Res Chem Intermed

DOI 10.1007/s11164-016-2712-7

Investigation on effects of operating conditions

in combined activation of char to develop
the pore structure

Soudeh Banivaheb1 • Hassan Hashemipour1

Received: 6 April 2016 / Accepted: 24 August 2016

 Springer Science+Business Media Dordrecht 2016

Abstract In this study, activation of pistachio shell char was investigated by

chemical, physical and combined activation experimentally using a TGA instrument
and lab scale reactor. The effects of operating conditions such as activation time,
temperature and impregnation ratio in the activation methods were studied via the
weight loss and iodine number of the samples. The results showed that combined
activation is the premium activation method. In addition, in this method. a tem-
perature of 800 C, a dwell time of 45 min and an impregnation ratio of 50 % are
the optimum conditions which give a product with an iodine number of 2433 mg/g
and a weight loss 52 %. A high iodine number of this activated carbon was
developed at a moderate temperature and dwell time, and the weight loss confirms
the advantages of the combined activation method. The products were characterized
by SEM and FTIR analyses.

Keywords Activated carbon  Combined activation  Pistachio shell  Iodine

number

Introduction

Activated carbon, due its wide variety of applications, has been used in many
industries such as oil, gas, petrochemical, food, medicine and also used in
environmental application. The reason is the ability to develop a high surface area
and a controllable pore structure in carbon in addition to other properties such as
carbon’s mechanical, thermal and chemical stability. It can be produced from
various carbonaceous material such as coal, synthetic materials and biomasses [1].
Agriculture wastes or lignocelluloses such as coconut shell [2], pistachio shell [3],

& Hassan Hashemipour

[email protected]
1
Chemical Engineering Department, Shahid Bahonar University of Kerman, Kerman, Iran

123
 S. Banivaheb, H. Hashemipour

almond shell [4], cotton stalk [5], walnut shell [6] and acorn shell [7, 8] are good
material to produce activated carbon because of a high concentration of fixed carbon
and low ash content [9, 10]. Generally, activated carbon can be produced via
chemical activation, physical activation and combined activation methods. In
physical activation, the precursor should be exposed to the oxidizing gases such as
CO2 and H2O at high temperatures [11]. In chemical and combined activation
methods, the precursor should be impregnated with the chemical agents which has a
main effect on the precursor reactivity and can influence the pore structure of
activated carbon. The most common chemical agents are ZnCl2, KOH and H3PO4
[12–14]. ZnCl2 acts as an electrolytic agent to swell the cellulose structure of the
biomass. This swelling causes initial porosity in the reactant char [15]. ZnCl2 can
also destroy the cellulosic structure of a biomass by stripping hydrogen and oxygen
atoms, and developing initial porosity [16]. On the other hand, the effectiveness of
KOH can be attributed to the ability to form intercalation compounds with carbon.
KOH is decomposed to K2O in the biomass and K2O easily diffuses into the pores
and is reduced to potassium using carbon to form pores [18].
Chemical and physical activation methods have been used by many researchers
by investigating the effects of several parameters on the final pore texture of
activated carbon [10, 17]. The combined activation (or catalytic physical activation)
method, however, is a newer method which has been less investigated than the other
methods. Hu [6] studied the activation of walnut shell impregnated with ZnCl2 in a
CO2 atmosphere. The study focused on the effect of impregnation ratio on the
activated carbon surface area. Laine et al. [2] investigated the effect the type of
potassium compound had on the reactivity of impregnated coconut shell char with
CO2 gas. Budinova [19] studied chemical and combined activation at two
temperatures (600 and 700 C) and the results showed that the combined method
is more effective than chemical activation at developing pore and surface area.
In this study, preparation of activated carbon from pistachio shell char via a
combined activation method is investigated in a tubular reactor and is compared
with other methods. The effects of the operating conditions in these methods on
pore development is studied and compared. The significance of operating conditions
such as chemical ratio, activation temperature, dwell time, chemical agents on
surface area and weight loss of the product are different and this signification is
studied in this work.

Materials and methods

Pistachio nut shells from Rafsanjan farms were charred at 550 C for 3 h under an
inert gas atmosphere and the char was used as the precursor. The ultimate analysis
of the char is given in Table 1. This analysis showed relatively high oxygen content
which facilitates the higher reactivity of the char. The BET surface area of this char
is around 470 m2/gr which shows initial porosity develops in the activation step. All
chemicals used in both the activation experiments and adsorption tests were
purchased from Merck & Co., Inc. and were laboratory grade ([99.99 %).

