9.1 Introduction

An ideal catalyst enables a reaction to proceed at low temperature and/or pressure, thereby consuming less energy. It speeds up (or on purpose slows down) a reaction and orientates the reaction pathway toward desired products, minimizing undesirable by‐products. Technologies involving chemical catalysis enable efficient and cost‐effective synthesis of chemicals and materials. Not only is the production efficiency remarkably increased but also the associated separation and purification procedures can be eliminated. Currently, up to 90% of all commercially available chemical products involve using catalysts at certain production stage [1], which unambiguously demonstrates the pivotal role of catalysis in various industries and the world economy. However, currently, most of the catalytic reactions on industrial scales are achieved by means of transition‐metal‐based catalysis. Many of them utilize expensive metals, suffer from limited natural abundance of transition metals, and generate toxic waste, presenting an enormous sustainability and environmental challenge.

Carbon is one of the most abundant and cheapest elements on Earth. Carbon nanomaterials are particularly interesting due to their physicochemical and mechanical properties, such as large surface area, outstanding electron conductivity, corrosion resistance, and thermal stability. Due to these wonderful properties, they have been widely used as excellent catalytic supports for metal‐based catalysts [2, 3]. It has been found that they can act as catalysts themselves instead of just acting as inert catalyst supports. Furthermore, their physicochemical and electronic properties, which in principle determine the catalytic properties of a material, can be tailored and fine‐tuned by molecular and atomic doping, which make them potentially attractive to replace some of the expensive and/or toxic transition‐metal‐based catalysts. Using one of the family members of carbon nanomaterials, graphene, as an example, defect‐free pristine graphene may not be a good catalyst. However, the introduction of heteroatoms into its carbon matrix produces distinct electronic and surface properties, which can be further fine‐tuned by different heteroatoms with various binding configurations in a graphene matrix. Synergistic actions of co‐dopants with an optimal ratio can simultaneously enhance charge polarization, spin density, and conductivity. All of these properties make graphene materials an attractive metal‐free catalyst for a wide range of chemical reactions. Wang et al. gave an excellent review article about heteroatom‐doped graphene materials [4]. The summary covered fabrication methods and properties. The applications of heteroatom‐doped graphene in electrochemical applications, such as supercapacitors, lithium ion batteries, fuel cells, solar cells, sensors, and gas storage, were also summarized. Numerous studies have reported and demonstrated that doping of heteroatoms into graphene matrix can give rise to an enhanced performance in electrocatalytic oxygen reduction reaction (ORR), oxygen evolution reaction (OER), and even hydrogen evolution reaction (HER) in terms of both activity and reaction kinetics, when compared with their undoped analogs. In addition, multiple dopants‐incorporated nanocarbons showed further improvement in electrocatalytic performances. In parallel, in an effort to elucidate the promoting effects of dopants in the ORR and to further increase the catalytic efficiency to replace precious metal catalysts, a large amount of theoretical calculations have been performed to provide understanding at microscopic atomic level of the catalytic origins and the catalytic sites. Recent reviews by Dai and coworkers provided a critical overview of this field with special emphasis on molecular design of efficient carbon‐based, metal‐free catalysts for clean energy conversion and storage and environmental protection [5, 6].

Compared to electrocatalytic studies, studies using doped and co‐doped carbon materials as metal‐free carbocatalysts for selective chemical catalytic reactions are in an early stage; however, great potential has already been demonstrated [7]. Several recent review articles timely summarized the developments in this research field [8–13]. However, it is clear that the majority of carbocatalysts are still a long way from practical industrial applications mainly due to the relative low efficiency and limited recyclability. This has been ascribed to the low density of catalytic active sites and/or the existence of “bad” sites. It is logical to think that if we can selectively remove the inactive sites, or even transform them to beneficial sites, and/or create additional beneficial active sites to have synergistic effects, we should be able to increase their catalytic efficiency. It has been widely accepted that the structure of carbon materials can be complicated and versatile. It has been challenging to precisely identify the “bad” and “good” sites. The recent reviews by the Kong group and the Garcia group summarized the active sites and catalytic mechanisms with an aim of providing possible guidelines to further design and improve graphene‐based catalysts for practical applications [1, 14].

Different allotropes of carbon materials share a lot of similarities in properties and, therefore, catalytic activities, even though differences do exist due to their different shape and/or different C hybridization (sp² or sp³). The recent review by Su et al. gave clear definitions and basic descriptions of carbon materials, which have been used as catalysts [11]. The structures of various carbon materials and their corresponding properties have been widely studies. The advantages of various carbon allotropes in developing advanced catalysts were also discussed [8]. To avoid repetition, this review will only give a brief introduction of the structure and properties of the carbocatalyst. The main focus will be to summarize several fundamental and industrially relevant catalytic reactions with different allotropes of carbon materials as metal‐free catalysts. The special emphasis is placed on catalytic mechanisms, identification of catalytic centers and detrimental sites or functional groups on carbocatalysts. Possible strategies to block or transform these sites to beneficial ones are included. Furthermore, the latest advances in the application of graphene‐derived materials as carbocatalysts in carbon–carbon bond coupling reactions are specifically reviewed to stimulate further research of carbocatalysts in more advanced organic synthesis. In the last section, perspective and suggestions are given for future research to transform the great potential of carbocatalysis to industrial domain.

9.2 Dehydrogenation

In 1960s, it was observed that iron oxide catalysts with carbon deposited (coking) on their surface were still active in oxidative dehydrogenation (ODH) of ethyl benzene [15]. It was realized for the first time that carbon itself might also be catalytically active in ODH. Due to the rapid development of petrochemical industry during the Second World War, dehydrogenation of hydrocarbons to produce valued‐added olefins had aroused great practical interest and, thus, this discovery draws intensive attention. However, the amorphous or disordered carbon materials involved in earlier studies showed low activity, low stability, and low oxidation resistance [8, 16–20]. In the past three decades, the discovery and the rapid development of nanocarbons with well‐ordered structures (fullerene, carbon nanotubes (CNTs), and graphene) have dramatically improved carbocatalysis in dehydrogenation performance. These materials exhibited much higher activity and stability as compared to classical carbon materials (carbon black, activated carbon (AC), or graphite). Further, compared to traditional metals or metal oxides, the well‐structured carbons have many advantages such as being highly resistant to acid and alkali, having large surface area, and exhibiting high electronic and/or thermal conductivity.

In fact, dehydrogenation of hydrocarbons, especially dehydrogenation of ethylbenzene, and light alkanes to styrene and their corresponding olefins are almost the first reactions to take advantage of these unique properties of nanocarbons [21, 22]. These reactions are also the most well studied and evaluated. Several reviews about this work have been published, in which the active‐site structure, the kinetic studies, and the tailoring of catalysts were systematically discussed [8, 17, 23, 24]. In general, there are two dehydrogenation processes: ODH and direct dehydrogenation (DDH). It is interesting that carbonyl groups in ketones and quinones were identified to be the catalytic centers for both ODH and DDH processes. However, the paths to regenerate these catalytic centers to finish the catalytic cycles are different [24]. The possible molecular mechanisms for ODH and DDH are illustrated in Figure 9.1. Diffusion and chemisorption of hydrocarbons take place on the ketonic CO groups of the carbocatalysts. It is possibly that two hydrogen atoms on the closely connected carbons of the reactant are first adsorbed onto the adjacent ketonic CO groups of the carbocatalyst, which abstract H from the reactant, leading to the generation of olefin product and the formation of hydroxyl groups on carbocatalyst surfaces. In ODH, reactivation of the catalyst is through oxidation of the hydroxyl groups back to ketonic groups via molecular oxygen. In contrast, in DDH, the hydroxyl groups thermally decompose back to ketone groups, which is a thermodynamically favorable process at high temperatures [25].

Due to the distinct pathways to regenerate the active sites, the “bad groups and sites” are different for these two processes. Accordingly, the strategies to further improve the catalytic performance are divergent. For ODH, except for the ketonic groups, almost all the other oxygen‐containing groups, such as carboxylic and anhydride groups, have negative effects. These groups are too active so that over‐oxidation occurs, resulting in low selectivity of the ODH process. The geometrical defects and edges existing on the carbocatalysts are also possibly detrimental to the selective production of alkenes. This is because molecular oxygen is required for regeneration of the catalytic centers; it is fed into the reaction system with the hydrocarbon reactant together. Molecular oxygen not only regenerates ketonic groups by oxidation of the hydroxyl groups as required to close the catalytic reaction cycle but it also unintentionally leads to the generation of electrophilic oxygen‐containing groups, such as O⁻ (oxide), O₂⁻ (superoxide), and O₂²⁻ (peroxide). Similar to the carboxylic and anhydride groups that originally exist on the carbocatalysts, these in‐situ‐generated groups were also reported to be responsible for oxidation and also catalyst combustion [26]. They are electron deficient and attack electron‐rich molecules, especially of the produced alkene, which has an electron‐rich π bond. These electrophilic oxygen species also induce combustion of the carbocatalysts, resulting in a decreased the recycleability of the catalyst. These side reactions happen because a number of structural defects and edges always exist in nanocarbon‐based catalysts, even in pristine CNTs produced via chemical vapor deposition. At reaction conditions (>600 °C), these defect sites or edges of graphene are efficient to activate molecular O₂ to generate these electrophilic oxygen species. It was reported that liquid‐phase oxidation, for example, fluxing CNTs in HNO₃ or H₂O₂ solutions, can fabricate oxidized CNTs (O‐CNTs) with higher density of ketonic CO than that achieved in gas‐phase oxidation. However, these liquid‐phase oxidation approaches also generate other oxygen‐containing groups, such as carboxylic and anhydride groups. At the reaction temperatures, desorption of less‐stable groups results in new graphitic defects that subsequently generate new electrophilic oxygen sites, thereby partially limiting the selectivity to alkenes on O‐CNTs. Therefore, these unstable electrophilic groups and structural defects and edges are very likely the “bad” groups and sites for an ODH process. As majority of the problems is associated with the use of strong oxidant O₂, Rao et al. also demonstrated that a soft oxidant, such as CO₂, turns these geometrical defects to be beneficial to the ODH process [27]. Strategies were also developed to block these “bad” groups and sites to improve the selectivity and cycle lifetime of the carbocatalysts. Su and coworkers modified the carbocatalysts with borate and phosphate. The modified carbocatalysts show dramatically improved selectivity toward the desired olefin products compared to the pristine CNTs and O‐CNTs [28, 29]. A high selectivity to alkenes was achieved for periods as long as 100 h, which improved the selectivity by suppressing the combustion of hydrocarbons. The combustion rate of the carbocatalysts is also largely suppressed, which results in largely prolonged catalyst lifetime. In contrast, the conventional carbons, in particular activated carbon, underwent unavoidable deactivations due to coking or combustion. The oxidative stability of phosphate‐modified CNTs far exceeds those of metal‐oxide catalysts, which only work well with excess oxygen (O₂/butane ≥2), a condition necessary to prevent severe deactivation due to coke deposition. Further mechanistic studies demonstrated that the modification did not affect the nature of active sites on the carbocatalysts. The enhanced selectivity was achieved by suppressing the combustion rate, rather than enhancing the formation rate of alkenes.

Considering that one of the remaining drawbacks is the low overall activity of carbon catalysts, to further advance the field for ODH toward practical applications, efforts should be made to develop new carbocatalysts with higher concentration and higher activity of the catalytic sites (nucleophilic ketone groups) and, most importantly, with the capability to avoid over‐oxidation under the reaction conditions. It might be beneficial to have carbocatalysts that can in‐situ generate beneficial sites but suppress the formation of “bad” groups. In this sense, Liu et al. found that ODH of n‐butane could induce lattice rearrangement of nanodiamond from cubic sp³‐hybidized to sp²‐hybridized fullerene shells, which gave a carbon surface that was highly selective in the ODH reaction [30]. The strongly curved and strained graphitic surface that contains carbon atoms with a certain degree of sp³ hybridization appears to be an appropriate matrix for the selective generation of surface quinoidic groups and effectively suppresses the formation of electrophilic oxygen species such as carboxylic acids and their anhydrides. However, doping N into the carbon matrix, especially in graphitic‐N bond configuration, also demonstrates the capability in enhancing the ODH activities [26]. This was ascribed to the fact that the graphitic‐N atoms donate electrons to the graphitic sheet and improve the electron density and mobility, which may facilitate propene to desorb from the surface due to repulsion between the electron‐rich graphitic surface and unsaturated CC bonds of propene. It is possible that it also speeds up the activation of molecular oxygen for regeneration of the catalytic centers, so that the overall activation energy of the ODH reaction is decreased [31].

With density functional theory (DFT) calculations, Tang et al. predicted the active sites and catalytic mechanisms for ODH of propane on graphene oxide (GO) [32]. Owing to their high electron density, the epoxy groups on GO surface may be the active sites for CH bond activation. The first CH bond breaking of propane through the H abstraction by epoxide leads to the formation of a propyl radical, which is the rate‐determining step for the conversion from propane to propene. The presence of OH groups around the active site can remarkably improve the activity of the epoxy group and facilitate the H abstraction. Furthermore, the sites of oxygen functional groups on the GO surface can be easily tuned by the diffusion of these groups under an external electric field, which increases the reactivity of GOs toward ODH of propane. The chemically modified GOs are thus quite promising in the design of metal‐free catalysis. However, as it is well known that GO is not stable at higher temperatures (even as low as 70 °C), the oxygen‐containing groups will diffuse and/or desorb, so it is not clear if these predictions can be practically realized [33]. However, hexagonal boron nitride and boron nitride nanotubes, normally an unreactive material, were recently reported by Grant et al. to exhibit unique and unanticipated catalytic properties in ODH of propane with high selectivity for the production of propene (77%). For reasons that are still under investigation, the BN materials are able to largely avoid the over‐oxidation reaction pathway to generate CO or CO₂, while the only by‐product is ethene (13%) and desired alkene [34]. Based on the catalytic experiments, spectroscopic insights, and ab initio modeling, a mechanistic hypothesis in which oxygen‐terminated armchair boron nitride edges is proposed to be the catalytically active sites. The intermediates are possibly a peroxo‐like >B–O–ON< armchair edge.

In carbocatalyst‐mediated DDH processes, as molecular oxygen is not required to regenerate the catalytic centers, the issues associated with the utilization of molecular oxygen are naturally eliminated, including low selectivity due to over‐oxidation, damage of catalyst structure by oxygen, and the risk of handling an explosive mixture containing hydrocarbons and oxygen. Another positive factor is that those geometrical defects are not “the problematic” sites anymore. In fact, they became beneficially active sites in DDH. Employing DFT, Su et al. calculated the dissociation energy of the first CH bond breaking of propane on the defect sites and found this process to be thermodynamically feasible for the activation of propane under the reaction conditions.

The hypothesis that the geometrical defects can directly activate hydrocarbon molecules was also experimentally demonstrated by Wang et al. [35]. A diamond/graphene composite with mixed sp²/sp³ hybrid character was fabricated and studied as a metal‐free carbocatalyst for DDH of propane. As the surface oxygen groups are desorbed or consumed during the DHH process, they found that CH bond in propane can be activated by uncovered defect sites on the surface, which then result in hydrogen abstraction from propyl intermediates to form propene. Finally, hydrogen desorption from the surface completes the catalytic cycle to form H₂. The mechanism is summarized in Figure 9.2. It is possible that the DH process is a combination of these mechanisms as summarized in Figures 9.1 and 9.2 as the carbocatalysts may have various kinds of active sites. In this work, they also found that the reactivity of ketone groups on the restructured graphitic surface is much higher than that initially found on an amorphous surface, which means that the reactivity of ketone oxygen functionalities on the carbon surface can be tunable by the chemical nature of defect/vacancy sites with different structural characteristics.

It is noteworthy to point out that the low selectivity in DDH is due to cracking reactions instead of over‐oxidation. As an example, the main by‐products from DDH of ethylbenzene were benzene and toluene, resulting from the cracking of ethylbenzene. As acid sites are active for cracking of hydrocarbons [17], it is well accepted that the surface phenolic hydroxy group and/or possible COOH groups may promote the cracking of ethylbenzene. Incorporation of electron‐rich N into the carbon matrix via N doping can increase the electron density of carbon materials and, therefore, strengthen the basic properties but weaken the acidity of the catalyst, which results in an improvement in the catalytic activity for styrene production and simultaneously suppress benzene and toluene formation [13]. Furthermore, the N doping possibly also enhances the nucleophilicity of the ketonic CO groups due to the increased electron density of the carbocatalysts; thus, increased activity was observed. It was recently demonstrated that N doping of graphene largely enhances the catalytic efficiency of reduced graphene oxide (rGO) in dehydrogenation of ethanol to acetaldehyde [36]. With all these developments, the problem of low catalytic activity and low selectivity of current carbocatalysts for the DDH process must be solved before it can be realized. The possible strategies to improve the catalytic performance of carbocatalysts are to enrich structural defects, increase content and activity of the surface ketonic CO groups, and increase basicity of the carbocatalysts by N doping [37, 38].

