Journal of Food Engineering 104 (2011) 632–638

Contents lists available at ScienceDirect

Journal of Food Engineering

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

The effects of freeze, dry, and wet grinding processes on rice flour properties
and their energy consumption
Peerapong Ngamnikom, Sirichai Songsermpong ⇑
Program of Food Engineering, Department of Food Science and Technology, Faculty of Agro-Industry, Kasetsart University, Chatuchak, Bangkok 10900, Thailand

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

Article history: This study attempted to replace the wet grinding process of rice with a freeze grinding process. The freeze
Received 29 September 2010 grinding process involved soaking the rice samples in liquid nitrogen before grinding in a dry grinding
Received in revised form 30 January 2011 machine. Three different types of grinders (hammer mill, roller mill, and pin mill) were used in both
Accepted 1 February 2011
the freeze and the dry grinding processes. Wet grinding resulted in significantly (P < 0.05) smaller aver-
Available online 21 February 2011
age particle size and a lower percentage of damaged starch than the alternative methods of grinding.
Freeze grinding, especially using the hammer mill significantly reduced both the average particle size
Keywords:
and the damaged starch content. Moreover, freeze grinding produced a higher yield after sieving in com-
Freeze grinding
Wet grinding
parison with dry grinding using an identical grinder. In particular, freeze grinding with the hammer mill
Dry grinding gave a significantly higher yield after sieving than dry grinding with the hammer mill. The wet grinding
Rice process had the significantly highest specific energy consumption (13,868 kJ/kg) due to the large con-
Rice flour sumption of electrical energy by the many machines in the process. The energy consumption of freeze
grinding was similar to dry grinding. Consequently, the freeze grinding process was a viable alternative
to the traditional wet grinding process.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction semidry grinding process, the properties of flour are intermediate

Rice (Oryza sativa L.) is one of the most important food crops in
the world. Broken rice, either from waxy rice or non waxy rice vari-
eties, is used to produce rice flour by different size reduction pro-
cesses. Rice flour is used to produce many kinds of food and
desserts such as noodles, breakfast cereals, unleavened breads,
snack food items, crackers, candies and baby foods (Bao and Berg-
man, 2004). Generally, there are three methods used to prepare
rice flour: wet grinding, semidry grinding, and dry grinding
(Chiang and Yeh, 2002). Wet grinding is a traditional method used
to prepare rice flour and incorporates five distinctively consecutive
processes: soaking, adding excess water during grinding, filtering,
drying, and sieving; this process includes the use of many ma-
chines and much manpower. The costs associated with the flour
loss, the high water consumption, the treatment of wastewater,
and the high energy consumption call for an alternative method
(Yeh, 2004). Dry grinding uses no water, does not generate waste
water, and moreover, consumes less energy. Broken rice is ground
with dry grinding machinery such as a hammer mill, pin mill, roller
mill, or disc mill, etc. However, the level of quality of many food
items (for instance noodles, made of dry ground flour) is not ade-
quate for many consumers (Naivikul, 2004; Yeh, 2004). In the
⇑ Corresponding author. Tel.: +66 2 562 5042; fax: +66 2 562 5021.
E-mail address: sirichai.so@ku.ac.th (S. Songsermpong).
to those of both dry and wet ground flour; in aspects of particle
size, viscosity, damaged starch, etc. The semidry grinding process
has three steps: soaking, drying to remove excess water (15–17%
wet basis; %wb), and grinding with dry grinding machinery (Nai-
vikul, 2004; Yeh, 2004). Nevertheless, such a method has some
drawbacks such as the extended duration necessary to adjust the
moisture content of the rice kernels, the excessive consumption
of energy needed for the drying procedure, the undue consumption
of water, and the generation of waste water. The type and method
of grinding potentially has a profound impact on the physicochem-
ical characteristics of the rice flours produced (Chen et al., 1999).
Yeh (2004) stated that both the model and design of the grinder af-
fect the performance, as well as the particle size of the flour. Gen-
erally, wet ground flour is better suited for the production of
traditional rice based products than dry ground flour and semidry
ground flour, regardless of whether the products are steamed or
baked, as wet ground flour results in the lowest amount of dam-
aged starch and the finest particle size (Chen et al., 1999).
When polished rice kernels are ground into rice flour, some
starch granules are damaged due to the mechanical action during
the grinding process (Nishita and Bean, 1982; Chen et al., 1999,
2003a; Solanki et al., 2005; Sharma et al., 2008). Chen et al.
