Doping Carbons Beyond Nitrogen (PDF)

Energy & Environmental Science...

Energy & Environmental Science View Article Online REVIEW View Journal | View Issue Doping carbons beyond nitrogen: an overview of advanced heteroatom doped carbons with boron, Published on 01 August 2013. Downloaded by University of Newcastle on 08/10/2017 16:22:50. Cite this: Energy Environ. Sci., 2013, 6, 2839 sulphur and phosphorus for energy applications Jens Peter Paraknowitsch* and Arne Thomas Heteroatom doped carbon materials represent one of the most prominent families of materials that are used in energy related applications, such as fuel cells, batteries, hydrogen storage or supercapacitors. While doping carbons with nitrogen atoms has experienced great progress throughout the past decades and yielded promising material concepts, also other doping candidates have gained the researchers' interest in the last few years. Boron is already relatively widely studied, and as its electronic situation is contrary to the one of nitrogen, codoping carbons with both heteroatoms can probably create synergistic effects. Sulphur and phosphorus have just recently entered the world of carbon synthesis, but already the first studies published prove their potential, especially as electrocatalysts in the cathodic compartment of fuel cells. Due to their size and their electronegativity being lower than those of carbon, structural distortions and changes of the Received 26th April 2013 Accepted 1st August 2013 charge densities are induced in the carbon materials. This article is to give a state of the art update on the most recent developments concerning the advanced heteroatom doping of carbon that goes beyond DOI: 10.1039/c3ee41444b nitrogen. Doped carbon materials and their applications in energy devices are discussed with respect to www.rsc.org/ees their boron-, sulphur- and phosphorus-doping. Broader context The research and design of novel materials that are applicable in various energy devices represent one of the central and most thriving subjects within today’s scientic work. The challenges do not only rely on the tremendous need to overcome traditional fossil fuel based energy recovery. While a lot of sustainable concepts already exist, these require the development of high performance materials that are able to cope with certain challenges, e.g. the dependence on highly expensive and rather not abundantly available noble metals, and also a lack of long term stability of those. Researchers have been carrying out numerous studies concentrating on nding alternative materials that can be applied in novel energy devices. Those alternative materials should most preferably be free of noble metals, avoid expensive precursor systems, be sustainable and not rely on the fossil energy sources that are to be replaced, and of course exhibit high activity in the devices they are designed for. One class of materials in whose development and understanding researchers have put strong effort is heteroatom-doped carbon materials. By heteroatom doping the properties are altered compared to crude carbon materials. The by far most intensely studied doping candidate is nitrogen, capable of not only increasing electric conductivity, but also the catalytic activity of carbons. Such N-doped carbons have advanced tremendously in the past few years, especially by proving their usefulness as electrocatalysts for the reduction of oxygen in fuel cell cathodes, or as electrode materials in super- capacitors. Meanwhile the spectrum of doping has been widened, and novel doped carbons have been reported, indicating the promising potential of such materials in the eld of novel forms of energy recovery. The envisaged aim of this research is contributing towards the solutions of the major global challenges of energy supply that are currently faced due to the severe problem of climate change, while at the same time a growing need for energy is encountered. This article is supposed to give an overview of the inuence of doped carbons on the progress of novel energy devices, pointing out the signicant importance of this class of materials for the future. Introduction candidates for certain applications. When it comes to different energy applications, such as electrocatalysis in fuel cells, an Graphite and its related carbon structures – such as carbon advanced modication of the pure carbon can lead to a signif- nanotubes1 or graphene2,3 – represent a thriving class of mate- icantly enhanced performance. One way to modify plain rials, with a high potential to be applied in numerous promising carbons is to dope the respective materials with heteroatoms, elds, such as sensors or photovoltaics.4–6 Nevertheless, the which basically means to chemically attach or incorporate them purest carbon materials are not always the most suitable into the backbone of the respective carbon material. Many studies have been performed; while – especially regarding the design of materials for energy related materials – nitrogen is by Technische Universität Berlin, Institute of Chemistry, Division of Functional Materials, far the most abundantly investigated heteroatom because – as a Hardenbergstr. 