journal of power sources 157 (2006) 11–27

Journal of Power Sources 157 (2006) 1127

Review

Carbon properties and their role in supercapacitors

Abstract

Supercstorage syhigh ratescapacitorseparatesof porouselectronicchemicalgenerallyto generagenerallywill contiwettability, and reduced inter-particle contact resistance. 2006 Published by Elsevier B.V.

Keywords:

Content

1. Int2. En3. ED4. Ele

4.1

4.24.3

CorrespE-mail a

This revbattery techrole of carb

0378-7753/$doi:10.10164.3.1. Carbon blacks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224.3.2. Carbon aerogels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224.3.3. Carbon fibres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224.3.4. Glassy carbons. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

onding author.ddress: [email protected] (A.G. Pandolfo).iew is one of a series dealing with the role of carbon in electrochemical energy storage. The review covering the role of carbon in valve-regulated lead-acidnology is also published in this issue, J. Power Sources, volume 157, issue 1, pages 310. The reviews covering the role of carbon in fuel cells and theon in graphite and carbon powders were published in J. Power Sources, volume 156, issue 2, pages 128150.

see front matter 2006 Published by Elsevier B.V./j.jpowsour.2006.02.065Supercapacitor; Ultracapacitor; Porosity; surface-area; Activated carbon; Capacitance

s

roduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12ergy storage in electrochemical double-layer capacitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13LC construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13ctrode material characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14. Carbon structure and form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

4.1.1. Engineered carbons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154.1.2. Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164.1.3. Electrical properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164.1.4. Carbon surface functionalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184.1.5. Double-layer capacitance of carbon materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

. Carbon surface-area and porosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

. Carbon forms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21A.G. Pandolfo , A.F. HollenkampCSIRO Division of Energy Technology, Box 312, Clayton South, Vic. 3169, Australia

Available online 4 April 2006

apacitors (also known as ultracapacitors) offer a promising alternative approach to meeting the increasing power demands of energy-stems in general, and of portable (digital) electronic devices in particular. Supercapacitors are able to store and deliver energy at relatively(beyond those accessible with batteries) because the mechanism of energy storage is simple charge-separation (as in conventional

s). The vast increases in capacitance achieved by supercapacitors are due to the combination of: (i) an extremely small distance thatthe opposite charges, as defined by the electric double-layer; (ii) highly porous electrodes that embody very high surface-area. A varietyforms of carbon are currently preferred as the electrode materials because they have exceptionally high surface areas, relatively highconductivity, and acceptable cost. The power and energy-storage capabilities of these devices are closely linked to the physical and

characteristics of the carbon electrodes. For example, increases in specific surface-area, obtained through activation of the carbon,lead to increased capacitance. Since only the electrolyte-wetted surface-area contributes to capacitance, the carbon processing is requiredte predominantly open pores that are connected to the bulk pore network. While the supercapacitors available today perform well, it isagreed that there is considerable scope for improvement (e.g., improved performance at higher frequencies). Thus it is likely that carbonnue to play a principal role in supercapacitor technology, mainly through further optimization of porosity, surface treatments to promote

12 A.G. Pandolfo, A.F. Hollenkamp / Journal of Power Sources 157 (2006) 1127

4.4. Carbon nanostructures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction

Supercapacitors, ultracapacitors and electrochemical double-layer capacitors (EDLCs) are commonly used names for a classof electrochemical energy-storage devices that are ideally suitedto the rapid storage and release of energy. The term supercapaci-tor is adopted in this paper. Compared with conventional capac-itors, the specific energy of supercapacitors is several orders ofmagnitude higher (hence the super or ultra prefix). Superca-pacitors also have a higher specific power than most batteries,but their specific energy is somewhat lower. Through appro-priate cell design, both the specific energy and specific powerranges fortude and thenergy suptem. This uspecific enposition beand Table 1

Supercaters complebatteries anstorage procan be bothexceedinginterest ingrowing, raics, hybrid[24].

Researcmain areasstorage, natrochemica

Fig. 1. Specifitrochemical c

In redoxtors), a revresulting corigin (henelectrostatian electrocan extent liable surfacpseudocaparutheniumpolypyrrolecharge stor

r is scom

a trapacit mdersacro

hat cjacen

stod bylectrrent

an

uchs, instor

highgdeenumbsupe

als [5st deariouidelyocusity.

apacbonof e

supplatiosupercapacitors can cover several orders of magni-is makes them extremely versatile as a stand-aloneply, or in combination with batteries as a hybrid sys-nique combination of high power capability and goodergy, allows supercapacitors to occupy a functionaltween batteries and conventional capacitors (Fig. 1).pacitors are particularly useful because their parame-ment the deficiencies of other power sources such asd fuel cells. Due to their highly reversible charge-cess, supercapacitors have longer cycle-lives andrapidly charged and discharged at power densities

1 kW kg1 [1]. These features have generated greatthe application of supercapacitors for a wide, andnge of applications that include: consumer electron-electric vehicles, and industrial power management

h into supercapacitors is presently divided into twothat are based primarily on their mode of energy

mely: (i) the redox supercapacitor and (ii) the elec-l double layer capacitor.

pacitoBy

way asSuperc(per uneral orplacelayer tthe adcan be(createarea e

is inheions tosteps, schangeenergya verycyclinhave b

A nogy ofmaterithe moin its vand wment fresistivsuperc

Cartrodestives,interca

for the conitance [16]suited as el. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

supercapacitors (also referred to as pseudocapaci-ersible Faradaic-type charge transfer occurs and theapacitance, while often large, is not electrostatic ince the pseudo prefix to provide differentiation fromc capacitance). Rather, capacitance is associated withhemical charge-transfer process that takes place tomited by a finite amount of active material or avail-e [5]. The most commonly investigated classes ofcitive materials are transition metal oxides (notably,

oxide) and conducting polymers such as polyaniline,or derivatives of polythiophene [69]. Given that

age is based on a redox process, this type of superca-omewhat battery-like in its behaviour.parison, the EDLC stores energy in much the sameaditional capacitor, by means of charge separation.itors can, however, store substantially more energy

ass or volume) than a conventional capacitor (by sev-of magnitude) because: (i) charge separation takes

ss a very small distance in the electrical double-onstitutes the interphase between an electrode andt electrolyte [5]; (ii) an increased amount of chargered on the highly extended electrode surface-area

a large number of pores within a high surface-ode material). The mechanism of energy storagely rapid because it simply involves movement ofd from electrode surfaces. In batteries, additionalas heterogeneous charge-transfer and chemical phasetroduce relatively slow steps into the process of

age and delivery. For similar reasons, EDLCs exhibitdegree of reversibility in repetitive chargedischargeemonstrated cycle lives in excess of 500,000 cyclesachieved [10].er of reviews have discussed the science and technol-rcapacitors for various configurations and electrode,1113]. The EDLC version of the supercapacitor isveloped form of electrochemical capacitor. Carbon,s forms, is currently the most extensively examinedutilised electrode material in EDLCs with develop-ing on achieving high surface-area with low matrixA number of carbon manufacturers are now targetingitors as a market for their products [14,15].materials have long been incorporated into the elec-nergy-storage devices as: electro-conductive addi-orts for active materials, electron transfer catalysts,n hosts, substrates for current leads, and as agentstrol of heat transfer, porosity, surface-area and capac-. For these reasons, carbons are of course also wellectrode materials for EDLCs.c energy and power capabilities of capacitors (electrostatic), elec-apacitors (supercapacitors), batteries and fuel cells [13].

It is clear that the ultimate performance of carbon-basedsupercapacitors will be closely linked to the physical and chemi-

A.G. Pandolfo, A.F. Hollenkamp / Journal of Power Sources 157 (2006) 1127 13

Table 1Comparison of typical capacitor and battery characteristics (adapted from Ref. [36])Characteristic Electrolytic capacitor Carbon supercapacitor Battery

Specific energy (Wh kg1)


Fig. 2. Reprecharged state)

trode, usuathat also in

The douis given by

Cdl = A4twhere isregion, A thof the electcombinatioextremelysible for thand power

E = 12CV

Pmax = V2

4Rwhere C isand R is th

The capacteristics oand the porrespondingcapacitance[32].

