die diameter of 9.5 mm was used to pellet
duced at the elevated temperature of 109°C.
corn stover and switchgrass. This resulted in pellets with di‐
The maximum switchgrass briquette durability obtained
ameter of 9.6 to 9.8 mm, length of 20.7 to 24.0 mm, and indi‐
in this study was 70%. We believe that higher switchgrass bri‐
vidual pellet density of 1029 to 1111 kg m-3. With a die
quette durability can be obtained in a commercial‐scale roll
diameter of 9.5 mm, switchgrass pellets with durability of
press briquetting machine. Kaliyan and Morey (2007b) stud‐
75% to 86% were produced in this study (table4). Colley et
ied several strategies to improve the durability of switchgrass
al. (2006) found that a die diameter of 4.76 mm produced
briquettes. They found that mixing of either 20% (wt.) corn
switchgrass pellets with durability of 78% to 97%. In addi‐
stover or 5% (wt.) lime at preheating temperatures of 75°C
to 100°C could increase the switchgrass briquette durability tion, for pelleting alfalfa, a die diameter of 6.1 to 6.4 mm was
found to produce good‐quality pellets (i.e., pellet durability
to >80%. Addition of 20% (wt.) corn stover would involve no
of up to 80%) (Hill and Pulkinen, 1988; Tabil and Sokhan‐
additional cost for briquetting of switchgrass if the price of
corn stover and switchgrass is about the same. sanj, 1996). In this study, we used a larger die diameter than
has been used in the past for pelleting biomass materials.
Selection of a larger die diameter would result in higher
throughput of pellets, and thus a die diameter of 9.5 mm could
be used for commercial production of corn stover and

Table 4. Properties of corn stover and switchgrass pellets.

Pelleting Conditions (mean ±SD, if given) Properties of Pellets (mean ±SD, if given)
Moisture Fines Throughput Hardness
Content Temp. of Output of Cleaned Temp. of Moisture Bulk (along
of Grind Grind Die along with Pellets Pellets[c] Content Density Durability diameter)
(% w.b.) (°C) L/D Pellets[b] (kg h‐1) (°C) (% w.b.)[d] (kg m‐3)[e] (%)[e] (N)[d]
(n = 3) (n = 1) Ratio[a] (n = 1) (n = 1) (n = 1) (n = 3) (n = 4) (n = 4) (n = 10)
Corn stover grind with particle size of 0.34 mm ±0.29 (hammer mill screen size = 2.4 mm (3/32 in.))[f]
19.8 ±0.2 12.2 5.3 1.1% 180.9 67.2 15.4 ±0.1 554.6 ±10.4 94.4 ±0.4 200.9 ±52.5
19.3 ±0.1 11.1 6.0 4.4% 225.5 75.6 14.2 ±0.2 609.9 ±8.9 95.2 ±1.1 224.2 ±99.3
Corn stover grind with particle size of 0.36 mm ±0.35 (hammer mill screen size = 4.0 mm (5/32 in.))[g]
21.8 ±0.3 11.1 6.0 1.2% 162.6 75.6 15.1 ±0.2 547.6 ±10.6 94.4 ±0.7 196.5 ±52.4
Switchgrass grind with particle size of 0.49 mm ±0.38 (hammer mill screen size = 2.4 mm (3/32 in.))[h]
20.8 ±0.3 11.1 5.3 3.9% 85.9 70.0 12.2 ±0.1 527.9 ±7.3 75.3 ±3.2 148.1 ±81.8
20.0 ±0.9 11.1 6.0 4.3% 176.3 81.1 10.7 ±0.2 570.0 ±7.1 85.6 ±1.2 216.3 ±61.6
[a] L = length of the holes in the ring‐die; D = diameter of the holes in the ring‐die (D = 9.5 mm).
[b] Fines output along with pellets (%) = (kg of fines × 100) / (kg of fines + kg of cleaned pellets). The size of fines is <6.4 mm.
[c] Temperature of pellets measured immediately after forming (°C).
[d] Moisture content of pellets and pellet hardness were measured after one week of storage of pellets at room temperature (about 23°C).
[e] Bulk density and durability of pellets are average values for the pooled data measured immediately after forming and after one week of storage of pellets at
room temperature (about 23°C).
[f] For corn stover with particle size of 0.34 mm, the die L/D ratio of 5.3 versus 6.0 produced significantly different bulk densities (P < 0.05), but not
durabilities (P > 0.05).
[g] At the die L/D ratio of 6.0, the corn stover grind with particle size of 0.34 mm versus 0.36 mm resulted in significantly different bulk densities (P < 0.05),
but not durabilities (P > 0.05).
[h] For switchgrass with particle size of 0.49 mm, the die L/D ratio of 5.3 versus 6.0 produced significantly different bulk densities and durabilities (P < 0.05).