40, 10623 Berlin, Germany. E-mail: jens.p.paraknowitsch@tu- berlin.de “neighbour” of carbon – it is chemically relatively easy to bring This journal is ª The Royal Society of Chemistry 2013 Energy Environ. Sci., 2013, 6, 2839–2855 | 2839  View Article Online Energy & Environmental Science Review the two types of atoms together. The result is a colourful electrochemically in the cathodic half cell, inspired by redox plethora of nitrogen containing carbon based materials that active enzymes.14 Thermal treatment of related catalysts nally can exhibit variable property proles. Basically, this class of increased both activity and stability;15 so the concept of N-doped material can be divided into two major groups with contrary carbons as electrocatalysts was born. It nevertheless took more properties: nitrogen rich carbon nitrides exhibiting a rather than a decade to prove that also precursors not related to bio- stoichiometric composition of CxNy$x on the one hand, and logical redox systems can successfully create ORR catalysts.16 Up nitrogen-doped carbon materials in which only a small to this point of time it yet remained unclear whether the N- percentage of the atoms in the carbon backbone are doping, the metals in the respective materials or a synergy of substituted by nitrogen atoms on the other hand. The devel- both are essential for good ORR activity. A complete overview of opment of stoichiometric nitrogen-rich carbon nitrides Published on 01 August 2013. Downloaded by University of Newcastle on 08/10/2017 16:22:50. the development of such a metal complex mimicking catalysts included the postulation of two different allotropes, namely can be derived from an accordant review provided by Zhang cubic C3N4 and graphitic C3N4, of which especially the cubic et al. in 2008.17 As further progress was achieved, the necessity structure was subject of theoretical studies, due to its pre- of metals in the ORR catalysts was more and more doubted, as dicted hardness being higher than that of diamond, while experimental results gave more hints on an intrinsic electro- feasible synthetic procedures remain unknown.7–9 Unlike its catalytic activity of N-doped carbons. In 2006 completely metal cubic counterpart, the graphitic allotrope g-C3N4 has been the free catalysts were proven to be active ORR catalysts by Ozkan subject of numerous synthetic approaches, based on the tri- et al., nally evidencing the benecial inuence on the activity. merisation of nitrile units, which actually dates back to 1834.10 While the possibility of a crucial inuence of metals used These carbon nitrides represent a class of materials on its own throughout the synthetic procedure – that were removed aer- that lies beyond the scope of this manuscript. The reader is wards – on the formation of C/N-sites with electrocatalytic thus referred to early reviews on the subject,11,12 and also to an activity was mentioned, the authors also already considered the overview focusing on the photocatalytic and generally hetero- idea of a favourably changed charge prole at the carbon atoms catalytic applications of carbon nitrides,13 allowing for neighbouring the more electronegative nitrogen atoms, which obtaining a complete picture of these semiconductors at the might ease the interaction with molecular oxygen. Also the edge of inorganic and organic materials. possibility that nitrogen atoms bound in pyridinic sites at the N-doped carbons for ORR and supercapacitors edges of carbon sheets play an important role was implied.18 From the semiconducting properties of graphitic carbon The design and improvement of now metal free catalysts for nitrides it already becomes obvious why the fundamentally ORR are too complex to be discussed here in full detail, and an contrary properties of N-rich carbon nitrides and N-doped accordant review article by Shao et al. is hence recommended.19 carbons have been stated previously. The merely doped mate- Later vertically aligned nitrogen-doped carbon nanotube arrays rials, with nitrogen contents of up to 10%, rather retain the were presented by Dai et al., remarkably exceeding the activity of properties of their non-doped analogues, nevertheless nely conventional Pt@C catalysts, while additionally exhibiting an tuning and thus – considering certain energy related applica- outstanding ORR selectivity avoiding deactivation by crossover tions – improving them. The roots of the development of N- effects in an alkaline medium.20 As for the rst time Pt@C doped carbons can be found in the eld of fuel cell research and catalysts are le behind, at least under alkaline conditions; this date back to 1964, when Jasinski et al. applied metal–phthalo- can be considered as one of the most important breakthroughs cyanine-complexes to catalyse the reduction of oxygen in the research on N-doped carbon based electrocatalysts. The Jens Peter Paraknowitsch Arne Thomas studied Chemistry studied chemistry in Marburg in Gieben, Marburg and Edin- and Valencia. During his Ph.D. burgh and received his PhD from (2007–2009) at Max-Planck- the Max Planck Institute for Institute of Colloids and Inter- Colloid and Interfaces in Pots- faces in Potsdam, under super- dam/Golm. Aer a postdoctoral vision of Professor Markus stay at the University of Cal- Antonietti and Professor Arne ifornia, Santa Barbara, as an Thomas, he started focusing on AvH fellow, he rejoined the MPI the chemistry of carbon mate- for Colloids and Interfaces as a rials. He contributed to the group leader. In 2009 he became development of nitrogen-doped a Professor for Inorganic Chem- carbon materials from dicyana- istry at the Technical University mide-based ionic liquids – a concept he is now applying and Berlin where he is leading the department of Functional Materials. extending for the synthesis of versatile mesoporous inorganic His research focuses on porous materials—from mesoporous materials and composites during his work as a scientist at Tech- inorganic materials to microporous organic frameworks. nical University of Berlin. 