Templating synthesis of porous carbons for energy-related applications: A review

GUAN Lu, HU Han, TENG Xiao-ling, ZHU Yi-fan, ZHANG Yun-long,CHAO Hui-xia, YANG Hao, WANG Xiao-shan, WU Ming-bo

（College of Chemical Engineering, College of New Energy, State Key Laboratory of Heavy Oil Processing,China University of Petroleum (East China), Qingdao 266580, China）

Abstract： Because of their large specific surface area, high chemical and thermal stability and good electrical conductivity, porous carbons have found wide applications in the fields of electrochemical energy storage and conversion.Their performance hinges heavily on their structure, making the structural control of porous carbons a research frontier in their development.In addition to the straightforward hard-templating processes, soft templating synthesis is considered another appealing strategy for the precise engineering of porous carbons.We review recent progress on synthesizing porous carbon materials for energy storage and conversion using templating processes.First, the rise of this method of preparing porous carbons is outlined by comparing it with the traditional hard templating methods.Soft templating methods are then classified into top-down, state-change and bottom-up templates based on the template formation processes.The performance of these materials in electrochemical energy storage and conversion is presented,highlighting the advantages of this synthesis method.Finally, possible obstacles and future prospects are provided.

Key words: Porous carbons；In-situ templates；Top-down；State-change；Bottom-up

1 Introduction

The highly tunable structure, as well as a myriad of excellent properties, makes porous carbons the choice for many current and emerged applications, for example, electrochemical energy storage and conversion[1-11].The most widely available porous carbons are activated carbons that are typically produced by carbonization and chemical or physical activation of biomass, polymer, heavy by-products from the petroleum refinery, etc[12-14].Generally, activated carbon shows a large surface area and high porosity and could accommodate a large number of molecules,ions, and other species.The pore size of the activated carbon is randomly distributed and the pore channels are always tortuous.In this context, the pore volume of the activated carbon could not be fully utilized and the diffusion of the adsorbed species shows a low efficiency, usually resulting in unsatisfied performance.To overcome this bottleneck of activated carbons in practical applications, the precise regulation of the porous structure of carbon materials has been highly recognized.Since the pioneering work of simultaneous engineering pore size and channel using ordered mesoporous silica, hard templating strategies have been widely employed to precisely maneuver the porous structures of carbon materials with remarkable progress[15-19].In a typical process, the hard templates are first controllably synthesized with the desired structures.Then, the precursors of the carbon materials are infiltrated into the pore of the hard templates to replicate their morphologies.The porous carbons with controlled structures are then obtained after the template removal[20-24].

Despite the straightforwardness, the hard templating synthesis of porous carbons encounters some drawbacks that heavily impede the wide and commercial application of this technology.The most obvious drawback is the tedious and laborious template construction and removal, making the hard templating synthesis cost-ineffective.In addition, the structure of the hard templates could only be engineered at the mi-cro-/nanoscale and more precise manipulation of the porous carbons at molecular and atomic scale remains challenging via the hard templating processes.A certain type of hard template could only be compatible with limited carbon sources that restrict the wide use of this technology[19].

Recently, the emerged in-situ templating methods offer excellent opportunities to overcome the aforementioned drawbacks.In contrast to the tedious hard templating synthesis of porous carbons, the insitu templating synthesis is realized by simply carbonizing the mixtures of carbon sources and the precursor of templates[25].In some cases, the template removal is unnecessary as the template could be self-decomposed during the formation of carbon materials.In addition, the precursors of the templates are usually small molecules and salts, making the molecular and/or atomic-level structural regulation possible[26,27].In addition to structural engineering, intrinsic defects and heteroatom doping, both of which are highly demanded for improved electrochemical energy storage and conversion, could also be manipulated during the in-situ templating processes.As a result, the in-situ templating synthesis represents a more versatile technology for the controlled construction of porous carbon materials with targeted structures.

Herein, we summarize the recent progress of regulating porous carbon via the in-situ templating processes.First, the in-situ templating methods are separately introduced based on the template formation processes including top-down, state-change, and bottomup.The effects of each template are elaborately discussed.Then, the focused attention is concentrated on the application of these porous carbons for energy storage and conversion, highlighting the structural merits of these materials.After that, the challenges and perspectives of this rising technology are analyzed, aiming to provide the readers an insight view of future directions in this prospective field.

2 In-situ templating synthesis of porous carbons

In the hard templating synthesis, the precursors infiltrated into the templates are generally converted into carbon via a solid-state carbonization process[28].To uniformly replicate the morphologies of the templates, their interaction with the precursors plays a key role.The hydrogen bonds, complexation effect, and so on are generally required.As for the in-situ templating synthesis, the templates are usually formed through decomposition, melt, and polymerization at elevated temperatures.Meanwhile, the raw materials of the porous carbons are melted and interact with insitu formed templates to produce carbon materials with different pore architectures[26,29-32].In the following section, the in-situ templates will be introduced based on their formation processes including topdown, state-change, and bottom-up.