550 TRANSACTIONS OF THE ASABE

switchgrass pellets. In general, increasing the L/D ratio of the grass pellets. In addition, Colley et al. (2006) manufactured
die from 5.3 to 6.0 increased the temperature of the pellets, switchgrass pellets (die diameter = 4.76 mm) with bulk densi‐
bulk density of the pellets, durability of the pellets, and hard‐ ty of 687 kg m-3, durability of 96%, and pellet hardness of
ness of the pellets (table 4). For the corn stover grind with par‐ 27N at a pellet temperature of 85°C and pellet moisture con‐
ticle size of 0.34 mm, the die L/D ratio of 6.0 resulted in 55.3 tent of 11% (w.b.). Likewise, in this study, switchgrass pellets
kg m-3 higher bulk density of pellets (P < 0.05), but similar (die diameter = 9.5 mm) with bulk density of 570 kg m-3, du‐
pellet durability (P > 0.05), than for the die L/D ratio of 5.3. rability of 86%, and pellet hardness of 216 N at a pellet tem‐
At the die L/D ratio of 6.0, the corn stover grind with particle perature of 81°C and pellet moisture content of 11% (w.b.)
size of 0.34 mm resulted in 62.3 kg m-3 higher bulk density were produced (table 4).
of pellets (P < 0.05), but similar pellet durability (P > 0.05), The main difference between this study and Colley et al.
than for the corn stover grind with particle size of 0.36 mm. (2006) is that Colley et al. (2006) steam conditioned the
For the switchgrass grind with particle size of 0.49 mm, the switchgrass grind to raise the temperature of the grind to
die L/D ratio of 6.0 resulted in 42.0 kg m-3 higher bulk densi‐ 75°C before pelleting in a lab‐scale, low‐power pelleting ma‐
ty and 10.3% points higher durability of pellets than for the chine (1.5 kW), whereas in this study, steam conditioning was
die L/D ratio of 5.3 (P < 0.05). Therefore, the L/D ratio of the not done before pelleting the switchgrass in a pilot‐scale,
die of 6.0 could be chosen for commercial‐scale pelleting of high‐power pelleting machine (29.8 kW). The results suggest
corn stover and switchgrass. that, considering the cost of grinding, corn stover grind with
A peripheral ring‐die speed of 4 to 5 m s-1 has been sug‐ a geometric mean diameter of 0.36 mm (i.e., hammer mill
gested to expel a large volume of air during pelleting low bulk screen size of 4.0 mm) could be chosen for producing high‐
density feed (Heinemans, 1991; Tabil and Sokhansanj, 1996; quality corn stover pellets in commercial‐scale pelleting ma‐
Kaliyan and Morey, 2009). Tabil and Sokhansanj (1996) pro‐ chines (table 4). For switchgrass, results from this study and
duced highly durable alfalfa pellets at a ring‐die speed of from Colley et al. (2006) showed that switchgrass grind with
250rpm (2.8 m s-1). In this study, the speed of the ring‐die geometric mean diameters of 0.49 mm to 0.87 mm (i.e., ham‐
used was 232 rpm, which is equivalent to a peripheral ring‐ mer mill screen sizes of 2.4 to 3.2 mm) could be used to pro‐
die speed of 4.3 to 4.4 m s-1. With this speed of the ring‐die duce highly durable switchgrass pellets in commercial‐scale
(i.e., 232 rpm), highly durable pellets were produced from pelleting machines.
corn stover (pellet durability of up to 95%) and switchgrass
(pellet durability of up to 86%). COMPARISON OF ROLL PRESS BRIQUETTING WITH
Although no steam conditioning was used for pelleting PELLETING
corn stover and switchgrass, the temperature of the pellets The briquettes produced were of almond shape with a
was measured at 67°C to 81°C due to the frictional heating maximum size of 31.3 mm length × 23.3 mm width ×
of the grinds during pelleting (table 4). This suggests that the 17.9mm depth, and the briquettes had a maximum of 5.0 mm