It is worthwhile to point out that “on‐purpose” light olefin production has gained renewed interest and has the potential to be a game‐changing technology in the chemical industry [34, 39]. First of all, light alkenes, such as propylene (propene) and ethylene, are important chemical building blocks and their demand has increased steadily over the past years [39, 40]. Currently, the most common approaches to produce light alkenes are steam creaking and fluid catalytic cracking (FCC) of naphtha, light diesel, and other oil by‐products. However, the high energy demand and low product selectivity of these processes and, most importantly, the continuous decrease in petroleum reserves are driving the petrochemical industry to search for a more economical feedstock and more efficient conversion technologies. However, the recent boom in shale gas production has turned high availability of relatively cheap gas liquids, specifically ethane, and resulted in a shift from oil‐based naphtha to shale‐based ethane for ethene production. However, relative to naphtha cracking, steam cracking of ethane produces a negligible amount of propylene and butadiene other than ethylene, which has caused dramatic supply drop and sharp price increase of propylene and butadiene, creating opportunities for catalytic dehydrogenation of propane (PDH) and dehydrogenation of butane (BDH). In addition, as the dehydrogenation is an on‐purpose technique, which yields exclusively propylene or butadiene instead of a mixture of products, it also represents an economical and environment‐friendly route compared to the traditional thermal or catalytic cracking of crude‐oil‐derived naphtha, which requires large amounts of energy and generates enormous CO₂ emissions. CrO_x and Pt‐based catalysts are the two main catalysts applied in PDH and BDH on a commercial scale for decades [39, 40]. Although the catalytic performance of these catalysts is relatively satisfactory, the harmful impacts of Cr and the high cost of Pt have limited their applications. However, even after decades of research to establish supported vanadium oxide (V/SiO₂) as the state‐of‐the‐art precious metal‐free and toxic metal‐free catalyst for oxidative PDH, selectivity to the olefin product remains too low to be commercially attractive because of over‐oxidation of propylene into CO and CO₂ (CO_x) [39]. More efficient and selective and more green and ecofriendly catalysts are still eagerly awaited.

9.3 Oxidation Reactions

Compared to dehydrogenation reactions, using metal‐free carbocatalysts for oxidation reactions are very broad. In general sense, ODH is also an oxidation reaction, which mainly happens in gas phase. The desired products are light olefins. In this section, we focus on liquid‐phase oxidation of hydrocarbons to produce value‐added oxygen‐containing functional compounds such as peroxides, aldehydes, alcohols, ketones, and acids or esters. These reactions are also significant for petrochemical industrial processes and fine chemical industry [41]. Currently, these processes are achieved mainly via transition‐metal‐based catalysts. Air or molecular oxygen is the most environmentally benign oxidant. However, due to the triplet ground‐state structure of oxygen, activation of di‐oxygen normally needs harsh conditions, such as relatively high reaction temperature and pressure. For some reactions, these conditions may lead to over‐oxidation and thus decrease in selectivity toward the desired products. In this scenario, stronger oxidants, such as H₂O₂, tert‐butyl hydroperoxide (TBHP), and other organic peroxides, are often used as attractive alternative oxidants. In fact, they are commonly used in synthetic organic reactions. Not surprisingly, carbon‐based catalysts have also shown their ability to activate these peroxide‐based oxidants in the catalysis of organic reactions, and various activation mechanisms and catalytic centers have been identified or proposed.

9.3.1 The π Electrons of Carbocatalysts

In 1992, Atamny et al. proposed that the oxygen molecule was able to chemisorb onto a graphitic basal plane (sp² C) and reduce to peroxo species by accepting π electrons from graphitic surface [42]. Similar mechanism was also suggested by Khorrampour et al. based on density functional theory studies on single‐walled carbon nanotubes (SWCNTs) [43]. In these cases, graphitic basal plan acts as an electron donor to oxygen molecule. Then, the thus formed peroxo species can migrate on the defect‐free surface of graphitic carbon until they reach reactant molecules absorbed on the carbon surface, which are also possibly activated by the carbon surface. An excellent example was demonstrated in aerobic oxidation of benzyl alcohol (BzOH) to benzaldehyde (BzH) using CNT or N‐CNT as the metal‐free catalysts without any additives/promoters [44]. As shown in Figure 9.3, O₂ was adsorbed and activated on CNT surface to form a peroxo radical. BzOH is also adsorbed onto the CNT surface via a weak π–π interaction between the graphene sheet of CNT and the benzene ring of BzOH. The formed peroxo radicals abstract the weakly bonded α‐H‐atom of the activated BzOH. As a result, a HO₂ radical and benzyl alcohol radical were generated, which then subsequently form benzyl alcohol peroxyl radicals (BAP) in the presence of free oxygen in the solution. It is noteworthy to point out that due to the π–π interaction between the BAP and graphene sheets of the CNTs, CNTs further catalyze the decomposition of BAP to eliminate a HO₂ radical (H₂O₂) to form the product BzH. Possible over‐oxidation of the BzH leads to the generation of benzoic acid as a by‐product.

To investigate the role of surface structures of the carbocatalysts in aerobic oxidation of BzOH, Luo et al. [44] modified the CNTs with heat treatment to introduce different levels of geometric defects and oxygen‐containing groups. They found that the surface carboxylic groups were particularly detrimental for the catalytic activity, which is ascribed to the reduction of the adsorption strength of the oxygen‐containing intermediate (BAP). Furthermore, the introduction of carboxylic group also decreased conductivity and electron mobility of CNTs. As a consequence, the electron transfer between CNTs and BAP radicals was diminished [45], which resulted in the decreased BAP elimination of HO₂ radical to form BzH. To confirm if the electronic property of CNT is critical for the catalytic reaction, they fabricated electron‐rich N‐doped CNTs and compared their catalytic activity with the non‐doped ones. They found that the BzOH oxidation reaction catalyzed by N‐CNTs (1.02 at.% N by XPS) appears to have much lower apparent activation energy (65.4 kJ mol⁻¹) than the non‐doped CNTs (81.3 kJ mol⁻¹). Note that these values are close or even lower than that reported for the ammonium‐molybdate‐catalyzed aerobic BzOH oxidation using H₂O₂ as the oxidant (84 kJ mol⁻¹) [46]. Outstanding recyclability was demonstrated, showing great potential for industrial applications of BzOH oxidation to BzH. Commercial CNTs were also used as catalysts for the aerobic oxidation of cumene to cumene hydroperoxide (CHP) [47]. Under optimal conditions, cumene conversion of 24.1% with CHP selectivity of 88.4% was obtained, which is very close to that obtained using metal catalysts. The CNT catalysts also showed good recyclability after five recycles. A similar hypothesis that CNTs activate molecule oxygen was proposed. Note that in these studies, a long induction period was observed before a fast conversion occurs, which was attributed to the fact that activation of molecular oxygen is not very efficient at these relatively low temperatures. Interestingly, when N‐doped CNT was used for this catalytic reaction [48], the reaction speed largely increased and the apparent activation energy decreased to 12.8 kJ mol⁻¹. Most impressively, the product pattern was also changed from CHP to acetophenone (AP) and 2‐benzyl‐2‐propanol (BP) with an extraordinarily high selectivity. The authors claimed that this was the first time that a metal‐free carbocatalyst can promote cumene aerobic oxidation in one‐step production of BP and AP with a total selectivity as high as 99% and a conversion higher than 80% under atmospheric oxygen pressure and low temperature (353 K). On pristine CNTs, the formed CHP is more inclined to desorb from the surface of the tubes and only a small portion is absorbed and further decomposed for the new catalytic cycle; therefore, the major product was CHP. Compared to the pristine CNTs, N‐doped CNTs have much higher electron density, which not only provides much stronger interactions with the formed CHP but dramatically improves the decomposition of CHP possibly via an electron transfer process. It was reported that SWCNTs can accelerate decomposition of organic peroxide to generate strong oxidative species [49]. The catalytic mechanism has been proposed to involve a related single‐electron transfer process. The similar mechanism has been previously proposed for electron‐rich aromatics [50] and C60 [51]. According to this mechanism, increasing electron richness of carbon materials is expected to improve their catalytic activity. The largely improved catalytic performance described above is consistent with this electron‐transfer‐induced decomposition mechanism.

Work function of materials has been widely used to measure the ability of materials to donate electrons. A material with low work function indicates its high capability to donate electrons. Average work function for a bulk material can be measured by UV photoelectron spectroscopy, and localized nanoscale work function of a material is normally measured by scanning probe microscope techniques, such as Kelvin probe scanning microscopy (KPSM). Peng and coworkers used UV photoelectron spectroscopy to measure the work function of Fe‐filled CNTs, which were used as a measure of the ability of CNT to donate electrons and correlate the specific catalytic activity of these CNTs in aerobic oxidation of cyclohexane [45]. It has been reported that when CNTs are filled with metals, charge transfer between the graphite walls and the encapsulated metals results in an increased electron density of the CNT's outside convex surfaces [52–54]. They indeed found that the specific activity of cyclohexane oxidation exponentially increases with the decrease in the work function with the following relationship:

where r is the specific conversion rate of cyclohexane (mmol m⁻² h⁻¹) and Φ is the work function (eV).

Even though this work involved encapsulated transition metals as catalytic promoters, the dependence described the structure–activity relationship so well that it became an effective index to define the CNTs' activity in the aerobic oxidation reactions. It indicates that it is possible to tailor the catalytic performance by tuning the electronic structure of the carbon materials. It is worthwhile to mention that Peng's group published a series of liquid‐phase aerobic catalytic oxidation of hydrocarbons to value‐added chemicals using nanocarbons (mainly CNTs and N‐doped CNTs) as metal‐free catalysts [47, 48, 55–59]. Outstanding catalytic performance was demonstrated. Some of them even outperformed the currently used transition‐metal‐based catalysts. A common feature of these series reactions is that they all have a free radical oxidation reaction pathway, including radical initiation, propagation, and termination. Among these elemental steps, the propagation is the rate‐determining step. Another interesting feature is that these reactions all use molecular oxygen as the terminal oxidant. However, the carbocatalyst did not activate molecular oxygen directly to generate reactive surface oxygen species (ROSs) nor they did play an important role in the radical initiation step. Instead, the carbocatalysts possibly accelerate decomposition of the organic peroxide formed during the first initiation step via an electron transfer process to form new radicals. These radicals are further stabilized by the large π‐conjugated system of graphene wall, which facilitates the propagation step toward the desired products. According to this catalytic mechanism, carbon nanomaterials with long range order and electron delocalization are preferred for these reactions. Their experiments indeed demonstrate that nanocarbons with more sp² carbon, such as CNTs, are more efficient than those dominated by sp³ carbons, such as diamond. Furthermore, doping with electron‐rich elements, such as N in their graphene matrix, enhances the catalytic activity. However, introduction of oxygenated functional groups on CNTs had a negative effect on their catalytic activities. It becomes very reasonable as introduction of these functional groups results in localization of electrons. Meanwhile, the mechanism that drives the decrease in catalytic activity of the O‐CNTs has also been revealed. It has been proposed that the O‐CNTs serve as free radical quenchers, capturing the free radical intermediates from the reaction and inhibiting the free radical chain transfer; thereby, the O‐CNTs reduce the catalytic activity of the free radical reaction in the liquid phase [58]. Very interestingly, they found that geometric defects, such as nanoholes on the basal planes of the graphene matrix, did not have negative effects. In fact, the existence of these surface defects slightly enhanced the catalytic performance.

Graphene can efficiently decompose H₂O₂ to various ROSs such as hydroperoxide and superoxide. This property of graphene has been explored to simulate the activity of peroxidase to develop enzyme‐free glucose detection sensors [60]. It was also reported that rGO exhibits high activity as a Fenton catalyst for HO^• radical generation and phenol degradation with efficiency similar to that achieved with transition metals, thereby showing the potential of carbocatalysts in heterogeneous Fenton reactions. It was proposed that hydroquinone/quinone‐like functional groups present in rGO have a similar potential to that of the Fe²⁺/Fe³⁺ redox pair. The quinone‐like groups promote the one‐electron reduction of H₂O₂ to efficiently generate HO^• radicals, which are directly involved in phenol degradation [61, 62]. Ma and coworkers reported that rGO can be used as efficient and highly selective metal‐free catalysts for one‐step oxidation of benzene to phenol with H₂O₂ as the oxidant [63]. The moderate H₂O₂ activation rate and the high benzene adsorption ability of the rGO provided a balanced kinetic control over the oxidation reactions, which are involved in the reaction system and are responsible for the outstanding catalytic performance, especially its excellent selectivity to phenol. The chemical identities of the catalytic centers on the rGO were not identified in this work. However, it was experimentally demonstrated that the good π characteristic of the rGO was required to efficiently activate H₂O_2. This is similar to that CHP oxidation reactions described above with CNTs. The zigzag edge of graphene was also mentioned as the possible active sites due to the existence of additional density of states at the Fermi level. Furthermore, the rGO was prepared via reduction of GO with hydrazine hydrate in ammonia solution, which is known to be accompanied by certain level of N doping [64]. Thus, it is not clear if N doping played any role in the observed catalytic behavior. Recently, the Su and coworkers [65] also studied this oxidation reaction with various carbon materials as carbocatalysts, including CNTs, activated carbon, flake graphite, nanodiamond, and acetylene black. Among these nanocarbons, CNTs gave the highest yield (5.8%); however, it is still lower than that of 24% reported by Ma and coworkers with rGO as carbocatalysts [63]. In this work, the possible catalytic centers for activation of H₂O₂ were carefully studied. In particular, a series of small model molecules, such as 9,10‐anthraquinone, 1,4‐benzenediol, phthalide, benzoic acid, and benzyl ether, were applied to study the catalytic roles of carbonyl, phenol, lactone, carboxylic acid, and ether groups in this oxidation reaction. The results indicated that these oxygenated functional groups did not play a catalytic role. Even carbonyl groups, which have been identified to be the catalytic centers for dehydrogenation reactions, did not show a positive effect in this reaction. To study the role of geometrical defects existing on the carbocatalysts, another two model compounds, phenanthrene and anthracene (Figure 9.4), were used to mimic two types of defects, namely armchair and zigzag edges, respectively. Phenanthrene was found to give a much higher phenol yield than anthracene, indicating that the armchair configuration might have some positive influence on the reaction.

9.3.2 Geometrical Defects (Point Defects, Vacancies, and Edge Defects)

As described earlier, in all the carbon materials including the high‐quality graphene and CNTs, certain levels of defects as point defects (one of two atom missing) or/and n‐member rings (non‐6‐member rings) exist in their basal planes. It was reported that the electronic transport properties of SWCNTs exhibited extremely high sensitivity to oxygen environments [66]. This has been explained by the fact that due to the presence of on‐tube defects, charge transfer occurs between molecular oxygen and SWCNTs, which induces p‐doping effects of the tubes. DFT calculations have suggested that molecular oxygen could be activated and dissociated with almost no barrier on the vacancies and edge defects on carbon materials to form various ROSs [67–69]. Based on these proposals, Su's group has developed strategies to prevent the formation of electrophilic ROSs by blocking these defect sites with phosphonated or borate groups. The modifications remarkably enhanced the alkene selectivity in ODH reactions of these carbocatalysts [28, 29, 70, 71]. At the same time, these modifications also largely improved the oxidative resistance to combustion [71, 72].

Compared with CNTs and nanodiamonds, graphene materials, especially graphene nanosheets, nanoribbons, and holey graphene (which are referred to as graphene materials with nanoholes existing in their basal planes), are much more abundant in edge defects [73–75]. The properties of the edges play an important role in their overall electronic properties and catalytic properties. It has been theoretically predicted and experimentally observed that localized nonbonding π electrons exist in the exterior zigzag edges of graphene [76, 77]. The localized edge‐state electrons are strongly spin‐polarized and coupled with the itinerant π carriers. They have been referred to as π‐electronic spin‐polarized states, or Fujita states, which have been the foundation to explain the experimentally observed exotic electronic and magnetic properties of graphene. A theoretical study demonstrated that the zigzag edge states in graphene nanoribbons chemically behave like a partial radical that is capable to activate a wide range of molecules, including molecular hydrogen and oxygen [69]. By depositing an air‐saturated electrolyte solution droplet (∼15 μm) either on the edge or on the basal plane on highly oriented pyrolytic graphite (HOPG), Dai, Wang, and coworkers individually studied the electrochemical catalytic behavior of the basal plane and the planes with edge atoms. As the basal plane of the HOPG is almost defect free, the results provided direct evidence for the catalytic roles of the edge carbon atoms in ORR [78]. In the chemical catalysis, an excellent example demonstrating this mechanism is the oxidative coupling of amines using porous GO as a metal‐free carbocatalyst reported by Loh and coworkers [79]. The porous GO, which was named as ba‐GO, was obtained by a sequential base and acid treatment of GO prepared by Hummers' method. With only 5 wt% loading of the carbocatalyst, the yield of imine reached 98% under solvent‐free, open‐air conditions. The performance is comparable or even superior to that of using transition metal catalysts. The low catalyst loading is unprecedented compared with most of the catalysis reaction using GO as the catalyst. The origin of the dramatically enhanced catalytic activity was probed and was contributed to the synergistic effect of the unique functionalities existing along the edge defects that were created during the base–acid treatment, specifically the unpaired electrons and carboxylic acid groups (Figure 9.5). During the base treatment, a large amount of nanoholes and their associated edge defects were created on the basal planes of GO. These edge sites with unpaired electrons constitute the active catalytic sites and afford enhanced kinetics for the trapping and activating molecular oxygen by a sequence of electron transport and reduction steps to superoxide radical (^•O₂)_, which acted as the true oxidant in this catalytic reaction. This hypothesis was supported by three nicely designed experiments: (i) electron paramagnetic resonance (EPR) studies to measure the localized spins created at the edge of ba‐GO, (ii) comparable catalytic studies to see if the catalytic activity decreased upon selectively blocking the unpaired electrons via a diazonium coupling reaction, and (iii) in‐situ spin‐trapping EPR studies to trap the in‐situ‐formed radicals. The following acid treatment recovers the carboxyl anionic groups positioned along the edges to carboxylic acid groups, which produced a synergistic effect for the coupling reaction. These carboxylic groups serve two important roles in the reaction: act as hydrogen bonding sites to the amines and facilitate proton transfer reactions due to acidic properties. In this work, it was also mentioned that the base reduction removed the hydroxyl groups from the GO, which has the effect of changing the dynamics of water solvation layer around GO, allowing greater access to the catalytic sites by the reactant molecules. The combination of solvent‐free and metal‐free catalysis using low catalyst loading is a good model for practical application. This work provides an excellent example that industrial‐relevant carbocatalysts can be developed.