2.1 Top-down templates

Top-down templates usually include a series of compounds that can decompose or be converted into thermally stable components with specific morphology and structure at elevated temperatures.Then, the carbon sources in the molten state or gas phase could be introduced into the as-formed templates to produce carbon through the thermal polymerization processes[21,33-35].In some cases, the templates could catalyze the thermal polymerization reactions, leading to the regulation of local crystallinity.

The most widely used templates are the thermally unstable chemical compounds including citrates, oxalates, basic carbonates, and acetates.Taking citrate as an example, Yang et al.prepared a carbon skeleton with a 3D structure, large interlayer spacing,and high conductivity by directly calcining sodium citrate[36].The interlayer distance and conductivity of 3D framework carbon (3DFC) could be adjusted by controlling the calcination temperatures (600, 700,800 and 1 000 °C).During the calcination of sodium citrate, the disappearance of crystal water, the formation of sodium carbonate, volatile hydrocarbons, and tar will sequentially occur.The in-situ formed sodium carbonate could act as a catalyst and hard template to promote the conversion of the organic compounds into a carbon nanosheet-assembled network as shown in Fig.1a.

In a similar attempt, Sevilla et al.synthesized highly porous carbon nanosheets by one-step carbonization of potassium citrate[37].Potassium citrate was firstly transformed into K2CO3(＜ 650 °C) and then decomposed into CO2and K2O at higher temperatures.The as-produced CO2could react with the in-situ formed carbon to create micropores.When rising the temperatures above 720 °C (the boiling point of metal K), K2O would react with C to form K vapor, a highly active species to regulate the microstructures of carbon materials.The desert rose-like morphology of the as-produced carbon materials (CK-850) shown in Fig.1b prevents the aggregation of flat carbon nanosheets, thus improving the ion diffusion rate.Despite the simplicity of this process, the carbon yield is very low that inevitably increases the cost of the asproduced carbon materials.To raise the carbon yield of the citrate-involved technology, we introduced petroleum asphalt as the extra carbon source[38].At elevated temperatures, asphalt will melt and uniformly mix with the template precursors to produce a nanosheet-assembled hierarchical carbon architecture(NHCA) (Fig.1c).The 3D hierarchically porous structure (Fig.1d) could be well maintained while the carbon yield is increased by twenty-fold (Fig.1e)compared to potassium citrate pyrolyzed alone(HCA).

Basic carbonates are another typical precursors serving as the in-situ template, among which basic magnesium carbonates are the most explored one.One key reason for their popularity is their facile synthesis processes and tunable morphologies.The basic magnesium carbonates will decompose into porous MgO templates with designed structures and guide the formation of the novel hierarchical carbon nanocages(hCNC), as illustrated in Fig.1f-i[39].Besides, oxalatesoffer a similar potential for this purpose.In a mixed solvent of ethanol and water, zinc ions and oxalate ions would electrostatically interact and then self-assemble with sodium lignosulfonate.Then embedded zinc oxalate could serve as the template and guide the lignosulfonate species to be converted into porous carbon nanosheets, as illustrated in Fig.1j[40].

Shao and co-workers took advantage of acetates for this purpose to prepare a series of porous carbons using various carbon precursors including resorcinol resin, polyvinyl alcohol, melamine resin, and coal tar pitch.The precursors of the templates, such as Mg(AC)2, would decompose into porous MgO according to the following reactions:

The carbon yield of this method is quite high and the as-obtained porous carbons generally possess large specific surface areas, hierarchical porosity and abundant oxygen functional groups, making them excellent electrode materials for supercapacitors[41].

In addition to the in-situ formed metal oxides, the nanostructured metals formed during the carbon formation could also serve as the templates.Traditionally,some metals are excellent catalysts to regulate the graphitic degree of carbon materials[42,43].Moreover,the morphology of metal could be regulated through the interaction of carbon precursors and metal ions.Through regulating the coordination and catalytic graphitization effect, Wang et al.realized the mass production of highly crystalline graphene nanosheets using Fe2+ions exchanged polyacrylic weak-acid resin (AC resin) (Fig.2a)[44].In a typical process, the AC resin is immersed into a FeCl2solution via which the Fe2+ions could compactly organize along the AC chain through the coordination effect.Due to the confinement of the AC chains, the reduced Fe species are organized into a two-dimensional structure and catalytic conversion of highly graphitized graphene during the carbonation process.Zhang et al.proposed the direct carbonization of the mixture of iron pentacarbonyl and ethanol where the in-situ formed Fe nanoparticles could contribute to the production of porous carbon with highly graphitic pore walls[45].Considering the excellent capability of Cu foils for catalytic growth of high-quality graphene nanosheets, Zhao et al.mixed basic copper carbonate with polymethyl methacrylate and annealed the mixture at elevated temperatures[46].The in-situ formed nanoporous Cu induces the formation of a core-shell architecture with a wide variety of open pores possessing interconnected micro-meso-macropores in the shell or core (Fig.2bg).