123
Investigation on effects of operating conditions in…

Table 1 The ultimate analysis of the pistachio shell char (ash free basis)
Elements C H N O S

Content (wt%) 47 7 0.4 45 0.6

In the combined activation method, the impregnation step was carried out with
zinc chloride solution. In this step, 20 gr of pistachio shell char with 150 ml of 1.5-
M salt solution (ZnCl2 and KOH) were mixed for 2 h. This gives an impregnation
ratio of 150 %. The impregnation ratios of 100 % and 50 % (100 and 50 ml of the
ZnCl2 solution, respectively) also were included in this study. The impregnated
chars were dried at 100 C for 24 h and then activated in a reactor under an
oxidizing atmosphere (CO2). In the chemical activation method, the carbonization
atmosphere of the impregnated sample is inert (N2) while in the physical activation
method, the carbonization atmosphere of the not impregnated sample is the
oxidizing gas.
The experimental setup of the activation process is shown in Fig. 1. The reactor
was a tubular furnace reactor equipped with a temperature-controlling system and a
continuous gas stream via an inlet and an outlet. The reaction zone was a stainless
steel 304 tube with an I.D. of 7 cm and a length of 60 cm. The inlet gas flow rate
was indicated with a gas flow meter. The gas flow rate (inert or oxidizer) was
adjusted to produce an excess amount of gas content in the reactor.
The samples were heated in the reactor from room temperature to the final
activation temperature at a heating rate of 30 C/min in an N2 atmosphere, and then
the samples were activated at the final temperature for 15, 45 or 75 min in the

Fig. 1 Schematic diagram of the experimental study

123
 S. Banivaheb, H. Hashemipour

desired atmosphere. After that, the activated products were cooled under N2 flow to
room temperature. The products were washed with distilled water at 100 C several
times until the pH number of the mixture reached above 6 and then the products
were dried at 100 C for 24 h.
Activated carbon was characterized by a weight-loss parameter and an iodine
number test. Weight loss is defined as the amount of burn-off of the char during the
activation step. This parameter is calculated by Eq. (1):
Weight loss ¼ ðw0  wÞ=w0 ð1Þ
w0, the initial weight of char; w, the weight of the end product.
The iodine number was measured to evaluate the surface area of the activated
carbon by titration based on the standard method (ASTM D 4607-94) [20]. In
addition, scanning electron microscopy (SEM, Philips XL30 model) and Fourier
transform infrared (FTIR Tensor 27) analyses were also used to characterized the
prepared activated carbons. The reactivity of the char in different conditions was
determined with a thermogravimetric analysis (TGA) instrument (model 409, PC
LUXX, NETZSCH Company). The methylene blue concentration in the methylene
blue number test was measured by an ultraviolet–visible light (UV–Vis)
spectrophotometer (UV-Vis Mecasys 3220).

Results and discussion

Thermogravimetric (TGA) analysis

The TGA results of chemical activation of the char with ZnCl2 and KOH are shown in
Fig. 2. As it is observed, weight loss starts to descend, in the case of ZnCl2, at
approximately 450 C, while for KOH, it starts about 570 C. Therefore, ZnCl2 is
more reactive than KOH in the activation reaction. Higher weight loss in ZnCl2
activation also confirms its greater reactivity. This difference can be attributed to the
roles of the reagents during the reaction. According to the definition of a Lewis acid
and Bronsted–Lowry base, KOH is a strong base, whereas ZnCl2 is a weak acid. The
acidic reagents are more able than the base reagents to interact with oxygen functional

Fig. 2 TGA results of the ZnCl2 KOH

120
ZnCl2 and KOH chemical
activation
WEIGHT LOSS%

100

80

60

40

20

0
0 200 400 600 800 1000
TEMPERATURE(C)

