Chemical Engineering Journal 158 (2010) 129–142

Contents lists available at ScienceDirect

Chemical Engineering Journal

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

Development of mesoporous structure and high adsorption capacity of

biomass-based activated carbon by phosphoric acid and zinc chloride activation
Tzong-Horng Liou ∗
Department of Chemical Engineering, Ming Chi University of Technology, 84 Gungjuan Rd., Taipei 24301, Taiwan

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

Article history: This paper reports the preparation of activated carbon from two different types of agricultural biomass
Received 6 August 2009 materials, sugar cane bagasse and sunflower seed hull, by phosphoric acid and zinc chloride activation.
Received in revised form 4 December 2009 The experiments in this study vary the pre- and post-treatment procedures, the impregnation ratio of
Accepted 10 December 2009
the activating agent, and the carbonization temperature. In recent years, the high surface area and high
mesopore proportion of carbon have attracted a lot of attention for potential applications in the green
Keywords:
resources such as hydrogen energy storage and carbon dioxide capture. However, the traditional methods
Activated carbon
for fabricating activated carbon produce a mainly microporous structure. The experimental results show
Mesoporous structure
Sugar cane bagasse
that the activated carbon produced by base-leaching has a mostly mesoporous structure, which effectively
Sunflower seed hull enhances its adsorption capacity. The carbon materials obtained from zinc chloride activation of both
Chemical activation sugar cane bagasse and sunflower seed hull have mesopore volumes as high as 1.07 and 0.95 cm3 /g, and
mesopore contents of 81.2 and 74.0%, respectively. The surface area and pore volume of carbon produced
using zinc chloride activation were higher than that produced using phosphoric acid activation. The
total activation process of bagasse and hull occurs in three reaction stages. This study also presents a
corresponding pyrolysis mechanism that agrees well with the experimental results. The proposed method
of preparing mesoporous activated carbon is not complicated, and is suitable to bulk production.
© 2009 Elsevier B.V. All rights reserved.

1. Introduction the CO2 sorption of activated carbon plays an important role in

reducing carbon emissions [8].
Activated carbon is an extremely versatile material with a high Sugar cane and sunflower are two of the most commonly culti-
surface area. Due to its excellent adsorption capability, it is widely vated plants in the world. Sugar cane bagasse is the waste material
used in industrial wastewater and gas treatment [1,2]. The pore produced by sugar factories after sugar juice extraction. Sunflower
structure of activated carbon in conventional applications is mainly seed hull is produced by oil extraction plants that manufacture bio-
microporous. Recently, the potential use of activated carbon with a oil products. The production of sugar cane in Brazil, which is the
high proportion of mesoporous structures has attracted increasing largest global producer, totaled 425 million tons for the 2006/2007
attention because it possesses a high pore volume and a wide range season [9]. China and India are the two largest sugar cane produc-
of pore size. The mesoporous carbon materials are very suitable for ers in Asia. Together, these countries produced 437 million tons
use in large molecule adsorption, including battery capacitors, cat- of sugar cane in 2008 [10]. On the other hand, global sunflower
alyst supports, biomedical engineering and adsorbents for bulky production reached 27 million tons in 2007 [11]. These two crops
pollutants [3–5]. In the future, gasoline and diesel fuel for vehicles are abundantly available. Sugar cane and sunflower are also very
will probably be replaced for economy advantages. An important good candidates for bio-ethanol and bio-diesel, which are used as
application of activated carbon is related to the gas storage for alternative fuels [12,13]. The production of sugar cane and sun-
natural gas vehicles [6]. Moreover, researchers have paid signifi- flower can be anticipated to increase significantly in the future.
cant attention to the use of carbon materials in hydrogen storage. The observations above show that sugar, oil, and ethanol manu-
Large pores are necessary for the fast mass transfer of adsorbate facturing processes produce a large amount of sugar cane bagasse
to the bulk of the carbon material [7]. Hence, the high surface area and sunflower seed hull wastes each year. These wastes are usually
of activated carbon and its well-balanced micro/mesoporosity can burnt in the open or left in the field, and only a small portion is used
provide an appropriate medium for energy storage. In particular, as paper pulp or fuel. This creates disposal and pollution problems.
The activated carbon obtained from these two agricultural wastes
can provide useful, value-added products in the sugar factories, oil
∗ Tel.: +886 2 29089899; fax: +886 2 29083072. refineries, and related industries. Activated carbon for commercial
E-mail address: [email protected]. utilization can be obtained by thermal treatment of low-cost and

1385-8947/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.cej.2009.12.016
130 T.-H. Liou / Chemical Engineering Journal 158 (2010) 129–142

low ash content materials. Sugar cane bagasse just conforms to the 2. Experimental
rule for it is a low ash content and availability. In contrast, sun-
flower seed hull contains a high proportion of ash content. This 2.1. Materials
study compares the two carbon products obtained from these two
biomass sources. Sugar cane bagasse and sunflower seed hull were obtained from
Mesoporous activated carbon can be synthesized by many a sugar mill and a rural area, respectively. These two raw materials
methods. Traditional methods use carbon precursors such as lig- are respectively denoted as SCB and SSH hereafter. They were first
nocellulosic materials, coals, and phenolic resins in the presence washed thoroughly with distilled water to remove dust particles,
of transition metals, followed by a combination of carbonization and then dried at 100 ◦ C, crushed, and sieved to obtain a uniform
and physical activation with stream or carbon dioxide, to obtain grain size approximately 10 mm long and 2 mm wide. ASTM stan-
mesoporous activated carbon powder [14–16]. Shen et al. [17] used dard test procedures were applied for ash content, volatile matter
commercially activated carbon produced by steam activation with content and moisture content [25]. Fixed carbon was calculated
cerium oxide as a catalyst to prepare mesoporous carbon, and from the difference. Table 1 lists the SCB and SSH contents in the
found that the catalyst restricted the formation of mesoporous proximate analysis. The elemental analysis section below illus-
structures at temperatures of 800–870 ◦ C. Besides, some recent trates the carbon, hydrogen, oxygen, and nitrogen contents of both
studies have developed mesoporous carbon with pores that can raw materials. The weight percentages of ash content for SCB and
be controlled using mesoporous ceramic templates, such as SBA or SSH are 3.82 and 11.79 wt%, respectively. The ash content of SSH is
MCM series [18,19]. Compared to the above methods, ZnCl2 and three times as much as that of SCB.
H3 PO4 are considered more effective and less expensive activating
agents for producing mesoporous carbon [20,21]. Hu and Srini- 2.2. Preparation of activation carbon
vasan [22] produced mesoporous activated carbon using coconut
shells and palm seeds as starting materials, with simultaneous Preparing activated carbon from SCB and SSH can be divided into
activation by ZnCl2 and CO2 . They found that at a low ZnCl2 con- three stages, including base-leaching, activation, and acid-washing.
centration, the main pore characteristic is microporous; whereas at First, the water-rinsed bagasse and hull were leached with a
a high ZnCl2 concentration, the main pore characteristic is meso- 0.5 M NaOH (Merck & Co.) solution and boiled in a glass round-
porous. Kennedy et al. [23] studied the H3 PO4 activation of rice husk bottom flask at 100 ◦ C for 1 h. The basic solution was drained and
through precarbonization and chemical activation. They found that the remaining solid was washed with distilled water to remove the
pores created at a higher activation temperature were enlarged base. This process reduced the pH level to about 7. The material was
or widened, leading to the formation of mesoporosity. Generally, then dried at 105 ◦ C for 24 h.
activated carbon has a mesopore volume and a mesopore surface Second, H3 PO4 or ZnCl2 of chemically pure quality (Merck & Co.)
area in the range of 0.1–0.5 cm3 /g and 100–200 m2 /g, respectively were used as activating agents. A known mass of activating agent
[22]. However, for advanced applications, carbon materials should was mixed with distilled water, and the bagasse and hull were then
possess not only a high surface area, but also a high mesopore con- impregnated in the acidic solution. The impregnation ratio defined
tent. In particular, treating sugar cane bagasse and sunflower seed as the mass ratio of activating agent to dried solid, was 1–4. The
hull with an adequate base-leaching procedure enlarges or widens impregnated samples were dried in a sand bath at high temper-
the micropore structure of the resulting carbon during the acti- ature to remove residual water, and then oven-dried for 24 h. A
vation process. This, in turn, increases the mesopore surface area weighed amount of impregnated samples was then placed in a
and mesopore volume. After activation, the carbon precursor must fixed bed reactor made of quartz. The reactor was purged with
be washed with water to remove the chemical activating agent. high-purity (99.95%) nitrogen gas (San-Fu Chem. Co.) at a high
However, traditional water-washing cannot effectively remove all flow rate to avoid sample oxidation, and then the nitrogen flow
chemical residues, and may result in the reduction of pore volume. was adjusted to the desired rate. The reactor was inserted into
In this case, acid-washing procedure is an adequate alternative an electric furnace at reaction temperatures ranging from 400 to
[24]. 800 ◦ C, and the reaction continued for 1 h. The activated products
Previous studies have already reported the preparation of meso- were subsequently removed from the furnace and cooled to room
porous activated carbon from lignocellulosic materials by physical temperature.
or chemical activation. However, the mesopore surface area and After activating the samples, 3 M hydrochloric acid was used
mesopore volume of these materials are not high enough, and many to remove the phosphorous and zinc compounds. The activating
of the proposed approaches are complicated or require expen- agent and carbon mixtures were refluxed with hot acidic solu-
sive equipments. In addition, there is a lack of information in tion for 1 h. The remaining solid was then washed with distilled
the literature about activated carbon prepared from sugar cane water until neutralization. The washed samples were dried at
bagasse and sunflower seed hull using H3 PO4 or ZnCl2 in a base- 105 ◦ C for 24 h, and then ground to form a porous carbon pow-
leaching or acid-washing process. This lack in existing literature is der.
a motivation for the present study. The primary objective of this Fig. 1 shows an experimental flow chart illustrating the methods
study is to compare activated carbon prepared from sugar cane above to help compare the effect of base-leaching and acid-washing
bagasse and sunflower seed hull through chemical activation to on the surface characteristics of the samples. These tests were
produce a highly mesoporous activated carbon. The study investi- applied to several samples, as described below. Raw bagasse or
gates the effects of base-leaching, acid-washing, the impregnation
ratio of activating agent, the kind of activating agent, and the car-
bonization temperature on the development of pore characteristics Table 1
of activated carbon. Furthermore, the reactants and products are Proximate analysis of sugar cane bagasse and sunflower seed
hull.
examined by several forms of analysis, including SEM, XRD, FTIR,
ICP-MS, EA, TGA and a N2 -adsorption analyzer. The method of Proximate analysis SCB (wt%) SSH (wt%)
activated carbon preparation proposed in this study is not compli- Volatile matter 77.80 66.42
cated, and well suited to mass production. Simultaneously, it is an Ash 3.82 11.79
effective way to enhance the mesopore surface area and mesopore Fixed carbon 14.75 16.30
Moisture 3.63 5.49
volume of activated carbon.
 T.-H. Liou / Chemical Engineering Journal 158 (2010) 129–142 131

