comprehensive inorganic chemistry ii || carbon

47
7.13 Carbon P Serp, Universite ´ de Toulouse, Toulouse, France ã 2013 Elsevier Ltd. All rights reserved. 7.13.1 Introduction 323 7.13.2 Formation of Carbon and Graphite Materials 324 7.13.2.1 Gas Phase 327 7.13.2.2 Liquid Phase 328 7.13.2.3 Solid Phase 329 7.13.3 Structure and Properties of Carbon and Graphite Materials 330 7.13.3.1 Structure and Texture of Carbon and Graphite Materials 330 7.13.3.1.1 Structure of carbon and graphite materials 330 7.13.3.1.2 Macrotexture of carbon and graphite materials 334 7.13.3.1.3 Microtexture of carbon and graphite materials 335 7.13.3.1.4 Nanotexture of carbon and graphite materials 338 7.13.3.2 Bulk and Physical Surface Properties 339 7.13.3.2.1 Bulk properties 339 7.13.3.2.2 Physical surface properties 340 7.13.4 Surface Chemistry of Carbon Materials 341 7.13.5 Carbon and Graphite Materials for Catalysis 345 7.13.5.1 Activated Carbons 347 7.13.5.2 Carbon Blacks 347 7.13.5.3 Graphite and Graphitized Material 348 7.13.5.4 Activated Carbon Fibers 348 7.13.5.5 Carbon Nanostructures 348 7.13.5.6 Carbon Aerogels 349 7.13.5.7 Glassy Carbons 349 7.13.5.8 Carbon Molecular Sieves 349 7.13.6 Carbon as Catalyst 349 7.13.6.1 Influence of Carbon Properties on Catalysis 350 7.13.6.2 Examples of Applications 351 7.13.6.2.1 Oxidative dehydrogenation 351 7.13.6.2.2 Dehalogenation 354 7.13.7 Carbon as Catalyst Support 355 7.13.7.1 Role of Carbon Properties on Performances of the Support 355 7.13.7.1.1 Surface area and porosity 355 7.13.7.1.2 Surface chemical properties 357 7.13.7.1.3 Inertness 358 7.13.7.2 Examples of Applications 359 7.13.7.2.1 Hydrotreating reactions 359 7.13.7.2.2 Selective hydrogenation reactions of a-b-unsaturated aldehydes 361 7.13.7.2.3 Photocatalysis 362 7.13.8 Conclusion 364 References 365 7.13.1 Introduction The word ‘carbon’ is derived from the Latin ‘carbo,’ which to the Romans meant charcoal. Carbon, in the form of charcoal, is a material discovered prehistorically and was familiar to many ancient civilizations. Carbon, in the form of diamonds, has been used since 500 BC in India, and was probably known as early as 2500 BC in China. 1 Today, the use of this word should refer to the chemical element number 6 of the periodic table of elements. 2 The element carbon is the fourth most abundant in the solar system, after hydrogen, helium, and oxygen, and is found mostly in the form of hydrocarbons and other compounds. On earth, carbon is present in the atmosphere (CO 2 ), in the biosphere (organic carbon), in the hydrosphere (hydrogencarbonate ions), and in the earth’s crust, including the pedosphere and the lithosphere (Figure 1). The biogeochemical cycle by which carbon is exchanged among these media is called the carbon cycle, a complex series of processes through which all of the carbon atoms in existence rotate, maintaining a global constant concentration. Carbon is found in the terrestrial biosphere reservoir mostly in the form of compounds, and only two polymorphs Comprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-097774-4.00731-2 323

Upload: p

Post on 14-Dec-2016

240 views

Category:

Documents


14 download

TRANSCRIPT

Page 1: Comprehensive Inorganic Chemistry II || Carbon

Co

7.13 CarbonP Serp, Universite de Toulouse, Toulouse, France

ã 2013 Elsevier Ltd. All rights reserved.

7.13.1 Introduction 3237.13.2 Formation of Carbon and Graphite Materials 3247.13.2.1 Gas Phase 3277.13.2.2 Liquid Phase 3287.13.2.3 Solid Phase 3297.13.3 Structure and Properties of Carbon and Graphite Materials 3307.13.3.1 Structure and Texture of Carbon and Graphite Materials 3307.13.3.1.1 Structure of carbon and graphite materials 3307.13.3.1.2 Macrotexture of carbon and graphite materials 3347.13.3.1.3 Microtexture of carbon and graphite materials 3357.13.3.1.4 Nanotexture of carbon and graphite materials 3387.13.3.2 Bulk and Physical Surface Properties 3397.13.3.2.1 Bulk properties 3397.13.3.2.2 Physical surface properties 3407.13.4 Surface Chemistry of Carbon Materials 3417.13.5 Carbon and Graphite Materials for Catalysis 3457.13.5.1 Activated Carbons 3477.13.5.2 Carbon Blacks 3477.13.5.3 Graphite and Graphitized Material 3487.13.5.4 Activated Carbon Fibers 3487.13.5.5 Carbon Nanostructures 3487.13.5.6 Carbon Aerogels 3497.13.5.7 Glassy Carbons 3497.13.5.8 Carbon Molecular Sieves 3497.13.6 Carbon as Catalyst 3497.13.6.1 Influence of Carbon Properties on Catalysis 3507.13.6.2 Examples of Applications 3517.13.6.2.1 Oxidative dehydrogenation 3517.13.6.2.2 Dehalogenation 3547.13.7 Carbon as Catalyst Support 3557.13.7.1 Role of Carbon Properties on Performances of the Support 3557.13.7.1.1 Surface area and porosity 3557.13.7.1.2 Surface chemical properties 3577.13.7.1.3 Inertness 3587.13.7.2 Examples of Applications 3597.13.7.2.1 Hydrotreating reactions 3597.13.7.2.2 Selective hydrogenation reactions of a-b-unsaturated aldehydes 3617.13.7.2.3 Photocatalysis 3627.13.8 Conclusion 364References 365

7.13.1 Introduction

The word ‘carbon’ is derived from the Latin ‘carbo,’ which to

the Romans meant charcoal. Carbon, in the form of charcoal,

is a material discovered prehistorically and was familiar to

many ancient civilizations. Carbon, in the form of diamonds,

has been used since 500 BC in India, and was probably known

as early as 2500 BC in China.1 Today, the use of this word

should refer to the chemical element number 6 of the periodic

table of elements.2 The element carbon is the fourth most

abundant in the solar system, after hydrogen, helium, and

mprehensive Inorganic Chemistry II http://dx.doi.org/10.1016/B978-0-08-09777

oxygen, and is found mostly in the form of hydrocarbons

and other compounds. On earth, carbon is present in the

atmosphere (CO2), in the biosphere (organic carbon), in the

hydrosphere (hydrogencarbonate ions), and in the earth’s crust,

including the pedosphere and the lithosphere (Figure 1). The

biogeochemical cycle by which carbon is exchanged among

these media is called the carbon cycle, a complex series of

processes through which all of the carbon atoms in existence

rotate, maintaining a global constant concentration.

Carbon is found in the terrestrial biosphere reservoir mostly

in the form of compounds, and only two polymorphs

4-4.00731-2 323

Page 2: Comprehensive Inorganic Chemistry II || Carbon

Atmosphere

Ocean

Terresterial biosphere

39120

2190 754

Figure 1 Magnitude of carbon reservoir (actively cycling CO2)in gigatons of carbon.

324 Carbon

(or allotropes) of carbon are found on earth as minerals:

natural graphite and diamond. Many of these natural com-

pounds, such as coals, hydrocarbons, and biomass, are essen-

tial tomodern human life for the production, often catalytic, of

synthetic organic and inorganic carbons.

Both organic and inorganic carbons play a central role in

catalysis. Organic carbon compounds form the huge and very

complex discipline of organic chemistry, and they are, in most

catalytic applications, the substrates and the products of the

process under consideration. In homogeneous catalysis, car-

bon is often the main constituent of the organic ligands sur-

rounding the metallic center. In enzymatic catalysis, it

constitutes the backbone of the active species. In heteroge-

neous catalysis, carbon materials are unique catalyst supports,

allowing the anchoring of the active phase, and can also act as

catalysts or catalyst poisons (carbon deposits) by themselves.3

As heterogeneous catalysis is the subject of the present volume,

in this chapter we focus on the polymorphs of carbon and not

on its compounds.

The capability of a chemical element to combine its atoms

to form such polymorphs is not unique to carbon. Other

elements in the fourth column of the periodic table (silicon,

germanium, and tin) also have this characteristic. However,

carbon is unique in the number and the variety of its poly-

morphs. A result of this diversity is that the carbon terminology

can be confusing for the nonspecialist. These allotropes

are composed entirely of carbon but have different physical

structures and, exclusively for carbon, have different names:

graphite, diamond, lonsdaleite, and fullerene, among others.

Additionally, carbon as a solid denotes all natural and

synthetic substances consisting mainly of atoms of the element

carbon, such as single crystals of diamond and graphite, as well

as the full variety of carbon and graphite materials. The termi-

nology used so far is mainly based on technological tradition

and on the standardized characterization methods derived

from decades of industrial experience. Because of the increas-

ing interdisciplinary importance of this group of materials in

science and technology, it is obvious that clear definitions of

the corresponding terms are required. In order to clarify the

terminology, we will refer to the recommended terminology

for the description of carbon as a solid (IUPAC Recommenda-

tions 1995).2 As already stated, when used by itself, the term

‘carbon’ should refer only to the element. For a description of

the various types of carbon as a solid, the latter should be used

only in combination with an additional noun or a clarifying

adjective (i.e., carbon material, carbon fiber (CF), graphitic

carbon, pyrolitic carbon, etc.). These materials have an sp2

atomic structure, and can be structurally in a nongraphitic

(carbon materials) or graphitic state (graphite materials).

Other materials with an sp3 atomic structure are, by common

practice, called by the name of their allotropic form, that is,

diamond, lonsdaleite, etc., and are not commonly referred to

as ‘carbon’ materials, although, strictly speaking, they are. The

existence of carbyne or the linear acetylenic carbon of chemical

structure –(C�C)n– with sp orbital hybridization as a carbon

allotrope is still controversial.4 The principal carbon poly-

morphs and polytypes are represented in Figure 2. As we

discuss in this chapter, these materials are very different both

in structure and in properties, and their structural as well as

surface chemistry is extremely complex.

As a consequence, for many years, carbon science was a very

specialized field, considered by many to be too complicated.

More recently, carbon science has gained high visibility5 with

the discovery of fullerenes in 19856 and the first high-resolution

transmission electronmicroscopy (HR-TEM) observations of car-

bon nanotubes (CNTs) in 1991.7 This visibility has been further

heightened by the 1996 Nobel Prize in Chemistry awarded to

R.F. Curl, H. Kroto, and R.E. Smalley for their discovery of

fullerenes and the 2010 Nobel Prize in physics awarded to

A. Geim and K. Novoselov for groundbreaking experiments

regarding the two-dimensional (2D) material graphene.

The catalytic behavior of solid carbons depends of course on

their surface properties, but these properties are, to a large extent,

a direct consequence of their bulk characteristics. Therefore in

this chapter, after a brief overview of solid carbon formation, we

discuss their structure before focusing on their physicochemical

surface properties. Finally, through selected case studies we high-

light the impact of these surface physicochemical properties on

the performances of carbon-based catalysts.

7.13.2 Formation of Carbon and Graphite Materials

As the unwanted formation of carbon deposits in catalyst

deactivation processes as well as the modification of primary

carbon surfaces during chemical treatments are frequent prob-

lems encountered in catalytic carbon chemistry, it is appropri-

ate to discuss some of the general mechanistic ideas of solid

carbon formation. All carbon and graphite materials are

formed in the gas, liquid, or solid phase; and the conditions

of their formation dictate their physicochemical properties.

The structure of these materials is also known to depend on

the temperature they have experienced. In order to obtain such

materials, in which carbon atoms are the principal constituent,

it is necessary to heat treat a carbon precursor, in an inert

atmosphere, a process usually called carbonization. The car-

bonization process (Figure 3), also known as ‘pyrolysis,’ can be

defined as the step in which the organic precursor is trans-

formed into a material that is essentially all carbon. Carboni-

zation is basically a heating cycle. The precursor is heated

slowly in a reducing or inert environment, over a temperature

range that varies with the nature of the precursor and may

extend to 1573 K. The organic material is decomposed into a

carbon residue and volatile compounds diffuse out to the

atmosphere. The process is complex and several reactions

may take place at the same time such as dehydrogenation,

Page 3: Comprehensive Inorganic Chemistry II || Carbon

spLinear geometry

sp2

Trigonal geometry sp3

Tetrahedral geometry

Carbyne Fullerenes Graphite Diamond Polymorphs

Polytypes

Carbolites

Family of molecules,isomers

rh-graphitea

HOPGb

kishturbostraticgraphene

Glassy carbon a-C:Hc

a-DLCd

Coke

Carbonaceous depositNanotubesNanofibersNanocarbon

Fullerene black Carbon black

Activated carbon

Chains Moleculesnanoarrays

Blocks

Molecular crystal Crystalline

a Rhombohedral graphite. b Highly orientated graphite. c Amorphous hydrogenated carbon. d Amorphous diamond-like carbon.

Planar sheetsBent sheets Ribbons, chains

Molecular crystal AmorphousAmorphous Crystallinenanocrystalline nanocrystalline

Figure 2 Polymorphs and polytypes of carbon.

Carbon 325

condensation, and isomerization. The carbon content of the

residue is a function of the nature of the precursor and the

pyrolysis temperature. It usually exceeds 90 wt% at 1173K and

99 wt% at 1573 K. The diffusion of the volatile compounds to

the atmosphere is a critical step and must occur slowly to avoid

disruption and rupture of the carbon network. As a result,

carbonization is usually a slow process. Its duration may vary

considerably, depending on the composition of the end prod-

uct, the type of precursor, and the thickness of the material,

among other factors.

The range of hydrocarbon feedstock used as carbon pre-

cursors is also dictated by these conditions; and seemingly

subtle changes often produce profound structural effects.

These are briefly discussed below.

In the beginning of pyrolysis, aliphatic molecules with low

relative molecular weight and subsequently relative low-

molecular-weight aromatics are released as gases, mainly

because the C–C bonds of the precursors are weaker than the

C–H ones. Cyclization and aromatization proceed in the resi-

due still associated with the release of low-molecular-weight

hydrocarbons, followed by the polycondensation of aromatic

molecules. At around 873 K, heteroatoms such as oxygen and

nitrogen are released as CO2, CO, N2, and HCN, together

with methane. At this stage, the nature of the residue

(gas, liquid, or solid) depends on the nature of the carbon

precursor. Above 1273 K, outgassing is mainly H2 because of

the polycondensation of aromatics. The residue may be called

a carbonaceous residue that still contains hydrogen. Above

1573 K, the residue is considered a carbon material. These

reaction processes during the changes from organic precursors

to inorganic materials, that is, pyrolysis, cyclization, aromati-

zation, polycondensation, and carbonization, depend strongly

on the carbon precursors and also on the conditions of the heat

treatment (temperature, heating rate, etc.). These processes

usually overlap with each other. Therefore, the whole process

to the final carbon or graphite material is often called ‘carbon-

ization.’ Details of the carbonization process of standard pre-

cursors can be found in the literature.8 In general, all processes

that homogeneously generate carbon from molecular precur-

sors and from homogeneous activation lead to spherical parti-

cles made from graphene layers, as in the case of carbon blacks

(CBs). The possible significant presence of curved sp2 nano-

structure in solid carbons such as soot, CBs, or nanotubes,

which can result in significant p-orbital rehybridization,9 can

have a significant impact in adsorption and catalysis.10 Solid

carbons with larger continuous graphene layers can be

obtained from pyrolytic polymerization of prepolymerized

hydrocarbons. Thus, polyacrylonitrile carbonization will pro-

duce CFs, whereas carbonization of biopolymers as coconut

shell, wood coal, or rice husks can be used to produce activated

carbon (AC). In this latter case, the structure of the carbon

precursor is preserved in the final product. Another important

aspect related to the nature of the carbon precursor is its

dependency on the efficiency of carbon formation. The idea

that the chemical structure of the molecule should be relevant

for solid carbon formation (aromatic molecules are better than

Page 4: Comprehensive Inorganic Chemistry II || Carbon

Heat treatmenttemperature (K)

473

773

873

1073

1773

2273

3273

Phase of reaction

Vapor Solid Liquid

Gas

volatilization

Chemical and physical Molecular

structures

Organic materials

Carbon materials

Graphite materials

Carbonaceousmaterials

Carb

on materials

Radicalpyrolysis

Cross-linking

Aromatizationpolycondensation

Coking

H2OLow mol. paraffins or olefinsLow mol. aromatic molecules

CH4, CO, NO2,H2S, CO2,H2, etc.

H2,CO, CO2,H2S,etc.

H2S,HCN,CS2,N2,etc.

H2,H2S,N2

Main chain rearrangement

Aromatization, CondensationPolymerization, Cross-linkingCoking

DevolatilizationCrack nucleationStacking startLoss of viscosity (inorganicmaterials)

Removal of heteroatomsDehydrogenationMicropore nucleationLa* increasing

Removal of heteroatomsLc* increasingReducing micropores

Removal of inorganic materialsFormation of 3D graphitic structures

Organic material

* La and Lc are lattice parameters

changes

Figure 3 Carbonization process.

326 Carbon

aliphatic ones) is incorrect in all situations where a pool of

primary building blocks is formed (in flames, under oxidative

polymerization, and under physical activation). However, it is

confirmed that a decrease in the average C–C bond energy

allows easier chemical fragmentation, which enhances the

rate of carbon formation.11

During the carbonization process, a fundamental scheme of

preferred orientation of basic structure units, that is, hexagonal

carbon layers, is established and is called nanotexture (see

Section 7.13.3.1.4). Therefore, this carbonization process is

the most important step for the preparation of carbon mate-

rials. However, carbon layers are generally still small and in

many cases are stacked randomly (turbostratic structure). In

order to make these carbon layers stack with high (graphitic)

regularity, heat treatment at high temperatures, often above

2773 K, is required. In some cases, graphitic structures are

developed and so this process has been called graphitization.

A structural model for CFs during graphitization12 is given in

Figure 4(a), and the effect of graphitization on CNT structure13

is shown in Figure 4(b). However, graphitization does not

occur always because the development of the graphitic struc-

ture depends strongly on the nanotexture formed during the

carbonization process. The nongraphitizable materials (or iso-

tropic carbons), as for example glassy carbons, are nongraphi-

tic carbons which cannot be transformed into graphitic carbon

solely by high-temperature treatment up to 3300 K under

atmospheric or lower pressure. A common result of heat treat-

ment at high temperatures (typically>1100 K) is the decrease

of the surface area of the material, due to cross-linking of

carbon atoms that reduces space between atoms and essentially

closes off porosity to an adsorptive gas.

We have seen that the carbon structure is produced through

a series of carbonization, graphitization, gasification, and suc-

cessive modification, as summarized in Figure 3. The first two

processes are included in the transformation of organic sub-

strates to graphitic materials through carbonaceous intermedi-

ates. The last two are the posttreatments of carbon into the

active surface materials. Such transformations include reaction

chemistry and phase change from the organic precursors to

solid carbon, defining the structure of carbon, and hence

their properties. Such series of chemical and physical trans-

formations must be governed by the properties of organic

substrates, solvent materials, and the presence of a catalyst.

Indeed, the catalyst is a very important tool to control the

molecular and compositional chemistry for the transformation

process of organic substrates into carbon, throughout carbo-

naceous intermediates.14 The catalyst can control the reactions

in terms of reaction chemistry and phases involved in all stages

of carbon preparation:

i) Catalytic synthesis of feed and intermediate for carbon, as

feeds for commercial carbons, are produced in the petro-

leum refining and coal tar industries;

ii) Catalytic carbonization where the catalyst governs and con-

trols the reaction phase as well as the chemistry involved in

carbonization;

Page 5: Comprehensive Inorganic Chemistry II || Carbon

Carbon 327

iii) Catalytic carbon growth, particularly for CNTs and nano-

fibers that currently attract the broad interest of researchers

and engineers to develop nanocarbon materials for much

higher performance or new applications of carbon;

iv) Catalytic activation of carbons, as gasifying catalysts and

oxidants are important to introduce pores regardless of the

type of carbon material involved; and

v) Catalytic graphitization, as highly graphitic materials, are

always to be prepared at lower temperatures.

2000 K

1473 KAs-produceda b c

d e f

1773 K

2273 K2123 K2023 K

1700

(a)

(b)

130110

Figure 4 (a) Structural model for CFs during graphitization – dotsrepresent heteroatoms (reproduced from Bunsell, A.R. FibreReinforcements for Composite Materials; Elsevier Science PublishersB.V: Amsterdam, 1988, pp. 120, with permission). (b) Evolution of CNTstructure upon graphitization. Scale bar¼5 nm (reproduced from Mattia,D.; Rossi, M. P.; Kim, B. M.; Korneva, G.; Bau, H.H.; Gogotsi, Y. J. Phys.Chem. B 2006, 110, 9850–9855, with permission from AmericanChemical Society).

FullerenesCondensationCondensation

Pyro

Carbo(s

Carbon nanotubes,carbon filaments

CVD on metallicsurfaces

Hydrocsourc

Pyro

Figure 5 Solid carbon materials formed from the gas phase. Adapted from SHoboken, NJ, 2009.

7.13.2.1 Gas Phase

Figure 5 summarizes the main carbon products that can be

produced in gas-phase reactions from a wide variety of carbon-

containing gases such as CO, CH4, C2H2, C3H8, and C6H6 or

volatile products of coal or biomass pyrolysis. Highly divided

carbonmaterials as soot or CBs, which play an important role in

electrocatalysis in devices such as batteries, supercapacitors, and

fuel cells, can be formed by heat treatment of hydrocarbons in

the absence of oxygen (thermal decomposition) or in its pres-

ence (flames). The mechanism of their formation in flames is

basically identical in premixed flames (the oxidizer has been

mixed with the fuel before it reaches the flame front) and in

diffusion flames (the oxidizer combines with the fuel by diffu-

sion). However, as the different chemical stages are well sepa-

rated, from a spatial point of view, in premixed flames (the

oxidizer combines with the fuel by diffusion), most investiga-

tions deal with this type of flame. In the oxidation zone of the

flame, a part of the hydrocarbon is burnt out while another part

undergoes complex reactions leading to polyacetylenes and later

to polycyclic hydrocarbons with lateral chains. In the same

region, ions are formed by chemiionization. In the zone of

formation of carbon particles, polycyclic hydrocarbons are dehy-

drogenated and give polyaromatic hydrocarbons; their partial

pressure increases until they reach a supersaturation state high

enough to inducenucleation of liquidmicrodroplets. Themicro-

droplets are formed by homogeneous nucleation as well, prob-

ably, as the nucleation on the positive ions formed in the

oxidation zone of the flame; they are converted by growth,

association, and chemical transformation into solid particles.

In thermal systems, the mechanisms involved are identical

except that nucleation on ions may be disregarded. In both

systems, nucleation of microdroplets is a fast and discontinuous

phenomenon. These carbon materials are relatively disordered

but not amorphous, nongraphitic, and nongraphitizable. It is

important to distinguish betweenCB that ismanufactured under

controlled conditions and soot that is randomly formed. These

two carbonmaterials can be distinguished on the basis of tar, ash

content, and impurities. Attempts in the literature to create a

general term, ‘aciniform carbon,’ which would cover both CB

and soot, are not yet generally accepted.

The much more ordered, nongraphitic but graphitizable

pyrolytic carbon, known as ‘pyrocarbon’ for short, is obtained

by chemical vapor deposition of carbon precursors under

oxygen-free conditions on a relatively ‘inert’ substrate. In the

lysis

Gas phasepyrolysis

n blackoot)

Pyrolyticcarbon

CVD on ceramicsurfaces

PolymerizationLC, SC

arbone GC Decomposition

(evaporation)

lysis

Polymerization

Decomposition(evaporation)

erp, P.; Figueiredo, J. L. Carbon Materials for Catalysis, J. Wiley & Sons:

Page 6: Comprehensive Inorganic Chemistry II || Carbon

>2273 K

GraphiteCarbon fibers(mesophase)Cokes

<2273 K Extrusion(orientation)

PolymerizationPolymerization

GCThermoplasticpolymers LCDecompositionDecomposition

CarbonizationCarbonization

Figure 6 Solid carbon materials formed from the liquid phase. Adaptedfrom Serp, P.; Figueiredo, J. L. Carbon Materials for Catalysis, J. Wiley &Sons: Hoboken, NJ, 2009.

328 Carbon

industrial steam cracking of ethane to produce ethylene, pyro-

lytic carbon (coke) deposits on the walls of the reaction tubes,

limiting heat transfer, and ultimately blocking flow.15 The

formation mechanism of deposits of pyrolytic carbon is a

complex process involving homogeneous reactions in the gas

phase, heterogeneous reactions at the substrate surface, and

subsequent dehydrogenation. All three processes are relevant

for the final carbon structure and other physical properties.

Surfaces play an important role in these processes, providing

sites for nucleation and growth as well as the adequate envi-

ronment for further reactions such as dehydrogenation and

ordering of the deposited materials. As the mechanisms

involve the formation of free radicals, it is expected that pyro-

lytic coke can be important only in the deactivation of catalysts

used in high temperature processes, such as steam reforming of

hydrocarbons. The structure of crystalline graphite can be

achieved upon simple heat treatment above ca. 2773 K.

On amore ‘reactive’metallic surface such as Fe, Co, orNi, the

deposition of carbon precursors typically results in hydrocarbon

decomposition, carbon dissolution, intra- and/or suprametal

diffusion, and precipitation in the form of nanotubes or fila-

ments. The deactivation of metal-supported catalysts by carbon

deposition has been reviewed in comprehensive books.16 It is

generally admitted that carbon may (1) chemisorb strongly

onto the metal to produce a monolayer, or to be physically

adsorbed by multilayers, thus hindering the access of reactants

to the metal surface sites; (2) fully encapsulate each metal par-

ticle and thereby completely deactivate the catalyst; and (3) plug

themicro- andmesopores so that the access of reactants tomany

crystallites inside the pores is suppressed. Finally, in the worst

cases, carbon filaments may build up in the catalyst grain poros-

ity to such an extent that they stress and fracture the support

material, ultimately causing disintegration of catalyst pellets and

cover the reactor walls.17

The novel aspect of this process of filamentous carbon for-

mation is the emergence of curvature of sp2-bonds in the grow-

ing graphene layers and at the nanotube tip due to the presence

of a metallic nanoparticle. The formation of carbon nuclei by

carbon precipitation onto the surface of metal particles is a

critical step, as this step should be at the origin of the

formation of curved sp2 carbon. The catalyzed formation of

carbon nanofibers (CNFs) and nanotubes over supported

metal nanoparticles has received much attention because it

may provide low-cost, large-scale synthesis of these materials,

and it is important to inhibit this in order to prevent breakdown

of industrial steam reforming catalysts for the production of

hydrogen and synthesis gas. Despite numerous studies, the

growth mechanisms are still subject to intense debate.18

7.13.2.2 Liquid Phase

Figure 6 shows the solid carbons obtained from liquid-phase

reactions using thermoplastic polymers (a polymer that turns

to a liquid when heated and freezes to a very glassy state when

cooled sufficiently). Natural precursors such as bituminous

coals or synthetic ones such as polyvinyl chloride, –(CH2Cl2–)n,

can be used. Mechanisms involved in carbonization in the

gas, liquid, and solid phases are very different. During the solid

phase carbonization, all structural changes involving atom

removal and replacement within a solid lattice remain essentially

rigid throughout the process. No bulk movement of materials

occurs within the solid. Liquid-phase carbonizations are very

different indeed. It is not the case of a liquid at some stage

solidifying to give a structure as disorganized as the preceding

liquid. In fact, the opposite is true. Indeed, under the most

commonly utilized carbonization conditions, typically <2273 K

in a largely nonreactivemedium, the degree of alignment and the

mobility of emerging, growing, and coalescing carbon crystallites

is considerable, but the consequent relative orientations of the

resulting graphene layers are insufficient to achieve the perfect

crystalline structure of graphite. In fact, behind the liquid phase

carbonization is the concept of growth of aromatic, nematic,

discotic liquid crystals.

Thus, heat treatment of highly aromatic feedstocks, such as

petroleum pitch, coal-tar pitch, etc., at temperatures between

623 and 773 K leads to the formation of an optically aniso-

tropic liquid crystalline phase, ‘mesophase carbon,’ from the

isotropic parent precursor. The formation of the mesophase

can be divided into several steps. First, polycondensed aro-

matic molecules form due to thermal decomposition and ther-

mal polymerization reactions. As the molecules grow larger,

the cohesive force exceeds the translational energy, the align-

ment of lamellar molecules results in the formation of nematic,

discotic liquid crystals. The generation of the mesophase

within a pyrolyzing system will only occur if several constraints

are verified, for example,: (1) the intermolecular reactivity of

constituent pyrolysis molecules is constrained to limited

molecular growth of polynuclear aromatic molecules, not

exceeding 900 amu on average; (2) the system must remain

fluid up to temperatures of 673–723 K such that the polynu-

clear aromatic molecules have sufficient mobility to establish

the liquid crystal system; and (3) the intermolecular reactivity

of constituent molecules of the mesophase is also constrained

to facilitate growth, coalescence, and movement within the

fluid liquid crystal system prior to solidification. The meso-

phase adopts a spherical shape to minimize surface energy.

Although chemically stable (relatively) at the temperature of

formation, increasing the high temperature treatment pro-

motes cross-linking into a macromolecular structure. A meso-

phasic order is the direct result of liquid crystal formation in

the fluid phase followed by molecular weight growth and

solidification in the ordered state to produce a structure with

long range order. Thus, the solids that emerge from a carbon-

izing liquid are extremely well organized and constitute the

graphitizable carbons (also called anisotropic carbons), that is,

Page 7: Comprehensive Inorganic Chemistry II || Carbon

Carbon 329

carbon capable of producing 3D x-ray diffraction (XRD) lines

(hkl) of the graphite lattice upon heating above 2273 K. Coke

is a solid with a high content of the element carbon and is

structurally in the nongraphitic state. It is produced by pyrol-

ysis of organic material which has passed, at least in part,

through a liquid or liquid-crystalline state during the carbon-

ization process.19 The manufacture of pitch CFs uses carbon

mesophase precursors in conjunction with the fiber melt spin-

ning process. These carbon mesophases are textured anisotropic

viscoelastic liquid crystalline materials formed by disc-like aro-

matic molecules. Therefore, this material is a discotic, nematic,

thermotropic liquid crystal, which exhibits orientational order

and positional disorder. The melt spinning process consists of

a flow sequence that induces unique textural transformations in

the mesophase. Fiber melt spinning of carbon mesophase pre-

cursors usually leads to a variety of cross-sectional fiber textures

that give specific properties to the fibers and its composites.

7.13.2.3 Solid Phase

In comparison with liquid-phase carbonization, carboniza-

tions in the solid phase do not allow any internal recrystalliza-

tion (although it exists catalytic graphitization), the resultant

structures being quite disordered but never disorganized and

constitute what are termed the nongraphitizable or isotropic

carbons. The carbon precursor, almost always being a macro-

molecular system, essentially ‘decomposes’ as the high temper-

ature treatment increases, the process being accompanied by

the evolution of gases and liquids of low molecular weight

resulting from the decomposition processes. Consequently,

the resultant carbon is a ‘pseudomorph’ of the parent material,

having more or less the same original shape but it is now of a

lower bulk density. Figure 7 summarizes the carbon formation

processes taking place in the solid phase, with thermosetting

carbon precursors such as low-rank coals, preoxidized bitumi-

nous coals and wood, or thermosetting polymers (polymer

material that cures irreversibly) such as polyvinylidene chlo-

ride, –(CHCl3–)n. The solid-phase carbonization process is a

progressive decomposition that will cease when the heat treat-

ment is stopped. Increasing further the temperature of the

treatment results in the formation of progressively more stable

intermediate structures. During the carbonization process,

when the parent macromolecular system is decomposing, the

Mostlybottlenecked

pores

Carbon molecular

sieves

Carbon fibers(activated)

Glassy carbon

Charring/activation Extrusion

(activation)

PolymerizationPolymerization

GCThermosettingpolymers SC

DecompositionDecomposition

Thermal decompositionThermal decomposition

Activatedcarbons (chars)

Figure 7 Solid carbon materials formed from the solid phase.Adapted from Serp, P.; Figueiredo, J. L. Carbon Materials for Catalysis,J. Wiley & Sons: Hoboken, NJ, 2009.

remaining carbon atoms of the macromolecular network move

short distances (probably <1 nm) within the network to posi-

tions of greater stability (e.g., formation of six-membered ring

systems), eventually creating a network of carbon atoms (with

residual hydrogen bonded to it), which constitutes the struc-

ture in such carbons. Thus, the carbonization process consists

essentially in the conversion, by progressive heating, of a 3D

organic macromolecular system (e.g., coals, woods, nutshells,

etc.) to a 3D ‘macro-atomic’ network of carbon atoms. As small

molecules, such as water, methanol, and carbon dioxide, are

eliminated from the organic system, so the resultant free radi-

cals (e.g., dangling bonds) are removed by movement of atoms

over short atomic distances to create an intermediate stable

phase of higher carbon content. It is emphasized again that this

process does not produce a continuous solid phase but one in

which spaces of nanometer dimensions result, that is, a net-

work of porosity is produced. As the carbonization tempera-

ture continues to rise so does the extent of this space network

as well as its dimensions, length, and breadth, to a maximum.