Another example was demonstrated in the study of rGO as metal‐free catalysts in oxidative desulfurization reactions using molecular oxygen as the terminal oxidant [80]. It was demonstrated that rGO catalyst is effective to remove broad range of sulfur‐containing compounds from fuels and exhibits excellent reusability. The mechanistic studies based on X‐ray photoelectron spectroscopy (XPS) analysis, chemical titration method, and a series of comparative experiments revealed that carbonyl groups played a crucial role during the oxidation process. Defects existing in the rGO, such as vacancies, are also beneficial to the catalytic performance because carbonyl groups could be generated in situ on these defects under the reaction conditions. However, the carbonyl groups themselves were not directly involved in the generation of ROS. Instead, the electron‐withdrawing properties of the carbonyl groups reduce the electron population of the carbon atoms at their adjacent positions, which facilitates absorption and activation of molecular oxygen. Then, the strongly adsorbed oxygen molecules convert to adsorbed super oxygen anion radicals (rGO–OO–^•). At the same time, the sulfur‐containing substrates turn into sulfur‐centered cation radicals, which react with the negative charged radicals of rGO–OO–^• and generate sulfones as the final products.

9.3.3 Heteroatom Doping

Apart from the graphitic basal plane and existing defect sites (naturally existing or intentionally introduced), the introduction of heteroatoms (O, N, B, P, S, etc.) into the carbon matrix brings new defects with tailored/controlled physicochemical and electronic properties for optimal catalytic performance [4, 7, 81].

9.3.3.1 O‐Doping

Even though oxygen as a heteroatom is impossible to be substitutionally doped into the matrix of carbon materials due to its strong electronegativity and large size, it brings a wide variety of oxygen‐containing functional groups on the basal planes and along the edges. These functional groups are either electron rich or deficient, acidic, or basic, which render carbon‐based materials catalytically active for a wide variety of catalytic reactions [82]. It is noteworthy to mention that the oxygen functional groups directly attacked the reactants before the reaction. Oxidants, such as molecular oxygen or peroxide, are required to regenerate the catalytic centers. The gas‐phase oxidation reactions as described in dehydrogenation reactions are clear examples. Here, we will list several liquid‐phase oxidation reactions, especially with graphene oxide. Bielawski and coworkers pioneered the application of GO as metal‐free catalysts. A wide range of catalytic oxidation reactions have been demonstrated, such as oxidation of alcohols, thiols, and sulfides [83]; CH oxidations [84]; hydrations of various alkynes [85]; olefin polymerization [86]; and among others. DFT and experimental studies were reported to propose that epoxide functional group plays a role in the oxidation reactions. Using oxidation benzyl alcohol as an example, in the first step of the reaction, hydrogen will transfer from CH₂ group of benzyl alcohol to GO surface followed by ring opening of epoxide group [87]. In the next step, aldehyde is formed followed by dehydration and GO reduction. This rGO catalyst can be partially reoxidized in the presence of molecular oxygen; however, it requires high amount of energy. In addition, O‐doped carbon catalyst is not stable at moderate temperatures (60–100 °C) at which it starts reducing, making it less feasible for catalytic applications [88, 89]. Recently, Hutchings group demonstrated that GO can be used as an efficient carbon catalyst for low‐temperature aerobic epoxidation of linear alkenes in the absence of solvent, initiator, and metals [90]. GO with a minimum of 15% and an optimum of 25 wt% oxidation level is required for the catalytic reaction. The beneficial oxygen‐containing functional groups are not identified, but they pointed out that GO prepared by the commonly used permanganate‐based Hummers method [91] is inferior to the one prepared by the less commonly chlorate‐based Hofmann method [92]. They also found that sulfur residues in the form of both inorganic and organosulfate acted as deactivators, so are the “bad” species for this catalytic reaction.

9.3.3.2 N, B, and NB Co‐doping

Different from oxygen, N and B have similar diameters as C atoms. Depending on the doping precursor and the temperatures applied during the doping process, both can be incorporated as surface functional groups or as substitutional doping with different bond configuration. Carbon materials doped with nitrogen (N) and N co‐doped with B have been explored for catalytic oxidation reactions. Much improved catalytic activity was observed compared to the material without doping. For some reactions, the catalytic activity can be further improved via co‐doping with B. Interestingly, B‐doped carbon materials have also been explored for oxidation reactions, but the catalytic efficiency was not improved, possibly due to the electron‐deficient nature of B, which induces electron deficiency of the graphene surface [93].

Long et al. explored nitrogen‐doped graphene nanosheets as metal‐free catalysts for aerobic selective oxidation of benzylic alcohols [94]. They found that among the three types of nitrogen species (pyridinic‐N, pyrrolic‐N, and graphitic‐N) doped in the graphene lattice, the graphitic N species were the catalytically active centers based on the good linear correlation with the activity results. Kinetic analysis showed that the catalytic oxidation has a moderate activation energy (56.1 ± 3.5 kJ mol⁻¹), which is close to that of precious‐metal‐based catalysts (Ru/Al₂O₃ catalyst has an apparent activation energy of 51.4 kJ mol⁻¹).

This work proposed that a transition state of sp² N−O₂ adduct wad formed, which oxidized alcohols directly to aldehydes without any by‐product, including H₂O₂ and carboxylic acids (Figure 9.6). Note that many efficient heterogeneous systems based on active noble metals, including Pt, Pd, Au, Ru, and their alloys, have been developed for selective oxidation of alcohols to aldehydes by molecular oxygen. However, the formation of a large amount of H₂O₂ as a by‐product seems to be unavoidable in these noble‐metal‐catalyzed systems and some TEMPO (2,2,6,6‐tetramethyl‐1‐piperidinyloxy)‐catalyzed systems. H₂O₂ can further react with oxygen‐containing products, leading to selectivity loss. This work demonstrated that carbocatalysts not only has the advantage of being sustainable but also highly selective. Watanabe et al. applied nitrogen‐doped activated carbon as metal‐free catalysts for aerobic oxidation of alcohols [95]. The graphite‐type nitrogen was also demonstrated as the catalytically active sites, which adsorb and activate oxygen molecules. However, a different mechanism was proposed as shown in Figure 9.7. Similar to the case of activation of oxygen molecules in electrochemical catalytic ORRs, the nitrogen dopant did not participate in the activation of oxidant; instead, the adjacent carbon atoms accomplish the task due to their increased charge density and/or spin density [96]. The authors also pointed out that the graphite‐type nitrogen species are likely to change to less‐active oxidized species during the reaction, causing catalyst deactivation during the oxidation reaction. Due to the nature of the reaction, it seems the deactivation cannot be avoided, which influences recyclability of the catalysts. Further, the turnover frequency (TOF) value is more than one order of magnitude smaller compared with that for carbon‐supported Ru and Pt catalysts. In addition, due to the planar structure of graphitic sp² N, these carbocatalysts also show difficulties in overcoming substrate steric hindrance effects, which cause limited catalytic reaction scope of N‐doped carbon materials.

During the study of selective oxidation of ethylbenzene in aqueous solution with TBHP as the oxidant and N‐doped graphene as metal‐free catalysts, Ma and coworkers also found that the graphitic‐N is critical for the remarkable activity in the CH oxidation reaction [97]. DFT calculation suggested that the nitrogen atoms are not able to host the peroxide species because of the high negative charge of nitrogen. Both the electronic charge and spin density on the ortho‐carbons are superior positions for the adsorption of reactive oxygen species such as peroxide. Using both C K‐edge and N K‐edge X‐ray absorption spectroscopy (XAS) to study N‐doped carbon catalysts before and after TBHP and ethylbenzene (the reactant) treatment, they clearly demonstrated that the graphitic nitrogen dopant modulated the electronic structure of sp² carbon material. The density of states intensities near the Fermi level for the adjacent ortho‐carbon are much stronger than those of undoped graphene carbon, which gives the nitrogen‐neighboring carbon a metal‐like d‐band electronic structures. These adjacent carbon atoms were the active centers for the activation of TBHP to form and host peroxide‐like species, which are the active oxygen species actually performing CH activation. Similarly, they also demonstrated that N‐doped graphene materials were able to give the highest recyclable catalytic activity for the epoxidation of trans‐stilbene, with 95.8% conversion and 94.4% selectivity to trans‐stilbene epoxide [98].

Annealed ultra‐dispersed nanodiamond (ADD) is regarded as an interesting sp²/sp³ hybrid material. It combines the remarkable surface properties of graphene‐based nanomaterials and the intrinsic properties of a diamond core [99]. Owing to its highly curved concentric graphitic shells, it exhibits more active surface physicochemical properties than a purely graphene material. Lin et al. applied three approaches to introduce N doping in the graphene shell of ADD, which was referred as nitrogen‐modified annealed nanodiamond (N‐ADD) [100]. The selective oxidation of benzylic alcohols to benzylic aldehydes in the presence of TBHP (^tBuOOH) was used as a probe reaction to study the catalytic activity of N‐ADD. Interestingly, in this work, a linear relationship was found between the pyridinic N:C atomic ratio (not the graphitic‐N) and the benzyl alcohol conversion rate. The detailed mechanistic study (Figure 9.8) of the catalytic reaction revealed that the pyridinic‐N did not activate TBHP directly; instead, it acts as a hydrogen bond acceptor due to its high localized electron density. The catalytic process initiates with hydrogen bond formation between the pyridinic‐N and ^•OH free radical, which is derived from decomposition of ^tBuOOH. The resulting pyridinic N–OH active species captures a hydrogen atom on the hydroxyl group of BzOH, which leads to the formation of a CH₂O group. In the same time, a water molecule was released that results in regeneration of the pyridinic‐N closing the catalytic loop. However, the CH bond of the formed CH₂O group was further activated by the intermediate free radical ^tBuOO^• that originated from the reaction of the decomposed ^tBuO^• with a portion of ^tBuOOH. Finally, the product BzH was formed under mild conditions.

Co‐doping with several different heteroatoms could synergistically interact and improve catalytic performance, which has been widely demonstrated in electrochemical catalytic studies, such as ORR, OER, and HER. Wang et al. demonstrated that incorporation of B and F into the matrix of mesoporous carbon nitride increases cyclohexane oxidation with H₂O₂ as the oxidant [101]. The catalytic centers were not mentioned in this work. Enhanced catalytic performance is observed compared to the materials without doping. The heteroatoms B and F played an important role in the much enhanced catalytic activities. The Garcia group applied N‐, B‐, and B,N‐co‐doped graphene (B,N‐G) as metal‐free carbocatalyst for the selective aerobic oxidation of benzylic hydrocarbons and cyclooctane to the corresponding alcohol/ketone mixture with more than 90% selectivity [102]. The most active material was co‐doped (B,N)G. In the absence of solvent and with a substrate:B,N‐G ratio of 200, 50% tetralin conversion was achieved in 24 h with a alcohol/ketone selectivity of 80%. Using O isotope labeling and an in situ Raman spectroscopic study, they found that hydroperoxide‐like species formed on the graphene sheet upon oxygen activation [80]. A free radical mechanism was proposed. The peroxo species formed on graphene sheets abstract a hydrogen atom from tetralin by generating a benzylic radical, which reacts with oxygen to give the product. The same catalysts were also explored for aerobic oxidation of styrene, resulting in around 45% styrene oxide (SO) selectivity under solvent‐free conditions [102]. The most likely reaction mechanism requires the generation of carbon‐centered radicals and epoxidation is favored by the presence of BA. A comparison of the optimal doped G catalysts used for aerobic oxidations of benzylic compounds and styrene showed significant differences. Although benzylic oxidation is promoted by all doped Gs, B,N‐G is the most active. In high contrast, B,N‐G catalysts are not active in styrene epoxidation, demonstrating that this reaction is more sensitive to the nature of bond configurations of the doping elements. Further experimental and theoretical work is required to fully rationalize these differences in the catalytic activities of doped G.

9.3.3.3 P, S, and P,S Co‐doping

Phosphorus has the same number of valence electrons as N, making P‐doped carbon materials also electron rich [103]. The polarity of the CP bond is opposite to that of the CN bond due to lower electronegativity of P atoms (2.19) than C (2.55) [4]. Furthermore, as the diameter of P is much larger than C, P doping results in more local structural distortion of the hexagonal carbon framework and in such a configuration, P protrudes out of the graphene plane [104]. All these characteristics empower P‐doped carbon materials to overcome the steric hindrance effects encountered in N‐doped carbon materials [94, 102]. However, experimentally, most of the approaches for P‐doping necessitate accompanying O doping, forming various P‐ and O‐containing functional groups [105–107]. These functional groups and the bonding configuration of P in a carbon matrix could influence the electronic property and therefore its catalytic performance [105]. Recently, He and coworkers reported an extremely simple and rapid (seconds) approach to directly synthesize gram quantities of P‐doped porous carbon materials from abundant biomass molecules [89]. The work function of P‐doped carbon materials and its connectivity to the P bond configuration in the carbon matrix have been studied via PeakForce Kelvin probe force microscopy (PF KPFM). The capability of the P‐doped carbon materials as metal‐free catalysts for aerobic oxidation reactions has been demonstrated for the first time. Unlike N‐doped carbon material, which can only catalyze primary benzyl alcohol oxidation due to steric effects, the P‐doped carbon materials can efficiently catalyze aerobic oxidation of both primary and secondary benzyl alcohols to the corresponding aldehydes or ketones. Furthermore, in direct contrast to N‐doped graphene, the P‐doped carbon materials with higher work function show high activity in catalytic aerobic oxidation. The selectivity trend for the electron‐donating and ‐withdrawing properties of the functional groups attached to the aromatic ring of benzyl alcohols is similar to that of the metal‐based catalysts, but very different from other metal‐free, carbon‐based catalysts. Based on the experimental results, a unique catalytic mechanism was proposed (Figure 9.9), which is different from both N‐doped and O‐doped carbon materials. In the first step of catalysis, condensation between the alcohol and PO moieties on PGc takes place, and an alcoholate intermediate is formed. The condensation is likely facilitated by the interaction of the alcohol with the PGc surface by π−π interactions with the graphitic domains and hydrogen bonding with the polar groups (such as P−OH). In the second step of the reaction, a rate‐determining H transfer takes place, possibly through a cyclic transition state. The aldehyde product and a water molecule are released simultaneously. Next, the generated P(III) groups on PGc react with molecular oxygen to regenerate the PO centers for further reactions, thereby completing the catalytic cycle. Similar role of molecular oxygen for regenerating the active site was also reported in the ODH of alkane and ethyl benzene catalyzed by other non‐doped or O‐doped carbon catalysts [25, 28, 108].

S‐doped carbon materials were reported as an electrochemical catalyst for ORR and OER [109, 110]. However, studies aimed at using them as catalysts for organic synthesis, especially for oxidation reactions, are rare [111, 112]. In the continuing work, the He and coworkers reported that S‐doped carbon can also activate molecular oxygen and shows catalytic role in benzylic alcohol oxidation [93, 113]. Furthermore, they found that the P,S dual‐doped Gc catalyst shows the most improved catalytic performance compared with single‐doped (S‐Gc and P‐Gc) and other P‐co‐doped carbon catalysts (P,B‐Gc and P,N‐Gc) [113]. The active centers originated from P doping and S doping to additively and/or synergistically catalyze the aerobic oxidation reactions through different pathways. For the first time, using X‐ray absorption near‐edge structure (XANES), the catalytic centers stemming from S doping were experimentally identified to be exocyclic S species (C–S–C, sulfur out of the carbon ring) (Figure 9.10), which are different from those proposed for electrochemical ORR with a 4e⁻ pathway and for OER. Notably, all of the catalytic sites from both P and S doping share a similar “protruding out” pyramid structure, which is in contrast to the planar structure of the catalytic sites in N‐ or B‐doped graphitic materials. The unique geometric structure of the catalytic sites can minimize substrate steric hindrance effects, allowing the P,S‐co‐doped catalysts to be characterized by a wide substrate scope and functional group tolerance. The fundamental knowledge gained from these comprehensive studies could provide valuable guidance for designing and developing more efficient catalysts to satisfy practical industrial applications.

9.4 Reduction Reactions

Carbocatalysts for reduction reactions have been much less studied than for oxidation. At present, several reductants or hydrogen resources, such as molecular hydrogen, hydrazine, NaBH₄, and liquid hydrogen, have been applied in the reduction of various organic substrates. Similar to oxidation reactions, the activation mechanisms, the active centers, and thus the reaction mechanisms and reduction pathways depend on the structure of the carbon catalysts as well as the molecular structure of the reductants and the reactants.