containing a 0.1 N iodine solution, and then shaken to maintain

equilibrium for 24 h. After adsorption, the adsorbent was removed
by filtration. The iodine adsorption capacity was determined from
the titration of the filtrate with a Na2 S2 O3 solution. The standard
deviations of the BET surface area and iodine adsorption capacity
are used as the basis for the error bars shown in the figures.

2.3.3. Thermogravimetric analysis

This study also used a Perkin-Elmer TGA7 thermogravimetric
analyzer to conduct experiments on the activation of SCB and SSH
impregnated with the activating agents [29]. For each experimental
run, a known weight of 6 ± 0.5 mg was placed on a platinum sample
pan. The samples were heated to 900 ◦ C at a rate of 10 ◦ C/min in a
nitrogen atmosphere. This study defines the remaining amount of
sample decomposed in nitrogen as W/Wo , where Wo and W repre-
sent the initial and instantaneous mass of the sample, respectively.
The TG and DTG curves were recorded simultaneously as the tem-
perature increased.
Fig. 1. Experimental flow chart expresses the methods used for treating the samples
as base-leaching or acid-washing process.
2.3.4. Calculation of activated carbon yield
hull only base-leached with NaOH, without chemical activation, The yield of activated carbon obtained from SCB and SSH for
is labeled as SB. Raw bagasse or hull that is first base-leached, and both H3 PO4 and ZnCl2 -activated samples, was calculated from the
then activated with H3 PO4 or ZnCl2 , and then washed with HCl solu- weight of resultant activated carbons divided by the weight of dried
tion, is labeled as SBAH. Samples in which acid-washing is replaced raw materials.
by a general water-washing are labeled as SBAW. Raw bagasse or
hull washed with distilled water, activated with H3 PO4 or ZnCl2 , 3. Results and discussion
and then washed with HCl solution is labeled as SWAH. Samples
in which acid-washing is replaced by a general water-washing are 3.1. Effect of base and acid treatment on activation
labeled as SWAW.
Table 2 shows the effects of thermal decomposition and acid-
2.3. Characterization washing on the purity of samples. Beside P and Zn elements, the
raw materials (S) for both SCB and SSH samples have the high-
2.3.1. Measure of physical and chemical properties est concentration of metallic ingredients, including Mg, Ca, Fe, Na,
To measure the amount of metallic impurities in the samples, K, and Cu. The SSH sample has a higher impurities content than
the reactant and product were dissolved in a solution of HNO3 and the SCB sample. These metal elements are clearly reduced when
HF, and heated at 180 ◦ C for 6 h. The amounts of metallic impurities samples are activated (SBAW or SBAH). It may be possible that the
in the solution were then determined with an inductively coupled metals are carried by the vaporization of organic matter during
plasma-mass spectrometer (ICP-MS) (Konton Plasmakon, model carbonization [30]. When raw samples are base-leached, activated,
S-35). The amount of fundamental organic element for various and water-washed (SBAW), the element P in carbons activated by
treatment procedures was determined using a Heraeus elemen- H3 PO4 is much higher than in raw material samples. A similar trend
tal analyzer. The dried sample was powdered to a 325 mesh size is also observed that Zn impurity in carbons activated by ZnCl2 is
(ASTM), and this powder was employed in the subsequent analysis. much higher than in raw material samples. This abrupt increase
The crystalline structures of the reactant and product were in the two impurities is because H3 PO4 and ZnCl2 remain in the
examined by an X-ray diffractometer (Siemens, Model D-500) using sample after activation, and cannot be effectively removed by a sim-
Cu-K␣ radiation with a scan speed of 2.5◦ min−1 . Fourier transform ple water-washing step. When the activated samples were treated
infrared (FTIR) spectra were also obtained using a spectrometer with a HCl solution (SBAH), the P and Zn elements were effectively
to determine the composition of samples (Shimadzu, model FTIR- reduced. Simultaneously, this process greatly decreased the con-
8300). The morphologies of the reactant and product were obtained centration of other metals. This decrease in impurities may be due
with a scanning electron microscope (Topcon, model AST-150S). to metals dissolving in the acidic solution, and subsequent leach-
ing and filtration. Experimental results show that the proportion
2.3.2. Characteristics of N2 and I2 adsorption of residual metals in the ZnCl2 -activated sample is lower than that
The samples were characterized in terms of specific surface in H3 PO4 -activated sample. In another words, Zn can be removed
area, pore volume, and pore diameter. Samples were measured much more easily than P. This finding can be explained by the FTIR
at −196 ◦ C using the adsorption of nitrogen (Micrometric, model spectrum below, which shows that the phosphate and polyphos-
ASAP2010). The specific surface area of each specimen was cal- phate species incorporated into the carbon mixture through the
culated using the BET method [26]. This study assumes that the C–O–P bands block the some of the pores in the carbon samples
across-sectional area of a nitrogen molecule is 0.162 nm2 . The sin- [31]. As a result, phosphorus compounds are not easily removed
gle point total pore volume was determined from the amount of with washing. However, zinc salt is highly soluble in water, and
adsorbed nitrogen expressed in liquid form at a relative pressure is therefore easily removed in the washing stage [30]. The final
of approximately 0.95. The DR equation was applied to determine purity of ZnCl2 -activated samples for both SCB and SSH is as high
the micropore volume from the N2 -adsorption data [27]. This study as 99.95%.
assumes that micropores are less than 2 nm wide, mesopores are Table 3 lists the elemental analysis for raw bagasse and hull,
2–50 nm wide, and macropores are more than 50 nm wide [28]. base-leached, and activated samples. The major organic elements
Adsorption tests were conducted using iodine adsorption capac- in raw materials (S) for both SCB and SSH are carbon and oxygen.
ity. The known amount of activated carbon was placed in a flask The SSH sample has a higher ash content than the SCB sample. The
132 T.-H. Liou / Chemical Engineering Journal 158 (2010) 129–142

Table 2
Amount of metallic ingredients before and after the activation of samples.