Cell volcific energvoltage ofstability. Aalkalis (e.g(up to 1 Shand, theyrange with[33]. Neversurface-are

icantly higher than that of the same electrode in non-aqueoussolutions owing to the higher dielectric constant that pertains to

s sy-aqu

hat a]. Sial tlyte

ningmploseityf mare thistanightor ar ofivelynce,rs inc

troninteent-cionicionicelect

ctro

attrses fties,

con

sur

d cortemsentation of an electrochemical double layer capacitor (in its.

lly derived from a three-electrode laboratory test cellcorporates a reference and counter electrode [30].)ble layer capacitance, Cdl, at each electrode interface

(2)

the dielectric constant of the electrical double-layere surface-area of the electrode and t is the thickness

rical double layer. In double-layer capacitors, it is then of high surface-area (typically >1500 m2 g1) withsmall charge separation (Angstroms) that is respon-eir extremely high capacitance [31]. The energy (E)(Pmax) of supercapacitors are calculated according to2 (3)

(4)

aqueouNon

oped t[33,35portionelectrocontaibeen elarly thresistivorder otherefonal res

A hcapacinumbecollectresistapacito

elec the

curr

the the the

4. Ele

Therial ariproper

high high goo highthe dc capacitance in Farads, V the nominal voltage,e equivalent series resistance (ESR) in ohms [10].acitance of a device is largely dependent on the char-f the electrode material; particularly, the surface-areae-size distribution. Due to the high porosity, and cor-ly low density of carbons, it is usually the volumetricof each electrode that determines the energy density

tage is also an important determinant of both the spe-y and the power of supercapacitors. The operatingsupercapacitors is usually dependant on electrolytequeous electrolytes, such as acids (e.g., H2SO4) and., KOH) have the advantage of high ionic conductivitycm1), low cost and wide acceptance. On the other

have the inherent disadvantage of a restricted voltagea relatively low decomposition voltage of 1.23 Vtheless, the specific capacitance (Farads g1) of higha carbons in aqueous electrolytes tends to be signif-

controlle processa relativel

In generthe construthe propertarea to be mto be the sureviewing tin more detand chemicof the role

4.1. Carbo

Carbon(sp3 bondinstems [34].eous electrolytes of various types have been devel-llow the use of cell operating voltages above 2.5 Vnce the specific energy of supercapacitors is pro-o the square of the operating voltage, non-aqueousmixtures such as propylene carbonate or acetonitrile,dissolved quaternary alkyl ammonium salts, have

oyed in many commercial supercapacitors, particu-targeting higher energy applications. The electrical

of non-aqueous electrolytes is, however, at least angnitude higher than that of aqueous electrolytes ande resulting capacitors generally have a higher inter-ce.

internal resistance limits the power capability of thend, ultimately, its application. In supercapacitors, asources contribute to the internal resistance and are

measured and referred to as the equivalent seriesor ESR [10,36]. Contributors to the ESR of superca-lude:

ic resistance of the electrode material;rfacial resistance between the electrode and theollector;(diffusion) resistance of ions moving in small pores;resistance of ions moving through the separator;

rolyte resistance.

de material characteristics

action of carbon as a supercapacitor electrode mate-rom a unique combination of chemical and physicalnamely:

ductivity,face-area range (1 to >2000 m2 g1),rosion resistance,perature stability,d pore structure,bility and compatibility in composite materials,y low cost.

al terms, the first two of these properties are critical toction of supercapacitor electrodes. As will be seen,ies of carbon allow both conductivity and surface-

anipulated and optimised. Such activities continuebject of a considerable amount of research. Prior tohe results of this research, it is useful first to considerail other aspects of carbon, e.g., its structural diversityal behaviour, so as to establish a better understandingof carbon materials in supercapacitors.

n structure and form

has four crystalline (ordered) allotropes: diamondg), graphite (sp2), carbyne (sp1) and fullerenes (dis-


torted sp2). While two carbon allotropes are naturally foundon earth as minerals, namely, natural graphite and diamond,the other forms of carbon are synthetic. Carbon is consideredunusual in the number of its allotropic structures and the diver-sity of structural forms, as well as in its broad range of physicalproperties [37,38].

Due to the wide range of carbon materials, and to avoid confu-sion, the term carbon is typically used to describe the elementrather than its form. To describe a carbon-based material, it isbest coupled with a qualifier such as carbon black, activatedcarbon, vitreous carbon and others. A comprehensive guideto the terminology and description of carbon solids, as used inthe science and technology of carbon and graphite materials, isavailable [39].

4.1.1. Engineered carbonsThe majority of commercial carbons used today can be con-

veniently described as engineered carbons. These are manu-factured carbons that have an amorphous structure with a moreor less disordered microstructure based on that of graphite [37].Amorphous carbons can be considered as sections of hexago-nal carbon layers with very little order parallel to the layers.The process of graphitisation consists essentially of the order-ing and stacking of these layers and is generally achieved byhigh-tempeof amorphmaterials cextent, for

The marich organi(a process rof these cae.g., the carcarbonizatiand the str

Summaries of the key processing conditions that lead tospecific carbon products (and their features) are outlined inTable 3.

During carbonization, carbon precursors go through a ther-mal decomposition (pyrolysis), which eliminates volatile mate-rials that include heteroatoms. With increasing temperature,condensation reactions are initiated and localized graphitic unitscommence to grow and become aligned into small graphite likemicrocrystallites, i.e., stacks of aromatic layers or graphenesheets [41]. The carbon precursor and its processing conditionswill determine the size of the graphene sheets (La), the stackingnumber of graphene sheets (Lc) and the relative orientation of thecrystallites. The size and orientation of the crystallites are veryimportant as they define the materials texture and the degree ofelectrical conductivity [42].

Some carbon precursors (e.g., petroleum pitch, coal pitch,some polymers) pass through a fluid stage (referred to asmesophase) during carbonization that allows large aromaticmolecules to align with each other and form a more extensivepre-graphitic structure. Upon further high-temperature treat-ment (>2500 C), these carbons can be converted into highlyordered graphite and are referred to as graphitising carbons [41].

Other carbon precursors retain a solid phase during car-bonization, and the limited mobility of the crystallites leads

formran

anno

aturs. Nh asnd me, P

nd coy nore w

Table 3Common prec s (adaCarbon mater Con

Gas phaseCarbon bla PrePyrolytic c DepVapour-gro PreFullerene ConNanotubes Con

Liquid phaseCokes SheGraphite HigCarbon fibr Spi

Solid phaseActivated c Car

Molecular SeleGlass-like c Slow

Carbon fibr SlowHighly orie Higrature treatment (>2500 C). Between the extremesous carbon and graphite, a wide variety of carbonan be prepared and their properties tailored, to somespecific applications.jority of carbon materials are derived from carbon-c precursors by heat treatment in inert atmosphereseferred to as carbonization). The ultimate propertiesrbons are dependent on a number of critical factors,bon precursor, its dominant aggregation state duringon (i.e., gas, liquid or solid), processing conditions,uctural and textural features of the products [40].

to thesist ofrials ctempercarbonals succoals achloridrigid aof manstructu

usors and controlling production factors for various classes of carbon material

ial (phase during aggregation) Common precursors

cks Hydrocarbon gas or liquidarbon Hydrocarbon gaswn carbon fibres Hydrocarbon gas

Graphite rodHydrocarbon vapour

Coals, petroleum pitchPetroleum coke

es (pitch derived) Coal pitch, petroleum pitch

arbons Biomass, coals, petroleum coke, selectedpolymers

sieve carbon Selected biomass, coals, polymersarbons Thermosetting polymers (e.g., polyfurfuryl

alcohol)es (polymer derived) Selected polymers (e.g., polyacrylonitrile)nted graphite Pyrolytic carbon, poly-imide filmation of a rigid amorphous structure that con-domly-oriented graphene layers [40]. These mate-t be readily converted to graphite by further high-e treatment and are referred to as non-graphitizingon-graphitizing carbons are produced from materi-

biomass (e.g., wood, nut shells, etc.), non-fusingany thermosetting polymers (e.g., polyvinylidene

VDC). The loss of volatiles and the retention of amplex molecular structure during the carbonizationn-graphitizable carbons can lead to a highly porousithout the need for further activation.

pted from Ref. [40])trolling production factor Structural/textural feature

cursor concentration Colloidal/nanosizedosition on a substrate Preferred orientation

sence of a catalyst Catalyst particle size/shapedensation of carbon vapour Nanosize moleculedensation of carbon vapour Single wall, multi-wall, chiral

ar stress Mesophase formation and growthh temperature Mesophase formation and growthnning Mesophase formation and growth

bonization/activation Nanosize pores

ctive pore development Nanosize pore/constrictionscarbonization Random crystallites, impervious

carbonization Random crystallites, non-poroush molecular orientation Highly oriented crystallites


4.1.2. ActivationAs noted earlier, one of the great attractions of using car-

bon as an einto a formally speaki(and porosis referredmaterials isa relativelymentary crthem. The iresidues (tathese porescarbon preperature, tiover the resof the interguard theirthe processactivation a

Thermavation, entgasification700 and 11as steam, cgasificationvolume ancarbon buucts. The legoverningby the tempactivation iactivity is adensity, red

Chemictemperaturaction of cand potassiis usually rinorganic rfrom the cExceptionabeen prepa[44].