2840 | Energy Environ. Sci., 2013, 6, 2839–2855 This journal is ª The Royal Society of Chemistry 2013  View Article Online Review Energy & Environmental Science reasons for the electrochemical activity of such materials are liquids from 2006 in which the potential of nitrile function- meanwhile seen in the charge prole induced in the carbon alised ionic liquids is indicated, as they do not decompose backbone by the electronegativity of the doping atoms, and also completely to volatile products under an inert gas.55 This led in the inuence of pyridinic nitrogen atoms with their free lone to profound and detailed studies on how these ionic liquids pair available for interaction with oxygen.18,21–24 Furthermore, N- can be used as a nitrogen-doped carbon source, ranging from doped carbon materials can not only be used as ORR catalysts mechanistic and fundamental points of view56–60 to more directly, but also improve the stability of classical Pt nano- application oriented studies, as the derived materials are particle catalysts. This could be proven on carbon nanotubes promising candidates, e.g. for fuel cell applications.25,61–63 Also modied with ionic liquid derived N-doped carbon using time supramolecular types of ionic liquids can form systems suit- resolved in situ small angle X-ray scattering, clearly pointing out able for N-doped carbon synthesis.64 Meanwhile also nitrogen Published on 01 August 2013. Downloaded by University of Newcastle on 08/10/2017 16:22:50. a signicant increase of stability of the particles against coars- rich organic networks and frameworks have been thermally ening and agglomeration achieved by the N-doping.25 As the treated to give nitrogen-containing carbon materials65–67 with development of noble metal free catalysts might still be facing promising properties, e.g. for supercapacitor applications66 or certain challenges, this concept shows that N-doped carbon can as cathodic materials in lithium batteries.67 also be applied in a kind of bridging technology. As we tried to show throughout this introduction, nitrogen- Apart from being applied as ORR catalysts, the second major doping is nowadays a widely applied concept in carbon energy related eld of applications for N-doped carbons is material research. Although it is almost impossible to provide found in supercapacitors. Carbon materials have been used an overview truly covering the entire eld, we tried to focus on widely as electrodes in supercapacitors, as they have been the most central steps in the history of N-doped carbons. reviewed frequently.26–30 It has been shown explicitly by Frack- While these N-doped carbons have become a well-established owiak et al. that nitrogen atoms induce favourable pseudoca- approach to face challenges in energy related applications, pacities, especially relying on the protonation of pyridinic interest has also arisen concerning the doping of carbons with nitrogen atoms at graphitic edges.31 This inuence of nitrogen other heteroatoms. The chemical preconditions of the other was also correlated with the porosity of the respective mate- heteroatoms might not be as perfectly tting for incorporation rials.32 It thus cannot be questioned that nitrogen-doping is one into a carbonaceous backbone, as is the case for nitrogen. of the central keys in the design of supercapacitor electrode Nonetheless the electronic situation in boron or the size of materials. second row elements like sulphur and phosphorus and thus induced structural effects represent a pool of possibilities to tune carbon materials into even more versatile directions. Synthetic routes towards N-doped carbons In the following chapters we will therefore give an overview The synthetic pathways leading to such N-doped carbons are of highly advanced carbon materials, doped with boron, nevertheless incredibly manifold, and are thus not limited to a sulphur or phosphorus. Therein we will highlight the rst single standard procedure. Classically, post-treatment of goals that could be reached in energy applications of such crude carbons with reactive nitrogen sources, such as urea, materials that are however still at the very beginning of their nitric acid or especially ammonia, is one way of obtaining successful development. N-doped carbons.33–37 Another dominant approach is the pyrolysis or chemical vapour deposition of nitrogen and Boron carbon containing precursors, such as heterocycles, melamine or aminated sugars, by which a direct incorporation of the Boron is an element with unique and basically incomparable nitrogen atoms into the forming carbon backbone becomes properties within the periodic table. It is thus a highly inter- possible.38–40 Actually the examples of nitrogen-doped carbons esting candidate for the doping of carbon materials, modifying are nearly uncountable, also due to numerous studies on the properties of pure carbons as will be discussed in detail nitrogen-doped graphene, which have been reviewed by Wang throughout this article. Early works on boron doped carbons et al. in 2012 (ref. 41), and on nitrogen-doped nanotubular were inspired by the fascination of stoichiometric boron nitride structures, about which overviews can also be found else- compounds that can form hexagonal patterns enabling sp2- where.42,43 A very recent overview of nitrogen-doped porous carbon-related structures, such as stacked sheets or nanotubes, carbons has just been presented by Shen and Fan.44 Never- as has been reviewed e.g. by Rhenzi et al. or Goldberg et al.68,69 theless we would like to introduce some more unconventional Early works on B/N/C-materials with graphite like structure and routes towards N-doped carbon materials that do not follow a composition of B0.35C0.3N0.35 have been reported by Bartlett the classical pathways. One way is hydrothermal carbon- et al. in 1987, who used a chemical vapour deposition approach isation that has meanwhile become a well-established proce- with boron trichloride, acetylene and ammonia as the precursor dure for deriving carbonaceous materials from carbohydrate mixture.70 Instead of chemical vapour deposition, Ajayan et al. rich biomass.45–47 Using nitrogen-containing biomass related applied an electric arc discharge base synthetic procedure, in precursors and treating them hydrothermally yield nitrogen- which elemental amorphous boron and elemental graphite containing carbonaceous materials that offer different possi- powder served as precursors, yielding a mixture of different bilities for further treatments and applications.48–54 Another boron containing carbon nanostructures, such as thin graphitic approach