SCB (ppm) SSH (ppm)

a b b c c
S SBAW SBAH SBAW SBAH Sa SBAWb SBAHb SBAWd SBAHd

Mg 623 155 48 100 ND 3,950 86 ND 460 ND

Ca 635 330 125 ND ND 12,300 470 350 200 120
Fe 477 106 ND 458 ND 1,480 95 ND 50 ND
Na 4089 504 251 265 69 190 390 330 320 160
K 778 97 39 ND ND 5,330 190 ND ND ND
Mn ND ND ND ND ND 49 ND ND ND ND
Cu 29 39 ND 37 ND 59 ND ND 40 ND
P 570 62100 36800 ND ND 1,850 47800 27100 ND ND
Zn ND ND ND 28200 50 88 38 240 43300 150
Total 7201 63331 37263 29060 119 25,296 49069 28020 44370 430

ND means not detected.

a
Samples were water-rinsed and un-heated.
b
Samples were activated with H3 PO4 at 400 ◦ C.
c
Samples were activated with ZnCl2 at 500 ◦ C.
d
Samples were activated with ZnCl2 at 600 ◦ C.

inorganic components in the ash are mainly silica and metal impu- ous treatment procedures. The BET surface areas are the lowest for
rities, which agrees with the results of other studies [32,33]. When non-base and non-acid treated samples (SWAW). The surface area
the raw materials are base-leached (SB), the ash content clearly obtained mainly comes from the vaporization of organic matter
decreases. This reduction in ash content is due to the removal of sil- during the activation process, leaving the porous surface. The sur-
ica in the sample. The silica reacts with the NaOH and forms sodium face area of the SCB-derived carbon obtained from SWAW process
silicate (Na2 SiO3 ) solution [34], and then is removed by filtration. is 886 m2 /g, which is very close to the value of 881 m2 /g reported by
When raw material samples were activated with H3 PO4 or ZnCl2 Castro et al. [36] in their study of the H3 PO4 activation of sugar cane
(SBAH), the percentage of carbon increased. However, hydrogen bagasse at 400 ◦ C. When SCB or SSH samples are treated with SWAH
and oxygen exhibit the opposite trend. As generally recognized, process, the surface areas of activated carbon increase significantly.
hydrogen and oxygen can react with each other or with carbon. As This can be explained by the metallic ingredient analysis in Table 2,
a result, the gaseous volatiles are formed and are then released from which shows that the residual activating agents adhered to the
the sample surface during pyrolysis [35]. The proportion of carbon interior of the carbon samples after activation block the pores. This
in SCB-derived carbon samples is very close to that of SSH-derived leads to a reduction in the sample’s adsorption capacities. However,
carbon samples. For both SCB or SSH samples, the ash content in the activating agent can be removed by an acid-washing process,
H3 PO4 -activated samples is higher than that in ZnCl2 -activated which increases the surface area. When SCB or SSH samples are
samples. This can be explained according by Table 2, which shows base-leached, activated, and water-washed (SBAW), the BET sur-
Zn salts are easily removed in the washing stage. However, P com- face area and iodine adsorption capacity are abruptly enhanced. The
pounds are not easily removed due to the bonding of phosphorus base-leaching process (SBAW) enhances the surface area more than
and carbon in the activated sample. the acid-washing process (SWAH). The element analysis in Table 3
Fig. 2 plots the BET surface area and iodine adsorption capac- shows that the raw bagasse and hull contain the ash content. The
ity of samples as a function of the base-leaching or acid-washing inorganic components in the ash probably prevent porosity evolu-
procedure. Experimental results show that the base-leaching and tion by blocking pore entrances during the activation process [37].
acid-washing have obvious effect on the adsorption performance of Hence, when the ash is removed by an adequate base-leaching step,
samples. For all samples, the iodine adsorption capacity increases as this increases the surface area of the activated carbon. Soleimani
the BET surface area increases. The BET surface areas (do not show in and Kaghazchi [38] reported that activated carbon with a high ash
the figure) of both SCB and SSH for the raw material samples (S) are content is undesirable because it reduces the mechanical strength
0.60 and 0.06 m2 /g, and that of base-leached samples (SB) are 1.56 and adsorption capacity of the carbon, which is consistent with
and 0.59 m2 /g, respectively. The base-leached samples can remove the present experimental results. The BET surface area and iodine
the ash and leave a number of tiny pores, and therefore increase the adsorption capacity reach their maximum when the activated SCB
surface area of carbon precursor. Fig. 2(a) and (b) shows the results and SSH samples are base-leached and acid-washed (SBAH). This
of SCB and SSH carbon samples for H3 PO4 activation using vari- indicates that combining both base-leaching and acid-washing pro-
cedures creates a dual influence after activation. Fig. 2(c) and (d)
illustrate the BET surface area and iodine adsorption capacity of
Table 3
Elemental composition and ash content before and after the activation of samples.
ZnCl2 -activated samples obtained from before and after treatment
procedures. The effect of ZnCl2 activation on the sample’s surface
SCB (wt%) SSH (wt%) characteristics is similar to that of H3 PO4 activation, which can be
Sa SBb SBAHc SBAHd Sa SBb SBAHc SBAHe enhanced by acid-washing or base-leaching processes.
C 45.03 44.88 63.68 84.45 41.60 45.81 62.74 86.51
The experimental results of this study show that base-leaching
H 6.88 6.05 4.67 3.56 5.20 5.58 4.22 2.68 offers several potential benefits: (i) reducing the ash content in the
Of 44.03 48.38 27.72 11.48 40.49 47.31 29.87 10.34 samples to avoid obstructing the pore development, (ii) helping the
N 0.24 0.06 0.08 0.06 0.92 0.14 0.12 0.22 activating agent enter the interior of the carbon tissue after remov-
Ash 3.82 0.63 3.85 0.45 11.79 1.16 3.05 0.25
ing the ash, which increases the contact area between the activating
a
Samples were water-rinsed and un-heated. agent and the carbon precursor, and (iii) allowing the base solution
b
Samples were base-leached and un-heated. to decompose or soften the organic matter in the raw material [39],
c
Samples were activated with H3 PO4 at 400 ◦ C.
d
Samples were activated with ZnCl2 at 500 ◦ C.
which facilitates the pyrolysis of bagasse or hull into carbon during
e
Samples were activated with ZnCl2 at 600 ◦ C. activation. For these reasons, base-leaching can increase the sur-
f
By difference. face area of activated carbons. The surface areas show an increase
 T.-H. Liou / Chemical Engineering Journal 158 (2010) 129–142 133

Fig. 2. Effect of base and acid treatment on the surface area and iodine adsorption capacity of samples at impregnation ratio of 2: (a) SCB-based carbon with H3 PO4 activation
at 400 ◦ C, (b) SSH-based carbon with H3 PO4 activation at 400 ◦ C, (c) SCB-based carbon with ZnCl2 activation at 500 ◦ C, and (d) SSH-based carbon with ZnCl2 activation at
600 ◦ C.