4.1.3. ElecThe ele

related to thity/conductboth to theproperties,ticulate forelectrode s

Most cativities >10bonded cais strongly

rapidly with heat treatment temperature up to 700 C and ata much slower rate for heat treatment above 700 C [45]. The

sing proportion of conjugated carbon in the sp2 state dur-rbonization progressively increases the conductivity ofrting material as electrons associated with -bonds arelized and become available as charge carriers [46]. Elec-onductivity increases as separate conjugated systems alsoe interconnected to form a conducting network. The tem-re range at which the conductivity of solid carbon initiallyto level off (600700 C) corresponds to the range incarbons lose acidic functionalities, primarily by forma-H2O and CO2 [45]. Thus heat treatment increases the

ctivity of carbons by altering the degree of structural dis-which can vary from nearly amorphous carbon to the neart crystals of graphite (formed at temperatures >2500 C).sistent with the anisotropic nature of graphite, the elec-

resistivity of carbons also varies with their degree of crys-aphic orientation [47]. The specific resistivity of graphites in the plane of the aromatic rings (a-axis direction) is cm, whereas perpendicular to this plane, along the c

he resistivity is in the vicinity of 102 cm [48]. Theity of selected representative carbon materials are shown

. 3 and can vary by at least three orders of magnitude,range from that of semiconductors to semi-metals [49].ilst tateriy, thes istact

us crrentbon/h theequ

esisttal gralyticarbonlectrode material is that it can be readily convertedthat has very high specific surface-area. Gener-

ng, the process employed to increase surface-areaity) from a carbonised organic precursor (a char)to as activation and the resulting broad group ofreferred to as activated carbons. Chars usually havelow porosity and their structure consists of ele-

ystallites with a large number of interstices betweennterstices tend to be filled with disorganized carbonrs) that block the pore entrances. Activation opensand can also create additional porosity. Varying the

cursor and activation conditions (particularly tem-me and gaseous environment) allows some controlulting porosity, pore-size distribution, and the naturenal surfaces. Although carbon manufacturers closelyprocedures for the activation of commercial carbons,es can be placed into two general categories: thermalnd chemical activation [38,43].

l activation, sometimes referred to as physical acti-ails the modification of a carbon char by controlled, and is usually carried out at temperatures between

00 C in the presence of suitable oxidising gases sucharbon dioxide, air, or mixtures of these gases. During, the oxidising atmosphere greatly increases the pore

d surface-area of the material through a controlledrn-off and the elimination of volatile pyrolysis prod-vel of burn-off is, perhaps, the most important factorthe quality of the activated carbon and is controllederature and duration of activation. A high degree of

s achieved by increased burn-off, but the additionalccompanied by a decrease in carbon strength, a loweruced yield and pore widening.

al activation is usually carried out at slightly loweres (400700 C) and involves the dehydratingertain agents such as phosphoric acid, zinc chlorideum hydroxide. Post-activation washing of the carbonequired to remove residual reactants as well as anyesidue (sometimes referred to as ash) that originatesarbon precursor or is introduced during activation.lly high surface-area materials (>2500 m2 g1) havered with potassium hydroxide activation techniques

trical propertiesctrical properties of carbon materials are directlyeir structure. For electrode applications, the resistiv-ivity of the material is a major concern. This appliesintrinsic properties of the carbon and to aggregateparticularly in the common situation where fine par-ms of carbon are bound together to form a compositetructure (v.i.).rbon precursors are generally good insulators (resis-12 cm) that contain a high proportion of or sp3

rbon structures. The conductivity of solid carbonsinfluenced by heat treatment in that it increases

increaing cathe stadelocatrical cbecomperatubeginswhichtion ofconduorder,perfec

Contronictallogrcrystal105axis, tresistivin Figwhich

Whbon mphologparticlthe cona porothe cuthe carthrougber of

Fig. 3. Rgle crys(4) pyroglassy che intrinsic (i.e., intra-particle) resistivity of a car-al is dependent on its chemical and structural mor-e electrical resistance of a packed bed of carbona function of both its intra-particle resistance and(or inter-particle) resistance [28,46]. In addition, for

arbon artefact, there is a resistance associated with-carrying path, through the carbon particle, acrosscarbon and carbon/metallic collector interfaces, andmetallic collector. Although there could be a num-

ivalent paths to a particular point, the value of the

ivity () vs. temperature plots for various forms of carbon: (1) sin-phite; (2) highly oriented pyrolytic graphite; (3) graphite whisker;graphite; (5) petroleum coke carbon; (6) lampblack carbon; (7); (8) carbon filmelectron beam evaporated [49].


Table 4Resistivities of several types of powdered carbons under differing compaction pressures [50]Sample

ElectrographiActive carbonCarbon-blackAcetylene-blaPetroleum cokPetroleum cokBabacu nut coEucalyptus lig

a Kilogram

resistance aposition wresistance o

Particletance of agboth the phshape, aggrpact the paelectrical pof compaccompactiononly at highof a packedintrinsic re

Packingbetter contthrough thesmaller paramount offormed betcapacitors trequiremenadvantagesness of theturn, enablecapacities.cles of appthat are we

The precarbon powentially forincrease thcrystallineincrease ththe exposuby prolongConversely(at 1000resistivity o

In prepait is standarbinding agHere, with

ntiald thebonctrodnce,tor Ere aingtivitectety an

treathwand isringationnot nre. Al tretiestionctedty alistanutor

sic) etor ESRutorimpentResistivity ( cm)200 kgfa cm2

te 1.40 103(Riedel de-Haen) 1.07(SRF) 3.93 101ck 1.78 101e (1400 C) 1.50 101e (1600 C) 6.47 102ke 1.24nite carbon 9.21 101

of force used in original reference work, kgf = 9.8 N.

ssociated with the path will be a function of: (i) theithin the bed, (ii) packing parameters, and (iii) thef the carbon material itself [28].contact resistance is a major contributor to the resis-gregated carbon powders and is highly dependent onysical morphology of the carbon particle (e.g., size,egation, etc.) and the pressure that is applied to com-rticles [28,46,50]. Espinola et al. [50], measured theroperties of a range of carbon powders as a functiontion pressure, see Table 4. It was found that greater

pressure leads to a decrease in resistivity and thatcompaction pressures does the measured resistivity

carbon powder begin to approach asymptotically thesistivity of the carbon material itself.pressure compresses the particle bed and produces

act between particles and reduces the path lengthbed. The effect of pressure is generally larger for

ticles. A closely-packed structure also minimizes theelectrolyte contained in the inter-particle void spaceween the irregular carbon particles, although in super-his needs to be carefully balanced against electrolytets to avoid electrolyte-depletion effects. One of theof the use of carbon powders is to allow the thick-carbon coating to be easily manipulated and this, ins the preparation of electrodes of varying shapes and

The use of thin coating films (containing small parti-ropriate size) generally results in low ESR electrodesll suited to high chargedischarge applications [28].sence of surface oxygen also affects the resistivity ofders [46,51]. Oxygen functionalities, which prefer-

m at the edges of the graphite like microcrystallites,

is esseto avoithe carthe eleresistacapaci

Theincreasconducbe expporosisimilartive pathis trecompaexplanity dostructuthermationaliconduccompaporosical rescontrib(intrincapaciinant Econtrib

Anattachme barrier for electrons to transfer from one micro-element to the next [52,53]. Therefore, processes thate surface oxygen content of carbons, for example byre of carbons to oxygen at elevated temperatures ored grinding, also increase their electrical resistivity., the removal of surface oxides, by heat treatmentC) in an inert atmosphere, decreases the electricalf carbon powders [48].ring electrodes based on powdered activated carbons,d practice to employ an organic (usually polymeric)

ent to establish and maintain an electrode structure.the maintenance of maximum conductivity in mind, it

collector. Tand collectlong-termmorphologthe metallipresence ofdesign (espbinders, itin a manneand maintachemical asubstrates i1000 kgfa cm2 2000 kgfa cm2