2.2 State-change templates

State-change templates are those salts which can melt at high temperature and effectively adjust the carbonization process of organic precursors to prepare the porous carbon materials with the desired structure[33].The fluid molten salts could mix with the carbon precursors at molecular and/or atomic levels,allowing the precise engineering of the defects, porosity, and surface chemistry of the as-generated carbon materials.Moreover, the salts can be easily recycled via a simple water washing process and the carbon materials are then collected.As a result, the molten salt-mediated synthesis could essentially reduce the cost of carbon production.The unique reaction medi-ums created by molten salts are highly tunable as the molten temperature and viscosity could be adjusted by changing the composition of the salts.As a typical kind of in-situ template, the general process of molten salt-mediated production of porous carbon materials is illustrated in Fig.3a[47].The carbon precursor and the salt were blended and then underwent the thermal treatment at a certain temperature in an inert atmosphere during which the salt could be melted.The carbon source dissolved in the molten salt is then induced to polymerize and convert into porous carbons.In addition to facilitate the polymerization process, the molten salt can also play the role of the in-situ template to engineer the porous structures.Finally, the porous carbons are obtained via a simple water washing process, and the recycled salt is ready for another use[48].

Wu’s team produced a series of porous carbon materials via the molten-salt technology by using petroleum asphalt, a low-cost by-product from petroleum refinery, as the carbon precursors[49-53].For example, a eutectic mixture of KCl and CaCl2with the melting point of 600 °C was adopted to create a reaction medium through which ultra-thin carbon nanosheets could be produced.Instead, irregular particles were obtained by exclusively carbonizing petroleum asphalt.Additionally, the carbon yield in the molten salt-mediated synthesis process is 10 wt%higher than that of direct carbonization of asphalt due to the strong affinity of molten salt towards organic species, allowing more precursors to be converted.Moreover, the structure of the as-obtained carbon materials is highly tunable, from porous carbon particles to thick carbon nanoplates and ultrathin carbon nanosheets through gradually increasing the heating temperatures, as shown in Fig.3b[52].

Another promising feature of the molten salt-mediated synthesis is the possibility of introducing other doping agents to regulate the composition of the porous carbon.In an attempt, Wu’s team proposed the incorporation of melamine to produce nitrogen-doped porous carbon nanosheets[53].At the melting temperature, the asphalt could be thermally polymerized and the melamine is decomposed to produce active nitrogen species for nitrogen-doping, leading to the production of nitrogen-doped ultrathin carbon nanosheets as illustrated in Fig.3c.The nitrogen-doped two-dimensional carbon structure can effectively improve the amorphous degree of asphalt-derived carbon, with increased interlayer spacing, high disorder degree and effective active surface, and further provides more active sites and defects.

In addition, other low eutectic salts, such asLiCl―KCl, NaCl―KCl, ZnCl2―KCl, NaOH―KOH, NaNO3―KNO3, Na2SO4―K2SO4, Li2SO4―Na2SO4, etc., could serve as the in-situ templates to regulate the porous structures of carbon materials.Because of the diverse choice of eutectic salts, a myriad of carbon materials with versatile structures could be obtained through simply regulating the composition of the eutectic systems[49,54-60].

2.3 Bottom-up templates

The in-situ template generated from the bottomup approaches represents the process by which small molecules are converted into nano/micro-structures.Among them, graphitic-carbon nitride (g-C3N4) are the most representative one, which is generally produced through the thermal polycondensation reactions of melamine, dicyandiamide, urea, etc[61-63].The in-situ produced g-C3N4possesses a two-dimensional morphology that will direct the formation of carbon nanosheets.Moreover, the g-C3N4could be self-decomposed at elevated temperatures, producing the active nitrogen species to dope and etch the as-formed carbon nanosheets[64].In this manner, the carbon nanosheets are generally highly porous and doped,both of which are highly demanded for electrochemical energy storage and conversion[65].Meanwhile, the structure of g-C3N4is also highly tunable, and the porous structures could be generated through the introduction of sacrificial agents, for example (NH4)2S2O8,which provide a reliable strategy to more precisely regulating the structure of carbon nanosheets.

