Carbon nanofibres and activated carbon nanofibres as electrodes in supercapacitors

Carbon 43 (2005) 551–557

www.elsevier.com/locate/carbon

Carbon nanofibres and activated carbon nanofibres as electrodesin supercapacitors

Cesar Merino a, Pablo Soto a, Eduardo Vilaplana-Ortego b, Jose M. Gomez de Salazar c,Fernando Pico d, Jose M. Rojo d,*

a Grupo Antolin Ingenieria S.A., Carretera Madrid-Irun km. 244.8, E-09007 Burgos, Spainb Departamento de Quimica Inorganica, Universidad de Alicante, E-03080 Alicante, Spain

c Departamento de Ciencia de Materiales e Ingenieria Metalurgica, Facultad de Ciencias, Universidad Complutense Madrid,

Ciudad Universitaria, E-28040 Madrid, Spaind Instituto de Ciencia de Materiales de Madrid, Consejo Superior de Investigaciones Cientıficas, Cantoblanco, E-28049 Madrid, Spain

Received 6 July 2004; accepted 13 October 2004

Available online 25 November 2004

Abstract

Carbon nanofibres have been prepared by a floating catalyst procedure at industrial scale in a metallic furnace. The nanofibres

(50–500nm diameter and 5–200lm length) are grown from the Fe particles used as catalyst. Soot appears together with the carbon

nanofibres. The sample has been chemically activated using KOH as activating agent. Scanning electron microscopy has shown a

smooth surface for the as-prepared carbon nanofibres but a rough surface for the activated ones. The specific surface area increases

from 13 to 212m2/g due to the activation. The volume of the micropores (in the 1–2nm range) and the mesopores (2–5nm range), as

deduced by density functional theory methods, also increases after the activation. Electrochemical behaviour of the as-prepared and

activated carbon nanofibres has been tested in a supercapacitor at laboratory scale using 6M KOH aqueous solution as electrolyte.

The specific capacitance, which is less than 1F/g for the as-prepared sample, increase up to �60F/g for the activated sample. Only aslight decrease in capacitance has been observed as the current density increases. Specific power of �100W/kg at specific energy of1Wh/kg has been found in some particular cases. We have compared the electrochemical parameters of our activated carbon nano-

fibres with those of activated carbon nanofibres coming from a commercial sample; the latter was activated by the same way as our

sample.

� 2004 Elsevier Ltd. All rights reserved.

Keywords: A. Catalytically grown carbon, Electrodes; B. Activation; C. Scanning electron microscopy; D. Electrochemical properties

1. Introduction

Carbon nanofibres are catalytically vapour-grown

carbon fibres (s-VGCF) of submicrometric size (diame-

ters of 50–500nm and lengths of 50–100lm) that showphysical properties between those of commercial ex-

PAN or pitch carbon fibres and carbon nanotubes. They

0008-6223/$ - see front matter � 2004 Elsevier Ltd. All rights reserved.

doi:10.1016/j.carbon.2004.10.018

* Corresponding author. Tel.: +34 913349000; fax: +34 913720623.

E-mail address: [email protected] (J.M. Rojo).

present a highly graphitic structure and can massively

be produced sacrificing product quality in some degree[1].

Two Japanese and two American companies are cur-

rently producing carbon nanofibres in moderate indus-

trial amount (several tens of tons per year) [2]. At the

European level, the Spanish Company Grupo Antolın

Ingenierıa, S.A. is involved in the industrial production

of carbon nanofibres. Its production capacity was of

half a ton per year in 2002; the production is expectedto increase to 10tons in 2004.

Fig. 1. Scheme of the furnace developed by Grupo Antolin for

continuous industrial production of carbon nanofibres.