increase in temperature of the grinds during pelleting could flashings. The pellets made were of cylindrical shape with a
have created temperatures in the range of glass transition of maximum size of 9.8 mm diameter × 24.0 mm length. High‐
the biomass grinds, and thus the natural binding components er bulk density, durability, and strength values were obtained
in the biomass materials would have been activated. for pellets than for briquettes. The briquettes had bulk densi‐
Throughput of the pelleting machine ranged from 163 to ties from a minimum of 1.8 to 2.6 times to a maximum of 3.5
226kg of cleaned pellets per hour for corn stover, and from to 5.3 times the bulk density of the bales (bulk density of the
86 to 176 kg of cleaned pellets per hour for switchgrass bales is 100 to 200 kg m-3). The pellets had bulk densities
(table4). In addition, 1.1% to 4.4% of fines (<6.4 mm) along from a minimum of 2.6 to 3.1 times to a maximum of 5.3 to
with pellets were collected as products (table 4). The fines 6.1 times the bulk density of the bales. The non‐uniform
were produced mainly by the action of the knife while cutting shape of the briquettes may have contributed to lower bulk
the pellets, the abrasion of pellets in the collection container, density of briquettes than that of pellets. In addition, during
and handling immediately after production before sufficient the durability test of briquettes, the loss of mass occurred due
curing could take place. In commercial pellet production, to the breakage of flashings around the almond‐shape of the
fines can be recycled to make good‐quality pellets. briquettes, but the almond‐shape was approximately pre‐
served. Thus, selection of a more uniform shape for the bri‐
Effect of Biomass Feedstock Variables
quette pockets on the rolls could result in higher bulk density
No previous pelleting study on corn stover was found in
the literature; however, Samson et al. (2000), Jannasch et al. and higher durability briquettes.
With no steam conditioning, the temperature of the bri‐
(2004), and Colley et al. (2006) studied the pelleting of
quettes was 51°C to 63°C, and the temperature of the pellets
switchgrass. Samson et al. (2000) and Jannasch et al. (2004)
produced switchgrass pellets with bulk density of 593 to was 67°C to 81°C. This rise in temperature of the products
was purely due to the heating provided by the friction/shear
641kg m-3 and hardness of >30 as measured by the Pfizer
between the biomass particles, and between the biomass par‐
tablet hardness tester. They found that switchgrass pellets
created large quantities of fines (about 43% of initial feed‐ ticles and the machine. In this study, we were unable to mea‐
sure the specific energy consumed by the roll press
stock dry matter) during the process of handling, screening,
briquetting machine. The specific energy consumed for the
and shaking before bagging of pellets; however, no pellet du‐
rability data were reported. Colley et al. (2006) reported that pelleting process was 0.95% to 1.31% of the energy in corn
stover (189 to 262 MJ t-1) and 2.37% to 2.43% of the energy
switchgrass grind with a moisture content of about 20%
in switchgrass (403 to 414 MJ t-1). The above specific energy
(w.b.) produced high‐quality pellets.
Similarly, in this study, a grind moisture content of about values did not include the energy consumed for running the
empty (i.e., no load) pelleting machine (Kaliyan and Morey,
20% (w.b.) resulted in good‐quality corn stover and switch‐
2007a). The power consumption for running the empty pel‐