With further increases in high temperature treatment (above

about 1073 K), accessibility to adsorbed gases decreases as

cross-linking of carbon atoms reduces space between atoms

and essentially closes off porosity to an adsorptive gas. Differ-

ent precursors will decompose in their own distinctive ways

to produce a specific type of carbon.

Glassy carbon (also referred to as vitreous or polymeric

carbon) is produced by the thermal degradation of selected

polymers, resins; typically, phenolic resins or furfuryl alcohol

are used. The precursor resin is cured (cross-linked), carbon-

ized very slowly, and then heated to elevated temperatures. The

physical properties of glassy carbons are generally dependent

on the maximum heat treatment temperature, which can vary

from 873 to 3273 K. Glassy carbons have little accessible sur-

face area and a relatively low density (�1.5 g cm�3, graphite

2.26 g cm�3), which is attributed to the presence of a signifi-

cant volume of isolated ‘closed’ pores (�30%v/v). The isolated

porosity of glassy carbons can be opened by thermal or elec-

trochemical oxidation processes (activation) to give a material

with a high specific surface area.

The procedure for processing AC typically consists of a

carbonization followed by an activation of carbonaceous

material from vegetable origin. The carbonization (673–

1073 K) converts the raw materials to carbon by minimizing

the content of volatile matter and increasing the carbon con-

tent of the material. This increases the material strength and

creates an initial porous structure, which is necessary if the

carbon is to be activated. Adjusting the conditions of carbon-

ization can affect the final product significantly. An increased

carbonization temperature increases reactivity, but at the same

time decreases the volume of pores present. This decreased

volume of pores is due to an increase in the condensation of

the material at higher temperatures of carbonization, which

yields an increase in mechanical strength. After the initial

porous structure has been created by carbonization, an oxida-

tion, referred to as activation, is carried out in order to create

micropores. Activation can be carried by oxidizing gases or by

chemical activation. In activation by oxidizing gases, such as

steam activation, carbon reacts with the oxidizing agent pro-

ducing oxides of carbon (COx). These oxides diffuse out of the

carbon resulting in a partial gasification that opens pores that

were previously closed and further develops the carbons

Page 8: Comprehensive Inorganic Chemistry II || Carbon

Table 1 Selected examples of carbon precursors and theresulting product

Precursors Products

Methane Pyrolytic graphiteHydrocarbonsFluorocarbonsAcetone, etc.

Diamond-like carbonPolycrystalline diamond

Hydrocarbons (þcatalyst) Filamentous carbonRayonPolyacrylonitrile

Carbon fibers

FCC tar, coal tar, ethylene cracking tar,and vegetable oil

Carbon black

PhenolicsFurfuryl alcohol

Carbon–carbonVitreous carbon

Petroleum fractionsCoal tar pitch

Molded graphitesCarbon fibers

Biomass CoalCoal (bituminous, sub-bituminous,and lignite), peat, wood, or nutshells

Activated carbon

330 Carbon

internal porous structure. In chemical activation, the carbon is

reacted at high temperatures with a dehydrating agent that

eliminates the majority of hydrogen and oxygen from the

carbon structure. Chemical activation often combines the car-

bonization and the activation step, but these two steps may still

occur separately depending on the process.

Carbon molecular sieves are formed by the controlled

pyrolysis of suitable polymeric materials (e.g., polyvinylchlor-

ide) or petroleum pitch materials at temperatures usually

above 673 K. They have a highly porous structure with almost

uniform micropores. They are comprised of very small crystal-

lites cross-linked to yield a disordered cavity-aperture structure.

To conclude, essentially any organic material can be ther-

mally transformed to carbon (see Table 1). The carbonization

process uses heat to convert organic precursors into a carbon

polymer. Some selected precursors can then be transformed into

a 3D graphite structure or near-graphite structure. Differences in

properties of the final carbon products depend on the raw

materials used, on the extent of completion of overall chemical

and physical ordering processes, and whether the thermal trans-

formation takes place from the gas, liquid, or solid phase.

7.13.3 Structure and Properties of Carbonand Graphite Materials

Carbon atoms have three different hybrid orbitals, sp3, sp2, and

sp, and give a variety of combinations of chemical bonds.

Figure 8 illustrates how these three hybrid orbitals of carbon

atoms give a large family of organic molecules and how the

inorganic carbon or graphite material, graphite, diamond, fuller-

enes, and carbynes, can result from the extension of these organic

materials through large molecules. A recent study suggests that

such a complex chemistry can occur in a simple candle flame,

where all the four known carbon forms (diamond, graphitic,

fullerenic, and amorphous particles) have been identified.20

The two allotropes of carbon with particularly well-defined

properties are hexagonal graphite and cubic diamond. Although

both well-crystallized forms will only very rarely occur in

catalytic systems, it is important to recall some details about

their structure and properties as these also prevail in the practical

forms of carbon. The fullerenes and carbynes, which are also

well-defined forms of carbon, are very rare in their pure forms in

catalytic materials. Because of the curvature of the surface, fuller-

ene hybridization falls between graphite (sp2) and diamond

(sp3) and these new carbon allotropes are therefore of inter-

mediate, and perhaps variable, hybridization. The specific

reactivity of such a surface is of great interest in catalysis as the

presence of curved sp2 nanostructures in cokes, CBs, and of

course filamentous carbon is significant.

7.13.3.1 Structure and Texture of Carbonand Graphite Materials

7.13.3.1.1 Structure of carbon and graphite materials7.13.3.1.1.1 Graphite

The mineral graphite is one of the allotropes of carbon. It was

named by Abraham Gottlob Werner in 1789 from the Ancient

Greek grάjo, ‘to draw/write,’ for its use in pencils. Unlike

diamond, graphite is an electrical conductor, a semimetal.

Graphite is composed of series of stacked parallel layer planes

shown schematically in Figure 9, with trigonal sp2 bonding.

Within each layer plane, the carbon atom is bonded to three

others, forming a series of continuous hexagons in what can be

considered as an essentially infinite 2D molecule. The bond is

covalent (s bond) and has a short length (0.141 nm) and high

strength (524 kJ mol�1). The hybridized fourth valence elec-

tron is paired with another delocalized electron of the adjacent

plane by a much weaker van der Waals bond (a secondary

bond arising from structural polarization) of only 7 kJ mol�1

(p bond). Carbon is the only element to have this particular

layered hexagonal structure. The spacing between the layer

planes is relatively large (0.335 nm) or more than twice

the spacing between atoms within the basal plane and

approximately twice the van der Waals radius of carbon. The

stacking of these layer planes occurs in two slightly different

ways: hexagonal and rhombohedral. The most common

stacking sequence of the graphite crystal is hexagonal with

an –ABABAB– stacking order; in other words, in the vertical

direction, perpendicular to the basal plane, another layer of

hexagons is formed, shifted in such a way that there is a carbon

atom at the center of each hexagon of the top layer (Figure 9

(b)). The next layer down is at the same position as the top

layer. The crystal lattice parameters, that is, the relative position

of its carbon atoms (along the orthohexagonal axes) are:

a¼0.246 nm and c¼0.6708 nm.

In rhombohedral graphite, the stacking order is

–ABCABCABC– (Figure 9(c)). The carbon atoms in every

third layer are superimposed. Rhombohedral graphite is ther-

modynamically unstable, and can be considered as an

extended stacking fault of hexagonal graphite. It is never

found in pure form but always in combination with hexagonal

graphite, at times up to 40% in some natural and synthetic

materials. In such structures, the hexagonal carbon rings pro-

vide the delocalized electrons, allowing easy electronic conduc-

tion within the planes. Conduction over the perpendicular

planes is much lower (around three orders of magnitude smal-

ler) – this is highly anisotropic behavior. This structure also

Page 9: Comprehensive Inorganic Chemistry II || Carbon

Carbyne(Polyyne)

Carbyne(cumulene)

Polyacetylene

Ovalene

Graphite

Graphene

1,3-butadiene

1-buten-3-yne

Adamantane

Propane

Diamond

Butadiyne

sp + 2π

sp2 + π

sp3

C

Benzene

Anthracene

Fluorene

Corannulene

Fullerenes

nanotubes

Figure 8 C–C bonds form a large number of hydrocarbons and their extension to carbon families. Adapted from Inagaki, M.; Feiyu, K. Carbon MaterialsScience and Engineering: From fundamental to applications, Tsinghua University Press, 2006.

Carbon 331

creates anisotropy in other properties of graphite, such as

thermal conductivity and thermal expansion (see Table 3).

Graphite materials, such as pyrolytic graphite, carbon fiber/

carbon matrix composites (carbon–carbon), vitreous carbon,

CB, and many others, are actually aggregates of graphite crys-

tallites, in other words, polycrystalline graphite. These turbos-

tratic quasicrystalline domains, usually referred to as graphitic

crystallites, nanocrystallites, or basic structural units (BSUs),

are characterized by the distance between graphene planes

(d002) and the average sizes of the crystallites, La and Lc,

where La and Lc are the crystallite sizes in the layer plane and

in its normal direction, respectively (Figure 10). Quantitative

characterization of the microstructure could be performed by

measuring BSU parameters, including the average distance of

the interlayer spacing (d002), the stacking layer length (La), and

the layer thickness (Lc) from the 002 lattice fringe images.

These important parameters can be calculated from XRD

patterns.21 The d002 values are determined from the (002)

diffraction peak positions. The interlayer spacing for graphitic

stacking is known to be 0.3354 nm and that for turbostratic

stacking is reported to be about 0.344 nm. Therefore, the

observed interlayer spacing d002 is an average value depending

on the relative ratio of graphitic to turbostratic stacking, and so

decreases gradually to the value of 0.3354 nm with structural

improvement by heat treatment.

From XRD data of d002, a graphitization index (gP) can be

derived by applying the following equation:

gP ¼ 0:3440� d0020:3440� 0:3354

It is used to characterize quantitatively the degree of simi-

larity between carbon material and a perfect single crystal of

graphite.

The dimensions of carbon crystallites are determined from

the analysis of XRD line broadening. According to the Scherrer

formula, Lc is equal to:

Lc ¼ 0:89l= B cos ycð Þwhere l is the x-ray wavelength, B is the angular width

(radians) of the (002) diffraction peak at half-maximum inten-

sity (corrected for instrumental broadening) and yc is the

Bragg angle for the reflection (002). The layer dimension La is

calculated by use of the equation:

Page 10: Comprehensive Inorganic Chemistry II || Carbon

Layer plane spacing (c/2)

A plane

B plane

A plane

a0

0.141 nm

a = 0.246 nm

(a)

A planeB plane

(b) (c)

A planeB planeC plane

c0.6708 nm

Figure 9 (a) Idealized crystal structure of graphite showing ABAB stacking sequence; (b) top view of a hexagonal graphite crystal; and (c) top view of arhombohedral graphite crystal (the gray circles showing the position of the carbon atoms do not represent the actual size of the atom. Each atom, in fact,contacts its neighbors).

d002

La

Lc

Figure 10 Crystallographic parameters d002, La, and Lc of a graphite

332 Carbon

La ¼ 1:84l= B cos yað Þwhere B and ya correspond to the reflection (100) (often

labeled (10) for 2D-structures).

The sizes of graphitic crystallites have also been estimated

from neutron scattering, atomic force microscopy, and Raman

spectroscopy measurements.22

These crystallites may vary considerably in size. For

instance, the apparent crystallite size perpendicular to the

layer planes (Lc) of some glassy carbons may be as small as

1.2 nm, which is the length of a few atoms, or up to 100 nm

found in highly ordered pyrolytic graphites. The layer planes

may or may not be perfectly parallel to each other, depending

on whether the material is graphitic or nongraphitic carbon.

crystallite.

Thus, the local molecular ordering of the BSU is very impor-

tant for graphitization. The graphitizing carbons tend to be soft

and nonporous, with relatively high densities, and can be

readily transformed into crystalline graphite by heating in the

range of 2473–3273 K. In contrast, the nongraphitizing car-

bons are hard, low-density materials and cannot be trans-

formed to the crystalline graphite even at 3273 K and above.

The low density of the materials is a consequence of their

microporous structure, which gives a high internal surface

area. Recently, it has been suggested that nongraphitizing car-

bons may have a microstructure related to that of fullerenes.23

One of the main reasons for believing that nongraphitizing

carbons may be fullerene-like is that they can be transformed

by high-temperature heat treatment into structures containing

many closed carbon cages.

The aggregates of crystallites also have widely different sizes

and properties. Some, such as soot, are extremely small and

contain only a few small crystallites. In such cases, the proper-

ties are mostly related to the surface area. Other aggregates may

be relatively large and free of defects and essentially parallel to

each other, in which case the structure and its properties closely

match those of the ideal graphite crystal. Such large aggregates

are often found in pyrolytic graphite. In other aggregates, the

crystallites have an essentially random orientation. This occurs

in turbostratic (i.e., showing no evidence of 3D order)

or amorphous carbon. In such cases, the bulk properties are

essentially isotropic.

Page 11: Comprehensive Inorganic Chemistry II || Carbon

Carbon 333

7.13.3.1.1.2 Diamond

The name ‘diamond’ is derived from the ancient Greek adάmaB(adamas), ‘proper,’ ‘unalterable,’ ‘unbreakable,’ ‘untamed.’

A diamond is a crystal of tetrahedrally bonded carbon atoms

(sp3 hybridization) that crystallizes into the diamond lattice

which is a variation of the face-centered cubic structure.

Diamond is a relatively simple substance in the sense that

its structure and properties are essentially isotropic, in contrast

to the pronounced anisotropy of graphite. However, unlike

graphite, it has several crystalline forms and polytypes.24

Each diamond tetrahedron combines with four other tetra-

hedral to form strongly bonded, 3D, and entirely covalent

crystalline structures. Diamond has two such structures, one

with a cubic symmetry (the more common and stable) and one

with a hexagonal symmetry found in nature as the mineral

lonsdaleite. Cubic diamond is by far the more common struc-

ture. The covalent link between the carbon atoms of diamond

is characterized by a small bond length (0.154 nm) and a high

bond energy of 711 kJ mol�1. Each diamond unit cell has eight

atoms. Hexagonal diamond is an allotropic form of carbon

which is close to cubic diamond in structure and properties. It

is a polytype of diamond, that is, a special form of polymorph

where the close packed layers ({111} for cubic and {100} for

hexagonal) are identical but have a different stacking

sequence.23 There are two ways to visualize the diamond struc-

ture (Figure 11). First, diamond may be visualized as a face-

centered cubic lattice with additional atoms in half of the

tetrahedral holes. The two models of the diamond structure

shown in Figure 11(a) are drawn from this perspective.

A second way of representing the diamond structure is shown

in Figure 11(b). In that case, this is the same carbon skeleton

as found in the hydrocarbon adamantane and consists of a six-

membered ring ‘capped’ by the center atom above. Repetition

(a)

(b)

Figure 11 (a) Models of the diamond structure from the cubicperspective. The dark gray bonds show the outline of the cell and cannotbe represented with the models. (b) On the left, a model of the carboncage from which the diamond structure can be constructed. On the rightis a larger-scale model.

of this carbon skeleton gives the larger model shown on the

right of Figure 11(b). When represented this way, the diamond

structure is easier to compare to the lonsdaleite structure. Just

remember that there is only one diamond structure; ultimately,

these two ways of representing diamond are equivalent.

In diamond, all four outer electrons of each carbon atom

are ‘localized’ between the atoms in covalent bonding. The

movement of electrons is restricted and diamond does not

conduct electricity. With its fourfold coordinated tetrahedral

(sp3 hybridization) bonds, the diamond structure is isotropic

and, except on the (111) plane, is more compact than graphite

(with its sp2 anisotropic structure and wide interlayer spacing).

Consequently, diamond has higher density than graphite

(3.515 vs. 2.26 g cm�3).

7.13.3.1.1.3 Fullerenes

The fullerenes were discovered in 1985 at Rice University in

Houston by Richard Smalley, Robert Curl, and Harry Kroto,

who shared a Nobel Prize in 1996 for the discovery. While

performing mass-spectroscopy analysis of carbon vapor, they

observed the presence of even-numbered clusters of carbon

atoms in the molecular range of C30–C100. The fullerenes are

generally arranged in the form of a geodesic spheroid and thus

were named after the inventor of the geodesic dome, the

renowned architect Buckminster Fuller. They are also known

as ‘buckyballs.’ Unlike graphite or diamond, the fullerenes are

not a single material, but a family of molecular, geodesic

structures in the form of cage-like spheroids, consisting of a

network of five-membered rings (pentagons) and six-

membered rings (hexagons). The buckyball molecule follows

Euler’s theorem, which specifies that any convex closed-caged

structure can be made up of any number of hexagons but must

include exactly 12 pentagons in order to provide the appropri-

ate curvature necessary to close the cage. Fullerenes can have a

variable number of hexagons (m), with the general composi-

tion: C20þ2m. Although C20 is theoretically possible, it is a

highly unlikely structure due to the fact that two pentagons

do not go together well structurally. This is due to added strain

on the geometry. Fullerenes with less than 60 carbon atoms are

rare. The next smallest fullerene is C70, a rugby-ball-shaped

molecule that is commonly found in soot. Other fullerenes

were discovered with more and fewer carbon atoms than in

the C60, ranging from 28 up into the hundreds, though C60

remains the easiest to produce.

In order to account for the bonding of the carbon atoms of

a fullerene molecule, the hybridization must be a modification

of the sp3 hybridization of diamond and the sp2 hybridization

of graphite. In this case, the s orbital no longer contains all of

the s-orbital character and the p orbital is no longer of purely

p-orbital character, as they are in graphite. Unlike the sp3 or the

sp2 hybridizations, the hybridization in fullerene is not fixed

but has variable characteristics depending on the number of

carbon atoms in the molecule.25 This number varies from

20 (C20) for the smallest geometrically (but not thermodynam-

ically) feasible fullerene, to infinity for graphite (which could

be considered as the extreme case of all the possible fullerene

structures). It determines the size of the molecule as well as the

angle y of the basic pyramid of the structure (the common

angle to the three s-bonds). The number of carbon atoms, the

pyramidization angle (y – 90�), and the nature of the

Page 12: Comprehensive Inorganic Chemistry II || Carbon

334 Carbon

hybridization are related and this relationship (in this case, the

s character in the p-orbital) is given in Figure 12.

The bond lengths of the fullerenes are reported as

0.145�0.0015 nm for the bonds between five- and six-

membered rings, and 0.140�0.0015 nm for the bond between

the six-membered rings. The C60 has a calculated diameter of

0.710�0.007 nm. Rehybridization plays an important role in

determining the electronic structure of the fullerene’s family

and it is the combination of topology and rehybridization that

accounts for the possible specific reactivity of all the curved sp2

nanostructures.

Both the bulk and the surface properties of carbonmaterials

are dependent on their structure/texture, that is, on the spatial

Pyramidalization angle (qσπ–90)°

s –

char

acte

r in

the

π-o

rbita

l

5 10 15

0.1

0.2

0.3

C500

C240

C100

C84

C76C70

C60

00 0° 4°

5.8°6.9°

9.8°10.3°

10.7°11.7°

qσπ

C∞(graphite)

Figure 12 Hybridization of fullerene molecules as a function ofpyramidization angle (ysp – 90�). ysp is the common angle of the threes bonds. Reproduced from Haddon, R.C. Acc. Chem. Res. 1992, 25,127–133, with permission from American Chemical Society.

(a)

(b)

(c)

(d)

(e)

(f)

Figure 13 Macro-texture of (a) and (c) fibrous carbon materials (bar¼100carbon materials (bar¼10 nm): (d) carbon beads; (e) CB; (f) carbon onions;(h) AC; and (i) carbon nanowalls.

arrangement of carbon atoms and/or BSU. The structure and

the texture are in turn dependent on the precursor used and

the conditions of formation, as outlined in Section 7.13.2.

Because of the prodigious variety of possible carbon atom

arrangements characteristic of the carbon and graphite mate-

rials, it is useful to distinguish between three levels of ‘texture’:

nano-, micro-, and macrotexture. We will limit this analysis

to the sp2 and curved sp2 structures.

7.13.3.1.2 Macrotexture of carbon and graphite materialsThis level of carbon atom arrangement confers the most readily

recognizable features to carbon materials; it is related to mor-

phology and is easily accessible by electron microscopy obser-

vations. Basically, we will distinguish between three types of

macrotextures (Figure 13): the fibrous, the spheroidal, and the

textured materials. Note that this macrotexture is not necessar-

ily related to the orientation of the graphene layers, and can be

correlated to the production process.

Thus, typical fibrous or filamentary carbon materials include

CNTs, vapor-grown carbon fibers (VGCFs), both obtained by

catalytic chemical vapor deposition processes, as well as CFs. For

the spheroidal carbon materials,26 there is in general a concen-

tric organization of the graphene layers, so that the nanotexture

dictates the macrotexture. Examples include CB, mesophase

pitch, and the graphitic particles in spherulitic graphite cast

iron. For years, the structure and formation of these particles

have been inadequately understood, as models of spheroidal

carbon particles tended to involve assemblies of flat graphene

fragments. The discovery of C60 allowed us to consider the

evidence that fullerene-like elements may be present in the

well-known forms of spheroidal carbon mentioned, as well as

in carbon onions.25 As already pointed out, it has also been

suggested that nongraphitizing carbons such as those produced

(g)

(h)

(i)

nm): (a) CFs, (b) vapor-grown carbon fiber, (c) CNTs; (d–f) spheroidaland (g–h) textured carbon materials (bar¼5 mm): (g) glassy carbon;

Page 13: Comprehensive Inorganic Chemistry II || Carbon

Carbon 335

by the pyrolysis of polyvinylidene chloride and sucrose may

have a microtexture which is related to that of fullerenes.23

Graphitization of spheroidal particles have been investigated

by high-resolution electron microscopy and 3D-TEM.27 It has

been shown that the outside region of the particle heat-treated at

3073 K has a stacking structure of aromatic layers with some

distribution of d002, while the center region consisted of non-

graphitic microtexture (Figure 14). Structure defects seemed to

be concentrated along the ridgelines of the polyhedronized

particles after heat treatment.

The textured carbon presents a pore structure that can be a

memory of the texture of the precursor (this is the case for

some ACs) or results from a template approach as for hierar-

chical carbon foams. Although this macroporosity will often

contribute weakly to the total porosity of these materials, it

constitutes a structural characteristic at the macrolevel.

Of course, various important carbon or graphite materials

cannot enter in this classification such as the very ordered

graphite, highly oriented pyrolytic graphite (HOPG), Kish

graphite, diamond, and some poorly textured structures. For

these materials, an inspection of the micro- or even the nano-

structure will be necessary.

7.13.3.1.3 Microtexture of carbon and graphite materialsHere we can consider the equivalent of grain boundaries

(region of mismatch between two adjacent grains in a poly-

crystalline microstructure) in carbons as the most recognizable

Central region

(a)

(b)

Concentric orientationThin carbon layers (10–20nm) with nesting texture

Figure 14 (a) Typical bright-field image and electron diffraction patterns of3073 K) and (b) structure model. Reproduced from Yoshizawa, N.; Tanaike, O2558–2564, with permission from Elsevier.

feature of carbon microtexture. Many carbons are polycry-

stalline materials, especially those of interest in catalytic appli-

cations. According to a variety of recently published

characterization results, the pore walls are made of a spatial

arrangement of BSUs composed of stacked graphene layers. As

the spaces between the BSUs constitute an important part of

the pores accessible for adsorption of gases and liquids, the

knowledge of carbon microtexture is of fundamental impor-

tance. However, the microscopic structure of these materials is

still a matter of debate. Due to the disordered nature of porous

carbons, the detailed microtexture cannot be deduced from

experimental techniques such as XRD or HR-TEM. Although

many models have been proposed to describe the microtexture

of carbonmaterials (Figure 15),28–35 some key issues that need

further efforts to understand the intimate microtexture of car-

bon materials are: (1) how to identify and quantify the BSU in

the carbon material of interest, (2) how to understand the

supramolecular, that is, supragraphene organization of the

various carbon materials, and (3) how to identify which

intra- and/or intermolecular forces are responsible for the for-

mation of the microtexture in these materials.

Thus, suitable atomistic models need to be developed that

realistically describe the atomic structure of these materials

to understand their different properties and to help in tailoring

them for specific purposes. In an attempt to elucidate the

actual structure and physical properties of nano- and mesopo-

rous carbons, theoretical investigations are required. Some

Outside region

Concentric orientationgraphitic structure alongsurface

20 nm

the whole particle for spheroidal particles (right image heat-treated at.; Hatori, H.; Yoshikawa, K.; Kondo, A.; Abe, T. Carbon 2006, 44,

Page 14: Comprehensive Inorganic Chemistry II || Carbon

(e)

(g)

(i)

(c)

(a) (b)

(d)

(f)

(h)

1 nm

(j)

Figure 15 Micro-texture models of various carbon materials: (a) Franklin’s representations of a graphitizing carbon; (b) Franklin’s representations of anon graphitizing carbon28; (c) model suggested by Crawford and Johnson for structure of PAN-derived carbon fibers29; (d) model by Jenkins andKawamura for structure of glassy carbon30; (e) model for structure of glassy carbon containing closed, fullerene-like particles23; (f) model forstructure of glassy carbon derived from phenol resin following heat treatment at 3073 K31; (g) model for structure of nongraphitizing carbons basedon fullerene-like species32; (h) model of PVDC carbon heat treated at 2223 K33; (i) model of a microporous carbon34; and (j) model of microporouscarbon made up of a carbonaceous structural unit produced by pyrolysis of a cellulosic precursor.35 (a, b) Reproduced from Franklin, R. E. Proc. R. Soc.Lond. A 1951, A209, 196–218, with permission from Royal Society. (c) Reproduced from Crawford, D.; Johnson, D. J. J. Microsc. 1971, 94, 51–62,with permission from John Wiley and Sons. (d) Reproduced from Jenkins, G. M.; Kawamura, K.; Ban, L. L. Proc. R. Soc. Lond. A 1972, A327, 501–517,with permission from Royal Society.

336 Carbon

approaches have been recently undertaken to apprehend the

structure of porous carbons in their full complexity in order to

provide a more realistic basis for further theoretical and for a

more insightful interpretation of the experimental data. The

first models for nanoporous carbons were developed by

Franklin (Figure 15(a) and 15(b)). The basic fragments for

both graphitizing and nongraphitizing carbons were taken to

be sheets of carbon atoms; however, in the latter case, ordering

of the planes was confined to a small domain. Cross-links

between adjacent domains led to disclinations, which

Page 15: Comprehensive Inorganic Chemistry II || Carbon

Carbon 337

contributed to the amorphous nature of these carbons. Subse-

quently developed models used randomly oriented carbon

planes to form a ribbon-like structure.30,33 After the discovery

of fullerenes, Mackay and Terrones36 suggested that negatively

curved graphite structures, known as Schwarzites (periodic

graphitic arrangements with heptagonal and octagonal rings

of carbon), could be formed in arc-synthesis experiments

together with fullerenes. The high stability (i.e., low energy)

of these structures was shown subsequently.37 Continuing with

the same approach, a later work by Terrones and Mackay38

proposed that curved layers were essentially structural features

of noncrystalline materials. Terrones and Mackay39 used triply

periodic minimal surfaces tessellated with seven- and eight-

membered rings to compute the stability of negatively curved

graphite compared with fullerenes. Finally, theoretical charac-

terization of several models of nanoporous carbon has shown

that if negative curvature surfaces are joined in space, then

porous structures can be formed (Figure 16(a)).40 This was

finally confirmed by Bourgeois and Bursill,41 who observed

negatively curved graphite using HR-TEM.

The model proposed by Harris and Tsang (Figure 15(g))

initiated structure formation with curved sheets of graphite.32

It has been confirmed that the curvature and the correlation

length seen in nongraphitizing carbons may be readily ex-

plained by constructing a model of carbon with a small frac-

tion of nonhexagonal rings (Figure 16(b)).42 The curvature is

introduced by including pentagons and heptagons, together with

hexagons, as starting fragments. It has been also shown that this

is not necessary, as both five- and seven-membered rings are

readily generated from hexagonal graphene sheets depending

on how the connections are made (Figure 16(c) and 16(d)).43

A recent work shows a process of fullerene formation from a

graphene sheet using aberration-corrected TEM.44 Quantum

(a) (b)

(c) (d)

Figure 16 (a) Negative curvature surfaces joined in space, formingporous structures41; (b) A model of structure consisting of fourturbostratically layered sheets. Each sheet contains about 1% non-hexagonal rings42; (c) formation of five-membered ring from BSU; and(d) carbon nanostructure formed by connecting atoms based on shortestdistance or random algorithm.43

chemical modeling explains four critical steps in a top-down

mechanism of fullerene formation: (1) loss of carbon atoms at

the edge of graphene, leading to (2) the formation of penta-

gons, which (3) triggers the curving of graphene into a bowl-

shaped structure, which (4) subsequently zips up its open edges

to form a closed fullerene structure. A theoretical study exploring

the possibility of the transformation of a graphene sheet into a

fullerene confirms that the formation of defects at the edge of

graphene is the crucial step in the process.45

The details regarding a 3D construction of the microtexture

of porous carbon materials are of paramount importance to

understand the properties of these materials. Indeed, the syn-

thesis and the manufacture of carbons with the desired adsorp-

tive properties require a reliable characterization of the internal

structure of the carbon, which can be used to predict adsorp-

tion equilibrium and kinetics from a mature understanding of

the behavior of fluids in confined spaces.

Classification of pores is one of the basic requisites of

comprehensive characterization of porous solids.46 Based on

their origins, the pores are classified into intraparticle and

interparticle pores. The intraparticle pores are further classified

into intrinsic and extrinsic intraparticle pores. The former owes

its origin to the crystal structure, such as the spaces developed

between graphene layers of graphite, which can accept small

atoms or ions, known as intercalation. Rigid interparticle pores

are typically present in AC where large amount of pores in

various nanometer sizes are formed because of the random

orientation of BSU. The classification of pores according to

their size was proposed by the International Union of Pure

and Applied Chemistry (IUPAC) (Table 2). If we consider their

state, open or closed, it is worth mentioning that closed pores

are not necessarily of small size like in glassy carbons, and

closed macropores can be obtained in some carbon foams.47

For catalysis, there is a need for tailored carbons with respect to

mean pore size (L0), but also pore size distribution (PSD) and

specific microporous volume (W0).

Table 2 Classification of pores in solid materials

(1) Based ontheir origin

Intraparticlepores

Intrinsicinttraparticle pores

Extrinsicinttraparticle pores

Interparticlepores

Rigid interparticlepores(agglomerated)

Flexible interparticlepores(aggregated)

(2) Based ontheir sizeMicropores <2 nm Ultramicropres <0.7 nm

Supermicropores 0.7–2 nmMesopores 2–50 nmMacropores >50 nm(3) Based on

their stateOpen poresClosed pores(Latent pores)

Page 16: Comprehensive Inorganic Chemistry II || Carbon

338 Carbon

7.13.3.1.4 Nanotexture of carbon and graphite materialsIn the carbon and graphite materials family, the fundamental

structure unit is the graphene layer, which has a strong anisot-

ropy because the bonds in the layer are covalent and those

between the layers are van der Waals-like. As a consequence,

the anisotropic layers tend to be oriented during their agglom-

eration to form specific shapes. Therefore, different ways to

agglomerate yield different structures and different degrees of

preferred orientation of the layers, resulting in a variety of

carbon and graphite materials, in addition to the mixing ratio

of ABAB and turbostratic stacking. A classification based upon

the scheme and degree of preferred orientation of anisotropic

layers has been proposed as illustrated in Figure 17.8 As these

textures are constructed by fundamental nanosized structural

units, they are called nanotextures. First, random and oriented

nanotextures are differentiated, and then the latter classified by

the orientation scheme: parallel to the reference plane (planar

orientation) or around the reference point (point orientation).