552 C. Merino et al. / Carbon 43 (2005) 551–557

Carbon nanofibres have been used to fabricate poly-

mer composites with improved tensile strength, tensile

modulus, electrical conductivity and thermal conductiv-

ity. Other applications are, for example, as additive in

tires instead of carbon black, or as anode material in

lithium ion batteries [3].On the other hand it is known that carbons can be

used in general as electrode material for supercapacitors

[4–18]. The capacitance, and hence the energy, comes

mainly from double layer mechanism. For instance, if

the carbon particles are positively polarized, the nega-

tive ions of the electrolyte are situated near those parti-

cles and it gives rise to a double layer. If the carbon

particles are negatively polarized, the double layer isformed with the positive ions of the electrolyte. Since

the double layer appears at the carbon/electrolyte inter-

face, the specific surface area of the carbon strongly af-

fects the capacitance; high specific surface area leads to

high specific capacitance, and hence to high specific en-

ergy. The electrical conductivity and the porosity of the

carbon are associated with the power; high conductivity

and large pores seem to lead to high specific power.The specific surface area and porosity of carbons can

significantly be modified by an activation process. It re-

moves part of the carbon atoms from the structure (pref-

erentially the most reactive ones) and increases porosity

and surface area. There are two methods to activate car-

bons [19]: the so-called physical and chemical activation.

The physical activation consists of carbonisation of a

carbonaceous precursor followed by a controlled gasifi-cation of the carbonised material, or, direct activation of

the raw material in presence of an activating agent such

as CO2 or steam, or both together. The other method

(chemical activation) consists of an impregnation of

the raw material with an activating agent followed by

pyrolisis. Then, the activated carbon is washed to re-

move the remained activating agent. Some activating

agents such as zinc chloride, phosphoric acid, potassiumand sodium hydroxides have been reported as suitable

agents for preparing activated carbons [20].

The porous texture of the activated carbons depends

mainly on the nature of the raw material and the activa-

tion process followed. Recently, some authors [21] have

reported on the difficulty to activate carbon materials

with high degree of graphitisation. Since this is the same

case as the carbon nanofibres, we have chosen a similarprocedure to activate our samples.

In this work we report on the electrochemical behav-

iour of carbon nanofibres and activated carbon nanofi-

bres as electrode materials for supercapacitors. The

carbon nanofibres have been produced at industrial

scale by Grupo Antolın Ingenierıa S.A. They have been

chemically activated by using KOH as activating agent.

Both as-prepared and activated carbon nanofibres havebeen examined by scanning electron microscopy and

their texture has been investigated. The specific surface

area and the pore size distribution have been deter-

mined. The two kind of carbon nanofibres have been

processed to get electrodes, and a supercapacitor at lab-

oratory scale has been mounted and tested under

charge/discharge experiments. We have determined the

specific capacitance, specific energy and specific power.We have compared these parameters with those we mea-

sured in another commercial sample of carbon nanofi-

bres in which chemical activation was carried out by

the same procedure followed with our sample.

2. Experimental

2.1. Preparation of carbon nanofibres

Based on the floating catalyst method reported else-

where [22–24], Grupo Antolın has developed a metallic

furnace to continuously produce carbon nanofibres. A

scheme of the furnace is shown in Fig. 1. It has an inner

diameter of 400mm and a reaction length of 4500mm,

the length being broken down in nine heating zones of500mm each zone. Catalyst, carbon source reagents

and hydrogen are introduced at the top of the furnace.

The amount of the gases is controlled with the help of

flowmeters.

The liquids are introduced into the furnace (at the

top) with the help of a peristaltic pump, which is placed

at the bottom of the furnace. The use of a pneumatic

piston allows the collection of the carbon nanofibres atthe bottom of the furnace without loosing gas tightness

in the furnace. By this way it was possible to produce

carbon nanofibres continuously for one week, stopping

the introduction of the catalyst, hydrogen, and carbon

source reagents about five times a day during less than

15min each time to collect the product.