(2003a) indicated that the damaged starch and the particle size
distribution were two key factors which affect the physicochemical
properties and the application suitability of rice flour. Additionally,

0260-8774/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jfoodeng.2011.02.001
 P. Ngamnikom, S. Songsermpong / Journal of Food Engineering 104 (2011) 632–638
Nomenclature
W wet grinding
R dry grinding with roller mill
H dry grinding with hammer mill
P dry grinding with pin mill
RF freeze grinding with roller mill
HF freeze grinding with hammer mill
PF freeze grinding with pin mill
several components, with the exception of carbohydrate, affect the
swelling of flour particles such as protein, ash, and lipids, etc., espe-
cially protein which is present in greater amount than the other
components in the rice kernel and can block gel formation of the
starch granule during heating (Naivikul, 2004). Damaged starch
is a big concern in flour production, as it is separate from the intact
granules impacting on both the solubility and the susceptibility to
enzymatic digestion. Yeh (2004) stated that some grinding
machinery causes higher temperatures than others and that is part
of the reason why such machinery yields flour that has more dam-
aged starch and does not function well. When the rice is ground,
amylopectin molecules are broken down into low molecular
weight fragments by the mechanical force applied. These amylo-
pectin fragments block the formation of structures of leached amy-
lose during heating which result in a less rigid structure (Han et al.,
2002). Rice flour with a high damaged starch content is rapidly hy-
drated and hydrolyzed by a- and b-amylase (Megazyme, 2008).
Flour with a fine particle size has more swelling power and is thus
more prone to form rigid gel structures than course particle rice
flour (Chen et al., 2003b).
Many parameters of the rice grinding process affect the charac-
teristics of flour such as the grinding method, grinding machinery,
the rice kernel hardness, the soaking process, rice cultivar (Chen
et al., 2003a). The important characteristics of rice flour which af-
fect the food quality, especially in rice noodles, are the particle size
distribution, the amount of damaged starch, and the chemical com-
position (such as amylose, protein, and amylopectin etc.).
A new method was investigated that uses a freeze grinding pro-
cess, which includes the freezing of rice with liquid nitrogen prior
to dry grinding. At extremely low temperature, the samples are
generally brittle and easily broken (Fan and Hsu, 1976; Gouveia
et al., 2002; Devi et al., 2009). The objectives of the present study
were to compare the performance of the freeze grinding, dry grind-
ing, and wet grinding processes on yield, damaged starch content,
average particle size, particle size distribution, microscopic struc-
tures, and energy consumption. Additionally, three types of grind-
ers (hammer mill, roller mill, and pin mill) were used in both the
freeze grinding and the dry grinding processes to compare the effi-
ciency of the different machinery.
2. Materials and methods
2.1. Materials
Thai milled rice (Leuang 11 variety) was used in this experi-
ment, with an amylose content of approximately 32–37% (dry ba-
sis) (Suksomboon and Naivikul, 2006). In each treatment in the
grinding processes, 2 kg of rice was ground.
2.2. Grinding machines
For wet grinding, a super mass colloider (MKPB6–2, Masuko
Sangyo, Japan) was used. For dry and freeze grinding, a hammer
633
A electric current (ampere)
V voltage (volt)
P.F. power factor
E electrical energy (kJ)
Es specific energy consumption (kJ/kg)
P electrical power (kW)
t time (sec)
mill (AP-S, Hosokawa, Japan), a roller mill (Quadrumat Junior,
C.W. Brabender, USA), and a pin mill (Ngow Huat Yoo Machinery,
Bangkok, Thailand) were used.
2.3. Grinding techniques
2.3.1. Wet grinding
The broken rice kernels were steeped in water for 4 h to soften
the rice kernels using ratio of water to rice of 2:1 (w/w). Then, the
samples (rice and water) were ground using the super mass collo-
ider. The flour slurry was poured into a thick cloth bag and centri-
fuged in a basket centrifuge for 10 min at 1492 rpm to remove the
excess water. The wet ground flour was dried in a hot air oven at
40 °C for 12 h to obtain a moisture content of approximately 12%.
The dried samples were ground by means of a hammer mill with
a 0.5 mm sieve. Flour samples were passed through a 100 mesh
sieve (Retsch, Germany) with 150 lm openings, packed in plastic
bags and stored in a bucket at ambient temperature (about 25–
28 °C).
