Effects of micropore development on the physicochemical properties of KOH-activated carbons

(1)

Effects of micropore development on the physicochemical

properties of KOH-activated carbons

Ru-Ling Tseng

a,

*

, Szu-Kung Tseng

b,c

, Feng-Chin Wu

d

,

Chi-Chang Hu

e,f

, Chen-Ching Wang

f

a

Department of Safety, Health and Environmental Engineering, National United University, No. 1 Lien-da, Kung-Ching Li, Miao-Li 360, Taiwan

b

Graduate Institute of Environmental Engineering, National Taiwan University, Taipei 106, Taiwan

c_{Department of Bioenvironmental Engineering, Chung Yuan Christian University, Chung Li 320, Taiwan} d_{Department of Chemical Engineering, National United University, Miao-Li 360, Taiwan} e_{Department of Chemical Engineering, National Tsing Hua University, Hsin-Chu 300, Taiwan} f_{Department of Chemical Engineering, National Chung Cheng University, Chia-Yi 621, Taiwan}

Received 23 August 2007; accepted 21 November 2007

Abstract

Effects of micropore development through varying the KOH/char ratio on the porous, electrochemical, electronic, and adsorptive properties for corncob-derived activated carbons (ACs) prepared by means of the KOH activation method were systematically compared. The pore properties of ACs, including BET surface area, total pore volume, micropore volume ratio, bulk density, and product yield based on the raw material were investigated to gain an understanding for the influence of KOH dosage on the pore development. Element analysis and temperature-programming desorption (TPD) were used to obtain the information of chemical composition and surface oxygen functional groups on ACs in order to propose the reaction mechanism of KOH activation. Based on the pore development, KOH-activated carbons can be classified into two groups: a combination of physical activation and chemical KOH etching at low KOH/char ratios (0.5–2) as well as chemically uniform etching at high KOH/ char ratios (3.0). From the adsorption study for five organics with molecular weights varying from 129 to 466 g/mol, the specific adsorption capacity of ACs for organics is independent of their specific surface area. The specific capacitance of ACs reached a maximum as the KOH/char ratio was equal to 3, attributed to a compromise between the specific surface area and electronic resistance of ACs.

Keywords: Activated carbons; KOH activation mechanism; Specific capacitance; Specific resistance; Adsorption

1. Introduction

Activated carbons (ACs) are important industrial materials for various applications because of their high specific surface area (Sp) (Fujimoto et al., 2000; Hsieh and Teng, 2000; Shi, 1996; Tsai et al., 1997; Tseng et al., 2003). In analyzing the pore properties and determining the specific surface area of ACs, N2adsorption/desorption isotherms at 77 K are the most

popular method (e.g., about 89% of 46 articles using this method to analyze the pore properties of ACs derived from plant raw materials in 2005). Data derived from N2adsorption/

desorption isotherms do not only provide Sp, total pore volume

(Vpore), and pore size distribution but also the micropore volume

(Vmicro) and exterior surface area (Sext) (Tamai et al., 1999; Wang et al., 2005). Moreover, from the shape and hysteresis loop of isotherms, the pore size distribution as well as the shape/ structure of pores inside ACs can be reasonably estimated (Sing et al., 1985; Suzuki, 1990). Accordingly, this measurement provides the basic physical properties of ACs. On the other hand, the performances of ACs in certain fields (e.g., electrochemical energy storage/conversion systems) do not only depend on the pore structure but also on the chemical properties of ACs, such as chemical composition and surface functional groups (Minkova et al., 2000; Suzuki, 1990). Accordingly, the influences of physicochemical properties of ACs on their performances in modern industries are very important (Gergove and Eser, 1996; Guo et al., 2005; Wu et al., 2001, 2005a).

The chemical composition of ACs, in terms of the weight percentage of C, H, and O, is usually measured by means of

www.elsevier.com/locate/jcice

Available online at www.sciencedirect.com

Journal of the Chinese Institute of Chemical Engineers 39 (2008) 37–47

* Corresponding author. Tel.: +886 37 381775; fax: +886 37 333187. E-mail address:trl@nuu.edu.tw(R.-L. Tseng).

(2)

element analyses (Minkova et al., 2000; Oh and Park, 2002). The chemical composition of ACs has been found to depend on the nature of raw materials, activation methods, the dosage of chemical species, and the specific surface area (Minkova et al., 2000; Oh and Park, 2002). However, the relationship between chemical composition and preparation condition as well as the activation mechanism is seldom discussed. Thus, changes in the chemical composition of materials during the activation process cannot be suitably described, which is worthy being investigated.

The surface oxygen functional groups on ACs can be detected by means of infra-red (IR) spectroscopy (Guo and Lua, 1999; Lillo-Rodenas et al., 2003), X-ray photoelectron spectroscopy (XPS) (Qu, 2002; Yoshizawa et al., 2000), and temperature-programming desorption (TPD) (Lillo-Rodenas et al., 2003; Molina-Sabio et al., 1996) while the spectroscopic methods usually provide qualitative information only. Although quantitative information can be obtained from curve fitting of XPS spectra, this technique is an external, superficial analysis tool which cannot obtain the complete distribution of functional groups within ACs. Accordingly, TPD through the evolution of

CO2 and CO in different temperature regions is often

recommended for the analysis of surface oxygen functional groups. For example, the steam-activated carbon was found to release more CO in comparison with the CO2-activated carbon

by Stones (Molina-Sabio et al., 1996). Carbon fiber (CFN) treated with nitric acid at various temperatures was reported to

release more CO2 when the treatment was under a low

temperature. Recently, we also found that KOH-activated carbons release more CO2 than the steam-activated carbons

(Wu et al., 2005b). The above facts indicate the importance of TPD in getting information of surface functional groups in the field of AC research.