C. Merino et al. / Carbon 43 (2005) 551–557 553

The H2 flow chosen was 7.4 l/min. The hydrocarbons

used as carbon source were methane (9.3 l/min), and n-

heptane (1 l/h). The catalysts used was a Fe-based com-

pound (ferrocene), which was introduced in the furnace

as a solution in n-heptane (25g/l). The operating furnace

temperature was 1100 �C.

2.2. Activation of carbon nanofibres

It was conducted by a chemical activation method in

which KOH was the activating agent. First the sample

and KOH lentils were mixed at room temperature; the

KOH/sample weight ratio was 3/1. Then carbonisation

of the mixtures was done in a horizontal furnace undernitrogen flow. The heating rate from room temperature

to the final carbonisation temperature (800 �C) was of20 �C/min. The temperature at 800 �C was held for 1h.

Then the furnace was cooled down to room tempera-

ture. The nitrogen flow was kept at 500ml/min along

the heating and cooling runs. After carbonisation the

samples were repeatedly washed with 5 N HCl aqueous

solution. In addition, they were washed with distilledwater to get samples free of chloride ions. Once the acti-

vating agent was removed in the washing step, the sam-

ples were dried at 110 �C for at least 12h. The samples

were weighted to know the final yield of the activation

process.

2.3. Surface characterization

The porous texture of the as-prepared and activated

carbon nanofibres was analysed by physical adsorption

of gases (N2 at 77K and CO2 at 273K) in two automatic

volumetric adsorption systems (Autosorb-6 and Auto-

sorb-6B, Quantachrome Corporation). All samples were

outgassed at 250 �C for 4h prior to the adsorption

measurements. The total micropore volume (pores

smaller than 2nm) was calculated by applying theDubbinin–Radushkevich (DR) equation to the N2 ad-

sorption data collected at 77K. The narrow micropore

volume (pores smaller than 0.7nm) was assessed from

CO2 adsorption data obtained at 273K using the DR

equation [25–27]. The specific surface area was mea-

sured in the relative pressure interval of 0.05–0.30 by

using the BET method. The micropore size distributions

were calculated by applying the density functional the-ory (DFT) to the N2 isotherms using the Autosorb Mul-

tistation software.

2.4. Scanning electron microscopy

The as-prepared and activated carbon nanofibres

were examined by scanning electron microscopy

(SEM) in a Jeol JSM 6345 F microscope. It allowed usto compare the surface of the carbon nanofibres before

and after the activation.

2.5. Electrochemical measurements

From the as-prepared and activated carbon nanofi-

bres, electrodes were processed as composites. Polyviny-

lidenefluoride (PVDF, MW � 534000) was used as an

inert binder. In all cases the content of the PVDF andas-prepared sample (or activated sample) was 10wt.%

and 90wt.%, respectively. The two components of the

composites were mixed and ground in an agate mortar.

Then cylindrical pellets of 13mm diameter and �0.5mmthickness were obtained by cold pressing at 38MPa.

With two equal electrodes a supercapacitor was

mounted in a Swagelok-type cell. The two electrodes

were separated by a glassy microfibre paper into whichthe liquid electrolyte (6M KOH aqueous solution) had

been impregnated. For comparison 2M H2SO4 aqueous

solution was used as electrolyte. The charge and dis-

charge of the supercapacitor cells were followed at room

temperature with a 1286 Solartron potentiostat/galvano-

stat. The galvanostatic and voltametric measurements

were carried out in the current range of 1–100mA and

in the voltage scan rate range of 2–50mVs�1,respectively.

3. Results and discussion

3.1. Activation of the carbon nanofibres

The sample produced as described in the experimen-tal section, was washed ultrasonically in acetone. It re-

moved polycyclic aromatic hydrocarbons that usually

accompany the sample due to the fast cooling of the

gas at the bottom of the furnace. Then a portion of

the sample was activated in KOH as described in the

experimental section. The yield of the activation process

was 58% by weight. The as-prepared and activated sam-

ple from Grupo Antolin are hereafter called S andS-K800, respectively.