The extreme case of planar orientation, that is, perfect orienta-

tion with large-sized planes, is a crystal of graphite. The plates

so-called HOPG have a very high degree of planar orientation

of hexagonal carbon layers, but the sizes of layers are not very

large. In other words, in the direction perpendicular to the

plate, the structure is close to perfect orientation, but in the

parallel direction it is polycrystalline. Various intermediates

between perfect planar orientation and random orientations

are found in pyrolytic carbons and coke particles depending

on the preparation conditions and the temperature of heat

treatment. Various highly oriented graphite materials, Kish

graphite, various pyrolytic carbons, and carbon films are

derived from some thermoplastic polymer precursors (see

Random orientation

Random nanotexture

Referenceplane

Referenceaxis

Referencpoint

Figure 17 Classification of nanotextures in carbon and graphite materials. Aand Engineering: from fundamental to applications, Tsinghua University Pres

Section 7.13.2.2). An axial orientation of layers is found in

some fibrous carbon materials as CNTs, some nanofilaments,

and VGCFs; the fibrous macrostructure morphology is possible

because of this axial orientation scheme. Note that this is not

the case for some CNFs, which present a fibrous texture and

graphene layers presenting an angle, or even being perpendic-

ular to the reference axis (see Section 7.13.5). The radial

alignment of layers can be found in one of the mesophase-

pitch-based CFs, having radial texture in the cross-section of the

fibers. In CNFs, most of which are grown through a chemical

vapor deposition process using fine catalyst particles, different

orientation schemes along the axis, from parallel (tubular type)

to perpendicular (platelet type) through herring-bone type or

cup-stacked type, can be differentiated. In point orientation,

concentric and radial alignments also have to be differentiated.

The extremes of concentric point orientation are the family of

fullerenes. They can also be found in carbon onions and in CB

particles formed by minute hexagonal carbon layers. A radial

alignment of layers to form a spheroidal macrostructure is

found in carbon spherules, which are formed from a mixture of

polyethylene and polyvinylchloride by pressure carbonization.

The structures of mesophase spheres are close to the radial point

orientation scheme,48 but in their centers the orientation of

layers is not radial. Inagaki proposed a theory in which the

interface between carbon and its surroundings during the ther-

mal decomposition plays an important role in the formation

of spheroidal carbon. According to Inagaki, a solid/liquid

interface supports the concentric growth of a sphere. A liquid/

liquid interface would lead to a radial growth of the sphere, and

solid/gas yields a random texture.49,50 The texture with random

orientation occurs in glass-like carbons and carbon materials

Oriented nanotexture

Planar orientation

Coaxial

Axial orientation

Point orientation

e

Radial

Concentric

Radial

dapted from Inagaki, M.; Feiyu, K. Carbon Materials Sciences, 2006.

Page 17: Comprehensive Inorganic Chemistry II || Carbon

Carbon 339

just after carbonization of some precursor polymers such as

phenol resin. Fundamental structural units composed mostly

of glass-like carbons are so small that they are difficult to observe

under TEM. Therefore, discussion on their nanotexture was

often based on TEM observations of high-temperature-treated

samples, where the layers became somewhat larger.

7.13.3.2 Bulk and Physical Surface Properties

7.13.3.2.1 Bulk propertiesThe bulk properties of carbon and graphite materials reflect

their structure, and particularly their degree of anisotropy

between the a and the b axes (very strong aromatic C–C

bonds), and the c direction, perpendicular to the basal plane

(very weak van der Waals or p–p interactions). For diamond,

an isotropic material, the sp3 tetrahedral arrangement of the

C atoms is responsible for its specific properties: a superhard,

electrical insulator with extremely high thermal conductivity.

Typical property values of principal carbon and graphite

materials of interest for catalysis are given in Table 3.

The electronic or electrical properties of carbons and graphite

materials are of practical interest in electrocatalysis.51,52 Anisot-

ropy and its degree of replication are responsible for the behav-

ior, from good conductors (graphite and CNT) to insulators

(AC). For single-walled CNTs (SWCNTs), theoretical calcula-

tions predicted that they could exhibit either metallic or semi-

conducting behavior depending only on diameter and helicity.

This ability to display fundamentally distinct electronic proper-

ties without changing the local bonding was experimentally

confirmed by scanning tunneling microscopy (STM), which is

able to resolve simultaneously, the atomic lattice and the elec-

tronic density of states of a material. Thermal properties often

follow electronic properties quite closely and are of importance

for thermal management of a given catalytic reaction, both on

Table 3 Physical bulk and surface properties of various carbon and grap

Property Graphite Diamond

d002 (nm) 0.3354 –Lc (nm) >100 –La (nm) >100 –Electrical resistivity (mΩ m)ab-Directionc-Direction

1–400.4>40

1011–1024

Thermal conductivity (W m�1 K�1)ab-Directionc-Direction

400<80

2000

SBET (m2 g�1) <10/500c <10/400d

Pore volume (cm3 g�1)MicroMeso

0.15–0.7f 0.8–1.2g

Packing (bulk) density (g cm3) >1 0.2g

AC, activated carbon; CB, carbon black; CF, carbon fibers; CNT, carbon nanotubes.aMWCNTs.bValues ranging between 0.345 and 0.339 nm can be obtained upon graphitization betweencThe second value is an upper value for high surface area graphite powders.dThe second value is an upper value and is for nanodiamond powders.eValues ranging between 800 and 1500 m2 g�1 are commonly obtained for activated carbonfFor high surface area graphite powders.gFor nanodiamond powders.hFor activated carbon fibers.

the macro- and the microscale.53 The room-temperature ther-

mal conductivity of carbonmaterials span an extraordinary large

range of orders of magnitude from the lowest in amorphous

carbons to the highest in graphene and CNTs. Carbon and

graphite materials such as graphite and glassy carbon absorb

light over a wide energy range, at least from deep ultraviolet

(UV) to radio frequencies. Crystalline diamond is optically

transparent due to its bandgap at �220 nm into the infrared,

but impurities and defects can result in weak absorption

throughout the visible region. Optical properties of interest for

catalytic applications are particularly those that reveal contribu-

tions of the interband transitions of p-electrons, because these

can, in principle, shed light on the electron-donating or

electron-accepting properties of carbons.54

Resistance to attrition and crushing are also important

parameters in catalysis as they will strongly impact the filtra-

tion behavior of supports as powdered or granular AC, acti-

vated carbon fibers (ACFs), and high-surface-area graphite

(HSAG). Thus, if carbon’s surface chemistry is duly utilized it

is not necessary to use as support a carbon adsorbent with very

high surface area and consequently presenting poor mechani-

cal resistance. In that respect, fibrous carbons offer advantages

not only in terms of mechanical properties but also as regards

better control of transport and other characteristics that can

ensure maximum accessibility of catalytically active sites, and

not only their maximum concentration.

Another additional issue is of special relevance here,

because of the impact of bulk properties on surface properties:

the effect of heteroatom incorporation into the carbon structure

(doped carbon) or between the graphene layers (intercalated

carbons). Heteroatom doping (e.g., boron, sulfur, phosphorus,

and nitrogen) of graphitic carbon lattices, either during the

production process or by a posttreatment, affects various phy-

sicochemical properties of sp2 carbon materials. In particular,

hite materials

AC CB CF CNTa

>0.344 0.35–0.37b 0.34–0.36 0.339–0.348<5 1–4 >5 –<5 1–3 5–50 –103–106 102 2–20 0.6–2

0.2–0.3 1–150 10–1100 300–3000

500–3000 10–1500 <10e 150–4500.5–1.4 0.2–3

(0–0.2)0.2–1.5h 0.2–2

0.6–1 0.2–0.5 1.4–2.2 0.1–0.3

1773 and 3773 K, respectively.

fibers.

Page 18: Comprehensive Inorganic Chemistry II || Carbon

340 Carbon

doping with boron or nitrogen has received growing attention

because significant changes in hardness, electrical conductivity

(additional electronic states around the Fermi level), and chem-

ical reactivity have been theoretically predicted and experimen-

tally observed. Through physical, chemical, or electrochemical

methods, ions or compounds can be intercalated between the

graphene layers. Intercalation provides to the host material a

means for controlled variation of many physical properties

over wide ranges. Because the free carrier concentration of the

graphite host is very low (�10�4 free carriers/atom at room

temperature), intercalation with different chemical species and

concentrations enables a wide variation of the free carrier con-

centration and thus of the electrical, thermal, and magnetic

properties of the host material.55

XRD and Raman spectroscopy are the most powerful single

tools for characterizing the bulk properties of carbon and

graphite materials.

7.13.3.2.2 Physical surface propertiesThese properties are related to adsorption, an important ele-

mental step both for catalyst preparation and for catalytic

reaction. The structure of the pores and PSD of carbon or

graphite materials are largely dictated by the nature of the

raw materials, their production processes and, for some of

them, the history of their carbonization. The pores in carbons

are scattered over a wide range of size (Table 2) and shape. The

pores are classified by their sizes usually into three groups:

(1) macropores having an average diameter of more than

50 nm, (2) mesopores with a diameter of 2–50 nm, and (3)

micropores having an average diameter of less than 2 nm. The

last are further divided into supermicropores (0.7–2.0 nm)

and ultramicropores of diameter less than 0.7 nm.

Template-based synthesis of porous carbons (Figure 18)

offers the opportunity to obtain orderedmaterials (microporous,

mesoporous, or macroporous) that combine a large specific

From zeolite

For microporous carbon

From mesoporous silica

For mesoporous carbon

From silica opal

For macroporous carbon

Figure 18 Ordered carbon and graphite materials from templatesyntheses. Reproduced from Lee, J.; Kim, J.; Hyeon, T. Adv. Mater. 2006,18, 2073–2094, with permission from John Wiley and Sons.

surface area with well-defined pore geometry.56 Hard template

methods use a regular nanostructure with free and accessible

empty spaces that can be infiltrated with a carbon precursor.

After carbonization, this structure is removed to generate pores

in the carbon material. Soft template methods, instead, use

amphiphilic molecule aggregates.

For CNTs, which can be either microporous (SWCNTs) or

mesoporous (multiwalled carbon nanotubes, MWCNTs), one

should consider two types of pores: the well-defined inner

cavity, cylindrical endocavities or a-pores,57 and the intertube

voids, interstitial pores or b-pores58 that will both contribute

to the PSD. In addition to the well-known a and b pores,

two types of newly defined pores with diameters of 2–10

and 8–100 nm have been evidenced for SWCNT arrays, inter-

bundle (packing of bundles) pores and interarray (bundle

arrays) pores.59 For MWCNTs, it was shown that the aggre-

gated (intertube) pores are much more important for adsor-

ption than the endocavities.60 It is worth noting that, as it

is difficult to describe the complex secondary structures of

such materials by the methods of Euclidean geometry, the

method of fractal geometry has been applied both for CNT61

and CB.62

For most of the carbon materials and particularly for AC,

the porosity can be tailored over a very wide range but it is

spatially random, and carbon atoms differ from each other in

their reactivity depending on their spatial arrangement. If the

origin of porosity is relatively well known and understood, the

adsorption of gases on these microporous carbons is still today

poorly understood, partly because the structure of these car-

bons is not well known. As already stated in Section 7.13.2, the

carbonization process involves thermal decomposition of car-

bonaceous material, eliminating noncarbon species, produc-

ing a fixed carbon mass (including disorganized carbon) and

rudimentary pore structure. These raw materials are activated

using steam or carbon dioxide (physical activation) or using

some chemical substances such as ZnCl2 or H3PO4 (chemical

activation). In the case of AC, physical activation first elimi-

nates the disorganized carbon, exposing the aromatic sheets to

the action of the activating agent, and leading to the develop-

ment of a microporous structure. As activation is associated

with weight loss of the host carbon, the extent of burn off of the

carbon material is taken as a measure of the degree of activa-

tion. Normally, in the first phase, the disorganized carbon is

burnt preferentially when the burn off is about 10%. This

results in the opening of blocked pores. Subsequently, the

carbon of the aromatic ring system starts burning, producing

active sites and wider pores. During themanufacturing process,

it is generally accepted that macropores are first formed by the

oxidation of weak points (edge groups) on the external surface

area of the raw material. Mesopores are then formed and are,

essentially, secondary channels formed in the walls of the

macropore structure. Finally, the micropores are formed by

the attack of the planes within the structure of the rawmaterial.

In the latter phase, excessive activation reaction results in

knocking down of the walls by the activated agents and a

weight loss of more than 70%. This results in an increase of

macropores and a decrease of micropores.

Moreover, the carbon atoms that are localized at the edges

and the periphery of the aromatic sheets, or those located

at defect positions and dislocations or discontinuities are asso-

ciated with unpaired electrons or have residual valances; these

Page 19: Comprehensive Inorganic Chemistry II || Carbon

Carbon 341

are rich in potential energy. Consequently, these carbon atoms

are more reactive and have a tendency to form surface oxygen

complexes during oxidative activation (see the next section).

These surface chemical groups promote adsorption and are

beneficial in certain applications. Alternatively, these surface

oxygen complexes break down and peel off the oxidized car-

bon from the surfaces as gaseous oxides, leaving behind new

unsaturated carbon atoms for further reaction with an activat-

ing agent. Thus, the activation mechanism can be visualized as

an interaction between the activating agent and the carbon

atoms which form the structure of an intermediate carbonized

product, resulting in a useful large internal surface area with

interconnected pores of the desired dimension and surface

chemical groups. After activation, the carbon will have acquired

an internal surface area between 700 and 1200 m2 g�1, depend-

ing on the operating conditions. Different techniques have been

used to determine PSD in porous carbons such as mercury

porosimetry, gas adsorption isotherms, and recently, STM. For

adsorption measurements, both the shape of the pores and the

pore connectivity are important parameters. However, despite

their importance in areas of technology such as catalysis and the

purification of gas and water supplies, the detailed structure of

ACs is still unknown. Indeed, due to their disordered nature, the

detailed microstructure cannot be deduced from experimental

techniques such as XRD or HR-TEM. Thus, suitable atomistic

models need to be developed that realistically describe the

atomic structure of these materials to understand their adsorp-

tion properties. Recently developed methods applying Monte

Carlo simulation63,64 led to interesting results in this field, the

most important conclusion of which is that the results predicted

for the ideal slit-like model of carbon pores (Figure 19(a)) are

far from reality.65 Computational simulation results from the

reverse Monte Carlo technique suggest that the model originally

proposed by Harris,32 based on small fragments of curved sp2

carbon sheets (fullerene-like species) can be successfully used to

describe PSD in AC (Figure 19(b)).66

As far as pore connectivity is concerned, two continuum

models are worth mentioning, the pore tree structure and the

random pore models. The random pore model considers that

the overlap between void elements of arbitrary shape can be

described exactly when the location of the voids is completely

random, that is, it obeys the Poisson distribution.67–69

Simons70 had proposed a tree or river system, for which they

assume that the pore structure occurs randomly and the pore

(a) (b)

Figure 19 (a) Slit-like model of carbon pores, (b) random porous structureJain, J. S. K.; Pellenq, R. J. M.; Pikunic, J. P.; Gubbins, K. E. Langmuir 2006,(c) the pore tree model proposed by Simons.70

length is determined by an arbitrary intersection with another

pore. At such an intersection, the smaller pore is terminated

and the larger pore retains its structure; then the largest pores

must terminate at either the exterior surface of the particle or at

an internal void (Figure 19(c)). This can be interpreted as

pores near to the external surface are larger than pores within

the particle. The pore tree model differs from the random pore

models generally used: “The primary difference between the

pore tree theory and the random pore model is that the ran-

dom pore model allows a single small pore to connect two

larger pores. This requires that pore aspect ratio (length to

diameter) be of order one hundred. The pore tree theory pre-

dicts that all pores possess an aspect ratio of order ten. Hence,

small pores may connect to larger pores only on one end and

all pores must branch from successively larger pores like a tree

or river system.”70

7.13.4 Surface Chemistry of Carbon Materials

The surface chemistry of carbon materials is governed by basal

and edge carbon atoms, as well as by the presence of defects

(i.e., structural carbon vacancies and nonaromatic rings).

These imperfections and defects along the edges of graphene

layers are the most active sites owing to high densities of

unpaired electrons. Heteroatoms, such as oxygen, hydrogen,

nitrogen, and sulfur, can be chemisorbed leading to stable

surface compounds, resulting in a complex surface chemistry,

when compared to oxides such as silica or alumina. Although a

direct analogy to the functional groups classified in organic

chemistry exists, given the chemical complexity of the carbon

surface and the fact that the heteroatoms are often located in a

confined space, one cannot fully predict the behavior of those

groups based on well-known organic chemistry reaction mech-

anisms. The concentration and the distribution of the surface

groups present on the carbon depend on the carbon type and

the pretreatment applied. Although adsorption onto carbons is

mainly of the dispersive interaction type (see Section 7.13.5.1),

surface chemistry plays an important role when specific

interactions are considered. The surface chemistry of carbons

determines their moisture content, catalytic properties, acid–

base character, and adsorptive properties. It is related to the

presence of heteroatoms other than carbon within the carbon

(c)

from the reverse Monte Carlo technique (reprinted with permission from22, 9942–9948. Copyright 2006 American Chemical Society), and

Page 20: Comprehensive Inorganic Chemistry II || Carbon

342 Carbon

matrix, and to the presence of carbene and carbyne structures71

at the edges of the graphene layer in sp2-hybridized carbon

materials. The most common heteroatoms are oxygen, nitro-

gen, phosphorus, hydrogen, chlorine, and sulfur. They are

bound to the edges of the graphite-like layers and form organic

functional groups such as carboxylic acids, lactones, phenols,

carbonyls, aldehydes, ethers, amines, nitro compounds, and

phosphates. These functional groups can be acidic, basic, or

neutral in character. The acidic behavior is often associated

with surface complexes or oxygen functionalities such as car-

boxyls, lactones, and phenols. However, functionalities such as

amines, pyrones, chromenes, ethers, and carbonyls are respon-

sible for the basic properties of the carbon surfaces. Basic prop-

erties associated with Lewis sites located at the p electron-rich

regions within the basal planes of graphitic microcrystals, away

from the edges, have also been proposed. We will concentrate

our analysis on oxygen- and nitrogen-containing surface groups,

as they are the most common and relevant for catalysis. It is

however worth mentioning that with the discovery of new

nanocarbon forms such as fullerenes or nanotubes, a rich car-

bon functionalization chemistry has recently emerged.72,73

Figures 20 and 21 represent the different types of oxygen- and

nitrogen-containing functionalities generally observed for car-

bon materials.

O

O

O

O

HC

O

O

O

O

R

O

(d)

(e)

(f)

(g)

(h) (i)

(n)

Figure 20 Oxygen-containing species, and specific sites generally present on(d) ether, (e) quinone, (f) aldehyde, (g) lactone, (h) chromene, (i) pyrone, (j) c(n) p electron density on carbon basal plane.

Carbon surface oxidation is the most popular way of carbon

surface modifications. It can be done either from a gaseous

(oxygen, ozone, air, or nitric oxides) or a liquid phase (nitric

acid, hydrogen peroxide, potassium permanganate, sulfuric

acid, and sodium peroxydisulfide).74 Although the conditions

of gas phase oxidation vary, usually it is done in an oven at

elevated temperatures between 473 and 623 K, with continu-

ous flow of the oxidant. Air oxidation is considered weak, and

as a result, various oxygen-containing groups are formed with

the predominant population of weakly acidic groups such as

phenols. Oxidation in the liquid phase is much more complex

and results in more severe changes to the carbon surface chem-

istry. The oxidations are usually carried out in open vessels

with oxidants in a wide range of concentrations, depending

on the desired effects. Another important factor in this type of

oxidation is temperature. The higher is the temperature and the

stronger is the oxidant, the more oxidized is the carbon surface.

In some cases, strong oxidation, as that obtained with nitric

acid at its boiling point, can totally destroy the carbon struc-

ture. It is generally accepted that oxidation with strong oxi-

dants such as nitric acid leads to carbons with a predominant

population of surface carboxylic groups, whereas treatment

with hydrogen peroxide increases mainly the population of

phenols. Nitric acid oxidation can also result in incorporation

O

O

OH O

OH

O

OH

OO

··

··

(a)

(c)

(b)

(j)(k)

(l)

(m)

carbon surface: (a) carboxylic acid, (b) phenol, (c) carboxylic anhydride,arbene-like species, (k) carbonyl, (l) lactol, (m) carbyne-like species, and

Page 21: Comprehensive Inorganic Chemistry II || Carbon

N

N

N

HN

NO2

HONO

O

H2N

O

HN

N

N

CNNH2N

H

··

(a)(b)

(c)

(d)

(e)

(f)(g)

(h)

(i)

(j)(k)

Figure 21 Nitrogen-containing species, and specific sites generally present on carbon surface: (a) nitroso group, (b) a-pyridone, (c) nitro group(abbreviated as N-X), (d) amide, (e) pyrrole type nitrogen (abbreviated as N-5), (f) amine, (g) pyridine-like group (abbreviated as N-6), (h) nitrile,(i) imine, (j) lactam, and (k) quaternary amine (abbreviated as N-Q).

Carbon 343

of few nitrogen as nitro groups, likely attached to the carbons

at the edges of graphene planes.75

The oxygen-containing functionalities are considered as

acidic or basic. The acidic surface groups (i.e., carboxylic acid,

lactone, and phenol groups) are formed when the carbon

surface is exposed to an oxidant via reactions with oxidizing

agents from solutions or gas phase, either at room or high

temperatures. However, basic groups are formed when an oxi-

dized surface is reduced by heating in inert atmosphere at high

temperatures. The decomposition of acidic groups results in

active sites at the edges of the graphene layers that upon cool-

ing in inert atmosphere and reexposure to air, attract oxygen-

forming basic functional groups such as chromene or

pyrone.76 Although there is a general agreement about the

type of surface functionalities that determines the acidic char-

acter of a carbon material, the nature of carbon basic surfaces

remains controversial and open to investigation. Generally

speaking, oxygen-containing functionalities (i.e., chromene,

pyrone, and quinones) and nonheteroatomic Lewis base

sites, characterized by regions of p electron density on the

carbon basal planes, govern carbon basicity.77

Similarly to oxygen, introduction of nitrogen in carbon

materials can be performed both in a liquid or a gas phase

using nitrogen-containing precursors. Ammonia is usually

used as a gaseous precursor of nitrogen at temperatures

between 673 and 1273 K. When the modifications are carried

out on the carbon samples, either preoxidized or not, in the

liquid phase compounds such as carbazole, nitrogen-enriched

polymers, acridine, melamine, or urea are used. The carbons

are impregnated with water or alcoholic solutions of the

nitrogen-containing compounds and then exposed to heat

treatment at temperatures between 673 and 1293 K. Usually,

the nitrogen content in carbon materials is very small, unless it

is present in the carbon precursor, as for instance carbazole,

nitrogen-enriched polymers, acridine, and melamine. So far, it

has been demonstrated that the presence of nitrogen can be a

key parameter for the performance of carbon materials as

catalyst supports and for catalytic activity.78

The type of nitrogen functionalities present on the carbon

surface is a function of the treatment applied. This includes the

type of nitrogen-containing precursor, the chemical activity of

the carbon surface and, the most important, the temperature of

heat treatment. The latter determines the type of chemistry

owing to the fact that some nitrogen-containing species are

unstable at high temperatures. According to the studies per-

formed, lactam and imide structures are mainly formed by

ammination, and amide upon ammoxidation; the former is

transformed to pyrrole and pyridine by heat treatments.79 Like

for oxygen, nitrogen-containing functionalities determine

the acidic or the basic character of carbons and thus its surface

chemical reactivity or catalytic activity. Treatments of carbo-

naceous surfaces with nitrogen-containing reagents at low

temperatures (less than 800 K) lead to the formation of

lactams, imides, and amines, slightly acidic in their nature.

However, treatments at high temperatures result in an increase

in the N quaternary (N atoms incorporated in the graphitic

layer in substitution of C atoms) pyridinic and pyrrole-type

structures. They are responsible for an increase in the surface

Page 22: Comprehensive Inorganic Chemistry II || Carbon

Table 4 Common binding energy assignments for thecarbon 1s peak

Functional group Binding energy (eV)

Hydrocarbon C–H, C–C 285.0Amine C–N 286.0Alcohol, ether C–O–H, C–O–C 286.5Carbonyl C¼O 288.0Amide N–C¼O 288.2Acid, ester O–C¼O 289.0Urea N–CO–N 289.2Carbamate O–C–N 289.9Carbonate O–C–O 290.3p–p* Shake-up satellite 291.0Plasmon 292.0

The observed binding energies will depend on the specific environment where the

functional groups are located. Most ranges are � 0.2 eV, but some can be larger.

Table 5 Common binding energy assignments for theoxygen 1s peak

Functional group Binding energy (eV)

Carbonyl C¼O, O–C¼O 532.2Alcohol, ether C–O–H, C–O–C 532.8Anhydride, carboxylic acid, lactone, and esterC–O–C¼O

533.7

Chemisorbed water or O2 535–536

The observed binding energies will depend on the specific environment where the

functional groups are located. Most ranges are �0.2 eV.

Table 6 Common binding energy assignments for thenitrogen 1s peak

Functional group Binding energy (eV)

Pyridinic nitrogen, Ar–N–Ar 398.6þ0.3–N–HPyrrolic nitrogenPyridine nitrogen

399.4þ0.3

–OC¼N Pyridine pyrrole 400.2þ0.1N, quaternary 401.3þ0.2Pyridine-N-oxide 402.5�403.8NOx, oxidized, N–O–C 404.5þ0.4

The observed binding energies will depend on the specific environment where the

functional groups are located.

344 Carbon

polarity of the carbon, and basicity as both pyridine and

pyrrole-type structures are considered as basic.

Very marked effects of carbon surface chemistry have been

reported on the adsorption of various species such as aromatics,

dyes, heavy metals, pharmaceuticals, polar species such as alco-

hols, acids, or aldehydes, and even small-molecule gases.80 In

those applications, the species present on the carbon surface can

enhance the specific interactions or even alter the porosity via

blocking of pore entrances for molecules to be adsorbed.

Specific interactions include hydrogen bonding, acid/base,

complexation, etc. These specific interactions are also very

important in catalysis. When carbon is used as support, the

surface chemistry governs the catalyst dispersion, its loading,

as well as the catalytic activity or selectivity. The presence of,

often confined, surface functional groups can influence surface

diffusion and desorption or provide a suitable environment for

gradient concentration that can affect both catalyst activity and

selectivity. Knowledge of the surface chemistry of carbon mate-

rials is also of great importance as the physicochemical proper-

ties of carbons are strongly influenced by the presence, even in

small amounts, of chemical species on the surface. Hence, many

of their applications are conditioned by their chemical charac-

teristics. In addition, these properties are known to change

during storage of carbons. Thus, the surface functional groups

determine the self-organization, the chemical stability, and the

reactivity in adsorptive and catalytic processes.

A complete characterization of carbon surface chemistry

necessitates the use of a broad battery of analytical techniques.81

The nature of the chemical surface groups is currently determined

by Fourier-transform infrared spectroscopy (FTIR), diffuse reflec-

tance infrared Fourier transform spectra (DRIFTS), and x-ray

photoelectron spectroscopy (XPS). Their quantitative determina-

tion can also be carried out by selective titrations in aqueous

solutions at room temperature. The results obtained by selective

titrations can be contrasted by other experimental methods.82

Among them, thermogravimetry, thermal programmed desorp-

tion (TPD), and, very frequently, calorimetric measurements are

used. Generally speaking, the methods used to characterize the

surface of carbonaceous materials are referred to as ‘wet’ and ‘dry’

techniques. The ‘wet’ technique involves Boehm83 and potentio-

metric titrations.84 In general, the Boehm titration, which has

been recently standardized,85 proceeds as follows: carbon is

added to three different bases (NaHCO3, Na2CO3 and NaOH),

the NaHCO3 and the Na2CO3 aliquots are then back-titrated

with NaOH, and the NaOH aliquots are titrated directly with

HCl. As the strongest base, NaOH is assumed to neutralize all

Brønsted acids, while Na2CO3 neutralizes carboxylic and lactonic

groups, and NaHCO3 neutralizes carboxylic acids. From this

knowledge, the types of oxygen surface groups can be deter-

mined. Acid–base potentiometric titration can also be used for

the determination of pH at the point of zero charge (pHPZC),86

that is, the pH above which the total surface of the carbon

particles is negatively charged. Such a determination is crucial

for adsorption of metallic ions during catalyst preparation.87

Generally speaking, the combination of these titration methods,

together with temperature-programmed desorption, is necessary

for an in-depth characterization.

‘Dry’ methods include diffuse reflectance FTIR, XPS, ther-

mal analysis, and TPD. XPS has proved to be a useful analytical

tool for monitoring the processing steps by providing

information on the relative amounts of different elements

with respect to carbon and their valence states. As other ana-

lytical techniques cannot distinguish between the sp2 and the

sp3 carbons very well, XPS can be useful in the semi-

quantitative analysis of carbon species on carbon materials.

Some common binding energy peak assignments for the car-

bon 1s, oxygen 1s, and the nitrogen 1s peaks are seen in

Tables 4–6.88

When applied to porous carbons for the determination of

the oxygen surface groups, XPS has the following drawbacks:

(1) the external surface area is only a small fraction of the total

surface area and it is not representative of all the material;

Page 23: Comprehensive Inorganic Chemistry II || Carbon

Carbon 345

(2) the holes in the surface can affect the final results because

the surface is not flat; (3) the analysis is made in high vacuum,

that is, under conditions quite different from those usually

used in the applications of the carbon catalyst, and a rearrange-

ment of the surface can occur; and (4) deconvolution of the

N 1s, O 1s, and C 1s peaks is not straightforward, and it is still

a matter of discussion.

Infrared (IR) spectroscopy has been widely used to charac-

terize the surface groups of different carbon materials.89,90

However, IR spectra of carbon materials are difficult to obtain

because of problems in sample preparation, poor transmis-

sion, uneven light scattering related to large particle size, and

so forth. Moreover, the electronic structure of carbon materials

results in a complete absorption band through the visible

region to the IR. Fortunately, some of these problems can be

overcome by improving sample preparation (e.g., carbon

films) as well as by using the most recently developed IR

techniques such as DRIFTS. The adsorption band and peaks

for oxygenated surface species on carbon materials are given in

Table 7.80 Besides technical difficulties in obtaining the IR

spectra of carbon materials, their interpretation is often an

additional problem because not all of the observed absorption

bands may be assigned unequivocally to specific functional

groups, most likely owing to the overlap of several bands. In

other cases, it is not uncommon that some band assignments

differ substantially among the recent IR studies on carbonmate-

rials. This is the case for the so-called ‘quinone band’ that has

been assigned to the 1660–167091 or 1550–1680 cm�1

interval.90 Another controversial assignment corresponds to

the band at 1600 cm�1, which is a prominent feature in the IR

spectra of carbon materials. The 1600-cm�1 band has been

attributed to either oxygen surface compounds or ring vibrations

of the basal plane. Intriguingly, the presence and the intensity of

Table 7 Infrared adsorption bands on carbon surfaces and theircorresponding assignments to oxygenated functionalities

Group/functionality Assignment (cm�1)

1000–1500 1500–2050 2050–3700

C–O stretch of ethers 1000–1300Ether bridge between rings 1230–1250Cyclic ethers containingCOCOC groups

1025–1141

Alcohols 1049–1276 3200–3640Phenolic groupsC–O stretch 1000–1220O–H bend/stretch 1160–1200 2500–3620Carbonates; carboxyl-carbonates

1000–1500 1590–1600

Aromatic C¼C stretching 1585–1600Quinones 1550–1680Carboxylic acids 1120–1200 1665–1760 2500–3300Lactones 1160–1370 1675–1790Anhydrides 980–1300 1740–1880Ketenes (C¼C¼O) 2080–2200C–H stretch 2600–3000

Reprinted from T. Bandosz, Surface chemistry of carbon materials. In Carbon Materials

for Catalysis; Ser, P.; Figueiredo, J. L., Ed.; J. Wiley & Sons: Hoboken, NJ, 2009;

pp 45–92, with permission.

this band is strictly related to the concentration of surface oxides,

but IR studies using 18O labeled carbon do not support an

assignment to carbonyl-type species.92 Thus, a widely accepted

hypothesis assigns the 1600 cm�1 band to the C¼C stretching

modes of polyaromatic systems (Table 7), whose IR intensity

would be reinforced by chemisorbed oxygen. However, the

relationship between the nature of carbon surface oxides and

the intensity of the 1600 cm�1 band is not clear.

Quantum chemical methods have been recently employed

to get a more detailed assignment of the IR absorption bands of

carbon materials.93 It was found that the frequencies of the

C¼O bonds present in acid functional groups were systemati-

cally lowered when phenolic groups were close enough to estab-

lish hydrogen bonds. Concerning the origin of the 1600 cm�1

band of carbons, it was found that, in the case of acid carbons,

this band can be assigned to the C¼C stretching of carbon rings

decorated mainly with phenolic groups. Cyclic ethers in basic

carbons would also promote absorption in the 1600 cm�1

region of the IR spectrum. The calculated (density functional

theory, DFT) vibrational frequencies in the 1400–1900 cm�1

range of the surface oxygen species are given in Figure 22.