2.3.2. Dry and freeze grinding
For dry grinding, the broken rice kernels were ground by the
hammer mill with a 0.5 mm sieve, the roller mill, and the pin mill.
Each sample was ground for two rounds to follow the manner of
wet grinding (using the super mass colloider in the first round
and the hammer mill in the second round).
For freeze grinding, the broken rice kernels were steeped in li-
quid nitrogen for approximately 1 min with a ratio of rice to liquid
nitrogen of 2:5 w/v. The frozen samples were ground with the
hammer mill, the roller mill and the pin mill. Subsequent to the
first round of grinding, the flour sample was dipped into liquid
nitrogen for 1 min prior to the second round of grinding.
Flour samples (dry and freeze ground) were passed through a
100 mesh sieve, packed in plastic bags and stored in a bucket at
ambient temperature.
2.4. Performance assessment
2.4.1. Sample temperatures and grinding times
A thermocouple (Rixen, T-60) was used to measure the tempera-
ture of each rice or flour sample. In the first round, the temperature
of the rice was measured before grinding and the temperature of the
flour was remeasured after grinding. The temperature of the flour
was measured again before grinding in the second round and remea-
sured after grinding in the second round. The temperature measure-
ments were closely monitored, as fluctuations in temperatures were
considered likely to affect the properties of flour. The grinding times
were recorded with a digital stop watch. The grinding times were
correlated with the temperature readings of samples.
2.4.2. Yield
The yield of rice flour was calculated using three formulas (Eqs.
(1)–(3)):
634 P. Ngamnikom, S. Songsermpong / Journal of Food Engineering 104 (2011) 632–638
Grinding yieldð%Þ ¼ ½Flour after grindingðgÞ=Weight of riceðgÞ  100
Sieving yieldð%Þ ¼ ½Flour after sievingðgÞ=Flour after grindingðgÞ  100
Process yieldð%Þ ¼ ½Flour after sievingðgÞ=Weight of riceðgÞ  100
The grinding yield denotes the yield of flour in the grinding part
of the process; the sieving yield represents the yield of flour that
passes through the 100 mesh sieve, whereas the process yield re-
fers to the yield of total flour that passes through the 100 mesh
sieve per 2 kg of rice.
2.4.3. Determination of the damaged starch contents
The damaged starch content in the flour samples was deter-
mined in accordance with the approved method 76-31 of the
American Association of Cereal Chemists (AACC, 2000), where
damaged starch granules are hydrated and hydrolyzed to malto-
saccharides plus a-limit dextrins by means of a carefully controlled
treatment with purified fungal a-amylase. The fungal a-amylase
treatment is designed to result in a near complete solubilization
of damaged granules with a minimum breakdown of the undam-
aged granules. This reaction is terminated with the addition of di-
lute sulfuric acid. Aliquots are treated with excess levels of purified
amyloglucosidase to cause the complete degradation of the starch
derived dextrin to glucose. The glucose levels in this experiment
were measured with a high purity glucose oxidase/peroxidase re-
agent mixture. With the application of a spectrophotometer (at
510 nm), the damaged starch was determined as a percentage of
flour weight on a dry basis.
2.4.4. Particle size analysis
The particle size distribution and average particle size were
determined with a Mastersizer (Malvern 2000, UK) with Scirocco
2000 units for the dry samples. The equipment uses a laser beam
to detect the individual particles. The flour samples were dried in
a hot air oven to reduce the moisture content to approximately
8–11%wb prior to the particle size analysis.
2.4.5. Energy consumption
Only electrical energy was used during the wet, dry, and freeze
grinding processes. The electrical current was measured with a
clamp meter (Kaise SK-7711, Japan), and the readings were used
to calculate the electrical energy consumption. Energy consump-
tion was calculated for all wet grinding processes. As for the dry
and freeze grinding processes, the analysis was performed with
the stated grinding machinery in each method.
The electrical power for single phase and three phase were cal-
culated using Eqs. (4) and (5), respectively:
P ¼ ðA  V  P:F:Þ=1000 for single phase
P ¼ ðA  V  1:73  P:F:Þ=1000 for three phase
The electrical energy consumption during grinding was calcu-
lated using Eq. (6):
E¼Pt
The specific energy consumption was calculated using Eq. (7):
Es ¼ E=WRS
The units are P = kW, A = ampere, V = volt, P.F. = no unit, E = kJ,
t = second, WRS (weight of rice sample) = kg, and Es = kJ/kg.