For comparison purposes a commercial sample of

carbon nanofibres from Applied Sciences-Pyrograf

Products Inc. (the product of reference Pyrograf III

24-PR PS) was activated according to the procedure

already mentioned. The yield of the activation process

was 52%. The starting and activated sample from Pyrog-

raf are hereafter called Pyr III and Pyr III-K800,respectively.

3.2. Microstructural characterization of the

as-prepared and activated carbon nanofibres

A SEM photograph of the S sample is shown in Fig.

2a. It can be seen carbon nanofibres that are grown from

the Fe particles used as catalyst. Soot is also observed asballs with Fe particles inside. The presence of soot is

indicative of a poor catalytic activity of the Fe particles

Fig. 2. (a) Scanning electron micrograph obtained for the S sample.

(b) Magnification of the arrow-marked carbon nanofibre.

Fig. 4. Scanning electron micrograph showing an activated carbon

nanofibre of the S-K800 sample.

554 C. Merino et al. / Carbon 43 (2005) 551–557

under the production conditions chosen. The fibre

marked with an arrow in Fig. 2a is shown in detail inFig. 2b. The fibre looks quite homogeneous with a diam-

eter that is nearly the same along the fibre axis and

shows a rounded shape at the end. The length of the ob-

tained carbon nanfibres shows values in the range 5–

200lm. The histogram of the diameters of the carbon

nanofibres is shown in Fig. 3 (open bars). It can be seen

that the distribution is rather broad. The average diam-

eter is 209nm; the standard deviation is 134nm.

0 100 200 300 400 500 600 700 800 900 10000

10

20

30

40

50

Frec

uenc

y (%

)

Diameter (nm)

Fig. 3. Diameter distribution of the carbon nanofibres of the samples:

S (open bars) and S-K800 (shaded bars).

The S-K800 sample was also examined by SEM. As

an example the photograph of an activated fibre is

shown in Fig. 4. Although we do not see the ladder-like

structure recently reported [28], our activated fibre

shows a rough surface contrary to the smooth surface

of the as-prepared fibre (Fig. 2b). The end of the acti-

vated carbon nanofibres (not shown) is also modifiedby the activation process. The fibre ends show a sharp

shape instead of the rounded shape of the as-prepared

fibres. A histogram of the diameters of the activated car-

bon nanofibres is shown in Fig. 3 (shaded bars). The his-

togram is similar to that of the S sample; the average

diameter (199nm) and the standard deviation (131nm)

of the activated fibres are nearly the same as the values

found for the as-prepared carbon nanofibres.

3.3. Pore texture of the as-prepared and activated

carbon nanofibres

N2 adsorption isotherms corresponding to the S

(open circles) and S-K800 (solid circles) samples are

shown in Fig. 5. For comparison the adsorption

0

40

80

120

160

200

240

280

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0P/Po

Volu

me

adso

rbed

(cc/

g) S

TP

Fig. 5. Nitrogen adsorption isotherms obtained at 77K for the as-

prepared S (open circles) and Pyr III (open triangles) samples, and for

the activated S-K800 (closed circles) and Pyr III-K800 (closed

triangles) samples.

10

15 (a)

charge

C. Merino et al. / Carbon 43 (2005) 551–557 555

isotherms obtained on the Pyr III (open triangle) and

Pyr III-K800 (solid triangle) samples are also presented

in that figure. For the as-prepared S and Pyr III samples

the volume adsorbed at low and medium relative pres-

sures is very low. At high relative pressures the S iso-

therm shows no significant adsorption but the Pyr IIIisotherm shows a hysteresis, which is typical of materials

containing mesopores [29]. The isotherms of the acti-

vated S-K800 and Pyr III-K800 samples show: (i) a

sharp adsorption at very low relative pressures that is

typical of materials having micropores, and (ii) a hyster-

esis at medium-high relative pressures indicating the

presence of mesopores.