of 480–605 m2 /g by H3 PO4 activation and 225–390 m2 /g by ZnCl2 when the activation temperature exceeds 500 ◦ C, the resulting sur-
activation, respectively. Yun et al. [37], who studied the carboniza- face area decrease may be due to contraction or collapse of pores at
tion of rice straw as a biomass precursor, reported similar findings. high temperature [40]. Under the same conditions, activated car-
They found that removing the inorganic component in the raw bon from sugar cane bagasse exhibits a higher BET surface area and
straw creates new pores and voids, which increase the surface area iodine adsorption capacity than carbon prepared from sunflower
and pore volume. According to the experimental results in Fig. 2, seed hull. This indicates that the performance of the resulting
the surface area and adsorption capacity of carbon can be enhanced carbon depends on the raw material used as the precursor. This
using a combination of base-leaching and acid-washing processes. discrepancy probably arises from the different constituents inher-
Hence, the remaining experiments in the study are conducted with ent in each precursor, including cellulose, hemicellulose and lignin.
both procedures. These organic matters can lead to different pore evolution mecha-
nism in the carbon [38,41]. The optimum surface areas for SCB and
3.2. Effect of H3 PO4 on activation reaction SSH samples for H3 PO4 activation at 500 ◦ C are 1611 and 1543 m2 /g,
respectively.
Fig. 3 shows the BET surface area and iodine adsorption capacity
of H3 PO4 -activated samples with different activation tempera- 3.3. Effect of ZnCl2 on activation reaction
tures and impregnation ratios. Fig. 3(a) and (b) shows the effect
of impregnation ratio on the surface characteristics of SCB and Fig. 4 illustrates the effect of ZnCl2 activation on the BET sur-
SSH samples. As the impregnation ratio increases from 1 to 2, the face area and iodine adsorption capacity of samples. The adsorption
adsorption capacities are enhanced. This is due to an increase in behavior of ZnCl2 -activated samples is the same as that of H3 PO4 -
the contact area between the raw material and the activating agent. activated samples. The highest surface area and adsorption capacity
This, in turn, increases the release of volatiles and increases the pore of samples occur at an activation temperature and impregnation
surface area. However, further increasing the impregnation ratio ratio of 500 ◦ C and 2, respectively. Yorgun et al. [25] studied the
from 2 to 4 causes a decrease in adsorption capacity. This obser- ZnCl2 activation of Paulownia wood. They pointed out that ZnCl2
vation is probably due to the contraction of pore structure caused worked as a dehydration regent during carbonization. This result
by the thermal breakdown of phosphate esters at a high ratio of leads to the charring and aromatization of the carbon skeleton and
activating agent to carbon precursor [36]. Like the impregnation creation of the porous structure. On the other hand, ZnCl2 can also
ratio samples above, Fig. 3(c) and (d) shows the surface area and restrict the tar formation. When the reaction temperature is lower
adsorption capacity for both SCB and SSH under various activation than 500 ◦ C, more volatiles are released by increasing the reaction
temperatures. These figures indicate that the highest adsorption temperature, and the surface area increases. However, a reaction
capacity occurs at the activation temperature of 500 ◦ C. Increasing temperature greater than 500 ◦ C, this decrease in surface area of
the activation temperature increases the thermal decomposition carbon can be attributed to the heat shrinking [42]. Lua and Yang
rate, and therefore increases the surface area of carbon. However, [43] reported a similar trend that increasing the impregnation ratio
134 T.-H. Liou / Chemical Engineering Journal 158 (2010) 129–142

Fig. 3. Effect of H3 PO4 activation on the surface area and iodine adsorption capacity of samples: (a) SCB-based carbon at 400 ◦ C, (b) SSH-based carbon at 400 ◦ C, (c) SCB-based
carbon at impregnation ratio of 2, and (d) SSH-based carbon at impregnation ratio of 2.

Fig. 4. Effect of ZnCl2 activation on the surface area and iodine adsorption capacity of samples: (a) SCB-based carbon at 500 ◦ C, (b) SSH-based carbon at 600 ◦ C, (c) SCB-based
carbon at impregnation ratio of 2, and (d) SSH-based carbon at impregnation ratio of 2.
 T.-H. Liou / Chemical Engineering Journal 158 (2010) 129–142 135

of ZnCl2 from 0.25 to 1.50 increases the pore surface area of carbon. Table 4 shows the specific surface area and pore volume of
A comparison of Figs. 3 and 4 shows that ZnCl2 -activated carbon carbon samples obtained from H3 PO4 activation at various tem-
always has greater surface areas and adsorption capacities than peratures. When the SCB and SSH carbons are obtained neither
H3 PO4 -activated carbon. The optimum surface areas of SCB and acid-treated nor base-treated (SWAW), they exhibit significantly
SSH carbons for ZnCl2 activation at 500 ◦ C are 2289 and 2240 m2 /g, lower surface areas. The two carbon samples have higher micro-
respectively. porous volume than mesoporous volume. However, the total pore
volumes of the carbon are only 0.448 and 0.475 cm3 /g, respectively.
A similar property is also observed by other studies [36] that acti-
3.4. Pore structure analysis
vated carbon obtained from H3 PO4 activation of non-base-treated
sugar cane bagasse, had the main pore characteristic of microp-
Fig. 5 plots the nitrogen adsorption–desorption isotherms of
orosity. Both the pore volume and specific surface area increase
H3 PO4 -activated samples at reaction temperatures ranging from
when the two raw materials are activated and followed by acid
400 to 700 ◦ C. The curve trends for volume adsorbed from SCB and
and base treatments (SBAH). This indicates that removing ash and
SSH are vary similar. When raw material samples are un-treated
residual activating agent effectively enhances the pore volume
(SWAW samples), the adsorption curves are Type I isotherms that
and surface area. Experimental results also show that the carbon
do not display an apparent desorption hysteresis loop, in accor-
obtained from SBAH process produces a higher mesopore volume
dance with IUPAC classification [44]. This adsorption behavior
and mesopore surface area. As mentioned earlier, base-leaching
implies that the carbon materials are highly microporous. When
raw material samples decreases the ash content and leaves a num-
samples are base- and acid-treated at various activation tempera-
ber of tiny pores. Simultaneously, NaOH decomposes the tissue
tures, the isotherms display an abrupt increase in adsorbed volume
of the carbon precursor, which also creates some new pores and
at a low relative pressure (P/Po ). This adsorption behavior exhibits
voids. The activating agent permeates these tiny pores or voids,
a Type I isotherm as characteristic of microporous structure [45].
which increase the contact area between the activating agent and
As the relative pressure increases, the nitrogen uptake increases
the carbon precursor. Consequently, this promotes the release of
across the entire pressure range. Furthermore, as the relative pres-
volatiles from the carbon structure and widens the micropores
sure increases to 0.4, the isotherms exhibit apparent hysteresis
in the original carbon structure, converting them into mesopores.
loops. This adsorption behavior exhibits a Type IV isotherm as
In addition, the activating agent probably deposits on the inte-
characteristic of mesoporous structure. Therefore, the activated
rior of the pores after water-washing, as Fig. 2 indicates. This may
carbon obtained from various activation temperatures show both
lead to pore contraction and microporosity formation. However,
microporous and mesoporous structures [22]. The samples’ nitro-
acid-washing removes the activating agent enlarging the micro-
gen adsorption capacities also change with activation temperature,
pores and increasing mesoporosity. An increase in the activation
and reach their maximum at 500 ◦ C.
temperature produces optimum surface area and pore volume val-
ues. The highest pore volumes, which occur at 500 ◦ C, are 1.233
and 0.999 cm3 /g for SCB- and SSH-derived carbon, respectively.
The optimum temperature of 500 ◦ C has also been reported by
other studies, which prepared activated carbon from lignocellu-
losic materials such as cherry stones using H3 PO4 activation [46].
Experimental results show that activated carbon prepared from
sugar cane bagasse possesses a higher pore volume and surface area
than sunflower seed hull. The trends of total pore volume and sur-
face area are reflected in the mesopore pore volume and mesopore
surface area. Consequently, both SCB and SSH carbon samples pro-
duced at 500 ◦ C posses mesopore volumes of 0.802 and 0.696 cm3 /g,
and mesopore surface areas of 1068 and 975 m2 /g, respectively.
The mesopore contents of 500 ◦ C H3 PO4 -activated carbon approach
65.0 and 69.7% for SCB and SSH, respectively. The mesopore con-
tents exceed the expected proportion of conventional microporous
carbon. The average pore diameters of H3 PO4 -activated samples for
both SCB and SSH range from 2.2 to 3.1 nm and 2.2 to 2.6 nm, respec-
tively. The carbon yield decreases as the reaction temperatures
increases. A higher reaction temperature releases more volatiles
and therefore produces a lower yield. The yield values of base-
and acid-treated carbon (SBAH) are lower than un-treated carbon
(SWAW). This is because base-leaching removes a portion of the
organic matter in the precursor (according to SEM observation),
and this causes a decrease in the carbon yield. Further, SSH-derived
samples yield more carbon than SCB-derived samples. This can be
explained by the fact that the mass loss of organic matter in SCB is
higher than that of SSH after base-leaching. This, in turn, causes a
reduction in the carbon yield of the SCB samples.
Fig. 6 illustrates the pore characteristics of samples activated
with ZnCl2 at various activation temperatures. These isotherms
resemble a combination of Type I and Type IV isotherms. All
samples show a broad knee and a pronounced desorption hystere-
sis loop. The slope of the plateaus in these isotherms gradually
Fig. 5. Adsorption–desorption isotherm of H3 PO4 activation samples at impregna- increases as increasing the relative pressure. These results indi-
tion ratio of 2: (a) SCB-based carbon and (b) SSH-based carbon. cate that the pore structure of ZnCl2 -activated carbon is mainly
136 T.-H. Liou / Chemical Engineering Journal 158 (2010) 129–142

Table 4
Pore volume, surface area, and activated carbon yield for the samples activated by H3 PO4 .

