Trans. Mat. Res. Soc.

Japan 37[4] 557-562 (2012)

Effects of Activation with CO 2 /KOH on the Pore Structures

of Coffee Grounds-derived Carbon

Keisuke Kikuchi*, Rie Yamashita* and Satoshi Sakuragawa*

*Industrial Research Institute of Shizuoka Prefecture, 2078 Makigaya, Aoi-ku, Shizuoka 421-1298, Japan
Fax: 81-54-278-3066, e-mail: [email protected]

In this study, coffee grounds were carbonized through activation with carbon dioxide
(hereinafter referred to as CO 2 ) or potassium hydroxide (hereinafter referred to as
KOH), and the effects of the respective activation conditions on the pore structures of
the carbon obtained were investigated. The specific surface area of the sample
carbonized in a CO 2 stream at 850˚C was 752 m 2 /g and that of the sample impregnated
with KOH and carbonized was 2549 m 2 /g when the concentration of KOH was 4.5 M.
In the sample carbonized in a CO 2 stream, the pores formed were 0.8, 1, and 3.1 nm in
size, which were enlarged with the increase in treatment temperature and started to
include numerous mesopores. In the sample impregnated with KOH and carbonized,
the pores formed were 0.7 nm and 1 to 1.1 nm in size, which were slightly enlarged
with the increase in concentration of KOH. It was confirmed that for coffee grounds,
activation with KOH had an effect of increasing specific surface area and that CO 2
activation had an effect of increasing pore size.
Key words: Coffee grounds, Carbon, Activation, KOH, CO2

1. INTRODUCTION micropores and of drastically increasing specific surface

Coffee is a popular beverage all over the world. area[4-6]. It has also been reported that combined use of
Millions of tons of coffee beans are consumed every physical and chemical activation can contribute to
year. This naturally leads to the production of a large further increases in specific surface area and to
amount of used coffee grounds. Though some of them controlling pore size[7].
are effectively used as agricultural resources such as While many other reports on activation treatment
feed and fertilizers, most coffee grounds are disposed of methods have been published, there are only a limited
by burning[1]. From the perspective of environmental number of reports on the activation of coffee grounds.
protection, however, it would be desirable to reuse them. Although KOH used for activation is strongly alkaline,
Carbonization is a method of reusing such coffee in most of published papers, the amount of KOH used
grounds as useful resources. Carbon properties are was 4 times as much as that of the material to be
measured mainly with the indexes of specific surface carbonized[7]. Using too much KOH can cause a number
area and pore size. Carbon with a high specific surface of problems such as increasing producing costs and
area (i.e., activated carbon) is used as material for corroding equipment. It is desirable to not perform
adsorption, catalysts, or electrodes. In general, the activation with an unnecessarily large amount of KOH.
higher the specific surface area, the higher the In this study, coffee grounds were carbonized
performance of activated carbon. It is likely to be a in several different ways in order to find suitable
useful material with a high adsorption rate and methods of obtaining a highly specific surface
efficiency if its pore size can be controlled depending on area and to control pore size. Firstly, coffee
the type of adsorbate. grounds were carbonized without employing
Specific surface area can be increased through activation treatment to examine their
activation, which is roughly classified into physical carbonization behavior. Next, carbonization was
activation and chemical activation. Physical activation performed through activation with CO 2 and with
includes CO2 activation and steam activation. Such KOH. The specific surface area of the carbon
activation has been reported to have effects of producing obtained was measured and the pore structures
micropores and then enlarging them to mesopores or were analyzed in order to clarify optimal
macropores. For example, Gonzalez J. F. et al. reported conditions for the respective activation treatments
that mesopores produced were increased in number and the effects of activation conditions on the
through activation as used tires were activated with formation of pore structures.
CO2[2]. Roman S. et al. reported that micropores were
increased in size by activation as olive stones were 2. EXPERIMENTAL
activated with CO2[3]. Chemical activation includes 2-1. Materials
activation with KOH and with zinc chloride (ZnCl2). Of Coffee grounds discarded by a beverage processing
these, activation with KOH is attracting particular plant were used as the carbon raw material. The coffee
attention because of its effects of producing numerous grounds were fine granules with a water content of 60%

557
558 Effects of Activation with CO2/KOH on the Pore Structures of Coffee Grounds-derived Carbon

Table 1 Impregnation rate of KOH to coffee grounds calculated from result of ICP -OES.