The BET surface areas and the micropore volumescalculated from N2 adsorption data at 77K [VDR (N2)]

and from CO2 adsorption data at 273K [VDR (CO2)]

are outlined in Table 1 for each sample. The comparison

between the total micropore volume [VDR (N2)] and the

narrow micropore volume [VDR (CO2)] gives an idea of

the micropore size distribution. Since the difference is

negligible [i.e. VDR (N2) � VDR (CO2)], the microporos-

ity is homogeneous in size [25–27,30]. Although the PyrIII sample has a BET surface area higher than the S

sample, the BET surface area developed by the chemical

activation is similar in both S-K800 and Pyr III-K800

samples.

Fig. 6 shows the DFT pore size distributions for the

as-prepared samples and the activated ones. We can

see that: (i) the two activated samples (solid symbols)

Table 1

Textural characteristics obtained from N2 and CO2 isotherms for the

original carbon nanofibres and the activated ones

Sample SBET(m2/g)

VDR(N2)

(cc/g)

VDR(CO2)

(cc/g)

VDR(N2)–VDR(CO2)

(cc/g)

S 13 – – –

S-K800 212 0.09 0.09 0.00

Pyr III 40 – – –

Pyr III-K800 183 0.08 0.09 0.01

0.000

0.001

0.002

0.003

0.004

0.005

0.006

0.007

0 10 20 30 40 50Pore width (Å)

Pore

vol

ume

(cc/

Å/g)

Fig. 6. DFT pore size distribution for the as-prepared S (open circles)

and Pyr-III (open triangles) samples, and for the activated S-K800

(closed circles) and Pyr III-K800 (closed triangles) samples.

show higher porosity than the two as-prepared ones

(open symbols), and (ii) the chemical activation has

led to similar pore size distributions in the two samples.

The activated samples present a mean micropore size at

around 11A, and an unimodal micropore size distri-

bution with most of the pores in the range 7–20A. Inaddition, they have mesopores with sizes in the range

20–50A. Therefore the chemical activation carried out

develops microporosity and mesoporosity.

3.4. Electrochemical behavior of the carbon nanofibres

and activated carbon nanofibres

From the S, S-K800 and Pyr III-K800 samples,electrodes were processed as composites and sup-

ercapacitors at laboratory scale were mounted. The vol-

tammetry curves obtained for the S-K800 sample are

shown as an example in Fig. 7a. The curves have been

obtained with two voltage scan rates: 2mVs�1 (solid

line) and 10mVs�1 (dash line), and in both cases a

nearly rectangular shape is observed; this shape is char-

acteristic of supercapacitors made with carbon-basedelectrodes. The supercapacitor was also tested under gal-

vanostatic conditions and the voltage vs. time plot is

0.0 0.2 0.4 0.6 0.8 1.0-15

-10

-5

0

5

dischargeCur

rent

(mA)

Voltage (V)

0 100 200 300 4000.0

0.2

0.4

0.6

0.8 (b)

∆ V 2

∆ V1

dischargecharg

e

Volta

ge(V

)

Time (s)

Fig. 7. (a) Voltammograms obtained at 2mVs�1 (solid line) and

10mVs�1 (dash line) for the supercapacitor having S-K800 as

electrode material and 6M KOH aqueous solution as electrolyte. (b)

Galvanostatic curves recorded at 5mA (solid line) and 10mA (dash

line) in the same supercapacitor.