2.4.6. Microscopic structures of the ground particles
Microscopic images of the rice flour granules were examined
with a scanning electron microscope (SEM; SU-1500, Hitachi,
Japan). The samples were mounted on aluminum stubs with a dou-
ble sided tape and a gold coated sputter and investigated under the
ð1Þ SEM at an accelerated voltage of 15 kV in accordance with the
method described by Suksomboon and Naivikul (2006).
ð2Þ
2.5. Statistical analysis
ð3Þ
A randomized complete block design (RCBD) was applied. The
SPSS for Windows program, version 12.0 (IBM, USA), was em-
ployed for analyzing the statistical results obtained from three rep-
lications. The mean ± standard deviation (SD) was calculated for
each treatment. Analysis of variance (ANOVA) and the Duncan’s
multiple range test (DMRT) were used to compare the differences
between mean values at the 5% confidence level.
The temperature of each sample and the grinding times were cal-
culated as the mean ± SD. The yield, damaged starch content, average
particle size, and energy consumption were analyzed statistically.
3. Results and discussion
3.1. Sample temperatures and grinding times
The temperatures and grinding times in Table 1 show that the
temperature increase was dependent on the type of grinder, the
type of sample, and the grinding time. The temperatures of the
samples after grinding in the first and the second round were in-
creased as the rice kernels or rice flour absorbed mechanical en-
ergy which was transformed into heat. However, the heat did not
affect the physicochemical properties of the rice flour, as the tem-
perature did not reach the gelatinization temperature of rice flour
(approximately 62–65 °C from Rapid Visco Analyzer; RVA), and
thus did not modify the structure of the starch granules. Prior to
the first grinding process, the rice kernels used in the wet grinding
process had a lower temperature than in the dry grinding process
because of they had been soaked in water as an initial part of the
process. In the dry grinding process, the samples after grinding in
the first round had a higher temperature due to the large particle
size of the broken rice. But, the temperature prior to and post
grinding in the second round did not appear to change due to the
small particle size of the rice flour. In the freeze grinding process,
the temperatures of the samples prior to grinding in both the first
and second rounds were obviously lower than those of the other
grinding processes (less than 170 °C) as the broken rice was
dipped into liquid nitrogen for 1 min. After grinding, the actual
temperature of the sample increased rapidly. However, after the
grinding process the flour had a low temperature. The extremely
low temperature of the sample at input represents the brittleness
of the sample and easy breakage.
The grinding time for each grinding process was different due to
differences in the feeding capacity of each grinder with regard to
ð4Þ the rice or flour sample. For dry grinding, both the hammer mill
ð5Þ and pin mill generated more heat but used less grinding time dur-
ing both rounds than the roller mill. These results indicated that
the design of both the hammer mill and the pin mill (which employ
impact force) resulted in the generation of more heat than the
ð6Þ compressive force present in the roller mill. In freeze grinding, a
significant reduction of the temperature of the flour resulted from
the addition of liquid nitrogen to the rice. The grinding time af-
ð7Þ
fected the increase in the temperature of the rice flour in the freeze
grinding process. In freeze grinding, using the hammer mill in both
rounds shortened the grinding time so the flour temperature after
grinding was low, which was in stark contrast to the dry grinding
process using the hammer mill that resulted in higher flour tem-
peratures. The first round of freeze grinding with the roller mill re-
quired less grinding time than the pin mill so, after grinding, the
flour had a lower temperature. However, in the second round,
 P. Ngamnikom, S. Songsermpong / Journal of Food Engineering 104 (2011) 632–638 635

Table 1
Temperature of flour samples and grinding time with various grinding processes.