9.4.1 Molecular Hydrogen as the Reductant or Hydrogen Resource

Hydrogen is an environmental‐friendly and nonpolluting resource. Reactions that use H₂ as a reductant (i.e. hydrogenations) are some of the most general organic reactions widely used in both pharmaceutical and fine chemical industries. Owing to the strength of the HH bond, dissociative adsorption of diatomic molecule H₂ on a catalytic surface is a central step in these catalytic processes. Accordingly, the major consideration in designing catalysts for hydrogenation reactions is the ease of dissociative hydrogen adsorption on the catalytic particle surface. In the past, metals and metal oxides have been the most widely used catalysts in industrial catalytic processes. However, these metal‐based catalysts often suffer from several disadvantages, including their high cost, susceptibility to gas poisoning, and detrimental effects on the environment. The development of inexpensive, metal‐free catalysts with high performance is, therefore, a highly desirable goal. However, it is still unclear if carbon material could activate molecular hydrogen to give atomic species even though both theoretical and experimental studies have been reported. The catalytic mechanisms, the catalytic centers, and the influence of the metal impurities in the observed catalytic behaviors are also under discussion so far.

As mentioned in the early sections of this chapter, due to their extremely large surface areas, carbon materials have been widely used as supports to prepare highly dispersed catalysts for various applications, including for hydrogenation reactions. The use of carbon supports themselves as catalysts was stimulated by the discovery of the double catalytic effects when transition metal catalysts were supported on carbon materials. In 1977, Keren and Soffer demonstrated that pure carbon could split gaseous hydrogen into atomic form through hydrogen chemisorption [114]. During examination of the hydrogenation activity of metal catalysts supported on activated carbon, Zhang et al. found that the AC support is not only catalytically active for H₂ dissociation itself but also for hydrogen transfer from donor tetralin to anthracene [115]. To rule out if the observed catalytic activities were not resulting from the inorganic ash content existing in the AC materials, a control experiment was performed in their following‐up work using demineralized AC catalyst. As the anthracene conversion and product distributions were found to have no significant change, they concluded that the observed catalytic activities originated from carbon itself. In this work, the catalytic activity of AC for H₂ dissociation and hydrogen transfer in the hydrogenation of anthracene was further examined in the temperature range of 300–400 °C with three kinds of hydrogen sources: hydrogen gas, hydrogen donor tetralin (i.e. transfer hydrogenation), and the combination of both [116]. With hydrogen gas alone, the hydrogen transfer rate increased with both hydrogen partial pressure and temperature, whereas in the case of tetralin, it depended on the concentration of hydrogen atoms formed on the surface of activated carbon by dehydrogenation of tetralin. When both hydrogen gas and tetralin were used together, the hydrogen transfer rate was higher than that obtained separately with hydrogen gas alone or tetralin alone, but much lower than the simple sum of both rates in the range of 350–400 °C. It seems that the hydrogen transfer over the activated carbon mainly occurred from tetralin to anthracene, whereas the function of hydrogen gas in this case was mainly to suppress the release of hydrogen atoms from the carbon surface. As a consequence, the amount of tetralin solvent was largely reduced when hydrogen gas was co‐fed during the reaction. It is noteworthy to mention that a different scenario is observed when using an activated‐carbon‐supported Ni catalyst. In this case, the hydrogen mainly transferred from the gas phase into anthracene through Ni metal via a spillover mechanism. When both tetralin and hydrogen gas were used together, the rate of hydrogen gas consumed in the hydrogenation was lower due to the competitive adsorption of tetralin on Ni metal, especially at lower temperatures. In 2004, Sun et al. applied AC as metal‐free catalysts for hydrogenation of benzene and four polycyclic arenes (PAs) at 300 °C with hydrogen pressure of 5 MPa [117, 118]. They found that the AC was able to selectively catalyze hydrogenation of these substrates and their reactivity toward hydrogenation was determined by their hydrogen‐accepting abilities, which can be evaluated by the super‐delocalizability(Sr) [119], steric hindrance effects, and their adsorption strengths on the AC catalyst surfaces. It is worthwhile to mention that these early studies using activated carbon, activated coal, and carbon black as catalysts were normally performed in high H₂ partial pressures (1–6 MPa) and high temperatures (>300 °C). Furthermore, the chemical identities of molecular structures responsible for H₂ chemical absorption association and hydrogen transfer are largely unknown.

Recent studies with carbon nanomaterials with more defined structures, such as CNTs, fullerene, and graphene, lead to different conclusions. Schimmel et al. studied the interaction of hydrogen with activated charcoal, carbon nanofibers, and SWCNTs [120]. They found that at room temperature and under ambient pressure, molecular hydrogen is just physisorbed on aromatic carbon surfaces with very short residence time. In general, one would not expect any noticeable charge transfer to occur between physisorbed species and its substrate. Therefore, a dissociative activation of molecular hydrogen via aromatic carbon surfaces to atomic hydrogen does not seem feasible [121]. In 2006, it was first reported that rigid molecules with frustrated Lewis acid–base pairs (FLPs) at appropriate distances have the ability to activate hydrogen molecules at low temperatures [122, 123]. In principle, if such FLPs exist in carbon materials, such carbon materials should also be able to activate molecular hydrogen and act as hydrogenation catalysts. However, this requires that the carbon skeleton accommodates Lewis basic and acidic sites and that the base–acid pairs exist in such a distance that they can polarize H₂. The Garcia group has experimentally demonstrated this possibility [124, 125]. However, the chemical identities of the FLPs on the carbon surfaces are still unknown. Furthermore, it was proposed that there were no obvious steric constrains of carbon materials that would prevent such discrete Lewis base and acids sites from forming Lewis acid–base adducts, unless the carbon materials were doped with different elements. Indeed, DFT calculations using periodic models from the Su and coworkers suggested that a mechanism could involve H₂ activation in doped graphenes [126]. The influence of H₂ activation in the presence of co‐dopants was also studied using DFT calculations [127]. According to these studies, pyridinic‐N‐ and B‐doped domains at the edges are involved in the activation of H₂. Hydrogen activation occurs at the carbon atom surrounding the dopant element rather than on the dopant itself. The higher activity can be attributed to the presence of less coordinating dopants on the edges rather than on the basal plane and their strong acidic/basic properties.

Experimentally, in 2009, Xu and coworker demonstrated that fullerene can activate molecular hydrogen as a novel metal‐free hydrogenation catalyst [128]. Hydrogenation of nitrobenzene to aniline was achieved with high yield and selectivity under one atmospheric pressure of H₂ and light irradiation at room temperature. A cooperative effect between C₆₀ and C₆₀⁻ anion was observed. With an optimum ratio of 2 : 1, 100% conversion with 100% selectivity was achieved, whereas no conversion was recorded in the absence of light. To study if the promoted catalysis was due to the presence of trace amount of metals, the catalyst was screened using inductively coupled plasma mass spectrometry (ICP‐MS). Trace amount of metals such as Co, Cr, Fe, Cu, and Ag were detected; however, the effect of these metals in the reduction of nitrobenzene was not significant when the metal ions were used alone without fullerene. Therefore, they concluded that the observed catalytic activity is from C₆₀ and not metal contaminations. Later, van Bokhoven and coworkers [129] studied this reaction in great detail and found that Ni residue possibly served as the real active site in this reaction. The balance of charge transfer between C₆₀⁻ and Ni²⁺, which generates atomically dispersed Ni atoms, contributes to high catalytic activity.

The first experimental work using rGO for hydrogenation of ethylene was reported by Perhun et al. [130]. The reaction required very high temperature (500 °C) and high pressure of hydrogen to proceed, which is less interesting from a practical point of view. In 2014, the Garcia and coworkers reported that graphene‐based materials can catalyze selective hydrogenation of acetylene in large excess of ethylene with molecular hydrogen [124]. Two types of graphene materials were found active: rGO and graphene. The rGO was fabricated via thermal reduction of GO (from Hummers method) in water at 150 °C in an autoclave and the graphene material was fabricated via pyrolysis of a biomass molecule, alginate, at 900 °C in an inert environment. These were graphene platelets of few layers with a dimension of around 2 μm or smaller and having a turbostratic structure. The hydrogenation of acetylene increases with an increase in temperature; 99% of acetylene conversion was reported together with 21% of ethylene conversion at a temperature of 110–120 °C with graphene. In this work, the effect of heteroatoms doping, such as N, S, P doping in the graphene matrix, for the catalytic hydrogenation of acetylene was also studied. Catalytic effects were not observed, which was ascribed to the instability of these doping elements under the reaction conditions. To consider the possibility of metal impurities (in this particular case, the presence of Mn was detected) toward the activity in hydrogenation, Mn and Pd metals were purposely incorporated into the graphene matrix. The presence of Mn decreased the activity (as was inactive for hydrogenation and the addition of Mn possibly blocks some of the active sites). However, inclusion of Pd improved the activity as Pd is active for hydrogenation. The application of these catalysts was further extended to study the hydrogenation of CC bonds in liquid phase and hydrogenation of nitro compounds [125]. A conversion of 16% with 30% selectivity to the corresponding amine was achieved with graphene.

The activity was rationalized by the presence of FLPs, which can activate H₂ by its polarization, followed by the formation of H⁺‐ and H⁻‐ like sites (Figure 9.11). The H₂‐TPD (temperature‐programmed desorption) results suggest the strong uptake of H₂ on Gr, leading to subsequent hydride transfer to the organic molecules having multiple CC bonds. The presence of acidic and basic sites was further confirmed by the titration of Gr and rGO with CO₂ and NH_3, where strong peaks contributing for the basic sites were observed at 680 °C in the CO₂ desorption profile. Further mechanistic and in situ studies are necessary to determine the chemical structures of the sites that activate molecular H₂. Nevertheless, these results demonstrated the possibilities to extend the application of carbocatalysts to hydrogenation reactions or reductions at lower temperatures. The Muhler group reported that N‐functionalized multiwalled carbon nanotubes (MWCNTs) can activate molecular hydrogen for hydrogenation of 1,5‐cyclooctadiene [131]. The activity was ascribed to the presence of nitrogen‐containing functional groups as well as surface defects, whereas the oxygen‐containing groups and their associated defects on the MWCNTs did not show clear contribution to the hydrogenation activity. Very recently, Xiao and coworkers [132] reported that N‐doped rGO exhibited high catalytic activity for deep hydrogenation of anthracene. N doping with different bond configurations in the graphene matrix played different roles in the catalytic reaction. First of all, incorporation of N atoms gave more vacancies in the graphene structures, which favor H₂ adsorption and dissociation from previous DFT calculations [133]. Furthermore, the pyridinic‐N atoms, which are preferentially located at the edges of the graphene structures, can modify the band and electronic structures of the neighbored carbon atoms. The high electronegativity of N creates a net positive charge on the adjacent carbon atoms, which in turn enhance their capabilities to attract electrons from molecular hydrogen, weakening the HH bond and facilitating its dissociation. At the same time, the synergy of the graphitic‐N and the sp² CC structure activates anthracene via enhanced π–π interactions. The cooperation of anthracene activation and hydrogen‐dissociative adsorption accounts for the observed catalytic activities. In addition, the N‐doped rGO is effective for hydrogenation of various polyaromatic hydrocarbons.

9.4.2 Hydrazine as the Reductant (Nitro Group Reduction)

As early as 1985, Han and Cho et al. reported that natural graphite could catalyze the reduction of nitrobenzene and substituted nitroarenes to the corresponding anilines with hydrazine as a reducing agent. However, relatively high temperature (>120 °C) and very high catalyst/reactant ratio (weight ratio > 2) were required [134].

9.4.2.1 Nitrobenzene Reduction Reaction Pathway

The reduction pathways of nitrobenzene reduction have been widely studied with various reductants. It can occur either via a direct route or a condensation pathway. In the direct route, the reduction of nitrobenzene occurs initially to give nitrosobenzene, with the subsequent reduction to phenyl hydroxylamine and aniline. In the condensation pathway, the produced nitrosobenzene couples with phenyl hydroxylamine to form azoxybenzene, which is then reduced to azobenzene, hydrazobenzene, and then aniline. Each step is a 2e⁻ reduction process (Figure 9.12).

Larsen et al. studied the catalytic mechanism using a carbon catalyst (Black Pearls L carbon 15%) in the reduction of nitrobenzene by hydrazine [135]. The intermediate in this chemical catalytic reaction was identified to be phenylhydroxylamine by NMR and other trapping methods. They also found that the two‐electron intermediate, nitrosobenzene, gave different products than nitrobenzene under the reaction conditions and could not be trapped. These results led them to conclude that the nitro group first undergoes an initial four‐electron reduction to phenylhydroxylamine, which was then reduced to aniline in the second step. In this study, they nicely demonstrated that hydrazine was oxidized via a two‐electron pathway to give diimide. It is well known that diimide does not directly reduce nitro groups. Moreover, without a catalyst, it must be impossible to execute a four‐electron reduction using a two‐electron reducing agent without forming a termolecular complex. All these results led to a hypothesis for the catalytic mechanism of this reaction: the carbon catalyst serves as an adsorbent to collect both nitrobenzene (the oxidant) and hydrazine molecules (the reductant) onto its surface. It also serves as an electrical conductor and mediator, so that electrons flow through the carbon to perform the reaction, therefore a four‐electron reduction process achieved by a two‐electron donor. This mechanistic hypothesis was in line with the proposal by Spiro et al. in carbon‐catalyzed redox reactions, in which carbon also serves as both an adsorbent and an electrode conductor [136].

In 2011, Bao and coworkers [137] reported the catalytic activity of rGO for the reduction of nitrobenzene with hydrazine hydrate as the reducing agent. Similar activity was observed with GO as a catalyst (≈94.2% of aniline). The ¹³C NMR of the used GO catalyst indicated that the GO catalyst was in situ reduced by the hydrazine in the reaction environment. Using ¹³C NMR to in situ monitor the nitrobenzene reduction process with rGO, they found that N‐phenyl hydroxylamine was possibly an intermediate in this reaction, suggesting the direct pathway for nitrobenzene reduction. The reusability of rGO for the reduction was up to nine cycles, maintaining 95% yield of aniline. The catalytic centers were proposed to be the zigzag edges of the rGO. To eliminate the possibility that the catalytic activity is achieved due to the presence of trace amount of Mn used in the GO synthesis, they prepared rGO with low content of manganese by a more intensive cleaning process. They also tested the catalytic activity with Mn‐free high‐quality graphene, which was synthesized via arc discharge technique. The former catalyst reported similar conversions, whereas the latter showed a decreased conversion. However, the catalytic activity was increased upon treatment of high‐quality graphene with concentrated nitric acid to create more defects and edges, further suggesting the catalytic role of defects and edges in rGO catalysts. This conclusion was supported by DFT calculations. It was observed that the carbon atoms at the zigzag edges interact with the terminal O atoms of nitro benzene, which weakens the NO bonds and, therefore, activates the nitrobenzene molecules. However, it is well accepted that oxygen‐containing groups still exist in the basal plane and along the edges [138, 139]. Furthermore, the rGO used in the work was obtained by a reduction process with hydrazine hydrate following a modified Wallace's method [140]. It is known that this reduction approach would unintentionally introduce N doping in rGO. Thus, the N doping and the other oxygen‐containing functional groups may also have catalytic effects on this reaction, which was not addressed in this work.

The catalytic behavior of different carbon materials, mainly CNTs, on the reduction of nitrobenzene by hydrazine hydrate was also studied by Su and coworkers [141, 142]. The reaction also proceeded through the direct route in which the intermediate nitrosobenzene was directly converted to aniline. With the help of TPD analysis, certain oxygen‐containing groups can be controllably removed and the catalytic activity studied as a function of the remaining functional groups. These studies helped them to elucidate which groups act as possible catalytic centers on the reaction and which groups may have negative effects. They found that the oxygenated groups, especially the carbonyl groups, were responsible for the catalytic activity. The conjugated π system was also necessary for electron transfer and nitrobenzene adsorption, which was another critical factor in the observed catalytic activity. However, the carboxylic group and anhydride adversely affected the reaction. Other carbon materials such as oxidized nanodiamonds (UDD), GO, rGO, and oxidized activated carbon (oAC) were also studied in this work. The results also indicated that both oxygen functional groups and conjugated π systems on the surface of carbocatalysts are critical. The surface area, pore structure, morphology, structural defects, and Fe impurities in the catalysts did not have a significant influence on the activity. Using H₂O₂ as a green oxidant to oxidize CNTs, the thus produced CNT catalysts exhibited much higher activities compared to those obtained by the commonly used HNO₃ oxidation approach [142]. The higher catalytic activity was ascribed to the comparatively lower content of the negative functionalities, such as the carboxylic and anhydride groups. However, in these works, it was not mentioned as to why these functional groups have reverse catalytic effects. However, in the work by Zhou et al., activated carbon treated with commonly used acids such as HCl, HNO₃, and H₂SO₄ [143] was explored as a catalyst for nitrobenzene reduction with hydrazine hydrate as the reducing agent. Unsurprisingly, the amount and the characteristic fingerprints of various oxygen functional groups on the ACs are different, depending on the acid applied in the treatment. Interestingly, they found that the oxygen functional groups could promote higher rate of decomposition of the reducing agent hydrazine hydrate, which leads to decreased catalytic reaction rate of nitrobenzene reduction. To further shed light on the diverse catalytic roles of the oxygen functional groups on carbocatalysts, Su and coworkers [144] studied the nitrobenzene reduction as a probe reaction with a series of model molecules. More specifically, 9,10‐anthraquinone, 1,4‐benzenediol, phthalide, benzoic acid, and benzyl ether were used as model catalysts to study the roles of carbonyl, phenol, lactone, carboxylic acid, and ether groups on carbocatalysts, respectively, some of which are depicted in Figure 9.13. Based on the experimental findings, it was concluded that the carbonyl groups act as the active centers for the catalytic reaction. The oxygen in carbonyl groups has an unpaired electron, which interacts with hydrogen atoms of hydrazine molecules, forming hydrogen bonds. As a result, the NH bonds in hydrazine were weakened, which largely facilitated hydrazine decomposition to generate activated hydrogens. The activated hydrogens were reported to be the real reducing species in this reaction rather than diimide as reported by Larsen et al. [135] and Zhou et al. [143]. This reaction mechanism has some similarities to the selective hydrogen transfer mechanism occurring in the reduction of nitrobenzene with isopropanol as the hydrogen donor. In this work, CO groups were also identified as the active centers for the catalytic reaction [145]. Furthermore, depending on the relative positions of the carbonyl groups with respect to one another, the selectivity of the catalytic reaction can be largely changed.