556 C. Merino et al. / Carbon 43 (2005) 551–557

shown in Fig. 7b. The solid and dash lines stand for the

results obtained at the currents of 5 and 10mA, respec-

tively. In both cases we can see at the beginning of the

charge and discharge a sharp change in voltage (DV1),

which is usually associated with the equivalent series

resistance (ESR) of the supercapacitor [9]. We also seea gradual variation in voltage (DV2) that increases for

the charge and decreases for the discharge. From DV2

in the discharge, the capacitance (C) of the supercapac-

itor was calculated according to C = I · Dtd/D V2, where

Dtd is the time spent along the discharge and I is the

current. The capacitance of each electrode (Ce) was cal-

culated as Ce = 2 · C because both electrodes are ar-

ranged in series within the supercapacitor. From thecapacitance of each electrode and its carbon mass, the

specific capacitance (in F/g) was also calculated.

Variation of the specific capacitance vs. current den-

sity is plotted in Fig. 8 for the S, S-K800 and Pyr III-

K800 samples. The specific capacitance is below 1F/g

for S (diamonds) in agreement with its low specific sur-

face area and low porosity (see Table 1 and Fig. 6).

However the specific capacitance is �60F/g for S-K800 (squares). The increase in specific capacitance

seems to be associated with the increase in specific sur-

face area and in porosity (micropores and mesopores)

due to the activation. The value of 60F/g is higher than

the specific capacitance reported for activated carbon

nanofibres when 1M solution of Et4NBF4 in polycar-

bonate was used as electrolyte [28]. That value is similar

to the specific capacitance reported for single-walled car-bon nanotubes and activated fibres [4,5,12,14,18] but

lower than the specific capacitance reported for other

activated carbons and also for mesoporous carbons

[8,13,17]. In our S-K800 sample we also observe that

the specific capacitance decreases slightly as the current

density increases (from �60F/g at 1mA/cm2 to �50F/g

0 20 40 60 80 1000

20

40

60

80

Spe

cific

capa

cita

nce

(Fg-1

)

Current density (mA cm-2)

Fig. 8. Specific capacitance vs. current density for the samples S, S-

K800, and Pyr III-K800 as electrodes. The results were obtained with

6M KOH and 2M H2SO4 as electrolytes. The symbols stand for as

follows: open diamonds for S and KOH, open squares for S-K800 and

KOH, stars for Pyr III-K800 and KOH, and open circles for S-K800

and H2SO4.

at 75mA/cm2). This trend, which indicates a good

behavior in power, has also been observed in carbons

having a significant amount of mesopores [6,8,17].

It is worth to note that our already mentioned specific

capacitances were obtained in a supercapacitor in which

the electrolyte was 6M KOH aqueous solution. We havealso checked the effect of the electrolyte on the specific

capacitance by doing the same electrochemical measure-

ments with 2M H2SO4 aqueous solution as electrolyte.

We measured a specific capacitance of �35F/g for theS-K800 sample (circles in Fig. 8). The lower specific

capacitance found agrees with similar results reported

for single-walled carbon nanotubes [18] and can be ex-

plained on the basis of the larger size of the hydratedH3O

+ and SO2�4 ions compared to the hydrated K+

and OH� ions [6,11].

For comparison purposes, the Pyr III-K800 sample

was also used as electrode material in a supercapacitor

in which 6M KOH was the electrolyte. The choice of

KOH was because it led to higher specific capacitances

as already mentioned. The specific capacitance found

and the dependence of the specific capacitance as a func-tion of the current density (stars in Fig. 8) are similar as

for the S-K800 sample. Therefore, the two activated S-

K800 and Pyr III-K800 samples show similar specific

capacitances in agreement with their similar surface

characteristics (specific surface area and pore size

distribution).

The specific power vs. specific energy is plotted in

Fig. 9 for the two activated samples (S-K800 and PyrIII-K800), and for the two electrolytes (KOH and

H2SO4). The energy was calculated as (1/2)C(DV2)2,

where C is the capacitance of the supercapacitor and

DV2 is the voltage decrement in discharge as already

mentioned. The power was calculated as energy/Dtd,

where td is the time spent in the discharge. The specific

energy and specific power were obtained from the en-

ergy and power, respectively, by dividing these magni-

10-1 100 101100

101

102

103

Spe

cific

pow

er(W

kg-1)

Specific energy (Wh kg-1)

Fig. 9. Specific power vs. specific energy for the samples S-K800 and

Pyr III-K800. KOH and H2SO4 were the electrolytes. The symbols

stand for as in the caption of Fig. 8.