is based on a study on the thermal stability of ionic sheets, tubes and laments.71 A similar pathway was followed This journal is ª The Royal Society of Chemistry 2013 Energy Environ. Sci., 2013, 6, 2839–2855 | 2841  View Article Online Energy & Environmental Science Review for B/N/C-nanotubular structures by Redlich et al.72 Over the carbon monoxide poisoning, for which a favoured interaction of years many boron-doped carbons and – due to the fundamental the CO with the boron sites on the host compared to the elec- proximity to the aforementioned stoichiometric boron nitride trocatalytically active particles is considered as responsible.82 materials – also boron- and nitrogen-codoped carbons have Of course boron-doped carbon materials are not only suit- been synthesised, some examples of which can be found in the able as hosts for rather classical noble metal catalysts, but also reference list, reaching from chemical vapour deposition and serve as fuel cell electrocatalysts themselves, as has been arc discharge approaches over pyrolytic procedures towards the demonstrated by Hu et al. whose Boron doped carbon nano- carbonisation of ionic liquid based precursors.73–80 Meanwhile tubes were directly tested for their ORR activity. Upon research groups have started focusing on not only fundamental increasing the boron content, also an increase of the ORR studies on boron-doping, but also on applying the obtained activity has been observed, indicating the importance of the Published on 01 August 2013. Downloaded by University of Newcastle on 08/10/2017 16:22:50. materials and exploiting their benecial properties in energy boron moieties in the catalytic process. Also excellent ORR related applications. In the following, we would like to briey selectivity and resistance against methanol crossover qualify the introduce the readers to boron-doped or boron- and nitrogen- B-doped nanotubes as promising ORR candidates. Additional codoped carbon materials used as fuel cell catalysts, as elec- theoretical calculations performed by the authors have closely trodes in supercapacitors and as materials for lithium interca- elucidated the reason for the benecial inuence of the boron lation in batteries. doping, which is based on two synergistic effects: on the one hand boron has a lower electronegativity than carbon, and the Boron-doping for fuel cell applications positively polarised boron atoms attract the negatively polarised When it comes to fuel cells, some studies have focused on the oxygen atoms leading to chemisorption. On the other hand, use of merely boron-doped carbon electrocatalysts. E.g. Sun boron sites can also act as electron donors for the reduction et al. presented boron-doped carbon nanorods derived by a reaction, as the electron density of the graphitic p-electron spray pyrolysis chemical vapour deposition process. Yet the system can be transferred to the free pz orbital of the boron.83 authors did not directly use these nanorods as ORR catalysts, This electronic tuning adds a general promising potential to but as a support for electrochemically active platinum nano- B-doped carbons as ORR catalysts, as could be further shown by particles. The boron doping succeeded in signicantly Xia et al. who annealed graphene oxide in the presence of B2O3 enhancing the durability of this noble metal based ORR cata- yielding B-doped graphene. This material has proven to exhibit lyst, in comparison to crude carbon hosts, which is attributed to remarkable ORR activity with long-term stability of the catalyst a direct Pt–B-interaction and the high stability of the nanorods under alkaline conditions.84 Also other ideas on the inuence of themselves: only 13.2% of the electrochemical surface area of boron-doping have been discussed, as those recently by Nabae the boron-free catalyst is retained aer 3000 electrochemical et al. Their resin-based B-doped carbons have shown improved cycles, while boron doping increases this value to 47.7%.81 A ORR activity compared to the non-doped equivalent material, similar idea has been adopted recently by Manthiram et al. who which is attributed to the enriched presence of oxygen species distributed platinum nanoparticles as catalysts for methanol due to the reactivity of their boron source. The reasons have oxidation in the anodic compartment of fuel cells on boron hence not yet been fully illuminated, and the concept seems to doped carbon nanotubes derived from a chemical vapour be limited to the very specic example of the type of material deposition process with toluene, ferrocene and triethyl borate presented in their study.85 as the precursor–catalyst-system. In comparison to non-doped reference nanotubes, boron functionalised tubes allowed for a Dual doping with boron and nitrogen for ORR more homogeneous distribution of the Pt particles with fewer tendencies towards agglomeration, which is illustrated in Fig. 1. While hitherto only merely B-doped carbons have been dis- Thus also the electrochemical surface area has been enhanced. cussed, also B- and N-codoped carbons have been developed Besides this effect the authors have most noteworthily observed and applied for ORR, exploiting synergistic effects of the a signicant increase of stability of the Pt catalyst against benecial inuences of the two contrarily proled hetero- atoms. Carbonisation of polymer particles with melamine and boron triuoride as reactive agents was e.g. used by Ozaki et al. to obtain B–N-codoped carbon with interesting ORR activity boosted in comparison to only nitrogen doped carbons. The activity does however not reach values of conventional Pt@C catalysts. The authors suggest that this effect is not only due to the addition of both doping effects, but also due to a crucial role of B–N–C-type moieties throughout the catalyst.86 The authors intensied their study by performing directly compara- tive measurements of purely N-doped, purely B-doped and B–N-codoped