SCB (◦ C) SSH (◦ C)
a b b b b
400 400 500 600 700 400a 400b 500b 600b 700b

Vt (cm3 /g) 0.448 0.846 1.233 0.900 0.814 0.475 0.838 0.999 0.798 0.761
Vmic (cm3 /g) 0.273 0.336 0.129 0.210 0.287 0.258 0.313 0.191 0.213 0.219
Vmeso (cm3 /g) 0.162 0.489 0.802 0.568 0.445 0.200 0.474 0.696 0.506 0.461
Vmeso /Vt (%) 36.16 57.80 65.04 63.11 54.67 42.11 56.56 69.67 63.41 60.58
SBET (m2 /g) 886 1521 1611 1373 1357 915 1470 1543 1252 1221
Smeso (m2 /g) 260 717 1068 773 623 313 694 975 687 645
Dp (nm) 2.02 2.23 3.06 2.62 2.40 2.07 2.28 2.59 2.55 2.49
Yield (%) 46.42 25.42 24.87 23.84 21.68 46.65 33.14 32.25 32.20 30.65

Vt = total pore volume, Vmic = micropore volume, Vmeso = mesopore volume, SBET = BET surface area, Smeso = mesopore surface area, Dp = average pore diameter calculated as
4V/A, where A is the BET surface area and V is the single point total pore volume.
a
Samples were treated with SWAW process.
b
Samples were treated with SBAH process.

Table 5
Pore volume, surface area, and activated carbon yield for the samples activated by ZnCl2 .

SCB (◦ C) SSH (◦ C)
a b b b b
500 400 500 600 700 600a 400b 500b 600b 700b

Vt (cm3 /g) 1.082 1.026 1.358 1.311 1.091 0.896 0.776 1.318 1.116 1.030
Vmic (cm3 /g) 0.298 0.419 0.237 0.153 0.202 0.273 0.384 0.292 0.283 0.182
Vmeso (cm3 /g) 0.667 0.545 1.046 1.065 0.836 0.541 0.312 0.952 0.790 0.762
Vmeso /Vt (%) 61.65 53.12 77.03 81.24 76.63 60.38 40.21 72.23 70.79 73.98
SBET (m2 /g) 1825 1872 2289 2060 1817 1527 1478 2240 1956 1676
Smeso (m2 /g) 1197 852 1567 1544 1233 983 492 1421 1199 1124
Dp (nm) 2.27 2.19 2.37 2.55 2.40 2.14 2.10 2.35 2.28 2.46
Yield (%) 41.08 21.29 19.27 19.07 17.60 43.50 29.62 29.11 27.32 25.04

Vt = total pore volume, Vmic = micropore volume, Vmeso = mesopore volume, SBET = BET surface area, Smeso = mesopore surface area, Dp = average pore diameter calculated as
4V/A.
a
Samples were treated with SWAW process.
b
Samples were treated with SBAH process.

mesoporous. At the same activation temperature, the nitrogen

adsorption capacities of un-treated samples are clearly lower than
that of base-treated samples. Nitrogen adsorption capacity reaches
its highest value at an activation temperature of 500 ◦ C. Table 5
illustrates the pore volume, surface area, and activated carbon yield
for the samples activated by ZnCl2 . The pore structure of carbon
prepared without base-leaching and acid-washing exhibits a mix-
ture of both micropore and mesopore. When the activated carbon
samples are acid-washed and base-leached, the mesopore volume
and mesopore surface area increase further. These results suggest
that removal of ash transforms the micropores into mesopores due
to pore widening. The SCB and SSH samples have the highest pore
volumes of 1.358 and 1.318 cm3 /g at 500 ◦ C, respectively. The SCB
and SSH carbon samples produced at 500 ◦ C possess mesopore vol-
umes of 1.046 and 0.952 cm3 /g and mesopore surface areas of 1567
and 1421 m2 /g, respectively. The proportion of mesopore volume
in SCB and SSH samples activated with ZnCl2 at 500 ◦ C approach
the values of 77.0 and 72.2%. These values are higher than those of
conventional activated carbon. Moreover, the mesopore content of
the SCB sample reaches a value of 81.2% at 600 ◦ C.
ZnCl2 -activated samples possess a higher mesopore volume
and mesopore surface area than H3 PO4 -activated samples. This
is because acid-washing removes ZnCl2 much more easily than
H3 PO4 , reducing the possibility of activating agent deposition on
the pores and sample surface. This, in turn, increases the meso-
pore volume and mesopore surface area. The ratio of micropores
to mesopores in activated carbon can be adjusted by changing
the before and after treatment procedure, the type of activating
agent, and the activation temperature. The average pore diameters
of ZnCl2 -activated samples range from 2.2 to 2.6 for SCB and 2.1
to 2.5 nm for SSH. The activated carbon yield of samples decreases
Fig. 6. Adsorption–desorption isotherm of ZnCl2 activation samples at impregna- as the reaction temperature increases, as can be seen in Table 5.
tion ratio of 2: (a) SCB-based carbon and (b) SSH-based carbon. The SSH-derived samples yield more carbon than the SCB-derived
 T.-H. Liou / Chemical Engineering Journal 158 (2010) 129–142 137

samples. These results are similar to those of the H3 PO4 activation of the graphic structures. This produces relatively well-organized
procedure (Table 4). aromatic carbon that is more stable than amorphous-like carbon.
Fig. 7 shows the pore size distribution determined using the Fig. 9 shows the FTIR spectroscopy of samples used to study the
BJH method for SCB and SSH samples activated with H3 PO4 and chemical compositions of the raw materials and carbon products.
ZnCl2 [17]. For un-treated samples (SWAW), the carbon obtained The spectra obtained form SCB and SSH have many similar features
from H3 PO4 activation shows a weak peak at a pore diameter of because they are lignocellulosic materials [43]. For the raw material
about 3.5 nm. A narrow peak of carbon obtained from ZnCl2 activa- samples of both SCB and SSH, a wide band located in the range
tion for the un-treated process approaches 3.2 nm. When samples of 3200–3400 cm−1 is a consequence of O-H groups and adsorbed
are treated with base-washing and acid-washing (SBAH), the peak water [49]. The band at 2918 cm−1 is due to C-H vibration. The
position shifts to 3.5 nm after ZnCl2 activation, and reaches a value band at 1737 cm−1 is related to the stretching of C O groups. The
of 3.9 nm for H3 PO4 -activated samples. This indicates that many band at 1646 cm−1 may be due to the stretch vibrations of C C
micropores are enlarged, and the structure eventually becomes bonds in olefinic structures. The two bands at 1505 and 1425 cm−1
mesoporous. are the skeletal C C vibrations of aromatic rings. The stretching
vibrations of –CH3 at 1378 cm−1 are related to methyl structures.
3.5. Analysis of physical properties The adsorption between 1000 and 1350 cm−1 may be assigned to
C–O vibrations. The O–H band is located at 665 cm−1 . The spectra
Fig. 8 shows the XRD patterns of raw materials and activated of activated carbon are different from those of raw materials. Many
carbon obtained from H3 PO4 and ZnCl2 activation at various reac- bands disappear, indicating the vaporization of organic matters. All
tion temperatures. In the raw SCB and SSH samples, a broad peak of the carbon samples show a wide band at about 3425 cm−1 , which
appearing at 2Â = 22.5◦ is a typical silica characteristic, which can is due to the O-H vibrations of hydroxyl groups. When SCB and
be attributed to the presence of disordered cristobalite [47]. The SSH are activated with ZnCl2 , the band located at about 1604 cm−1
peak disappears when raw materials are base-treated and activated is attributable to C C vibrations. The bands located at 1165 and
with H3 PO4 and ZnCl2 at all temperatures, indicating the removal 864 cm−1 are due to C–O and C–H vibrations [43]. When SCB and
of ash. Two broad peaks appear at approximately 2Â = 25◦ and 45◦ SSH are activated with H3 PO4 , the bands that appear between 900
after activation with H3 PO4 and ZnCl2 , indicating the formation of and 1200 cm−1 may be attributed to the presence of phosphorus
turbostratic structure of disordered carbon. The peaks at 2Â = 45◦ species in the samples [48]. The peak at 1175 cm−1 can be assigned
became stronger as the reaction temperature increases. This indi- to C–O stretching vibrations in a C–O–P linkage. The shoulder at
cates that carbon tends to crystallize at an elevated temperature. 1053 cm−1 may represent a chain of P–O–P vibrations.
The peak intensity for samples using the ZnCl2 activation proce- Fig. 10 shows electron microscope images of raw materials and
dure is higher than that using the H3 PO4 activation procedure. carbon samples obtained from SCB and SSH. Fig. 10(a) and (b)
Kennedy et al. [48] reported that peaks at 2Â = 45◦ are due to the cre- shows the outer epidermis and inner surface of original sugar cane
ation of pores by the decomposition of carbon along the direction bagasse, respectively. The outer epidermis of the bagasse has a