Grinding Temperature (°C) Surrounding Operated time (min)

process temperature (°C)
Before grinding After grinding 1st Before grinding After grinding 2nd round 1st round 2nd round
1st round round 2nd round
W 27.1 ± 0.8 31.1 ± 0.3 33.7 ± 0.6 39.5 ± 0.7 (with Hammer mill) 32.7 ± 2.3 7.87 ± 1.44 5.99 ± 1.24
R 31.9 ± 4.1 33.7 ± 0.9 33.5 ± 1.0 33.9 ± 0.8 32.0 ± 1.4 22.93 ± 8.88 23.10 ± 5.94
H 32.7 ± 2.9 46.5 ± 2.0 45.1 ± 2.1 44.7 ± 1.5 33.3 ± 2.1 8.98 ± 2.23 5.30 ± 0.89
P 31.3 ± 2.6 43.6 ± 3.1 41.3 ± 2.2 39.4 ± 0.7 33.1 ± 2.1 13.36 ± 1.93 15.72 ± 3.08
RF <170 6.7 ± 10.7 <170 13.7 ± 6.3 31.7 ± 1.8 10.70 ± 3.43 18.96 ± 5.85
HF <170 16.0 ± 14.8 <170 33.5 ± 15.1 30.8 ± 1.5 7.08 ± 0.95 8.59 ± 3.10
PF <170 25.2 ± 8.4 <170 3.6 ± 5.4 31.4 ± 0.5 13.34 ± 5.39 9.21 ± 2.49

Values shown are the mean ± standard deviation; (n = 3).

The sieving yield of wet grinding, dry grinding with the roller
Table 2
Flour yields with various grinding processes. mill, and freeze grinding with the roller mill were significantly
higher than the other grinding processes with yields of 84.74%,
Grinding process Grinding yield (%) Sieving yield (%) Process yield (%)
91.19%, and 91.24%, respectively. These three methods yielded
W 79.02c ± 6.62 84.74a ± 8.84 66.51b ± 3.54 flour with the greatest percentage of particles less than 150 lm,
R 88.97ab ± 9.25 91.19a ± 3.85 80.18a ± 12.60
yet the flour obtained from these three methods did not have a
H 96.06a ± 1.18 53.64c ± 10.13 50.40c ± 7.64
P 92.38ab ± 4.60 39.46d ± 3.25 35.68d ± 4.71
smaller average particle size than that obtained from the other
RF 86.39b ± 5.02 91.24a ± 1.67 78.83a ± 5.09 methods (see further elaboration below in the particle size analy-
HF 89.68ab ± 4.91 75.34b ± 6.32 67.22b ± 1.95 sis). The freeze grinding process reduced the particle size. The siev-
PF 91.58ab ± 1.61 34.04d ± 3.82 27.36d ± 4.89 ing yield of flour obtained by freeze grinding with the hammer mill
Values shown are the mean ± standard deviation; (n = 3). resulted in a significantly higher yield than that of the flour ob-
Means with different letters within the same column differ significantly (P < 0.05). tained by the dry grinding process with an identical grinder; how-
ever no significant divergence was observed for flour obtained
from the grinding processes using the roller mill and the pin mill.
Table 3 The process yield is an indication of the total flour obtained
Average particle size and range and damaged starch from various grinding processes.
from each process. The yields of dry ground flour and freeze
Grinding Average particle Range of particle Damaged starch ground flour produced using the roller mill were significantly high-
process size (lm) size (lm) (%db) er than the yields of flour from the other grinding processes. This
W 61.293d ± 3.579 1.905–239.883 2.78c ± 0.12 observation suggests that the roller mill was able to produce a
R 112.242a ± 1.238 1.259–316.228 10.73a ± 0.37 large amount of flour with a particle size less than 150 lm. How-
H 95.054bc ± 0.461 1.445–316.228 11.43a ± 1.10
ever, this does not confirm that these two flours had the smallest
P 99.752b ± 8.484 1.096–316.228 12.36a ± 1.61
RF 110.897a ± 1.997 2.512–316.228 7.21b ± 1.01 particle size. Both the wet ground flour and the freeze ground flour
HF 87.873c ± 4.913 1.445–316.228 5.88b ± 0.90 with the hammer mill had comparable yields. The dry ground flour
PF 104.768ab ± 8.513 3.802–363.078 12.27a ± 1.67 with the hammer mill, the dry ground flour with the pin mill, and
Values shown are the mean ± standard deviation; (n = 3). the freeze ground flour with the pin mill all had low yields.