Xie and Wang et al.proposed a low-cost and convenient route to fabricate nitrogen-doped porous carbon nanosheets using coal tar pitch and melamine/urea as the carbon and template precursors,respectively.The typical synthesis processes were presented in Fig.4a, e.Firstly, the coal tar pitch,melamine and KOH were ground and mixed uniformly, which were then calcinated in a nitrogen atmosphere.At 380-400 °C, KOH melts and reacts with aromatics hydrocarbons.With the temperature rising to 300-500 °C, melamine converges and condenses into g-C3N4nanosheets (Fig.4b, c) as a self-sacrificing template.Meanwhile, the coal tar pitch melts and covers the surface of g-C3N4to form a layered carbon skeleton.The further increase of the temperature to 800 or 900 °C could result in the complete decomposition of g-C3N4to produce layered microporous carbons (LMPCs) (Fig.4d)[66].The only difference in Wang et al.‘s approach is that the additional time of KOH, and they all end up with nitrogen-doped porous carbon nanosheets[65].

Because of the high tunability of the g-C3N4involved synthesis, other functional components could be simultaneously incorporated to produce the porous carbon nanosheet-based composites[67].Liu et al.have demonstrated the simultaneous introduction of vana-dium nitride during the synthesis of nitrogen-doped carbon nanosheets using dicyandiamide, NH4VO3,glucose as the precursors of template, nanostructured VN, and carbon, respectively[68].First, dicyandiamide,glucose, and NH4VO3were mixed and calcined in the nitrogen atmosphere.

At 600 °C, the dicyandiamide was thermally condensed to produce g-C3N4nanosheets, which could guide the formation of two-dimensional carbon nanosheets and facilitate the uniform dispersion of VN nanoparticles.By increasing the temperature to 800 °C, the in-situ formed g-C3N4templates were decomposed, leaving the VN decorated N-doped carbon nanosheets behind.Zhang and coworkers produced the Co nanoparticles/Co, N, S tri-doped graphene with urea, sucrose, Co(NO3)2, and sulfur as raw materials(Fig.4f)[62].Concretely, the first three raw materials were uniformly mixed and calcined in sulfur vapor at 600 °C (SSUCo-p) for a while and then further heated at 900 °C (SSUCo-900) in inert atmosphere to produce the desired product.

Another typical example of the bottom-up template is the in-situ formed ice, which is used to prepare a large variety of porous materials.The ice crystals formed when freezing serve as the templates which are removed via a subsequent freeze-drying to produce the porous network.The molecules, colloidal particles, and gels in the solution, suspension, and hydrogels will be excluded from the in-situ formed ice for templating.Tamon et al.have systematically conducted ice-templating synthesis of porous carbon networks using resorcinol-formaldehyde resin as the precursors[69-72].By comparing the porous structures prepared under different drying conditions, it is found that the shrinkage during hot-air drying is much larger than freeze drying.Therefore, freeze drying is more conducive to maintain the gel structure comparing with hot-air drying and vacuum drying (Fig.5a).The adsorption and desorption isotherms of N2on carbon aerogel at 77 K indicates that the micropore volume of cryogels is higher than that of aerogels(Fig.5b).So after the ice-templating, freeze-drying,and pyrolysis, porous carbons with specific surface area larger than 800 m2g?1and mesoporous volume larger than 0.55 cm3g?1could be obtained.The frozen solvents show a strong impact on the mesoporosity.By replacing the water in the gel with t-butanol, the porous carbons with a higher content of mesopores could be produced.The resorcinol-formaldehyde resin could be replaced with other polymers containing heteroatoms, such as nitrogen and sulfur, to produce heteroatom-doped porous carbon networks.For example, the use of melamine-formaldehyde resin is capable of producing nitrogen-doped porous carbon.The organic gel formed through the reaction between resorcinol and 2-thiophenecarboxaldehyde has been proposed to produce sulfur-doped carbons.In addition, CNTs and graphene nanosheets could be assembled into monolithic structures using the ice-templating processes.The dispersions of CNTs and graphene are frozen where the uniformly dispersed subunits are concentrated among boundaries of the insitu formed ice crystal as shown in Fig.5a[73].Through controlling the freezing process, the porous structurescan be regulated and the elaborately engineered porous structure could be highly compressible, and the monolithic structures could remain stable without collapse even after thousands of cycles[74].Chen et al.controlled the size and shape of ice crystals to determine the porosity of graphene aerogels through two freezing methods (slow cooling in the refrigerator and rapid cooling in liquid nitrogen)[75].The samples frozen in liquid nitrogen underwent a large temperature gradient to promote the rapid nucleation growth of ice, resulting in a large number of small ice crystals and isotropic structural ice templates (graphene foam).In contrast, the slower freezing resulted in a large and anisotropic ice template that allowed the nanosheets in the hydrogel to be arranged into macroporous structures with several micrometers in thickness, named graphene sponges.Moreover, the ice-templating process is highly versatile and could be compatible with other technology, providing an extra possibility to regulate the porous carbon.Liao et al.coupled the electrospraying and the ice-templating of the graphene oxide dispersion to produce the graphene-based microspheres[76].Barg et al.proposed that the combination of wet spinning and ice-templating, the as-spun fibers possess ordered cellular pores[77].