7.70 103 6.81 1032.90 101 1.79 1016.10 102 2.44 1024.86 102 2.64 1023.02 102 1.96 1021.74 102 1.35 1022.72 101 1.21 1011.90 101 9.57 102

that the binder be well distributed and used sparinglyformation of an electrically insulating layer between

particles. Excessive amounts of binder will increasee resistance, largely through a greater particle contactand will ultimately contribute to an increase in theSR [54].re varying reports in the literature on the effect ofsurface-area and porosity on the intrinsic electronicy of compacted carbon powders. Intuitively, it wouldd that the volumetric conductivity decreases as thed surface-area increase (for carbons experiencingtment temperatures) since the number of conduc-ys would also be expected to decrease. Generally,infrequently observed, but is more noticeable whensimilar carbon materials [51,55]. The most likelyis that carbons with a similar surface-area or poros-

ecessarily have the same physical, or even chemical,s outlined earlier, variations in carbon precursor,

atment, the presence (or absence) of surface func-and associated changes in the mode of electronic[46], are more likely to influence the resistivity ofcarbon powders than variations in surface-area or

one. It should also be acknowledged that the electri-ce of the carbon material is one of several possibles to the capacitor ESR [10,28]. Whilst the carbonlectrical resistance needs to be kept to a minimum theSR may ultimately be governed by other, more dom-contributors that can sometimes overshadow minors.ortant factor in determining capacitor ESR is the

of the carbon material to the metallic current-

he interfacial electrical resistance between the carbonor can be a major contributor to capacitor ESR andstability. It is influenced by factors such as carbony, particle size, surface functionalities, cleanliness ofc substrate (i.e., free from surface impurities), thea binder, electrode form/shape, electrolyte, and cell

ecially containment pressure). Even with the use ofis challenging to adhere carbons to metallic surfacesr that does not substantially increase capacitor ESRins stability over the lifetime of the capacitor. Variousnd physical treatments are often applied to metallicn order to improve carbon-binder adhesion and stabi-


lize the carbon/metal interface [56]. The use of highly etched oreven porous metallic substrates (e.g., metal gauze, metal foams)also assistsin, the caran increase

4.1.4. CarFor carb

surface grofacial statethat includresistance,characterisas being msites: basalto be morewith unpaivation thatpurity grapsites [45]. Cvary with thwith the ratdegree of d

Porous cciable concgen and hyhalogens [4ing materiaresult of inrated on toexposure toadsorption)sub-ambienversible ad(e.g., ambious oxygenCarbonoxgroup on c

Three tneutral, havmined by ttemperaturAcidic surfoxygen betsolutions asidered to blactonic anoxides aretend to formpounds bytemperaturcally inertwettabilityspecific capand greater

The exteular oxyge

op: relationship between concentration of acidic functional groups onctrode and leakage current of corresponding double-layer capacitor [62].influence of the oxygen content of activated carbons used for electrode

citance after application of 2.8 V at 70 C for 1000 h [63].

ly the rate and mechanism of capacitor self-discharge (ore) [5]. In particular, carbons with a high concentration ofsurface functionalities are prone to exhibit high rates ofscharge (Fig. 4(a)). The increase in leakage current sug-hat the oxygen functional groups may serve as active sites,can catalyze the electrochemical oxidation or reductioncarbon, or the decomposition of the electrolyte compo-6264]. Conversely, the removal of oxygenated surfacenalities by high-temperature treatment in an inert envi-

nt results in lower leakage currents [62,65]. The presencegenated functional groups can also contribute to capacitorlity, which results in an increased ESR and deteriorationacitance [6365]. The removal of oxygen from the car-sed in supercapacitor electrodes generally improves theirty (Fig. 4(b)).carbon adhesion by physically locking, or keyingbon to the porous metal collector, and by providingin contact surface-area [54,57].

bon surface functionalitieson-based double-layer capacitors, the presence ofups is known to influence the electrochemical inter-of the carbon surface and its double-layer properties

e: wettability, point of zero charge, electrical contactadsorption of ions (capacitance), and self-dischargetics [5,58]. Graphitic carbon surfaces can be regardedade up of (at least) two chemically different kinds ofand edge carbon sites [45]. Edge sites are consideredreactive than basal sites as they are often associatedred electrons. This view is supported by the obser-the reactivity of edge sites towards oxygen, for highhite, is an order of magnitude greater than that of basalonsequently, the chemical properties of carbons alsoe relative fraction of edge sites and basal plane sites;io of edge to basal sites generally increasing with theisorder.arbons are almost invariably associated with appre-entrations of heteroatoms, which are primarily oxy-drogen, and to a lesser degree, nitrogen, sulfur and3,48]. These heteroatoms are derived from the start-ls and become part of the chemical structure as acomplete carbonization. They may also be incorpo-the carbon surface during subsequent treatment orair. Carbons can readily physisorb (i.e., reversiblemolecular oxygen upon exposure to air, even at

t temperatures [49]. Oxygen chemisorption (irre-sorption) also begins to occur at low temperaturesent) and increases with temperature to form vari--based functionalities on the carbon surface [48,59].ygen complexes are by far the most important surfacearbons.ypes of surface oxides, namely, acidic, basic ande been proposed to form on carbon surfaces as deter-he formation history of the carbon material and thee at which it was first exposed to oxygen [46,60].ace oxides are formed when carbons are exposed toween 200 and 700 C or by reactions with oxidizingt room temperature. These surface groups are con-e less stable and include groups such as carboxylic,

d phenolic functionalities. Basic and neutral surfaceconsidered to be more stable than acidic oxides and

after a carbon surface, freed from all surface com-heat treatment, comes in contact with oxygen at lowes [46,60]. Functional groups that are electrochemi-in the potential range of operation can enhance theof carbon electrodes and, consequently, increase theacitance of the carbon through improved pore accesssurface utilization [61].nt of oxygen retention as physically adsorbed molec-n, or as surface complexes, is believed to influence

Fig. 4. TACF eleBottom:on capa

strongleakagacidicself-digests twhichof thenents [functioronme

of oxyinstabiof capbons ustabili


Oxygen surface functional groups have also been shown toinfluence the rest potential of activated carbons [66]. The restpotential of activated carbons was found to be proportional to thelogarithm of the oxygen content or to the concentration of acidicsurface sites. Carbons with a high rest potential will experienceundesirably higher voltages when charged and this could lead togas generation.

Whilst charge storage on carbon electrodes is predominantlycapacitive, there are also contributions from surface functionalgroups that can be rapidly charged and discharged and giverise to pseudocapacitance [13,67]. Various oxidative treatmentshave been applied to carbons [58,6870], but it is often dif-ficult to isolate the chemical changes from the physical andstructural changes that frequently accompany oxidative treat-ments. Hsieh and Teng [64] used a relatively mild oxidativetechnique to oxidize polyacrylonitrile-activated carbon fabricswithout introducing significant physical changes to the carbon. Itwas found that the Faradaic current, a direct measure of pseudo-capacitance, increased significantly with the extent of oxygentreatment, while the change in double-layer capacitance wasonly minor120 to 15increases in

4.1.5. DouIt is usu

bon (expresurface-areobserved [simplificatiwhat is oftetions in thein the specogy; (iii) vpseudocapain the cond

The surfcommonly77 K) and u

mate of apparent surface-area. Despite its widespread use, theapplication of this approach to highly porous (particularly micro-porous) and heterogeneous materials has some limitations [74],and is perhaps more appropriately used as a semi-quantitativetool. Possibly the greatest constraint in attempting to correlatecapacitance with BET surface-area, is the assumption that thesurface-area accessed by nitrogen gas is similar to the surfaceaccessed by the electrolyte during the measurement of capac-itance. While gas adsorption can be expected to penetrate themajority of open pores down to a size that approaches the molec-ular size of the adsorbate, electrolyte accessibility will be moresensitive to variations in carbon structure and surface proper-ties [75]. Electrolyte penetration into fine pores, particularly bylarger organic electrolytes, is expected to be more restricted (dueto ion sieving effects) and vary considerably with the electrolyteused [76]. Variations in electrolyteelectrode surface interac-tions that arise from differing electrolyte properties (viscosity,dielectric constant, dipole moment) will also influence wettabil-ity and, hence, electrolyte penetration into pores.

The specific double-layer capacitance (expressed per unit ofrea, 2le 5.ighl

uble-rtedayertanceplanehighouldlow

te isand

-conn o

ise tde ofial exevel

Table 5Typical value

Carbonaceous er ca

Activated carb

Carbon black

Carbon fiber c

GraphiteBasal planeEdge plane

Graphite powGraphite clothGlassy carbonCarbon aerog

a Values ba. The large increase in specific capacitance (from0 F g1) was, however, accompanied by undesirableinternal resistance and leakage current.

ble-layer capacitance of carbon materialsally anticipated that the capacitance of a porous car-ssed in F g1) will be proportional to its availablea (in m2 g1). Whilst this relationship is sometimes44,71], in practice it usually represents an over-on [30,72,73]. The major factors that contribute ton a complex (non-linear) relationship are: (i) assump-measurement of electrode surface-area; (ii) variationsific capacitance of carbons with differing morphol-ariations in surface chemistry (e.g., wettability andcitive contributions discussed above); (iv) variationsitions under which carbon capacitance is measured.ace areas of porous carbons and electrodes are mostmeasured by gas adsorption (usually nitrogen atse BET theory to convert adsorption data into an esti-