carbons, all derived from the carbonisation of Fig. 1 Scanning transmission electron micrographs of Pt particles loaded on polymerised furfuryl alcohol under the inuence of the respective non-doped nanotubes (a) and B-doped nanotubes (b). [Reprinted with permis- reactive additives. Once more the pronounced synergistic effect sion from ref. 82; copyright ª 2012 Royal Society of Chemistry.] of the codoping on the B–N–C-type moieties was pointed out.87 2842 | Energy Environ. Sci., 2013, 6, 2839–2855 This journal is ª The Royal Society of Chemistry 2013  View Article Online Review Energy & Environmental Science Additional computational studies have implied that indeed certain atoms in the resulting material. The electrochemical measure- B–N–C-sites may cause enhanced ORR activity; yet the activity is ments of the ORR activity of the as-synthesised materials offer a highly sensitive to the actual arrangement of these moieties within profound insight into the mechanistic details; measurements the carbon backbone, e.g. boron and nitrogen atoms directly were performed with varying doping degrees of boron, in bound to each other do not seem to represent a favourable binding comparison to purely N-doped analogous materials. The exclu- concept.88 Woo et al. who used a chemical vapour deposition sively N-doped materials have shown – as expected – enhanced approach for the synthesis of B–N-codoped carbon using a ORR activities compared to pristine CNTs. Upon a stepwise precursor and catalyst system composed of dicyandiamide, boric increase of the boron content, a decrease of the maximum acid and iron or cobalt chloride, also observed a benecial effect of reduction current has been observed in the case of the doping the codoping on the ORR activity of their materials.89 Benecial concept with neighbouring boron and nitrogen atoms. Published on 01 August 2013. Downloaded by University of Newcastle on 08/10/2017 16:22:50. effects on ORR activity of B–N-codoping in carbons have also been Contrarily, an enhancement of the current has been observed in observed by Chisaka et al., nevertheless depending on the presence the case of boron and nitrogen atoms incorporated into the of iron species.90 carbon isolated from each other. These data explain that the B–N-codoping for ORR has furthermore been applied benecial effects of both boron and nitrogen on the ORR recently for two highly advanced carbon allotropes: graphene activity of carbons sensitively depend on their chemical envi- and carbon nanotubes. Regarding the latter, Dai et al. have used ronment. Due to the opposite properties of both dopants, their melamine diborate as a single precursor for vertically aligned effects can be mutually compensated, while most fruitful carbon nanotubes doubly doped with boron and nitrogen, synergies can be achieved when thoughtfully designing the nally synthesised by a chemical vapour deposition approach doping procedures.92 This is in good agreement with afore- with metal catalysts, while metal residues were eliminated from mentioned calculations.88 The effects have also been similarly the nal product by acid treatment. The aligned nanotubes (see observed in studies on B–N-codoped graphene: Dai et al. Fig. 2 for electron micrographs) have been successfully tested as have published their approach that allows for the synthesis of electrocatalysts in ORR, in which the B–N-codoped systems B–N-codoped graphenes by thermally treating graphene oxide clearly exceeded both mono-doped nanotubes used as a refer- with boric acid in an atmosphere of ammonia. The authors ence system; the current densities of the dually doped tubes get have thereby obtained BCN-materials with stoichiometries of close to those of conventional Pt@C electrodes. Given the fact B38C28N34, B7C87N6 and B12C77N11 and outstanding thermal that XPS measurements show that no metal residues are present stabilities. Already XPS measurements have indicated the easy in the catalyst, this observation clearly underlines the syner- adsorption of oxygen on the graphene surface, pointing out a gistic effect of both dopants.91 The results of Hu et al. further high potential of the materials as ORR catalysts. B38C28N34 allow for an even more detailed elucidation of the B–N-codop- nevertheless could not be shown to be a useful electrocatalyst, ing synergistic effects. The authors have therefore prepared two probably due to a lack of conductivity, attributed to the low types of codoped carbon nanotubes. One approach was per- carbon content, and a mutual compensation of the effects of B formed by chemical vapour deposition under direct incorpora- and N atoms directly bound to each other – in agreement with tion of boron and subsequent post-treatment of the tubes under the previously discussed aspects of the binding environment of ammonia to additionally introduce nitrogen atoms, preserving B and N in doped nanotubes. Hence the accordant electro- the antecedently established binding sites of boron. Thus B and chemical measurements for the other candidates could indeed N are separate and not directly bound to each other in this carbon show the outstanding behaviour of the B/N-graphenes as ORR material. In contrast, the authors used a chemical vapour depo- catalysts under alkaline conditions, while B12C77N11 even sition procedure using precursors allowing for a simultaneous exceeds the current density of a conventional Pt@C catalyst and direct inclusion of both heteroatoms into the forming throughout the major part of the scanned potential range. Also, carbon nanotube structures, yielding neighbouring N and B good electrochemical stability, high ORR selectivity and an electron transfer number of almost 4 (see Fig. 3) complete the Fig. 2 (a) Scanning electron micrograph and (b) transmission electron micro- Fig. 3 RRDE testing on a BCN graphene (B12C77N11) sample in an oxygen- graph of boron/nitrogen-doped vertically aligned carbon nanotubes. [Reprinted