Fig. 7. Pore size distribution of samples treated with SWAW or SBAH process: (a) SCB-based carbon with H3 PO4 activation at 400 ◦ C, (b) SSH-based carbon with H3 PO4
activation at 400 ◦ C, (c) SCB-based carbon with ZnCl2 activation at 500 ◦ C, and (d) SSH-based carbon with ZnCl2 activation at 600 ◦ C.
138 T.-H. Liou / Chemical Engineering Journal 158 (2010) 129–142

Fig. 8. X-ray diffractogram of samples activated with H3 PO4 or ZnCl2 : (a) SCB-based carbon and (b) SSH-based carbon.

smooth and pitted surface. The inner surface of the bagasse is a long the base-leaching procedure. Rao et al. [50] reported that soaking
and narrow strip. Fig. 10(c) shows the inner surface of base-leached raw bagasse in a NaOH solution can remove the lignin content,
bagasse. The surface characteristic of the sample is slightly altered which is similar to the present experimental results. Fig. 10(d)
by the base-leaching process, and appears clear-cut, indicating that shows that the morphology of SCB-derived carbon powder acti-
a portion of the organic tissue of the SCB sample is dissociated after vated with H3 PO4 is a long column. However, activated carbon
produced by ZnCl2 activation has an irregular shape, as Fig. 10(e)
indicates. Fig. 10(f) shows the outer epidermis of a sunflower seed
hull, which is orderly tubular tissue. Fig. 10(g) shows the inner sur-
face of sunflower seed hull. The morphology of hull is similar to
that of the SCB (Fig. 10(b)), which is a long narrow strip. Fig. 10(h)
shows the outer epidermis of base-leached hull. The surface of the
outer epidermis is peeled off and contains a number of rectangular
tissues, indicating the removal of part of the hull’s organic matter.
Fig. 10(i) and (j) shows the morphology of carbon powders obtained
from SSH and activated with H3 PO4 and ZnCl2 , respectively. These
figures show that the two activation procedures do not significantly
affect the surface morphology of the resulting carbons, which has
an irregular grain shape. However, ZnCl2 -activated carbon has a
few small pores distributed within the solid.

3.6. Analysis of thermal characteristic and reaction path

Fig. 11 shows the TG and DTG curves of samples activated with

H3 PO4 and ZnCl2 at various impregnation ratios. Fig. 11(a) and (b)
shows the TGA curves of H3 PO4 -activated SCB and SSH, respec-
tively. The TG curves exhibit three temperature zones. This mirrors
the fact that the activation of bagasse and hull with H3 PO4 occurs in
three main stages. At a constant activation temperature, the mass
loss caused by specimen decomposition increases as the impreg-
nation ratio increases. In the first stage, the small amount of mass
loss in the TG curves as the temperature reaches ∼300 ◦ C is due
to the pyrolysis of organic matter in the carbon precursor. The
mass loss of the heating curves in the second stage ranged from
300 to 550 ◦ C, which is smaller than that in the first stage. This
Fig. 9. FTIR spectrogram of samples activated with H3 PO4 or ZnCl2 : (a) SCB-based loss may be associated with a transition period in the activation.
carbon and (b) SSH-based carbon. A significant mass loss occurs in the third stage, as the temper-
 T.-H. Liou / Chemical Engineering Journal 158 (2010) 129–142 139

Fig. 10. Scanning electron microscope images of samples: (a) and (b) outer epidermis and inner surface of SCB, (c) inner surface of base-leached SCB, (d) and (e) SCB-based
carbon powders by H3 PO4 and ZnCl2 activation, respectively, (f) and (g) outer epidermis and inner surface of SSH, (h) outer epidermis of base-leached SSH, (i) and (j) SSH-based
carbon powders by H3 PO4 and ZnCl2 activation, respectively.
140 T.-H. Liou / Chemical Engineering Journal 158 (2010) 129–142

ature increases from 550 to 900 ◦ C. The DTG curves in Fig. 11(a) These observations also suggest that the ZnCl2 activation occurs
and (b) shows that two peaks appear in each heating curve. The more quickly and easily than H3 PO4 activation.
temperature of the first peak increases as the impregnation ratio Due to the complexity of this activation, researchers do not fully
increases, while the temperature of the second peak decreases as understand the pyrolysis paths of activation. The different pore
the impregnation ratio increases. Furthermore, the height of the characteristics of carbon activated by either H3 PO4 or ZnCl2 acti-
peak, which is associated with the instantaneous rate of pyrolysis, vation depend on the process conditions and the composition of
also increases as the impregnation rate increases. This indicates raw materials. However, this study reveals the main mechanism
that the reaction rate can be enhanced by increasing the amount for this activation. Essentially, the organic constituent of SCB and
of chemical activating agent. Experimental results also show that SSH is a biopolymer including cellulose, hemi-cellulose and lignin
SCB samples have slightly lower activation temperatures than SSH [38,51]. The hemi-cellulose and lignin are probably embedded in
samples. the interior of the cellulose matrix [40]. Thermogravimetric analy-
Fig. 11(c) and (d) shows the TGA plots of ZnCl2 -activated SCB sis (Fig. 11) shows that a small amount of mass loss appears in the
and SSH obtained at various impregnation ratios of activating initial stage of both ZnCl2 and H3 PO4 activation. This indicates that
agent. The mass loss in these samples occurs in three tempera- the two chemicals first attack the organic matter in the bagasse or
ture zones, namely ∼200, 200–450 and 450–800 ◦ C. Qian et al. [35] hull [52]. In this case, chemical reactions between the activating
reported that three stages of thermal decomposition were observed agent and biomass cause the decomposition or depolymerization
in the chemical activation of cattle-manure compost with ZnCl2 , of celluloses and lignin into intermediates with a smaller molar
which is similar to the present experimental observations. The mass. Simultaneously, gaseous volatiles such as H2 O, CO, CO2 , CH4
DTG curves reveal two peak temperatures for SCB and SSH sam- and aldehydes are released from the sample surface [35]. This is
ples activated at various impregnation ratios. The temperature of consistent with FTIR spectra (Fig. 9) observations, in which many
the first peak increases as the impregnation ratio increases, while bands disappear after activation, indicating the volatilization of
the temperature of the second peak increases as the impregna- organic matter. Jagtoyen and Derbyshire [40] reported that the
tion ratio decreases. These results are similar to those of the H3 PO4 mass loss occurring during the primary activation is due to a loss
activation procedure. Thermogravimetric analysis shows that the of aliphatic character, which simultaneously leads to an increase in
thermal decomposition mechanism of SCB is very similar to that of aromaticity by rearranging the atomic structure. As the activation
SSH for both H3 PO4 and ZnCl2 activation. However, SCB activation temperature increases, the mass loss in the second stage begins
has a higher instantaneous rate of thermal decomposition and a to transform gently. The main characteristic of the reaction at this
lower peak temperature, indicating that the SCB activation occurs stage is probably due to an expansion of the volumetric structure
more easily than SSH activation. A comparison of H3 PO4 -activated of the samples, which in turn reduces the release of volatiles [36].
carbon and ZnCl2 -activated carbon shows that ZnCl2 activation pro- The XRD curves in Fig. 8 reflect the same trends in the volumetric
duces a lower peak temperature and shorter duration of pyrolysis, expansion, in that peak intensity increases as the reaction tem-
as well as a higher instantaneous rate of thermal decomposition. perature increases. This indicates an increase in crystalline size or