123
Investigation on effects of operating conditions in…

groups and—thus catalyze the dehydration and dehydrogenation reactions. Therefore,

lignocellulosic materials with high oxygen content are generally preferred in
chemical activation with ZnCl2 [12, 21].
The TGA analyses were carried out on non-impregnated samples under both inert
(N2) and oxidizing atmospheres (CO2; samples 1 and 3, respectively). The analysis
was repeated on the impregnated samples under the same inert (N2) and oxidizing
atmospheres (CO2: samples 2 and 4, respectively). The results are shown in Fig. 3.
The weight loss of all samples at temperatures less than 200 C is because of
water evaporation. This loss in samples 2 and 4 is higher than in samples 1 and 3
because of impregnation. The TGA results show similar weight loss in a
temperature range of 550–700 C for samples 2 and 4 while there is no weight
loss in this temperature range for samples 1 and 3. This weight loss indicates a
chemical effect on the activation progress. In addition, comparison of the TGA
behavior of samples 3 and 4 in this temperature region indicates that the gas
atmosphere has no important effect on the activation progress. The weight loss of
samples 3 and 4 at temperatures ranging from 700–900 C indicates the effect of the
char reaction with CO2 as the oxidizing gas. The TGA data of these samples have
the same slope and this suggests a negligible effect of the chemical (ZnCl2) in this
temperature range. This weight loss is not observed for samples 1 and 2. These
analyses suggest that a temperature range of 500–700 C for the chemical effect and
a range of 700–900 C for the oxidizing gas effect are appropriate.

Experimental results

Several experiments have been carried out in the laboratory reactor setup to study
the effect of operating conditions in the three activation methods. The results are
summarized and categorized according to the activation methods.

Fig. 3 TGA results of samples 1–4

123
 S. Banivaheb, H. Hashemipour

Chemical activation

The chemical activation experiments were conducted at temperature range

500–700 C, dwell time 15–75 min and impregnation ratios 50 and 100 %.The
weight loss of sample during activation and the Iodine number of the product are
reported in the Table 2.
As is shown in Table 2, an increase of temperature, dwell time and impregnation
ratio causes an increase of weight loss of the samples. In addition, the data show that
increasing of the activation temperature and time have undesirable effects on the
improvement of the iodine number of synthesized activated carbon in the higher
range. Indeed, a weight loss increase has no relevance with surface area
improvement. In other word, progress of the reaction at higher activation time
and temperature causes pore overlapping and this decreases the iodine number of
the product [22].
Comparison of impregnation ratios in the activation show that a higher
impregnation ratio is caused by reducing the iodine number and this can be related
to destruction of the pores at a higher chemical content [22]. Reduction of the iodine
number and, therefore, surface area at a high impregnation ratio is confirmed in the
literature [3].

Physical activation

As mentioned in the discussion of the TGA results, the reaction of char with CO2
progresses at temperatures over 700 C. The physical activation of the chars with
CO2 is studied in the reactor setup at different reaction temperatures and times, and
the results are presented in Table 3. The effects of operating conditions (activation
time and temperature) on the weight loss and iodine number of the activated carbon
in the physical activation are the same as the chemical activation. This behavior is
observed in the other studies [11].

Table 2 Results of chemical activation

T (C) Impregnation ratio (100 %) Impregnation ratio (50 %)

Time (min) Weight loss Iodine number Weight loss Iodine number

500 15 53.125 918.97 31.25 1053.01

500 45 58.125 974.82 37.37 1231.72
500 75 59.375 1058.59 39.62 1120.03
600 15 53.375 1002.74 38.12 1231.72
600 45 57.125 1041.84 42.5 1427.20
600 75 58.375 972.58 56.25 1287.57
700 15 56.875 1080.93 50.62 1388.10
700 45 59.875 1382.52 52.37 1633.84
700 75 62.5 1265.23 61.87 1611.50

123
Investigation on effects of operating conditions in…

Table 3 Results of physical

T (C) Time (min) Iodine number Weight loss %
activation
700 15 880.1 15
700 45 1050.4 20
700 75 1080.3 24.87
800 15 1480.1 22.12
800 45 1724.3 37.25
800 75 1742.1 48.25
900 15 1801.4 37.37
900 45 2002 49.7
900 75 1946.4 54

The optimum conditions of physical activation, in which a maximum iodine

number is achieved at a minimum weight loss, are a temperature of 900 C and a
dwell time of 45 min. In this condition, the activated carbon with an iodine number
of 2002 mg/g is synthesized with a weight loss of 49 %.