3 Applications

3.1 Porous carbons for energy storage

3.1.1 Lithium-ion batteries

Lithium-ion batteries (LIBs), featuring high energy densities, have aroused widespread attentions in our increasingly electrified society[78-82].Carbon materials are commonly used as anode materials for LIBs.However, the theoretical capacity of current commercial graphite is approaching its theoretical limit(372 mAh g?1), which has denied the further performance improvement with the state-of-the-art electrode materials.Therefore, the search for a new generation of LIBs electrode materials is of paramount importance.Because of the structural merit of porous carbon materials including large specific surface area, high physical and chemical stability, and high conductivity,they offer great potential as the anode materials for LIBs[16,83,84].

Wang et al.prepared carbon nanomaterials using petroleum asphalt as a carbon source through a statechange templating process[51].Typically, the low eutectic combination of NaCl/KCl was mixed with asphalt in toluene.After the solvent removal, the dark brown powder was obtained and then carbonized in a tubular furnace at 800 °C in inert atmosphere for 2 h.Finally, the samples named MSC were obtained after the salt templates were removed by water washing(Fig.6a).When employed as the anode materials of LIBs, as shown in Fig.6b, a reversible specific capacity of 729 mAh g?1can be obtained at 100 mA g?1.In addition, the rate performance (280 mAh g?1at 5 A g?1), and cyclability is also promising compared with the directly carbonized petroleum asphalt (PAC)as illustrate in Fig.6c, d.

Recently, defect engineering has also been taken into account as an powerful strategy to promote the electrochemical performance of electrode materials[85,86].The vacancies and/or edges defects of carbon materials offer improved affinity towards lithium ions, thereby essentially boosting the lithium storage performance[85,87-90].Niu et al.presented a low-cost and green method to prepare porous carbon (PC) with high defect density by carbonization of cattle bone,primarily composed of hydroxyapatite (Cax(PO4,CO3)y(OH) and collagen, at different temperatures,generating CaO and releasing CO2, CO, etc.[91].The as-obtained porous carbon (PC-1100) exhibits excellent performance at various current densities due to its unique pore structure and rich defect sites (Fig.6e, f).The reversible capacities of PC-1100 were 1 230 and 281 mAh g?1at 1 and 30 A g?1, respectively, and maintains great performance and high coulomb efficiency even after long cycling (Fig.6g).

When the ravens had thus conversed29 they fled onward30, but Trusty John had taken it all in, and was sad and depressed31 from that time forward; for if he were silent to his master concerning what he had heard, he would involve him in misfortune; but if he took him into his confidence, then he himself would forfeit32 his life

3.1.2 Lithium-sulfur batteries

In addition to develop advanced electrode materials with high lithium storage capability, an alternative strategy is a search for newly configurated electrochemical energy devices, among which lithium-sulfur batteries are deemed as the most promising candidates[92].The key advantages of Lithium-sulfur (Li-S)batteries are the high theoretical specific capacity, approaching to 1 672 mAh g?1, and wide availability of the cathode material, namely sulfur[93-95].However,the electronic insulation of sulfur species, large volume expansion of electrode, and shuttle effect of the intermediate species seriously deny the possibility of fully exploring the theoretical potential.Great efforts have been made to design suitable structures to alleviate the aforementioned issues[96-103].Long et al.synthesized a hierarchical porous carbon (HPC) containing micropores, mesopores, and macropores by an in-situ templating process with lithium citrate as a carbon source and template precursor simultaneously[48].The collaborative strength of hierarchically porous structure and high electronic conductivity render the excellent stability and outstanding rate performance as the host for active sulfur.However, it cannot be ignored that too high specific surface area is not conducive to the initial Coulombic efficiency, so it is crucial to control the amount of template.