Fig. 11. TG and DTG thermogram of samples at impregnation ratios of 2 and 4: (a) SCB-based samples with H3 PO4 activation, (b) SSH-based samples with H3 PO4 activation,
(c) SCB-based samples with ZnCl2 activation, and (d) SSH-based samples with ZnCl2 activation.
 T.-H. Liou / Chemical Engineering Journal 158 (2010) 129–142 141

aromatic structure of carbon when reaction temperature increases. be illustrated experimentally. The proposed method of preparing
Olivares-Marin et al. [53] further showed that the mass loss decel- activated carbon using an adequate base-leaching process is inex-
eration at 250–500 ◦ C for the carbon precursors activated with pensive, simple and well-suited to mass production, especially for
H3 PO4 may be attributed to the formation of phosphate esters via application in sustainable fuel and the adsorption of bulky pollu-
cellulose phosphorylation. These observations indicate that a cross- tant.
linking reaction combines carbon, hydrogen, and oxygen in the
carbon precursors. Simultaneously, cross-linking reactions begin
Acknowledgement
to control the activation of samples rather than decomposition and
depolymerization reactions [46]. This situation increases the arom-
The author expresses thanks to the National Science Council of
atization of the carbonized char and also improves porosity. In the
Taiwan for its financial support under Project No. NSC 94-2214 –
initial period of the second stage reaction, the pores are formed in a
E131 – 003.
short retention time and therefore are mostly microporous. As the
reaction temperature increases, the carbon structure continues to
expand, which leads to an increase in mesoporosity. References
However, as the activation temperature continues to rise, Fig. 11
shows a significant mass loss in the third stage of both H3 PO4 [1] D. Kalderis, D. Koutoulakis, P. Paraskeva, E. Diamadopoulos, E. Otal, J.O. del Valle,
C. Fernández-Pereira, Adsorption of polluting substances on activated carbons
and ZnCl2 activation. The observation can be explained by the prepared from rice husk and sugarcane bagasse, Chem. Eng. J. 144 (2008)
fact that the boiling point of H3 PO4 is 407 ◦ C [46] and the chem- 42–50.
ical transformation of H3 PO4 into P2 O5 occurs in the temperature [2] O. Ioannidou, A. Zabaniotou, Agricultural residues as precursors for activated
carbon production—a review, Renew. Sust. Energ. Rev. 11 (2007) 1966–2005.
range of 580–585 ◦ C [54]. The melting point of ZnCl2 is 283 ◦ C [35] [3] Y.J. Kim, B.J. Lee, H. Suezaki, T. Chino, Y. Abe, T. Yanagiura, K.C. Park, M.
and the vaporization of ZnCl2(l) occurs in the temperature range Endo, Preparation and characterization of bamboo-based activated carbons
of 480–550 ◦ C [55]. Therefore, this mass loss can be attributed to as electrode materials for electric double layer capacitors, Carbon 44 (2006)
1592–1595.
the release of residual activating agents and a few volatiles. Due [4] H. Tamai, U. Nobuaki, H. Yasuda, Preparation of Pd supported mesoporous
to the cross-linking of the phosphate esters in carbon precursor activated carbons and their catalytic activity, Mater. Chem. Phys. 114 (2009)
during H3 PO4 activation process should increase the pyrolysis tem- 10–13.
[5] R.V. Ramanujan, S. Purushotham, M.H. Chia, Processing and characterization of
perature. Thermogravimetric analysis shows that H3 PO4 activation activated carbon coated magnetic particles for biomedical applications, Mater.
requires a longer heating duration than ZnCl2 activation to increase Sci. Eng. C: Bio. Sci. 27 (2007) 659–664.
the thermal degradation temperature. As a result, the severe gasi- [6] M.J. Prauchner, F. Rodríguez-Reinoso, Preparation of granular activated carbons
for adsorption of natural gas, Micropor. Mesopor. Mater. 109 (2008) 581–584.
fication of volatiles for both H3 PO4 and ZnCl2 activation at high
[7] C. Vix-Guterl, E. Frackowiak, K. Jurewicz, M. Friebe, J. Parmentier, F. Béguin,
reaction temperatures cause a sintering reaction or bond cleavage, Electrochemical energy storage in ordered porous carbon materials, Carbon 43
which in turn causes pore shrinkage or collapse. The observations (2005) 1293–1302.
above are consistent with the analysis results of pore characteristics [8] B. Guo, L. Chang, K. Xie, Adsorption of carbon dioxide on activated carbon, J.
Nature Gas Chem. 15 (2006) 223–229.
in Figs. 5 and 6, which indicate that the development of porosity of [9] M.C.R. Halmeman, S.N.M. de Souza, P.T. Oyama, Potential and costs for the pro-
carbon reaches its maximum value at a carbonization temperature duction of electrolytic hydrogen in alcohol and sugar cane plants in the central
of 500 ◦ C. Lua and Yang [43] further pointed out that the decreasing and south regions of Brazil, Int. J. Hydrog. Energy 33 (2008) 6858–6864.
[10] G. Kostka, C. Polzin, J. Scharrer, The future of sugar cane in China and
porosity of carbon activated with ZnCl2 at higher reaction temper- India—supply constraints and expansion potential, Appl. Energy 86 (2009)
atures is due to the sintering of volatiles, which causes the pore S100–S107.
structure to shrink. When the activation is complete, activated car- [11] C. Pickardt, S. Neidhart, C. Griesbach, M. Dube, U. Knauf, D.R. Kammerer, R.
Carle, Optimisation of mild-acidic protein extraction from defatted sunflower
bon with a porous structure can be obtained after the final process (Helianthus annuus L.) meal, Food Hydrocolloids 23 (2009) 1966–1973.
of water- or acid-washing. [12] M. Balat, M. Balat, Recent trends in global production and utilization of bio-
ethanol fuel, Appl. Energy 86 (2009) 2273–2282.
[13] G. Antolín, F.V. Tinaut, Y. Briceño, V. Castaño, C. Pérez, A.I. Ramírez, Optimisation
4. Conclusions
of biodiesel production by sunflower oil transesterification, Bioresour. Technol.
83 (2002) 111–114.
Using sugar cane bagasse and sunflower seed hull as raw mate- [14] K. Nakagawa, S.R. Mukai, K. Tamura, H. Tamon, Mesoporous activated carbons
from phenolic resins, Chem. Eng. Res. Des. 85 (2007) 1331–1337.
rials, this study reports the preparation of activated carbon with
[15] E. Lorenc-Grabowska, G. Gryglewica, Adsorption characteristics of Congo red
a highly mesoporous surface area. Both H3 PO4 and ZnCl2 serve as on coal-based mesoporous activated carbon, Dyes Pigment 74 (2007) 34–40.
chemical activating agents. The starting raw materials, before and [16] H. Tamai, U. Nobuaki, H. Tasuda, Preparation of Pd supported mesoporous
after treatment processes, and conditions employed for carbon syn- activated carbons and their catalytic activity, Mater. Chem. Phys. 114 (2009)
10–13.
thesis, strongly affect the final pore structure and the adsorption [17] W. Shen, J. Zheng, Z. Qin, J. Wang, Preparation of mesoporous carbon from
behavior of the resulting carbon. The proposed base-leaching pro- commercial activated carbon with steam activation in the presence of cerium
cedure can effectively enhance the surface area and pore volume oxide, J. Colloid Interface Sci. 264 (2003) 467–473.
[18] M.J. Lázaro, L. Calvillo, E.G. Bordejé, R. Moliner, R. Juan, C.R. Ruiz, Function-
of samples. The removal of ash from the activated carbon results in alization of ordered mesoporous carbons synthesized with SBA-15 silica as
pore widening and the development of mesoporosity. The largest template, Micropor. Mesopor. Mater. 103 (2007) 158–165.
pore volume and surface area of activated carbon prepared from [19] C.N. Mbileni, F.F. Prinsloo, M.J. Witcomb, N.J. Coville, Synthesis of mesoporous
carbon supports via liquid impregnation of polystyrene onto a MCM-48 silica
SCB and SSH samples using H3 PO4 and ZnCl2 activation processes template, Carbon 44 (2006) 1476–1483.
occur at 500 ◦ C and an impregnation ratio of 2. Simultaneously, the [20] Z. Zhu, A. Li, L. Yan, F. Liu, Q. Zhang, Preparation and characterization of highly
mesopore volume and mesopore surface area are also accompa- mesoporous spherical activated carbons from divinylbenzene-derived polymer
by ZnCl2 activation, J. Colloid Interface Sci. 316 (2007) 628–634.
nied by an increase in the largest amounts. The surface areas of
[21] M. del Mar Gómez-Tamayo, A. Macías-García, M.A. Díez, E.M. Cuerda-Correa,
the activated carbon are as high as 1611 and 2289 m2 /g for SCB, Adsorption of Zn(II) in aqueous solution by activated carbons prepared from
and 1543 and 2240 m2 /g for SSH, for H3 PO4 and ZnCl2 activation, evergreen oak (Quercus rotundifolia L.), J. Hazard. Mater. 153 (2008) 28–36.
[22] Z. Hu, M.P. Srinivasan, Mesoporous high-surface-area activated carbon, Micro-
respectively. The SCB-based carbon shows a slightly higher surface
por. Mesopor. Mater. 43 (2001) 267–275.
area and pore volume than SSH-based carbon. Thermogravimetric [23] L.J. Kennedy, J.J. Vijaya, K. Kayalvizhi, G. Sekaran, Adsorption of phenol from
analysis shows that the thermal decomposition mechanism of SCB- aqueous solutions using mesoporous carbon prepared by two-stage process,
based samples is similar to that of SSH-based samples. However, Chem. Eng. J. 132 (2007) 279–287.
[24] D. Lozano-Castello, M.A. Lillo-Rodenas, D. Cazorla-Amoros, A. Linares-Solano,
SCB-based activation occurs more easily than SSH-based activation. Preparation of activated carbons from Spanish anthracite. I. Activation by KOH,
The reaction mechanism for both H3 PO4 and ZnCl2 activation can Carbon 39 (2001) 741–749.
142 T.-H. Liou / Chemical Engineering Journal 158 (2010) 129–142