Recently, the relationship between the porous structure and electrochemical behavior of ACs is very important because carbons in various forms have been widely used as electrode

materials for electrochemical energy storage/conversion sys-tems, especially for electric double layer capacitors (EDLCs) (Frackowiak and Beguin, 2001; Hsieh and Teng, 2002; Jurewicz et al., 2004; Lufrano and Staiti, 2004; Wei et al., 2005; Wu et al., 2004, 2005b). Although the charge storage of EDLCs is based on the Helmholtz double layers, which is highly reversible, their performances are significantly influ-enced by the physicochemical properties of ACs, such as pore structure, surface functional groups, and chemical composition, etc. The specific capacitance of ACs can even reach above 100 F/g (Qu, 2002; Wu et al., 2004, 2005b) meanwhile the raw materials of ACs are generally cheap and available. Accord-ingly, this type of energy storage devices is commercially attractive.

Chemical activation of chars with KOH is very powerful in developing micropores (Wu et al., 2005c), which is recognized to favor the formation of surface oxygen functional groups (Lua and Guo, 2001). In this work, element analysis, TPD, specific capacitance, and specific resistance of KOH-activated carbons in various KOH/char ratios (Tseng and Tseng, 2005) are used to find the influences of micropore development on the chemical composition, distribution of surface functional groups, specific capacitance, and specific adsorption capacity of ACs. Through a comparison of the above information and the results of pore properties measured previously, the influences of physico-chemical properties of ACs on the performances in adsorption of organics and EDLCs will be clarified.

2. Materials and methods

2.1. Preparation of carbons by KOH activation

Carbons were prepared from chars activated with KOH, which were carbonized from corncob; a typical two-step process. The preparation procedure has been described in our previous work (Tseng and Tseng, 2005). The weight ratio of KOH/char is set at 0.5, 1, 2, 3, 4, and 6 and the resultant ACs are denoted as CobK0.5, CobK1, CobK2, CobK3, CobK4, and CobK6, respectively.

2.2. Measurements of physicochemical properties

The BET surface area (Sp) of ACs was obtained from the N2

adsorption isotherm at 77 K with a sorptiometer (Porous Materials, BET-202A). The measurement steps and data processing were also described previously (Tseng et al.,

2006). Element analysis of ACs was performed using an

elemental analyzer (Elementar Co., model Vario EL III). 2.3. Specific resistance measurement

The carrier template (1.5 cm 1.5 cm) for preparing the test plate samples in the specific resistance measurement was prepared from degreased glass plates. Activated carbon powders were well mixed with 2 wt.% polyvinylidene difluoride (PVdF) binders for 30 min and then N-methy-2-pyrolidinone (NMP) was dropped into the above mixture and Nomenclature

CS average specific capacitance (F/g)

KF Freundich constant (mol/kg)(mol/m3)n

KL Langmuir constant (m3/mol)

qe amount of adsorption at equilibrium (mol/kg)

qmon amount of adsorption corresponding to

mono-layer coverage (mol/kg) RS surface resistivity (V/cm2)

Smicro micropore surface area (m2/g)

Sp total BET surface area (m2/g)

V volume of the solution (m3)

Vmicro micropore volume (cm3/g)

Vpore total pore volume (cm 3

/g)

Yp weight ratio of final carbons to the initial dried

raw materials Greek symbol

(3)

ground for 30 min. The gas bubbles in the slurry were removed by means of an ultrasonic vibrator for 10 min. Then, the slurry was spread on the template with thickness of 1 0.1 mm, which was dried in an oven at 85 8C for 8 h to form the test plates. Four point probe surface resistor meter (Laresta-EP MCP-T360; Mitsubish Chemical Co.) was used to measure the surface resistance of these test plates and the average value of five replications for each sample was reported in this work.

2.4. Electrode preparation

Activated carbon powders were well mixed with 2 wt.% PVdF binders for 30 min and NMP was dropped into the above mixture and ground to form the coating slurry. This slurry was smeared onto the pretreated graphite substrates and dried in a vacuum oven at 50 8C overnight. In order to avoid any unexpected influences, the total amount of AC paste on each electrode was kept approximately constant (ca. 2 mg/cm2). The pretreatment procedure of the

10 mm 10 mm 3 mm graphite supports (Nippon Carbon

EG-NPL, NCK) completely followed our previous work (Wu et al., 2005b). The exposed geometric area of these pretreated graphite supports is equal to 1 cm2 while the other surface areas were insulated with polytetrafluorene ethylene (PTFE) coatings.

2.5. Capacitance measurements

The electrochemical measurements were performed by means of an electrochemical analyzer system, CHI 633A (CH Instruments). All experiments were carried out in a three-compartment cell. An Ag/AgCl electrode (Argenthal, 3 M KCl, 0.207 V vs. SHE at 25 8C) was used as the reference and a

platinum wire with an exposed area equal to 4 cm2 was

employed as the counter electrode. A Luggin capillary was used to minimize errors due to iR drop in the electrolytes. The electrolytes used for the electrochemical measurement were degassed with purified nitrogen gas for 25 min before measurements and this nitrogen was passed over the solutions during the measurements. The solution temperature was maintained at 25 8C by means of a water thermostat (Haake DC3 and K20).

2.6. Procedure for adsorption experiments

Five solutes, acid blue 74 (AB74), basic brown 1 (BB1), methylene blue (MB), 2,di chloro-phenol (2,DCP), and 4-chlorophenol (4-CP) with their molecular weight of 466.4, 419.4, 284.3, 163.0, and 128.6 g/mol, respectively, are from Merck. The molecular structures of MB, BB1, and AB74, and the characteristics of chlorinated phenols and dyes have been detailed previously (Wu et al., 2005c). The steps for the adsorption equilibrium experiments completely follow the procedure expressed in our previous study (Tseng et al., 2003).