saturated 0.1 M KOH solution (a) and the corresponding electron transfer number with permission from ref. 91; copyright ª 2011 Wiley-VCH Verlag GmbH & Co. of ORR on the BCN graphene (b). [Reprinted with permission from ref. 93; KGaA, Weinheim.] copyright ª 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.] This journal is ª The Royal Society of Chemistry 2013 Energy Environ. Sci., 2013, 6, 2839–2855 | 2843  View Article Online Energy & Environmental Science Review picture of a remarkable ORR catalyst, based on heteroatom triuoride and subsequent freeze-drying. The morphology of the doping of carbon. The authors have further supported their interconnected doped graphene sheets giving monolithic struc- experimental data by calculations, showing that in the case of tures allows for a targeted shaping of the material to t the B12C77N11 the values of maximum spin densities and maximum requirements of the supercapacitor device. Electrochemical charge densities are higher than those for the other tested measurements performed in direct comparison to N-doped, graphenes. This is thus considered as one crucial factor for B-doped and pristine derivatives of the material have pointed out ORR activity enhancement by B–N-codoping.93 In a similar that signicant synergies can be achieved by dual doping, as the approach, with varied N- and B-sources during thermal treatment highest specic capacitances of up to >60 F g 1 could be observed of graphene oxide, Woo et al. have presented a work supporting for the dually doped samples. Nevertheless, also pure B-doping is the previous results discussed here. Also their B–N-codoped shown to exhibit a remarkable supercapacitor performance, Published on 01 August 2013. Downloaded by University of Newcastle on 08/10/2017 16:22:50. graphene has proven itself as a good ORR catalyst, even under getting close to the results of the codoped sample.97 This result acidic conditions.94 further proves that not only is codoping with boron and nitrogen a It can be summarised that boron-doped carbons represent a promising concept for the design of novel supercapacitor elec- promising class of ORR catalyst candidates. The results dis- trodes, but also the mere doping with boron can pave the way cussed throughout this article speak for themselves, giving towards interesting electrode materials, as Cheng et al. have numerous examples of advanced ORR catalysts based on this shown using an ordered mesoporous carbon derived from a hard concept, using different ways to achieve synergies of the two templating approach with sucrose as the carbon source and SBA- contrary doping heteroatoms boron and nitrogen. 15 as the template. Already minimal boron doping in amounts of 0.2 atom% – achieved by boric acid used as an additive during synthesis – was capable of boosting the specic capacitance B-doped and B–N-codoped carbons for supercapacitors and strongly.98 Furthermore Park et al. could recently point out similar lithium ion batteries results in their study on boron-doped graphene nanoplatelets. The Due to their interesting property proles, such materials have authors have derived their B-doped graphene from a solution heretofore also been applied in other energy related elds, e.g. process, simply giving dispersed graphene oxide the possibility to as electrodes in supercapacitors. A main focus has also been on react with a borane–tetrahydrofurane adduct (as reducing agent boron- and nitrogen-codoped carbon materials, as presented and dopant) and subsequently drying the sample in a vacuum e.g. by Gao et al. The authors have followed a relatively uncon- under gentle conditions. This mild procedure yielded materials ventional synthetic pathway using citric acid as the carbon with 1 atom% of boron incorporated into the graphene platelets. source, boric acid as the boron source and claim to have suc- They thereby reached a conclusion that specic capacitance can ceeded in the incorporation of nitrogen into their structures by reach values of >200 F g 1, which can be almost fully retained aer simply executing the carbonisation reaction in a nitrogen 4500 electrochemical cycles.99 atmosphere that is – despite its estimated role in the reaction – As a last example of energy related applications of boron- referred to as an inert gas atmosphere. Application of the doped carbon materials, we would like to provide a short derived hierarchically porous materials as supercapacitor elec- overview of lithium ion batteries. Nevertheless, in comparison trodes has been reported as successful, indicating that the to the aforementioned application elds, studies of lithium ion favourable pseudocapacitance effects are not only achievable by batteries using B-doped carbons as electrodes are still relatively nitrogen-, but also by boron-doping.95 As mentioned before, Dai rare. An early overview has been provided in 2000 by Endo et al., et al. have synthesised B/C/N-nanotubes from chemical vapour pointing out difficulties that have been faced with B-doped deposition using melamine diborate as a single precursor. The carbons as lithium battery anodes, as in many cases the authors have – additionally to the fuel cell application tests – formation of boron carbide or nitride sites in carbonaceous focused on the applicability of the material in supercapacitors, materials has prevented benecial effects of real boron too. The rst interesting observation is a signicant enhancement doping.100 Later some examples have shown that boron-doping of the specic capacitance of the nanotubes, from 83.8 F g 1 for obtained by graphitisation of pitches or fossil coals with boron pristine nanotubes to 162.5 F g 1 for the doped tubes (in acidic sources at temperatures >2000  C can yield potential anode media). This is attributed to induced pseudocapacitance contri- materials for lithium ion batteries, still partially encountering butions, and also a thickened electrochemical double layer due to the