Treatment Potassium content (wt%) Impregnation ratio of KOH

Unteated 0.369 (originally included) ‐
Immersed in KOH 0.5 M 7.55 0.12
Immersed in KOH 2 M 17.4 0.33
Immersed in KOH 4.5 M 24.9 0.56

KOH-impregnated
coffee grounds

Carbonization

1 cm

1 cm

Fig.1 Appearance of KOH-activated carbon (4.5 M)

to 70%. They were dried at 105˚C for not less than 12 desiccator. The carbonization temperatures were
hours and sieved to particle sizes of 0.5 to 2 mm. Yields, specified as 750˚C to 900˚C in increments of 50˚C, and
which will be given later, were calculated by dividing the rate of increase in the furnace temperature was set to
the mass of the carbon obtained by the mass of the 10˚C/minute.
sieved coffee grounds.
2-2-3. Carbonization of KOH-impregnated sample
2-2. Carbonization Approximately 5 g of the coffee grounds prepared as
2-2-1. Control carbonization treatment described in 2-1 were placed in a beaker containing 100
Approximately 5 g of the coffee grounds prepared as mL of KOH aqueous solution, and were agitated for 30
described in 2-1 were placed in a crucible, which was minutes. The concentrations of the KOH aqueous
placed at the center of an electric furnace (SBA-2035; solution were specified as 0.5 to 4.5 M in increments of
MOTOYAMA Co., Ltd.). All the air inside the furnace 0.5 M. After impregnation, suction-filtration was
was evacuated and replaced with nitrogen. performed to remove excess solution from each sample,
Carbonization was then performed in a nitrogen stream which was then dried at 105˚C for not less than 12
of 2 L/minute. The furnace temperature was increased at hours. After drying, the potassium content in each
the rate of 10˚C/minute to the carbonization sample was measured using an inductively-coupled
temperatures, specified as 500˚C to 1100˚C, in plasma optical emission spectrometer (ICP-OES),
increments of 100˚C. Each of the specified carbonization Optima 3300DV (PerkinElmer Japan Co., Ltd.), and the
temperatures were maintained for 1 hour. After cooling mass of KOH was calculated based on the potassium
in the nitrogen stream, the carbonized samples were content measured. KOH impregnation rates were
stored in a desiccator. calculated by dividing the mass of KOH by the mass of
dried coffee grounds (see Table 1). Each sample was
2-2-2. Carbonization in a CO2 stream placed in a crucible, which was placed at the center of an
Approximately 5 g of the coffee grounds prepared as electric furnace. All the air inside the furnace was
described in 2-1 were placed in a crucible, which was evacuated and replaced with nitrogen. Carbonization
placed at the center of an electric furnace. All the air was then performed in a nitrogen stream of 2 L/minute.
inside the furnace was evacuated and replaced with The furnace temperature was increased at a rate of
nitrogen. The furnace temperature was then increased in 10˚C/minute until reaching the set carbonization
a nitrogen stream of 2 L/minute until it reached temperature, 800˚C, which was maintained for 1 hour.
pre-specified carbonization temperatures. After this, the After cooling in a nitrogen stream, each sample was
nitrogen stream was changed to a CO2 stream of 1.5 neutralized with 2 M hydrochloric acid and washed well
L/minute and the each of the set temperatures was with distilled water. After drying at 105˚C for not less
maintained for 1 hour. After cooling in a nitrogen than 12 hours, the sample was stored in a desiccator.
stream, the carbonized samples were stored in a
 K. Kikuchi et al. Trans. Mat. Res. Soc. Japan 37[4] 557-562 (2012) 559

2-3. Characterization
2-3-1. Specific surface area
(a) 900 35

specific surface area S (m2/g)

29.9 S
800 28.3 27.5 30
Nitrogen adsorption of a mixed gas of nitrogen 700 26.1 yield
(24%) and helium (76%) was performed at liquid 600
22.9 25

yield (wt%)
nitrogen temperature using a FLOWSORB II 2300 500 18.3 17.2 20
(Shimadzu Corporation) and specific surface areas 400 313 461 15
408
were calculated by the single point BET method. 300
175 10
200
2-3-3. Scanning electron microscope observation 100 5
6 10 2
Cross sections of the carbonized samples were 0 0
400 500 600 700 800 900 1000 1100 1200
examined with a scanning electron microscope
T (℃)
(hereinafter referred to as SEM), TM-1000 (Hitachi
High-Technologies Corporation). (b) 900 35

specific surface area S (m2/g)