As an example, both 9,10‐anthraquinone and phenanthraquinone have two carbonyl groups (Figure 9.13). However, due to the different relative position and reduction potential of the carbonyl groups, the selectivity of the catalytic reaction changes. In the case of 9,10‐anthraquinone, a nitrobenzene molecule (the reactant) interacts with the catalyst through π–π interactions. The two oxygen atoms of the nitro group abstract two hydrogen atoms, which were activated by one of the carbonyl group. This leads to the formation of nitrosobenzene, which then non‐catalytically converts to aniline. However, in the case of phenanthraquinone, as the two carbonyl groups are present on the same side, four activated hydrogens are available for the interaction with the nitro group, which leads to the generation of hydroxylamine. As its conversion to aniline was slow, large amount of hydroxylamine was accumulated in the final product. These findings provide mechanistic guidance for designing efficient and selective carbon catalysts.

The effect of heteroatom doping into rGO in the catalytic reduction of nitrobenzene was observed by Ma and coworkers [146]. The incorporation of N, P, B into the rGO graphitic network enhanced the catalytic activity compared to rGO. Among the three doping elements, B doping resulted in the highest catalytic activity. As the major focus of this work was to study the microwave adsorption behavior and microwave‐assisted heteroatom doping of graphene materials, the correlation between heteroatom doping induced morphological and electronic structural changes and the observed catalytic activities was not discussed. However, the electronic effect of N doping was clearly illustrated in the selective catalytic reduction of nitrophenol (Nip) to aminophenol by Chen and coworkers [147]. Interestingly, they found that for the N‐doped graphene (NG), the catalytic reaction follows a pseudo‐zero‐order kinetics, which is completely different from all of the reported pseudo‐first‐order reactions catalyzed by metal‐nanoparticle‐based catalysts. Theoretical studies demonstrated that the carbon atoms next to the doped N atoms are active to catch Nip, owing to its weakened conjugation as well as higher positive charge density. Due to the limited catalytic centers, the number of Nip absorbed on NG sheets was not determined by their concentration, but by the number of active sites on NG. Therefore, pseudo‐zero‐order kinetics was observed, which deviates from the transition‐metal‐based catalysis (pseudo‐first‐order kinetics that indicate ample active centers). Compared to nitrobenzene, the nitrophenol has an OH group on ortho position. Based on the in situ FTIR results and theoretical studies, it was suggested that Nip interacts with the catalytic centers via the O atom of the hydroxyl group, but not via the nitro groups, which are far apart from the NG planar surface. In this work, sodium borohydride was used as a hydride donor for the nitrophenol reduction. The reduction mechanism of nitroarenes using sodium borohydride and transition metals has been widely studied. According to the literature, the borohydride reacts and transfers a hydride to the metal catalytic surface, where the adsorbed nitrophenol incorporates the hydride and then undergoes reduction. The catalytic mechanism with NG catalysts was not elucidated, but it was proposed that a similar mechanistic pathway could take place. In this work, it was also mentioned that all types of N configurations exhibited similar adsorption energies. Based on the widely accepted studies on the electronic structures of carbon atoms on a graphene sheet upon N doping with different bonding configuration, this claim might need further studies to clarify. In summary, various carbocatalysts have been used for the reduction of nitroarenes. As described in the above section, the catalytic reaction mechanisms and the selectivity of these processes depend on the reductants used and the catalyst centers.

9.5 Carbon–Carbon Coupling

Metal‐free heterogeneous carbon‐based materials have been used as catalysts for synthetic organic reactions since as early as the 1930s [148]. There are several types of carbon‐based materials used for catalysis, such as graphite, graphene, and CNTs. Graphene oxide has gained increasing interest as a carbocatalyst to promote carbon–carbon bond formation for several reasons: (i) Graphene oxide provides a benign, cheap, and readily available alternative to metal catalysts currently being used in the field. Importantly, graphene oxide combines the benefits of green synthesis with heterogeneous reaction conditions, which greatly simplifies work‐up conditions and is particularly attractive from a practical standpoint [10, 11, 149]. (ii) Catalysis using graphene oxide heavily benefits from the natural abundance of carbon, thereby obviating the challenges associated with limited natural resources of transition metals. (iii) Graphene oxide is an acidic [150] two‐dimensional aromatic scaffold [79] containing hydroxyls, epoxides, carboxylic acids, and occasionally lactone functional groups [151–155], in addition to holes along with hydrophobic regions [88]. As such, graphene oxide merges the slightly acidic (pH of c. 4.5) with strong oxidizing properties, which can be exploited in carbon–carbon bond forming reactions via acidic, isomerization, oxidative, and oxidative coupling pathways. And (iv) graphene oxide matches or even surpasses the efficiency of metal catalysis, while at the same time its structure can be rationally modified by surface modifications and heteroatom doping to customize to specific synthetic applications with chemoselectivity and regioselectivity that is often difficult to achieve using traditional catalysts. Schematic representation of graphene oxide is shown in Figure 9.14.

Recently, graphene oxide has been found to successfully catalyze various types of carbon–carbon bond forming reactions [88, 156–184]. Herein, we will discuss the scope and application of graphene oxide for the formation of carbon–carbon bonds to show distinct ways in which GO can be used as a carbocatalyst to promote synthetically valuable carbon–carbon bond couplings through (i) Friedel–Crafts reactions, (ii) multicomponent reactions (MCRs), (iii) biaryl synthesis, (iv) Michael addition, (v) aldol condensation, and (vi) miscellaneous reactions. It should be noted that while the high efficiency of GO has been demonstrated for various CC bond coupling reactions listed above, different functional groups and catalytic centers on the surface may contribute to the observed reactivity.

9.5.1 Carbon–Carbon Coupling Reactions Catalyzed by Graphene Oxide

9.5.1.1 Friedel–Crafts Reactions

Friedel–Crafts alkylation and acylation reactions represent synthetically attractive processes for arene functionalization that are extensively utilized in both academic and industrial laboratories. Graphite and graphene have been used as effective catalysts for Friedel–Crafts acylation [155, 156] and alkylation with alkyl halides [183]. More recently, there has been an increased interest in using GO as a carbocatalyst for alkylation using unactivated olefins and alcohols (Figure 9.15) [88].

In this reaction, arenes and heteroarenes are alkylated with unactivated olefins and alcohols to furnish diarylmethane products with excellent regioselectivity using acidic sites on GO. There are several ways in which GO can be prepared; in each case, the resulting GO features different properties, such as different quantities of polar functional groups and a variable quantity of holes in its basal plane. During reaction optimization, various types of GO and conditions were tested. The highest yield was observed using GO prepared by a modified Hummer's method in CHCl₃ at 100 °C. Interestingly, application of other graphene‐based materials, such as microwave‐enabled holey GO (h‐GO) and rGO, gave much lower conversions. The scope of the reaction is broad with tolerance for olefins containing neutral, electron‐deficient, and electron‐rich substituents and halides. Moreover, 1,3‐dimethoxybenzene was reacted with dihydronaphthalene and indene, demonstrating that cyclic olefins are suitable reaction partners. Norbornene was also tested to demonstrate that the reaction is not only limited to styrenyl olefins. Likewise, the arene substrate scope includes various electron‐rich, sterically hindered, and even unactivated aromatic hydrocarbons.

The proposed mechanism involves the following steps: (i) anchoring of both coupling partners to the GO surface, (ii) activation of the olefin through hydration with the concomitant reduction of the GO surface and arene prepositioning by π‐stacking interactions, and (iii) concerted CC bond forming step (Figure 9.16). The authors demonstrated that polar functional groups anchored on the GO surface (carboxylic acid, hydroxyl, and epoxide) are critical to the high activity in the reaction. The catalytic cycle is facilitated by the presence of polar and aromatic functional groups on a single surface.

Using the optimized conditions for the coupling of arenes with olefins, He, Szostak, and coworkers demonstrated that alcohols could also serve as suitable coupling partners with excellent efficiency [88]. In this case, the proposed mechanism involves a temporary tethering of an alcohol moiety to the carbocatalyst surface giving a stabilized cation. Interestingly, unconventional selectivity in the coupling was demonstrated in that typically more reactive acetate‐leaving groups were recovered unchanged from the reaction. This opposite selectivity to the classical alkylations highlights the potential of chemoselective manipulation of substrates by graphene oxide. This reaction demonstrates how both aromatic and polar domains present on the GO surface offer a low‐cost, nontoxic alternative to conventional catalysts.

Rama Rao reported graphite‐oxide‐catalyzed Friedel–Crafts reaction of indoles with α,β‐unsaturated ketones and nitrostyrenes (Figure 9.17) [157]. The scope of this transformation includes several functional groups sensitive to transition metal catalysis, such as chloro, bromo, iodo, nitro, and cyano. The authors demonstrated that Fe, Co, Cu, and Pb were below the detection limit in the GO catalyst, whereas Mn was present in less than 30 ppb. Both activated charcoal and graphite showed much lower catalytic activity than graphite oxide. Moreover, common acid catalysts such as p‐TsOH (p‐toluenesulfonic acid), HCl, H₂SO₄, and acidic Al₂O₃, as well as nanocatalysts such as Fe₂O₃, Fe₃O₄, CuO, and MgO, gave lower reactivity than GO.

Acocella et al. reported GO‐catalyzed Friedel–Crafts reaction of indoles with epoxides that occur with high regioselectivity and enantioselectivity (Figure 9.18) [158]. Graphene oxide and exfoliated graphene oxide (e‐GO) were selected for this transformation. The reaction with e‐GO proceeded with similar efficiency and selectivity but in shorter reaction time than with GO, while large‐surface‐area graphite and carbon black were inactive in the reaction. The catalytic activity was assigned to the presence of carboxylic acid and hydroxyl groups on the surface. The reaction showed general scope for the coupling of styrenyl epoxides at the benzylic carbon. Importantly, the rearrangement to the isomeric carbonyl was not observed. A particularly interesting feature of this protocol is complete inversion of stereochemistry using chiral epoxides, green reaction conditions, and the potential to recycle the e‐GO catalyst.

9.5.1.2 Multicomponent Reactions

MCRs are an efficient method to rapidly build up molecular complexity, wherein three or more compounds react to form a single product. Graphene oxide has been used as a catalyst in several examples of MCRs including highly useful Mannich‐type reactions, synthesis of heterocycles, and functionalization of barbituric acids.

Kapoor reported GO‐catalyzed solvent‐free synthesis of 1‐amidoalkyl‐2‐naphthols and 1,2‐dihydro‐1‐arylnaphth[1,2‐e][1,3]oxazin‐3‐ones (Figure 9.19) [159]. These GO‐catalyzed reactions present several advantages, such as high yields, short reaction time, ease of work‐up, and environmentally safe profile. Importantly, the GO catalyst can be easily recovered and reused. In a related development, Shabaani and coworkers reported the synthesis of xanthenes and benzoxanthenes catalyzed by graphene oxide or sulfated graphene nanosheets using the intermolecular coupling between 1‐ or 2‐napthols, 1,3‐diketones, and carbonyl compounds and both catalysts demonstrated recyclability [160].

Khalili demonstrated the four‐component synthesis of 2‐amino‐3‐cyanopyridines in water using GO as a carbocatalyst (Figure 9.20) [161]. The reaction affords medicinally relevant 2‐aminopyridines in a multicomponent coupling of an aldehyde, ketone, malononitrile, and ammonium acetate. As mentioned in the section on Friedel–Crafts reactions, various types of GO can be employed as a carbocatalyst; in this case, the best activity was observed for a GO catalyst prepared by a modified Hummer's method and exfoliation in an aqueous solution. The reaction was optimized using acetophenone, ammonium acetate, 4‐methylbenzaldehyde, and malononitrile. The methodology was then applied with the synthesis of a range of pyridines by modifying either the aldehyde or the ketone component. Aromatic aldehydes including electron‐donating or ‐withdrawing groups readily furnished the coupling products. Moreover, different ketones could be used, including aromatic (4‐methoxy, 4‐chloro‐, and 4‐phenyl), aliphatic, and cyclic. The authors proposed that carboxylic acid groups on the GO are responsible for the reactivity. The catalyst could be reused up to six times while maintaining high activity. In a related multicomponent method, Mirza‐Aghayan et al. reported GO‐catalyzed synthesis of pyridines using methyl 3‐aminocrotonate and various aldehydes (not shown) [162].

Karami group demonstrated that the one‐pot, three‐component coupling between 4‐hydroxycoumarin, aryl glyoxals, and malononitrile was efficiently catalyzed by GO nanosheets (GO‐NSs) to produce pyranocoumarins (Figure 9.21) [163]. In a related process, Das and coworkers have utilized the reaction of 4‐hydroxycoumarin with diverse chalcones catalyzed by GO nanosheets to synthesize 4‐H‐pyrans (not shown) [164]. Karami and coworker also described a mechanistically related, three‐component coupling between barbituric acids (R,R = H,H or Me,Me), 4‐hydroxycoumarin, and a wide range of aryl aldehydes to obtain novel barbituric acid derivatives (Figure 9.22) [184]. The recyclability of the respective GO catalysts in these works was demonstrated. It is likely that in these Knoevenagel‐type condensations, acidic groups on the GO act as catalytically active sites. It should be noted that in contrast to the challenging Friedel–Crafts reaction with unactivated olefins and alcohols (Figure 9.15), these reactions employ activated substrates, which facilitate the coupling.

9.5.1.3 Synthesis of Biaryls

Recently, significant progress has been achieved in the synthesis of biaryls via radical pathways enabled by GO‐based materials. Nishina and coworkers demonstrated the use of GO as a catalyst for the dimerization of anisoles and derivatives, while also showing that the reaction follows a free radical mechanism (Figure 9.23) [165]. The reaction conditions were initially optimized in the oxidative coupling of 3,4‐dimethoxytoluene to the corresponding dimer. It was found that GO in conjunction with BF₃OEt₂ afforded the biaryl product in excellent yield (99%). rGO and activated carbon were unreactive. Other acids, including Fe(OTf)₃, TFA, H₂SO₄, and TfOH, could also be used, however furnished the dimer in lower yields. Importantly, the authors demonstrated that the developed conditions are superior to the use of a common hypervalent iodine reagent, PhI(OAc)₂, and proceed under very mild conditions. Impressively, the substrate scope was shown to include halogen‐containing substrates, which could be used as handles for traditional cross‐coupling reactions. In this case, the yield decreases with more electronegative halogen substituents (Br > Cl > F). The authors also demonstrated trimerization of 1,2‐dimethoxybenzene as well as polymerization of 1,3‐ and 1,4‐dimethoxybenzene under the developed conditions. Two examples of intermolecular cross‐coupling using 2‐methoxynaphthalene were reported. These reactions are particularly promising and highlight the potential of this GO methodology for the synthesis of unsymmetrical biaryls.

Regarding the mechanism, a radical scavenger, TEMPO, inhibited the reaction, thereby suggesting that a radical species may be involved in the hydrogen abstraction. Electron spin resonance (ESR) was performed on GO, and no signal was produced. When 3,4‐dimethoxytoluene and BF₃OEt₂ were analyzed together by ESR, a negligible signal appeared. However, when GO, 3,4‐dimethoxytoluene, and BF₃OEt₂ were analyzed, a signal appeared consistent with a carbon radical. Using FT‐IR, it was found that BF₃OEt₂ activates GO to form a radical species. The proposed mechanism involves radical abstraction from the anisole substrate to form an aryl radical, which then reacts with another molecule of the starting material. In contrast, the reaction using PhI(OAc)₂/BF₃OEt₂ proceeds via a radical cation intermediate. The authors nicely demonstrated the high stability of the radical intermediate in the GO‐catalyzed process by deuterium‐labeling experiments, in which significant scrambling was observed. A related homo‐coupling of 2‐naphthols using oxygen‐doped carbon materials to afford biaryls has been reported [166].

More recently, the Nishina group reported radical coupling between aryldiazonium salts and electron‐rich, five‐membered heterocycles catalyzed by rGO (not shown) [167]. The reaction provides rapid access to 2‐arylfurans, thiophenes, and pyrroles under mild conditions and the rGO catalysts can be reused for several times. Another inventive coupling for the metal‐free synthesis of biaryls involving a free radical mechanism using GO as a catalyst was reported by Wang, Ma, and coworkers (Figure 9.24) [168]. In this method, a biaryl compound is formed by reacting benzene with an aryl iodide or bromide in the presence of a strong base. GO showed much higher activity than CNTs, active carbon, carbon black, and natural graphite. Electron‐rich aryl iodides are more reactive than those containing electron‐neutral or unconjugated substituents. By extensive mechanistic studies and DFT calculations, it was proposed that the negatively charged oxygen atoms on the GO surface promote the coupling by activating K⁺ ions, which then react with the CI bond to give aryl radical. The reaction of this intermediate with benzene followed by proton transfer furnishes the biaryl product. The graphene π system facilitates the coupling as the aromatic coupling partners are easily adsorbed through π–π interactions.