Although XPS and DRIFTS provide excellent qualitative

information about the carbon surface, the quantitative insight

is not straightforward and requires special mathematical treat-

ment using many approximations.

Boehm and potentiometric titrations provide qualitative

and quantitative information on the carbon surface. However,

the information on acidic groups is limited to such comp-

ounds as phenols, lactones, and carboxylic acids, yet neglecting

any other groups present. Temperature-programmed desorp-

tion allows one to study functional groups that decompose

below 1250 K, but it does not provide direct qualitative results

on the carbon chemistry.94 Surface oxygen complexes on car-

bon materials decompose upon heating by releasing CO and

CO2; thus, the TPD peaks of CO and CO2 at different temp-

eratures correspond to specific oxygen groups. For example,

CO2 is released by decomposition of carboxylic groups at

373–673 K, or lactone groups at 463–923 K. Both CO and

CO2 peaks originate from the decomposition of carboxylic

anhydrides in the temperature range of 623–900 K. Phenols,

ethers, carbonyls, and quinones give rise to CO at 973–1253 K.

The quantities of CO and CO2 released during the TPD experi-

ments correspond to the total amount of surface oxygen

groups. The decomposition temperature is related to the

bound strength of specific oxygen-containing groups. Thus,

the position of the peak maximum at a defined temperature

corresponds to a specific oxygen complex at the surface.

Deconvolution of the TPD profiles gives quantitative informa-

tion about surface oxygen groups.

7.13.5 Carbon and Graphite Materials for Catalysis

The physical and chemical properties of carbon materials, as

their tunable porosity and surface chemistry, make them suit-

able for application in many catalytic processes. Traditionally,

carbon materials have been used as supports for catalysts in

heterogeneous catalytic processes, although their use as cata-

lysts on their own is becoming more and more common.95–97

Although several kind of carbon materials have been studied,

Page 24: Comprehensive Inorganic Chemistry II || Carbon

C

OHO

O

O

OOO

Anhydride

1740–1760 (C=O, symetric)1780–1810 (C=O, antisymetric)

Carboxyl1740–1750 (C=O)

Cyclic ketone Cyclic ether1580–1620 (C ring)

O

OO

O

5-membered ringlactone

6-membered ringlactone

~1820 (C = O) ~1790 (C = O)

O

CO O

O

OOOH

Anhydride1670–1770 (C=O)

Carboxyl1660 (C=O)

HO OO

O

H

O

CO O

O

O

O

O

5-membered ringlactone

6-membered ringlactone

~1740 (C=O) ~1710 (C=O)

Ketone and ether rings (pyrones)

1660–1700 (C=O)1450–1640 (C ring)

OHOH OH O O O

O

Cyclic ether (zigzag)1600–1700 (C = C and C ring)

Cyclic ketones (zigzag) 1690–1710 (C = O)

Zigzag phenol groups

~1590 (C ring)1420–1480 (C–O–H and C = C deformation)

1650–1700 (C = O)1550–1600 (C ring)

H

H

O

O

O O

Figure 22 Structure of carbon surface oxides and their corresponding IR assignments in the 1400–1900 cm�1 region according to B3LYP/6-31G*calculations on polyaromatic systems. Reproduced from Fuente, E.; Menendez, J. A.; Dıez, M. A.; Suarez, D.; Montes-Moran, M. A. J. Phys. Chem. B2003, 107, 6350–6359, with permission from American Chemical Society.

346 Carbon

AC and CBs are the most commonly used carbon supports. The

typically large surface area and high porosity of AC catalysts

favor the dispersion of the active phase over the support

and increase its resistance to sintering at low metal loadings.

The PSD can also be adjusted to suit the requirements of

several reactions. The surface chemistry of carbon catalysts

also influences their performance as catalysts and catalyst sup-

ports. Carbon materials are normally hydrophobic and they

usually show a low affinity toward polar solvents, such as

water, and a high affinity toward nonpolar solvents, such as

Page 25: Comprehensive Inorganic Chemistry II || Carbon

(a) (b)

(c)

5 nm

10 nm

5 nm

10 nm

(d)

Figure 23 Conventional HRTEM micrographs of (a) fresh AC; (b) ACfollowing heat treatment at 2273 K (reproduced from Harris, P. J. F.; Liu,

Carbon 347

acetone. Although their hydrophobic nature may affect the

dispersion of the active phase over the carbon support, we

have seen that the surface chemistry of carbon materials can

easily be modified, for example by oxidation, to increase their

hydrophilicity and favor ionic exchange. Apart from an easily

tailorable porous structure and surface chemistry, carbon

materials present other advantages: (1) metals on the support

can be easily reduced; (2) the carbon structure is resistant to

acids and bases; (3) the structure is stable to high temperatures

(even above 1023 K under an inert atmosphere); (4) porous

carbon catalysts can be prepared in different physical forms as

granules, cloth, fibers, pellets, etc.; (5) the active phase can be

easily recovered; and (6) the cost of conventional carbon sup-

ports is usually lower than that of other conventional supports,

such as alumina and silica. Nevertheless, carbon supports pre-

sent also some disadvantages, such as, they can be easily gasi-

fied, which makes them difficult to use in hydrogenation and

oxidation reactions, and their reproducibility can be poor,

especially AC catalysts, as different batches of the same mate-

rial can contain varying ash amounts. In the following part of

this section, we will briefly review the main carbon and graph-

ite materials relevant to catalysis.

Z.; Suenaga, K. J. Phys. Condens. Matter 2008, 20, 362201, withpermission from IOP); (c) untreated CB particle; and (d) CB particle heattreated at 2973 K.

7.13.5.1 Activated Carbons

The term ‘activated carbon’ defines a group of materials with

highly developed internal surface area and porosity, and hence

a large capacity for adsorbing chemicals from gases and

liquids.98 The adhesion to the surface is due to van der Waals

or London dispersion forces. This force is strong over short

distances, equal between all carbon atoms, and not dependent

on outside parameters such as pressure or temperature. Thus,

adsorbed molecules will be held most strongly where they are

surrounded by the most number of carbon atoms. The area

presenting a high density of graphitic plates will favor a high

adsorption. High temperature treatment (>1500 K) of AC can

favor the adsorption sites by increasing the density of ‘p-sites’present on partly graphitized structure (Figure 23).99,100 The

main adsorbance performance criterion of AC is its iodine

number. It is determined by the amount of iodine it can adsorb

in a given solution (given in mg g�1).101 Almost all precursors

containing a high fixed carbon content can potentially be

activated. The most commonly used raw materials are coal

(anthracite, bituminous, and lignite), coconut shells, wood

(both soft and hard), peat, and petroleum-based residues.

Most carbonaceous materials do have a certain degree of poros-

ity and an internal surface area in the range of 10–15 m2 g�1.

The carbonization process is completed by heating thematerial

at elevated temperature in an oxygen-deficient atmosphere that

cannot support combustion. The carbonized particles are

‘activated’ by exposing them to an activating agent, such as

steam at high temperature. During the activation process, the

internal surface is developed and extended by controlled

oxidation of carbon atoms. After activation, the carbon will

have acquired an internal surface area between 700 and

1200 m2 g�1, depending on the processing conditions. The

internal surface area must be accessible to the passage of reac-

tants for adsorption. Thus, it is necessary that an AC has not

only a highly developed internal surface but accessibility

to that surface via a network of pores of different diameters

(see Section 7.13.3.2.2). All ACs contain micropores, meso-

pores, and macropores within their structures but the relative

proportions vary considerably according to the rawmaterial. In

general, it can be said that macropores are of little value in their

surface area, except for the adsorption of unusually large mol-

ecules and are, therefore, usually considered as an access point

to micropores. Mesopores do not generally play a large role in

adsorption, except in particular carbons where the surface area

attributable to such pores is appreciable (usually 400 m2 g�1 or

more). Thus, it is the micropore structure of an AC that plays

an effective role in adsorption. It is, therefore, important that

AC not be classified as a single product but rather a range of

products suitable for a variety of specific applications. This is

also true concerning the presence of ashes in the AC. The ash

content is an important physical characteristic of AC. This is

the inorganic, commonly inert, amorphous, and unusable part

present in the AC. This ash comes initially from the basic

material. The lower the ash content, the better the AC. The

practical limit for the level of ash content allowed in the AC

varies within 2–5%. As these impurities on the AC surface often

result in undesirable side reactions, AC should be derived from

rawmaterials with a high degree of purity. In those cases where

additional purity is required, AC can be acid washed before

use. Other important physical characteristics of AC with respect

to catalysis are: hardness, apparent density, moisture, pHPZC

value, and PSD. Powdered AC is made up of crushed or ground

carbon particles, 95–100% of which will pass through a desig-

nated mesh sieve or sieves. Granular AC can be either in the

granular form or extruded.

7.13.5.2 Carbon Blacks

CBs (a very pure form of soot) are a group of materials that

are characterized by having near-spherical carbon particles

Page 26: Comprehensive Inorganic Chemistry II || Carbon

348 Carbon

of colloidal size, which are produced by three significant

processes: the furnace process, the channel process, and the

acetylene process. The products made by each process have

unique characteristics. During production, the colloidal car-

bon particles coalesce into chemically fused aggregates and

agglomerates (groups of aggregates) with varying morphol-

ogies. Their fundamental properties vary with feedstock and

manufacturing conditions, and they are usually classified

according to their method of preparation or intended applica-

tion. The key properties for CBs are considered to be fineness

(primary particle size), structure (aggregate size/shape), poros-

ity, and surface chemistry. The Brunauer, Emmett, and Teller

(BET) surface area of CB covers a wide range from few tens for

acetylene or thermal blacks to greater than 1000 m2 g�1 for

Ketjen black. The porosity in CB also varies from mild surface

pitting to the actual hollowing out of particles. Additional poros-

ity is also created by the intra- and interaggregate voids that are

formed between the small, fused, primary carbon particles. The

surface area of CB is generally considered as being more acces-

sible than other forms of high surface-area carbon.

Graphitized CB is another support material of interest to

catalyst manufactures. This high-surface material is obtained

by recrystallization of the spherical CB particles at 2773–

3273 K (Figure 23). The partially crystallized material pos-

sesses well-ordered domains. The degree of graphitization is

determined by process temperature. Highly conductive CBs are

characterized by a high structure (i.e., aggregates with a highly

branched, open structure), high porosity, small particle size,

and a chemically clean (oxygen-free) surface. The conductivity

of CBs is typically in the range 10�1–102 (O cm)�1 and is

influenced by the relative ability of electrons to jump the gap

between closely spaced aggregates (electron tunneling) and by

graphitic conduction via touching aggregates. CBs are conven-

tional support for fuel cell electrocatalysis.102

7.13.5.3 Graphite and Graphitized Material

Hardly any of the vast amount of natural or manufactured

graphite is currently used as catalyst support. This is particu-

larly due to the low surface area of graphite of only 10–

50 m2 g�1. Additionally, the relatively low reaction ability of

natural graphites toward the activation agents (steam, oxygen,

and carbon dioxide) can hinder the preparation of porous

carbon supports on their basis. However, graphites are able

to form graphite intercalated compounds that can sufficiently

(by hundreds of times) increase their volume after a high

temperature treatment allowing a significant expansion of the

material along the crystallographic c-axis to occur.103 This

unique property of intercalated graphites was used for

manufacturing thermally expanded graphites. Such a material,

developing a specific surface area of 300 m2 g�1, has been used

as catalyst support.104 Additionally, HSAG is available from

graphitized material by a special grinding process. Surface

areas of 100–300 m2 g�1 make this graphite an interesting

support material for precious metal catalysts.

(a) (b) (c) (d)

Figure 24 Different carbon nanostructures produced by catalyticchemical vapor deposition: (a) multi-walled carbon nanotubes; (b)ribbon-carbon nanofibers (r-CNFs); (c) fishbone-carbon nanofibers(f-CNFs); and (d) platelet-carbon nanofibers (p-CNFs).

7.13.5.4 Activated Carbon Fibers

Among the latest addition to porous carbons are the ACFs.

ACFs with high and most effective pore structure have been

used as support for catalytic applications.105,106 The pore

structures and the mechanical properties of ACFs are a function

of the fiber precursor, preparation method, and the chemical

activation process. An advantage of ACFs as catalyst support

over zeolites is that these can be prepared with larger pore

volumes, while the pore size can be easily adjusted during

preparation and treatment or by post-preparation modifica-

tion. Furthermore, by appropriate preactivation impregnation

with, for example, phosphates or boric acid, it is also possible

to introduce different types of mesoporosity.107 Another dif-

ference between ACFs and zeolites is the geometry of the

micropores; in zeolites the pores are shaped channels and

cavities whereas in ACFs the pores are slit shaped. Conversely,

narrow and long pores might hinder transport of reactant and/

or products, which generally is undesirable. As the pores in

ACF are microporous and their PSD is narrow, it is thought

that they could show molecular sieving and shape selectivity

effects in the catalytic process.108,109 Another advantage of ACF

is the ability to prepare the products in different shapes such as

woven clothes and nonwoven mats. ACFs have also a number

of advantages over granular ACs. In granular ACs, the adsorbed

gas molecules always have to reach micropores by passing

through macropores and mesopores, whereas in ACFs, most

micropores are exposed directly to the surface of the fibers and

hence to the adsorbed gas. Therefore, the adsorption rate as

well as the amount of adsorption of gases into the ACFs are

much higher than those into granular ACs.

7.13.5.5 Carbon Nanostructures

CNTs and nanofibers are produced by the catalytic decompo-

sition of certain hydrocarbons.110,111 By careful manipulation

of various parameters, it is possible to generate nanostructures

in assorted conformations and also to control their crystalline

order (Figure 24). Recently, there is an increasing interest in

the application of CNTs or CNFs as supports for catalysis as the

nanoscale tubular morphology of these materials can offer a

unique combination of low electrical resistivity and high

porosity in a readily accessible structure.112–114

Thus, CNTs represent an interesting alternative to conven-

tional supports for a number of reasons,115,116 including:

(1) their high purity that eliminates self-poisoning; (2) their

impressive mechanical properties, high electrical conductivity,

Page 27: Comprehensive Inorganic Chemistry II || Carbon

Carbon 349

and thermal stability; (3) the high accessibility of the active

phase and the absence of any microporosity (for MWCNT and

CNFs), thus eliminating diffusion and intraparticle mass trans-

fer in the reaction media; (4) the possibility of macroscopic

shaping of the support; (5) the possibility of tuning the specific

metal–support interactions, which can directly affect the cata-

lytic activity and selectivity; and (6) the possibility of confine-

ment effects in their inner cavity. Additionally, compared

to conventional supports, CNTs have a high flexibility for

the dispersion of the active phase as it is possible to: (1)

modulate their specific surface area (10–250 m2 g�1 for

CNFs, 50–500 m2 g�1 for MWCNTs, and 400–900 m2 g�1

for SWCNTs) or their internal diameter (5–100 nm for

MWCNTs); (2) easily functionalize chemically their surfaces;

(3) change their chemical composition (nitrogen- or boron-

doped CNTs); and (4) deposit the catalytic phase either on

their external surface or in their inner cavity.

Besides CNTs and CNFs, the catalytic applications of other

carbon nanomaterials such as fullerenes and carbon nano-

onions,117 or recently nano-diamonds118,119 and carbon

nanohorns120 have been much less studied.

7.13.5.6 Carbon Aerogels

Carbon aerogels are nanostructured carbons obtained from the

carbonization of organic aerogels, which are prepared from the

sol–gel polycondensation of certain organic monomers.121

They are usually synthesized by the polycondensation of res-

orcinol and formaldehyde, via a sol–gel process, and subse-

quent pyrolysis. By varying the experimental conditions during

the sol–gel process, the macroscopic properties of aerogels

(density, pore size, and form (shape/size)) can be controlled.

The aerogel solid matrix is composed of interconnected

colloidal-like carbon particles or polymeric chains. After pyrol-

ysis, the resulting carbon aerogels are more electrically conduc-

tive than most ACs. Carbon aerogels derived from the pyrolysis

of resorcinol–formaldehyde are preferred as they tend to have

the highest porosity, high surface area (400–1000 m2 g�1),

uniform pore sizes (largely between 2 and 50 nm), and high

density. They can also be produced as monoliths, composites,

thin films, powders, or microspheres. The versatility of the

sol–gel process and the diversity of forms enable the construc-

tion of carbon electrodes from aerogel powders using a binder,

or the manufacture of monolithic, binder-less electrodes.

The activation of carbon aerogels results in a large increase

in BET surface area, from �650 to �1300 m2 g�1. All these

properties make them promising materials for application in

adsorption and (electro)catalysis.122,123

7.13.5.7 Glassy Carbons

Glassy carbon, also referred to as vitreous or polymeric carbon,

is produced by the thermal degradation of selected polymer

resins. The precursor resin is cured, carbonized very slowly,

and then heated to elevated temperatures. The physical proper-

ties of glassy carbon, a nongraphitizing carbon, are generally

dependent on the maximum heat treatment temperature, which

can vary from 873 to 3273 K. In addition to very high thermal

stability, its distinguishing properties include its extreme resis-

tance to chemical attack: it has been demonstrated that the rates

of oxidation of glassy carbon in oxygen, carbon dioxide, or

water vapor are lower than those of any other carbon. Glassy

carbons have little accessible surface area and a relatively low

density (�1.5 g cm�3). This is attributed to the presence of a

significant volume of isolated ‘closed’ pores (�30%v/v). These

pores are typically 1–5 nm in size and are formed by the cavities

created by randomly oriented and intertwined graphene sheets.

The resulting structure is very rigid and provides glassy carbons

with tensile and compressive strengths that are typically higher

than those for graphite. Glassy carbon also has a very low

electrical resistivity ((�3–8)�10�4 Ocm) and is therefore par-

ticularly suited for electrocatalysis.124,125 Another attractive fea-

ture of glassy carbon is that it can be produced as free-standing

films or thin sheets as well as powders. The isolated porosity of

glassy carbons can be opened by thermal oxidation processes to

give a material with a high specific surface area.126

7.13.5.8 Carbon Molecular Sieves

Carbon molecular sieves are a special class of ACs having small

pore sizes with a sharp distribution in a range of micropores, as

compared with other ACs. They are used for adsorbing and

eliminating gas and liquid phase adsorbates of very low con-

centration like ethylene gas adsorption to keep fruits and veg-

etables fresh; or filtering of hazardous gas in power plants. The

most important application of the carbon molecular sieves is

in gas-separation systems (the swing adsorption method).

These materials, exhibiting molecular sieving between

branched and unbranched hydrocarbons, have been used for

selective hydrogenation of alkene mixtures.127–129 Carbon

molecular sieves are also active for the oxidative dehydrogena-

tion and dehydration of a variety of substrates.130

7.13.6 Carbon as Catalyst

The physical and chemical properties of carbon materials,

mainly porosity and surface chemistry, make them suitable for

application in many catalytic processes. Traditionally, carbon

materials have been used as supports for catalysts in heteroge-

neous catalytic95,97,131,132 or electrocatalytic52 processes,

although their use as catalysts on their own is becoming more

and more common. The catalytic application of carbon mate-

rials could be backdated to the use of ACs in the treatment of

wastewater and gas. It has been proved that ACs display good

catalytic performance in the dechlorination and desulfation of

the waste gases. Thus, AC catalysts have been used for a long

time in the production of phosgene133 and sulfur halides.134

Anhydrous chlorine gas is reacted with high-purity carbonmon-

oxide in the presence of an AC catalyst producing phosgene,

some unwanted by-products, and considerable heat. The pro-

duction process is continuous with the raw materials carefully

metered and excess heat removed. The most unwanted by-

products are other chlorinated hydrocarbons such as carbon

tetrachloride. The corresponding technologies are well estab-

lished, although the mechanistic details are not known in

detail.135 A variety of AC has also been found to catalyze a

highly selective reaction between phosgene and formaldehyde

to produce dichloromethane and CO2.136 The production of

sulfuryl chloride by the reaction of chlorine with sulfur dioxide

Page 28: Comprehensive Inorganic Chemistry II || Carbon

Table 8 Reactions catalyzed by carbon catalysts

General classification Example

Oxidation–reduction SO2þ½ O2!SO3

NOþ½ O2!NO2

deNOx2 H2SþO2!S2þ2H2OC6H5C2H5þ½ O2!C6H5C2H3þH2OToxin oxidation (creatinine)Oxidation of industrial effluents (oxalicacid)

Hydrogenation–dehydrogenation

RXþH2!RHþHX (X¼Cl, Br)HCOOH!CO2þH2

CH3CHOHCH3!CH3COCH3þH2

Combination with halogens H2þBr2!2 HBrCOþCl2!COCl2 (phosgene)C2H4þ5 Cl2!C2Cl6þ4 HClSO2þCl2!SO2Cl2C6H5CH3þCl2!C6H5CH2ClþHCl

Decomposition 2 H2O2!2 H2OþO2

CH4!Cþ2H2

Dehydration, isomerization,and polymerization

HCOOH!H2OþCO3 C2H2!C6H6

a-Olefines!poly(a-olefines)a-Oxime!b-oxime

Emerging applications CMS for shape selectivity reactionsCWAO of organic pollutantsGasification of organics and biomassCO2 reforming of methaneMethylamines synthesis

Adapted from Rodriguez-Reinoso, F. Carbon 1998, 36, 159–175; Stuber, F; Font, J;

Fortuny, A; Bengoa, C; Eftaxias, A; Fabregat, A. Top. Catal. 2005, 33, 3–50;

Coughlin, R W., Ind. Eng. Chem. Prod. Res. Dev. 1969, 8, 12–23; Fidalgo, B.;

Angel Menendez, J. Chinese J. Catal. 2011, 32, 207–216.

350 Carbon

in the presence of a carbon catalyst is a well-known process.

Typically, chlorine and sulfur dioxide are dissolved in a solvent

as sulfuryl chloride before contact with the catalyst. Thionyl

chloride is typically made by treating sulfur dioxide with chlo-

rine and sulfur dichloride in the presence of a carbon catalyst.

Carbon catalysts are known to degrade during such processes.

Another important industrial application of carbon cata-

lysts is in fuel gas cleaning. The dry desulfurization, denitrifi-

cation, and air toxics removal process using activated coke was

originally researched and developed during the 1960s by

Bergbau Forschung in Germany. In these processes, the active

carbon acts simultaneously as an adsorbent and as a catalyst in

the temperature range 383–443 K, that is, under conditions

where the material is stable in the presence of oxygen.137,138

From 1992, Mitsui Mining developed a technology to produce

activated coke used in the dry DeSOx/DeNOx/Air Toxic

removal process based on the activated coke. The low temper-

ature process was developed over many years, culminating in

both pilot and demonstration plants. These tests have proven

the system’s capability of removing over 99% of the SOx. This

activated coke has a surface area of 150–500 m2 g�1 that is less

than a conventional AC. The quantity of SO2 adsorbed is

45–110 mg SO2 per g activated coke.139 Another important

use of carbon catalysts for environmental cleaning is the

removal of halogen from halogen-containing compounds.

Two basic approaches can be used: (1) gas phase-catalyzed

oxidation of the halogen-containing compounds to carbon

dioxide and the corresponding halogen-containing acid140

and (2) catalytic dehalogenation.141

Most of the reactions that are catalyzed by carbon catalysts

can be classified into one of the following groups: (1) oxidation–

reduction; (2) hydrogenation–dehydrogenation; (3) combina-

tion with halogens; and (4) decomposition. There are also

examples of the catalysis of dehydration, isomerization, and

polymerization reactions. Some of the most relevant reactions

catalyzed by carbon catalysts are summarized in Table 8.97,142–144

7.13.6.1 Influence of Carbon Properties on Catalysis

At the end of the 1960s, Coughlin143 suggested that the cata-

lytic activity and selectivity of carbon catalysts was related to

their electronic properties. As discussed in Section 7.13.3.2.1,

carbon materials can exhibit the properties of a conductor,

semiconductor, or insulator, depending on the methods of

pretreatment and preparation. By applying the right treatment,

the catalytic properties of carbon can be adjusted to suit a

specific application. Thermal and graphitization treatments

would favor metallic behavior; oxidation would tend to

localize p-electrons and lead to semiconductivity; and other

treatments can result in highly disordered carbon with an

insulator-type behavior. Moreover, in view of the strong elec-

tronic anisotropy of graphitic carbon, a broad spectrum of

crystalline properties can be possible within a given carbon

catalyst, which may explain the poor selectivity that is some-

times observed with carbon catalysts. More recently, the influ-

ence of heteroatoms (N, B, and P) on the catalytic activity

of carbons was interpreted in terms of semiconductor

properties.145 The insertion of nitrogen atoms into the graphite

lattice lowers the bandgap, leading to higher electron mobility

and lowering the electron work function at the carbon/fluid

interface. Thus, N-doped carbons exhibit enhanced catalytic

activity in electron transfer reactions, in the same way as semi-

conducting metal oxide catalysts.

Later on, the catalytic activity of carbon materials came to

be associated with their surface area. However, in several cases,

the correlation was not found when using ACs with different

surface areas and PSD for a given catalytic reaction.97 It means

that the catalytic performances of carbon materials should also

be related with the chemical nature of the surface of carbon

catalysts. The relationship between the chemical properties of

carbon materials and the catalytic activity has been studied for

several decades. Generally, two approaches have been widely

followed in the surface chemistry of carbon; one is a ‘solid state

chemistry’ approach and the other is an ‘organic surface

groups’ approach. The former focuses on the crystalline micro-

structure of carbon materials whereas the latter focuses on the

organic character of the surface groups. In the ‘solid state

chemistry’ approach, the defects on the surface of carbon

materials are considered as active sites as the edge-side carbon

atoms are more chemically reactive. The ‘organic surface

groups’ approach deals with the nature and the functionality

of surface complexes of oxygen and other compounds chemi-

sorbed at the surface defects. Obviously, the combination of

both approaches could lead to a deeper insight into the real

reaction process taking place on the surface of carbon mate-

rials. For instance, the dependence of the chemical nature of

Page 29: Comprehensive Inorganic Chemistry II || Carbon

Carbon 351

AC on the raw material and the preparation history was always

observed, suggesting that the microstructure should be the key

factor for the reaction activity. However, AC annealed in H2

exhibited no activity for the dehydration and dehydrogenation

of alcohols, while the oxidation treatment by nitric acid con-

siderably increased the activity of the same carbon by two

orders of magnitude, suggesting that the catalytic activity

should be attributed to the surface functionalities.146 Table 9

provides an overview of the various reactions that can be

catalyzed with carbon, together with the type of surface chem-

istry required or the nature of the active sites, when they have

been identified. The redox couple quinone/hydroquinone was

found to be involved in the oxidative dehydrogenation of

hydrocarbons, while carboxylic acid groups are the active

sites for the dehydration of alcohols. Carbon materials have

also been used in environmental catalysis for the removal of

SOx, NOx, and H2S from gaseous streams. Basic carbons were

the most active in these processes, particularly N-doped car-

bons. Linear correlations between the catalytic activity and the

concentration of pyridinic groups have been reported in the

oxidation of SO2 and in the selective catalytic reduction of NO

with ammonia. The oxidation of organic compounds in liquid

effluents is another environmental application of carbon cata-

lysts, using air, oxygen, ozone, or hydrogen peroxide as oxi-

dants. The reaction mechanisms in the liquid phase are more

complex; nevertheless, the following general conclusions can

be drawn: (1) the reaction mechanisms involve free radical

species; (2) basic carbons are the best catalysts; and (3) oxida-

tion of the organic compounds may occur both in the liquid

phase (homogeneous reaction) and on the catalyst surface

(heterogeneous reaction).

Of course, to be complete, a correlation between carbon

material properties and their performances as catalyst should

take into account the nature, the concentration, and the acces-

sibility of the active sites. Thus, such studies should use a series

Table 9 Various reactions catalyzed by carbon materials, and thetype of surface chemistry required or the nature of the active sites

Reactions Surface chemistry/active sites Reference

Gas phaseOxidativedehydrogenation

Quinones 148–150

Dehydration ofalcohols

Carboxylic acids 152

Dehydrogenationof alcohols

Lewis acids and basic sites 153

NOx reduction (SCRwith NH3)

Acidic surface oxides (carboxylicand lactone)þbasic sites(carbonyls or N5, N6)

154,155

NO oxidation Basic sites, vacancies 156SO2 oxidation Basic sites, pyridinic-N6 157H2S oxidation Basic sites 158Dehydrohalogenation Pyridinic nitrogen sites 159Liquid phaseHydrogen peroxidereactions

Basic sites 160

Catalytic ozonation Basic sites 161,162Catalytic wet airoxidation

Basic sites 163,164

of carbon materials of the same nature and origin, which are

produced with very similar textural properties and different

amounts of surface functional groups. Concerning the nature

of the active sites, the basal planes are not much reactive;

therefore, we may expect to find active sites essentially at the

edges of the graphene layers, where the unsaturated carbon

atoms may chemisorb oxygen, nitrogen, or sulfur, originating

surface groups such as those discussed in Section 7.13.4. These

groups can act as active sites in various acid–base or redox

reactions. The concentration of these active sites depends to a

large extent on the microcrystalline structure of the carbon

material. Small crystallites can expose more edges; therefore,

more surface groups can be formed, while the role of the basal

planes and the p electrons will become more important with

large crystallites. Thus, the orientation of the graphene layers

and the ratio between prismatic and basal plane areas may

affect the catalytic performance.147 As far as the accessibility

of the active sites is concerned, it will be conditioned by the

pore sizes, especially in the case of microporous materials, such

as AC, for which pore diffusion limitations become important,

particularly in liquid phase reactions. Therefore, deactivation

and diffusion phenomena will in general affect more strongly

the performance of microporous carbons. As a result, there has

been a need to develop mesoporous carbon catalysts such as

CNTs, aerogels, xerogels, and templated carbons. Finally, as

technical carbonmaterials may comprise significant amount of

impurities, as sulfur or various oxides, modification of catalytic

activity is sometimes observed.

7.13.6.2 Examples of Applications

7.13.6.2.1 Oxidative dehydrogenationThe styrene monomer is extensively used in the manufacture of

various important polymers or copolymers, so that industrial

nonoxidative dehydrogenation of ethylbenzene (DE) to sty-

rene is one of the ten most important industrial processes

producing as much as 20�106 t year�1 of styrene monomer.

This is a high energy-costing process, currently carried out in

vapor phase at 873–973 K on potassium-promoted iron oxide

catalysts, in the presence of a large excess amount of super-

heated steam.165 The active sites in this catalyst are KFeO2. As

indicated by the positive standard enthalpy of the reaction

(eqn [1]), this conversion requires high energy consumption.

Further, due to the thermodynamic equilibrium, themaximum

conversion of ethylbenzene is limited below 70%. Coke depo-

sition on the active catalyst sites, loss and redistribution of the

potassium promoter, changes in the oxidation state of iron,

and physical degradation of the catalyst structure are the main

causes of catalyst deactivation. Lastly, the high steam-to-

hydrocarbon (10:1) ratio results in the formation of by-

products such as benzene and toluene that limits the yield

and the purity of the resulting styrene.

DEð ÞC6H5C2H5 þH2O $ C6H5C2H3 þH2,DH0298

¼ þ117:5 kJ mol�1 [1]

ODEð ÞC6H5C2H5 þ 1�2 O2 ! C6H5C2H3 þH2O,DH0

298

¼ �124:3 kJ mol�1 [2]

The oxidative dehydrogenation of ethylbenzene (ODE), an

exothermic reaction (eqn [2]), is an alternative route to the

Page 30: Comprehensive Inorganic Chemistry II || Carbon

352 Carbon

production of styrene that presents as the main advantage the

fact that conversions are not limited by equilibrium and can be

carried out at lower temperatures (623–723 K). During the last

three decades, a group of catalysts was reported, such as alu-

mina and various metal oxides and phosphates, which showed

an activity and selectivity in the ODE to styrene comparable

to the iron catalyst. Evidence was gradually accumulated that

the active sites were not located on the initial catalyst surface,

but on a carbonaceous overlayer. This carbonaceous overlayer

is initially deposited on the surface, and it is one reason

why the catalysts exhibited induction periods in their

activities.166–171 Thus, a large variety of carbon or graphite

materials have been studied in the literature for ODE, includ-

ing ACs,149–151,172–175 graphites,176 CBs,177 ACFs,148 highly

ordered mesoporous carbon,178 CNFs,179 nanotubes,180–182

and ultradispersed diamond and onion-like carbons.183

During the first studies, the results were rationalized mainly

in terms of textural properties of the carbon catalysts, that is,

their surface area and pore sizes. Then, comparing the activity

of various carbon materials, Grunewald and Drago proposed

that in addition to their large surface area, other factors might

play a role, such as the surface structure and the adsorption

capacity.184 In a subsequent paper, the same team studied

various carbon materials differing in their texture. No direct

relationship could be found between the activity and the sur-

face area of the carbon materials, and it was observed that the

most active catalysts were those with more mesopores and

higher external surface area.174 The importance of the texture

of the carbon material was then addressed in various

papers.148,175,185,186 These reports show that materials with

larger pores, consisting mainly of mesopores and macropores,

show better performances in the ODE, the smaller micropores

being quickly blocked by carbon deposits. The most relevant

conclusions of these works are reported by Pereira et al.186 (1)

no direct proportionality between surface area and activity in

the ODE reaction was observed; (2) the pores of very small

dimensions were quickly blocked due to coke deposition dur-

ing the process; (3) narrowing the pore sizes of the original ACs

by coke deposition led to lower catalytic activity in the ODE

reaction; and (4) the textural effects were found to be impor-

tant up to an average pore width of 1.2 nm; for larger pores

sizes, the surface chemistry controls the catalyst performance.