800 752
30
26.1 S
700
2-3-3. Pore size distribution 600
23.0 yield 25

yield (wt%)
The isotherms of nitrogen adsorption for the 500 17.4 20
samples carbonized with the control treatment 400
494
15
method at 800˚C, 900˚C, and 1100˚C were 300
10
measured at liquid nitrogen temperature using a 200
176

BELSORP-max (BEL JAPAN Inc.) and pore 100 5

0.9
size distribution was obtained through NLDFT 0 0
analysis [8, 9]. The nitrogen adsorption isotherms 700 750 800 850 900 950
T (℃)
of the samples carbonized in a CO 2 stream at
800˚C and 850˚C, and the samples impregnated (c) 3,000 35
specific surface area S (m2/g)

2460 2528 2549

with KOH (0.5, 2, and 4.5 M) and carbonized, 2,500 26.1 2263 2265
30
were measured at liquid nitrogen temperature 1922 1995 25
2,000 S
using an AUTOSORB IQ (Quantachrome

yield (wt%)
1520 yield 20
Instruments) and pore size distribution was 1,500 14.7
13.5 12.6 12.3 12.9 15
obtained through QSDFT analysis [10, 11]. 1,000
11.5 11.9 10.6
10.0
1044 10
3. RESULTS AND DISCUSSIONS 500 5
175
The samples carbonized with the control treatment 0 0
method and the samples carbonized in a CO2 stream 0 1 2 3 4 5
were granular in appearance. The samples concentration of KOH (M)
impregnated with KOH and carbonized were different, Fig.2 BET specific surface area and yield.
however, with the sample impregnated with 0.5 M KOH (a)Control carbonization treatment
remaining granular, while the samples impregnated with (b)Carbonization in a CO 2 stream
1 to 4.5 M KOH changed from a granular to a foam-like (c)Carbonization of KOH-impregnated sample
shape (a shape like leavened bread) (Fig.1).
The yields and specific surface areas of the sample temperature and reached 752 m2/g at 850˚C. In general,
carbonized with the control treatment method (a), the CO2 activation is believed to progress at temperatures
sample carbonized in a CO2 stream (b), and the sample above 850˚C in accordance with the following reaction
impregnated with KOH and carbonized (c) are shown in formula (1) [13].
Fig.2. C+CO2→2CO (1)
The yield of the sample carbonized with the control In this experiment, the fact that the specific surface area
treatment method decreased with the increase in was considerably increased at 850˚C suggested that the
carbonization temperature. The specific surface area of reaction progressed smoothly. It was considered that
the sample rapidly increased with temperature for 900˚C was an excessive temperature and the reaction
carbonization temperatures above 700˚C. It kept progressed too rapidly at 900˚C to be controlled.
increasing monotonically with the increase in The yield of the sample impregnated with KOH and
temperature and reached 461 m2/g at 1100˚C. The rapid carbonized decreased with increasing KOH solution
increase of the specific surface area at 700˚C may have concentration. The specific surface area initially rapidly
been associated with some action of a certain factor increased with the increase in KOH solution
inducing activation. Residual autolysis gas or air concentration and increased more slowly roughly after
released from packing material used for the furnace may the concentration exceeded 2 M. It reached 2549 m2/g at
have acted as an activation gas[12]. a concentration of 4.5 M. The mechanism of reaction of
The yield of the sample carbonized in a CO2 stream activation with KOH has been explained in several
decreased with the increase in carbonization different ways by various researchers[14-18]. Using
temperature. It rapidly decreased with increasing common factors presented in all their articles, Sakazaki
temperature for temperatures exceeding 850˚C, and the et al. proposed the following reaction formulas[7]. Firstly,
sample was incinerated at 900˚C. The same action was the following reaction (2) progresses at 630˚C and
observed when the retention time was changed to 30 K2CO3 is generated.
minutes. The specific surface area increased 6KOH+2C→2K+3H2+2K2CO3 (2)
monotonically with the increase in carbonization
560 Effects of Activation with CO2/KOH on the Pore Structures of Coffee Grounds-derived Carbon