aforementioned problem.101–104 A recent breakthrough for an improved wettability of the doped structures. A second nding B-doped carbons could be presented in a study by Cheng et al. is that not only the doping itself, but also the nanomorphology The authors have focused on graphene to be doped with boron of the electrode material inuences the capacitance drasti- for enhanced performance as an anode material in lithium ion cally. Using vertically aligned B/C/N-nanotubes a specic batteries. Thus graphene is post-functionalised at 800  C using capacitance of 312.0 F g 1 is reached, which is almost doubled boron trichloride as a doping agent, yielding a doping level of compared to the non-aligned tubes.96 A new step in the 0.88 atom% in the material. High electrode capacities of 1549 ongoing process of B/N-doped carbon materials for super- mA h g 1 at low charge/discharge rates and 235 mA h g 1 at very capacitors has further been reached by Müllen et al. who pre- high rates represent a remarkable result, attributed to sented a study on using a B–N-codoped graphene aerogel in all- numerous effects of the boron doping regarding electrode/ solid-state supercapacitors. The graphene aerogels were obtained electrolyte wetting properties, interlayer distances, electrical by hydrothermally treating graphene oxide with ammonia boron conductivities or heteroatom induced defect sites, thus leading 2844 | Energy Environ. Sci., 2013, 6, 2839–2855 This journal is ª The Royal Society of Chemistry 2013  View Article Online Review Energy & Environmental Science to enhanced performance of the material concerning its Li+ in the eld of energy as hydrogen storage media. The hydrogen absorption and diffusion properties.105 It can be followed that uptake in these S-doped carbons can reach 5.71 wt% at 196  C boron doping is indeed a promising concept for the design and under a pressure of 20 bar and 2.41 wt% at a pressure of 1 bar. development of new anode materials for lithium ion batteries, The calculated hydrogen uptake density of this specic sample and that there certainly are ways to overcome the previously was 9.5 mmol m 2. The materials with surface areas lower than encountered problems of the formation of unfavourable boron 3000 m2 g 1 can reach up to 12.7 mmol m 2. Although a sulphur- binding situations. Boron is thus already a relatively profoundly doped carbon type material has been used in this study, the studied dopant for carbon materials that has entered numerous authors tend to assign the hydrogen uptake behaviour to the energy related applications and will play a signicant role in the porous structure and the high active surface area, but do not see a further ongoing process within this eld. clear benecial inuence provided specically by the sulphur Published on 01 August 2013. Downloaded by University of Newcastle on 08/10/2017 16:22:50. atoms within the material.123 This has been discussed in different manners by Xia et al. Their approach included a template based Sulphur synthesis of microporous sulphur-doped carbon using 2-thio- In comparison to boron, sulphur doping in carbon materials is phenemethanol as a polymerisable precursor and a zeolite as a hitherto still quite rare and represents an emerging eld within hard template. The as synthesised S-doped carbons exhibited one carbon material research. While nowadays the high potential of of the highest hydrogen uptake densities reported for nanoporous such materials for energy applications in fuel cells, super- carbons at this point of time of 14.3 mmol m 2. According to the capacitors or batteries is continuously discovered and exploited, authors, this is due to the sulphur atoms that induce a stronger until only a few years ago, only little had been known about such interaction between the carbon host and the hydrogen molecules, sulphur-doped carbonaceous species. The knowledge mainly which has been derived from measurements of the isosteric heat ranged from graphite–sulphur-composite materials and their of the hydrogen adsorption that is – in the case of the template superconductive behaviour106,107 towards theoretical studies about derived S-doped carbon – up to 8.9 kJ mol 1 signicantly higher the effects of single sulphur atoms in carbon nanotubes or gra- than those for comparable N-doped or non-doped carbons and phene sheets.108–112 Sulphur–carbon composites have also entered also those for metal organic frameworks.125 Thus the real inu- the eld of lithium–sulphur batteries.113–118 Nonetheless, synthetic ence of the sulphur atoms is still controversially discussed; concepts establishing doped carbons with sulphur atoms rmly yet their promising potential as a dopant is obvious and will and covalently incorporated into the carbon structures have only be subject of future studies. Other works of S-doped activated been developed since 2011, when Schmidt et al. used a micro- carbon – with S-doping in the ppm range – have been per- porous polymer network containing thienyl building blocks as a formed by Bandosz et al., focusing on other types of applica- precursor for intrinsically microporous S-doped carbon with tions as adsorbent materials126,127 or photoactive substances variable sulphur contents of 7 wt% up to 20 wt% depending for light harvesting.128 on the carbonisation temperature.119–121 Comparable to this concept, Spange et al. developed an approach linking thienyl S-doped graphene based materials for ORR monomers covalently to silica precursors. This dually function- alised precursor system allows for the synthesis of silica/S-doped Another carbon allotrope that is currently the subject of carbon composites yielding additional porosity aer silica different studies in the eld of sulphur doping is graphene. The removal. The authors further induced mesoporosity within this general feasibility of S-doping in graphene has been predicted type of material applying hard templates.122 theoretically by Denis et