Vol. 52(2): 543-555 551

leting machine was about 18 kW, which was equal to an aver‐ for corn stover and switchgrass pellets ranged from 19° to
age of 344 MJ t-1 for corn stover and 554 MJ t-1 for 21°. For comparison, the angle of repose of shelled corn
switchgrass. The energy content of the corn stover and ranges from 16° to 27° (Mohsenin, 1986). The angle of re‐
switchgrass was 20 and 17 MJ kg-1, respectively (Lemus et pose values provide an indication of flow‐ability of the bri‐
al., 2002; Pordesimo et al., 2005). quettes and pellets (Mohsenin, 1986). The moisture content
Komarek (1963) reported that the specific energy required of the corn stover and switchgrass briquettes ranged from 6%
for roll press briquetting of organic solids (i.e., biomass mate‐ to 14% (w.b.) (table 3). The moisture content of the corn stov‐
rials) was 16 to 32 MJ t-1. Drzymala (1993) reported that er and switchgrass pellets ranged from 11% to 15% (w.b.)
large‐scale roll press briquetting (up to 60 t h-1) of ores, metal (table 4). The storage stability of the briquettes and pellets is
chips, and coals required specific energy of 20 to 60 MJ t-1. determined by their moisture contents (Maier et al., 1992;
Dec (2002) found that roll press briquetting machines con‐ Colley et al., 2006). Future work is required to determine the
sumed about 18 to 36 MJ t-1 at optimum briquetting condi‐ safe storage moisture contents for briquettes and pellets made
tions for sodium chloride, quick lime, and municipal sludge. from corn stover and switchgrass.
Samson et al. (2000) reported that pelleting alfalfa, switch‐
grass, and straw required specific energy of 108, 162, and BINDING MECHANISMS OF CORN STOVER AND SWITCHGRASS
300MJ t-1, respectively, for pellet mill operation. Jannasch According to Rumpf (1962) and Pietsch (2002), bonding
et al. (2004) found that specific energy consumption by a pel‐ between particles can take place through a solid bridge or
leting mill was 268 MJ t-1 for pelleting switchgrass in a com‐ through inter‐particle attraction forces when there is no solid
mercial pelleting plant (2 t h-1). Thus, specific energy bridge. Solid bridges are developed due to partial melting of
consumption is likely to be less for both briquettes and pellets components, crystallization of soluble substances, chemical
produced in commercial‐scale systems optimized for pro‐ reaction, hardening of binders, adsorption layers due to high‐
duction. Since roll press briquetting relies primarily on nor‐ ly viscous binders (adhesion and cohesion forces due to the
mal compression with a small amount of friction, while binder layer between particles), and mechanical interlocking
ring‐die pelleting requires some normal compression and rel‐ of particles. When there is no solid bridge between particles,
atively high friction/shear (fig. 1), it is likely that lower spe‐ but the particles are brought together, short‐range attraction
cific energy consumption may be possible in briquetting forces such as molecular (valence forces, for instance, due to
compared to pelleting in commercial‐scale systems opti‐ free chemical bonds; hydrogen bridges; and van der Waals'
mized for production. forces), electrostatic, and magnetic forces can cause the par‐
The angle of repose values for corn stover and switchgrass ticles to adhere to each other.
briquettes ranged from 21° to 24°. The angle of repose values

(a1) Corn stover grind (particle size = 0.34 mm) (b1) Cross-section of a corn stover briquette (c1) Cross-section of a corn stover pellet
before briquetting or pelleting (no coatings of made with a particle size of 0.34 mm, made with a particle size of 0.34 mm,
natural binders on most of the particles). grind moisture content of 15% (w.b.), grind moisture content of 19% (w.b.), die
and no steam conditioning L/D ratio of 6, and no steam conditioning
(with coatings of natural binders). (with coatings of natural binders).