[25] S. Yorgun, N. Vural, H. Demiral, Preparation of high-surface area activated car- [41] M.M. Yeganeh, T. Kaghazchi, M. Soleimani, Effect of raw materials on properties
bons from Paulownia wood by ZnCl2 activation, Micropor. Mesopor. Mater. 122 of activated carbons, Chem. Eng. Technol. 29 (2006) 1247–1251.
(2009) 189–194. [42] K. Mohanty, M. Jha, B.C. Meikap, M.N. Biswas, Preparation and characterization
[26] S. Brunauer, P.H. Emmett, E. Teller, Adsorption of gases in multimolecular lay- of activated carbons from Terminalia Arjuna nut with zinc chloride activation
ers, J. Am. Chem. Soc. 60 (1938) 309–319. for the removal of phenol from wastewater, Ind. Eng. Chem. Res. 44 (2005)
[27] M.M. Dubinin, Fundamentals of the theory of adsorption in micropores of car- 4128–4138.
bon adsorbents: characteristics of their adsorption properties and microporous [43] A.C. Lua, T. Yang, Characteristics of activated carbon prepared from pistachio-
structures, Carbon 27 (1989) 457–467. nut shell by zinc chloride activation under nitrogen and vacuum conditions, J.
[28] G.H. Oh, C.R. Park, Preparation and characteristics of rice-straw-based porous Colloid Interface Sci. 290 (2005) 505–513.
carbons with high adsorption capacity, Fuel 81 (2002) 327–336. [44] S.J. Gregg, K.S.W. Sing, Adsorption, Surface Area and Porosity, Second ed., Aca-
[29] T.H. Liou, Kinetics study of thermal decomposition of electronic packaging demic Press, London, 1982.
material, Chem. Eng. J. 98 (2004) 39–51. [45] T.H. Liou, Evolution chemistry and morphology during the carbonization and
[30] T.H. Liou, S.J. Wu, Characteristics of microporous/mesoporous carbons pre- combustion of rice husk, Carbon 42 (2004) 785–794.
pared from rice husk under base- and acid-treated conditions, J. Hazard. Mater. [46] M. Olivares-Marín, C. Fernández-González, A. Macías-García, V. Gómez-
171 (2009) 693–703. Serrano, Porous structure of activated carbon prepared from cherry stones
[31] S. Timur, I.C. Kantarli, E. Ikizoglu, J. Yanik, Preparation of activated carbons from by chemical activation with phosphoric acid, Energy Fuels 21 (2007)
Oreganum stalks by chemical activation, Energy Fuels 20 (2006) 2636–2641. 2942–2949.
[32] D. Kalderis, S. Bethanis, P. Paraskeva, E. Diamadopoulos, Production of acti- [47] T.H. Liou, F.W. Chang, J.J. Lo, Pyrolysis kinetics of acid-leached rice husk, Ind.
vated carbon from bagasse and rice husk by a single-stage chemical activation Eng. Chem. Res. 36 (1997) 568–573.
method at low retention times, Bioresour. Technol. 99 (2008) 6809–6816. [48] L.J. Kennedy, J.J. Vijaya, G. Sekaran, Effect of two-stage process on the prepa-
[33] M.J. Fernández Llorente, J.E. Carrasco García, Comparing methods for predicting ration and characterization of porous carbon composite from rice husk by
the sintering of biomass ash in combustion, Fuel 84 (2005) 1893–1900. phosphoric acid activation, Ind. Eng. Chem. Res. 43 (2004) 1832–1838.
[34] T. Li, T. Wang, Preparation of silica aerogel from rice hull ash by drying at [49] Y. Guo, D.A. Rockstraw, Physicochemical properties of carbons prepared from
atmospheric pressure, Mater. Chem. Phys. 112 (2008) 398–401. pecan shell by phosphoric acid activation, Bioresour. Technol. 98 (2007)
[35] Q. Qian, M. Mochida, H. Tatsumoto, Preparation of activated carbons from 1513–1521.
cattle-manure compost by zinc chloride activation, Bioresour. Technol. 98 [50] M. Rao, A.V. Parwate, A.G. Bhole, Removal of Cr6+ and Ni2+ from aqueous solu-
(2007) 353–360. tion using bagasse and fly ash, Waste Manage. 22 (2002) 821–830.
[36] J.B. Castro, P.R. Bonelli, E.G. Cerrella, A.L. Cukierman, Phosphoric acid activation [51] A. Demirbas, Effect of temperature on pyrolysis products from four nut shells,
of agricultural residues and bagasse from sugar cane: influence of the experi- J. Anal. Appl. Pyrolysis 76 (2006) 285–289.
mental conditions on adsorption characteristics of activated carbons, Ind. Eng. [52] P.T. Williams, A.R. Reed, Development of activated carbon pore structure via
Chem. Res. 39 (2000) 4166–4172. physical and chemical activation of biomass fibre waste, Biomass Bioenergy 30
[37] C.H. Yun, Y.H. Park, G.H. Oh, C.R. Park, Contribution of inorganic components in (2006) 144–152.
precursors to porosity evolution in biomass-based porous carbons, Carbon 41 [53] M. Olivares-Marin, C. Fernandez-Gonzalez, A. Macias-Garcia, V. Gomez-
(2003) 2009–2012. Serrano, Thermal behavior of lignocellulosic material in the presence of
[38] M. Soleimani, T. Kaghazchi, Agricultural waste conversion to activated carbon phosphoric acid. Influence of the acid content in the initial solution, Carbon
by chemical activation with phosphoric acid, Chem. Eng. Technol. 30 (2007) 44 (2006) 2330–2356.
649–654. [54] D. Montané, V. Torné-Fernández, V. Fierro, Activated carbons from lignin:
[39] T.H. Usmani, T.W. Ahmad, A.H.K. Yousufzai, Preparation and liquid-phase char- kinetic modeling of the pyrolysis of Kraft lignin activated with phosphoric acid,
acterization of granular activated-carbon from rice husk, Bioresour. Technol. 48 Chem. Eng. J. 106 (2005) 1–12.
(1994) 31–35. [55] E. Oniyama, P.G. Wahlbeck, Application of transpiration theory to TGA data:
[40] M. Jagtoyen, F. Derbyshire, Activated carbons from yellow poplar and white calcium carbonate and zinc chloride, Thermochim. Acta 250 (1995) 41–53.
oak by H3 PO4 activation, Carbon 36 (1998) 1085–1097.