BET ain Tabto be hthe dois repobasal lcapacibasalwith aratio) c

ThegraphiRandina semitributiogives rbon sipotent(Csc) d

s for electrochemical double-layer capacitance of carbonaceous materials [5]material Electrolyte Double-lay

on 10% NaCl 19

1 M H2SO4 831 wt.% KOH 10

loth 0.51 M Et4NBF4 in propylene carbonate 6.9

0.9 N NaF 30.9 N NaF 5070

der 10% NaCl 350.168 N NaCl 10.70.9 N NaF 13

el 4 M KOH 23

sed on estimates. For a comprehensive discussion see Ref. [48].in F cm ) of a range of carbon materials is listedThe reported values vary considerably and appear

y dependent on carbon morphology. Most notably,layer capacitance of the edge orientation of graphiteto be an order of magnitude higher than that of the[61,77]. One determinant of specific double-layercould therefore be the relative density of edge andgraphitic structures in carbon materials. Carbons

er percentage of edge orientations (i.e., high Lc/Labe expected to exhibit a higher capacitance.capacitance recorded when the basal layer of

exposed to solution has been examined in detail byYeager [78,79]. By treating basal plane carbon as

ductor, they were able to show that there is a dis-f charge carrier concentration (charge density) thato a semi-conducting space-charge region on the car-

the interface. This means that some of the appliedtends into the carbon and a space-charge capacitanceops. This capacitance is in series with other capac-

pacitance (F cm2)a RemarksSurface-area 1200 m2 g1

Surface-area 80230 m2 g1

Surface-area 1630 m2 g1

Highly oriented pyrolytic graphite

Surface-area 4 m2 g1Surface-area 630 m2 g1SolidSurface-area 650 m2 g1


itive components of the double-layer, namely, the Helmholtz(CH) and diffuse (Cdiff) double-layer contributions. In effect,the storedtional dieleis then repr

1C

= 1Csc

+

For thecharge captherefore dRandin anzero chargthe space-ctrast, the eddensity sinspace-charoverall intCH.

The oriexaminedmetal and ament produit also allothe behaviauthors nothe conductronic statenation forboth vary sthe potentiis the concby the numspace-chareither sidetrode increauthors alsramificatiocapacitor egreater degeffective thto the dimFermi lenglayer). In tand less caexplanationtance oftenextremely lwalls) [81,

The medent on thFig. 5 dema range ofcharge curare particuto the greanarrower p

Relationship between capacitance and BET surface-area of activatedmeasured at different current densities [135]. Top: low discharge current; bottom: high discharge current (160 mA).

des, particularly when used for comparative purposes,be measured and compared at a fixed current den-

arbon surface-area and porosity

ous carbons (particularly activated forms) are character-extremely large BET surface areas, that range from 500

00 m2 g1. This surface-area largely arises from a com-terconnected network of internal pores. The IUPAC [85]es pores into three classes: micropores (diameters lessnm), mesopores (diameters between 2 and 50 nm) and

pores (diameters greater than 50 nm).ropores have a high surface-area to volume ratio and,uently, when present in significant proportions are acontributor to the measured area of high surface-areaed carbons. Micropore sizes extend down to molecu-ensions and play an important role in the selectivity

orption-based processes, through restricted diffusion andular sieve effects. Fine micropores also exhibit a greaterentadsorbate affinity due to the overlap of adsorptioncharge, at a given potential, is spread over an addi-ctric element. The capacitance of the electrode (C)esented as

1CH

+ 1Cdiff

(5)

semi-conducting basal graphite layer, the space-acitance is significantly less than Cdiff and CH, andominates the overall interface capacitance. In fact,d Yeagers calculations show that at the point ofe essentially all of the potential is associated withharge region within the carbon electrode. By con-ge orientation of graphite has a higher charge carrierce its conductivity is more like that of a metal. Thege capacitance is correspondingly greater so that theerface capacitance is now dominated by Cdiff and

ginal work by Randin and Yeager was later re-by Gerischer [80], this time treating carbon as applying density-of-states arguments. While this treat-ced similar capacitance results to the earlier work,wed Kotz and Hahn [81,82] to explain some of

our of carbon in greater detail. In particular, theseted that the magnitude of both the capacitance andtivity can be directly related to the density of elec-s. Further, this treatment also provides an expla-the observation that capacitance and conductivitytrongly with potential and exhibit minima close toal of zero charge (pzc). Central to this behaviourlusion that both properties are essentially limitedber of available charge carriers within the carbon

ge region. This limitation is alleviated at potentialsof the pzc because charging of the carbon elec-

ases the numbers of available charge carriers. Theo indicate that these relationships have importantns for the future development of carbon-based super-lectrodes. For example, as carbons are activated torees, the pore walls become thinner. As a result, theickness of the electrode material becomes similar

ensions of the space-charge region (defined by theth, and approximately equivalent to one graphenehis situation, space-charge regions begin to overlappacitance is generated per unit electrode area. Thisis consistent with the smaller than expected capaci-observed for highly activated carbon materials witharge surface areas (and correspondingly thinner pore83].asurement of carbon capacitance is also very depen-e experimental conditions employed. For example,onstrates that the capacitance values measured forporous carbons can vary substantially with dis-

rent [84]. The capacitance of microporous carbonslarly affected by variations in discharge current dueter possibility of restricted electrolyte diffusion inores. Ideally, reported capacitance values of carbon

Fig. 5.carbons(10 mA)

electroshouldsity.

4.2. C

Porized byto 30plex inclassifithan 2macro

Micconseqmajoractivatlar dimof adsmolecadsorb


forces from opposing pore walls [86]. Accordingly, adsorptionin fine pores can occur via a pore filling mechanism rather thansolely by sBET calcusion of adsapplicationsurface-are

Mesopolarger sizeviding widemake a negbons and ththe interior

While aurally occusize classeactivation cative contrisurface-areactivated cthe total po

As discucontributetrochemicathe double-of electrolylimitation oThe surfacedependenteffectivelycapacitance

Studiesconcludedelectro-adsstudy also clayer capacbetween 0.tance obsethe correspference is gthe solvateaccess to smthe lower selectrolytesreported foicant portioalso be accproposed threduce ioninto smalle

Narrowciable solelectrolytecontributeresponse),dation ofing/dischar

ffectorous

ution, plectrdable deutedde wpacit). Ae ofed a

systeionsnal sa m

lyteed. Sresisjoritizedarly,to aon-banduenentlyals with a tailored pore-size distribution to yield hightance and low resistance electrodes.

arbon forms

a large extent, carbons used in commercial supercapaci-ve remained proprietary, but most would appear to be af activated carbon which is often blended with a conduc-rbon black or graphite. Activated carbons obtained fromized phenolic resins or petroleum cokes appear to be par-ly suitable, especially for organic-based systems [15,63].forms, or blends of carbons, are also employed and theties of selected carbons are discussed in greater detailurface coverage (as is assumed by the Langmuir andlations of surface-area). In such cases, the conver-orption data into an estimate of surface-area, by theof the BET equation, can lead to unrealistically higha estimates.res also contribute to surface-area and their relativelyalso allows improved adsorbate accessibility by pro-r transport pores for diffusion. Macropores generallyligible contribution to the surface-area of porous car-eir main function is to act as transport avenues intoof carbon particles.

ctivated carbons, particularly those derived from nat-rring precursors, tend to contain pores from all threes, careful selection of the carbon precursor and theonditions does allow significant control over the rel-bution of each size class. Materials with the highesta are consistently obtained from highly microporousarbons, with high pore volumes, in which >90% ofre volume arises from microporosity.ssed earlier, while the vast majority of open pores can

to the measured surface-area, not all pores are elec-lly accessible. Ultimately, pore sizes will approachlayer dimensions, with the result that the movementte will be restricted and, eventually, there will be an the ability of the electrolyte to form a double-layer.-area arising from pores in this size range (which areon electrolyte molecular dimensions) would not beutilized and is unlikely to contribute to double-layer.

on micropore accessibility in aqueous solvents havethat, generally, pores >0.5 nm are available for theorption of simple hydrated ions [76,87]. A furtheroncluded that the optimal pore-size range for double-itance of a carbon aerogel in aqueous H2SO4 was8 and 2.0 nm [88]. The electrode material capaci-rved with organic electrolytes is generally less thanonding value in an aqueous electrolyte [7]. This dif-enerally attributed to the larger overall diameter of

d organic electrolyte ions, which results in limitedaller pores. Whilst there is considerable debate over

ize limit of pores that can be accessed by organic, it is apparent from the high capacitances regularlyr highly microporous carbons [30,44] that a signif-n of carbon microporosity (i.e., pores


4.3.1. Carbon blacksCarbon blacks are a group of materials that are character-

ized by havwhich are pposition ofthe gas phaparticles coerates (grofundamentconditions,method oferties of caparticle sizface chemi

Carbontypes of bHighly constructure (iture), high(oxygen freically in thby the relaclosely-spaconductionis of criticainter-aggrethe conducpercolationthe conducporosity anweight andaggregates