(a2) Switchgrass grind (particle size = 0.49 mm) (b2) Cross-section of a switchgrass briquette (c2) Cross-section of a switchgrass pellet
before briquetting or pelleting (some natural made with a particle size of 0.49 mm, made with a particle size of 0.49 mm,
binders are expressed on the particles grind moisture content of 12% (w.b.), grind moisture content of 20% (w.b.), die
due to size reduction). and steam conditioning to 685C L/D ratio of 6, and no steam conditioning
(with coatings of natural binders). (with coatings of natural binders).

Figure 4. Scanning electron microscopy (SEM) images of corn stover and switchgrass grinds, briquettes, and pellets (magnification at 600×). The SEM
images of the briquettes and pellets show that the particles are coated with natural binders, which are pressed out of the biomass cells during briquetting
or pelleting.

552 TRANSACTIONS OF THE ASABE

 In the roll press briquetting and pelleting machines, bio‐ cate that the natural binders that were coated on the particles
mass particles could have experienced pressures of 100 to in the briquettes or pellets were primarily lignin and protein
200 MPa (Dec, 2002; Kaliyan and Morey, 2009). Because of compounds (fig. 5). The molecules that did not fluoresce
the application of high pressures, particles were brought (i.e., black color) could be carbohydrates (cellulose, hemi‐
close together, causing inter‐particle attraction forces, and cellulose, simple sugars, and starch) and fat (fig. 5). The UV
the natural binding components in the corn stover and switch‐ auto‐fluorescence images of the shiny appearance of pellets
grass were squeezed out of the cells, which made solid confirmed that lignin was heavily coated on the outer surface
bridges between the particles. After cooling, these solid of the pellets (Kaliyan and Morey, 2008). Because lignin is
bridges hardened (i.e., curing process). This caused the bri‐ hydrophobic in nature, the lignin coating on the outer sur‐
quettes and pellets to become strong and durable. This postu‐ faces of the briquettes/pellets can make them water resistant
late on the binding mechanisms was verified through SEM (Anglés et al., 2001).
and UV auto‐fluorescence images. The SEM images of briquettes and pellets shown in fig‐
Figures 4a1 and 4a2 show that most of the corn stover par‐ ure4 are for the briquetting or pelleting conditions that re‐
ticles appear to be bare without any sign of coating of natural sulted in the highest durability and strength values. The SEM
binding components; however, on the switchgrass particles, images of briquettes and pellets made under conditions that
expression of some natural binding components (due to size resulted in lower durability and strength values also appeared
reduction) can be observed. The SEM images of the bri‐ to have similar coatings of natural binding components (Kali‐
quettes and pellets show that the particles are covered with yan and Morey, 2008). Therefore, the SEM images can only
a layer of natural binders (fig. 4). When viewed with light mi‐ explain the squeezing of natural binders from the particles
croscopy (Nikon SMZ 1500), these coatings appeared as due to the high pressure, but not the differences in the durabil‐
glassy/white sugar‐like coatings on the particles, and high ac‐ ity and strength values. The differences in the durability and
cumulation of these binding components was observed at the strength values can be explained by the extent of activation
junctions of particles (Kaliyan and Morey, 2008). The UV of the expressed natural binders through moisture or
auto‐fluorescence images of the briquettes and pellets indi‐ temperature or both. The amount of moisture required de-

(a1) Corn stover grind before briquetting (b1) Cross-section of a corn stover briquette (c1) Cross-section of a corn stover pellet
or pelleting (particle size = 0.34 mm). made with a particle size of 0.34 mm, made with a particle size of 0.34 mm,
grind moisture content of 15% (w.b.), grind moisture content of 19% (w.b.), die
and no steam conditioning. L/D ratio of 6, and no steam conditioning.

(a2) Switchgrass grind before briquetting (b2) Cross-section of a switchgrass briquette (c2) Cross-section of a switchgrass pellet
or pelleting (particle size = 0.49 mm). made with a particle size of 0.49 mm, made with a particle size of 0.49 mm,
grind moisture content of 12% (w.b.), grind moisture content of 20% (w.b.), die
and steam conditioning to 685C. L/D ratio of 6, and no steam conditioning.