3. Results and discussion

3.1. Physical properties of activated carbon

The specific (BET) surface area (Sp), total pore volume

(Vpore), and micropore pore volume ratio (Vmicro/Vpore) of ACs

were estimated from the adsorption/desorption isotherms of N2

at 77 K in this study. Fig. 1(a)–(f), respectively, show the dependence of Sp, Vpore, Vmicro/Vpore, yield (Yp), bulk density

(rb), and (Yp Sp) on the KOH/char ratio. There exists a

discontinuity in these properties when the KOH/char ratio is changed from 2 to 3. Based on these phenomena, ACs are qualitatively divided into two groups: group I ACs with the KOH/char ratio varying from 0.5 to 2 and group II ACs with the KOH/char ratio from 3 to 6 (Tseng and Tseng, 2005). We believed that chars were activated by a combination of physical activated and chemical KOH etching to form group I ACs while they were homogeneously etched with KOH resulting in the development of micropores for group II ACs. The above statement is supported by the SEM photographs observed previously; i.e., a twisted outer morphology for group I ACs but smooth, flat surface for group II ACs (Tseng and Tseng, 2005). In Fig. 1(a), the specific surface area rises abruptly when the KOH/char is changed from 2 to 3 but Vporemonotonously and

gradually increases in Fig. 1(b) although the slope for Vpore

against KOH/char is started to change at the same range of KOH/char. The above phenomena are attributable to the fact that micropore development is the main reaction in the chemical activation process of chars as the KOH/char ratios are 3. This statement is supported by the discontinuity of Vmicro/

Vporein the same KOH/char range inFig. 1(c). Moreover, from

the two straight lines in Fig. 1(c), two distinct mechanisms should involve in the activation process, resulting in the presence of groups I and II ACs.Fig. 1(d) shows that the yield of ACs at the KOH/char ratio of 0.5 is lower than that as the KOH/char ratio is equal to 1. This is probably due to the competition of the physical surface pyrolysis and chemical KOH etching when the KOH/char ratio is between 0.5 and 2. Note that a higher coverage of KOH onto the chars (i.e., high KOH/char ratio) will inhibit the physical surface pyrolysis meanwhile the chemical process of KOH etching will be promoted. Thus, yield reaches the highest value at the KOH/ char ratio = 1.Fig. 1(e) shows that the bulk density of group I

ACs monotonously decrease from 266 to 135 g/m3 with

increasing the KOH/char ratio while no significant variation is found for group II ACs (between 94 and 112 g/m3). The lower bulk density of group II ACs is attributed to its pure KOH etching process which generates many micropores within the AC cake bread, observed previously in the SEM photographs (Tseng and Tseng, 2005).

It is important and meaningful to gain the information of the yield of BET surface area per unit mass of raw material in the AC researches because it provides the information for the economical assessment of AC preparation and the under-standing on the pore variation during the activation process. Fig. 1(f) shows the yield of BET surface area per unit mass of raw material, expressed as Yp Sp. In general, curves

(4)

corresponding to groups I and II ACs are parabolic. When there is no dosage of KOH (KOH/char ratio = 0), chars are completely activated through the physical pyrolysis process, and the yield of BET surface area per unit mass of raw material is only 46 m2/g. When the KOH/char ratios equal 1 and 2, on the other hand, values of (Yp Sp) are 239 and 238 m2/g,

respectively, which are larger than that of the purely physical pyrolysis AC and that of the steam-activated carbons derived from corncob (between 98.1 and 102 m2/g) (Wu et al., 2001) or pinewood (between 67 and 103 m2/g) (Tseng et al., 2003). The above enhancement in (Yp Sp) is reasonably attributed to the

simultaneous occurrence of the physical activation and chemical KOH etching for group I ACs. Moreover, it also suggests that (Yp Sp) of a chemical activation process is much

higher than those of a physical activation process (e.g., steam

activation). This reveals that chemical etching (by KOH) is much more powerful for micropore development in the interior of chars in comparison with the physical (steam) activation. This statement is supported by the much higher value of (Yp Sp) for the pure KOH-activated carbons (i.e., group II

ACs). In addition, the yield of BET surface area per unit mass of raw material reaches the highest value of 435 m2/g as the KOH/ char ratio equals 4. When the ratio is above 4, an over-etching phenomenon is found and harmful to the generation of surface area of ACs.

3.2. Chemical composition of activated carbon

The results of element analysis are listed inTable 1where corncob, cob-char, and CobK0 are respectively indicative of

Fig. 1. (a) BET surface area (Sp), (b) total pore volume (Vpore), (c) micropore pore volume ratio (Vmicro/Vpore), (d) yield (Yp), (e) bulk density (rb), and (f) yield of BET

(5)

raw material dried at 130 8C, carbonized char at 450 8C, and the carbon with physical activation at 780 8C with no KOH addition. The C content is increased from 46 to 87% with the carbonization of raw material and physical surface pyrolysis of chars but an opposite result is found for the contents of H and O, decreased respectively from 6.9 to 2.5% and from 44 to 7.8%. Actually, the composition of five wooden fiber materials (Minkova et al., 2000), with their C content of 47.0–49.5%, H content of 4.6–6.2%, and O content of 43.7– 46.7%, is close to that of corncob used in this work. For ACs derived from rice straw with pyrolysis in the oxygen-exhausted condition (Oh and Park, 2002), the contents of both H and O significantly decreased while the C content increased. The above studies and our results shown in this work reveal that carbonization of plant raw materials and physical pyrolysis of chars will cause the significant loss of H and O (probably in the forms of H2O vaporization and/or CO2/

CO evolution). This effect increasing the C content is most likely due to the limited supply of oxygen.

There are few articles describing the element variations of ACs before and after the chemical activation. The contents of C, H, and O for ACs prepared from corncob and chemically activated with ZnCl2(Tsai et al., 1998) were reported to be

90.31, 0.81, and 8.85%, respectively. In addition, the contents of C, H, and O for ACs prepared from palm shells and soaked with H2SO4(Guo et al., 2005) were found to be 83.3, 0.4, and

16.1%, respectively. The results from the above two examples indicates that the content of C is increased and both H and O are decreased when raw materials were chemically activated to form ACs. This trend is similar to that of ACs with the physical activation treatment. However, a reverse trend is found in this study; the C content is decreased from 80 to 56% and that of H and O is respectively increased from 2.9 to 6.3% and from 15 to 36% when the KOH/char ratio is increased from 0 to 6. The above result suggests that the C elements in chars will react with KOH to form oxygen-enriched surface functional groups during the chemical activation process.

For convenient evaluation of the above property, the element residue ratio (dXi) is defined as the ratio of weight

residue of a certain element in char during activation to the weight content of that element in char (Xi,char), which can be

calculated as

dXi¼

Xi Yi;p

Xi;char

(1)

where Xiindicates the mass of element i (i.e., C, H, or O) per

unit mass of AC. Xi,char is the mass of the corresponding

element per unit mass of char. Xi Yi,p is indicative of the

residual mass of the element i per unit mass of char. Based on the above definitions, di will be equal to 1 if no loss in mass

during the activation process is found for the element under study. A low value of dXiindicates a more loss of the element,

implying its higher reactivity to KOH.