The surfi.e., from 8 kW kg1.to improve their capacitance, carbon nanotubes have

ically activated with potassium hydroxide. The treat-ubstantially increase the surface-area of carbon nan-a factor of two to three times) whilst maintaininglar morphology [125,129]. Whereas the treatment

have little effect on nanotube diameter, it can consid-ten the length and is accompanied by the developmentnd surface irregularities through a partial erosion ofarbon layers [129]. Frackowiak et al. [125] man-reased the BET surface-area of MWNTs from anof 430 to 1035 m2 g1 by activation with potassiumWhilst the product still maintained a high degreeosity, the activation process also introduced consid-

roporosity. The specific capacitance of the material1 (8.7F cm2) in alkaline media but was limited(6.2F cm2) in non-aqueous media.have been prepared as composites with polyvinyli-

de (PVDC). After carbonisation, the composite elec-bited a maximum specific capacitance of 180 Fg1sured power density of 20 kW kg1 in potassium[121,122,130]. The high specific capacitance, for anaterial with a surface-area of only 357 m2 g1, was

o an increased surface-area and a redistribution ofsize to smaller, more optimal, values near 35 nm

ons in the effective surface-area of CNTs hasstudies into the preparation of nanocomposites ofconducting polymers [121,131]. These composites

rge capacitance, by combining the electric double-citance of the CNTs and the redox capacitance ofting polymer [120]. They are typically prepared bychemical polymerization of a suitable monomer

polypyitancegreatePpy alproducIn thesand conal res

Asare usu

enrichods ofbetweetive fogrowinapproamateritrode f

Cargrownresultiitancestabiliwell-aThese110tor, cogave apower

Altstructewell ahighlyally losurfacity and

5. Sum

Carined aContinwith loBET sous foand naof capparticlalso inmance

good epriatelstored

Theengin(Ppy), have been prepared with a specific capac-170 F g1 [123,124]. This value is considerably

n the corresponding values for pristine MWNTs orSimilarly, SWNT-Ppy nanocomposites have beenith specific capacitances up to 265 F g1 [120,121].ocomposites, the Ppy also acts as a conducting agentutes to the specific capacitance by reducing the inter-ce of the supercapacitor [121].other forms of particulate carbon, carbon nanotubesformed into electrodes using rather tedious binder-r binderless, fabrication methods. Both these meth-trode fabrication result in a high contact resistancee active material and the current collector. An attrac-of carbon nanotube electrode can be obtained bye nanotubes directly on to a conductive substrate. This

inimizes the contact resistance between the actived the current-collector and greatly simplifies elec-ation [132,133].nanotubes, uniformly 50 nm in diameter, have beengraphite foil [132,134]. Subsequent testing of themposite electrode confirmed a high specific capac-

15.7 F g1 (1 M H2SO4) and good electrochemicalmmenegger et al. [133] have also successfully grownd carbon nanotube films on aluminium supports.tubes were multi-walled with lengths in the range

nd diameters between 5 and 100 nm. A supercapaci-cted from the nanotube-coated aluminium electrodemetric capacitance of 120 F cm3 and a very highity.h there are good indications that EDLC devices con-m carbon nanotubes may offer power capabilities8 kW kg1 [122,128], their specific energies are

endent on their method of preparation and are gener-hat the levels achieved from more conventional higha carbons. This, in addition to their limited availabil-h cost, presently limits their commercial utilization.

ry

in its various forms, is the most extensively exam-idely utilised electrode material in supercapacitors.efforts are aimed at achieving higher surface-area

matrix resistivity at an acceptable cost. Carbons withe areas of up to 3000 m2 g1 are available in vari-viz., powders, fibres, woven cloths, felts, aerogels,bes. Even though surface-area is a key determinantnce, other factors such as carbon structure, pore size,e, electrical conductivity and surface functionalitiesce capacitance and ultimately supercapacitor perfor-bon materials that have both a high surface-area andolyte accessibility (favourable distribution of appro-ed pores) allow the optimum amount of charge to bedelivered.jority of commercial carbons can be described as

carbons. These are manufactured and have an


amorphous structure with a more or less disordered microstruc-ture, which is based on that of graphite and can be viewed assections ofparallel tothe processsheets, theiand the relaentation ofproperties aitance and

The preelectrochemdouble-layadsorptiontics. Whilstporous carbdocapacitaand, aftergressive deseries resis

Considethe developdistributionresistance.oxides or creceiving itance, as isof operatinenhance thwithout depower.

Acknowled

The authCSIRO Enetive.

Reference

[1] A.K. S[2] P.J. Ma[3] A. Nis[4] J. Nick

ble LaBeach,

[5] B.E. Ctals an

[6] J.P. Zh[7] M. Ma

Scrosatdemic/

[8] M. Ma493.

[9] K. Naocal Soc1996, p

[10] X. AndTechno

[11] S. Sarangapani, B.V. Tilak, C.-P. Chen, J. Electrochem. Soc. 143 (1996)3791.

[12] A. Burke, J. Power Sources 91 (2000) 37.. Kot. Rei

ologyance

. Otsg theigh P003..S. F. von. Gou.L. C. Ster.C. GO. Bo5..I. Be,800,6.A. Ratent.L. B

apacitemina.L. Bnited. En001).F. R

osium.M. Nalia,. Qu,.K. S656.. Bur

nd Suresenteptem. Haa.E. C.F. R994).F. B

ge andeld, A

ty, Pit. Mcechno.O. Poyes. Fitz7 (19. Ina. Ma. Bouateria.C. Bew Y. Lozishi,.A. Lol. 24. Binind Ph25.hexagonal carbon layers with little long-range orderthe layers. The selection of the carbon precursor anding conditions will influence the size of the graphener degree of stacking (into graphitic micro-crystallites)tive orientation of these crystallites. The size and ori-the graphitic crystallites strongly influences carbonnd defines the texture, porosity, surface-area, capac-

electrical conductivity.sence of surface groups on carbons modifies theical interfacial state of the carbon surface and its

er properties, e.g., wettability, point of zero charge,of ions (capacitance) and self-discharge characteris-the presence of oxygenated species on the surface ofons is often linked to increases in capacitance (pseu-

nce), this increase often leads to irreversible changesprolonged cycling, can be accompanied by a pro-terioration of capacitance and increases in equivalenttance and self-discharge rates.rable research is presently being directed towardsment of carbon materials with a tailored pore-sizeto yield electrodes with high capacitance and low

The incorporation of redox materials (e.g., metalonducting polymers) into carbon electrodes is also

ncreased attention as a means of increasing capaci-the tailoring of electrodes and electrolytes capable

g at higher voltages (>3 V). Clearly, the goal is toe specific energy of carbon-based supercapacitorstracting from the present very high levels of specific

gements

ors wish to acknowledge financial support from thergy Transformed National Research Flagship initia-

s

hukla, Resonance 6 (2001) 72.hon, C.J. Drummond, Aust. J. Chem. 54 (2001) 473.

hino, J. Power Sources 60 (1996) 137.erson, Proceedings of the Ninth International Seminar on Dou-yer Capacitors and Similar Energy Storage Devices, DeerfieldFL, 1999.

onway, Electrochemical SupercapacitorsScientific Fundamen-d Technological Applications, Kluwer, New York, 1999.eng, Electrochem. Solid-State Lett. 2 (1999) 359.stragostino, F. Soavi, C. Arbizzani, W. van Schalkwijk, B.i (Eds.), Advances in Lithium-ion Batteries, Kluwer Aca-Plenum Publishers, 2002, p. 481.stragostino, C. Arbizzani, F. Soavi, Solid State Ionics 148 (2002)

i, Y. Oura, H. Tsujimoto, Proceedings of the Electrochemi-iety, vol. 96-25, The Electrochemical Society, Pennington, NJ,. 120.rieu, Energy Storage Syst. Electron.: New Trends Electrochem.l. 1 (2000) 521.

[13] R[14] M

n

m

[15] KinH2

[16] A[17] H[18] G[19] D[20] O[21] D[22] J.