Figure 5. UV auto‐fluorescence images of corn stover and switchgrass grinds, briquettes, and pellets (magnification at 145×). Kaliyan (2008) also docu‐
mented these images in color. Auto‐fluorescence color interpretation (Rost, 1995): green or yellow‐green for protein compounds; brilliant blue or
bluish‐white for lignin, cutin, suberin, or phenolic acids such as ferulic acid; and whitish fluorescence for cutin (cuticle). Pure carbohydrates (cellulose,
hemicellulose, and starch) and lipid/fat molecules do not fluoresce (Rost, 1995). The UV auto‐fluorescence images for briquettes and pellets indicate
that the natural binders that are coated on the particles in the briquettes and pellets are primarily lignin and protein compounds.

Vol. 52(2): 543-555 553

pends on the amount of natural binder (e.g., water‐soluble the pellets produced with the conventional pelleting ma‐
carbohydrates) available that requires moisture for activa‐ chine. Future work is required to compare the cost and energy
tion. For example, as discussed previously, at high moistures involved for commercial‐scale briquetting and pelleting of
(15% to 20% w.b.), corn stover resulted in stronger and more corn stover and switchgrass. Moreover, roll press briquetting
durable briquettes than switchgrass because of the presence appears to be a promising low‐cost, low‐energy, high‐
of a higher amount of water‐soluble carbohydrates than capacity approach for densifying corn stover and switchgrass
switchgrass. for renewable energy applications.
Glass transition temperature is the minimum temperature
required for activation of natural binders. The glass transition ACKNOWLEDGEMENTS
in corn stover and switchgrass starts at 50°C (Kaliyan and The authors wish to acknowledge the Initiative for Re‐
Morey, 2006). Therefore, in addition to extracting the natural newable Energy and the Environment (IREE), University of
binding components from the particles (due to the application Minnesota, for providing support for this study; Bepex Inter‐
of high pressures by the densification machines), the natural national LLC, Minneapolis, Minnesota, for providing access
binding components should be fully activated to use their to their facility and assistance with the roll press briquetting
binding functionality by providing enough moisture and a study; and the Agricultural Utilization Research Institute
temperature in the range of the glass transition for the bio‐ (AURI), Waseca, Minnesota, for providing access to their fa‐
mass materials. cility and assistance with the size reduction of biomass feed‐
stocks and pelleting study. In addition, the authors would like
to thank Mr. Matthew De Kam (University of Minnesota) for
SUMMARY AND CONCLUSIONS his assistance with this study.
In this research, pilot‐scale roll press briquetting and pel‐
leting behaviors of corn stover and switchgrass were studied.
The optimum machine variables (screw feeder and roll REFERENCES
speeds in the roll press briquetting machine, and length to di‐ Anglés, M. N., F. Ferrando, X. Farriol, and J. Salvadü. 2001.
ameter (L/D) ratio of the die in the pelleting machine) and Suitability of steam exploded residual softwood for the
biomass feedstock variables (particle size, moisture content, production of binderless panels. Effect of the pre‐treatment
and steam conditioning temperature) that could produce severity and lignin addition. Biomass and Bioenergy 21(3):
strong and durable briquettes and pellets in commercial‐scale 211‐224.
ASABE Standards. 2006a. S358.2: Moisture measurement—
briquetting and pelleting machines were determined.
Forages. St. Joseph, Mich.: ASABE.
The briquettes made were of almond shape with a maxi‐ ASABE Standards. 2006b. S319.3: Method of determining and