Systematic studies on the role of surface chemistry in the ODE

were reported by Pereira et al.149–151 The main results reported

support the view that carbonyl/quinone groups on the surface

are the active sites for this reaction, in agreement with earlier

proposals.187,188 These studies, based on the comparison of

carbon materials with similar textural properties but different

amounts of various oxygenated groups, have permitted to estab-

lish a linear correlation between the amount of surface car-

bonyl/quinone groups and the catalytic activity for ODE:

a ¼ 3:87� 10�4 Q½ � þ 0:71

where the activity a is in mmol g�1 s�1 and [Q] is the concen-

tration of carbonyl/quinone groups (mmol g�1). This correla-

tion is not a straight line through the origin, as active sites are

generated on the carbon surface by oxygen present in the

reacting mixture.150 The experimental results were well

described by a kinetic model based on a redox mechanism of

the Mars-van-Krevelen type, where the quinone surface groups

are reduced to hydroquinone by adsorbed ethylbenzene, and

reoxidized back to quinone by oxygen, as shown in Figure 25.

Investigations by quasi-in situ XPS confirm that the carbonyl/

quinone and hydroxyl groups are involved in this reaction.189

Recent works on the use of nanocarbons in ODE basically

confirm the proposed mechanism based on a redox cycle

involving the quinone surface groups.147,190,191 This mecha-

nism involves several steps: (1) ethylbenzene adsorption at the

graphite step edges, (2) ethylbenzene reaction with the oxy-

genated species also located at the graphite step edges leading

to the DE to styrene, (3) the simultaneous transformation of

the dehydrogenating oxygen species to hydroxyl groups, which

remain at the graphite edges, (4) styrene desorption from the

carbon surface, (5) gas-phase oxygen activation on the basal

planes of the graphene layers, (6) oxygen diffusion to the pris-

matic planes with the hydroxyl groups, (7) reformation of the

basic, chinoidic oxygen functionalities from the activated oxy-

gen and the hydroxyl groups, (8) water desorption (Figure 25).

Due to their high active surface area, ACs were mainly

studied as promising catalysts for the ODE. In spite of appar-

ently promising results, the long-term stability of the catalyst is

poor, as the ODE occurs in the presence of oxygen at relatively

high temperatures, which affects the physicochemical pro-

perties of AC. In addition, coke is always formed on the

catalysts.149 Thus, commercialization of the AC as catalysts

for ODE to styrene is not possible. Accordingly, carbon cata-

lysts, if they are to be used in industrial applications, should be

further improved to achieve high catalytic activity and selectiv-

ity, and stability during long time of operation. The remarkable

properties of CNTs and CNFs, and particularly their high sur-

face area and thermal stability, have attracted the interest of the

catalysis community.177–180,182,183,189–191 It was reported that

sp2-bound carbon is required for the selective styrene forma-

tion, as sp3-bound carbon led to the production of benzene

instead of styrene. It has been shown that the microstructure of

sp2-bound carbon materials is of great importance in order to

obtain high and stable efficiencies. CNFs have shown the high-

est styrene yields at the highest ethylbenzene conversions as

compared to CB and graphite. The comparative study of CNFs

and CNTs of different structure has shown that more perfect

CNTs, produced by the arc-discharge technique, are the most

active catalysts in terms of reaction rates. Onion-like carbon

was found to be the most efficient catalyst for the ODH reac-

tion on a mass-referenced basis. CNTs and onions have shown

a high and stable efficiency in the ODE reaction. The radius of

curvature of the basic structural element of CNTs and onions

and also their high aspect ratio seem to provide a high density

of functional surface groups under reaction conditions. The

perfectness of these carbon nanostructures provides also

enough stability toward oxidation and is essential for gas

phase oxygen activation. Nevertheless, the good performances

that some of these carbon nanomaterials claim to offer must be

put into perspective. Indeed, most of the results have been

obtained at temperatures in the range 793–823 K, whereas

ACs give high styrene yields at 623 K, a temperature at which

CNTs show no significant activity. The main advantage of

CNTs over ACs is their higher stability toward oxidation. How-

ever, it is still hardly conceivable that these materials will

become an alternative to the existing catalysts used for the

Page 31: Comprehensive Inorganic Chemistry II || Carbon

Carbon 353

industrial production of styrene, which give conversions of 60–

75% and a selectivity of 85–95% in the temperature range

813–923 K. Besides ethylbenzene, the ODH of other alkenes

such as butadiene192 or 10-dihydroanthracene193 has also

been reported on CNT catalysts.

It is noteworthy that effort in the ODH reaction was mostly

put on the conversion of ethylbenzene, and only few works

have been done on the conversion of alkanes. One possible

reason is that the intermediate product obtained in the ODH

of ethylbenzene is much more stable than those obtained in

the ODH of alkanes, while the radical is stabilized by the

delocalized p bonds. The catalytic performances of various

carbon catalysts are shown in Table 10.

Table 10 Catalytic performance of carbon catalysts in the oxidative dehy

Catalysts Reactants Products

Coal Butane Butene, butadienCNTsCNTsP-doped CNTs

PropanePropaneButane

PropenePropeneButene

AC iso-Butane iso-ButeneP-doped graphitic mesoporous carbon iso-Butane iso-Butene

O

O

CH CH2

H

H

O

O

H2CCH3

H2O

Figure 25 Proposed mechanism for ODE on carbon materials. Reproducedfrom Elsevier.

The catalytic activities of coals were predominantly related

with the reaction temperature and the best catalytic perfor-

mance of about 7% yield C4 was achieved at 973 K. The com-

bustion of the catalyst was also observed in the work,

suggesting that the stability of carbon catalysts should be care-

fully considered.194 Various ACs were used for the ODH of

isobutane to isobutene, revealing a correlation between the

catalytic activity and the amount of oxygenated surface

groups.199 However, the formation of coke was also found in

the used catalysts. The catalytic activity of CNTs was also tested

for the ODH of propane to propene. It was found that a

significant catalytic performance could be also achieved at

high temperature, associated with the gasification of catalysts.

drogenation of light alkanes

T (K) Conv. (%) Selec. (%) Yield (%) Reference

e 973 �40 �17 7 194773673723

42620

402560

171.512

195196197,198

648 25 60 15 199673 2.1 90 1.9 200

HC CH2

HO

OH

½ O2

from Emig, G.; Hofmann, H. J. Catal. 1983, 182, 15–26, with permission

Page 32: Comprehensive Inorganic Chemistry II || Carbon

354 Carbon

However, phosphoric oxide addition could remarkably

decrease the reaction temperature with respect to the high

catalytic performance. For the ODH of butane to butene on

CNTs, it was confirmed that the catalytic behavior should be

attributed to the quinone groups as active sites. The adsorption

of butane and the following cleavage of C–H bonds is the rate-

determining step for the dehydrogenation of butane. Phospho-

ric addition could efficiently inhibit the total oxidation, con-

sequently improving the catalytic selectivity.

(b)(a)

H H H H HHHH

Figure 26 Chemisorbed hydrogen species on (a) zigzag graphite edgesites and (b) armchair graphite edge sites (the third valence is not used).

7.13.6.2.2 DehalogenationAnother important use of carbon catalysts for environmental

cleaning is the removal of halogen from halogen-containing

compounds. In technologically advanced countries, thermal

fast oxidation (incineration) is a primary choice for the disposal

of halogenated waste. However, halogen-containing aromatics

are recalcitrant to combustion, and incomplete oxidation yields

harmful products, including chloroarenes such as (poly)chlori-

nated benzenes, phenols, biphenyls (PCBs), dibenzo-p-dioxins

(PCDDs), and dibenzofurans (PCDFs). Strictly regulated, very

low emission levels must, and have been, achieved by increasing

processing temperatures and by using advanced, expensive air

pollution control technologies.

Two basic alternative approaches could be used: (1) gas

phase-catalyzed oxidation of the halogen-containing com-

pounds to carbon dioxide and the corresponding halogen-

containing acid and (2) catalytic dehalogenation. An important

difference between oxidative dehalogenation and hydro-

dehalogenation or dehydrohalogenation is that in the oxidative

method the carbon in the halogen-containing compounds is

transformed into carbon dioxide, while in hydrodehalogenation

or dehydrohalogenation the halogen removal from an aromatic

hydrocarbon is selective and the hydrocarbon skeleton is

maintained. The two methods address two different environ-

mental situations, that is, the removal of a high-concentration

contaminant (oxidation), or the selective removal of the halogen

from a hydrocarbon mixture containing some halogen com-

pounds. Catalytic hydrodehalogenation (eqn [3]) is receiving

increasing attention.141,201–203

R � X þH2 ! R �HþH� X [3]

where X¼halogen. The dehalogenation of chlorinated aromatic

compounds such as chlorobenzene or substituted chloroben-

zene can be achieved in the presence of ACs at temperatures

between 473 and 823 K. Dehalogenation of chlorobenzene to

HCl is complete at 763 K, rather than the 1173 K needed for the

mere gas phase reaction. Besides the industrial relevance of this

reaction, it is also interesting from a fundamental point of view

since its mechanism should involved hydrogen and Ph-X acti-

vation on carbon materials. The hydrogen activation is of par-

ticular importance for carbon and carbon supported catalysts in

hydroprocessing,204 and for hydrogen storage in carbon

materials.205 A proposed mechanism for dehalogenation of

substituted chlorobenzene involves the following steps202:

ZPhClþH2AC ! ZPhClH ð Þ þHAC ð Þ! ZPhHþ Cl þHAC ð Þ [4]

ZPhClþHAC ð Þ ! ZPhClH ð Þ þ AC! ZPhHþ Cl þAC [5]

ZPhClþ AC ! ZPh� AC Hð Þ þ Cl [6]

ZPhClþHAC ð Þ ! ZPh� AC Hð Þ þ Cl [7]

ZPhClþ AC ! ZPh� AC ð Þ þ Cl� AC ð Þ [8]

AC represents the polyaromatic carbon surface, also par-

tially hydrogenated to hydroaromatic radicals (HAC(•)) and

molecules (H2AC). It should be emphasized that carbons must

be able to activate hydrogen and to facilitate its transfer to

reactant molecules in order to be catalytically active in hydro-

processing reactions. The availability of such sites should

increase with increasing irregularities (graphite edge sites) of

carbon materials. Today, few experimental data are available

on the role of surface properties of carbons during hydrogen

activation. They suggest, as well as theoretical calculations, that

the reactivity follows the order zigzag edge>armchair

edge>basal plane (Figure 26).206 Chlorine atoms, if formed,

yield HCl by fast hydrogen abstraction from the carbon

surface. Equations [4] and [5] lead to the dehalogenated

benzenes – via cyclohexadienyl-type radicals, formed upon

ipso-addition of H stemming from the carbon – while eqns [6]

and [7] explain the chemical attachment of the arylmoiety to the

AC surface. Dissociative adsorption eqn [8] would lead to

intermediate (radical) species with both the aryl moiety and Cl

chemically attached to the surface. Depending on bond dissoci-

ation enthalpies, such intermediates may lead to irreversible

bonding with the surface, or to ZPhHþHCl. Discerning which

mechanism(s) prevails is a complicated task, which surely

deserves further experimental and theoretical studies.

A limiting factor of AC for this reaction is the relatively fast

catalyst deactivation, apparently by coking. Hence, possible

applications should be restricted to the conversion of streams

with low concentrations of (toxic) halogenated compounds.

Carbon-supported catalysts, particularly Ni207,208 and

Pd,209,210 yield higher activity and stability.

Another interesting process leading to dehalogenation is

the direct conversion of 1,2-dichloroethane into vinyl chloride

(dehydrochlorination). The thermal reaction at�820 K is gen-

erally used for the production of vinyl chloride. However, the

process suffers from heavy coke deposition on the reactor

walls, so that catalytic reactions operating at lower tempera-

tures were investigated. Carbons were found to catalyze the

dehydrochlorination of alkyl chlorides to the corresponding

alkenes; however, the activity of the carbon catalysts decreases

with time onstream. A considerable mass increase of the cata-

lyst and decrease of surface area after reaction has been made

responsible for blocking of the active sites.211 Rapidly, it was

discovered that carbon materials containing significant

amount of nitrogen such as polyacrylonitrile (PAN)-based

Page 33: Comprehensive Inorganic Chemistry II || Carbon

N

CH2ClCH2Cl CH2CHCl + HCl

CH2ClCH2Cl CH2CHCl + HCl

N+

N

HC

l CH2CHCl

PVC

Polyvinyl chloride

Acid–base reaction Radical reaction Radical polymerization

Radical polymerization

Carbonization

Figure 27 Reaction mechanisms proposed for dehydrochlorinationand catalyst deactivation on nitrogen-containing carbons. Reproducedfrom Sotowa, C.; Watanabe, Y.; Yatsunami, S.; Korai, Y.; Mochida, I.Appl. Catal. A: Gen. 1999, 180, 317–323, with permission from Elsevier.

Carbon 355

ACFs159 and NH3-treated AC or CB79 are particularly active for

this reaction. Selectivity for vinyl chloride of 99.9% was

obtained at 633 K. The side products were ethylene, acetylene,

and 1,1-dichloroethene (each <0.1%), and butadiene

appeared in a yield <0.01%.212,213 The results obtained with

a variety of PAN-ACFs suggest that a base-catalyzed reaction

(Figure 27) occurs on pyridine-like N-6 sites,159 while a radical

reaction is promoted on vacant sites on the ACF surface. The

HCl produced deactivates the basic sites, while vinyl chloride

polymerization and carbonization leads to coking and deacti-

vation, as shown in Figure 27. XPS of a PAN-ACF catalyst

showed signals of N-Q (higher intensity) and N-6 (see

Figure 21). Their intensity, in particular that of the N-6 peak,

decreased after reaction. This might be due to a thin overlayer

of PVC or ‘coke,’ and the stronger decrease of the N-6 signal

was associated with strong binding of HCl on the carbon

(strong Cl 2p signals in XPS). After heating to 873 K under

nitrogen, the N-6 peak was partly restored, but not the catalytic

activity.159 Alkyl bromides react at lower temperatures in

dehydrohalogenation reactions than the chlorides, but the

reactions are complex as brominated precursor molecules

appear, indicating that some bromine is formed from the

hydrogen bromide79 (Figure 27).

7.13.7 Carbon as Catalyst Support

The application of carbon and graphite materials in catalysis

is mainly as support for active phases. Carbon-supported

metal catalysts are employed in a number of applications

including hydrodesulfurization (HDS) of petroleum, hydrode-

nitrogenation (HDN), dehydrohalogenation, hydrogenation

of CO, hydrogenation of halogenated nitroaromatics com-

pounds and nitro compounds, hydrogenation of unsaturated

fatty acids, hydrogenation of alkenes and alkynes, oxidation of

organic compounds and organic pollutants, as well as for fuel

cells.95–97,112–115,131,204,214,215 The large-scale synthesis

of vinyl acetate and vinyl chloride and the desulfurization

of natural gas on AC impregnated with ZnO, CuO, or Fe2O3

are important technical applications.216 Additionally, in the

Catalytic Reaction Guide published by Johnson Matthey, of

69 reactions of industrial interest catalyzed by noble metals,

50 used carbon materials as catalyst support. In the past, the

lack of a fundamental understanding of many aspects of

the use of carbon in catalysis caused a limited application of

carbon as catalyst and still more as catalyst support. However,

the continuous studies to better understand all aspects of the

physical and chemical characteristics of carbon material, espe-

cially for AC (surface area and porosity) and even more, the

possibility of controlling the surface chemistry of such mate-

rials, are the origin of important researches carried out in

industrial chemistry during the last decade. The main advan-

tages of using carbon or graphitematerials in catalysis, as well as

the main carbon and graphite materials relevant to catalysis

have been reviewed in Section 7.13.5. Among the carbonmate-

rials, ACs, with their large surface area are by far the most used

in catalysis, followed by CBs for fuel cell applications. The

recent interest for nanocarbons, which constitute interesting

model supports and show specific properties of interest for

electro- and photocatalysis, is worth noting. In this section, we

concentrate on the role of carbon properties on the design of

the supported catalyst and on the final catalytic performances.

7.13.7.1 Role of Carbon Properties on Performancesof the Support

In heterogeneous catalysis, carbon materials have been used

for a long time because they can be used directly as catalysts,

and moreover, they can satisfy most of the properties desired

for a suitable support. It has been shown that although the

surface area and the shape of carbon porosity may be very

important in the preparation and the properties of the corre-

sponding catalysts, the role of carbon surface chemistry is also

extremely important. The number and the strength of the

surface groups influence the apparent acidity or the basicity

of the carbon surface. Moreover, as already stated a few percent

of inorganic matter from the organic precursor (e.g., coal or

wood) could be present on carbons and this fact is considered

important for their performance, especially when they are used

as catalyst supports. However, although the catalytic effect is

mainly headed by the chemical properties of the active phase,

the dispersion and the local distribution of the active phase

across the AC support as well as the interaction between the

active phase and the support are significantly important. These

are the aspects of carbon supports that make them so attractive

for heterogeneous catalysis. These parameters can be modified

to satisfy any specific requirement during catalyst preparation

making the surface not only physically but also chemically

accessible to the precursor and diminishing the deactivation

by sintering. It seems obvious that the future is promising once

it is understood that not only the surface area and the porosity

of carbon materials are the key for their use, but also both

physical and chemical surface properties of carbon support

have to be taken into account when designing a solid catalyst.

7.13.7.1.1 Surface area and porosityBoth high surface area and a well-developed porosity are very

important for achieving a high dispersion (fraction of metal

atoms that are on the surface of the support in relation to the

total metal loading) of the active phase in the catalyst. Carbon

materials, especially AC, exhibit surface areas significantly

Page 34: Comprehensive Inorganic Chemistry II || Carbon

356 Carbon

higher than other conventional oxide catalyst supports

(Table 11). However, a great part of this surface area may be

due to microporosity and, in this case, it may not be available

to precursors or reactants.

Many studies report the effect of porosity and surface area on

metal dispersion and catalytic activity. As far as metal dispersion

is concerned, it is difficult to separate the influence of surface

area from that of surface chemistry. Indeed, irrespective of the

model applied to describe the adsorption properties of a metal-

precursor during ion adsorption and impregnation, all authors

agree on the fact that the surface composition of carbon plays a

crucial role in the adsorption of metal ions. Often, oxygen

groups are introduced via oxidation of the surface of the carbon

(see Section 7.13.4) to enhance the adsorption of the metal in

order to obtain a high dispersion and a high metal loading.

However also the role of basic sites, being either oxygen-

containing groups such as chromene and pyrone-like structures

or p-sites, is claimed to be of importance.217

An interesting study shows the separate influence of poros-

ity, surface chemistry, and ash contents on the performances of

MoS2/AC catalysts.218 Several sulfided Mo catalysts supported

on ACs were prepared in order to show the effect of the poros-

ity, ash content, and oxygen surface groups of the carbon

supports on the preparation and the thiophene hydrodesul-

furation (HDS) activity of the catalysts. The five original ACs

were washed with HCl to eliminate the ash content, reduced in

hydrogen at high temperature to eliminate oxygen surface

groups, and oxidized with hydrogen peroxide to introduce

new oxygen surface groups, the porosity remaining constant

after these treatments. In this way, the effect of the surface

properties of the support in both the preparation and the

activity of the catalysts could be evaluated separately. The effect

of these properties on the preparation (adsorption from

ammonium heptamolybdate) of the catalysts is as follows:

(1) increasing surface area and porosity of the carbon facilitates

the loading of Mo; (2) the ash content has a relatively small

(but significant) negative effect in the adsorption of the metal

precursor; and (3) oxygen surface groups of the carbon support

increase the adsorption of the metal precursor. The effect on

HDS activity is: (1) specific activity increases with surface area

Table 11 Main features of conventional catalyst supports

Carrier Features

Activatedcarbon

Surface area: 800–1500 m2 g�1

Heat stabilityAlumina Surface area: 100–300 m2 g�1

Type a, g, and Z are often used as carriersReasonable priceHeat resistanceAlkali resistance

Silica Surface area: 200–600 m2 g�1

Zeolite Surface area: 350–900 m2 g�1

Type A, X, Y mordenite, erionite, and ZSM-5 areoften used as carriers

High controllability of pore sizeTitania Surface area: 40–100 m2 g�1

Magnesia Surface area: 50–200 m2 g�1

BasicityStrong adsorption of carbon dioxide and water in air

of the support up to about 1000 m2 g�1, remaining almost

constant thereafter; (2) the ash content has a negative effect

on the catalytic activity; and (3) the oxygen surface groups of

the carbon support have a negative effect on the catalytic

activity. Another study deals with different 4–5 wt% Pd/AC

catalysts prepared by the deposition/precipitation method on

ACs of different textural properties but similar surface chemis-

try. These catalysts differ by the Pd dispersion from 0.13 to

0.39, that is to say, Pd particles from ca. 8.5–2.8 nm, respec-

tively. It was shown that the prime parameter that determines

the Pd particle size is the extent of AC surface in the meso- and

macropores. A large value favors the formation of Pd particles

of small sizes (2–3 nm).219 Mesopore volume andmeanmeso-

pore size have also been identified as being important param-

eters to control Pt particle size and dispersion in carbon

aerogel-supported catalysts containing 2 wt% Pt.220 For elec-

trocatalysis (direct methanol fuel cells) on PtRu catalysts

deposited on Sibunit (a mesoporous carbon support devel-

oped in Boreskov Institute of Catalysis, Novosibirsk, Russia),

a high PtRu surface utilization and facilitated diffusion in

macropores was achieved by using supports of low specific

surface area (SBET between 20 and 70 m2 g�1) instead of high

surface area (200–400 m2 g�1).221 For the oxygen reduction

reaction on Pt/AC, an increase in the specific surface area and

mean pore diameter increased the catalytic activity due to the

enhanced mass transfer in the pores of the support.222

Studies on confinement effects in carbon-based catalysts

have recently been published.114,223,224 The confinement effects

that influence chemical reactions can be classify into three

groups: (1) shape-catalytic effects, that is, the effect of the

shape of the confining material and/or the reduced dimension-

ality of the porous space, (2) physical (or ‘soft’) effects including

the influence of dispersion and electrostatic interactions with

the confining material, and (3) chemical (or ‘hard’) effects –

interactions that involve significant electron rearrangement,

including the formation and breaking of chemical bonds with

the confiningmaterial.225 The last is usually considered to be the

actual catalytic effect, and it is the one that has the most obvious

influence on the reaction rates, as it alters the reaction mecha-

nism. However, the first and second types of effects can also

have a strong influence on both the rates and equilibrium yields,

as has been shown in several recent theoretical calculations226

and experimental studies.227 However, in some cases, a high

surface area of the carbon support may be detrimental if it is

confined in narrow micropores which are not accessible to the

reactant molecules. This is especially important in processes

where large molecules are involved, as in the treatment of petro-

leum feedstocks, and in liquid phase reactions in which diffu-

sion of reactants and/or products may be hindered by the

narrow porosity. Some authors have found that the shape of

the pores can play an important role in the catalytic process

when it is used as support. Laine et al. have used AC as support

for NiMo catalysts for HDS.228,229 On the basis of their results,

these authors suggested that the narrow slit-shaped pores in AC

are able to lower the vapor pressure of sulfur to such an extent

that they create a driving force for sulfur transfer from the active

compound to the micropores, this process forming active

vacancies in the metal sulfide. This so-called ‘sink effect’ was

not observed in microporous silica, whose pores are not slit-

shaped. Shape selectivity has also been reported for a variety of

Page 35: Comprehensive Inorganic Chemistry II || Carbon

Carbon 357

microporous carbon, as for hydrogenation,230–232 methanol

decomposition,233 or Fischer–Tropsch reaction.234 For the

methanol decomposition reaction, CH3OH is known to be

decomposed to CO and H2 first, then CH4, CO2, and H2O are

produced through reactions between CO andH2. It was possible

to produce only CO and H2, by using nickel supported on a

molecular sieving carbon catalyst with sharp pore distributions

around 0.45 nm in diameter.228

There are also reports indicating that the carbon structural

properties do not affect either the dispersion or the catalytic

activity. An important factor influencing the dispersion of the

active phase is the nature of the precursor. Rodrıguez-Reinoso

et al. used two different iron precursors (iron nitrate in aque-

ous solution and iron pentacarbonyl in organic solution) to

prepare Fe/AC with different PSDs.235 They obtained an

increase in iron dispersion with the support surface area for

the nitrate series, but a high and unaffected dispersion was

found for the [Fe(CO)5] series. Similarly, for iron supported

on a carbon nanosphere, iron acetate yielded higher dispersion

than iron nitrate, allowing better performances for Fischer–

Tropsch synthesis.236 In the case of bimetallic catalysts, an

influence of the nature of the precursor has also been noticed.

Devillers et al. studied the influence of metallic precursors on

the properties of carbon-supported bismuth-promoted palla-

dium catalysts for the selective oxidation of glucose to gluconic

acid.237 They used different precursors for Pd and Bi deposition

and found that the best catalytic activities were obtained when

using acetate precursors instead of classical salts as chlorides or

nitrates. However, for Ni-W/AC catalysts used in phenol hydro-

deoxygenation, it was shown that the effect of a tungsten pre-

cursor (silicotungstic, phosphotungstic, and tungstic acids) on

catalytic performances was negligible.238 In another study,

Li et al. investigated the catalytic behavior of ruthenium catalysts

supported on carbon materials with different porous and gra-

phitic structures in the catalytic ammonia decomposition.239

They found that the catalytic activity followed the trend: Ru/GC

(graphitic carbon)>Ru/CNTs>Ru/CB-C>Ru/CMK-3 (meso-

porous carbon)¼Ru/AC. It was concluded that the graphitic

structure of the carbons was critical to the activity of the sup-

ported ruthenium catalysts, whereas the surface area and the

porosity were less important.

It appears that a large surface area formed by accessible

pores is important to obtain highly dispersed and active cata-

lysts. However, there are some other carbon characteristics that

have to be taken into account in order to explain the catalytic

behavior of carbon-supported catalysts. One of the most

important is the surface chemistry.

7.13.7.1.2 Surface chemical propertiesThe presence of surface groups that contains heteroatoms

(O, N, and H) can affect the preparation of carbon-supported

catalysts, as they confer to the carbon surface an acid–base and

hydrophilic character. The presence of these groups can also

affect the adsorption/desorption phenomena and, thus,

impact the catalytic activity.

The first attempt to clarify the influence of surface func-

tional groups on the catalytic behavior of carbon-supported

catalysts was made in the landmark paper of Derbyshire et al.

on the influence of surface functionality on the activity of

carbon-supported molybdenum sulfide catalysts.240 In this

work, it is stressed that the affinity between a particular carbon

surface and a selected catalyst precursor will depend upon the

compatibility of the two chemical structures. Apart from the

effects of wetting and PSD, carbon surface functionality gov-

erns the extent of adsorption of the catalyst precursor and the

extent of its reduction or conversion to active state. First, it

provides the anchoring sites for the metal precursor. For exam-

ple, carboxyl groups are the sites where ion exchange with

cationic precursors takes place. Second, the optimum surface

chemistry allows favorable electrostatic (coulombic) interac-

tion between the support and the metal precursor. For exam-

ple, adsorption of anions occurs only when the pH of the

solution containing the catalyst precursor is lower than the

point of zero charge of the carbon. Third, the optimum surface

chemistry prevents excessive catalyst mobility on the support

surface. For example, the relatively stable carbonyl groups

appear to be quite effective for this purpose. Fourth, the opti-

mum surface chemistry also facilitates catalyst conversion into

its catalytically active state. This was nicely illustrated in the

work of Prado-Burguete et al. who reported studies on the role

of surface oxygen groups on the dispersion and resistance to

sintering of Pt/C catalysts.241,242 They prepared a number of

CB supports with similar porosities but different amounts of

oxygen surface groups, which were impregnated with H2PtCl6solutions. It was observed that the more acidic groups created

by the oxidizing treatment with H2O2, the less hydrophobic

was the CB surface and that made the surface more accessible

to the aqueous solution of the metal precursor upon the

impregnation procedure. Thus, the platinum dispersion

increased with increasing amount of oxygen surface groups.

However, the less acidic and more thermally stable surface

groups favored the interaction between the metal precursor

or the metal particle with the carbon surface, this minimizing

the sintering of the platinum particles. Similar results have

been obtained in the preparation of K-promoted Ru/C catalysts

for ammonia synthesis, although in this case the AC was oxi-

dized with HNO3.243 The authors concluded that the presence

of oxygen surface groups improved the hydrophilic character of

the carbon surface, thus enhancing the dispersion of K and Ru

and the catalytic activity. However, it has to be taken into

account that some oxygen groups may not be stable under the

heat treatment conditions (i.e., reduction) to which the catalysts

are subjected during the preparation stage to obtain the active

phase. Thus, if oxygen-containing groups play a crucial role in

wetting of carbon supports whether they also function as

anchoring sites for metal particles is still a matter of debate.217

The presence or absence of surface functionalities can also

directly affect the catalytic behavior of the active phase.

Although we will discuss this topic in detail in the next section

(Section 7.13.7.2.2), we will give hereafter some representative

recent examples. The effect of acidic treatments on N2O reduc-

tion over Ni catalysts supported on AC was systematically

studied by Lu et al.244 It was found that surface chemistry

plays an important role in the N2O-carbon reaction catalyzed

by a Ni catalyst. HNO3 treatment produces significant amount

of active acidic surface groups such as carboxyl and lactone,

resulting in uniform catalyst dispersion and a higher catalytic

activity. HCl treatment of the pristine AC decreases active

acidic groups and increases the inactive groups, playing

an opposite role in catalyst dispersion and catalytic activity.

Page 36: Comprehensive Inorganic Chemistry II || Carbon

358 Carbon

A strong effect of the AC surface chemistry on the activity of Au/

AC catalysts was also reported for glycerol oxidation.245 Gold

particles with similar average sizes resulted in different perfor-

mances, depending on the amount of oxygenated groups on

the surface of the support used. Basic oxygen-free supports,

characterized by a high density of free p-electrons, lead to an

enhancement of the gold catalyst activity. These characteristics

can easily be achieved by thermal treatments at high tempera-

tures, which remove the oxygen-containing surface groups. The

role of the AC surface chemistry was explained by considering

the capability of oxygen-free supports to promote electron

mobility. The mobility of the electrons from and to the gold

surface promotes both adsorption and regeneration of hydrox-

ide ions, which are necessary for the oxidation reaction. A clear

effect of the oxygen surface groups created by HNO3 oxidation

was also reported for Pt/CNT catalysts involved in different

probe reactions, including both steam reforming and liquid

phase reforming of hydrocarbon oxygenates and dehydroge-

nation of alkanes in the liquid and gas phases.246 Compared to

the pristine CNT supports, Pt dispersion is improved on

HNO3-treated supports, but the turnover frequency of aqueous

phase reforming decreases by half. Higher turnover frequencies

can be obtained by removing the oxygen surface groups via

high temperature annealing. A comparison of the results

obtained in the different reactions suggests that the oxygen

surface groups are only detrimental to reactions in a binary

mixture with two components of different hydrophilicities due

to their competitive adsorption onto the catalyst supports.