Combined activation

In this method, the samples impregnated with ZnCl2 were conducted to the
oxidizing gas. The range of operating conditions is selected according to TGA
description, with temperatures of 500, 600, 700 and 800 C, dwell times of 15, 45,
75 and 105 min and impregnation ratios of 50, 100 and 150 %. Investigation of the
effect of impregnation ratio on weight loss and iodine number is shown in Figs. 4
and 5.
Generally, both chemical and physical reactions of the activation step progress
gradually with time, and, therefore, variation of weight loss with the dwell time has
a positive slope (Figs. 4, 6, 8). Since variation of the iodine number with the dwell
time depends on the range of reaction conditions, the curve of the iodine number

Fig. 4 Weight loss percentage 100% 150% 50%

of sample with impregnation 80
percentage at 700 C

60
Weight Loss (%)

40

20

0
0 20 40 60 80
Time (min)

123
 S. Banivaheb, H. Hashemipour

Fig. 5 Iodine number with 150% 100% 50%

different impregnation 2500
percentage at 700 C

Iodine Number (mg/gr)

2000

1500

1000

500

0
0 20 40 60 80
Time (min)

Fig. 6 Weight loss percentage 800 700 600 500

of impregnated sample with
90
ZnCl2 100 %
80
70
Weight loss %

60
50
40
30
20
10
0
0 20 40 60 80 100
time (min)

versus time should be studied at different reaction temperatures and impregnation

ratios.
The weight loss and iodine number of the samples are compared at three levels of
impregnation ratio. As it is observed, weight loss is increased with increasing the
impregnation ratio from 50 to 150 %, while the iodine number is increased with
changing the impregnation ratio from 50 to 100 % and then is decreased with
changing the impregnation ratio from 100 to 150 %. This is related to destruction of
the pores at a high chemical content of the samples. Therefore, an impregnation
ratio of 150 % is deleted from the following experiments. Comparison of weight
loss and iodine number at different operating conditions of temperature, time and
impregnation ratio are shown in the Figs. 6, 7, 8 and 9.
As is observed via the TGA results (Fig. 3), in the lower temperature range (500
and 600 C), the chemical step is more dominant than the physical step. Therefore,
the variation of weight loss and iodine number in Figs. 6 and 7 is observed during a
similar chemical activation process (Table 3). While at a higher temperature range

123
Investigation on effects of operating conditions in…

Fig. 7 Iodine number of sample 600 700 800 500

impregnated with ZnCl2 at
3000
100 %
2500

iodin number (m2/gr)

2000

1500

1000

500

0
0 50 100
time (min)

Fig. 8 Weight loss percentage 800 700 500 600

of sample impregnated with 60
ZnCl2 at 50 %

50
Weight Loss (%)

40

30

20
0 20 40 60 80
Time (min)

Fig. 9 Iodine number of sample 800 700 600 500

impregnated with ZnCl2 at 50 % 3000

2500
Iodin Number (mg/gr)

2000

1500

1000

500

0
0 20 40 60 80
Time (min)

123
 S. Banivaheb, H. Hashemipour

(700 and 800 C), the CO2 reaction with the char is added to the mechanisms of
pore development; therefore, the maximum iodine number is achieved which is
about 2460 mg/g. This product is attained at a temperature of 800 C and an
impregnation ratio of 100 % for 75 min with a weight loss of 75 %.
As Figs. 8 and 9 show, an increase of temperature causes an increase in both
chemical and physical activation rates, which is the reason for weight loss at higher
temperatures. It is noticable that the pore development mechanism is related to the
reaction rates (and, consequently, to temperature) and, therefore, at lower
temperatures (500 and 600 C), the iodine number follows an ascending behavior,
while at higher temperatures (700–800 C), the iodine number shows an ascending–
descending behavior. The descending portion of this behavior is because of pore
overlapping mechanism which occurs at a higher temperature and shorter time.
According to these results, the maximum iodine number is 2433 mg/g with a weight
loss of 56 % observed at a temperature of 800 C and an impregnation ratio of 50 %
for 45 min.
The comparison of two optimum conditions (impregnation ratios of 100 and
50 %) shows that the synthesis at the optimum condition of a 50 % impregnation
ratio is better because of lower weigh loss and shorter duration time.
As is mentioned, the combined activation method is a new procedure to activate
char, and the char type has a main effect in finding the optimum condition of the
high-surface area product. Comparison of the optimum condition of the high-surface
area product for several biomass precursors is presented in Table 4. The table shows
that the pistachio shell char is better than the others because of higher reactivity
(lower activation time) and higher surface area.