C. Merino et al. / Carbon 43 (2005) 551–557 557

tudes by the mass of the carbon in the supercapacitor. In

Fig. 9 we can see that: (i) the maximum specific power

(�200W/kg) is similar for both S-K800 (squares) andPyr III-K800 (stars) and is not affected by the electrolyte

chosen, i.e. either KOH (squares) or H2SO4 (circles), (ii)

the maximum specific energy (�1Wh/kg) is similar forS-K800 and Pyr III-K800, and (iii) the maximum spe-

cific energy is higher for KOH (squares and stars) com-

pared to H2SO4 (circles).

4. Conclusions

Carbon nanofibres have been produced in a continu-ous way at industrial scale. The as-prepared carbon

nanofibres show low specific surface area and low poros-

ity as well as negligible specific capacitance as electrode

material.

The chemical activation carried out using KOH as

activating agent gives rise to an alteration of the surface

of the carbon nanofibres as observed by SEM. The spe-

cific surface area and porosity (micropores of 1–2nmsize and mesopores of 2–5nm size) increase and the spe-

cific capacitance is of �60F/g even at high current den-sity. In some particular cases specific energy of 1Wh/kg

at specific power of 100W/kg has been found.

Acknowledgment

F. Pico thanks the Red de Pilas de Combustible del

CSIC for the fellowship got.

References

[1] Morita T, Inoue H, Suhara Y. Fine carbon fiber and method for

producing the same. US Patent 2002/0058139 A1, 2002.

[2] Tibbetts GG. Vapor-Grown Carbon Fiber research and applica-

tions: achievements and barriers. In: Biro LP, Bernardo CA,

Tibbetts GG, Lambin Ph, editors. Carbon filaments and nano-

tubes: common origins, differing applications?. Dordrecht: Klu-

wer Academic Publishers; 2001. p. 1–9.

[3] Lake ML. Novel applications of VGCF including hydrogen

storage. In: Biro LP, Bernardo CA, Tibbetts GG, Lambin Ph,

editors. Carbon filaments and nanotubes: common origins,

differing applications?. Dordrecht: Kluwer Academic Publishers;

2001. p. 331–41.

[4] Shi H. Activated carbons and double layer capacitance. Electro-

chim Acta 1996;41:1633–9.

[5] Niu C, Sichel EK, Hoch R, Moy D, Tennent H. High power

electrochemical capacitors based on carbon nanotube electrodes.

Appl Phys Lett 1997;70:1480–2.

[6] Qu D, Shi H. Studies of activated carbons used in double-layer

capacitors. J Power Sources 1998;74:99–107.

[7] Lin C, Ritter JA, Popov BN. Correlation of double-layer

capacitance with the pore structutre of sol–gel derived carbon

xerogels. J Electrochem Soc 1999;146:3639–43.

[8] Yoon S, Lee J, Hyeon T, Oh SM. Electric double-layer capacitor

performance of a new mesoporous carbon. J Electrochem Soc

2000;147:2507–12.

[9] Conway BE. Electrochemical supercapacitors. N.Y.: Kluwer

Academic; 1999. 528-532.

[10] Kotz R, Carlen M. Principles and applications of electrochemical

capacitors. Electrochim Acta 2000;45:2483–98.

[11] Endo M, Takeda T, Kim YJ, Koshiba K, Ishii K. High power

electric double layer capacitor (EDLC�s); from operating principle

to pore size control in advance activated carbons. Carbon Sci

2001;1:117–28.

[12] Frackowiak E, Beguin F. Carbon materials for the electrochem-

ical storage of energy in capacitors. Carbon 2001;39:937–50.