The K2CO3 generated reacts with carbon in accordance Low magnification High magnification
with the following reaction formulas, (3) and (4), and
generates K2O. Based on reaction formula (5), K2O is
generated also from K.
K2CO3+C→K2O+2CO (3)
K2CO3→K2O+CO2 (4) (a)
2K+CO2→K2O+CO (5)
The K2O generated ultimately reacts with carbon at
about 800˚C in accordance with reaction formula (6).
K2O+C→2K+CO (6)
In this experiment, in which the carbonization
temperature was specified as 800˚C, reaction by
activation with KOH was thought to have progressed to
(b)
the final stage, described by reaction formula (6).
Fig.3 shows the SEM images of carbonized samples.
The sample carbonized with the control treatment
method at 1100˚C and the sample carbonized in a CO2
stream at 850˚C maintained the tissue structure
consisting of coffee bean parenchyma cells [Fig.3, (c)
(a)(b), left]. Although the sample impregnated with 0.5
M KOH and carbonized also maintained the tissue
structure of coffee beans [Fig.3, (c), left], its cell walls
were thinner than those of the sample carbonized with
the control treatment method or the sample in a CO2
stream [Fig.3, (a)(b)(c), right]. The carbon content
constituting the cell walls was likely to be consumed in (d)
the course of activation with KOH. In the sample
impregnated with 4.5 M KOH and carbonized, the tissue 40 μm 2 μm
structure of coffee beans was broken [Fig.3, (d), left]
and numerous fine pores with a diameter of 1 µm or Fig.3 SEM micrograhs for prepared coffee-derived
smaller were observed [Fig.3, (d), right], indicating carbon.
considerable progress of activation with KOH.
(a)Control carbonization treatment (1100 ˚C)
The isotherms of nitrogen adsorption were measured
(b)Carbonization in a CO 2 stream (850˚C)
in terms of the sample carbonized with the control
(c)Carbonization of KOH-impregnated (0.5M)
treatment method (a), the sample carbonized in a CO2
(d)Carbonization of KOH-impregnated (4.5M)
stream (b), and the sample impregnated with KOH and
carbonized (c). The measurement results are shown in
nm was produced at 850˚C in the sample carbonized in a
Fig.4 above. The results demonstrated that all the carbon
CO2 stream than was produced at 900˚C in the sample
displayed Type I isotherms and that they contained
carbonized with the control treatment method. While
numerous micropores.
small peaks were also observed at 900˚C for pore sizes
Fig.4 below shows the results of the NLDFT analysis
of 2.7 nm and 4.7 nm in the sample carbonized with the
of the sample carbonized with the control treatment
control treatment method, they occurred at a much lower
method (a) and those of the QSDFT analysis of the
frequency. Even after taking into account differences in
sample carbonized in a CO2 stream (b), and the sample
measurement devices and analysis methods, the results
impregnated with KOH and carbonized (c). The abscissa
suggested that CO2 activation promoted the development
has a logarithmic scale.
of mesopores. In agreement with certain existing
In the sample carbonized with the control treatment
literature[2],[3]), the development of mesopores with the
method, a marked peak was observed at 800˚C for the
progress of CO2 activation was also confirmed for coffee
pore size of 0.6 to 0.7 nm. The peak became even
grounds.
greater at 900˚C while showing other small peaks for the
In the sample impregnated with KOH and carbonized,
pore sizes of 2.7 nm and 4.7 nm. The peak for the pore
peaks were observed in the micropore area around the
size of 0.6 to 0.7 nm declined at 1100˚C to the similar
sizes of 0.7 nm and 1 to 1.1 nm. The heights of these
level of that at 800˚C, while the peaks for the pore sizes
peaks were increased when the concentration of KOH
of 1.6 nm and 4.7 nm increased. These results
was between 0.5 M and 2 M. When the concentration of
demonstrated that micropores with sizes of about 0.6 to
KOH was between 2 M and 4.5 M, the number of pores
0.7 nm were mainly produced at temperatures between
with a size of about 0.7 nm slightly decreased and pores
800˚C and 900˚C and that the development of
developed with a size of about 1 to 1.1 nm. Peaks were
micropores at a size of 1.6 nm and mesopores at a size
slightly shifted in the direction of larger-sized pores,
of 4.7 nm was promoted at 1100˚C.
demonstrating that the size of micropores became larger
In the sample carbonized in a CO2 stream, peaks were
with increasing KOH concentration. Using beer lees
observed in the micropore area around 0.8 nm and 1 nm
impregnated with KOH and carbonized, Hayashi et al.
and in the mesopore area around 3.1 nm. All of these
reported that the higher the carbonization temperature
peaks were higher and wider at 850˚C than those at
(in particular, at 800˚C or higher) and the KOH
800˚C. A greater amount of mesopores with a size of 3.1
concentration (impregnation rate ≥ 2, impregnation rate
 K. Kikuchi et al. Trans. Mat. Res. Soc. Japan 37[4] 557-562 (2012) 561