9.5.1.4 Michael Addition

Graphene oxide has been used as a recyclable phase transfer catalyst by the Lee and coworkers (Figure 9.25) [169]. They developed the coupling between 2,4‐pentanedione as a nucleophile and trans‐β‐nitrostyrene as an electrophile as a model system. It was found that the reaction proceeded efficiently within 10 min at room temperature, affording the Michael addition product in up to 83% yield, when using an aqueous GO solution and KOH in methylene chloride. Under the optimized conditions, various nitrostyrenes react with 1,3‐diketones in high yields. Moreover, 1,3‐diesters, 1,3‐ketoesters, and 1,3‐diketones are suitable coupling partners for this reaction. Functional group tolerance includes halogens (Br, Cl, and F), heterocycles, tert‐butyl esters, and cyclic ketones.

Mechanistically, they proposed that the efficiency of GO as a phase transfer catalyst was due to the chelation of cations by polar functional groups on the surface. Polar functional groups act as cation holders, resulting in an increase in the strength of hydroxide anions. Importantly, the catalyst could be recycled up to nine times affording yields ranging from 83% to 76%. The mild reaction conditions do not interfere with the oxygen functional groups on the GO surface, preserving the catalytic activity.

Acocella et al. reported GO and e‐GO as efficient catalysts for Mukaiyama–Michael coupling of 2‐(trimethylsilyloxy)furan with β‐nitroalkenes (Figure 9.26) [170]. In this reaction design, in contrast to traditional catalysts, the anti‐diastereoisomer is obtained with high diastereoselectivity (dr up to 85 : 15). GO and e‐GO exhibited comparable activity and diastereoselectivity in this process (85–90% yield, dr = 75 : 25–77 : 23). Interestingly, graphite oxide after reduction with ascorbic acid, carbon black, and large‐surface‐area graphite also exhibited high activity in the process, albeit resulted in somewhat lower yields of the final product (65–70%), while maintaining high anti‐diastereoselectivity. DFT calculations were performed to investigate the mechanism. The high anti‐selectivity of the process was proposed to arise from a combination of (i) π‐stacking interactions between β‐nitrostyrene and graphite and (ii) van der Waals interaction between the TMS group and graphite.

9.5.1.5 Aldol Condensation

Bielawski and coworkers demonstrated that GO could act as a catalyst for the aldol condensation between acetophenones and BzHs to afford chalcones in high yields (Figure 9.27) [171]. In this report, GO was also used as a tandem oxidation–aldol, tandem hydration–aldol, and tandem oxidation–hydration–aldol catalyst to couple acetophenones with benzylic alcohols, alkynes with aldehydes, and alkynes with benzylic alcohols, respectively. The CC coupling was proposed to proceed mainly due to acidic properties of GO.

Acocella et al. reported that graphite oxide efficiently promotes Mukaiyama aldol reaction between 2‐(trimethylsilyloxy)furan with various BzHs in generally high diastereoselectivity (Figure 9.28) [172]. A practical advantage of this method involves solvent‐free conditions. The authors proposed that the catalytic activity of GO was associated with the presence of carboxylic acid and hydroxyl functional groups on the surface. Activation of aromatic aldehydes by π‐stacking interactions was proposed to explain the low reactivity of aliphatic substrates.

An aldol‐type condensation was reported by Saha, Vlachos, and coworkers in the GO‐catalyzed carbon–carbon bond coupling of 2‐methylfuran (2‐MF) with biomass‐derived aldehydes and ketones (not shown) [173]. The authors found that improved graphene oxide (I‐GO) prepared by a refined Hummer's method and characterized by stronger acidic properties as well as improved hydrophilicity exhibits high activity in the reaction. The coupling between 2‐MF and furfural proceeds in 95% yield at 60 °C under neat conditions, while the coupling of 2‐MF with butanal, acetone, hydroxymethylfurfural, and acetol gave 83–90% yields. The catalyst could be regenerated to regain full activity. Extensive characterization demonstrated that surface oxidation, high degree of Brønsted acidity, and defect sites on the surface and edges are responsible for high catalytic activity in this process.

9.5.1.6 Miscellaneous Reactions

Chauhan reported the use of GO and graphite oxide as catalysts for the synthesis of 5,5‐dialkyldipyrromethanes and calix[4]pyrroles at room temperature (Figure 9.29) [174]. The key reaction involves the condensation of two pyrrole molecules and a ketone, in which GO was proposed to act as a solid acid. Dipyrromethanes were formed in 73–99% yields, while calix[4]pyrroles were obtained in up to 50% yields starting directly from pyrrole.

Mechanistically related condensations were reported by Karami and coworkers using 4‐hydroxycoumarin and aryl glyoxals (Figure 9.30) [175], and Wu and coworkers using indole and aldehydes (Figure 9.31) [176]. In these reactions, GO is proposed to catalyze the coupling as a solid acid catalyst, which can be recycled up to four times without a decrease in efficiency.

Islam et al. reported that graphene oxide catalyzes aldol and Knoevenagel condensations under solvent‐free condition (not shown) [177]. Compared with the reported heterogeneous catalysts, GO exhibited higher activity and could be recycled without a significant loss of efficiency.

Baeza, Alonso, and coworkers reported GO‐catalyzed pinacol rearrangement of 1,2‐diols that afforded α‐quaternary ketones with high selectivity (Figure 9.32) [178]. Graphene oxide was found to be the best catalyst for reaction with rGO, whereas carboxylic‐acid‐functionalized graphene oxide (GO‐CO₂H) and graphite were unproductive in the process. High yield of the rearrangement product was also obtained with graphite oxide containing acidic sulfate groups. Meinwald rearrangement and direct nucleophilic substitution of allylic alcohols were also reported (not shown). These latter reactions were found to be more efficient using GO‐CO₂H as the catalyst to facilitate alcohol activation.

Cid and coworkers developed a new catalyst system for carbon–carbon bond coupling by immobilizing piperazine on rGO support (not shown) [179]. The resulting nucleophilic catalyst was successfully used in Knoevenagel, Michael, and aldol condensations. The catalytic activity was proposed to result from stabilization of positively charged intermediates by the surface.

The Loh group reported the synthesis of α‐aminonitriles utilizing the oxidative properties of highly porous graphene oxide (p‐GO) (Figure 9.33) [180]. In this method, the coupling of aniline with benzylamines affords N‐benzylideneaniline, which is subsequently intercepted by cyanide in a one‐pot process. Related methods for the synthesis of α‐aminonitriles have been published [181, 182]. This process, which could be formally regarded as CH functionalization, holds a significant promise for the development of even more general α‐functionalization methods with unactivated carbon nucleophiles.

9.5.1.7 Conclusion

In this section, we have presented an overview of carbon–carbon couplings catalyzed by graphene oxide. As demonstrated in this review, this field has seen a rapid growth in recent years. In these reactions, GO acts as a catalyst, cocatalyst, or a phase transfer catalyst operating through various reaction pathways. For Friedel–Crafts reactions, it was shown that both polar and aromatic regions are essential for the catalytic activity via isomerization mechanism. In the case of condensation reactions, carboxylic acid and hydroxyl groups on the GO surface are crucial for these reactions to proceed. The biaryl synthesis showcases the capacity of GO to form free radicals as a cocatalyst en route to important synthetic motifs. The Michael addition features the use of oxygen‐containing groups as the main driving force of the mechanism as a phase transfer catalyst. In two of these four classes of reactions, recyclability of GO has been demonstrated. It is worth noting that in the case of Friedel–Crafts reactions and oxidative coupling, the produced rGO can be successfully employed in other applications or variations of the same reactions, thereby enhancing the utilization of GO.

The recent progress in better understanding of the underlying mechanisms of several GO‐catalyzed reactions is noteworthy. The major advantages of using GO as a carbocatalyst for carbon–carbon bond coupling include ready availability, abundance, low price, lack of toxicity, ease of removal, and selectivity complementary to other more established catalysts. The formation of carbon–carbon bonds represents one of the most important processes in organic synthesis. In light of rapid progress in the surface modification and heteroatom doping of graphene‐based materials, we anticipate that more refined GO‐based carbocatalysts will soon find widespread application as heterogeneous catalysts for carbon–carbon coupling in synthetic chemistry.