Fig. 2shows the dependence of dC, dH, and dO on the KOH/ char ratio. When the KOH/char ratio is increased from 0 to 1, the coverage of KOH onto the chars is increased, which gradually inhibits the physical activation on the surface of chars. Thus, the residue ratios of H and O are obviously increased from 0.24 to 0.53 and from 0.18 to 0.72, respectively. On the other hand, as the KOH/char ratio is increased from 0 to 0.5, the residual ratio of C is also increased obviously from 0.49 to 0.64 while a decrease in this property is found when the KOH/char ratio is above 0.5. This phenomenon is considered as a sign for the significant contribution of another activation process, i.e., the chemical etching of chars by KOH. This chemical etching is considered to mainly consume C elements in chars and a decrease in the residual ratio of C is found. Note the relatively high dosage of KOH as the KOH/char ratio ranges from 3 to 6 while the residual ratios of both H and O are nearly constant (from 0.58 to 0.65 and from 0.82 to 0.87, respectively). This phenomenon suggests that the consumption (and reaction mechanism) of H and O during the KOH activation are approximately constant when the KOH/char ratio is between 3 and 6. However, the residual ratio of C significantly drops from 0.47 to 0.34 in the same region of KOH/char ratio. The above results suggest that a chemical reaction between KOH and C elements in the char favor the development of micropores. Investigate elements H, O, and C. In the range of KOH/char ratios from 0 to 1, H and O increased with increased KOH consumption. This showed that increased KOH coverage could decrease physical activation; and in the range of KOH/char

Table 1

Chemical composition of corncob, char, and activated carbons

Materials Elemental composition (wt.%) H/C O/C

C H N S Odiff Corncob 46.33 6.87 2.49 0.246 44.06 0.150 0.951 Cob-char 74.19 4.38 2.26 0.438 18.73 0.059 0.252 CobK0 87.09 2.47 2.46 0.176 7.80 0.028 0.090 CobK0.5 80.16 2.87 1.57 0.146 15.26 0.036 0.190 CobK1 73.23 3.67 1.45 0.081 21.57 0.050 0.295 CobK2 72.34 3.97 0.97 0.083 22.63 0.055 0.313 CobK3 64.17 5.25 1.12 0.136 29.33 0.082 0.457 CobK4 63.20 5.03 1.34 0.059 30.38 0.080 0.481 CobK6 55.79 6.31 1.15 0.281 36.47 0.110 0.653 Odiff: the oxygen is assessed by difference.

Fig. 2. Element residue ratios of H, O, and C (dH, dO, and dC) of corncob-derived, KOH-activated carbons.

(6)

ratios from 3 to 6, H and O were almost constant. It meant that sufficient KOH coverage inhibited physical activation. 3.3. Activation mechanism

Cf–KOH reaction is described as (Yang and Lua, 2003):

2KOH! K2O þ H2OðdehydrationÞ (2)

Cfþ H2O! H2þ CO ðwatergas reactionÞ (3)

COþ H2O! H2þ CO2ðwatergas shift reactionÞ (4)

K2O þ CO2 ! K2CO3ðcarbonate formationÞ (5)

It is believed that after a serial of reactions during KOH activation, the last product is K2CO3.

Metal potassium (K) is produced during K2O activation at

temperature exceeding 700 8C. The possible reactions might be as follow:

K2O þ H2 ! 2K þ H2Oðreduction by hydrogenÞ (6)

K2O þ Cf ! 2K þ CO ðreduction by carbonÞ (7)

To K2O, Eqs. (5)–(7) are parallel reaction mechanisms. That

which reaction mechanism is the main reaction is not easy to be determined. Because the activation temperature was 780 8C, the boiling point of potassium, in the study, if the reaction were according to Eqs.(6) and (7), the product element potassium would lose in large amount through vent and exhaust of reaction furnace. From our experimental observations, up to 83% of potassium could be recovered. That is 17% of potassium was lost and 83% of it remained in the final products. This revealed that the main reaction mechanism was Eqs.(2)–(5)in this work. Furthermore, from element residue ratio of H, O, and C, it was found that C loss was much higher the other two elements. It can be inferred that reaction between KOH and C is the main reaction. Thus this paper neglects description on reaction mechanism of H and O.

3.4. Surface functional groups of activated carbons

The main surface functional groups present in the char were reported to be carbonyl groups (e.g. ketone and quinone) and aromatic rings (Arriagada et al., 1997; Gome-Serrano et al., 1996; Lua and Guo, 2001) from the findings of chars derived from peach stone (Arriagada et al., 1997) roockose ( Gome-Serrano et al., 1996), and oil-palm stones (Lua and Guo, 2001). For ACs activated at high temperatures (e.g., 900 8C) and long activation time (e.g., 60 min) with CO2, only aromatic rings

remained (Lua and Guo, 2001) since most oxygen-containing functional groups will be desorbed. Due to the fact that significant amount of H and O elements have been found in the KOH-activated carbons in the element analysis results, their surface functional groups should be more fruitful than those described in previous studies. Moreover, the distribution of surface functional groups is usually dependent on the preparation variables for the chemically activated carbons. This property of our KOH-activated carbons will significantly

influence their performances in several applications (Sun et al., 1996; Wu et al., 2005b). Based on this reason, the surface functional groups of ACs are semi-quantitatively analyzed by means of temperature-programming desorption (TPD) in this work (Cheng and Teng, 2003). Typical evolution profiles of

CO2and CO determined by the TPD method for CobK1 and

CobK4 are shown inFig. 3.