Recently, defect-rich carbon materials have been demonstrated to effectively accelerate a series of electrochemical reactions by virtue of a stronger polarity than heteroatom-doped carbon materials[104,105].In this regard, Guan and co-workers reported a hierarchically porous carbon material (DHPC) with abundant intrinsic defect, which affords strong adsorption and catalytic conversion capabilities for polysulfides, essentially improving the electrochemical performances of lithium-sulfur batteries[106].Using 1,10-phenanthroline and potassium hydroxide as raw materials,honeycomb nitrogen-doped porous carbon structures(NHPCs) were prepared by in-situ template technology (Fig.7a), and then the NHPCs were annealed at higher temperatures to introduce the intrinsic defects,namely DHPCs.After loading sulfur, the as-prepared electrodes exhibit a high specific capacity of 1 182 mAh g?1at a current density of 0.5 C (Fig.7b,c), improved rate performance and outstanding cycling stability with a low capacity decay rate of 0.06%per cycle (Fig.7d).Lyu et al.devised an in-situ MgO template method with self-structured 3D layered 4MgCO3·Mg(OH)2·5H2O as the precursor, which can be converted into 3D layered MgO templates after thermal decomposition[39,107].The final hierarchical carbon nanocages (hCNC) are composed of the interconnected cuboidal hollow nanocages as shown in Fig.7e.The practical application demonstrated that such a structure is suitable for anchoring and catalytic conversion of lithium polysulfide.The sulfur loading-within this structure is as high as 79.8 wt.%, and the composites show a discharge capacity of 1 214 mAh g?1at the current density of 0.2 A g?1, indicating a high utilization of sulfur (Fig.7f, g).Moreover, the S@hCNC cathode achieved excellent cycling performance at different current rates.when evaluated at 1 A g?1, a discharge specific capacity of 558 mAh g?1was still maintained after 300 cycles with a high coulombic efficiency (Fig.7h).

3.1.3 Supercapacitors

Supercapacitors representing another widely investigated energy storage devices feature high power output and ultralong cycle lifespan[108,109].They have gathered widespread attention and become the choice of power sources for intelligent electronic equipment,standby power supply, and hybrid electric vehicle, etc.Basically, the electrode materials have crucial influences on the performance of supercapacitors, and an ideal electrode material for electronic double layer capacitor (EDLC) should meet the following four requirements simultaneously: (I) large specific surface area (SSA) to offer ample space for charge adsorption,(II) uniform pore size distribution to facilitate the electrolyte transport, (III) high conductivity to assure high rate capacity and power density, (IV) high wettability to promote ion diffusion and increase the surface area accessibility[46,110-115].Basically, the porous carbon nanomaterials could be rationally designed to fulfill the aforementioned issues.Zhao’s team reported a variety of in-situ template methods to prepare carbon materials, and investigated their capability as electrode materials for supercapacitors[116].Typically,benzene was used as the precursor to prepare carbon nanocage (CNC) in one step by using the in-situ MgO template method.The size of the MgO particles could be manipulated through tuning the temperatures, so as the SSAs of the carbon materials as illustrated in Fig.8a-d.The electrochemical investigation suggested that CNC700 offered the optimal specific capacitance (251 F g?1) and its CV curves exhibited approximately rectangular shapes without distinct redox peaks even at the scan rate of 1 000 mV s?1(Fig.8e, f),because of its high electrical conductivity, maximum specific surface area, and even pore size distribution[117].In addition, there are numerous examples of preparing porous carbon materials through in-situ templating process[118,119].Ma et al.converted the silicon in the palm shell into molten salt and employed it as the template to synthesize porous carbon.When used as the electrode material for supercapacitors theas-obtained porous carbons showed excellent electrochemical properties with a specific capacitance of 326 F g?1at a current density of 0.5 A g?1[120].Pampel et al.achieved the customization of specific surface area and pore size distribution by simply adjusting the molar composition of molten salt KCl/ZnCl2mixture as the porogen[121].The premise of pore formation is that the precursor can be uniformly dispersed in the molten salt.The low KCl region of the KCl/ZnCl2phase diagram reveals relatively low melting points,which is necessary for stable dispersion/dissolution(Fig.8g).They found that the increase of KCl content resulted in the gradual opening of pores and the increase of mesoporous volume, leading to the increase of mass transport porosity along with a linear decrease of specific surface area (Fig.8h).The electrochemical performance of the supercapacitor was tested in 1 mol L?1H2SO4electrolyte, and it was found that the higher the KCl content (the sample names are Glu XX, where XX represents the molar amount of KCl in KCl / ZnCl2mixture), the lower the gravimetric capacity and the higher the areal capacity (Fig.8i,j), which was attributed to the opening of the pores,providing improved accessibility and usage of the active sites.

The combination of in-situ template carbonization and activation is a reliable method to prepare carbon materials with an ideal hierarchical porous structure and high specific surface area[122].For instance,the in-situ nano oxide templates can produce mesoporous and/or macropores, and the activation process secures plenty of micropores[41,123].Fu and his team reported a unique porous lignin-derived carbon quasinanosheets (PLC) by gas-exfoliation and in-situ templating carbonization technique with ZnO nanoparticles as the template, which exhibits a remarkable specific capacitance of 320 F g?1at 1 A g?1[40].Shao et al.proposed a dual in-situ template method including MgO and CaCO3to prepare carbon materials with high specific surface area (805-1 525 m2g?1) andwell-developed pore structure, which endow the asprepared materials with remarkable capacitive performance[41].