Means with different letters within the same column differ significantly (P < 0.05). The grinding time (Table 1) did not relate to the yield of flour, as
%db = percentage on a dry basis. the action zone of the grinder was open on either side of both the
inlet and the outlet, so that the samples were not ground in a
closed chamber. If the samples were ground in a closed chamber
the roller mill required more grinding time so, after grinding, the for a long time, that would affect the flour properties, including
flour had a higher temperature because, in the second round, the yield. The yield was dependent on (1) the sample fed into the grin-
flour agglomerated after being treated with liquid nitrogen and der which determines whether there is any lagging of supply to the
that made it rather hard to feed the flour into the roller mill. action zone, which affects any size reduction; (2) the design of the
action zone of the machinery which differed between machines;
3.2. Yield (3) the actual grinding system; and (4) the properties of the rice
cultivar.
Table 2 shows that both dry and freeze grinding resulted in sig-
nificantly (P < 0.05) higher grinding yields than wet grinding. The 3.3. Particle size analysis and damaged starch content
wet ground flour yield was 79.02%, whereas the freeze and dry
ground flour yields were between 86% and 96%, respectively, as Particle size analysis was performed with a mastersizer to re-
in the wet grinding process, more flour was lost in the step of soak- cord the particle size distribution and to assess the average particle
ing, grinding, removing of excess of water, drying, and regrinding. size (Table 3 and Fig 1). Flour samples were assessed after the siev-
Yields were not significantly different for the dry and freeze grind- ing process, so the particles size did not exceed 150 lm. The results
ing but the sieving yield diverged significantly. Freeze grinding revealed that the wet ground flour was composed of the finest
gave a slightly lower yield than dry grinding with the same grinder, average particle size and had the smallest range in particle size.
because flour spurted out of the container during the pouring of li- The average particle size of wet ground flour was 61.293 lm. The
quid nitrogen into the flour prior to the grinding process in the sec- flour particles resulting from the wet grinding process were found
ond round. Flour obtained after grinding in the second round of to be finer than from the dry grinding process, which confirmed the
each process was sieved with a 100 mesh sieve to obtain the fine findings reported by Chen et al. (1999), Solanki et al. (2005), Suk-
flour particle size necessary for the production of many rice based somboon and Naivikul (2006), and Yeh (2004), and can be clearly
products. The physicochemical properties were subsequently seen in the particle size distribution graph in Fig 1. The wet ground
examined. flour consisted of two particle types, with size of 2–20 lm and 20–
636 P. Ngamnikom, S. Songsermpong / Journal of Food Engineering 104 (2011) 632–638
Fig. 1. Particle size distribution of rice flour from various grinding types.
Table 4
Energy used by each grinding process.
Grinding process Electrical energy (kJ) Specific energy
consumption (kJ/kg)
W 26,013a ± 158 13,868a ± 5516
R 1153bc ± 319 801b ± 526
H 658d ± 135 420b ± 123
P 1170b ± 204 795b ± 466
RF 769bcd ± 199 534b ± 333
HF 714cd ± 189 440b ± 59
PF 1144bc ± 303 718b ± 170
Values shown are the mean ± standard deviation; (n = 3).
Means with different letters within the same column differ significantly (P < 0.05).
150 lm. Additionally, the particle size distribution showed an in-
creased volume of fine particle sizes compared with both the dry
grinding process and the freeze grinding processes. Since wet
grinding involves a soaking process, rice kernels are softened and
easily broken (Chiang and Yeh, 2002). As for dry and freeze grind-
ing processes, the freeze grinding process not only reduced the
average size of the particles but furthermore, increased the volume
of small particles compared with the dry grinding process with
both the hammer mill and the roller mill. Conversely, the pin mill
might not provide sufficient levels of efficiency. The results in Ta-
ble 3 show that dry and freeze grinding, with the application of
an identical grinder, did not produce significantly different average
particle sizes. Both the impact force (the main force) and the shear
force (the minor force) in a hammer mill produce a smaller average
particle size than the compressive force of a roller mill (Barbosa-
Cánovas et al., 2005). Generally, a fine particle size has the highest
swelling power and form a rigid gel structure which results in
additional strength in comparison with a coarse particle size (Chen
et al., 2003b; Hatcher et al., 2002).
Flour from wet grinding had significantly the lowest damaged
starch content of 2.78% on a dry basis (%db), as the water in the
wet grinding process absorbed the heat during grinding, and the
plentiful moisture content of the rice kernels resulted in a soft
structure. Therefore, the flour is composed of fine particles and
has a low content of damaged starch (Chiang and Yeh, 2002).