mum size of 31.3 mm length × 23.3 mm width × 17.9 mm expressing fineness of feed materials by sieving. St. Joseph,
depth, and the briquettes had a maximum of 5.0 mm flash‐ Mich.: ASABE.
ings. The pellets made were of cylindrical shape with a maxi‐ ASABE Standards. 2006c. S269.4: Cubes, pellets, and crumbles—
mum size of 9.8 mm diameter × 24.0 mm length. Corn stover Definitions and methods for determining density, durability, and
briquettes had bulk density of 422 to 481 kg m-3, durability moisture content. St. Joseph, Mich.: ASABE.
of 67% to 90%, and crushing strength of 88 to 277 N, whereas Back, E. L., and N. L. Salmen. 1982. Glass transitions of wood
corn stover pellets had bulk density of 548 to 610 kg m-3, du‐ components hold implications for molding and pulping
rability of 94% to 95%, and hardness of 197 to 224 N. In addi‐ processes. Tappi J. 65(7): 107‐110.
Colley, Z., O. O. Fasina, D. Bransby, and Y. Y. Lee. 2006. Moisture
tion, switchgrass briquettes had bulk density of 351 to 527 kg
effect on the physical characteristics of switchgrass pellets.
m-3, durability of 39% to 70%, and crushing strength of 28 Trans. ASABE 49(6): 1845‐1851.
to 171 N, whereas switchgrass pellets had bulk density of 528 Dec, R. T. 2002. Optimization and control of roller press operating
to 570 kg m-3, durability of 75% to 86%, and hardness of 148 parameters. Powder Handling and Processing 14(3): 222‐225.
to 216 N. The bulk densities of corn stover and switchgrass DOE. 2005. Biomass as feedstock for a bioenergy and bioproducts
briquettes ranged from 1.8 to 5.3 times the bulk density of industry: The technical feasibility of a billion‐ton annual supply.
bales (bulk density of bales is 100 to 200 kg m-3). The bulk Oak Ridge, Tenn.: U.S. Department of Energy, Office of
densities of corn stover and switchgrass pellets ranged from Scientific and Technical Information. Available at:
2.6 to 6.1 times the bulk density of bales. www.osti.gov/bridge. Accessed 18 June 2005.
This study proved that high‐quality briquettes and pellets DOE. 2007. Biomass feedstock composition and property database.
Washington, D.C.: U.S. Department of Energy. Available at:
could be produced from corn stover and switchgrass without
www1.eere.energy.gov/biomass/feedstock_databases.html.
adding chemical binders (i.e., additives) by activating (soft‐ Accessed 25 April 2007.
ening) the natural binding components, such as water‐soluble Drzymala, Z. 1993. Industrial Briquetting—Fundamental and
carbohydrates, lignin, protein, starch, and fat, in the corn Methods. Studies in Mechanical Engineering, Vol. 13. Warsaw,
stover and switchgrass through moisture and temperature. In Poland: PWN‐Polish Scientific Publishers.
addition, this study showed that if the temperature rise of the Heinemans, H. 1991. The interaction of practical experience and the
biomass grinds due to frictional heating in the briquetting and construction of new pelleting and cooling machinery. Advances
pelleting machines is in the range of glass transition in Feed Tech. 6: 24‐38.
(i.e.,softening) temperature of the biomass materials Hill, B., and D. A. Pulkinen. 1988. A study of the factors affecting
(i.e.,>75°C), then strong and durable briquettes and pellets pellet durability and pelleting efficiency in the production of
dehydrated alfalfa pellets. Special report. Tisdale, Saskatchewan,
could be produced without steam conditioning.
Canada: Saskatchewan Dehydrators Association.
Briquettes produced with the roll press briquetting ma‐ Holley, C. A. 1983. The densification of biomass by roll
chine had bulk densities, durabilities, and crushing strengths briquetting. Proc. Institute for Briquetting and Agglomeration
that were somewhat less than, but in a range comparable to, (IBA) 18: 95‐102.