It is well known that CO2mainly desorbs under relatively

low temperatures (i.e., <550 8C), which has been attributed to the presence of anhydrides, lactones, and carboxyl groups (Molina-Sabio et al., 1996; Qu, 2002; Yoshizawa et al., 2000). On the other hand, desorption of CO species, attributable to quinone, hydroxyl, and carbonyl groups, occurs predominately at relatively higher temperatures (ca. above 500 8C) (Cheng and Teng, 2003; Kinoshita, 1988; Otake and Jenkins, 1993). Hence, the distribution of oxygen functional groups can be roughly estimated from the TPD results. A comparison of the curves in Fig. 3reveals several features. First, the relative intensities of CO2and CO evolution are increased with increasing the KOH/

char ratio. Accordingly, the density of all surface functional groups on CobK4 should be higher than that of CobK1. Second, the increase in the area under the curve corresponding to the CO desorption is more obvious than that that corresponding to the CO2evolution when the KOH/char ration is changed from 1 to

4. This suggests that CobK4 should be enriched with the quinone, hydroxyl, and/or carbonyl groups which are electro-chemically active. The above results reveal that ACs activated with KOH at a higher KOH/char ratio reserve more fruitful functional groups, especially for the functional groups evolving CO. On the other hand, due to the fact that activation of chars at a low KOH/char ratio has to suffer the physical activation, certain amount of the surface functional groups will lose during the activation process, somewhat similar to the results of physically activated carbons (Lua and Guo, 2001).

3.5. Equilibrium adsorption isotherms for phenols and dyes

Adsorption of organics is studied in this work because ACs with high specific surface area is appropriate for the adsorption of solutes in liquid phase (Tseng and Tseng, 2005). In this work,

Fig. 3. Evolution profiles of CO2 and CO by temperature-programming

(7)

adsorption isotherms of dyes (MB, BB1, and AB74) and chlorinated phenols (4-CP and 2,4-DCP) on corncob-derived ACs at 30 8C are measured to fit suitable theoretical and empirical equations. Tables 2 and 3list the fitting values of parameters for the Langmuir and Freundlich isotherm equations, respectively, for the adsorption of the above five organics in aqueous solution. The parameter symbols and data processing of Langmuir and Freundlich isotherm equations follow the same way of our previous study (Wu et al., 2005d). In order to quantitatively compare the validity of the two isotherm equations under study, a normalized standard deviation, Dqe(in %), is defined as follow:

Dqeð%Þ ¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi P_½ðq

e;exp qe;calÞ=qe;exp 2

N 1

s

100 (8)

Table 2shows that the normalized standard deviation, Dqe, of

Langmuir equation is larger (from 1.74 to 9.09%) than that of Freundlich equation (from 0.65 to 4.90%). Table 3 indicates that the results of phenol adsorption is better described by the Langmuir equation with Dqefrom 0.69 to 3.96% in comparison

with the Freundlich equation with Dqe from 1.03 to 12.5%.

Although the empirical Freundlich equation more suitably describes the adsorption data of dyes with larger molecular weights, for studying the relationship between the adsorption capacity and KOH/char ratio, the physically meaningful Lang-muir parameter, qmon, is employed for a comparison purpose.

Adsorption capacity of ACs was found to depend on the features of (a) the physicochemical properties of adsorbents (e.g., surface area, pore size distribution, and functional groups), (b) the physicochemical properties of adsorbates (e.g., molecular weight and size, solubility, polarity, hydrophobic property, functional groups, and pKafor weak acids or bases),

and (c) solution properties (e.g., pH, adsorbate concentration, temperature, presence of competitive solutes, polarity of solvent, etc.). In this work, a parameter, qmon/Sp, was used to

evaluate the adsorption capacity per unit surface area (i.e., specific adsorption capacity) of ACs activated in different dosages of KOH for various organics. Based on this definition, the influence of feature (a), the physicochemical properties of ACs, on the specific adsorption capacity for all adsorbates can be systematically investigated since features (b) and (c) are considered to be unchanged.

Fig. 4shows the dependence of qmon/Spfor dyes (MB, BB1,

and AB74) and chlorinated phenols (4-CP and 2,4-DCP) on the KOH/char ratio. The qmon/Spvalues for the AC prepared at the

KOH/char ratio = 0.5 are very different from the others, probably due to the fact that activation of chars at the KOH/ char = 0.5 is mainly contributed by the physical pyrolysis process. For the ACs with their KOH/char ratios from 1 to 6, on the other hand, the qmon/Spvalues are approximately the same

Table 2

Fitting results for the parameters of Langmuir and Freundlich equations for the adsorption of MB, BB1, and AB74 solutes on corncob-derived ACs at 30 8C

Solute Adsorbent Langmuir Freundich

qmon(mol/kg) Dqe(%) 1/n KF(mol/kg) (mol/m 3 )n Dqe(%) MB CobK0.5 0.44 2.05 0.012 0.44 0.65 CobK1 1.14 2.75 0.025 1.11 0.81 CobK2 1.47 1.74 0.023 1.43 1.63 CobK3 2.34 1.98 0.045 2.27 1.50 CobK4 2.49 4.37 0.049 2.42 2.45 CobK6 3.56 5.04 0.078 3.67 4.60 BB1 CobK0.5 0.93 5.93 0.092 0.86 2.54 CobK1 1.80 4.26 0.149 1.60 1.53 CobK2 1.96 3.60 0.131 1.86 2.49 CobK3 2.81 6.61 0.227 2.69 3.63 CobK4 3.16 4.86 0.285 3.10 3.80 CobK6 4.08 6.58 0.273 4.39 3.49 AB74 CobK0.5 0.25 6.61 0.239 0.22 4.68 CobK1 0.43 2.36 0.192 0.42 2.24 CobK2 0.53 5.47 0.176 0.53 3.12 CobK3 1.19 2.91 0.146 1.31 1.70 CobK4 1.51 6.46 0.186 1.01 4.64 CobK6 1.86 9.09 0.200 2.52 4.90 Table 3

Fitting results for the parameters of Langmuir and Freundlich equations for the adsorption of 4-CP and 2,4-DCP solutes on corncob-derived ACs at 30 8C

Solute Adsorbent Langmuir Freundlich qmon(mol/kg) KL(m 3 /mol) Dqe(%) Dqe(%) 4-CP CobK0.5 1.61 26.3 1.43 7.04 CobK1 2.26 9.54 1.52 1.90 CobK2 2.61 35.8 1.97 7.48 CobK3 4.01 11.1 3.96 12.5 CobK4 4.69 5.32 2.09 7.92 CobK6 5.87 7.15 3.73 8.43 2,4-DCP CobK0.5 2.27 8.39 1.65 2.34 CobK1 3.42 6.46 2.23 3.73 CobK2 3.66 25.5 0.69 1.03 CobK3 5.95 28.1 1.09 4.25 CobK4 7.84 140 1.13 3.77 CobK6 8.47 76.6 1.62 3.52