Methylene blue analysis

Methylene blue and iodine number analysis are usually used to characterize the
adsorption performance of activated carbon. The molecular size of iodine is smaller
than methylene blue; therefore, the surface area covered by methylene blue is lower
than iodine which indicates super-micropores and mesopores. As a result,
comparison of these two criteria can help estimation of the pore size range of the
activated carbon [9]. To detect this range, the methylene blue numbers were
obtained for the three samples A, B and C at an initial concentration of 200 ppm of

Table 4 Comparison of combined activation of different raw materials

Raw material Atmosphere Temperature Time Chemical Surface area References
(C) (min) agent (m2/gr)

Coconut shell CO2 800 60 K3PO4 1550 [2]

Walnut shell CO2 800 120 ZnCl2 2450 [6]
Woody biomass H2O 700 120 H3PO4 1360 [19]
birch
Pistachio shell CO2 800 45 ZnCl2 2486 This work

123
Investigation on effects of operating conditions in…

Table 5 Methylene blue

Sample MBN (mg/g) IN (mg/g) (MBN/IN) 9
analysis of activated carbon
100 (%)
samples
A 78 941.7 8.3
B 110 1544.7 7.2
C 223 2433 9.1

methylene blue. These samples are selected based on different iodine numbers that
are activated at the same time and at different temperatures of 500, 700 and 800 C,
respectively. The results are presented in Table 5. Comparison of the methylene
blue number (MBN) and iodine number (IN) shows that more than approximately
90 % of the adsorption was carried out on the micropores.

Sample characterization

SEM analysis

To study the morphology of the synthesized activated carbon, pistachio shell char
and the activated carbon synthesized at an optimum condition of impregnation ratio
of 50 % are analyzed by scanning electron microscopy (SEM). The images are
shown in Figs. 10 and 11. Comparison of these images shows that the activation
process creates a porous structure in the sample in the macro size. It should be
mentioned that macropores are divided to the smaller size in depth and, therefore,
the mesopores and micropores were not observed in the SEM images.

Surface chemistry

The FTIR spectrum of the char pistachio shell is shown in Fig. 12. The band at
about 3423 cm-1 is attributed to m (O–H) vibrations in the hydroxyl groups. The

Fig. 10 SEM image of the as-synthesised pistachio shell char with magnification 9100 (a) and 9500 (b)

123
 S. Banivaheb, H. Hashemipour

Fig. 11 SEM image of the as-synthesised activated carbon with magnification 92000 (a) and 920,000
(b)

Fig. 12 FTIR spectrum of the pistachio shell char

location of hydrogen-bonded OH groups is usually in the range of 3200–3650 cm-1

for alcohols, phenols and carboxylic acids. The bond located around 2923 cm-1
corresponds to m (C–H) vibrations in the methyl and methylene groups. This band is
contrary to the d (C–H) vibrational bands for –CH3– and –CH2=, which are located
around 1380 and 1465 cm-1. The three characteristic bands at 873, 813 and 749 are
assigned to aromatic (C–H), and the band around 1694 cm-1 is attributed to m
(C=O) vibrations in the carbonyl group. Figures 13, 14, 15 show the effect of
activation temperature on the FTIR spectra of chemically, physically and combined
activated carbons which were washed. In all three methods, after the activated
carbon was thoroughly washed with HCl and distilled water, an intense band

123
Investigation on effects of operating conditions in…

Fig. 13 FTIR spectrum of an activated carbon prepared via chemical activation

Fig. 14 FTIR spectrum of an activated carbon prepared via physical activation

appeared in the region 1549–1565 cm-1, which could be attributed to m (C=C)

vibration in the aromatic rings. In th physical method, a little bond at about
1692 cm-1 corresponds to a carbonyl group, which shows the hydrolysis was not
complete. The disappearances of the carbonyl and carboxylate ion groups in the

123
 S. Banivaheb, H. Hashemipour

Fig. 15 FTIR spectrum of an activated carbon prepared via combined activation

chemical and combined methods were due to the hydrolysis effect of the HCl,
resulting in the decomposition of these groups and the subsequent release of their
by-products as volatile matter. Absorptions due to c (C–H) bending occur at 882,
821, 765, and 753 cm-1, while c (O–H) bending is attributed to the absorptions at
618 and 552 cm-1. In the chemical method, the stretching vibration bond of
aliphatic C–H appears at about 2922 cm-1. Disappearance of this bond in the
chemical method shows that the aliphaticity in the combined method was decreased
[23–25].