9.6 Perspective and Future Work

References

1. 1 Hu, H., Xin, J.H., Hu, H. et al. (2015). Metal‐free graphene‐based catalyst‐insight into the catalytic activity: a short review. Appl. Catal., A 492: 1–9.
2. 2 Schaetz, A., Zeltner, M., and Stark, W.J. (2012). Carbon modifications and surfaces for catalytic organic transformations. ACS Catal. 2 (6): 1267–1284.
3. 3 Serp, P., Corrias, M., and Kalck, P. (2003). Carbon nanotubes and nanofibers in catalysis. Appl. Catal., A 253 (2): 337–358.
4. 4 Wang, X., Sun, G., Routh, P. et al. (2014). Heteroatom‐doped graphene materials: syntheses, properties and applications. Chem. Soc. Rev. 43 (20): 7067–7098.
5. 5 Hu, C.G. and Dai, L.M. (2016). Carbon‐based metal‐free catalysts for electrocatalysis beyond the ORR. Angew. Chem. Int. Ed. 55 (39): 11736–11758.
6. 6 Liu, X. and Dai, L.M. (2016). Carbon‐based metal‐free catalysts. Nat. Rev. Mater. 1 (11): 16064.
7. 7 Kong, X.‐K., Chen, C.‐L., and Chen, Q.‐W. (2014). Doped graphene for metal‐free catalysis. Chem. Soc. Rev. 43 (8): 2841–2857.
8. 8 Su, D.S., Perathoner, S., and Centi, G. (2013). Nanocarbons for the development of advanced catalysts. Chem. Rev. 113 (8): 5782–5816.
9. 9 Navalon, S., Dhakshinamoorthy, A., Alvaro, M., and Garcia, H. (2014). Carbocatalysis by graphene‐based materials. Chem. Rev. 114 (12): 6179–6212.
10. 10 Chua, C.K. and Pumera, M. (2015). Carbocatalysis: the state of “metal‐free” catalysis. Chem. Eur. J. 21 (36): 12550–12562.
11. 11 Su, D.S., Wen, G., Wu, S. et al. (2017). Carbocatalysis in liquid‐phase reactions. Angew. Chem. Int. Ed. 56 (4): 936–964.
12. 12 Tang, P., Hu, G., Li, M.Z., and Ma, D. (2016). Graphene‐based metal‐free catalysts for catalytic reactions in the liquid phase. ACS Catal. 6 (10): 6948–6958.
13. 13 Yang, J.H., Yang, D., Tang, P., and Ma, D. (2016). Liquid‐phase heterogeneous catalytic reactions by metal‐free graphene‐based catalysts. Acta Phys. Chim. Sin. 32 (1): 75–84.
14. 14 Navalon, S., Dhakshinamoorthy, A., Alvaro, M. et al. (2017). Active sites on graphene‐based materials as metal‐free catalysts. Chem. Soc. Rev. 46 (15): 4501–4529.
15. 15 Iwasawa, Y., Nobe, H., and Ogasawara, S. (1973). Reaction mechanism for styrene synthesis over polynaphthoquinone. J. Catal. 31: 444–449.
16. 16 Centi, G., Perathoner, S., and Su, D.S. (2014). Nanocarbons: opening new possibilities for nano‐engineered novel catalysts and catalytic electrodes. Catal. Surv. Jpn. 18 (4): 149–163.
17. 17 Su, D.S., Zhang, J., Frank, B. et al. (2010). Metal‐free heterogeneous catalysis for sustainable chemistry. ChemSusChem 3 (2): 169–180.
18. 18 Pereira, M.F.R., Orfao, J.J.M., and Figueiredo, J.L. (1999). Oxidative dehydrogenation of ethylbenzene on activated carbon catalysts. I. Influence of surface chemical groups. Appl. Catal., A 184 (1): 153–160.
19. 19 Pereira, M.F.R., Orfao, J.J.M., and Figueiredo, J.L. (2000). Oxidative dehydrogenation of ethylbenzene on activated carbon catalysts 2. Kinetic modelling. Appl. Catal., A 196 (1): 43–54.
20. 20 Pereira, M.F.R., Orfao, J.J.M., and Figueiredo, J.L. (2001). Oxidative dehydrogenation of ethylbenzene on activated carbon catalysts 3. Catalyst deactivation. Appl. Catal., A 218 (1–2): 307–318.
21. 21 Zhang, J., Su, D.S., Zhang, A.H. et al. (2007). Nanocarbon as robust catalyst: mechanistic insight into carbon‐mediated catalysis. Angew. Chem. Int. Ed. 46 (38): 7319–7323.
22. 22 Mestl, G., Maksimova, N.I., Keller, N. et al. (2001). Carbon nanofilaments in heterogeneous catalysis: an industrial application for new carbon materials? Angew. Chem. Int. Ed. 40 (11): 2066–2068.
23. 23 Qi, W. and Su, D. (2014). Metal‐free carbon catalysts for oxidative dehydrogenation reactions. ACS Catal. 4 (9): 3212–3218.
24. 24 Zhao, Z.K., Ge, G.F., Li, W.Z. et al. (2016). Modulating the microstructure and surface chemistry of carbocatalysts for oxidative and direct dehydrogenation: a review. Chin. J. Catal. 37 (5): 644–670.
25. 25 Zhang, J., Su, D.S., Blume, R. et al. (2010). Surface chemistry and catalytic reactivity of a nanodiamond in the steam‐free dehydrogenation of ethylbenzene. Angew. Chem. Int. Ed. 49 (46): 8640–8644.
26. 26 Pelech, I., Soares, O., Pereira, M.F.R., and Figueiredo, J.L. (2015). Oxidative dehydrogenation of isobutane on carbon xerogel catalysts. Catal. Today 249: 176–183.
27. 27 Rao, R., Yang, M., Ling, Q. et al. (2014). A novel route of enhancing oxidative catalytic activity: hydroxylation of MWCNTs induced by sectional defects. Cat. Sci. Technol. 4 (3): 665–671.
28. 28 Zhang, J., Liu, X., Blume, R. et al. (2008). Surface‐modified carbon nanotubes catalyze oxidative dehydrogenation of n‐butane. Science 322 (5898): 73–77.
29. 29 Frank, B., Zhang, J., Blume, R. et al. (2009). Heteroatoms increase the selectivity in oxidative dehydrogenation reactions on nanocarbons. Angew. Chem. Int. Ed. 48 (37): 6913–6917.
30. 30 Liu, X., Frank, B., Zhang, W. et al. (2011). Carbon‐catalyzed oxidative dehydrogenation of n‐butane: selective site formation during sp³‐to‐sp² lattice rearrangement. Angew. Chem. Int. Ed. 50 (14): 3318–3322.
31. 31 Chen, C.L., Zhang, J., Zhang, B.S. et al. (2013). Revealing the enhanced catalytic activity of nitrogen‐doped carbon nanotubes for oxidative dehydrogenation of propane. Chem. Commun. 49 (74): 8151–8153.
32. 32 Tang, S.B. and Cao, Z.X. (2012). Site‐dependent catalytic activity of graphene oxides towards oxidative dehydrogenation of propane. Phys. Chem. Chem. Phys. 14 (48): 16558–16565.
33. 33 Schwartz, V., Fu, W.J., Tsai, Y.T. et al. (2013). Oxygen‐functionalized few‐layer graphene sheets as active catalysts for oxidative dehydrogenation reactions. ChemSusChem 6 (5): 840–846.
34. 34 Grant, J.T., Carrero, C.A., Goeltl, F. et al. (2016). Selective oxidative dehydrogenation of propane to propene using boron nitride catalysts. Science 354 (6319): 1570–1573.
35. 35 Wang, R., Sun, X.Y., Zhang, B.S. et al. (2014). Hybrid nanocarbon as a catalyst for direct dehydrogenation of propane: formation of an active and selective core–shell sp²/sp³ nanocomposite structure. Chem. Eur. J. 20 (21): 6324–6331.
36. 36 Li, S.M., Wang, W.P., Liu, X. et al. (2016). Nitrogen‐doped graphene nanosheets as metal‐free catalysts for dehydrogenation reaction of ethanol. RSC Adv. 6 (16): 13450–13455.
37. 37 Zhao, Z.K., Dai, Y.T., Ge, G.F., and Wang, G.R. (2015). Explosive decomposition of a melamine‐cyanuric acid supramolecular assembly for fabricating defect‐rich nitrogen‐doped carbon nanotubes with significantly promoted catalysis. Chem. Eur. J. 21 (22): 8004–8009.
38. 38 Zhao, Z.K., Dai, Y.T., Lin, J.H., and Wang, G.R. (2014). Highly‐ordered mesoporous carbon nitride with ultrahigh surface area and pore volume as a superior dehydrogenation catalyst. Chem. Mater. 26 (10): 3151–3161.
39. 39 Sattler, J., Ruiz‐Martinez, J., Santillan‐Jimenez, E., and Weckhuysen, B.M. (2014). Catalytic dehydrogenation of light alkanes on metals and metal oxides. Chem. Rev. 114 (20): 10613–10653.
40. 40 Nawaz, Z. (2015). Light alkane dehydrogenation to light olefin technologies: a comprehensive review. Rev. Chem. Eng. 31 (5): 413–436.
41. 41 Suresh, A.K., Sharma, M.M., and Sridhar, T. (2000). Engineering aspects of industrial liquid‐phase air oxidation of hydrocarbons. Ind. Eng. Chem. Res. 39 (11): 3958–3997.
42. 42 Atamny, F., Blöcker, J., Dübotzky, A. et al. (1992). Surface chemistry of carbon: activation of molecular‐oxygen. Mol. Phys. 76 (4): 851–886.
43. 43 Khorrampour, R., Esrafili, M.D., and Hadipour, N.L. (2009). Density functional theory study of atomic oxygen, O₂ and O₃ adsorptions on the H‐capped (5,0) single‐walled carbon nanotube. Physica E 41 (8): 1373–1378.
44. 44 Luo, J., Yu, H., Wang, H. et al. (2014). Aerobic oxidation of benzyl alcohol to benzaldehyde catalyzed by carbon nanotubes without any promoter. Chem. Eng. J. 240: 434–442.
45. 45 Yang, X.X., Yu, H., Peng, F., and Wang, H.J. (2012). Confined iron nanowires enhance the catalytic activity of carbon nanotubes in the aerobic oxidation of cyclohexane. ChemSusChem 5 (7): 1213–1217.
46. 46 Chaudhari, M.P. and Sawant, S.B. (2005). Kinetics of heterogeneous oxidation of benzyl alcohol with hydrogen peroxide. Chem. Eng. J. 106 (2): 111–118.
47. 47 Liao, S., Peng, F., Yu, H., and Wang, H. (2014). Carbon nanotubes as catalyst for the aerobic oxidation of cumene to cumene hydroperoxide. Appl. Catal., A 478: 1–8.
48. 48 Liao, S., Chi, Y., Yu, H. et al. (2014). Tuning the selectivity in the aerobic oxidation of cumene catalyzed by nitrogen‐doped carbon nanotubes. ChemCatChem 6 (2): 555–560.
49. 49 Engel, P.S., Billups, W.E., Abmayr, D.W. et al. (2008). Reaction of single‐walled carbon nanotubes with organic peroxides. J. Phys. Chem. C 112 (3): 695–700.
50. 50 Zhou, Z.B., He, H.Y., Weng, Z.Y., and Zhao, C.X. (2000). Modification of linear triblock copolymer SBS via olefinic per(poly)fluoroalkylation by per(poly)fluorodiacyl peroxides. J. Fluor. Chem. 104 (2): 285–290.
51. 51 Yoshida, M., Morinaga, Y., Iyoda, M. et al. (1993). Reaction of C₆₀ with diacyl peroxides containing perfluoroalkyl groups – the 1st example of electron‐transfer reaction via C₆₀⁺ in solution. Tetrahedron Lett. 34 (47): 7629–7632.
52. 52 Duffy, D.M. and Blackman, J.A. (1998). Magnetism of 3D transition‐metal adatoms and dimers on graphite. Phys. Rev. B 58 (11): 7443–7449.
53. 53 Banhart, F. (2009). Interactions between metals and carbon nanotubes: at the interface between old and new materials. Nanoscale 1 (2): 201–213.
54. 54 Gao, H.L. and Zhao, J.J. (2010). First‐principles study of Ru atoms and clusters adsorbed outside and inside carbon nanotubes. J. Chem. Phys. 132 (23): 234704.
55. 55 Yu, H., Peng, F., Tan, J. et al. (2011). Selective catalysis of the aerobic oxidation of cyclohexane in the liquid phase by carbon nanotubes. Angew. Chem. Int. Ed. 50 (17): 3978–3982.
56. 56 Yang, X.X., Wang, H.J., Li, J. et al. (2013). Mechanistic insight into the catalytic oxidation of cyclohexane over carbon nanotubes: kinetic and in situ spectroscopic evidence. Chem. Eur. J. 19 (30): 9818–9824.
57. 57 Cao, Y.H., Yu, H., Peng, F., and Wang, H.J. (2014). Selective allylic oxidation of cyclohexene catalyzed by nitrogen‐doped carbon nanotubes. ACS Catal. 4 (5): 1617–1625.
58. 58 Chi, Y.M., Zhu, M.L., Li, Y.H. et al. (2016). The effect of surface oxygenated groups of carbon nanotubes on liquid phase catalytic oxidation of cumene. Catal. Sci. Technol. 6 (7): 2396–2402.
59. 59 Luo, J., Peng, F., Yu, H. et al. (2013). Aerobic liquid‐phase oxidation of ethylbenzene to acetophenone catalyzed by carbon nanotubes. ChemCatChem 5 (6): 1578–1586.
60. 60 Song, Y.J., Qu, K.G., Zhao, C. et al. (2010). Graphene oxide: intrinsic peroxidase catalytic activity and its application to glucose detection. Adv. Mater. 22 (19): 2206–2210.
61. 61 Espinosa, J.C., Navalon, S., Alvaro, M., and Garcia, H. (2016). Reduced graphene oxide as a metal‐free catalyst for the light‐assisted fenton‐like reaction. ChemCatChem 8 (16): 2642–2648.
62. 62 Espinosa, J.C., Navalon, S., Primo, A. et al. (2015). Graphenes as efficient metal‐free fenton catalysts. Chem. Eur. J. 21 (34): 11966–11971.
63. 63 Yang, J.H., Sun, G., Gao, Y.J. et al. (2013). Direct catalytic oxidation of benzene to phenol over metal‐free graphene‐based catalyst. Energy Environ. Sci. 6 (3): 793–798.
64. 64 Wu, P., Cai, Z.W., Gao, Y. et al. (2011). Enhancing the electrochemical reduction of hydrogen peroxide based on nitrogen‐doped graphene for measurement of its releasing process from living cells. Chem. Commun. 47 (40): 11327–11329.
65. 65 Wen, G., Wu, S., Li, B. et al. (2015). Active sites and mechanisms for direct oxidation of benzene to phenol over carbon catalysts. Angew. Chem. Int. Ed. 54 (13): 4105–4109.
66. 66 Collins, P.G., Bradley, K., Ishigami, M., and Zettl, A. (2000). Extreme oxygen sensitivity of electronic properties of carbon nanotubes. Science 287 (5459): 1801–1804.
67. 67 Lee, S.M., Lee, Y.H., Hwang, Y.G. et al. (1999). Defect‐induced oxidation of graphite. Phys. Rev. Lett. 82 (1): 217–220.
68. 68 Allouche, A. and Ferro, Y. (2006). Dissociative adsorption of small molecules at vacancies on the graphite (0001) surface. Carbon 44 (15): 3320–3327.
69. 69 Jiang, D., Sumpter, B.G., and Dai, S. (2007). Unique chemical reactivity of a graphene nanoribbon's zigzag edge. J. Chem. Phys. 126: 134701.
70. 70 Sun, X., Ding, Y., Zhang, B. et al. (2015). Insight into the enhanced selectivity of phosphate‐modified annealed nanodiamond for oxidative dehydrogenation reactions. ACS Catal. 5 (4): 2436–2444.
71. 71 Huang, R., Xu, J.Y., Wang, J. et al. (2016). Oxygen breaks into carbon nanotubes and abstracts hydrogen from propane. Carbon 96: 631–640.
72. 72 Puziy, A.M., Poddubnaya, O.I., and Ziatdinov, A.M. (2006). On the chemical structure of phosphorus compounds in phosphoric acid‐activated carbon. Appl. Surf. Sci. 252 (23): 8036–8038.
73. 73 Boukhvalov, D.W., Osipov, V.Y., Shames, A.I. et al. (2016). Charge transfer and weak bonding between molecular oxygen and graphene zigzag edges at low temperatures. Carbon 107: 810–810.
74. 74 Panich, A.M., Shames, A.I., Tsindlekht, M.I. et al. (2016). Structure and magnetic properties of pristine and Fe‐doped micro‐ and nanographenes. J. Phys. Chem. C 120 (5): 3042–3053.
75. 75 Ritter, K.A. and Lyding, J.W. (2009). The influence of edge structure on the electronic properties of graphene quantum dots and nanoribbons. Nat. Mater. 8 (3): 235–242.
76. 76 Nakada, K., Fujita, K., Dresselhaus, G., and Dresselhaus, M.S. (1996). Edge state in graphene ribbons: nanometer size effect and edge shape dependence. Phys. Rev. B 54: 17954.
77. 77 Joly, V.L.J., Takahara, K., Takai, K. et al. (2010). Effect of electron localization on the edge‐state spins in a disordered network of nanographene sheets. Phys. Rev. B 81: 115408.
78. 78 Shen, A.L., Zou, Y.Q., Wang, Q. et al. (2014). Oxygen reduction reaction in a droplet on graphite: direct evidence that the edge is more active than the basal plane. Angew. Chem. Int. Ed. 53 (40): 10804–10808.
79. 79 Su, C.L., Acik, M., Takai, K. et al. (2012). Probing the catalytic activity of porous graphene oxide and the origin of this behaviour. Nat. Commun. 3 (1): 1298.
80. 80 Gu, Q.Q., Wen, G.D., Ding, Y.X. et al. (2017). Reduced graphene oxide: a metal‐free catalyst for aerobic oxidative desulfurization. Green Chem. 19 (4): 1175–1181.
81. 81 Fan, X., Zhang, G., and Zhang, F. (2015). Multiple roles of graphene in heterogeneous catalysis. Chem. Soc. Rev. 44 (10): 3023–3035.
82. 82 Su, C. and Loh, K.P. (2013). Carbocatalysts: graphene oxide and its derivatives. Acc. Chem. Res. 46 (10): 2275–2285.
83. 83 Dreyer, D.R., Jia, H.‐P., Todd, A.D. et al. (2011). Graphite oxide: a selective and highly efficient oxidant of thiols and sulfides. Org. Biomol. Chem. 9 (21): 7292–7295.
84. 84 Jia, H.‐P., Dreyer, D.R., and Bielawski, C.W. (2011). C–H oxidation using graphite oxide. Tetrahedron 67 (24): 4431–4434.
85. 85 Dreyer, D.R., Jia, H.‐P., and Bielawski, C.W. (2010). Graphene oxide: a convenient carbocatalyst for facilitating oxidation and hydration reactions. Angew. Chem. Int. Ed. 49 (38): 6813–6816.
86. 86 Dreyer, D.R. and Bielawski, C.W. (2012). Graphite oxide as an olefin polymerization carbocatalyst: applications in electrochemical double layer capacitors. Adv. Funct. Mater. 22 (15): 3247–3253.
87. 87 Boukhvalov, D.W., Dreyer, D.R., Bielawski, C.W., and Son, Y.‐W. (2012). A computational investigation of the catalytic properties of graphene oxide: exploring mechanisms by using DFT methods. ChemCatChem 4 (11): 1844–1849.
88. 88 Hu, F., Patel, M., Luo, F. et al. (2015). Graphene‐catalyzed direct Friedel–Crafts alkylation reactions: mechanism, selectivity, and synthetic utility. J. Am. Chem. Soc. 137 (45): 14473–14480.
89. 89 Patel, M.A., Luo, F., Khoshi, M.R. et al. (2016). P‐doped porous carbon as metal free catalysts for selective aerobic oxidation with an unexpected mechanism. ACS Nano 10 (2): 2305–2315.
90. 90 Pattisson, S., Nowicka, E., Gupta, U.N. et al. (2016). Tuning graphitic oxide for initiator‐ and metal‐free aerobic epoxidation of linear alkenes. Nat. Commun. 7: 12855.
91. 91 Hummers, W.S. and Offeman, R.E. (1958). Preparation of graphitic oxide. J. Am. Chem. Soc. 80: 1339.
92. 92 Hofmann, U. and Konig, E. (1937). Untersuchungen uber graphitoxyd. Z. Anorg. Allg. Chem. 234: 311.
93. 93 Patel, M., Savaram, K., Keating, K., and He, H. (2015). Rapid transformation of biomass compounds to metal free catalysts via short microwave irradiation. J. Nat. Prod. Res. Updat. 1: 18–28.
94. 94 Long, J., Xie, X., Xu, J. et al. (2012). Nitrogen‐doped graphene nanosheets as metal‐free catalysts for aerobic selective oxidation of benzylic alcohols. ACS Catal. 2 (4): 622–631.
95. 95 Watanabe, H., Asano, S., Fujita, S. et al. (2015). Nitrogen‐doped, metal‐free activated carbon catalysts for aerobic oxidation of alcohols. ACS Catal. 5 (5): 2886–2894.
96. 96 Gong, K.P., Du, F., Xia, Z.H. et al. (2009). Nitrogen‐doped carbon nanotube arrays with high electrocatalytic activity for oxygen reduction. Science 323 (5915): 760–764.
97. 97 Gao, Y., Hu, G., Zhong, J. et al. (2013). Nitrogen‐doped sp²‐hybridized carbon as a superior catalyst for selective oxidation. Angew. Chem. Int. Ed. 52 (7): 2109–2113.
98. 98 Li, W., Gao, Y., Chen, W. et al. (2014). Catalytic epoxidation reaction over N‐containing sp² carbon catalysts. ACS Catal. 4 (5): 1261–1266.
99. 99 Petit, T., Arnault, J.C., Girard, H.A. et al. (2012). Oxygen hole doping of nanodiamond. Nanoscale 4 (21): 6792–6799.