(8)

for any specified organic. This indicates that all ACs exhibit very similar specific adsorption capacity for the above five organics and that the specific adsorption capacity for any one of the five organics is independent of the physicochemical properties of ACs although ACs prepared at different KOH/ char ratios generally show very different physicochemical properties. The above phenomenon is attributed to the formation of uniform micropores in the ACs chemically activated by KOH, which is approximately independent of the dosage of KOH when the KOH/char ratio is above a critical value (e.g., 1.0 in this study). In our previous study (Wu et al., 2005d), ACs derived from firwoods with steam activation showed the BET surface area ranging from 527 to 1131 m2/g. The qmon/Sp values were approximately constant for the

adsorption of MB while a loss in the qmon/Spvalue up to 32%

with increasing the BET surface area for the adsorption of 4-CP was found. This different phenomenon was attributed to the decrease in the proportion of micropores for the steam-activated carbons with increasing the BET surface area, which is harmful to the adsorption of small molecular substances (Wu et al., 2005d).

3.6. Capacitive characteristics of activated carbons

The electrochemical behavior of all KOH-activated carbons was measured to demonstrate the dependence of capacitive performances on the physicochemical properties of ACs. Group I ACs (with the KOH/char ratios of 0, 0.5, 1, and 2) measured in 1 M NaNO3at 25 mV/s are shown inFig. 5(a). Note that all

curves in the whole potential region are capacitive-like, indicating that capacitance mainly comes from the double-layer charge/discharge process. In addition, all i–E curves on the positive sweeps are symmetric to those on their corresponding negative sweeps although an irreversible oxidation is found at potentials positive than ca. 0.8 V. The former results indicate that all ACs exhibit the excellent capacitive property in NaNO3. The latter result is attributable to

the oxygen evolution reaction. Note that voltammetric currents are gradually increased with increasing the KOH/char ratio from 0 to 2, due to the increase in the specific surface of ACs

with increasing the KOH/char ratio (seeFig. 1(a)). Also note the low ESR of these ACs since the voltammetric currents rapidly reach the plateau values on the positive sweeps of all curves when the direction of potential sweeps is changed from negative to positive. A low ESR should come from a combination of a good electronic conductivity of electrode materials (i.e., ACs) and a low ionic resistance of the electrolyte within the pores of ACs during the charge/discharge tests. From all the above results and discussion, KOH-activated carbons from corncob with the KOH/char ratio varying from 0.5 to 2 show the promising potential to supercapacitors.

The CV responses of group II ACs measured in 1 M NaNO3

at 25 mV/s are shown in Fig. 5(b), which are very similar in shape to the i–E curves in Fig. 5(a). As found previously, a higher specific surface area, a larger pore volume, and a higher density of surface function groups were obtained when the KOH/char ratio was increased. This result should promote the capacitive current density of ACs with increasing the KOH/char ratio from 3 to 6 but an opposite result is obtained. Reasons responsible for this conflict will be clarified in the next section. The influence of KOH dosage during the activation of chars on the capacitive current density of resultant ACs measured in 0.5 M H2SO4(see Fig. 6) is very similar to that observed in Fig. 5. On the other hand, the capacitive current densities obtained in H2SO4 are generally higher than that of the

corresponding AC measured in the neutral electrolyte (Fig. 5).

Fig. 5. Cyclic voltammograms measured at 25 mV/s in 1.0 M NaNO3for (a)

CobK0, CobK0.5, CobK1, and CobK2; (b) CobK3, CobK4, and CobK6. Fig. 4. The amount of adsorption corresponding to monolayer coverage of MB,

BB1, and AB74 and 4-CP, and 2,4-DCP on corncob-derived, KOH-activated carbons.

(9)

This phenomenon has been found in our previous work (Wu et al., 2004, 2005b, 2006), which was found to be independent of the activation methods. The above difference in voltam-metric currents is attributable to (i) the better conductivity of acidic electrolytes due to the proton hopping mechanism (Wu et al., 2004); (ii) the increase in the electrochemically accessible surface areas of these highly porous carbons because of the excellent mobility of protons in the acidic medium; (iii) the redox reactions of surface functional groups involving the exchange of electrons and protons. Recently, effects of pH on the specific capacitance of carbon cloth electrodes have been systematically investigated by Anderas and Conway (2006). A 30% loss in specific capacitance has been attributed to the disappearance of the redox transition of surface quinone groups when pH of the electrolytes is varied from 0 to 11, further supporting our statements.

The effect of KOH dosage in the char during the chemical activation on the specific capacitance (CS) of ACs is clearly

shown inFig. 7. In this work, the specific capacitance of ACs is estimated from the CV curves according to the following equation (Hu and Wang, 2004):

CS¼

q

DE w (3)

where CS, q*, DE, and w indicate the specific capacitance,

voltammetric charges integrated from the negative sweeps, potential window of CV, and mass of ACs, respectively. Note

that CSreaches a maximum value when the KOH/char ratio is

equal to 3 and that the highest values of CS in NaNO3 and

H2SO4are equal to 106 and 127 F/g, respectively. Actually,

ACs may be over etched by the formation of too many micro-pores, which will reduce their electronic conductivity, probably resulting in the lower specific capacitance. Hence, the specific resistance (RS) of all ACs was measured and represented in Fig. 7. In this figure, the specific resistance of ACs is approxi-mately independent of the KOH dosage when the KOH/char ratio is equal to/below 2. On the other hand, as the KOH/char ratio is increased from 2 to 6, RS rises very obviously. The

increase in RSis attributable to the breakage of carbon/graphite

chains due to an over-etching of ACs as well as the simulta-neous formation of surface oxygen functional groups when the KOH/char ratios are above 2. Based on the above results and discussion, CSwill increase with the continuous increase in Sp

for group I ACs because their specific resistance is approxi-mately constant. For group II ACs, a compromise between increasing Spand decreasing RSresults in the maximum CSfor

the AC with the KOH/char ratio equal to 3 and CS will

monotonously decrease with the continuous development of micropores.