Conclusion

The experimental data of pistachio shell char activation show that this precursor is a
suitable material to develop high surface area under the controlled conditions. In
addition, the activation method type and its operating condition influence the pore
structure of the product. Comparison of the experimental data shows that the
combined activation method is better than the chemical and/or physical activation
methods because of advantages such as synthesis of an activated carbon with a high
iodine number in moderate conditions. The optimum synthesis condition in the
combined activation method is a temperature 800 C, a dwell time of 45 min, and
an impregnation ratio of 50 %, which gives a product with an iodine number of
2433 mg/g and 52 % weight loss. The SEM images of this sample show the porous
morphology of the product, but a portion of the micropore volume in the total pore
volume can be approximated using the iodine number as compared to the methylene
blue number. The FTIR analysis illustrates that the carbonyl functional group was

123
Investigation on effects of operating conditions in…

removed from the char because of activation regarding of activation method type. In
addition, aliphatic functional groups decrease with activation, but the combined
activation method causes more aliphaticity loss.

References
1. E. Schröder, K. Thomauske, C. Weber, A. Hornung, V. Tumiatti, J. Anal. Appl. Pyrol. 79, 1–2 (2007)
2. J. Laine, A. Calafat, Carbon 29, 7 (1991)
3. H. Dolas, O. Sahin, C. Saka, H. Demir, Chem. Eng. J. 166, 1 (2011)
4. M.T. Izquierdo, A. Martinez, B. de Yuso, M. Rubio, R. Pino, Biomass Bioenergy 351, 1235 (2011)
5. M. Ozdemir, T. Bolgaz, C. Saka, O. Sahin, Anal. Appl. Pyrol. 92, 1 (2011)
6. Z. Hu, M.P. Srinivasan, Y. Ni, Carbon 39, 6 (2001)
7. O. Sahin, C. Saka, Bioresour. Technol. 136, 163–168 (2013)
8. C. Saka, Anal. Appl. Pyrol. 95, 1 (2012)
9. M. Özmak, Ph.D. Thesis in Chemical Engineering, Ankara University, Ankara, Turkey (2010)
10. M.A. Yahya, Z. Al-Qodah, C.W.Z. Ngah, Renew. Sust. Energy Rev. 46, 218–235 (2015)
11. H. Kamandari, H. Hashemipour, L. Saeednia, H. Najarzade, Res. Chem. Intermed. 40, 2 (2014)
12. H. Kamandari, H. Hashemipour, H. Najarzade, Z. Eksiri, Res. Chem. Intermed. 41, 1 (2015)
13. M. Jagtoyen, F. Derbyshire, Carbon 36, 7–8 (1997)
14. M.J. Ahmed, J. Eng. 17, 311–323 (2011)
15. M. Ozdemir, T. Bolgaz, C. Saka, O. Sahin, Anal. Appl. Pyrol 92, 1 (2011)
16. S.J. Allen, L. Whitten, G. McKay, Dev. Chem. Eng. Miner. Process. 5, 231–261 (1998)
17. H. Muhlen, K.H. Van Heek, W. Patrick Jr., Porosity in carbons: Characterization and Applications
(Halsted Press, London, 1995)
18. B. Viswanathan, P. Indra Neel, T.K. Varadarajan, Methods of Activation and Specific Applications of
Carbon Materials, 1st edn. (National Centre for Catalysis Research, Indian Institute of Technology,
Madras, 2009), pp. 19–26
19. T. Budinova, E. Ekinci, F. Yardim, A. Grimm, E. Björnbom, V. Minkova, M. Goranova, Fuel
Process. Technol. 87, 10 (2006)
20. IUPAC (International Union of Pure and Applied Chemistry Physical Chemistry Division Com-
mission on Colloid and Surface Chemistry Including Catalysis(, Pure. Appl. Chem. 57, 4 (1985(
21. A.N.A. El-Hendawy, Appl. Surf. 255, 6 (2009)
22. C.A. Nunes, M.C. Guerreiro, Quim Nova 34, 3 (2011)
23. A.C. Lua, T. Yang, J. Colloid Interface Sci 255, 6 (2009)
24. A.C. Lua, T. Yang, J. Colloid Interface Sci. 290, 2 (2005)
25. A.C. Lua, T. Yang, Colloid Interface Sci. 274, 2 (2004)

123