100. 100 Lin, Y.M. and Su, D.S. (2014). Fabrication of nitrogen‐modified annealed nanodiamond with improved catalytic activity. ACS Nano 8 (8): 7823–7833.
101. 101 Wang, Y., Zhang, J.S., Wang, X.C. et al. (2010). Boron‐ and fluorine‐containing mesoporous carbon nitride polymers: metal‐free catalysts for cyclohexane oxidation. Angew. Chem. Int. Ed. 49 (19): 3356–3359.
102. 102 Dhakshinamoorthy, A., Primo, A., Concepcion, P. et al. (2013). Doped graphene as a metal‐free carbocatalyst for the selective aerobic oxidation of benzylic hydrocarbons, cyclooctane and styrene. Chem. Eur. J. 19 (23): 7547–7554.
103. 103 Some, S., Kim, J., Lee, K. et al. (2012). Highly air‐stable phosphorus‐doped n‐type graphene field‐effect transistors. Adv. Mater. 24 (40): 5481–5486.
104. 104 Qiao, X., Liao, S., You, C., and Chen, R. (2015). Phosphorus and nitrogen dual doped and simultaneously reduced graphene oxide with high surface area as efficient metal‐free electrocatalyst for oxygen reduction. Catalysts 5 (2): 981–991.
105. 105 Sun, X.Y., Xu, J.Y., Ding, Y.X. et al. (2015). The effect of different phosphorus chemical states on an onion‐like carbon surface for the oxygen reduction reaction. ChemSusChem 8 (17): 2872–2876.
106. 106 Liu, Z.‐W., Peng, F., Wang, H.‐J. et al. (2011). Phosphorus‐doped graphite layers with high electrocatalytic activity for the O₂ reduction in an alkaline medium. Angew. Chem. Int. Ed. 50 (14): 3257–3261.
107. 107 Latorre‐Sanchez, M., Primo, A., and Garcia, H. (2013). P‐doped graphene obtained by pyrolysis of modified alginate as a photocatalyst for hydrogen generation from water–methanol mixtures. Angew. Chem. Int. Ed. 52 (45): 11813–11816.
108. 108 Chen, D., Holmen, A., Sui, Z., and Zhou, X. (2014). Carbon mediated catalysis: a review on oxidative dehydrogenation. Chin. J. Catal. 35 (6): 824–841.
109. 109 Yang, Z., Yao, Z., Li, G. et al. (2012). Sulfur‐doped graphene as an efficient metal‐free cathode catalyst for oxygen reduction. ACS Nano 6 (1): 205–211.
110. 110 El‐Sawy, A.M., Mosa, I.M., Su, D. et al. (2016). Controlling the active sites of sulfur‐doped carbon nanotube‐graphene nanolobes for highly efficient oxygen evolution and reduction catalysis. Adv. Energy Mater. 6 (5): 1501966.
111. 111 Dhakshinamoorthy, A., Latorre‐Sanchez, M., Asiri, A.M. et al. (2015). Sulphur‐doped graphene as metal‐free carbocatalysts for the solventless aerobic oxidation of styrenes. Catal. Commun. 65: 10–13.
112. 112 Meng, Y., Voiry, D., Goswami, A. et al. (2014). N‐, O‐, and S‐tridoped nanoporous carbons as selective catalysts for oxygen reduction and alcohol oxidation reactions. J. Am. Chem. Soc. 136 (39): 13554–13557.
113. 113 Patel, M.A., Luo, F., Savaram, K. et al. (2017). P and S dual‐doped graphitic porous carbon for aerobic oxidation reactions: enhanced catalytic activity and catalytic sites. Carbon 114: 383–392.
114. 114 Keren, E. and Soffer, A. (1977). Simultaneous electronic and ionic surface conduction of catalyst supports: a general mechanism for spillover: the role of water in the Pd‐catalyzed hydrogenation of a carbon surface. J. Catal. 50: 43–45.
115. 115 Zhang, Z.G., Okada, K., Yamamoto, M., and Yoshida, T. (1998). Hydrogenation of anthracene over active carbon‐supported nickel catalyst. Catal. Today 45 (1–4): 361–366.
116. 116 Zhang, Z.G. and Yoshida, T. (2001). Behavior of hydrogen transfer in the hydrogenation of anthracene over activated carbon. Energy Fuels 15 (3): 708–713.
117. 117 Sun, L.B., Zong, Z.M., Kou, J.H. et al. (2004). Activated carbon‐catalyzed hydrogenation of polycyclic arenes. Energy Fuels 18 (5): 1500–1504.
118. 118 Sun, L.B., Wei, X.Y., Liu, X.Q. et al. (2009). Selective hydrogen transfer to anthracene and its derivatives over an activated carbon. Energy Fuels 23: 4877–4882.
119. 119 Wei, X.Y., Ogata, E., and Niki, E. (1992). FeS₂‐catalyzed hydrocracking of alpha, omega‐diarylalkanes. Bull. Chem. Soc. Jpn. 65 (4): 1114–1119.
120. 120 Schimmel, H.G., Kearley, G.J., Nijkamp, M.G. et al. (2003). Hydrogen adsorption in carbon nanostructures: comparison of nanotubes, fibers, and coals. Chem. Eur. J. 9 (19): 4764–4770.
121. 121 Ulbricht, H., Moos, G., and Hertel, T. (2003). Interaction of molecular oxygen with single‐wall carbon nanotube bundles and graphite. Surf. Sci., Proc. Pap. 532: 852–856.
122. 122 Welch, G.C., Juan, R.R.S., Masuda, J.D., and Stephan, D.W. (2006). Reversible, metal‐free hydrogen activation. Science 314 (5802): 1124–1126.
123. 123 Stephan, D.W. and Erker, G. (2015). Frustrated Lewis pair chemistry: development and perspectives. Angew. Chem. Int. Ed. 54 (22): 6400–6441.
124. 124 Primo, A., Neatu, F., Florea, M. et al. (2014). Graphenes in the absence of metals as carbocatalysts for selective acetylene hydrogenation and alkene hydrogenation. Nat. Commun. 5: 5291.
125. 125 Trandafir, M.M., Florea, M., Neatu, F. et al. (2016). Graphene from alginate pyrolysis as a metal‐free catalyst for hydrogenation of nitro compounds. ChemSusChem 9 (13): 1565–1569.
126. 126 Li, B., Sun, X.Y., and Su, D.S. (2015). Calibration of the basic strength of the nitrogen groups on the nanostructured carbon materials. Phys. Chem. Chem. Phys. 17 (10): 6691–6694.
127. 127 Sun, X., Li, B., Liu, T. et al. (2016). Designing graphene as a new frustrated Lewis pair catalyst for hydrogen activation by co‐doping. Phys. Chem. Chem. Phys. 18 (16): 11120–11124. doi: 10.1039/C5CP07969A.
128. 128 Li, B. and Xu, Z. (2009). A nonmetal catalyst for molecular hydrogen activation with comparable catalytic hydrogenation capability to noble metal catalyst. J. Am. Chem. Soc. 131 (45): 16380–16382.
129. 129 Pacosová, L., Kartusch, C., Kukula, P., and van Bokhoven, J.A. (2011). Is fullerene a nonmetal catalyst in the hydrogenation of nitrobenzene? ChemCatChem 3 (1): 154–156.
130. 130 Perhun, T.I., Bychko, I.B., Trypolsky, A.I., and Strizhak, P.E. (2013). Catalytic properties of graphene material in the hydrogenation of ethylene. Theor. Exp. Chem. 48 (6): 367–370.
131. 131 Chen, P.R., Chew, L.M., Kostka, A. et al. (2013). Purified oxygen‐ and nitrogen‐modified multi‐walled carbon nanotubes as metal‐free catalysts for selective olefin hydrogenation. J. Energy Chem. 22 (2): 312–320.
132. 132 Liu, R., Li, F., Chen, C. et al. (2017). Nitrogen‐functionalized reduced graphene oxide as carbocatalysts with enhanced activity for polyaromatic hydrocarbon hydrogenation. Cat. Sci. Technol. 7 (5): 1217–1226. doi: 10.1039/C7CY00058H.
133. 133 Liao, T., Sun, C.H., Sun, Z.Q. et al. (2013). Chemically modified ribbon edge stimulated H₂ dissociation: a first‐principles computational study. Phys. Chem. Chem. Phys. 15 (21): 8054–8057.
134. 134 Han, B.H., Shin, D.H., and Cho, S.Y. (1985). Graphite catalyzed reduction of aromatic and aliphatic nitro‐compounds with hydrazine hydrate. Tetrahedron Lett. 26 (50): 6233–6234.
135. 135 Larsen, J.W., Freund, M., Kim, K.Y. et al. (2000). Mechanism of the carbon catalyzed reduction of nitrobenzene by hydrazine. Carbon 38 (5): 655–661.
136. 136 Austin, J.M., Groenewald, T., and Spiro, M. (1980). Heterogeneous catalysis in solution. Part 18. The catalysis by carbons of oxidation–reduction reactions. J. Chem. Soc., Dalton Trans. (6): 854–859.
137. 137 Gao, Y., Ma, D., Wang, C. et al. (2011). Reduced graphene oxide as a catalyst for hydrogenation of nitrobenzene at room temperature. Chem. Commun. 47 (8): 2432–2434.
138. 138 Bonanni, A., Ambrosi, A., and Pumera, M. (2012). On oxygen‐containing groups in chemically modified graphenes. Chem. Eur. J. 18 (15): 4541–4548.
139. 139 Ambrosi, A., Chua, C.K., Bonanni, A., and Pumera, M. (2012). Lithium aluminum hydride as reducing agent for chemically reduced graphene oxides. Chem. Mater. 24 (12): 2292–2298.
140. 140 Li, D., Muller, M.B., Gilje, S. et al. (2008). Processable aqueous dispersions of graphene nanosheets. Nat. Nanotechnol. 3 (2): 101–105.
141. 141 Wu, S., Wen, G., Zhong, B. et al. (2014). Reduction of nitrobenzene catalyzed by carbon materials. Chin. J. Catal. 35 (6): 914–921.
142. 142 Wu, S.C., Wen, G.D., Schlogl, R., and Su, D.S. (2015). Carbon nanotubes oxidized by a green method as efficient metal‐free catalysts for nitroarene reduction. Phys. Chem. Chem. Phys. 17 (3): 1567–1571.
143. 143 Zhou, H.Y., Shi, L., and Sun, Q. (2012). Reduction of nitrobenzene with hydrazine hydrate catalyzed by acid‐treated activated carbon. Chin. J. Catal. 33 (9): 1463–1469.
144. 144 Wu, S., Wen, G., Liu, X. et al. (2014). Model molecules with oxygenated groups catalyze the reduction of nitrobenzene: insight into carbocatalysis. ChemCatChem 6 (6): 1558–1561.
145. 145 Yang, H., Cui, X., Dai, X. et al. (2015). Carbon‐catalysed reductive hydrogen atom transfer reactions. Nat. Commun. 6: 6478.
146. 146 Tang, P., Hu, G., Gao, Y. et al. (2014). The microwave adsorption behavior and microwave‐assisted heteroatoms doping of graphene‐based nano‐carbon materials. Sci. Rep. 4: 5901.
147. 147 Kong, X.‐K., Sun, Z.‐Y., Chen, M. et al. (2013). Metal‐free catalytic reduction of 4‐nitrophenol to 4‐aminophenol by N‐doped graphene. Energy Environ. Sci. 6 (11): 3260–3266. doi: 10.1039/C3EE40918J.
148. 148 Dreyer, D.R. and Bielawski, C.W. (2011). Carbocatalysis: heterogeneous carbons finding utility in synthetic chemistry. Chem. Sci. 2 (7): 1233–1240.
149. 149 Dhakshinamoorthy, A. and Garcia, H. (2014). Metal–organic frameworks as solid catalysts for the synthesis of nitrogen‐containing heterocycles. Chem. Soc. Rev. 43 (16): 5750–5765.
150. 150 Dimiev, A.M., Alemany, L.B., and Tour, J.M. (2013). Graphene oxide. Origin of acidity, its instability in water, and a new dynamic structural model. ACS Nano 7 (1): 576–588.
151. 151 Lerf, A., He, H.Y., Forster, M., and Klinowski, J. (1998). Structure of graphite oxide revisited. J. Phys. Chem. B 102 (23): 4477–4482.
152. 152 He, H., Riedl, T., Lerf, A., and Klinowski, J. (1996). Solid‐state NMR studies of the structure of graphite oxide. J. Phys. Chem. 100 (51): 19954–19958.
153. 153 Lerf, A., He, H., Riedl, T. et al. (1997). ¹³C and ¹H MAS NMR studies of graphite oxide and its chemically modified derivatives. Solid State Ionics 101: 857–862.
154. 154 Cai, W.W., Piner, R.D., Stadermann, F.J. et al. (2008). Synthesis and solid‐state NMR structural characterization of ¹³C‐labeled graphite oxide. Science 321 (5897): 1815–1817.
155. 155 Kodomari, M., Suzuki, Y., and Yoshida, K. (1997). Graphite as an effective catalyst for Friedel–Crafts acylation. Chem. Commun. (16): 1567–1568. doi: 10.1039/A703947F.
156. 156 Chua, C.K. and Pumera, M. (2012). Friedel–crafts acylation on graphene. Chem. Asian J. 7 (5): 1009–1012.
157. 157 Vijay Kumar, A. and Rama Rao, K. (2011). Recyclable graphite oxide catalyzed friedel–crafts addition of indoles to α,β‐unsaturated ketones. Tetrahedron Lett. 52 (40): 5188–5191.
158. 158 Acocella, M.R., Mauro, M., and Guerra, G. (2014). Regio‐ and enantioselective Friedel–Crafts reactions of indoles to epoxides catalyzed by graphene oxide: a green approach. ChemSusChem 7 (12): 3279–3283.
159. 159 Gupta, A., Kour, D., Gupta, V.K., and Kapoor, K.K. (2016). Graphene oxide mediated solvent‐free three component reaction for the synthesis of 1‐amidoalkyl‐2‐naphthols and 1,2‐dihydro‐1‐arylnaphth[1,2‐e][1,3]oxazin‐3‐ones. Tetrahedron Lett. 57 (43): 4869–4872.
160. 160 Shaabani, A., Mahyari, M., and Hajishaabanha, F. (2014). The synthesis of xanthenes and benzoxanthenes on graphene oxide and sulfated graphene nanosheets in water. Res. Chem. Intermed. 40 (8): 2799–2810.
161. 161 Khalili, D. (2016). Graphene oxide: a reusable and metal‐free carbocatalyst for the one‐pot synthesis of 2‐amino‐3‐cyanopyridines in water. Tetrahedron Lett. 57 (15): 1721–1723.
162. 162 Mirza‐Aghayan, M., Asadi, F., and Boukherroub, R. (2014). Graphite oxide‐promoted one‐pot synthesis and oxidative aromatization of Hantzsch 1, 4‐dihydropyridines. Monatsh. Chem. Chem. Mon. 145 (12): 1919–1924.
163. 163 Khodabakhshi, S. and Karami, B. (2014). Graphene oxide nanosheets as metal‐free catalysts in the three‐component reactions based on aryl glyoxals to generate novel pyranocoumarins. New J. Chem. 38 (8): 3586–3590.
164. 164 Kausar, N., Ghosh, P.P., Pal, G., and Das, A.R. (2015). Graphene oxide nanosheets: a highly efficient and reusable carbocatalyst catalyzes the Michael‐cyclization reactions of 4‐hydroxycoumarins, 4‐hydroxypyrone and 4‐hydroxy‐1‐methylquinolinone with chalcone derivatives in aqueous media. RSC Adv. 5 (74): 60199–60207.
165. 165 Morioku, K., Morimoto, N., Takeuchi, Y., and Nishina, Y. (2016). Concurrent formation of carbon–carbon bonds and functionalized graphene by oxidative carbon–hydrogen coupling reaction. Sci. Rep. 6: 25824.
166. 166 Periasamy, M., Shanmugaraja, M., Reddy, P.O. et al. (2017). Synthetic transformations using molecular oxygen‐doped carbon materials. J. Org. Chem. 82 (9): 4944–4948.
167. 167 Morimoto, N., Morioku, K., Suzuki, H. et al. (2017). Carbocatalytic reductive coupling reactions via electron transfer from graphene to aryldiazonium salt. Chem. Commun. 53 (53): 7226–7229.
168. 168 Gao, Y., Tang, P., Zhou, H. et al. (2016). Graphene oxide catalyzed C−H bond activation: the importance of oxygen functional groups for biaryl construction. Angew. Chem. Int. Ed. 55 (9): 3124–3128.
169. 169 Kim, Y., Some, S., and Lee, H. (2013). Graphene oxide as a recyclable phase transfer catalyst. Chem. Commun. 49 (50): 5702–5704.
170. 170 Acocella, M.R., Mauro, M., Falivene, L. et al. (2014). Inverting the diastereoselectivity of the Mukaiyama–Michael addition with graphite‐based catalysts. ACS Catal. 4 (2): 492–496.
171. 171 Jia, H.P., Dreyer, D.R., and Bielawski, C.W. (2011). Graphite oxide as an auto‐tandem oxidation–hydration–aldol coupling catalyst. Adv. Synth. Catal. 353 (4): 528–532.
172. 172 Acocella, M.R., De Pascale, M., Maggio, M., and Guerra, G. (2015). Graphite oxide as catalyst for diastereoselective Mukaiyama aldol reaction of 2‐(trimethylsilyloxy) furan in solvent free conditions. J. Mol. Catal. A: Chem. 408: 237–241.
173. 173 Dutta, S., Bohre, A., Zheng, W. et al. (2017). Solventless C–C coupling of low carbon furanics to high carbon fuel precursors using an improved graphene oxide carbocatalyst. ACS Catal. 7 (6): 3905–3915.
174. 174 Singh Chauhan, S.M. and Mishra, S. (2011). Use of graphite oxide and graphene oxide as catalysts in the synthesis of dipyrromethane and calix[4]pyrrole. Molecules 16 (9): 7256–7266.
175. 175 Khodabakhshi, S., Karami, B., Eskandari, K. et al. (2014). Graphene oxide nanosheets promoted regioselective and green synthesis of new dicoumarols. RSC Adv. 4 (34): 17891–17895.
176. 176 Wang, Y., Sang, R., Zheng, Y. et al. (2017). Graphene oxide: an efficient recyclable solid acid for the synthesis of bis(indolyl)methanes from aldehydes and indoles in water. Catal. Commun. 89: 138–142.
177. 177 Islam, S.M., Roy, A.S., Dey, R.C., and Paul, S. (2014). Graphene based material as a base catalyst for solvent free aldol condensation and Knoevenagel reaction at room temperature. J. Mol. Catal. A: Chem. 394: 66–73.
178. 178 Gómez‐Martínez, M., Baeza, A., and Alonso, D.A. (2017). Pinacol rearrangement and direct nucleophilic substitution of allylic alcohols promoted by graphene oxide and graphene oxide CO₂H. ChemCatChem 9 (6): 1032–1039.
179. 179 Rodrigo, E., Alcubilla, B.G., Sainz, R. et al. (2014). Reduced graphene oxide supported piperazine in aminocatalysis. Chem. Commun. 50 (47): 6270–6273.
180. 180 Su, C., Tandiana, R., Balapanuru, J. et al. (2015). Tandem catalysis of amines using porous graphene oxide. J. Am. Chem. Soc. 137 (2): 685–690.
181. 181 Pan, Y., Wang, S., Kee, C.W. et al. (2011). Graphene oxide and rose Bengal: oxidative C–H functionalisation of tertiary amines using visible light. Green Chem. 13 (12): 3341–3344.
182. 182 Sengupta, A., Su, C., Bao, C. et al. (2014). Graphene oxide and its functionalized derivatives as carbocatalysts in the multicomponent Strecker reaction of ketones. ChemCatChem 6 (9): 2507–2511.
183. 183 Sereda, G.A. (2004). Alkylation on graphite in the absence of Lewis acids. Tetrahedron Lett. 45 (39): 7265–7267.
184. 184 Eskandari, K. and Karami, B. (2016). Graphene oxide nanosheets‐catalyzed synthesis of novel benzylbarbiturocoumarin derivatives under green conditions. Monatsh. Chem. Chem. Mon. 147 (12): 2119–2126.
185. 185 Schiros, T., Nordlund, D., Palova, L. et al. (2012). Connecting dopant bond type with electronic structure in N‐doped graphene. Nano Lett. 12 (8): 4025–4031.
186. 186 Jeyaraj, V.S., Kamaraj, M., and Subramanian, V. (2015). Generalized reaction mechanism for the selective aerobic oxidation of aryl and alkyl alcohols over nitrogen‐doped graphene. J. Phys. Chem. C 119 (47): 26438–26450.
187. 187 Ampelli, C., Perathoner, S., and Centi, G. (2014). Carbon‐based catalysts: opening new scenario to develop next‐generation nano‐engineered catalytic materials. Chin. J. Catal. 35 (6): 783–791.
188. 188 Chen, X.X. and Chen, B.L. (2016). Direct observation, molecular structure, and location of oxidation debris on graphene oxide nanosheets. Environ. Sci. Technol. 50 (16): 8568–8577.
189. 189 Li, X.‐H., Chen, J.‐S., Wang, X. et al. (2011). Metal‐free activation of dioxygen by graphene/g‐C₃N₄ nanocomposites: functional dyads for selective oxidation of saturated hydrocarbons. J. Am. Chem. Soc. 133 (21): 8074–8077.
190. 190 Deng, D., Novoselov, K.S., Fu, Q. et al. (2016). Catalysis with two‐dimensional materials and their heterostructures. Nat. Nanotechnol. 11 (3): 218–230.