[13] Weng TC, Teng H. Characterization of high porosity carbon

electrodes derived from mesophase pitch for electric double-layer

capacitors. J Electrochem Soc 2001;148:A368–73.

[14] Emmenegger Ch, Mauron Ph, Sudan P, Wenger P, Hermann V,

Gallay R, et al. Investigation of electrochemical double-layer

(ECDL) capacitors electrodes based on carbon nanotubes and

activated carbon materials. J Power Sources 2003;124:321–9.

[15] Lozano-Castello D, Cazorla-Amoros D, Linares-Solano A, Shi-

raishi S, Kurihara H, Oya A. Influence of pore structure and

surface chemistry on electric double layer capacitance in non-

aqueous electrolyte. Carbon 2003;41:1765–75.

[16] Babel K, Jurewicz K. KOH activated carbon fabrics as superca-

pacitor material. J Phys Chem Solids 2004;65:275–80.

[17] Fuertes AB, Pico F, Rojo JM. Influence of pore structure on

electric double-layer capacitance of template mesoporous carbons.

J Power Sources 2004;133:329–36.

[18] Pico F, Rojo JM, Sanjuan ML, Anson A, Benito AM, Callejas

MA, et al. Single-walled carbon nanotubes as electrodes in

supercapacitors. J Electrochem Soc 2004;151:A831–7.

[19] Bansal RC, Donnet JB, Stoeckli F. Active carbon. New

York: Marcel Dekker; 1998.

[20] Lozano-Castello D, Lillo-Rodenas MA, Cazorla-Amoros D,

Linares-Solano A. Preparation of activated carbons from Spanish

anthracite. I. Activation by KOH. Carbon 2001;39(5):741–9.

[21] Macia-Agullo JA, Moore JC, Cazorla-Amoros D, Linares-Solano

A. Chemical activation by KOH and NaOH of carbon materials

with different cristallinity. In: Extended Abstracts, Carbon 2003:

International Conference, Oviedo, Spain, July 6–10. Spanish

Carbon Group, 2003, p. 197.

[22] Arakawa K. Process for preparing fine carbon fibers in a gaseous

phase reaction. US Patent 4,572,813, 1986.

[23] Arakawa K. Process for preparing fine fibers. US Patent

4,640,830, 1987.

[24] Tibbetts GG, Gorkiewicz DW. A new reactor for growing carbon

fibers from liquid- and vapor-phase hydrocarbons. Carbon 1993;

31(5):809–14.

[25] Rodriguez-Reinoso F, Linares-Solano A. Microporous structure

of activated carbons as revealed by adsorption methods. In:

Thrower PA, editor. Chemistry and physics of carbon, vol.

21. New York: Marcel Dekker; 1988. p. 1–146.

[26] Cazorla-Amoros D, Alcaniz-Monge J, Linares-Solano A. Char-

acterisation of activated carbon fibres by CO2 adsorption.

Langmuir 1996;12(11):2820–4.

[27] Cazorla-Amoros D, Alcaniz-Monge J, De la Casa-Lillo MA,

Linares-Solano A. CO2 as an adsorptive to characterise carbon

molecular sieves and activated carbons. Langmuir 1998;14(16):

4589–96.

[28] Yoon SH, Lim S, Song Y, Ota Y, Qiao W, Tanaka A, et al. KOH

activation of carbon nanofibres. Carbon 2004;42:1723–9.

[29] Sing KSW, Everett DH, Haul RAW, Moscou L, Pierotti RA,

Rouquerol J, et al. Reporting physisorption data for gas/solid

systems with special reference to the determination of surface area

and porosity. Pure Appl Chem 1985;57(4):603–19.

[30] Linares-Solano A, Salinas-Martınez de Lecea C, Alcaniz-Monge

J, Cazorla-Amoro s D. Further advances in the characterisation of

microporous carbons by physical adsorption of gases. Tanso

1998;185:316–25.