Although less studied than oxygen, nitrogen surface groups

can also influence metal dispersion and catalytic activity. The

effect of these nitrogen functionalities depends on the system

studied. In this way, Derbyshire et al. obtained more active Mo

catalysts for HDS by prenitriding the carbon support,240 and

Guerrero-Ruiz et al. found the same effect for Fe/C and Ru/C

catalysts.247 Pd nanoparticles have been deposited on N-doped

nanocarbons and compared with undoped catalysts prepared

on CB for the direct synthesis of H2O2.248 The Pd on N-doped

CNT gives high productivities to H2O2. The introduction of

nitrogen in the CNT structure favors not only the dispersion of

Pd but also the specific turnover. Stone-fruit AC and modified

supports containing acidic oxygen (HNO3 oxidation) and

basic nitrogen groups (treatment with nitrogen-containing

compounds) have been used to prepare palladium catalysts

by wet impregnation.249 The influence of the nature of the

functional groups on the dispersion and the oxidation state

of palladium and its activity in hydrogen oxidation have been

investigated. Pd dispersion was found to increase with the

basic strength of functional groups on the support. XPS studies

have shown that introduction of amine surface groups results

in an increased proportion of Pd0, which is resistant to reox-

idation. Palladium catalysts supported on AC modified by

diethylamine groups are found to exhibit the highest metal

dispersion and greatest activity in hydrogen oxidation. In

another study, ruthenium catalysts were supported on two

different carbon materials, CNT and bamboo-like CNT doped

with nitrogen.250 The Ru catalysts were tested in the catalytic

ammonia decomposition reaction. The catalytic activity of Ru

particles was significantly improved when supported on CNTs

doped with nitrogen. However, Wachowski et al. studied the

polymerization of styrene with the [CpTiCl2(OC6H4Cl-p)]

catalyst supported on carbon materials with different degrees

of coalification, and analyzed the effect of the modification of

the support by nitrogen on the efficiency of the catalytic system

in the polymerization of styrene.251 It was found that the

introduction of nitrogen functionalities on the carbon surface

lowered the catalytic activity of these systems.

7.13.7.1.3 InertnessCompared to conventional oxide supports such as silica,

alumina, titania, or ceria, it is clear that the carbon surface

(in spite of the presence of surface hetero-atoms) is less reac-

tive. Thus, Milone et al.,252 who investigated the influence of

the support (SiO2 25–360 m2 g�1 or AC 1100 m2 g�1) on the

hydrogenation of citronellal on Ru-supported catalysts, show

that the main products obtained on the Ru/SiO2 catalyst were

the unsaturated cyclic alcohols produced by citronellal isom-

erization on the SiO2 surface. However, the main reaction

products on Ru/AC were the open chain hydrogenated prod-

ucts, and this was attributed to the low activity of the carbon

surface toward the isomerization reaction.

The carbon surface inertness may also significantly impact

catalyst preparation. An analysis of the literature suggests that

there are at least three different characteristics of carbon that

can be utilized to generate metal surfaces not found on refrac-

tory oxide supports.253 First, on graphitic carbon many metals

interact very weakly, allowing bimetallic particles to form

structures identical to those anticipated for bulk materials. Of

particular significance is the formation of true alloys, both in

the bulk and on the (catalytic) surface of the bimetallic parti-

cles. In contrast, on conventional refractory-oxide supports

these same structures will not form for certain base-metal/

noble-metal pairs. Instead, a preferential and strong interac-

tion between the more ‘base’ metal and the support generally

leads to preferential segregation of that metal to the refractory

oxide interface and, concomitantly, dominance of the catalytic

interface by the ‘more noble’ metal. As a result of these struc-

tural differences, the catalytic chemistry, both activity and

selectivity, of some bimetallic particles supported on refractory

oxides and graphitic carbons is dramatically different. One

clear example is the preparation of bimetallic PtSn254,255

or PtRu256 catalysts for selective hydrogenation of a,b-unsaturated aldehydes. The catalytic behavior of this system

is determined by, at least, three aspects that determine the

catalytic activity and the selectivity toward the desired product:

(1) the oxidation state of the promoter (Ru or Sn) in the

catalyst; (2) the possibility of formation of alloyed phases;

and (3) the extent of metal–support interactions. It is particu-

larly interesting to obtain the alloyed phase, and this is deter-

mined by the easiness of reduction of the tin species which, in

turn, depends on the interaction between the tin precursor and

the support. For PtSn/TiO2 catalysts it was shown by XPS that

even after reduction under hydrogen at 773 K, the main part of

tin (about 78%) was in an oxidized state, thus limiting the

possibility of formation of alloy phases. However, a nearly

complete reduction of tin to the metallic state can be achieved

by using carbon materials as supports, depending on the

preparation method and on the relative amount of tin.254,255

Second, it is clear that it is possible to directly bond metals

to unsaturated active sites on high surface area CBs, AC, etc.

This has been demonstrated to yield thermally stable particles

Page 37: Comprehensive Inorganic Chemistry II || Carbon

Carbon 359

of a unique structure.257 Thus, on CB it was demonstrated that

iron particles are more resistant to sintering if deposited on an

oxygen-free support than on a functionalized support. How-

ever, for Pd nanoparticles deposited on SWCNTs the disper-

sion of Pd nanoclusters was enhanced by creating defects on

the nanotube walls, which lead to a stronger metal–support

interaction.258 DFT calculations have shown that the binding

energy of Pd is significantly enhanced when the CNT surface is

oxygen-functionalized, compared to the case of the pristine

SWCNT surface. The electronic interaction of Pd atoms with

oxygen at the defect sites results in a stronger bonding. These

calculations were consistent with experimental measurements,

as microscopy images clearly show that the functionalized

SWCNT surface is muchmore effective than the pristine surface

in anchoring Pd nanoclusters. However, strong interaction

between Pd clusters and the carbon surface can be effective

without functionalization, and a molecular orbital study has

shown that supported Pd0 atoms and clusters are strongly

bound to unsaturated and defect surface sites of the AC

surface.259 In such positions, the interaction of Pd atoms

with the support is much stronger than that with each other.

That provides the driving force for the atomic dispersity of Pd/

AC catalysts. Similar conclusions were drawn from DFT calcu-

lation and TEM observations in the case of Au clusters depos-

ited on CNTs.260 Indeed, both experimental and theoretical

studies show that surface defects are the anchoring sites of Au

nanoparticles. Very weak binding is identified between Au

clusters and pristine CNTs, or defective CNTs when Au clusters

are located far from the vacancy. It is only when Au clusters are

directly located on the defect site that very strong interactions

are found. The strong interaction is caused by the charge trans-

fer from Au to the defect. This study suggests that defects of

CNTs can be used for the grafting of Au nanoparticles without

requiring surface oxidation. Potentially, the dispersion of Au

clusters can be controlled or regulated by controlling the den-

sity of defects. For Co/CNF catalysts, the interaction of precip-

itated cobalt oxides with the CNF support was shown to be

influenced to a large extent by heat treatment in an inert

atmosphere.261 Heat treatment of the CNF at 573 K resulted

in an increase in the interaction between the cobalt particles

and the support. Under these conditions, a small amount of

cobalt carbide and cobalt metal was detected by the XRD and

XPS analyses. Heat treatment at 873 K resulted in a further

increase in the interaction between the metal and the support

leading to increasing amounts of cobalt carbide and cobalt

metal. On refractory oxides, strong interaction generally leads

to the creation of complex, ionic-bonded ‘interface’ phases.

Third, the carbon structure can be manipulated to generate

shape-selective supports. This can also be done with refractory

oxides, but only carbon surfaces are neutral. Thus, only on

carbon will reduced metal readily form. The easiest metal

reduction on carbon support compared to conventional

oxide supports has been evidenced in the case of iron cata-

lysts.262–264 The inertness of the carbon surface facilitates the

presence of zero-valent iron in the catalyst, which is more

difficult in the case of other supports, such as alumina, on

which the reduction of the oxidized iron species in hindered.

Vannice et al. carried out an extensive study of carbon-

supported iron catalysts using different carbons and pre-

paration methods, and concluded that highly dispersed

Fe/C catalysts could be prepared on high surface area carbons

due to the weak chemical interactions between oxidized iron

precursors and the carbon surface.262–264 There is surprisingly

little research on these phenomena, suggesting there are many

opportunities to create unique metal surfaces using carbon as a

support. Finally, it is worth noting that an ‘inert,’ signifying not

functionalized, carbon materials can present electron conduc-

tivity provided the size of the graphite-like crystallites is large

enough. An increase in the concentration of oxygen-containing

surface functionalities will result in an increase of the work

function and concomitant increase in the resistance of com-

pacted carbon powders.52,265

7.13.7.2 Examples of Applications

7.13.7.2.1 Hydrotreating reactionsThe industrial hydrotreating reactions HDS and HDN are com-

monly carried out on cobalt–molybdenum and nickel–

molybdenum systems. The majority of all metal sulfide hydro-

treating catalysts are distributed as adsorbed particles over a

support before being applied as pellets in a hydrotreating

reactor. The purpose of the support is to increase the activity

of the catalyst by increasing the exposure of the active sites to

the reactants while still maintaining the mechanical strength of

the material. g-Alumina is almost exclusively used as a hydro-

treating catalyst support in industry. Significant efforts have

been made in an attempt to improve upon its effectiveness.

A review by Luck266 outlines five objectives for finding a supe-

rior catalyst support: (1) improving the dispersion of the bime-

tallic sulfides; (2) reducing the strong interaction between the

active component and the support while the active component

is in the initial oxide phase; (3) decreasing the spinel phase

concentration of g-alumina, increasing the usability of the cat-

alyst promoters; and (4) improving the recovery potential for

catalyst metals. Carbon- or graphite-supported catalysts have

been used in several studies of hydroprocessing of model com-

pounds and real feeds.204 The most frequently used carbons,

such as AC and CBs, are themselves active for hydrogenation,

HDS, HDN, and hydrodemetallization. This activity is attrib-

uted to the ability of carbons to facilitate H2 activation. After

desorption, active hydrogen is transferred to reactantmolecules

(by surface migration or spill-over) to initiate hydroprocessing

reactions. The H2 activation by carbons and consequently the

hydrogenation activity increases with increasing temperature.

This is one of the reasons for the different behavior of carbon-

supported catalysts compared with traditionally used g-Al2O3-

supported catalysts containing the same amount of active

metals. Because of the active hydrogen present, carbon support

is much more resistant to deactivation by coke deposition than

g-Al2O3 support.204 Other potential advantages, which are

based on the textural and the surface properties of the carbon

supports, includes weaker metal–support interactions and

higher activities per unit mass of catalyst.267

Concerning the strength of the metal–support interactions

under HDS conditions, DFT calculations have permitted us to

establish the following rank order of the supports for molyb-

denum catalysts, representing strong interaction tendency:

SiO2<carbon<Al2O3<TiO2<ZrO2<Y2O3.268 These strong

interactions make difficult the complete sulfidation and acti-

vation of the catalyst metal to occur due to the formation of

Page 38: Comprehensive Inorganic Chemistry II || Carbon

360 Carbon

metal aluminates.269 The strongmetal–support interactions also

contribute to accelerated catalyst deactivation occurring due to

sintering of the catalyst’s active phase.270 There is a significant

difference between the interactions of active metals with carbon

supports compared with g-Al2O3 supports. On Co–Mo catalysts,

it was observed that the activity of the carbon-supported catalyst

was similar to that of the alumina-supported sample, although

the butane yield was much higher on the former.271 The hydro-

genation activity of this systemhas been related to the edge plane

of MoS2, whereas the desulfurization activity depends on the

basal plane.272 The obtained results can be rationalized on the

basis of the epitaxial growth of MoS2 crystallites, with a high

basal-to-edge area ratio, on the Al2O3 support. This phenome-

non is favored by the interaction between the metal precursors

and the alumina surface, and it has been also evidenced in

graphite-supported catalysts. Furthermore, other authors have

found that carbon-supported Co–Mo catalysts were more active

than the alumina-supported counterparts for both dibenzothio-

phene and 4,6-dimethyldibenzothiophene HDS.273 It seems

clear that the weak interaction between the carbon surface and

the metal precursors is the origin of the higher activity of these

systems, as it facilitates a higher degree of sulfurization and the

formation of themore activeCo–Mo–S type II structure.274More

recently, it has been suggested that carbon could stabilize MoS2particles, avoiding the sintering of small crystallites and yielding

a higher dispersion on the carbon support.275

The works by Prins, de Beer and coworkers represented a

great contribution to the knowledge on the structure of carbon-

supported hydrotreating catalysts, as well as on the kinetics

and the mechanisms of these reactions. They first reported

that the activity of carbon-supported molybdenum and tung-

sten catalysts was higher than that of their counterparts sup-

ported on alumina or silica.276,277 They also compared the

catalytic behavior of sulfided Co–Mo, Fe, and Mo catalysts

supported on alumina, carbon black composite (CBC), and

AC,278 and observed that, for a given metal, very important

differences in activity were found both between the carbons

and alumina, and even between the carbons themselves, the

highest activity being obtained for the AC-supported catalyst.

All these effects were explained on the basis of the relatively

low interaction between the metal sulfide and the surface of

the carbon support, which facilitates the sulfurization of the

precursor and the dispersion of the active phase.279

Because of limited information on long-term performance,

the stability of the active phase on carbon supports has not yet

been fully determined. Thus, because of the diminished inter-

action, sintering of active metals is more likely to occur on

carbon-supported catalysts than on the g-Al2O3-supported cat-

alysts unless a bonding between the active phase and the

carbon is facilitated. This would result in the presence of a

new type of active phase such as Co–Mo–C(S).280 Some exper-

imental and theoretical evidences support the coexistence of

such phases, that is, Co–Mo–S and Co–Mo–C(S) phases, in

hydroprocessing carbon-supported catalysts.281–283 The cata-

lysts comprising other oxidic supports (e.g., SiO2, SiO2–Al2O3,

TiO2, zeolites, etc.) in comparison with carbon-supported cat-

alysts have been used to a much lesser extent.

The studies on carbon-supported catalysts for hydroproces-

sing have been dominated by conventional metals such as

Mo, W, Ni, and Co. Among carbon supports, AC, CB, and CB

composites receive the most attention. Recently, a growing

interest in novel carbon supports such as CNTs,284–286 CNFs,287

and mesoporous carbons288,289 has arisen.

The active phase in fresh catalysts, as well as during the

experiments and at the end of experiments has been character-

ized using spectroscopic techniques, temperature programmed

adsorption/desorption methods, surface science techniques,

etc. with the aim to define the structure and involvement of

the active phase during hydroprocessing reactions. Attention

has been paid to factors that are causing the decline in catalyst

activity. Although to a lesser extent, nonconventional metals

such as Pt, Pd, Ru, Rh, Re, and Ir supported on carbon supports

were also studied as catalysts for hydroprocessing of model

feeds and real feed. These catalysts are much more active than

conventional metal-containing catalysts; however, their high

cost prevents their commercial utilization for HDS. However,

as classical HDS catalysts are not sufficiently active for deep or

ultradeep HDS processes, there is a need for new catalytic

formulations.290,291 In this sense, one promising candidate to

participate in these new formulations is rhenium sulfide, for

which activities an order of magnitude higher than Mo-based

catalysts have been reported.292–295

Novel catalytic phases such as metal carbides and

phosphides,280,296–299 mostly containing conventional active

metals exhibited a good activity and stability when combined

with carbon supports.

As far as the effects of the textural properties of the carbon

support on catalyst performances is concerned, the general

trend observed is the higher the surface area the higher the

activity.218,300–302

Concerning the PSD, the presence of mesopores favors the

HDS of bulky molecules,273 and delays the phenomenon of

pore blocking and metal sintering.295

The effect of surface oxygenated groups on catalytic perfor-

mances has not been studied in great detail. It was reported

that the HDS activity of sulfided Mo catalysts supported on

CBCs is affected by the kind of functional groups present on

the CBC surface.303 The oxidation of CBCs with (NH4)2S2O8

produces functionalities with the highest acid strength and the

corresponding catalyst exhibits the highest HDS activity. The

acidic strength of the support could play an important role in

determining the hydrogenolysis/hydrogenation ratio of the

catalysts.304 It has also been reported that acidic carbon sur-

faces permit higher adsorption capacity for sulfur compounds

in diesel fuel. The adsorption of the sulfur compounds over AC

from the liquid hydrocarbon fuel may involve an interaction of

the acidic oxygen-containing groups on AC with the sulfur

compounds.305

The HDS activity of Ni/AC catalysts was shown to be

enhanced by acid treatments of the carbon support.306 In this

case, introduction of oxygen-containing functional groups

leads to a strong interaction of O(s)–Ni during impregnation,

which becomes essential to achieving and preserving high

nickel dispersion. It has been mentioned that the interaction

between the metal precursors and the carbon surface is very

important in achieving highly dispersed and stable active

phases. In this sense, the choice of the precursors and the

impregnation conditions are of paramount importance, and

they should be selected in relation with the surface chemistry

of the carbon supports. Thus, the presence of oxygen surface

Page 39: Comprehensive Inorganic Chemistry II || Carbon

Carbon 361

groups makes the surface more hydrophilic and enhances its

wettability. However, if these groups have an acidic character,

the surface will be negatively charged over a wide range of pH

values of the impregnating solution. In this case, impregnation

with an anionic precursor such as ammonium heptamolybdate

has to be carried out at very low pH values, to render the

surface positively charged. Nevertheless, under these condi-

tions, polymerization of molybdenum species takes place and

a very low dispersion is obtained.307 These problems may

be avoided by using neutral precursors in less polar solvents.

This is the case, for example, of Farag et al.273 and Sakanishi

et al.,308 who used methanolic solutions of cobalt and molyb-

denum acetylacetonates.

The activity of the carbon-supported catalysts has been

determined using model compounds and real feeds. The

model compound studies consistently show higher HDS and

hydrodeoxygenation activities of carbon-supported catalysts

than that of g-Al2O3-supported catalysts. Further, the deactiva-

tion of the former catalysts by coke deposition was much less

evident. However, most of the model compound studies on

HDS and hydrodeoxygenation were conducted in the absence

of nitrogen compounds, although the poisoning effects of such

compounds on hydroprocessing reactions has been well

known.309 In fact, the information on HDN of model com-

pounds over carbon-supported catalysts is rather limited com-

pared with other hydroprocessing reactions.310 The advantages

of carbon-supported catalysts determined using model com-

pounds were less evident for real feeds of petroleum origin

except for hydrodemetallization.311 For feeds of biomass ori-

gin, carbon-supported catalysts exhibited much higher activity

and stability than the g-Al2O3-supported catalysts.312–314 It

should however be noted that only a limited number of studies

involved long-run testing of carbon-supported catalysts. With-

out the long-run performance determined, the potential of

carbon-supported catalysts for commercial applications cannot

be established. Indeed, the hydroprocessing catalysts discussed

operate often at relatively high temperatures. Under these

hydrogenating conditions, a major practical problem could

be gasification of the support itself to give methane or other

hydrocarbon species, and this should not be overlooked when

considering carbon-supported catalysts for hydrogenation

applications involving high hydrogen partial pressures and

high temperatures.

The kinetics of hydroprocessing reactions, as well as the

kinetics of deactivation over carbon-supported catalysts, have

been investigated under a wide range of experimental condi-

tions. In most studies, the g-Al2O3-supported catalysts have

been included for comparison. The difference between deter-

mined kinetic parameters could be interpreted in terms of

different mechanisms of hydroprocessing reactions. Thus, the

radical-like mechanism is more likely to be part of hydropro-

cessing reactions on carbon-supported catalysts than on the

g-Al2O3-supported catalysts.204

In spite of the good activity and the stability of the carbon-

supported catalysts observed during the laboratory and bench-

scale studies, there is little evidence supporting the use of these

catalysts in commercial hydroprocessing operations. In this

regard, additional information on long-term performance

using pilot plant reactors is needed. A significant difference

between the specific activity of a carbon support and that of a

g-Al2O3 support deserves attention when commercial applica-

tions are considered. Thus, on a weight basis, much more of

the carbon-supported catalysts than g-Al2O3-supported

catalysts may be required to achieve a similar performance.

Nevertheless, the evidence indicates the potential of carbon-

supported catalysts during hydroprocessing, particularly in

deep HDS and hydrodemetallization.

7.13.7.2.2 Selective hydrogenation reactionsof a-b-unsaturated aldehydesIn the fine chemical and pharmaceutical industry, a relative

large amount of carbon-supported catalysts are used, particu-

larly for hydrogenation reactions, which constitute the most

important industrial application of carbon-supported catalysts.

Recently, significant efforts have been devoted to the study of

the hydrogenation of carbon oxides (CO and CO2) to yield

methane and hydrocarbons (Fischer–Tropsch synthesis).315–319

For those reactions that involved hydrogen adsorption, acti-

vation, and often transfer by the support, carbon and graphite

materials are particularly interesting. Indeed, it has been shown

that carbon itself can activate hydrogen and facilitate its transfer

to reactant molecules.204,320,321 For selective hydrogenation, the

relative inertness of the carbon surface is of importance as the

catalytic systems are usually constituted bymore than onemetal-

lic component. The carbon inertness facilitates the interaction

between the metals and/or between the metals and the pro-

moters, yielding more active and selective catalysts than those

supported onother common supports. The role of carbon surface

chemistry and of carbon structure on the catalytic performances

will also be discussed. Carbon and graphite materials have

been used as supports in various reactions including the selective

hydrogenationof: alkynes as acetylene322 or hexyne323; alkenes as

styrene324; mixtures of hydrocarbons such as acetylene–

ethylene,325 3-methyl-1-butene hydrogenation of levulinic

acid326; nitrate327; D-glucose328; chloronitrobenzene329,330; and

2,4-dinitrotoluene.331 Hydrogenation of a-b-unsaturated alde-

hydes, and particularly cinnamaldehyde, has been particularly

studied. The selective hydrogenation of a,b-unsaturated alde-

hydes to unsaturated alcohols, which dates back to 1925,332 has

been of growing interest because unsaturated alcohols are impor-

tant intermediates for the production of fine chemicals and phar-

maceutical precursors. Cinnamyl alcohol (COL) is industrially

produced by reduction of cinnamaldehyde (CAL) either with

isopropyl or benzyl alcohol in the presence of the corresponding

aluminum alcoholate, or using Os/C catalysts, or with alkali

borohydrides. The chemoselective catalytic hydrogenation is a

difficult task, since thermodynamics favor the hydrogenation of

the C¼C bond over the C¼O bond by ca. 35 kJ mol�1 and

because for kinetic reasons, the reactivity of the C¼C bond is

higher than that of the C¼Obond. For supported catalysts, both

metal and support influence activity and selectivity of the final

catalyst toward the formation of unsaturated alcohols. It should

be pointed out that the final catalyst properties are also a result of

specific catalyst pretreatment and an active catalyst can be

obtained via fine-tuning the catalyst pretreatment procedures.333

The selectivity to unsaturated alcohols can in general be increased

via increasing the number of active sites activating the carbonyl

group. Furthermore, several metals or some metal alloys can be

used as active components and electronic properties can be tuned

by adding promoters.334 The characteristics of the support

Page 40: Comprehensive Inorganic Chemistry II || Carbon

362 Carbon

material, such as porosity, PSD, and reducibility are important

parameters in catalyst preparation. Various carbon and graphite

materials have been used for selective hydrogenation of a-b-unsaturated aldehydes, such as CNFs, CNTs, graphite, AC, CBs,

and fullerenes. In general, the catalytic systems obtained by depo-

sition on CNTs or CNFs are more active than their counterparts

onACor oxides.335–337Nanostructured carbon catalysts present a

dynamic mesoporous structure that should limit clogging and

enhance diffusion phenomena. Catalysts based on graphitic

(nano)materials are also more selective toward COL than AC.

Richard et al. have interpreted this in terms of an electronic ligand

effect.338,339 Since metal particles are preferentially located on

steps and edges of graphitic planes, the p-electrons of the gra-

phitic planes can be easily extended to the metal particles, thus

increasing the charge density of the metal. In fact, the graphitic

support acts like amacro-electron-donating ligand. The increased

charge density on the nanoparticles decreases the probability of

adsorption via the C¼Cbond, so that the selectivity towardCOL

increases. Palladium340 and nickel341 systems are very active in

the selective hydrogenation of the C¼C bonds, while Os,338

Ir,342 Pt,343 and bimetallic Pt-based systems256,344,345 permit

hydrogenating the C¼O bond.

The presence or not of surface oxygenated groups also

directly affects the catalytic behavior of the active phase. In a

study on vapor-phase crotonaldehyde (2-butenal) hydrogena-

tion on Pt/C catalysts, Coloma et al. used a demineralized AC

oxidized with H2O2 to introduce oxygen surface functionalities

(ACOx) and then heat treated under helium at 773 K to remove

the less stable surface groups (ACOxT). The specific activity

of 1%Pt/AC catalysts was high, and followed the order

Pt/AC<Pt/ACOxT<Pt/ACOx. The selectivity to the unsaturated

alcohol was low for catalyst Pt/AC and much larger for the

others, especially for catalyst Pt/ACOx. The catalytic behavior

was also determined after reduction at different temperatures

(623–773 K). An increase in catalytic activity was noticed when

increasing the reduction temperature, which was much more

important for catalysts Pt/ACOxT and Pt/ACOx. With regard to

selectivity, the increase in the reduction temperature did not

affect the selectivity of the catalyst Pt/AC, but it enhanced the

selectivity of the other two catalysts. The beneficial effect

of using preoxidized carbon as support was related to the

decomposition–reduction of oxygen surface groups upon the

heat treatment in hydrogen to which the catalysts were sub-

jected before the catalytic measurements.346 Toebes et al. used

Ru/CNF catalysts to study the influence of oxygen surface

groups on catalytic performance for the hydrogenation of

CAL.347 The CNFs were oxidized to introduce the oxygen func-

tionalities. After reduction, the catalysts were heat treated in

nitrogen at different temperatures to control the number of

oxygen surface groups. They observed a narrow and stable

particle size distribution (1–2 nm) even after the heat treat-

ment at 973 K. The overall specific activity increased by a factor

of 22 after the treatment at this high temperature, which was

related to the decreased number of oxygen surface groups. In

this case, the selectivity to COL decreased from 48% to 8%. In a

similar study carried out with Pt/CNF catalysts,348,349 the

authors also found an increase in catalytic activity with increas-

ing thermal treatment temperature, and a linear decrease in the

hydrogenation activity with an increase of the number of acidic

groups on the CNF surface. They suggested that the

hydrogenation process was favored by the adsorption of CAL

on the carbon support after removal of the oxygen-containing

surface complexes. Similar results have been obtained for Pt/

CNT.246 For bimetallic PtRu/CNT systems reduced at 623 K,

the removal of oxygen surface groups by heat treatment under

an inert atmosphere (up to 1273 K) induces an increase of

both activity and selectivity toward COL.256 In a first stage,

heat treatment under argon at 973 K does not affect the metal

particle size (2.7–2.9 nm). This treatment allows an increase of

both activity and selectivity. The activity increase was attrib-

uted: (1) to a better adsorption of CAL, the CNT surface thus

acting as a reservoir of substrate where a high local concentra-

tion of CAL is readily accessible for the catalyst and (2) to a

better diffusion of CAL and COL on the support. The selectivity

increase was attributed to an enrichment of the surface of the

particles in ruthenium after the heat treatment, which act as an

electropositive metal that increases the electron density on the

Pt, and as electrophilic or Lewis sites for the adsorption and

activation of the C¼O bond. The possible favored adsorption

of COL on the heat-treated surface of the catalyst during the

reaction, which creates a steric effect that inhibits the hydroge-

nation of the C¼C bond of CAL, has also been proposed.342 At

higher temperatures (1123 or 1273 K), the PtRu particle size

increases up to 13 nm, and a slight increase in selectivity is

noticed due to this well-known particle size effect.334 The

positive or negative impact on catalytic performances of sur-

face oxygen group removal after supported catalyst preparation

has also been investigated for other reactions such as oxidation

of benzyl alcohol over Ru/CNFs,350 hydrodechlorination of

chlorobenzenes over Pd/AC, Pd/graphite, and Pd/ CNFs,351

and sorbitol hydrogenolysis to glycols on Ru/CNF.352 It is

also worth mentioning that the effect of the gradual thermal

decomposition of surface oxygen groups on the chemical and

catalytic properties of oxidized AC has also been reported.353

7.13.7.2.3 PhotocatalysisPhotocatalysis applications of wide-reaching importance

include water splitting for hydrogen generation, degradation

of environmental pollutants in aqueous contamination and

wastewater treatment, carbon dioxide remediation, self-

cleaning activity, and air purification.354 An ideal photocataly-

tic material would combine high activity regarding the relevant

process of interest with high efficiency of solar energy conver-

sion. By far the most researched photocatalytic material is

titanium dioxide, because it has provided the most efficient

photocatalytic activity, highest stability, lowest cost, and lowest

toxicity.355 Despite these many attempts at enhancement, effi-

cient, and commercially viable photocatalysts for strategic pro-

cesses such as water splitting and degradation of various

pollutants are yet to be realized. The band model (Figure 28)

is generally used to describe photocatalysis, and the main

involved processes are: (i) photon absorption and electron–

hole pair generation; (ii) charge separation and migration;

(iia) to surface reaction sites or (iib) to recombination sites;

and (iii) surface chemical reaction at active sites. At least two

reactions should occur simultaneously: the oxidation from

photogenerated holes and the reduction from photogenerated

electrons. The first key process (1) is the absorption of photons

to create electron–hole pairs. The energy of the incident light

must be greater than the bandgap of the TiO2 (3.0 and 3.2 eV

Page 41: Comprehensive Inorganic Chemistry II || Carbon

Energy level

Eg

h+

e−

(i)

TiO2Acceptor

Acceptor.−

Reduction (iii)

(iia)

(iib)

Donor

Donor.+

Oxidation (iii)

(iia)

(iib)

Figure 28 Main processes in semiconductor photocatalysis: (i) photonabsorption and electron–hole pair generation; (ii) charge separation andmigration; (iia) to surface reaction sites or (iib) to recombination sites;and (iii) surface chemical reaction at active sites. Reproduced fromLeary, R.; Westwood, A. Carbon 2011, 49, 741–772, with permissionfrom Elsevier.

Carbon 363

for rutile and anatase, respectively). A primary reason for the

current, low solar photoconversion efficiency of TiO2 is that it

can only be excited under irradiation of UV light at wavelengths

<380 nm, which covers only �5% of the solar spectrum. An

alternative to modifying the TiO2 bandgap is to employ a nar-

row bandgap photosensitizer, which is excited by lower energy

visible wavelengths, and is capable of transferring excited elec-

trons or holes to the TiO2. For the second key process (2), it has

been shown that electron–hole trapping or recombination rates

are extremely fast so that they can substantially lower the photo-

catalytic activity. It is therefore of interest to develop nanocata-

lysts, for which the distance that the photogenerated electrons

and holes need to travel to surface reaction sites is reduced,

thereby reducing the recombination probability. Concerning

the final step, the surface chemical reactions, both surface chem-

istry (active sites) and surface area are important. Increased

surface area may be obtained by using highly porous materials,

and/or reducing their size.

As a first approach to facilitate the use of TiO2 as photo-

catalyst, immobilization on a suitable support is desired.

Therefore, a great deal of effort has been taken to increase the

photoefficiency by dispersing TiO2 over high surface areamate-

rials. Due to their unique pore structure, adsorption capacity,

acidity, and electronic properties, carbon-based materials,

including carbon nanomaterials of different origins, can be

used for this purpose.356 Of potential significance is a recent

study that suggests that AC itself is capable of a significant level

of photocatalytic activity.357 Most of the studies combining

TiO2 and carbon or graphite materials have evidenced that by

introducing materials derived from carbon in a TiO2 matrix,

beneficial effects in the photocatalytic activity of TiO2 can be

obtained, in addition to the expected increased dispersion.

In the early works, the observed synergistic effect was

ascribed to the preferential adsorption of the organic molecule

onto AC, followed by surface transfer to TiO2.358,359 The

adsorption equilibrium constant of aromatics, such as phenol,

on AC is more than one order of magnitude greater than in bare

TiO2. After accumulation of the organic over the carbon surface,

the driving force for surface diffusion is the gradient of surface

concentration between the two materials. In order to achieve

effective transfer, there must be a close contact interface between

the two solid phases. In such a way, the organic is transferred

from the AC to the TiO2, and then undergoes immediate photo-

catalytic degradation, which originates the synergistic effect.