(a) 500
(b) 500 850℃
(c) 500
1100℃
900℃ 800℃
4.5 M
400 800℃ 400 400
2M

Va (cm3(STP)/g)

Va (cm3(STP)/g)
Va (cm3(STP)/g)

0.5 M
300 300 300

200 200 200

100 100 100

0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
p/p0 p/p0 p/p0

2.0 2.0 2.0

1.8 1100℃ 1.8 850℃ 1.8 4.5 M
1.6 900℃
dVp/dW (cm3/nm)
1.6 800℃ 1.6 2M
0.5 M

dVp/dW (cm3/nm)
1.4 800℃ 1.4 1.4
dVp/dW (cm3/nm)

1.2 1.2 1.2

1.0 1.0 1.0
0.8 0.8 0.8
0.6 0.6 0.6
0.4 0.4 0.4
0.2 0.2 0.2
0.0 0.0 0.0
0.4 0.8 1.6 3.2 6.4 12.8 25.6 0.4 0.8 1.6 3.2 6.4 12.8 25.6 0.4 0.8 1.6 3.2 6.4 12.8 25.6
W (nm) W (nm) W (nm)

Fig.4 N 2 adsorption isotherm and pore size distribution of coffee-derived carbon.

(a)Control carbonization treatment
(b)Carbonization in a CO 2 stream
(c)Carbonization of KOH-impregnated sample

= mass of KOH/mass of dried material), the larger the 4. CONCLUSION

amount of mesopores produced[19]. This is considered to In order to clarify their carbonization behaviors,
be deeply associated with CO2 generated in the process coffee grounds were carbonized at several different
of activation with KOH (4) mentioned earlier. It is carbonization temperatures and under various activation
assumed that the higher the carbonization temperature, conditions and the properties of the carbon obtained
the higher the effect of CO2 activation and that the were evaluated. In the sample carbonized with the
higher the content of KOH, the larger the amount of CO2 control treatment method, the higher the carbonization
generated, resulting in an increase in the effect of CO2 temperature, the lower the yields and the higher the
activation. Mesopore formation was not observed under specific surface area became, being 17.2 wt% and 461
the experiment conditions in this study. It will be m2/g, respectively, at 1100˚C. In the sample carbonized
necessary to investigate activation with KOH at 850˚C, in a CO2 stream, the higher the carbonization
the temperature at which the highest degree of the CO2 temperature, the lower the yield and the higher the
activation effect can be observed. specific surface area became, being 17.4 wt% and 752
Based on the results given above, it is m2/g, respectively, at 850˚C. In the sample which was
recommended to use the effect of CO 2 activation impregnated with KOH and carbonized, the higher the
to enlarge pore size, increase carbonization KOH concentration, the lower the yield and the higher
temperature, and raise KOH impregnation rate in the specific surface area became, being 10 wt% and
order to increase micropores to mesopores, and to 2549 m2/g, respectively, when the concentration was 4.5
control pores. Considering the possibility of M. However, since the rate of increase in specific
increasing the degree of alkalinity by increasing surface area became slower at KOH concentrations of 2
KOH impregnation rate, it may be desirable to M or higher, it is recommended to perform activation
combine CO 2 activation and temperature control with KOH at a concentration of not higher than 2 M in
in a balanced way. However, it has been reported consideration of workability and cost. Regarding pore
that simultaneous activation with CO 2 and KOH structures, the number of mesopores was increased in
leads to the CO 2 consuming the alkali and to the sample carbonized in a CO2 stream with the progress
decreasing the effect of KOH activation [20]. By of reaction. Micropores in the sample impregnated with
performing micropore formation through KOH and carbonized were slightly increased in size with
activation with KOH at a low concentration and the progress of reaction.
mesopore formation through CO 2 activation with By combining the effect of activation with
a multistage operation method, functional carbo n KOH to increase specific surface area and the
with high specific surface area and numerous effect of CO 2 activation to increase pore size, it
mesopores may be obtained. will be aimed to develop coffee grounds-derived
carbon to be used as material for water
562 Effects of Activation with CO2/KOH on the Pore Structures of Coffee Grounds-derived Carbon