al. S-doping should be more difficult than doping with e.g. nitrogen, taking into account the size and the different binding behaviour of sulphur atoms. Nevertheless S-doped activated carbons for hydrogen storage it was calculated that S-doping should allow for a targeted Fuertes and Sevilla et al. acknowledged the potential of the thienyl tuning of the graphene bandgap, depending on the amount of group as a central tool in the synthesis of S-doped carbon, and sulphur atoms incorporated into the sheets.110,129 Practically it is polymerised thiophene using an oxidative approach, followed by e.g. possible to synthesise S-doped graphene by a chemical carbonisation under activation of the forming carbon with vapour deposition approach using a solution of elemental potassium hydroxide. Thus a rather classical approach for the sulphur in hexane as the precursor.130 As can be derived from preparation of activated carbons has been successfully transferred data published by Yang et al., another way to obtain sulphur- to the design of modern materials. The as derived S-doped doped graphene is to blend graphene oxide homogeneously carbons exhibit a microspherical morphology and remarkable with benzyl disulde, followed by subsequent annealing microporosities with BET surface areas of up to 3000 m2 g 1 and (compare the scheme in Fig. 4). micropore surface areas of up to 2600 m2 g 1, depending on the The result is an interesting material, composed of graphene polythiophene–KOH-ratio and the carbonisation temperature sheets exhibiting partially wrinkled and folded morphological applied. The sulphur atoms are rmly bound to the carbon motifs. The sulphur with contents of up to 1.5 wt% is homo- backbone, mainly in C–S–C binding motifs, in amounts from 4 geneously distributed throughout these sheets, also at their wt% up to 14 wt% – again depending on the reaction temper- edges, according to the elemental mapping data provided. The ature.123,124 Due to the high surface areas and pore volumes of chemical environment of the sulphur atoms could be further their materials, the author initiated rst tests on their application elucidated by XPS, indicating covalent bonding between carbon This journal is ª The Royal Society of Chemistry 2013 Energy Environ. Sci., 2013, 6, 2839–2855 | 2845  View Article Online Energy & Environmental Science Review Published on 01 August 2013. Downloaded by University of Newcastle on 08/10/2017 16:22:50. Fig. 4 Schematic representation of a S-doped graphene synthetic approach using graphene oxide and benzyl disulfide as precursors. Application as an ORR catalyst is also indicated. [Reprinted with permission from ref. 131; copyright ª 2012 American Chemical Society.] and sulphur and thus the usefulness of this method for the graphene is one part of a composite, while the other part is a synthesis of S-doped graphene. The authors have further sulphur-doped porous carbon. This composite was obtained by reported the good performance of their S-doped graphene for ionothermal condensation of glucose in a sulphate containing ORR catalysis under alkaline conditions. Therefore, different ionic liquid with the addition of graphene oxide. The thus doped graphenes obtained at different reaction temperatures in obtained gel was carbonised in an atmosphere of an inert gas at the synthetic process were compared to their non-doped 800  C to yield the S-doped porous carbon–graphene composite. equivalents, synthesised without the addition of benzyl disul- With a surface area >900 m2 g 1 and a bimodal pore size de. In all cases, the ORR activity was signicantly enhanced by distribution – with a hierarchical structure exhibiting micro- S-doping. The electrocatalytic activity of S-doped graphene and mesopores (more details about the nanomorphology of the synthesised at 1050  C even exceeds the activity of a conven- composites are available in Fig. 5) – the material exhibits ideal tional Pt@C catalyst, while additionally exhibiting higher preconditions for its application in lithium ion batteries. Accor- selectivity for oxygen reduction and thus avoiding well crossover dant experiments have revealed good performance of this mate- effects. Koutecky–Levich-plots further revealed that an almost rial, exhibiting a stable reversible capacity of 1400 mA h g 1, a ideal four electron transfer occurred in the S-doped graphene long life cycle and remarkable rate performance. The authors catalysts, once more accentuating the powerful character of the attribute this behaviour at least partially to the sulphur doping S-doping method within the eld of fuel cell catalyst research. that is supposed to increase interlayer distances in the stacking The reason is suggested to be found in the increased spin of graphene sheets and to increase the amount of nanopore density in the graphene achieved by sulphur doping.131 Müllen moieties. Nevertheless, in this complex composite material, of et al. applied a similar approach in their S-doped graphene course sulphur doping is not the only factor that plays a crucial synthesis, using graphene oxide as the starting material. In a rst step, ultrathin silica–graphene oxide composite sheets were produced. In a second step, annealing in an atmosphere of H2S allowed for the formation of S-doped graphene that was isolated by subsequent removal of the silica. The tests on the materials' ORR activities were successful; like Yang et al. the authors observed good electrocatalytic activities, electron transfer numbers close to 4 and – especially in comparison to conventional Pt@C catalysts – good resistance against crossover effects, tested by the addition of methanol. Thus S-doped gra- phene also in this study is accentuated as a promising candi- date for electrocatalysis; nevertheless the authors discuss a bit more considerately that – in comparison to nitrogen doping – a homogeneous doping with sulphur atoms is still difficult; sulphur atoms – due to their different