5[23] H

2[24] R

P[25] D

c

S[26] D

U[27] M

(2[28] M

p[29] A

tr[30] D[31] A

1[32] A

a

PS

[33] O[34] B[35] M

(1[36] A

a

He

[37] BT

[38] HN

[39] E6

[40] M[41] H[42] X

M[43] R

N[44] D

ra

[45] Cv

[46] Sa

1z, M. Carlen, Electrochim. Acta 45 (2000) 2483.merink, Advanced Capacitor World Summit: Building the Tech-Applications and New Business Opportunities for High Perfor-Electrochemical Capacitors (ECs), Washington, DC, 2003.uka, C.L. Segal, Advanced Capacitor World Summit: Build-Technology Applications and New Business Opportunities for

erformance Electrochemical Capacitors (ECs), Washington, DC,

ialkov, Russ. J. Electrochem. 36 (2000) 389.Helmholtz, Ann. Phys. (Leipzig) 89 (1853) 211.y, J. Phys. 9 (1910) 457.hapman, Phil. Mag. 25 (1913) 475.n, Zeit. Elektrochem. 30 (1924) 508.rahame, Chem. Rev. 41 (1947) 441.ckris, M.A. Devanathan, K. Muller, Proc. R. Soc. A274 (1963)

cker, Low voltage electrolytic capacitor, United States Patent16 (1957).ightmire, Electrical energy storage apparatus, United States

3,288,641 (1966).oos, S.D. Argade, International Seminar on Double Layer Super-ors and Similar Energy Storage Devices, Florida Educationalrs, Deerfield Beach, FL, 1991, p. 1.oos, Electrolytic capacitor having carbon paste electrodes,States Patent 3,536,963 (1970).

do, T. Takeda, Y.J. Kim, K. Koshiba, K. Ishii, Carbon Sci. 1117.ose, Proceedings of the 33rd International Power Sources Sym-, Pennington, NJ, 1988, p. 572.amisnyk, B.E. Thesis, University of Technology, Sydney, Aus-

2003.H. Shi, J. Power Sources 74 (1998) 99.hukla, S. Sampath, K. Vijayamohanan, Curr. Sci. 79 (2000)

ke, M. Arulepp, Proceedings of the Symposium on Batteriespercapacitors, San Francisco, CA, 2001, p. 576 (Notes: Papered at the Electrochemical Society Meeting in San Francisco,ber 25, 2001).s, E.J. Cairns, Annu. Rep. Prog. Chem., Sect. C 95 (1999) 163.onway, J. Solid State Electrochem. 7 (2003) 637.ose, C. Johnson, T. Owens, B. Stephens, J. Power Sources 47303.

urke, T.C. Murphy, Materials for Electrochemical Energy Stor-ConversionBatteries, Capacitors and Fuel Cells: Symposiumpril 1720, 1995, San Francisco, CA, Materials Research Soci-

tsburgh, PA, 1995, p. 375.Enaney, T.D. Burchell (Eds.), Carbon Materials for Advancedlogies, Pergamon, 1999, p. 1.ierson, Handbook of Carbon, Graphite, Diamond and Fullerenes,Publications, NJ, USA, 1993.er, K.-H. Kochling, H.P. Boehm, H. Marsh, Pure Appl. Chem.95) 473.gaki, L.R. Radovic, Carbon 40 (2002) 2263.rsh, Introduction to Carbon Science, Butterworths, 1989.rrat, H. Marsh, F. Rodriguez-Reinoso (Eds.), Science of Carbonls, Universidad de Alicante, 2000, p. 1.ansal, J.B. Donnet, F. Stoeckli, Active Carbon, Marcel Dekker,ork, 1988 (Chapter 2).ano-Castello, D. Cazorla-Amoros, A. Linares-Solano, S. Shi-H. Kurihara, A. Oya, Carbon 41 (2003) 1765.eon y Leon, L.R. Radovic, Chemistry and Physics of Carbon,, Marcel Dekker, New York, 1994, p. 213.ak, A. Swiatkowski, M. Pakula, L.R. Radovic (Eds.), Chemistryysics of Carbon, vol. 27, Marcel Dekker, New York, 2001, p.


[47] T.D. Tran, K. Kinoshita, Proceedings of the Electrochemical Society,vol. 98-15, The Electrochemical Society, Pennington, NJ, 1998, p. 548.

[48] K. Kinoshita, Carbon: Electrochemical and Physiochemical Properties,Wiley

[49] I.L. Spof Car

[50] A. Esp33.

[51] W.M.Carbon

[52] K.H. R[53] M. Po

(1997)[54] L. Bon

Source[55] X. And[56] P.L. Ta

A292[57] J.H. K

Lee, J.[58] K. Kin

ical CaNJ, 19

[59] P.E. Fa[60] B.R. P

6, Mar[61] D. Qu,[62] A. Yos[63] T. Mor

60 (19[64] C.-T. H[65] T. Mor

ceeding93-23,

[66] M.W. V[67] B.E. C

Capaci1995, p

[68] M. IshSource

[69] T. Mom60 (19

[70] K. Jure[71] A. Yos

(1996)[72] S. Shir[73] M. En

M.S. D[74] S. Low

Chapm[75] J. Kore[76] G. Sali

Soc. 14[77] R.L. M

LaboraDekker

[78] J.-P. R36 (19

[79] J.-P. R58 (19

[80] H. Ger[81] R. Kot

the 13tEnergy2002.

[82] M. HaElectro

[83] D. Qu,

[84] H. Tamai, Kouzu, M. Masayuki, H. Morita, Yasada, Electrochem. SolidState Lett. 6 (2003) A214A217.

[85] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J.ouquW. P. Shi. Lin,. De. Dehemisork, 1.G. P.E. C.-I. Pochem. Tayuctione PubB. DoologyP. Zh. Osa. Bec001). Ric.W. P. Pro. Fisater.. Sch77.. Pro. Petr. Escruroca. Li,. Salharac.H. Yds.),

. 287.. Nisociety. Bra. Goe. Oyauction61..M. J

nd Ch.G.

lectroennin. Braarbon. Bra. Haa. Alliolid-S.M. R. Mcegla..H. Aochem.H. Loon,.H. Aee, YInterscience, New York, 1988.ain, P.L. Walker, P.A. Thrower (Eds.), Chemistry and Physics

bon, vol. 16, Marcel Dekker, New York, 1981, p. 119.inola, P.M. Miguel, M.R. Salles, A.R. Pinto, Carbon 24 (1986)

Hess, C.R. Herd, J.B. Donet, R.C. Bansal, M.J. Wang (Eds.),Black, 2nd ed., Marcel Dekker, New York, 1993, p. 89.

adeke, K.O. Backhaus, A. Swaitkowski, Carbon 29 (1991) 122.lovina, B. Babic, B. Kaluderovic, A. De Kanski, Carbon 351047.nefoi, P. Simon, J.F. Fauvarque, C. Sarrazin, A. Dugast, J. Powers 79 (1999) 37.rieu, L. Josset, Proc. Electrochem. Soc. 95-29 (1996) 181.

berna, P. Simon, J.F. Fauvarque, J. Electrochem. Soc. 150 (2003)A300.im, K.H. Shin, C.S. Jin, D.K. Kim, Y.G. Lim, J.H. Park, Y.S.S. Joo, K.H. Lee, Electrochemistry 69 (2001) 853.oshita, X. Chu, Proceedings of the Symposium on Electrochem-pacitors, vol. 95-25, The Electrochemical Society, Pennington,95, p. 171.nning, M.A. Vannice, Carbon 31 (1993) 721.uri, P.L. Walker (Eds.), Chemistry and Physics of Carbon, vol.cel Dekker, New York, 1970, p. 191.J. Power Sources 109 (2002) 403.

hida, I. Tanahashi, A. Nishino, Carbon 28 (1990) 611.imoto, K. Hiratsuka, Y. Sanada, K. Kurihara, J. Power Sources96) 239.sieh, H. Teng, Carbon 40 (2002) 667.imoto, K. Hiratsuka, Y. Sanada, K. Kurihara, T. Jimbo, Pro-s of the 183rd Meeting of the Electrochemical Society, vol.Honolulu, Hawaii, 1993, p. 49.erbrugge, B.J. Koch, J. Electrochem. Soc. 146 (1999) 833.onway, Proceedings of the Symposium on Electrochemical

tors, vol. 95-29, The Electrochemical Society, Pennington, NJ,. 15.

ikawa, A. Sakamoto, M. Morita, Y. Matsuda, K. Ishida, J. Powers 60 (1996) 233.

ma, X. Liu, T. Osaka, Y. Ushio, Y. Sawada, J. Power Sources96) 249.wicz, E. Frackowiak, Mol. Phys. Rep. 27 (2000) 36.hida, S. Nonaka, I. Aoki, A. Nishino, J. Power Sources 60213.aishi, H. Kurihara, A. Oya, Carbon Sci. 1 (2001) 133.do, Y.J. Kim, T. Takeda, T.H.T. Maeda, K. Koshiba, H. Hara,resselhaus, J. Electrochem. Soc. 148 (2001) A1135A1140.ell, J.E. Shields, Powder Surface Area and Porosity, 3rd ed.,an & Hall, London, 1991.sh, A. Soffer, J. Electrochem. Soc. 124 (1977) 1379.tra, A. Soffer, L. Eliad, Y. Cohen, D. Aurbach, J. Electrochem.7 (2000) 2486.cCreery, K.K. Cline, P.T. Kissinger, W.R. Heineman (Eds.),

tory Techniques in Electroanalytical Chemistry, 2nd ed., Marcel, New York, 1995, p. 293.andin, E. Yeager, Electroanal. Chem. Interfacial Electrochem.72) 257.andin, E. Yeager, Electroanal. Chem. Interfacial Electrochem.75) 313.ischer, J. Phys. Chem. 89 (1985) 4249.z, M. Hahn, O. Barbieri, J.-C. Sauter, R. Gallay, Proceedings ofh International Seminar on Double Layer Capacitors and SimilarStorage Devices, December 810, 2003, Deerfield Beach, FL,

hn, M. Baertschi, O. Barbieri, J.-C. Sauter, R. Kotz, R. Gallay,chem. Solid State Lett. 7 (2004) A33A36.H. Shi, J. Power Sources 74 (1998) 99.