Wet ground flour normally has a lower volume of damaged starch
in the flour than dry ground flour, as reported by Chen et al. (1999),
Solanki et al. (2005), Suksomboon and Naivikul (2006), and Yeh
(2004). The freeze grinding process reduced the damaged starch
content in the flour after using an identical grinder to the one used
in the dry grinding process. The roller mill reduced the damaged
starch content from 10.73%db to 7.21%db for the dry grinding pro-
cess and freeze grinding process, respectively. The damaged starch
content using the hammer mill reduced from 11.43%db to 5.88%db
for the dry grinding process and freeze grinding process, respec-
tively. Using the pin mill did not produce a significant between
the damaged starch content for the dry grinding process or the
freeze grinding process. Since, the rice before grinding in the freeze
grinding process has an extremely low temperature, it is more brit-
tle than rice at ambient temperature; it is easily broken and re-
ceives less mechanical action (Fan and Hsu, 1976; Gouveia et al.,
2002; Devi et al., 2009). Therefore, the flour from the freeze grind-
ing process had a lower damaged starch content and finer particle
size than the dry grinding process based on the results from partic-
ular grinding machines (only the hammer mill in this research).
3.4. Energy consumption
Energy results, based only on the grinding process, are shown in
Table 4. Wet grinding consumed significantly more electrical en-
ergy and specific energy consumption than the other grinding pro-
cesses because this method had several processes that used
electrical energy, including grinding with the wet grinding
machinery, removal of excess of water by centrifuging, drying to
adjust the moisture content, and regrinding with the hammer mill.
The electrical energy and specific energy consumption of dry and
freeze grinding with the same grinder were not significantly differ-
ent. Dry and freeze grinding with the hammer mill seemed to use
less electrical energy and have lower specific energy consumption
than the dry and freeze grinding methods. Moreover, flour made by
freeze grinding with the hammer mill had closely fine particle size
and lower damaged starch content with wet ground flour.
Wet grinding used more energy and water than dry and freeze
grinding; therefore, it had a higher cost. Industry needs to treat the
wastewater which results in additional extra cost. Freeze grinding
tended to produce flour with a finer particle size and a lower dam-
aged starch content than dry grinding. The cost of liquid nitrogen
was less than wastewater treatment for mass production. Addi-
 P. Ngamnikom, S. Songsermpong / Journal of Food Engineering 104 (2011) 632–638
Fig. 2. Microscopic structure of rice flour by SEM.
tionally, with the freeze grinding process, the plant area will be
cleaner than with wet grinding because there is no water. There-
fore, freeze grinding is a suitable process to use.
3.5. Microscopic structure
Photographs of wet, dry, and freeze ground rice flour are pre-
sented in Fig. 2. The microscopic structure of wet ground flour is
composed of both small and large particles. There were a greater
number of smaller particles than larger particles. The small parti-
cles (starch granules) were generally round in shape with a smooth
surface, whereas the large particles (flour particles) had a rough
surface and irregular shape. Dry and freeze ground flour samples
from each grinder were significantly different in appearance. The
freeze ground flour had more sharp breakage angles than the dry
ground flour, resulting from freezing the rice samples before
grinding.
4. Conclusions
The study showed that the freeze grinding process reduced the
average particle size and the damaged starch content due to the
extremely low temperature of the sample prior to grinding. In par-
ticular, freeze grinding using the hammer mill resulted both in a fi-
ner particle size and a lower content of damaged starch for the dry
and freeze grinding processes. In the freeze grinding process, the
rice had high brittleness and breakage. The freeze grinding resulted
in a smaller flour particle size than from dry grinding resulting in a
greater yield after sieving. The roller mill produced the highest pro-
cess yield of all methods. The freeze grinding process used less
water in the process and thus, produced less environmentally
problematic wastewater. Additionally, the energy consumption of
637
freeze grinding was lower than from wet grinding but the freeze
grinding and dry grinding consumed the same energy consump-
tion with the same grinder. Freeze grinding with the hammer mill
consumed a low amount of energy and resulted in a fine particle
size and low damaged starch content. In summary, the freeze
grinding process provided a suitable alternative to produce rice
flour as it uses no water and consumes less energy than the wet
grinding process and thus enables an environmentally friendly
production of rice flour. Additionally, the use of high performance
equipment in the freeze grinding process may produce a finer par-
ticle size and a lower damaged starch content than was achieved in
this experiment.