Based on the above findings, ACs with low RSis ideal for the

supercapacitor application. Note that CSof group II ACs with

the KOH/char ratio of 3, 4, and 6 was expected to be 148, 188, and 269 F/g in 1.0 M NaNO3if their RSwas kept as low as that

of group I ACs. Although the problem for the high specific resistance of ACs is not solved in this work, the above information is helpful in screening the raw materials and activation methods for searching ACs with high specific surface area and low specific resistance for the application of supercapacitors.

4. Conclusions

According to the discontinuity in several physical properties of ACs chemically activated at various KOH/char ratios, two competitive reaction paths for the pore development are proposed to dominate the activation process: physical activated and etching of carbon with KOH. Very low and high yields of

Fig. 6. Cyclic voltammograms measured at 25 mV/s in 0.5 M H2SO4for (a)

CobK0, CobK0.5, CobK1, and CobK2; (b) CobK3, CobK4, and CobK6.

Fig. 7. The dependence of specific capacitance (CS) measured in NaNO3and

H2SO4, and specific resistance (RS) of corncob-derived, KOH-activated carbons

(10)

BET surface area per unit mass of raw material (in terms of Sp Yp), 46 and 435 m2/g, are found for the former and latter

paths, respectively. From the element residue ratio of C, H, and O, physical activation causes a large drop in the residue ratio of H and O while KOH etching in chars shows high O and H residual ratios, attributable to the main consumption of C in this activation method. From TPD, KOH etching promotes the formation of surface oxygen functional groups, especially for quinine, hydroxyl, and carbonyl groups. The specific resistance of ACs starts to rise with increasing the dosage of KOH at the KOH/char ratio of 2, attributable to the over-etching of chars and the formation of surface oxygen functional groups. For the specific capacitance, the maximum CSof the AC with the KOH/

char ratio equal to 3 results from a compromise between increasing Spand decreasing RS. However, this property does

not affect the specific adsorption capacity of ACs for all organics studied in this work.

Acknowledgement

Financial support of this work by the National Science Council of the Republic of China under contract no. NSC 95-2221-E-239-022 is gratefully acknowledged.

References

Anderas, H. A. and B. E. Conway, ‘‘Examination of the Double-Layer Capacitance of an High Specific-Area C-Cloth Electrode as Titrated from Acidic to Alkaline pHs,’’ Electrochim. Acta, 51, 6510 (2006).

Arriagada, R., R. Garcia, M. Molina-Sabio, and F. Rodriguez-Reinoso, ‘‘Effect of Steam Activation on the Porosity and Chemical Nature of Activated Carbons from Eucalyptus Globulus and Peach Stones,’’ Micropor. Mater., 8, 123 (1997).

Cheng, P. Z. and H. Teng, ‘‘Electrochemical Responses from Surface Oxides Present on HNO3Treated Carbons,’’ Carbon, 41, 2057 (2003).

Frackowiak, E. and F. Beguin, ‘‘Carbon Materials for the Electrochemical Storage of Energy in Capacitors,’’ Carbon, 39, 937 (2001).

Fujimoto, H., A. Mabuchi, K. Tokumitsu, and T. Kasuh, ‘‘Relationship between the Charge Capacity of a Turbostratic Anode for a Li Secondary Battery and its Structure,’’ Carbon, 38, 871 (2000).

Gergove, K. and S. Eser, ‘‘Effects of Activation Method on the Pore Structure of Activated Carbons from Apricot Stones,’’ Carbon, 34, 879 (1996). Gome-Serrano, V., J. Pastor-Villegas, C. J. Duran-Valle, and C.

Valeneula-Calahorro, ‘‘Heat Treatment of Rockrose Char in Air. Effect on Surface Chemistry and Porous Texture,’’ Carbon, 34, 533 (1996).

Guo, J. and A. C. Lua, ‘‘Textural and Chemical Characterizations of Activated Carbon Prepared from Oil-Palm Stone with H2SO4 and KOH

Impregnation,’’ Micropor. Mesopor. Mater., 32, 111 (1999).

Guo, J., W. S. Xu, Y. L. Chen, and A. Lua, ‘‘Adsorption of NH3onto Activated

Carbon Prepared from Palm Shells Impregnated with H2SO4,’’ J. Colloid

Interface Sci., 281, 285 (2005).

Hsieh, C. T. and H. Teng, ‘‘Influence of Mesopore Volume and Adsorbate Size on Adsorption Capacities of Activated Carbons in Aqueous Solutions,’’ Carbon, 38, 863 (2000).

Hsieh, C. T. and H. Teng, ‘‘Influence of Oxygen Treatment on Electric Double-Layer Capacitance of Activated Carbon Fabrics,’’ Carbon, 40, 667 (2002). Hu, C. C. and C. C. Wang, ‘‘Effects of Electrolytes and Electrochemicalical Pretreatments on the Capacitive Characteristics of Activated Carbon Fabrics for Supercapacitors,’’ J. Power Sources, 125, 299 (2004).

Jurewicz, K., K. Babel, A. Ziolkowski, and H. Wachowska, ‘‘Capacitance Behaviour of the Ammoxidised Coal,’’ J. Phys. Chem. Solids, 65, 269 (2004).

Kinoshita, K., Carbon, Electrochemical and Physicochemical Properties, John-Wiley and Sons, Inc., New York, U.S.A. (1988).

Lillo-Rodenas, M. A., D. Cazorla-Amoros, and A. Linares-Solano, ‘‘Under-standing Chemical Reactions between Carbons and NaOH and KOH an Insight into the Chemical Activation Mechanism,’’ Carbon, 41, 267 (2003). Lua, A. C. and J. Guo, ‘‘Preparation and Characterization of Activated Carbons from Oil-Palm Stones for Gas-Phase Aadsorption,’’ Colloids Surf. A, 179, 151 (2001).