R[86] J.[87] H[88] C[89] R[90] R

c

Y[91] W[92] B[93] S

tr[94] R

dd

[95] J.n

[96] J.[97] T[98] F

(2[99] R

[100] R[101] H[102] U

M[103] C

2[104] H[105] R[106] S

E[107] W[108] R

C[109] S

(Ep

[110] AS

[111] AG

[112] Ad5

[113] Ga

[114] MEP

[115] AC

[116] AO

[117] DS

[118] N[119] B

W[120] K

tr[121] Y

M[122] K

Lerol, T. Siemieniewska, Pure Appl. Chem. 57 (1985) 603.atrick, Porosity in Carbons, Halsted Press, New York, 1995., Electrochim. Acta 41 (1996) 1633.

J.A. Ritter, B.N. Popov, J. Electrochem. Soc. 146 (1999) 3639.Levies, Electrochim. Acta 8 (1963) 751.Levie, P. Delahay, C.W. Tobias (Eds.), Advances in Electro-try and Electrochemical Engineering, vol. 6, Interscience, New967, p. 329.ell, B.E. Conway, J. Power Sources 96 (2001) 57.onway, J. Electroanal. Chem. 524/525 (2002) 4.yun, C.-H. Kim, S.-W. Kim, J.-H. Kim, J. New Mater. Elec-

. Syst. 5 (2002) 289.lor, H. Marsh, E.A. Heintz, F. Rodriguez-Reinoso (Eds.), Intro-

to Carbon Technologies, Universidad de Alicante, Secretaridolicaciones, 1997, p. 167.nnet, R.C. Bansal, M.J. Wang, Carbon Black Science and Tech-

, 2nd ed., Marcel Dekker, New York, 1993.eng, T.R. Jow, J. Power Sources 62 (1996) 155.ka, X. Liu, M. Nojima, J. Power Sources 74 (1998) 122.k, M. Dolata, E. Grivei, N. Probst, J. Appl. Electrochem. 31845.

hner, S. Muller, A. Wokaum, Carbon 40 (2002) 307.ekala, J. Mater. Sci. 24 (1989) 3221.bstle, M. Wiener, J. Fricke, J. Porous Mater. 10 (2003) 213.cher, R. Saliger, V. Bock, R. Petricevic, J. Fricke, J. Porous4 (1997) 281.mitt, H. Probstle, J. Fricke, J. Non-Cryst. Solids 285 (2001)

bstle, C. Schmitt, J. Fricke, J. Power Sources 105 (2002) 189.icevic, M. Glora, J. Fricke, Carbon 39 (2001) 857.ibano, S. Berthon, J.L. Ginoux, P. Achard, Extended Abstracts:rbon98, Strasbourg, France, 1998, pp. 841842.G. Reichenaur, J. Fricke, Carbon 40 (2002) 2955.

iger, G. Reichenauer, J. Fricke, in: K.K. Unger, et al. (Eds.),terization of Porous Solids V, Elsevier, 2000, p. 381.oon, Y. Korai, I. Mochida, H. Marsh, F. Rodriguez-ReinosoSciences of Carbon Materials, Universidad de Alicante, 2000,

hino, Proceedings of the 183rd Meeting of the Electrochemical, vol. 93-23, Honolulu, Hawaii, 1993, p. 1.un, M. Bartsch, B. Schnyder, R. Kotz, O. Haas, H.-G. Haubold,rigk, J. Non-Cryst. Solids 260 (1999) 1., H. Marsh, E.A. Heintz, F. Rodriguez-Reinoso (Eds.), Intro-to Carbon Technologies, Universidad de Alacante, 1997, p.

enkins, K. Kawamura, Polymeric CarbonsCarbon Fibre, Glassar, Cambridge University Press, Cambridge, 1976.

Sullivan, M. Bartsch, R. Kotz, O. Haas, Proceedings of thechemical Socioety, vol. 96-25, The Electrochemical Society,gton, NJ, 1996, p. 192.un, M. Bartsch, B. Schnyder, R. Kotz, O. Haas, A. Wokaun,40 (2002) 375.

un, M. Bartsch, O. Merlo, B. Schnyder, B. Schaffner, R. Kotz,s, A. Wokaun, Carbon 41 (2003) 759.ata, P. Haring, O. Haas, R. Kotz, H. Siegenthaler, Electrochem.tate Lett. 2 (1999) 33.odriguez, J. Mater. Res. 8 (1993) 3233.

Enaney, Prace Naukowe Insttutu Chemii i Technologii Nafty iPolitechniki Wroclawskiej, Konferencje 57 (2002) 10.n, K.K. Jeon, J.K. Heo, S.C. Lim, D.J. Bae, Y.H. Lee, J. Elec-. Soc. 149 (2002) A1058A1062.

ee, K.H. An, S.C. Lim, W.S. Kim, H.J. Jeong, C.-H. Doh, S.-I.New Diamond Frontier Carbon Technol. 12 (2002) 209.n, W.S. Kim, Y.S. Park, J.-M. Moon, D.J. Bae, S.C. Lim, Y.S.

.H. Lee, Adv. Funct. Mater. 11 (2001) 387.


[123] E. Frackowiak, K. Jurewicz, S. Delpeux, F. Beguin, L.M. Liz-Marzan,M. Giersig (Eds.), Low-dimensional Systems: Preparation, and SomeApplications, Kluwer Academic Publishers, 2003, p. 305.

[124] E. Frackowiak, K. Jurewicz, K. Szostak, S. Delpeux, F. Beguin, FuelProcess. Technol. 77/78 (2002) 213.

[125] E. Frackowiak, S. Delpeux, K. Jurewicz, K. Szostak, D. Cazorla-Amoros, F. Beguin, Chem. Phys. Lett. 361 (2002) 35.

[126] E. Frackowiak, F. Beguin, Carbon 39 (2001) 937.[127] F. Frackowiak, K. Metenier, V. Bertagna, F. Beguin, Appl. Phys. Lett.

77 (2000) 2421.[128] C. Niu, E.K. Sichel, R. Hoch, D. Moy, H. Tennent, Appl. Phys. Lett.

70 (1997) 1480.[129] Q. Jiang, M.Z. Qu, G.M. Zhou, B.L. Zhang, Z.L. Yu, Mater. Lett. 57

(2002) 988.

[130] K.H. An, W.S. Kim, Y.S. Park, Y.C. Choi, S.M. Lee, D.C. Chung, D.J.Bae, S.C. Lim, Y.H. Lee, Adv. Mater. 13 (2001) 497.

[131] M. Hughes, G.Z. Chen, M.S.P. Shaffer, D.J. Fray, A.H. Windle, Chem.Mater. 14 (2002) 1610.

[132] J.H. Chen, W.Z. Li, D.Z. Wang, S.X. Yang, J.G. Wen, Z.F. Ren, Carbon40 (2002) 1193.

[133] Ch. Emmenegger, P. Mauron, A. Zuttel, Ch. Nutzenadel, A.Schneuwly, R. Gallay, L. Schlapbach, Appl. Surf. Sci. 162/163 (2000)452.

[134] J.H. Chen, W.Z. Li, Z.P. Huang, D.Z. Wang, S.X. Yang, J.G. Wen,Z.F. Ren, Proceedings of the 197th Meeting of the ElectrochemicalSociety, Toronto, Canada, 2000.

[135] H. Tamai, K. Masayuki, M. Morita, H. Yasada, Electrochem. SolidState Lett. 6 (2003) A214A217.

Carbon properties and their role in supercapacitorsIntroductionEnergy storage in electrochemical double-layer capacitorsEDLC constructionElectrode material characteristicsCarbon structure and formEngineered carbonsActivationElectrical propertiesCarbon surface functionalitiesDouble-layer capacitance of carbon materials

Carbon surface-area and porosityCarbon formsCarbon blacksCarbon aerogelsCarbon fibresGlassy carbons

Carbon nanostructures

SummaryAcknowledgementsReferences

journal of power sources 157 (2006) 11–27

Documents

role of carbon

carbon surfacearea

carbon electrodes

carbon processing

carbon powders

carbon capacitancesroduction

carbon forms21a

carbon structure