3.2 Porous carbons for energy conversion

Oxygen reduction reactions (ORRs) are the cornerstone for many electrochemical energy conversion processes, such as fuel cells, metal-air batteries,and so on[124,125].The cutting-edge catalysts for ORR are mainly nanostructured noble metals.Despite the excellent performance, their large-scale applications remain painfully stagnant owing to the low reserves and expensive prices.Therefore, the search for earthabundant materials to promote the ORR has become an important research hotspot[126].Due to the advantages including high specific surface area, highly tunable structure, high chemical stability, and outstanding electrical conductivity, various carbon nanomaterials, especially those doped with heteroatoms, have demonstrated excellent capability for ORR[127].

One key process to promote the catalytic activity is to design and synthesize hierarchical porous structures.Basically, micropores are conducive to exposing a large number of active sites, while macropores and mesopores are favorable for the fast transport of oxygen and electrolyte[128-131].

The ORR catalytic activity of carbon-based materials is also closely relevant to the nitrogen doping level, which have a significant effects on the number of active sites, the electrical conductivity and electrophilicity.Li reported the large-scale preparation of nitrogen-doped carbon nanosheets by pyrolysis of dicyandiamide (DCDA) and glucose mixture[61].The asobtained oxygen-containing nitrogen-doped porous carbon nanosheets (ONC) (Fig.9a) possessed abundant active sites and defects, which displayed excellent electrochemical performance, including better tolerance to methanol than commercial Pt/C catalyst(Fig.9b).ONC samples also possessed long-term durability, which maintaining 98.8% performance after 1 000 cycles (Fig.9c).Besides, the synergistic effect between nitrogen and non-noble metals can be used to enhance the reaction performance.Because of the similar radius and higher electronegativity compared to carbon, nitrogen atoms could serve as the nucleation and anchoring sites for metals[129].Liu proposed the preparation of Co/N co-doped mesoporous carbon nanomaterials by in-situ self-templating method, in which metal salts are used as templates and additives,and the amount of Co doping is controlled to realize the transformation of nanocarbon from polymer microspheres to polymer nanosheets (Fig.9d)[132].Furthermore, the doped Co, the special structure and the large specific surface area synergistically reinforced the catalytic activity and improve the ORR efficiency.As depicted in Fig.9e, the electrocatalytic activity of Co/N-CLPC is far superior to other catalysts in terms of onset potential (E0) of 0.805 V (vs RHE) and halfwave potential (E1/2) of 0.686 V (vs RHE).Currenttime (I-t) chronoamperometric measurement shows that the current of Co/N-CLPC remains unchanged with the addition of methanol into the electrolyte, indicating its prominent methanol tolerance (Fig.9f).This is primarily due to the following three points: (I)Metal nanoparticles improve electrocatalytic activity through electronic interaction with carbon lattice.(II)The doping of nitrogen increases the degree of graphitization.(III) The large specific surface area and unique pore size distribution of the catalysts provide abundant active sites for the ORR reaction and increase the catalytic activity.Fe/N/S co-doped hierarchically porous carbon (FeNS/HPC) was prepared by an in-situ generated dual-template strategy(Fig.9g)[129].In the preparation process, NaCl crystallites were first formed by means of the freeze-drying technology and employed as the template to generate the graphene-like carbon layer.With the pyrolysis of sucrose, thiourea, and FeCl3·6H2O, the Fe3O4nanoparticles generated in-situ produced mesopores on the macropore walls as the secondary template (Fig.9h,i).The special porous structure resulted in a large specific surface area of 938 m2g?1, making the as-made catalyst possess rich active sites.As shown in Fig.9j,k, the FeNS/HPC catalyst exhibited much more positive half-wave potential and onset potential than those of NS/HPC in acid and alkaline medium.