Lufrano, F. and P. Staiti, ‘‘Performance Improvement of Nafio Based Solid State Electrochemical Supercapacitor,’’ Electrochim. Acta, 49, 2683 (2004). Minkova, V., S. P. Marinov, R. Zanzi, E. Bjornbom, T. Budinova, M. Stefanova,

and L. Lakov, ‘‘Thermochemical Treatment of Biomass in a Flow of Steam or in a Mixture of Steam and Carbon Dioxide,’’ Fuel Process. Technol., 62, 45 (2000).

Molina-Sabio, M., F. Rodriguez-Reinoso, F. Caturla, and M. J. Selles, ‘‘Devel-opment of Porosity in Combined Phosphoric Acid–Carbon Dioxide Activation,’’ Carbon, 34, 457 (1996).

Oh, G. H. and C. R. Park, ‘‘Preparation and Characteristics of Rice-Straw-Based Porous Carbons with High Adsorption Capacity,’’ Fuel, 81, 327 (2002). Otake, Y. and R. G. Jenkins, ‘‘Characterization of Oxygen-Containing Surface

Complexes Created on a Microporous Carbon by Air and Nitric Acid Treatment,’’ Carbon, 31, 109 (1993).

Qu, D., ‘‘Studies of Activated Carbons Used in Double-Layer Supercapacitors,’’ J. Power Sources, 109, 403 (2002).

Shi, H., ‘‘Activated Carbons and Double Layer Capacitance,’’ Electrochem. Acta, 41, 1633 (1996).

Sing, K. S. W., D. H. Everett, R. A. W. Haul, L. Moscou, R. A. Pierotti, and J. Rouquerol, ‘‘Reporting Physisorption Data for Gas/Solid Systems with Special Reference to the Determination of Surface Area and Porosity,’’ Pure Appl. Chem., 57, 603 (1985).

Sun, J., M. J. Rood, M. Rostam-Abadi, and A. A. Lizzio, ‘‘Natural Gas Storage with Activated Carbon from a Bituminous Coal,’’ Gas Sep. Purif., 10, 91 (1996).

Suzuki, M., Adsorption Engineering, Kodansha, Tokyo (1990).

Tamai, H., T. Yoshida, M. Sasaki, and H. Yasuda, ‘‘Dye Adsorption on Mesoporous Activated Carbon Fiber Obtained from Pitch Containing Yttrium Complex,’’ Carbon, 37, 983 (1999).

Tsai, W. T., C. Y. Chang, and S. L. Lee, ‘‘Preparation and Characterization of Activated Carbons from Corncob,’’ Carbon, 35, 1198 (1997).

Tsai, W. T., C. Y. Chang, and S. L. Lee, ‘‘A Low Cost Adsorbents from Agricultural Waste Corn Cob by Zinc Chloride Activation,’’ Bioresour. Technol., 64, 211 (1998).

Tseng, R. L., F. C. Wu, and R. S. Juang, ‘‘Liquid-Phase Adsorption of Dyes and Phenols Using Pinewood-Based Activated Carbons,’’ Carbon, 41, 487 (2003).

Tseng, R. L. and S. K. Tseng, ‘‘Pore Structure and Adsorption Performance of the KOH-Activated Carbons Prepared from Corcob,’’ J. Colloid Interface Sci., 287, 428 (2005).

Tseng, R. L., S. K. Tseng, and F. C. Wu, ‘‘Preparation of High Surface Area Carbons from Corncob Using KOH Combined with CO2Gasification for the

Adsorption of Dyes and Phenols from Water,’’ Colloids Surf. A, 279, 69 (2006).

Wang, S., Z. H. Zhu, A. Coomes, F. Haghseresht, and G. Q. Lu, ‘‘The Physical and Surface Chemical Characteristics of Activated Carbons and the Adsorp-tion of Methylene Blue from Wastewater,’’ J. Colloid Interface Sci., 284, 440 (2005).

Wei, Y. Z., B. Fang, S. Iwasa, and M. Kumagai, ‘‘A Novel Electrode Materode Material for Electric Double-Layer Capacitors,’’ J. Power Sources, 141, 386 (2005).

Wu, F. C., R. L. Tseng, and R. S. Juang, ‘‘Adsorption of Dyes and Phenol from Water on the Activated Carbons Prepared from Corncob Wastes,’’ Environ. Technol., 22, 205 (2001).

Wu, F. C., R. L. Tseng, C. C. Hu, and C. C. Wang, ‘‘Physical and Electro-chemical Characterization of Activated Carbons Prepared from Firwoods for Supercapacitors,’’ J. Power Sources, 138, 351 (2004).

Wu, F. C., R. L. Tseng, and R. S. Juang, ‘‘Preparation of Highly Microporous Carbons from Fir Wood by KOH Activation for Adsorption of Dyes and Phenols from Water,’’ Sep. Purif. Technol., 47, 10 (2005a).

(11)

Wu, F. C., R. L. Tseng, C. C. Hu, and C. C. Wang, ‘‘Effects of Pore Structure and Electrolyte on the Capacitive Characteristics of Steam and KOH Activated Carbons for Supercapacitors,’’ J. Power Sources, 144, 302 (2005b).

Wu, F. C., R. L. Tseng, and C. C. Hu, ‘‘Comparisons of Properties and Adsorption Performance of KOH-Activated and Steam-Activated Carbons,’’ Micropor. Mesopor. Mater., 80, 95 (2005c).

Wu, F. C., R. L. Tseng, and R. S. Juang, ‘‘Comparisons of Porous and Adsorption Properties of Carbons Activated by Steam and KOH,’’ J. Colloid Interface Sci., 283, 49 (2005d).

Wu, F. C., R. L. Tseng, C. C. Hu, and C. C. Wang, ‘‘The Capacitive Characteristics of Activated Carbons—Comparisons of the Activation Meth-ods on the Pore Structure and Effects of the Pore Structure and Electrolyte on the Capacitive Performance,’’ J. Power Sources, 159, 1532 (2006). Yang, T. and A. C. Lua, ‘‘Characteristics of Activated Carbons Prepared from

Pistachio-Nut Shells by Physical Activation,’’ J. Colloid Interface Sci., 267, 408 (2003).

Yoshizawa, N., K. Maruyama, Y. Yamada, and M. Zielinska-Blajet, ‘‘XRD Evaluation of CO2Activation Process of Coal and Coconut Shell-Based