purification or electrodes, which require a high 26, 293-297(2000)(in Japanese)

specific surface area and optimal mesopore [20] S. Kobayashi, H. Yuan, K. Kuragane, K.
distribution. Gomi, Y. Takiguchi, K. Onoe and T. Ymaguchi,
Nippon kagakukaishi, 5, 329-333(2000) (in
5. ACKNOWLEDGEMENTS Japanese)
Sincere gratitude is expressed to Dr. Y. Saito
(The University of Tokyo) for her advice on
sample preparation. (Received June 7, 2012; Accepted July 16, 2012)

6. REFERENCES
[1] N. Kawasaki, T. Tokimoto, T. Nakamura, K.
Satoh, M. Kabayama and S. Tanaba, Journal of
The Surface Science Society of Japan , 25,
382-386(2004)(in Japanese)
[2] J. F. Gonzalez, J. M. Encinar, C. M.
Gonzalez-Garcia, E. Sabio, A. Ramiro, J. L.
Canito, J. Ganan, Appl. Surf. Sci., 252,
5999-6004(2006)
[3] S. Roman, J. F. Gonzalez, C. M. Gonzalez-Garcia, F.
Zamora, Fuel Process. Technol., 89, 715-720(2008)
[4] M. A. Lillo-Rodenas, D. Cazorla-Amoros, A.
Linares-Solano, Carbon, 41, 267-275(2003)
[5] D. Lozano-Castello, M. A. Lillo-Rodenas, D.
Cazorla-Amoros, A. Linares-Solano, Carbon, 39,
741-749(2001)
[6] E. Raymundo-Pinero, P. Azais, T.
Cacciaguerra, D. Cazorla-Amoros, A.
Linares-Solano, F. Beguin, Carbon, 43,
786-795(2005)
[7] T. Sakazaki, K. Yoshimura and M. Machida,
Mokuzai Gakkaishi, 56, 251-257(2010)(in
Japanese)
[8] C. Latoskie, K. Gubbins and N. Quirke, J.
Phys. Chem., 97, 4786(1993)
[9] C. Latoskie, K. Gubbins and N. Quirke, Langmuir, 9,
2693(1993)
[10] G. Y. Gor, M. Thommes, K.A. Cychosz, and
A.V. Neimark, Carbon, 50, 1583-1590(2012)
[11] A.V. Neimark, Y. Lin, P.I. Ravikovitch, and
M. Thommes, Carbon, 47, 1617-1628(2009)
[12] R. Yamashita, Y. Saito and S. Sakuragawa, Journal
of Wood Science, 55, 446-452(2009)
[13] Y. Sanada, M. Suzuki and K. Fujimoto,
Shinpan Kasseitan, Kodansha Co., Ltd., 2-12-21
Otowa, Bunkyo-ku Tokyo, 1992, 48-49 (in
Japanese)
[14] M. A. Lillo-Rodenas, D. Cazorla-Amoros, A.
Linares-Solano, Carbon, 41, 267-275(2003) [3]
[15] E. Raymundo-Pinero, P. Azais, T.
Cacciaguerra, D. Cazorla-Amoros, A.
Linares-Solano, F. Beguin, Carbon, 43,
786-795(2005)
[16] E. N. El Qada, S. J. Allen, G. M. Walker,
Chem. Eng. J., 142, 1-13(2008)
[17] R. L. Tseng, S. K. Tseng, F. C. Wu, C. C. Hu,
C. C Wang, J. Chin. Inst. Chem. Eng., 39,
37-47(2008)
[18] J. Alcainz-Monge, M. J. Illan-Gomez, J.
Colloid Interface Sci., 318, 35-41(2008)
[19] J. Hayashi, A. Kubo, A. Furukawa and K.
Muroyama, KAGAKU KOGAKU RONBUNSHU,