554 TRANSACTIONS OF THE ASABE

Jannasch, R., Y. Quan, and R. Samson. 2004. A process and energy mechanical properties of biomass pellets from grasses. Biomass
analysis of pelletizing switchgrass. Ste. Anne de Bellevue, and Bioenergy 30(7): 648‐654.
Quebec, Canada: Resource Efficient Agricultural Production Mohsenin, N. N. 1986. Physical Properties of Plant and Animal
(REAP‐Canada). Available at: www.reap‐canada.com. Materials. New York, N.Y.: Gordon and Breach Science
Kaliyan, N. 2008. Densification of biomass. Unpublished PhD diss. Publishers.
Twin Cities, Minn.: University of Minnesota, Department of Pietsch, W. 2002. Agglomeration Processes—Phenomena,
Bioproducts and Biosystems Engineering. Technologies, Equipment. Weinheim, Germany: Wiley‐VCH.
Kaliyan, N., and R. V. Morey. 2006. Densification characteristics of Pordesimo, L. O., B. R. Hames, S. Sokhansanj, and W. C. Edens.
corn stover and switchgrass. ASABE Paper No. 066174. St. 2005. Variation in corn stover composition and energy content
Joseph, Mich.: ASABE. with crop maturity. Biomass and Bioenergy 28(4): 366‐374.
Kaliyan, N., and R. V. Morey. 2007a. Roll press briquetting of corn Raghavan, J. K., and H. N. Conkle. 1991. Physical characteristic
stover and switchgrass: A pilot‐scale continuous briquetting measurements for reconstituted coal pellets. Proc. Institute for
study. ASABE Paper No. 076044. St. Joseph, Mich.: ASABE. Briquetting and Agglomeration (IBA) 22: 85‐96.
Kaliyan, N., and R. V. Morey. 2007b. Strategies to improve Rost, F. W. D. 1995. Fluorescence Microscopy. Vol. II. New York,
durability of switchgrass briquettes. ASABE Paper No. 076182. N.Y.: Cambridge University Press.
St. Joseph, Mich.: ASABE. Rumpf, H. 1962. The strength of granules and agglomerates. In
Kaliyan, N., and R. V. Morey. 2008. Binding mechanisms of corn Agglomeration, 379‐418. W. A. Knepper, ed. New York, N.Y.:
stover and switchgrass in briquettes and pellets. ASABE Paper John Wiley and Sons.
No. 084282. St. Joseph, Mich.: ASABE. Samson, R., P. Duxbury, M. Drisdelle, and C. Lapointe. 2000.
Kaliyan, N., and R. V. Morey. 2009. Factors affecting strength and Assessment of pelletized biofuels. Ste. Anne de Bellevue,
durability of densified biomass products. Biomass and Quebec, Canada: Resource Efficient Agricultural Production
Bioenergy 33(3): 337‐359. (REAP‐Canada). Available at: www.reap‐canada.com.
Komarek, K. R. 1963. The roll type briquette machine. Proc. Shinners, K. J., B. N. Binversie, and P. Savoie. 2003. Harvest and
Institute for Briquetting and Agglomeration (IBA) 8: 35‐37. storage of wet and dry corn stover as a biomass feedstock.
Köser, H. J. K., G. Schmalstieg, and W. Siemers. 1982. ASAE Paper No. 036088. St. Joseph, Mich.: ASABE.
Densification of water hyacinth‐basic data. Fuel 61(9): 791‐798. Shinners, K. J., G. C. Boettcher, R. E. Muck, P. J. Wiemer, and M.
Lemus, R., E. C. Brummer, K. J. Moore, N. E. Molstad, C. L. D. Casler. 2006. Drying, harvesting, and storage characteristics
Burras, and M. F. Barker. 2002. Biomass yield and quality of 20 of perennial grasses as biomass feedstocks. ASABE Paper No.
switchgrass populations in southern Iowa, USA. Biomass and 061012. St. Joseph, Mich.: ASABE.
Bioenergy 23(6): 433‐442. Sitkei, G. 1986. Mechanics of Agricultural Materials. New York,
Maier, D. E., R. L. Kelley, and F. W. Bakker‐Arkema. 1992. In‐line, N.Y.: Elsevier.
chilled air pellet cooling. Feed Mgmt. 43(1): 28‐32. Tabil, L., Jr., and S. Sokhansanj. 1996. Process conditions affecting
Mani, S., L. G. Tabil, and S. Sokhansanj. 2006. Effects of the physical quality of alfalfa pellets. Applied Eng. in Agric.
compressive force, particle size, and moisture content on 12(3): 345‐350.

Vol. 52(2): 543-555 555

556 TRANSACTIONS OF THE ASABE