Such an effect has been described for several studies dealing

with AC–TiO2 composites.360–363 On the negative side, while

AC can enhance the formation of smaller nanosized TiO2

particles364 and contains nanosized pores, analysis suggests

that the smaller pores are rarely infiltrated, with the TiO2

remaining on the outer macropores.365 Furthermore, the need

for bandgap tuning of TiO2 remains untackled by using AC

which generally does not chemically interact with TiO2. Indeed,

the formation of Ti–O–C bonds has been rarely reported.366

Of course, when irradiation takes place at wavelengths

shorter than the absorption edge of the TiO2, it is impossible

to quantify the effect on exciting the carbon phase. This is

because below that threshold both phases compete for photon

absorption. However, when the light used is in the range

l>400 nm, photodegradation can only arise from a photosen-

sitized process.375 In this case, two major pathways can be

considered: (1) the carbon phase acts as a photosensitizer with-

out the contribution of TiO2; (2) the carbon phase is excited and

transfers an electron to the conduction band of TiO2, triggering

conventional semiconductor photocatalysis.367,368 Literature

values of measured and calculated TiO2/dopant state bandgap

energies induced by carbon doping show that carbon doping

induced shifts in absorption threshold varying from values as

little as�0.1 to as much as�1.05 eV.356 Of special interests are

the recent attempts to modify TiO2 with nanostructured carbo-

naceous materials such as CNT and graphene.356 The enhanced

photoactivity reported in these studies is suggested to originate

from the creation of an efficient route for the electron transport,

thereby reducing the recombination rate.369,370

Since CNTs show the potential to contribute to all three of

the routes of increasing photocatalytic activity (i.e., high sur-

face area and high quality active sites, retardation of electron–

hole recombination, and visible light catalysis by modification

of band-gap and/or sensitization), their use as cocatalyst has

been particularly studied. When functionalized CNTs are used

as a carbon phase in the composite catalysts, photosensititiza-

tion is expected to contribute to enhance the photoactivity

more efficiently than with AC. The photodegradation of phe-

nol in aqueous suspension in the presence of a multiwalled

CNT–TiO2 composite catalyst under visible light was found to

follow a pseudo-first-order kinetics and a synergy factor of

1.8 was measured.371 In this case, the synergistic effect of

CNT on the activity of the composite catalysts was explained

in terms of its action mainly as a photosensitizer, and partially

as an adsorbent. An additional role as dispersing agent pre-

venting TiO2 from agglomeration was also accounted for. Since

the gradual increase in the amount of CNT in the composite

catalysts does not provide significant increase in their adsorp-

tion capacities (measured in terms of the adsorption equilib-

rium constant), the synergistic effect cannot be merely due to

CNT acting as an adsorbent.

Page 42: Comprehensive Inorganic Chemistry II || Carbon

364 Carbon

Considering the semiconductive properties of CNT upon

light absorption, transfer of the photo-induced electron (e�)into the conduction band of the TiO2 particles occurs

(Figure 29). This electron transfer between the carbon phase

and the metal oxide phase was experimentally observed by mea-

suring the enhanced photocurrent in some other systems.372–375

Thus, the analysis of transient photocurrent responses reflects

CNTs as fast electron transfer channels in chemically bonded

CNT–TiO2 composites with low CNT loading.376 Simulta-

neously, a positive charged hole (hþ) might be formed by elec-

tron migrating from the TiO2 valence band to CNT.

Thus, the role played by CNTs is to provide electrons to the

TiO2 conduction band under visible light irradiation and to

trigger the formation of very reactive radicals such as the super-

oxide radical ion O2•¯. Back electron transfer from adsorbed

OH� closes the electrocatalytic cycle and provides a source for

the hydroxyl radical HO•, which is responsible for the degra-

dation of the organic molecules (Figure 29(a) and 29(b)).

Besides this electron transfer processes, it should be stressed

that the addition of CNTs in the TiO2 matrix leads to a broader

absorption toward the visible part of the spectrum.371 If the

absorption of the low energy photons is efficiently converted

in charge separation at the semiconductor phase, this will lead

to an increase in the relative photonic efficiency of heteroge-

neous photocatalysis.

Different CNT loadings have been investigated. In most

cases of TiO2 nanoparticles loaded onto or randomly mixed

VB

CB

e−

CNT

h+ HO.(c)

e−

X

VB

CB

e−

CNT

h+

OH−+ H+ H2O

HO + H+

(b)

VB

CBe−

TiO2

TiO2

TiO2

CNT

e− O2

O2.−

(a)

.

Figure 29 (a) CNT acting as photosensitizer in the composite catalystin the process of electron injection into the conduction band of TiO2

semiconductor; (b) electron back transfer to CNT following electrontrapping by a hole in the valence band of TiO2; and (c) proposedmechanism for the enhanced electron transfer in CNT/TiO2 composites.

with CNTs, photocatalytic activity increased up to �85 wt%

CNT, after which it decreased.377,378 However, the optimum

percentage of CNT appears to be highly dependent on the

morphology of the photocatalyst: for mixtures/composites of

CNTs loaded onto larger TiO2 nanoparticles/nanotubes opti-

mal activity has been found at around 20 wt% CNT.379,380

Conversely, Yen et al. also found �20 wt% CNT to be opti-

mum for TiO2 nanoparticles loaded on CNTs.381 In either case,

there exists a compromise between increased synergistic effect

from higher CNT loadings, and insufficient amounts of TiO2

and/or blockage of TiO2 active sites. The apparently contradic-

tory findings could be partially explained based on whether the

CNT is acting as an electron sink (Figure 28(a)) or as a pho-

tosensitizer (Figure 28(c)). If the former mechanism is active,

since the TiO2 is the photoactive phase, it may be beneficial to

have higher percentages of TiO2 and to promote exposed TiO2

surface area. If the latter mechanism is active, the CNT is the

photoactive phase, and optimal activity may be achieved at

higher CNT loadings and promotion of exposed CNT surface

area. However, since the photocatalytic activity also relates to

other factors such as the nature of the interphase contact,

variations of relative exposed surface areas in different mor-

phologies, and the variable importance of the CNT for adsorp-

tion of reactants (according to the different applications and

reactants), general guidelines for optimum mass fractions or

relative exposed surface areas are still difficult to come by based

on the current literature.

The reaction mechanism of photocatalytic degradation is

believed not to be changed by the introduction of the carbon

phase,358,359 thus proceeding by the electron transfer from

the water solvent molecules (H2O and OH�) to the positive

charged holes (hþ) forming the very oxidizing nonselective

HO• radical on the surface of the catalyst. Molecular oxygen,

which must be present, is adsorbed onto the surface of the

catalyst and acts as an electron acceptor, forming the superox-

ide radical anion O2•¯. In the presence of an organic molecule,

the generated HO• radical reacts by adduct formation, which

then breaks down into several intermediates till, eventually,

total mineralization. If the organic reactant is able to compete

with adsorbed oxygen and water molecules for the active sites

of the photocatalyst, then direct electron transfer to an active

hole is also conceivable, as the first step of an oxidative degra-

dation producing adsorbed organic radicals. This overall

process is commonly accepted367,368 and can be described by

a set of sequential and concurrent multielectron transfers.

7.13.8 Conclusion

Carbon is an extraordinary element that exists in a remarkable

variety of forms. Graphite, coal, soot, and diamonds are all

nearly pure, naturally occurring, forms of carbon. Carbon and

graphite materials can also be produced in various forms such

as ACs, CBs, CNTs, fullerenes, or graphene. A large variety of

carbon materials is industrially produced. The richness and the

complexity of carbon physical chemistry make the design of

structurally controlled materials a challenging task. For cataly-

sis, the description of carbon surfaces (chemistry and pore

description) is a key point toward a better control of the

catalytic performances. For most catalytic applications, the

Page 43: Comprehensive Inorganic Chemistry II || Carbon

Carbon 365

surface chemistry of the carbons has a major influence on the

performance of the material, for example, dispersions in liq-

uids, the dispersion of metallic compounds on carbon sur-

faces, the adsorption of reactants/products onto carbons, etc.

The catalytic reactions are often carried out in meso- and

microporous carbon or graphite materials, which can enhance

or reduce reaction yields through a host of different effects,

including an increase in the surface area per unit volume, the

selective adsorption of reactants and/or products, and confine-

ment and shape-catalytic effects, among others. A fundamental

understanding of the role of each one of these different effects

should lead to a systematic procedure for the design of

improved catalytic materials that take advantage of all of

them simultaneously.

References

1. Lu, P. J.; Nan, Y.; So, J. F.; Harlow, G. E.; Lu, J.; Wang, G.; Chaikin, P. M.Archaeometry 2005, 47, 1–12.

2. Fitzer, E.; Kochling, K. H.; Boehm, H. P.; Marsh, H. Pure Appl. Chem. 1995,67, 473–506.

3. Serp, P.; Figueiredo, J. L. Carbon Materials for Catalysis. J. Wiley & Sons:Hoboken, NJ, 2009.

4. Kroto, H. RSC Chemistry World November 2010.5. Demming, A. Nanotechnology 2010, 21, 300201.6. Kroto, H. W.; Heath, J. R.; O’Brien, S. C.; Curl, R. F.; Smalley, R. E. Nature 1985,

318, 162–163.7. Iijima, S. Nature 1991, 354, 56–58.8. Inagaki, M.; Feiyu, K. Carbon Materials Science and Engineering: From

fundamental to applications. Tsinghua University Press: Beijing, 2006.9. Dumitrica, T.; Landis, C. M.; Yakobson, B. I. Chem. Phys. Lett. 2002, 360,

182–188.10. Rao, A. M.; Dresselhaus, M. S. In Nanostructured Carbon for Advanced

ApplicationsNATO Science Series II; ; Benedek, G., Milani, P., Ralchenko, V. G.,Eds.; 24; Kluwer: Dordrecht, 2001; pp 3–26.

11. Schlogl, R. In Preparation of Solid Catalysts; Ertl, G., Knozinger, H., Weitkamp, J.,Eds.; Wiley-VCH: Weinheim, 1999; pp 150–240.

12. Bunsell, A. R. Fibre Reinforcements for Composite Materials. Elsevier SciencePublishers B.V.,: Amsterdam, The Netherlands, 1988; p.120.

13. Mattia, D.; Rossi, M. P.; Kim, B. M.; Korneva, G.; Bau, H. H.; Gogotsi, Y. J. Phys.Chem. B 2006, 110, 9850–9855.

14. Mochida, I.; Yoon, S. H.; Qiao, W. J. Braz. Chem. Soc. 2006, 17, 1059–1073.15. Glasier, G. F.; Pacey, P. D. Carbon 2001, 39, 15–23.16. Farrauto, R. J.; Bartholomew, C. H. Fundamentals of Industrial Catalytic

Processes. Blackie Academic and Professional: London, 2003.17. Froment, G. F. Rev. Chem. Eng. 1990, 6, 293–328.18. Kumar, M.; Ando, Y. J. Nanosci. Nanotechnol. 2010, 10, 3739–3758.19. Marsh, H.; Menendez, R. Fuel Process. Technol. 1988, 20, 269–296.20. Su, Z.; Zhou, W.; Zhang, Y. Chem. Commun. 2011, 47, 4700–4702.21. Kinoshita, K. Carbon: Electrochemical and physicochemical properties.

John Wiley and Sons: New York, 1988; pp. 31–34.22. Ungar, T.; Gubicza, J.; Tichy, G.; Pantea, C.; Zerda, T. W. Compos. Part A 2005,

36, 431–436.23. Harris, P. J. F. Int. Mater. Rev. 1997, 42, 206–218.24. Pierson, H. O. Handbook of carbon, graphite and fullerenes: Properties,

processing and applications. Noyes Publications: Park Ridge, NJ, 1993;pp. 244–276.

25. Haddon, R. C. Acc. Chem. Res. 1992, 25, 127–133.26. Harris, P. J. F. Carbon Nanotubes and Related Structures, New materials for the

twenty first century. Cambridge university press: Cambridge, 1999; pp. 235–262.27. Yoshizawa, N.; Tanaike, O.; Hatori, H.; Yoshikawa, K.; Kondo, A.; Abe, T. Carbon

2006, 44, 2558–2564.28. Franklin, R. E. Proc. R. Soc. Lond. A 1951, A209, 196–218.29. Crawford, D.; Johnson, D. J. J. Microsc. 1971, 94, 51–62.30. Jenkins, G. M.; Kawamura, K.; Ban, L. L. Proc. R. Soc. Lond. A 1972, A327,

501–517.31. Yoshida, A.; Kaburagi, Y.; Hishiyama, Y. Carbon 1991, 29, 1107–1111.32. Harris, P. J. F.; Tsang, S. C. Philos. Mag. A 1997, 76, 667–677.

33. Ban, L. L.; Crawford, D.; Marsh, H. J. Appl. Crystallogr. 1975, 8, 415–420.34. Stoeckli, H. F. Carbon 1990, 28, 1–6.35. Byrne, J. F.; Marsh, H. In Porosity in carbons: Characterisation and applications;

Patrick, J. W., Ed.; Arnold: London, 1995; pp 1–48.36. Mackay, A. L.; Terrones, H. Nature 1991, 352, 762.37. Lenosky, T.; Gonze, X.; Teter, N.; Elser, V. Nature 1992, 355, 333–335.38. Terrones, H.; Mackay, A. L. Carbon 1992, 30, 1251–1260.39. Terrones, H.; Mackay, A. L. Chem. Phys. Lett. 1993, 207, 45–50.40. Valencia, F.; Romero, A. H.; Hernandez, E.; Terrones, M.; Terrones, H. New J.

Phys. 2003, 5, 123.1–123.16.41. Bourgeois, L. N.; Bursill, L. A. Philos. Mag. A 1997, 76, 753–768.42. Smith, M. A.; Foley, H. C.; Lobo, R. F. Carbon 2004, 42, 2041–2048.43. Acharya, M.; Strano, M. S.; Mathews, J. P.; Billinge, S. J. L.; Petkov, V.;

Subramoney, S.; Foley, H. C. Philos. Mag. B 1999, 79, 1499–1518.44. Chuvilin, A.; Kaiser, U.; Bichoutskaia, E.; Besley, N. A.; Khlobystov, A. N. Nat.

Chem. 2010, 2, 450–453.45. Lozovik, Y. E.; Popov, A. M. Uspekhi Fizicheskikh Nauk 1997, 167, 751–774.46. Zdravkov, B. D.; Cermak, J. J.; Sefara, M.; Janku, J. CEJC 2007, 5, 385–395.47. Gross, A. F.; Nowak, A. P. Langmuir 2010, 26, 11378–11383.48. Pol, V. G.; Motiei, M.; Gedanken, A.; Calderon-Moreno, J.; Yoshimura, M. Carbon

2004, 42, 111–116.49. Inagaki, M. Solid State Ion. 1996, 86–88, 833–838.50. Inagaki, M. Carbon 1997, 35, 711–713.51. McCreery, R. L. Chem. Rev. 2008, 108, 2646–2687.52. Maillard, F.; Simonov, P. A.; Savinova, E. R. Carbon materials as supports for fuel

cells electrocatalysts. In Carbon Materials for Catalysis; Serp, P., Figueiredo, J. L.,Eds.; J. Wiley & Sons: Hoboken, NJ, 2009; pp 429–480.

53. Chin, Y. H.; Hu, J.; Cao, C.; Gao, Y.; Wang, Y. Catal. Today 2005, 110,47–52.

54. Radovic, L. Physico-chemical properties of carbon materials: a brief overview.In Carbon Materials for Catalysis; Serp, P., Figueiredo, J. L., Eds.; J. Wiley &Sons: Hoboken, NJ, 2009; pp 1–44.

55. Dresselhaus, M. S.; Dresselhaus, G. Adv. Phys. 2002, 51, 1–186.56. Lee, J.; Kim, J.; Hyeon, T. Adv. Mater. 2006, 18, 2073–2094.57. Furmaniak, S.; Terzyk, A. P.; Gauden, P. A.; Rychlicki, G. J. Colloid Interface Sci.

2006, 295, 310–317.58. Furmaniak, S.; Terzyk, A. P.; Gauden, P. A.; Harris, P. J. F.; Wiesniewski, M.;

Kowalczyk, P. Adsorption 2010, 16, 197–213.59. Sun, C. H.; Li, F.; Cheng, H. M.; Lu, G. Q. Appl. Phys. Lett. 2005, 87, 243109.60. Yang, Q. H.; Hou, P. X.; Bai, S.; Wang, M. Z.; Cheng, H. M. Chem. Phys. Lett.

2001, 345, 18–24.61. Chiang, Y. C.; Lee, C. Y. J. Mater. Sci. 2009, 44, 2780–2791.62. Eltekova, N. A.; Razd’yakonova, G. I.; Eltekov, Y. A. Pure Appl. Chem. 1993, 65,

2217–2221.63. Pikunic, J. C.; Clinard, N.; Cohaut, K. E.; Gubbins, K. E.; Guet, J. M.;

Pellenq, R. J. M.; Rannou, I.; Rouzaud, J. N. Langmuir 2003, 19, 8565–8582.64. Jain, J. S. K.; Pellenq, R. J. M.; Pikunic, J. P.; Gubbins, K. E. Langmuir 2006, 22,

9942–9948.65. Terzyk, A. P.; Furmaniak, S.; Gauden, P. A.; Harris, P. J. F.; Włoch, J.;

Kowalczyk, P. J. Phys. Condens. Matter 2007, 19, 406208.66. Terzyk, A. P.; Furmaniak, S.; Harris, P. J. F.; Gauden, P. A.; Włoch, J.;

Kowalczyk, P.; Rychlicki, G. Phys. Chem. Chem. Phys. 2007, 9, 5919–5927.67. Gavals, G. R. AIChE J. 1980, 26, 575–577.68. Srinivasalu Gupta, J.; Bhatia, S. K. Carbon 2000, 38, 47–58.69. Bhatia, S. K.; Vartak, B. J. Carbon 1996, 34, 1383–1391.70. Simons, G. A. Symp. Int. Combust. 1982, 19, 1067–1076.71. Radovic, L. R.; Bockrath, B. J. Am. Chem. Soc. 2005, 127, 5917–5927.72. Hirsch, A.; Vostrowski, O. Top. Curr. Chem. 2005, 245, 193–237.73. Tasis, D.; Tagmatarchis, N.; Bianco, A.; Prato, M. Chem. Rev. 2006, 106,

1105–1136.74. Stein, A.; Wang, Z.; Fierke, M. A. Adv. Mater. 2009, 21, 265–293.75. Salame, I. I.; Bandosz, T. J. J. Colloid. Interf. Sci. 2001, 240, 252–258.76. Menendez, J. A.; Phillips, J.; Xia, B.; Radovic, L. R. Langmuir 1996, 12,

4404–4410.77. Fuente, E.; Menendez, J. A.; Suarez, D.; Montes-Moran, M. A. Langmuir 2003,

19, 3505–3511.78. Stohr, A.; Boehm, H. P.; Schlogl, R. Carbon 1991, 29, 707–720.79. Boehm, P. Catalytic properties of nitrogen-containing carbons. In Carbon

Materials for Catalysis; Serp, P., Figueiredo, J. L., Eds.; J. Wiley & Sons:Hoboken, NJ, 2009; pp 45–92.

80. Bandosz, T. Surface chemistry of carbon materials. In Carbon Materials forCatalysis; Serp, P., Figueiredo, J. L., Eds.; J. Wiley & Sons: Hoboken, NJ, 2009;pp 45–92.

Page 44: Comprehensive Inorganic Chemistry II || Carbon

366 Carbon

81. Burg, P.; Cagniant, D. In Chemistry & Physics of Carbon; Radovic, L. R., Ed.; 30,CRC Press: UK, 2008 pp 129–175.

82. Domingo-Garcıa, M.; Lopez Garzon, F. J.; Perez-Mendoza, M. J. J. Colloid Interf.Sci. 2002, 248, 116–122.

83. Boehm, H. P. Carbon 1994, 32, 759–769.84. Moreno-Castilla, C. Carbon 2004, 42, 83–92.85. Goertzen, S. L.; Theriault, K. D.; Oickle, A. M.; Tarasuk, A. C.; Andreas, H. A.

Carbon 2010, 48, 1252–1261.86. Babic, B. M.; Milojic, S. M.; Polovina, M. J.; Kaludjerovic, B. V. Carbon 1999, 37,

477–481.87. Stelzer, J. B.; Nitzsche, R.; Caro, J. Chem. Eng. Technol. 2005, 28, 182–186.88. Ratner, B. D.; Castner, D. G. In Surface Analysis: The Principal Techniques, 2nd

ed.; Vickerman, J. C., Gilmore, I., Vickerman, J. C., Gilmore, I., Eds.; John Wiley &Sons: Chichester, 2009; pp 47–112.

89. Kohl, S.; Drochner, A.; Vogel, H. Catal. Today 2010, 150, 67–70.90. Fanning, P. E.; Vannice, M. A. Carbon 1993, 31, 721–730.91. Sellitti, C.; Koening, J. L.; Ishida, H. Carbon 1990, 28, 221–228.92. Socrates, G. Infrared Characteristic Group Frequencies, 2nd ed.; Wiley & Sons:

Chichester, England, 1994.93. Fuente, E.; Menendez, J. A.; Dıez, M. A.; Suarez, D.; Montes-Moran, M. A. J. Phys.

Chem. B 2003, 107, 6350–6359.94. Pereira, M. F. R.; Soares, S. F.; Orfao, J. J. M.; Figueiredo, J. L. Carbon 2003,

41, 811–821.95. Auer, E.; Freund, A.; Pietsch, J.; Tacke, T. Appl. Catal. A 1998, 173, 259–271.96. Arunajatesan, V.; Chen, B.; Mobus, K.; Ostgard, D. J.; Tacke, T.; Wolf, D. Carbon

Supported Catalysts for the Chemical Industry. In Carbon Materials for Catalysis;Serp, P., Figueiredo, J. L., Eds.; J. Wiley & Sons: Hoboken, NJ, 2009;pp 535–572.

97. Rodriguez-Reinoso, F. Carbon 1998, 36, 159–175.98. Marsh, H.; Rodrıguez-Reinoso, F. Activated carbons. Elsevier: Oxford, 2006.99. Harris, P. J. F.; Liu, Z.; Suenaga, K. J. Phys. Condens. Matter 2008, 20, 362201.100. Kang, M.; Bae, Y. S.; Lee, C. H. Carbon 2005, 43, 1512–1516.101. Nowicki, H.; Sherman, B. Water Cond. Purif. 2006, 48, 28–35.102. Carmo, M.; Dos Santos, A. R.; Poco, J. G. R.; Linardi, M. J. Power. Sources

2007, 173, 860–866.103. Chung, D. D. L. J. Mater. Sci. 1987, 22, 4190–4198.104. Li, W.; Han, C.; Liu, W.; Zhang, M.; Tao, K. Catal. Today 2007, 125, 278–281.105. Wang, F. L.; Hwang, C. Appl. Catal. 2004, 276, 9–16.106. Matatov-Meytal, Y.; Sheintuch, M. Appl. Catal. A 2002, 231, 1–16.107. Carrott, P. J. M.; Nabais, J. M. V.; Ribeiro Carrott, M. M. L.; Pajares, J. A. Carbon

2001, 39, 1543–1555.108. Landau, M. V.; Kogan, S. B.; Tavor, D.; Herskowitz, M.; Koresh, J. E. Catal. Today

1997, 36, 497–510.109. Jin, H.; Park, S. E.; Lee, J. M.; Ryu, S. K. Carbon 1996, 34, 429–431.110. Bartholomew, C. H. Catal. Rev. Sci. Eng. 1982, 16, 67–112.111. Rodriguez, N. M. J. Mater. Sci. 1993, 8, 3233–3250.112. De Jong, K. P.; Geus, J. W. Catal. Rev. Sci. Eng. 2000, 42, 481–510.113. Serp, P.; Corrias, M.; Kalck, P. Appl. Catal. A: Gen. 2003, 253, 337–358.114. Serp, P.; Castillejos, E. ChemCatChem 2010, 2, 41–47.115. Serp, P. Carbon nanotubes and nanofibers in catalysis. In Carbon Materials for

Catalysis; Serp, P., Figueiredo, J. L., Eds.; J. Wiley & Sons: Hoboken, NJ, 2009;pp 309–372.

116. Castillejos, E.; Serp, P. Carbon Nanotubes for Catalytic Applications. In CarbonNanotubes and Related Structures: Synthesis, Characterization, Functionalization,and Applications,; Martin, N., Guldi, D. M., Eds.; Wiley-VCH: Weinheim, 2010;pp 321–348.

117. Coq, B.; Planeix, J. M.; Brotons, V. Appl. Catal. 1998, 173, 175–183.118. Zhang, J.; Su, D. S.; Blume, R.; Schlogl, R.; Wang, R.; Yang, X.; Gajovic, A.

Angew. Chem. Int. Ed Engl. 2010, 49, 8640–8644.119. Vershinin, N. N.; Efimov, O. N.; Bakaev, V. A.; Aleksenskii, A. E.; Baidakova, M. V.;

Sitnikova, A. A.; Ya Vul’, A. Fuller. Nanotub. Carbon Nanoclust. 2011, 19, 63–68.120. Zhu, S.; Xu, G. Nanoscale 2010, 2, 2538–2549.121. Pekala, R. W.; Fricke, J. In Encyclopedia of Materials: Science and Technology;

Buschow, K. H. J., Cahn, R., Flemings, M., Ilschner, B., Kramer, E., Mahajan, S.,Veyssiere, P., Eds.; Pergamon: Oxford, 2001; pp 6502–6506.

122. Moreno-Castilla, C.; Maldonado-Hodar, F. J. Carbon 2005, 43, 455–465.123. Moreno-Castilla, C. Carbon gels in catalysis. In Carbon Materials for Catalysis;

Serp, P., Figueiredo, J. L., Eds.; J. Wiley & Sons: Hoboken, NJ, 2009;pp 373–399.

124. Jovanovic, V. M.; Terzic, S.; Tripkovic, A. V.; Popovic, K. D.; Lovic, J. D.Electrochem. Comm. 2004, 6, 1254–1258.

125. Stevanovic, S.; Panic, V.; Tripkovic, D.; Jovanovic, V. M. Electrochem. Comm.2009, 11, 18–21.

126. Braun, A.; Bartsch, M.; Schnyder, B.; Kotz, R.; Haas, O.; Wokaun, A. Carbon2002, 40, 375–382.

127. Schmitt, J. L., Jr.; Walker, P. L., Jr. Carbon 1971, 9, 791–796.128. Schmitt, J. L., Jr.; Walker, P. L., Jr. Carbon 1972, 10, 87–92.129. Schmiit, J. L. Carbon 1991, 29, 743–745.130. Grunewald, G. C.; Drago, R. S. J. Am. Chem. Soc. 1991, 113, 1636–1639.131. Radovic, L. R.; Rodrıguez-Reinoso, F. In Chemistry and Physics of Carbon;

Thrower, P. A., Ed.; 25, Marcel Dekker: New York, 1997 pp 243–358.132. Figueiredo, J. L.; Pereira, M. F. R. Carbon as Catalyst. In Carbon Materials for

Catalysis; Serp, P., Figueiredo, J. L., Eds.; J. Wiley & Sons: Hoboken, NJ, 2009;pp 177–217.

133. Schneider, W.; Diller, W. In Ullmann’s Encyclopedia of Industrial Chemistry.Wiley-VCH Verlag GmbH & Co: KGaA, 2002; Online Edition.

134. Lauss, H. D. In Ullmann’s Encyclopedia of Industrial Chemistry. Wiley-VCHVerlag GmbH & Co: KGaA, 2002; Online Edition.

135. Rase, H. F. Handbook of Commercial Catalysts Heterogeneous Catalysts. CRCPress: Boca Raton, 2000.

136. Ryan, T. A.; Stacey, M. H. Fuel 1984, 63, 1101–1106.137. Juntgen, H.; Kuhl, H. In Chem. Phys. Carbon; Thrower, P. A., Ed.; Marcel Dekker:

New York, 1989; pp 145–195.138. Tsuji, K.; Shiraishi, I. Fuel 1997, 76, 555–560.139. Li, J.; Kobayashi, N.; Hu, Y. J. Environ. Eng. 2007, 2, 740–751.140. Petrosius, S. C.; Drago, R. S. J. Chem. Soc. Chem. Commun. 1992, 4, 344–345.141. Farcasiu, M.; Petrosius, S. C.; Ladner, E. P. J. Catal. 1994, 146, 313–316.142. Stuber, F.; Font, J.; Fortuny, A.; Bengoa, C.; Eftaxias, A.; Fabregat, A. Top. Catal.

2005, 33, 3–50.143. Coughlin, R. W. Ind. Eng. Chem. Prod. Res. Dev. 1969, 8, 12–23.144. Fidalgo, B.; Angel Menendez, J. Chinese J. Catal. 2011, 32, 207–216.145. Strelko, V. V.; Kuts, V. S.; Thrower, P. A. Carbon 2000, 38, 1499–1503.146. Szymanski, G. S.; Rychlicki, G. Carbon 1991, 29, 489–498.147. Zhao, T. J.; Sun, W. Z.; Gu, X. Y.; Ronning, M.; Chen, D.; Dai, Y. C.; Yuan, W. K.;

Holmen, A. Appl. Catal. A: Gen. 2007, 323, 135–146.148. Pereira, M. F. R.; Orfao, J. J. M.; Figueiredo, J. L. Carbon 2002, 40, 2393–2401.149. Pereira, M. F. R.; Orfao, J. J. M.; Figueiredo, J. L. Appl. Catal. A: Gen. 2001, 218,

307–318.150. Pereira, M. F. R.; Orfao, J. J. M.; Figueiredo, J. L. Appl. Catal. A: Gen. 1999, 184,

153–160.151. Pereira, M. F. R.; Orfao, J. J. M.; Figueiredo, J. L. Appl. Catal. A: Gen. 2000, 196,

43–54.152. Moreno-Castilla, C.; Carrasco-Marin, F.; Parejo-Perez, C.; Ramon, M. V. L.

Carbon 2001, 39, 869–875.153. Szymanski, G. S.; Rychlicki, G. Carbon 1993, 31, 247–257.154. Teng, H. S.; Tu, Y. T.; Lai, Y. C.; Lin, C. C. Carbon 2001, 39, 575–582.155. Szymanski, G. S.; Grzybek, T.; Papp, H. Catal. Today 2004, 90, 51–59.156. Mochida, I.; Kawabuchi, Y.; Kawano, S.; Matsumura, Y.; Yoshikawa, M. Fuel

1997, 76, 543–548.157. Raymundo-Pinero, E.; Cazorla-Amoros, D.; Linares-Solano, A. Carbon 2003, 41,

1925–1932.158. Adib, F.; Bagreev, A.; Bandosz, T. J. Langmuir 2000, 16, 1980–1986.159. Sotowa, C.; Watanabe, Y.; Yatsunami, S.; Korai, Y.; Mochida, I. Appl. Catal. A:

Gen. 1999, 180, 317–321.160. Khalil, L. B.; Girgis, B. S.; Tawfik, T. A. M. J. Chem. Technol. Biotechnol. 2001,

76, 1132–1140.161. Sanchez-Polo, M.; von Gunten, U.; Rivera-Utrilla, J. Water Res. 2005, 39,

3189–3198.162. Rivera-Utrilla, J.; Sanchez-Polo, M. Langmuir 2004, 20, 9217–9222.163. Aguilar, C.; Garcia, R.; Soto-Garrido, G.; Arriagada, R. Appl. Catal. B-Environ.

2003, 46, 229–237.164. Rocha, R. P.; Sousa, J. P. S.; Silva, A. M. T.; Pereira, M. F. R.; Figueiredo, J. L.

Appl. Catal. B-Environ. 2011, 104, 330–336.165. Kochloefl, K. In Handbook of Heterogeneous Catalysis; Ertl, G., Knozinger, H.,

Eds.; 5, Wiley-VCH: Weinheim, 1997 pp 2151–2159.166. Alkhazov, T. G.; Lisovskii, A. E. Kinet. Catal. 1976, 17, 434–439.167. Emig, G.; Hofmann, H. J. Catal. 1983, 182, 15–26.168. Schraut, A.; Emig, G.; Sockel, H. G. Appl. Catal. 1987, 29, 311–326.169. Cadus, L. E.; Arrua, L. A.; Gorriz, O. F.; Rivarola, J. B. Ind. Eng. Chem. Res. 1988,

27, 2241–2246.170. Vrieland, G. E.; Menon, P. G. Appl. Catal. 1991, 77, 1–8.171. Lisovskii, A. E.; Aharoni, C. Catal. Rev. Sci. Eng. 1994, 36, 25–74.172. Alkhazov, T. G.; Lisovskii, A. E.; Ismailov, Y. A.; Kozharov, A. I. Kinet. Catal. 1978,

19, 611–614.173. Grunewal, G. C.; Drago, R. S. J. Mol. Catal. 1990, 58, 227–233.174. Drago, R. S.; Jurczyk, K. Appl. Catal. A: Gen. 1994, 112, 117–124.

Page 45: Comprehensive Inorganic Chemistry II || Carbon

Carbon 367

175. Guerrero-Ruiz, A.; Rodrıguez-Ramos, I. Carbon 1994, 32, 23–29.176. Maksimova, N. I.; Roddatis, V. V.; Mestl, G.; Ledoux, M. Eurasian Chem. Tech. J.

2000, 2, 231–236.177. Maksimova, N.; Mestl, G.; Schlogl, R. Stud. Surf. Sci. Catal. 2001, 133,

383–389.178. Su, D. S.; Delgado, J. J.; Liu, X.; Wang, D.; Schlogl, R.; Wang, L.; Zhang, Z.;

Shan, Z.; Xiao, F. S. Chem. Asian J. 2009, 4, 1108–1113.179. Mestl, G.; Maksimova, N. I.; Keller, N.; Roddatis, V. V.; Schlogl, R. Angew. Chem.