Acknowledgments
This research was financially supported by The Office of the
Higher Education Commission, Bangkok, Thailand. The scanning
electron microscope photographs were supported by the Agricul-
tural and Agro-Industrial Product Improvement Institute, Kasetsart
University, Bangkok, Thailand, and the COAX Group Corporation
Ltd., Bangkok, Thailand. The particle size analysis was carried out
with a mastersizer, Malvern 2000 that was supported by the Insti-
tute of Food Research and Product Development, Kasetsart Univer-
sity, Bangkok, Thailand, and Diethelm Ltd., Bangkok, Thailand.
References
American Association of Cereal Chemists (AACC), 2000. Approved Methods of the
American Association of Cereal Chemists. Methods: 76-31, 10th ed. American
Association of Cereal Chemists, St. Paul.
Bao, J., Bergman, C.J., 2004. The functionality of rice starch. In: Eliasson, A. (Ed.),
Starch in Food: Structure, Functional and Application. CRC Press, Cambridge, pp.
258–292.
638 P. Ngamnikom, S. Songsermpong / Journal of Food Engineering 104 (2011) 632–638
Barbosa-Cánovas, G.V., Enrique, O.R., Pablo, J., Hong, Y., 2005. Food Powders:
Physical Properties, Processing, and Functionality. Kluwer academic/Plenum
publishers, New York. pp. 157–173.
Chen, J.J., Lu, S., Lii, C.Y., 1999. Effect of milling on the physicochemical
characteristics of waxy rice in Taiwan. Cereal Chemistry 76 (5), 796–799.
Chen, J.J., Lu, S., Lii, C.Y., 2003a. Physicochemical and morphological on damaged
rice starch. Journal of Food and Drug Analysis 11 (4), 283–799.
Chen, Z., Schols, H.A., Voragen, A.G.J., 2003b. Starch granule size strongly determines
starch noodle processing and noodle quality. Journal of Food Science: Food
Chemistry and Toxicology 5 (68), 1584–1589.
Chiang, P.Y., Yeh, A.I., 2002. Effect of soaking on wet-milling of rice. Journal of Cereal
Science 35, 85–94.
Devi, A.F., Fibrianto, K., Torley, P.J., Bhandari, B., 2009. Physical properties of
cryomilled rice starch. Journal of Cereal Science 1, 1–7.
Fan, L.T., Hsu, K.H., 1976. A note on the characterization of cryomilled wheat flour
by scanning electron microscopy. Cereal Chemistry 53 (6), 979–982.
Gouveia, S.T., Lopes, G.S., Fatibello-Filho, O., Nogueira, A.R.A., Nóbrega, J.A., 2002.
Homogenization of breakfast cereals using cryogenic grinding. Journal of Food
Engineering 51, 59–63.
Han, Z.H., Campanella, O.H., Mix, N.C., Hamaker, B.R., 2002. Consequence of starch
damage on rheological properties of maize starch paste. Cereal Chemistry 79
(6), 897–901.
Hatcher, D.W., Anderson, M.J., Desjardins, R.G., Edwards, N.M., Dexter, J.E., 2002.
Effect of flour particle size and starch damage on processing and quality of
white salted noodles. Cereal Chemistry 79 (1), 64–71.
Megazyme International Limited, 2008. Starch Damage Assay Procedure. Megazyme
International Ireland Limited. pp. 1–11.
Naivikul, O., 2004. Rice: Science and Technology. Kasetsart University Press,
Bangkok.
Nishita, K.D., Bean, M.M., 1982. Grinding methods: their impact on rice flour
properties. Cereal Chemistry 59 (1), 46–49.
Sharma, P., Chakkaravarthi, A., Singh, V., Subramanian, R., 2008. Grinding
characteristics and batter quality of rice in different wet grinding systems.
Journal of Food Engineering 88 (1), 499–506.
Solanki, S.N., Subramanian, R., Singh, V., Ali, S.Z., Manohar, B., 2005. Scope of colloid
mill for industrial wet grinding for batter preparation of some Indian snack
foods. Journal of Food Engineering 69 (1), 23–30.
Suksomboon, A., Naivikul, O., 2006. Effect of dry-and wet-milling processes on
chemical, physicochemical properties and starch molecular structures of rice
starches. Kasetsart Journal 40, 125–134.
Yeh, A.-I., 2004. Preparation and applications of rice flour. In: Champagne, E.T. (Ed.),
Rice: Chemistry and Technology, 3rd ed. The American Association of Cereal
Chemists, Inc., pp. 495–539.