Int. Ed. 2001, 40, 2066–2068.180. Yu, D.; Nagelli, E.; Du, F.; Dai, L. J. Phys. Chem. Lett. 2010, 1, 2165–2173.181. Pereira, M. F. R.; Figueiredo, J. L.; Orfao, J. J. M.; Serp, P.; Kalck, P.; Kihn, Y.

Carbon 2004, 42, 2807–2813.182. Delgado, J. J.; Chen, X.; Tessonnier, J. P.; Schuster, M. E.; Del Rio, E.;

Schlogl, R.; Su, D. S. Catal. Today 2010, 150, 49–54.183. Su, D.; Maksimova, N. I.; Mestl, G.; Kuznetsov, V. L.; Keller, V.; Schlogl, R.;

Keller, N. Carbon 2007, 45, 2145–2151.184. Grunewald, G. C.; Drago, R. S. J. Mol. Catal. 1990, 58, 227–233.185. Kane, M. S.; Kao, L. C.; Mariwala, R. K.; Hilscher, D. F.; Foley, H. C. Ind. Eng.

Chem. Res. 1996, 35, 3319–3331.186. Pereira, M. F. R.; Orfao, J. J. M.; Figueiredo, J. L. Colloid. Surf. A 2004, 241,

165–171.187. Iwasawa, Y.; Nobe, H.; Ogasawara, S. J. Catal. 1973, 31, 444–449.188. Schraut, A.; Emig, G.; Hofmann, H. J. Catal. 1988, 112, 221–228.189. Macia-Agullo, J. A.; Cazorla-Amoros, D.; Linares-Solano, A.; Wild, U.; Su, D. S.;

Schlogl, R. Catal. Today 2005, 102–103, 248–253.190. Delgado, J. J.; Su, D. S.; Rebmann, G.; Keller, N.; Gajovic, A.; Schlogl, R. J. Catal.

2006, 244, 126–129.191. Delgado, J. J.; Vieira, R.; Rebmann, G.; Su, D. S.; Keller, N.; Ledoux, M. J.;

Schlogl, R. Carbon 2006, 44, 809–812.192. Liu, X.; Su, D. S.; Schlogl, R. Carbon 2008, 46, 547–549.193. Begin, D.; Ulrich, G.; Amadou, J.; Su, D. S.; Pham-Huu, C.; Ziessel, R. J. Mol.

Catal. A Chem. 2009, 302, 119–123.194. Maldonado-Hodar, F. J.; Madeira, L. M.; Portela, M. F. Appl. Catal. A: Gen. 1999,

178, 49–60.195. Sui, Z. J.; Zhou, J. H.; Dai, Y. C.; Yuan, W. K. Catal. Today 2005, 106,

90–94.196. Rinaldi, A.; Zhang, J.; Frank, B.; Su, D. S.; Abd Hamid, S. B.; Schlogl, R.

ChemSusChem 2010, 22, 254–260.197. Zhang, J.; Liu, X.; Blume, R.; Zhang, A.; Schlogl, R.; Su, D. Science 2008,

322, 73–77.198. Liu, X.; Frank, B.; Zhang, W.; Cotter, T. P.; Schlogl, R.; Su, D. S. Angew. Chem.

Int. Ed Engl. 2011, 50, 3318–3322.199. Velasquez, J. D. D.; Suarez, L. A. C.; Figueiredo, J. L. Appl. Catal. A: Gen. 2006,

311, 51–57.200. Schwartz, V.; Xie, H.; Meyer, H. M., III; Overbury, S. H.; Liang, C. Carbon 2011,

49, 659–668.201. Arends, I. W. C. E.; Ophorst, W. R.; Louw, R.; Mulder, P. Carbon 1996, 34,

581–588.202. Santoro, D.; Louw, R. Carbon 2001, 39, 2091–2099.203. Santoro, D.; de Jong, V.; Louw, R. Chemosphere 2003, 50, 1255–1260.204. Furimsky, E. Carbons and Carbon Supported Catalysts in Hydroprocessing.

In RSC Catalysis Series; Spivey, J. J., Ed.; The Royal Society of Chemistry:Cambridge, 2008; vol. 1.

205. Strobel, R.; Garche, J.; Moseley, P. T.; Jorissen, L.; Wolf, G. J. Power. Sources2006, 159, 781–801.

206. Zhu, Z. Molecular Simulations applied to adsorption on and reaction with carbon.In Carbon Materials for Catalysis; Serp, P., Figueiredo, J. L., Eds.; J. Wiley &Sons: Hoboken, NJ, 2009; pp 93–129.

207. Wu, W.; Xu, J. Catal. Commun. 2004, 5, 591–595.208. de Jong, V.; Louw, R. Appl. Catal. A: Gen. 2004, 271, 153–163.209. Chen, L.; Yang, K.; Liu, H.; Wang, X. Carbon 2008, 46, 2137–2139.210. Urbano, F. J.; Marinas, J. M. J. Mol. Catal. A Chem. 2001, 173, 329–345.211. Zlotnick, J.; Prinsloo, J. J.; van Berge, P. C. J. Catal. 1978, 53, 106–115.212. Sotowa, C.; Kawabuchi, Y.; Mochida, I. Chem. Lett. 1996, 25, 967–968.213. Mochida, I.; Yasumoto, Y.; Watanabe, Y.; Fujitsu, H. H.; Kojima, Y.; Morita, M.

Chem. Lett. 1994, 23, 197–200.214. Calvino-Casilda, V.; Lopez-Peinado, A. J.; Duran-Valle, C. J.; Martın-

Aranda, R. M. Catal. Rev. 2010, 52, 325–380.215. Bitter, J. H. J. Mater. Chem. 2010, 20, 7312–7321.216. Juntgen, H. Fuel 1986, 65, 1436–1446.217. Bitter, H.; de Jong, K. P. Preparation of carbon-supported metal catalysts.

In Carbon Materials for Catalysis; Serp, P., Figueiredo, J. L., Eds.; J. Wiley &Sons: Hoboken, NJ, 2009; pp 429–480.

218. Martın-Gullon, A.; Prado-Burguete, C.; Rodrıguez-Reinoso, F. Carbon 1993, 31,1099–1105.

219. Cabiac, A.; Cacciaguerra, T.; Trens, P.; Durand, R.; Delahay, G.; Medevielle, A.;Plee, D.; Coq, B. Appl. Catal. A: Gen. 2008, 340, 229–235.

220. Dawidziuk, M. B.; Carrasco-Marın, F.; Moreno-Castilla, C. Carbon 2009, 47,2679–2687.

221. Rao, V.; Simonov, P. A.; Savinova, E. R.; Plaksin, G. V.; Cherepanova, S. V.;Kryukova, G. N.; Stimming, U. J. Power. Sources 2005, 145, 178–187.

222. Maruyama, J.; Sumino, K.-i.; Kawaguchi, M.; Abe, I. Carbon 2004, 42,3115–3121.

223. Li, Z.; Chen, Z. X.; Kang, G. J.; He, X. Catal. Today 2011, 165, 25–31.224. Chen, Z.; Guan, Z.; Li, M.; Yang, Q.; Li, C. Angew. Chem. Int. Ed. 2011, 50,

4913–4917.225. Goettmann, F.; Sanchez, C. J. Mater. Chem. 2007, 17, 24–30.226. Santiso, E. E.; Buongiorno Nardelli, M.; Gubbins, K. E. J. Chem. Phys. 2008,

128, 034704.227. Santiso, E. E.; Kostov, M. K.; George, A. M.; Buongiorno Nardelli, M.;

Gubbins, K. E. Appl. Surf. Sci. 2007, 253, 5570–5579.228. Laine, J.; Severino, F.; Labady, M. J. Catal. 1994, 147, 355–357.229. Laine, J.; Labady, M.; Severino, F.; Yunes, S. J. Catal. 1997, 166, 384–387.230. Trimm, D. L.; Cooper, B. J. J. Chem. Soc. D 1970, 477–478.231. Kogan, S.; Landau, M. V.; Herskowitz, M.; Koresh, J. E. Stud. Surf. Sci. Catal.

1993, 78, 353–359.232. L’Argentiere, P. C.; Quiroga, M. E.; Liprandi, D. A.; Cagnola, E. A.; Roman-

Martınez, M. C.; Dıaz-Aunon, J. A.; Salinas-Martınez de Lecea, C. Catal. Lett.2003, 87, 97–101.

233. Miura, K.; Hayashi, J.; Kawaguchi, T.; Hashimoto, K. Carbon 1993, 31, 667–674.234. Lafyatis, D. S.; Foley, H. C. Chem. Eng. Sci. 1990, 45, 2567–2574.235. Rodrıguez-Reinoso, F.; Salinas-Martınez de Lecea, C.; Sepulveda-Escribano, A.;

Lopez-Gonzalez, J. D. Catal. Today 1990, 7, 287–298.236. Xiong, H.; Moyo, M.; Motchelaho, M. A. M.; Jewell, L. L.; Coville, N. J. Appl.

Catal. A: Gen. 2010, 388, 168–178.237. Wenkin, M.; Touillaux, R.; Ruiz, P.; Delmon, B.; Devillers, M. Appl. Catal. A: Gen.

1996, 148, 181–199.238. Echeandia, S.; Arias, P. L.; Barrio, V. L.; Pawelec, B.; Fierro, J. L. G. Appl. Catal. B

Env. 2010, 101, 1–12.239. Li, L.; Zhu, Z. H.; Yan, Z. F.; Lu, G. Q.; Rintoul, L. Appl. Catal. A: Gen. 2007, 320,

166–172.240. Derbyshire, F. J.; de Beer, V. H. J.; Abotsi, G. M. K.; Scaroni, A. W.; Solar, J. M.;

Skrovanek, D. J. Appl. Catal. 1986, 27, 117–131.241. Prado-Burguete, C.; Linares-Solano, A.; Rodrıguez-Reinoso, F.; Salinas-Martınez

de Lecea, C. J. Catal. 1989, 115, 98–106.242. Prado-Burguete, C.; Linares-Solano, A.; Rodrıguez-Reinoso, F.; Salinas-Martınez

de Lecea, C. J. Catal. 1991, 128, 397–404.243. Han, W.; Liu, H.; Zhou, H. Catal. Commun. 2007, 8, 351–354.244. Zhu, Z. H.; Wang, S.; Lu, G. Q.; Zhang, D. K. Catal. Today 1999, 53, 669–681.245. Rodrigues, E. G.; Pereira, M. F. R.; Chen, X.; Delgado, J. J.; Orfao, J. J. M.

J. Catal. 2011, 281, 119–127.246. Wang, X.; Li, N.; Webb, J. A.; Pfefferle, L. D.; Haller, G. L. Appl. Catal. B Env.

2010, 101, 21–30.247. Guerrero-Ruiz, A.; Rodrıguez-Ramos, I.; Rodrıguez-Reinoso, F.; Moreno-

Castilla, C.; Lopez-Gonzalez, J. D. Carbon 1988, 26, 417–423.248. Abate, S.; Arrigo, R.; Schuster, M. E.; Perathoner, S.; Centi, G.; Villa, A.; Su, D.;

Schlogl, R. Catal. Today 2010, 157, 280–285.249. Radkevich, V. Z.; Senko, T. L.; Wilson, K.; Grishenko, L. M.; Zaderko, A. N.;

Diyuk, V. Y. Appl. Catal. A: Gen. 2008, 335, 241–251.250. Garcıa-Garcıa, F. R.; Alvarez-Rodrıguez, J.; Rodrıguez-Ramos, I.; Guerrero-

Ruiz, A. Carbon 2010, 48, 267–276.251. Wachowski, L.; Skupinski, W.; Hofman, M. Appl. Catal. A: Gen. 2006, 303,

230–233.252. Milone, C.; Gangemi, C.; Ingoglia, R.; Neri, G.; Galvagno, S. Appl. Catal. A: Gen.

1999, 184, 89–94.253. Phillips, J.; Weigle, J.; Herskowitz, M.; Kogan, S. Appl. Catal. A 1998, 173,

273–287.254. Coloma, F.; Sepulveda-Escribano, A.; Fierro, J. L. G.; Rodrıguez-Reinoso, F. Appl.

Catal. A: Gen. 1996, 136, 231–248.255. Coloma, F.; Sepulveda-Escribano, A.; Fierro, J. L. G.; Rodrıguez-Reinoso, F. Appl.

Catal. A: Gen. 1996, 148, 63–80.256. Teddy, J.; Falqui, A.; Corrias, A.; Carta, D.; Lecante, P.; Gerber, I.; Serp, P.

J. Catal. 2011, 278, 59–70.257. Chen, A.; Vannice, M. A.; Phillips, J. J. Phys. Chem. 1987, 91, 6257–6269.258. Prasomsri, T.; Shi, D.; Resasco, D. E. Chem. Phys. Lett. 2010, 497, 103–107.259. Efremenko, I.; Sheintuch, M. J. Catal. 2003, 214, 53–67.

Page 46: Comprehensive Inorganic Chemistry II || Carbon

368 Carbon

260. Lv, Y. A.; Cui, Y. H.; Li, X. N.; Song, X. Z.; Wang, J. G.; Dong, M. Physica E2010, 42, 1746–1750.

261. Keyser, M. M.; Prinsloo, F. F. Stud. Surf. Sci. Catal. 2007, 163, 45–73.262. Jung, H. J.; Walker, P. L., Jr.; Vannice, M. A. J. Catal. 1982, 75, 416–422.263. Niemanstsverdriet, J. W.; van der Kraan, A. M.; Delgass, W. N.; Vannice, M. A.

J. Phys. Chem. 1989, 89, 67–72.264. Chen, A. A.; Kaminsky, M.; Geoffroy, G. L.; Vannice, M. A. J. Phys. Chem. 1986,

90, 4810–4819.265. Pantea, D.; Darmstadt, H.; Kaliaguine, S.; Roy, C. Appl. Surf. Sci. 2003, 217,

181–193.266. Luck, F. Bull. Soc. Chim. Belg. 1991, 100, 781–800.267. Abotsi, G. M. K.; Scaroni, A. W. Fuel Process. Technol. 1989, 22, 107–133.268. Joshi, Y. V.; Ghosh, P.; Daage, M.; Delgass, W. N. J. Catal. 2008, 257, 71–80.269. Breysse, M.; Geauntet, C.; Afanasiev, P.; Blanchard, J.; Vrinat, M. Catal. Today

2008, 130, 3–13.270. Besenbacher, F.; Brorson, M.; Clausen, B. S.; Helveg, S.; Hinnemann, B.;

Kibsgaard, J.; Lauritsen, J. V.; Moses, P. G.; Nørshov, J. K.; Topsøe, H. Catal.Today 2008, 130, 86–96.

271. Bridgewater, A. J.; Burch, R.; Mitchell, P. C. H. Appl. Catal. 1982, 4, 267–273.272. Chadwick, D.; Breysse, M. J. Catal. 1981, 71, 226–227.273. Farag, H.; Mochida, I.; Sakanishi, K. Appl. Catal. A: Gen. 2000, 194–195,

147–157.274. Topsoe, H.; Clausen, B. S.; Massoth, F. E. Hydrotreating catalysts. In Catalysis

Science and Technology; Anderson, J. R., Boudart, M., Anderson, J. R.,Boudart, M., Eds.; 11, Springer: Berlin, 1996 pp 1–310.

275. Chianelli, R. R.; Pecoraro, T. A. J. Catal. 2001, 198, 9–19.276. Duchet, J. C.; van Oers, E. M.; de Beer, V. H. J.; Prins, R. J. Catal. 1983,

80, 386–402.277. Thomas, R.; van Oers, E. M.; de Beer, V. H. J.; Medema, J.; Moulijn, J. A. J. Catal.

1982, 76, 241–253.278. de Beer, V. H. J.; Derbyshire, F. J.; Groor, C. K.; Prins, R.; Scaroni, A. W.;

Solar, J. M. Fuel 1984, 63, 1095–1100.279. Scheffer, B.; Arnoldy, P.; Moulijn, J. A. J. Catal. 1988, 112, 516–527.280. Puello-Polo, E.; Brito, J. L. Catal. Today 2010, 149, 316–320.281. Wen, X. D.; Cao, Z.; Li, Y. W.; Wang, J.; Jiao, H. J. Phys. Chem. B 2006,

110, 23860–23869.282. Chianelli, R. R.; Berhault, G. Catal. Today 1999, 53, 357–366.283. Kibsgaard, J.; Lauritsen, J. V.; Lægsgaard, E.; Clausen, B. S.; Topsøe, H.;

Besenbacher, F. J. Am. Chem. Soc. 2006, 128, 13950–13958.284. Shang, H.; Liu, C.; Xu, Y.; Qiu, J.; Wei, F. Fuel Process. Technol. 2007, 88,

117–123.285. Eswaramoorthi, I.; Sundaramurthy, V.; Das, N.; Dalai, A. K.; Adjaye, J. Appl. Catal.

A: Gen. 2008, 339, 187–195.286. Nakhaei Pour, A.; Rashidi, A. M.; Jozani, K. J.; Mohajeri, A.; Khorami, P. J. Nat.

Gas Chem. 2010, 19, 91–95.287. Yu, Z.; Fareid, L. E.; Moljord, K.; Blekkan, E. A.; Walmsley, J. C.; Chen, D. Appl.

Catal. B Env. 2008, 84, 482–489.288. Tan, Z.-l; Xiao, H.-n; Zhang, R.-d; Zhang, Z. S.; Kaliaguine, S. New Carbon Mat.

2009, 24, 333–343.289. Hussain, M.; Ihm, S. K. Ind. Eng. Chem. Res. 2009, 48, 698–707.290. Stanislaus, A.; Marafi, A.; Rana, M. S. Catal. Today 2010, 153, 1–68.291. Song, C.; Ma, X. Appl. Catal. B Env. 2003, 41, 207–238.292. Pecoraro, T.; Chianelli, R. R. J. Catal. 1981, 67, 430–445.293. Ledoux, M. J.; Michiaux, O.; Agostini, J.; Panissod, P. J. Catal. 1986, 102,

275–288.294. Lacroix, M.; Boutarfa, N.; Guillard, C.; Vrinat, M.; Breysse, M. J. Catal. 1989, 120,

473–477.295. Lagos, G.; Garcıa, R.; Lopez Agudo, A.; Yates, M.; Fierro, J. L. G.; Gil-

Llambıas, F. J.; Escalona, N. Appl. Catal. A: Gen. 2009, 358, 26–31.296. Furimsky, E. Appl. Catal. A: Gen. 2003, 240, 1–28.297. Abu, I. I.; Smith, K. J. Appl. Catal. A: Gen. 2007, 328, 58–67.298. Oyama, S. T.; Gott, T.; Zhao, H.; Lee, Y. K. Catal. Today 2009, 143, 94–107.299. Shu, Y.; Oyama, S. T. Carbon 2005, 43, 1517–1532.300. Vissers, J. P. R.; Lensing, T. J.; de Beer, V. H. J.; Prins, R. Appl. Catal. 1987, 30,

21–31.301. Kouzu, M.; Kuriki, Y.; Hamdy, F.; Sakanishi, K.; Sugimoto, Y.; Saito, I. Appl. Catal.

A: Gen. 2004, 265, 61–67.302. Lee, J. J.; Han, S.; Kim, H.; Koh, J. H.; Hyeon, T.; Moon, S. H. Catal. Today 2003,

86, 141–151.303. Gheek, P.; Suppan, S.; Trawczynski, J.; Hynaux, A.; Sayag, C.; Djega-

Mariadssou, G. Catal. Today 2007, 119, 19–22.304. Hubaut, R.; Altafulla, J.; Rives, A.; Scott, C. Fuel 2007, 86, 743–749.305. Zhou, A.; Ma, X.; Song, C. Appl. Catal. B Env. 2009, 87, 190–199.

306. Calafat, A.; Laine, J.; Lopez-Agudo, A.; Palacios, J. M. J. Catal. 1996, 162,20–30.

307. Solar, J. M.; Derbyshire, F. J.; de Beer, V. H. J.; Radovic, L. R. J. Catal. 1991,129, 330–342.

308. Sakanishi, K.; Nagamatsu, T.; Mochida, I.; Whitehurst, D. D. J. Mol. Catal. AChem. 2000, 155, 101–109.

309. La Vopa, V.; Satterfield, C. N. J. Catal. 1988, 110, 375–387.310. Sayag, C.; Suppan, S.; Trawczynski, J.; Djega-Mariadassou, G. Fuel Process.

Technol. 2002, 77–78, 261–267.311. Altajan, M. A.; Kriz, J. F.; Ternan, M. The effect of diffusion on the deactivation

of catalysts used for hydrocraking Athabasca bitumen. In Catalyst deactivation;Bartolomew, C. H., Butt, J. B., Eds.; Elsevier: Amsterdam, 1991; pp 315–322.

312. Centeno, A.; Laurent, E.; Delmon, B. J. Catal. 1995, 154, 288–298.313. Laurent, E.; Delmon, B. Appl. Catal. 1994, 109, 97–115.314. Bulushev, D. A.; Ross, J. R. H. Catal. Today 2011, 171, 1–13. http://dx.doi.org/

10.1016/j.cattod.2011.02.005.315. Bahome, M. C.; Jewell, L. L.; Padayachy, K.; Hildebrandt, D.; Glasser, D.;

Datye, A. K.; Coville, N. J. Appl. Catal. A: Gen. 2007, 328, 243–251.316. Moreno-Castilla, C.; Carrasco-Marın, F. J. Chem. Soc. Faraday Trans. 1995, 91,

3519–3524.317. Bezemer, G. L.; Radstake, P. B.; Koot, V.; van Dillen, A. J.; Geus, J. W.; de

Jong, K. P. J. Catal. 2006, 237, 291–302.318. Bezemer, G. L.; Radstake, P. B.; Falke, U.; Oosterbeek, H.; Kuipers, H. P. C. E.; van

Dillen, A. J.; de Jong, K. P. J. Catal. 2006, 237, 152–161.319. Bezemer, G. L.; Bitter, J. H.; Kuipers, H. P. C. E.; Oosterbeek, H.; Holewijn, J. E.;

Xu, X.; Kapteijn, F.; van Dillen, A. J.; de Jong, K. P. J. Am. Chem. Soc. 2006, 128,3956–3964.

320. Sun, L. B.; Zong, Z. M.; Kou, J. H.; Zhang, L. F.; Ni, Z. H.; Yu, G. Y.; Chen, H.;Wei, X. Y. Energy Fuel 2004, 18, 1500–1504.

321. Sun, L. B.; Wei, X. Y.; Liu, X. Q.; Zong, Z. M.; Li, W.; Kou, J. H. Energy Fuel 2009,23, 4877–4882.

322. Augustyn, W. G.; McCrindle, R. I.; Coville, N. J. Appl. Catal. A: Gen. 2010,388, 1–6.

323. Semagina, N.; Renken, A.; Kiwi-Minsker, L. Chem. Eng. Sci. 2007, 62,5344–5348.

324. Badano, J. M.; Betti, C.; Rintoul, I.; Vich-Berlanga, J.; Cagnola, E.; Torres, G.;Vera, C.; Yori, J.; Quiroga, M. Appl. Catal. A: Gen. 2010, 390, 166–174.

325. Bazzazzadegan, H.; Kazemeini, M.; Rashidi, A. M. Appl. Catal. A: Gen. 2011, 399,184–190.

326. Upare, P. P.; Lee, J. M.; Won Hwang, D.; Halligudi, S. B.; Hwang, Y. K.;Chang, J. S. J. Ind. Eng. Chem. 2011, 17, 287–292.

327. Sakamoto, Y.; Kamiya, Y.; Okuhara, T. J. Mol. Catal. A Gen. 2006, 250, 80–86.328. Hoffer, B. W.; Crezee, E.; Mooijman, P. R. M.; van Langeveld, A. D.; Kapteijn, F.;

Moulijn, J. A. Catal. Today 2003, 79–80, 35–41.329. Xing, L.; Qiu, J.; Liang, C.; Wang, C.; Mao, L. J. Catal. 2007, 250, 369–372.330. Jiang, L.; Gu, H.; Xu, X.; Yan, X. J. Mol. Catal. A Chem. 2009, 310, 144–149.331. Neri, G.; Rizzo, G.; Milone, C.; Galvagno, S.; Musolino, M. G.; Capannelli, G.

Appl. Catal. A: Gen. 2003, 249, 303–311.332. Tuley, W. F.; Adams, R. J. J. Am. Chem. Soc. 1925, 45, 3061–3068.333. Maki-Arvela, P.; Hajek, J.; Salmi, T.; Murzin, D. Y. Appl. Catal. A: Gen. 2005, 292,

1–49.334. Gallezot, P.; Richard, D. Catal. Rev. Sci. Eng. 1998, 40, 81–126.335. Pham-Huu, C.; Keller, N.; Charbonniere, L. J.; Ziessel, R.; Ledoux, M. J. Chem.

Commun. 2000, 19, 1871–1872.336. Pham-Huu, C.; Keller, N.; Ehret, G.; Charbonniere, L. J.; Ziessel, R.; Ledoux, M. J.

J. Mol. Catal. A Gen. 2001, 170, 155–163.337. Vu, H.; Goncalves, F.; Philippe, R.; Lamouroux, E.; Corrias, M.; Kihn, Y.; Plee, D.;

Kalck, P.; Serp, P. J. Catal. 2006, 240, 18–22.338. Giroir-Fendler, A.; Richard, D.; Gallezot, P. Stud. Surf. Sci. Catal. 1988, 4,

171–178.339. Richard, D.; Fouilloux, P.; Gallezot, P. In Proceeding of the 9th International

Congress on Catalysis, Phillips, M. J.; Ternan, M., Eds.; Calgary: Canada,1988,p. 1074 –1082.

340. Harada, T.; Ikeda, S.; Miyazaki, M.; Sakata, T.; Mori, H.; Matsumura, M. J. Mol.Catal. A Chem. 2007, 268, 59–64.

341. Asedegbega-Nieto, E.; Bachiller-Baeza, B.; Guerrero-Ruız, A.; Rodrıguez-Ramos, I. Appl. Catal. A: Gen. 2006, 300, 120–129.

342. Breen, J. P.; Burch, R.; Gomez-Lopez, J.; Griffin, K.; Hayes, M. Appl. Catal. A:Gen. 2004, 268, 267–274.

343. Jung, A.; Jess, A.; Schubert, T.; Schutz, W. Appl. Catal. A: Gen. 2009, 362,95–105.

344. Li, Y.; Li, Z. G.; Zhou, R. X. J. Mol. Catal. A Chem. 2008, 279, 140–146.345. Li, Y.; Lai, G. H.; Zhou, R. X. Appl. Surf. Sci. 2007, 253, 4978–4984.

Page 47: Comprehensive Inorganic Chemistry II || Carbon

Carbon 369

346. Coloma, F.; Sepulveda-Escribano, A.; Rodrıguez-Reinoso, F. Appl. Catal. A: Gen.1997, 150, 165–183.

347. Toebes, M. L.; Prinsloo, F. F.; Bitter, J. H.; van Pillen, A. J.; de Jong, K. P. J. Catal.2003, 214, 78–87.

348. Toebes, M. L.; Zhang, Y.; Hajek, J.; Nijhuis, T. A.; Bitter, J. H.; Jos van Pillen, A.;Murzin, D. Y.; Koningsberger, D. C.; De Jong, K. P. J. Catal. 2004, 226,215–225.

349. Toebes, M. L.; Nijhuis, T. A.; Hajek, J.; Bitter, J. H.; van Dillen, A. J.; Murzin, D. Y.;de Jong, K. P. Chem. Eng. Sci. 2005, 60, 5682–5695.

350. Tang, T.; Yin, C.; Xiao, N.; Guo, M.; Xiao, F. Catal. Lett. 2009, 127, 400–405.351. Amorim, C.; Keane, M. A. J. Chem. Technol. Biotechnol. 2008, 83, 662–672.352. Zhao, L.; Zhou, J.; Chen, H.; Zhang, M.; Sui, Z.; Zhou, X. Korean J. Chem. Eng.

2010, 27, 1412–1418.353. Szymanski, G. S.; Karpinski, Z.; Biniak, S.; Swiatkowski, A. Carbon 2002, 40,

2627–2639.354. Fujishima, A.; Rao, T. N.; Tryk, D. A. J Photochem. Photobiol. C Photochem. Rev.

2000, 1, 1–21.355. Faria, J. L.; Wang, W. Carbon materials in photocatalysis. In Carbon Materials for

Catalysis; Serp, P., Figueiredo, J. L., Eds.; J. Wiley & Sons: Hoboken, NJ, 2009;pp 481–506.

356. Leary, R.; Westwood, A. Carbon 2011, 49, 741–772.357. Velasco, L. F.; Parra, J. B.; Ania, C. O. Appl. Surf. Sci. 2010, 256, 5254–5258.358. Matos, J.; Laine, J.; Herrmann, J. M. Appl. Catal. B Environ. 1998, 18, 281–291.359. Matos, J.; Laine, J.; Herrmann, J. M. Carbon 1999, 37, 1870–1872.360. Zhang, X.; Zhou, M.; Lei, L. Carbon 2005, 43, 1700–1708.361. Lee, D. K.; Kim, S. C.; Kim, S. J.; Chung, I. S.; Kim, S. W. Chem. Eng. J. 2004,

102, 93–98.362. Arana, J.; Dona-Rodrıguez, J. M.; Tello Rendon, E.; Garriga i Cabo, C.; Gonzalez-

Dıaz, O.; Herrera-Melian, J. A.; Perez-Pena, J.; Colon, G.; Navıo, J. A. Appl. Catal.B Environ. 2003, 44, 153–160.

363. Xue, G.; Liu, H.; Chen, Q.; Hills, C.; Tyrer, M.; Innocent, F. J. Hazard. Mater.2011, 186, 765–772.

364. Zhang, X.; Zhou, M.; Lei, L. Carbon 2006, 44, 325–333.365. Gao, B.; Yap, P. S.; Lim, T. M.; Lim, T. T. Chem. Eng. J. 2011, 171, 1098–1107.366. Lin, X.; Rong, F.; Ji, X.; Fu, D. Microp. Mesop. Mater. 2011, 142, 276–281.367. Hoffmann, M. R.; Martin, S. T.; Choi, W. Y.; Bahnemann, D. W. Chem. Rev. 1995,

95, 69–96.368. Mills, A.; LeHunte, S. J. Photochem. Photobiol. A 1997, 108, 1–35.369. Sellappan, R.; Galeckas, A.; Venkatachalapathy, V.; Kuznetsov, A. Y.; Chakarov, D.

Appl. Catal. B Environ. 2011, 106, 337–342. http://dx.doi.org/10.1016/j.apcatb.2011.05.036.

370. Wang, Y.; Shi, R.; Lin, J.; Zhu, Y. Appl. Catal. B Environ. 2010, 100, 179–183.371. Wang, W. D.; Serp, P.; Kalck, P.; Faria, J. L. J. Mol. Catal. A Chem. 2005,

235, 194–199.372. Kamat, P. V.; Bedja, I.; Hotchandani, S. J. Phys. Chem. 1994, 98, 9137–9142.373. Banerjee, S.; Wong, S. S. Nano Lett. 2002, 2, 195–200.374. Jung, K. H.; Hong, J. S.; Vittal, R.; Kim, K. J. Chem. Lett. 2002, 31, 864–865.375. Lettmann, C.; Hildenbrand, K.; Kisch, H.; Macyk, W.; Maier, W. F. Appl. Catal. B

Environ. 2001, 32, 215–227.376. Yu, J.; Ma, T.; Liu, S. Phys. Chem. Chem. Phys. 2011, 13, 3491–3501.377. Yu, Y.; Yu, J. C.; Yu, J. G.; Kwok, Y. C.; Che, Y. K.; Che, Y. K.; Zhao, J. C.;

Ding, L.; Ge, W. K.; Wong, P. K. Appl Catal A: Gen 2005, 289, 186–196.378. Ahmmad, B.; Kusumoto, Y.; Somekawa, S.; Ikeda, M. Catal. Commun. 2008,

9, 1410–1413.379. Zhu, Z.; Zhou, Y.; Yu, H.; Nomura, T.; Fugetsu, B. Chem. Lett. 2006, 35,

890–891.380. Yao, Y.; Li, G.; Ciston, S.; Lueptow, R. M.; Gray, K. A. Environ. Sci. Technol.

2008, 42, 4952–4957.381. Yen, C. Y.; Lin, Y. F.; Hung, C. H.; Tseng, Y. H.; Ma, C. C.; Chang, M. C.; Shao, H.

Nanotechnology 2008, 19, 045604.