3.2.2 Hydrogen evolution reaction

As the energy carrier, hydrogen processes a wealth of advantages including abundant resources,environmental benign, and wide application[133-135].After years of research, there are several ways to produce hydrogen at present: (I) Hydrogen production from fossil fuels, (II) Hydrogen production from solar energy, (III) Biological hydrogen production, (IV)Hydrogen production by electrolysis of water.Hydrogen production from fossil fuels is a widely used method, and the raw materials are mainly coal and petroleum.Unfortunately, the reserves of fossil fuels are restricted, and the hydrogen production process produces a large amount of carbon and nitrogen oxides that pollute the environment.Hydrogen production by solar energy and biological are newly developed hydrogen production technologies.Although they have the advantages of energy-saving and clean,they still have problems of low conversion efficiency and small hydrogen production.At present, the technology of hydrogen production by electrolysis of water has been widely recognized among all the technologies of converting renewable resources into hydrogen energy, which is simple, pollution-free and efficient[136,137].Among them, Hydrogen Evolution Reaction (HER) is a key step during water electrolysis, so it is very important to develop highly active hydrogen evolution catalysts.The most efficient HER catalysts are as yet platinum-based catalysts, which can perform hydrogen evolution reactions at a potential very close to the electromotive force of the thermodynamic reaction.However, metal platinum resources arelimited and expensive, so it is of paramount importance to explore low-cost and high-efficiency electrocatalytic materials for hydrogen evolution.

A variety of transition metals have been used in electrochemical HER, such as Co, Ni, Fe, Mo, and the corresponding derivatives[138-141].In recent years,many researchers have combined transition metalbased catalysts with carbon materials, which can vigorously enhance the electrocatalytic activity of HER.Yu et al.developed a molten salt strategy, using various salts (such as KCl/LiCl, CaCl2/ZnCl2, and KCl/NaCl, etc.) as templates, to prepare Co9S8nanoparticles (NPs) and N, S co-doped mesoporous carbons for HER[142].The specific method is shown in Fig.10a.CoCl2, thiourea, glucose, and salt were mixed in a certain mass ratio and ground uniformly,then heating in inert atmosphere at elevated temperatures.The products obtained after washing and drying with deionized water are named GTCoTemp-Salts,where Temp and Salt stand for the carbonization temperature and the type of molten salt, respectively.These two factors are the most important to determine the porosity of the electrocatalysts.Using a threeelectrode system to study the electrocatalytic performance of the entire pH value on HER, it is found that GTCo900-KCl/NaCl possesses the best HER activity,providing a current density of 10 mA cm?2in an acid medium of 54 mV, a neutral medium of 142 mV and an alkaline medium of 103 mV (Fig.10b-d).

Currently, most effective HER electrocatalysts are based on transition metals, but they are inherently corrosion sensitive to acidic proton exchange membrane electrolysis.Some carbon materials with twodimensional ordered structures, for example graphene,possess good electrical conductivity, adjustable molecular structure, and strong resistance to acid/alkali environments, all of which make up the drawbacks of metals and their compounds electrocatalysts, so they could serve as excellent alternative to the non-metal catalysts in the HER.Zheng et al.presented a method of coupling GO and g-C3N4derived from dicyandiamide to synthesize a metal-free catalyst containing only carbon and nitrogen, which showed the electrocatalytic HER activity comparable to that of current metal catalysts.Therefore, this provided a new idea for searching for catalysts that could replace precise metals[143].

4 Conclusion

Porous carbon materials play vital and diverse roles in solving global energy and environmental problems by virtue of various overwhelming advantages including high electronic and ionic conductivity,high physical and chemical stability, high gas-liquid permeability, and so on.In the fields of energy storage and conversion, heterogeneous catalysis and water treatment, preparation of hierarchically porous carbon is generally considered as an effective approach offering superior performance.Meanwhile, economic yet facile preparation methods are urgently needed to achieve the large-scale application of porous carbon materials.The template method is the most widely used, and the traditional template method mainly includes the following three processes: (1) preparation of templates, (2) carbonization at high temperature,(3) template removal.In recent years, a more effective and convenient templating technology has emerged, namely in-situ template method, in which the template formed simultaneously during the carbon production processes, and, in some cases, could be directly decomposed.

In this review, the in-situ templates are categorized into three types based on their formation processes: (1) Top-down templates: chemicals could decompose or be converted into thermally stable compounds with special morphologies as templates at elevated temperatures.Then, the carbon sources in the molten state or gas phase could be introduced into the as-formed templates to produce carbon through the thermal polymerization processes.(2) State-change templates: the fluid molten salts could mix with the carbon precursors at molecular and/or atomic levels,allowing the precise engineering of the defects, porosity, and surface chemistry of the as-generated carbon materials.(3) Bottom-up templates, which is a process of transforming small molecules into nano/microstructures, where in-situ formation of g-C3N4and ice are the most representative examples.

For specified applications, the specific surface area and pore structure characteristics should be rationally combined.From the viewpoint of real applications, the precursors should be cheap enough and the process is required to be simple.Although the in-situ template method is controllable and multifunctional and has huge potential in the preparation of three-dimensional porous carbon nanomaterials with tunable specific surface area, large porosity, high mass transfer ability, and high conductivity, there are still some obstacles hindering their practical applications.(1) In some cases, the templates need to be removed using corrosive solutions which not only cause serious pollution to the environment but also increase the production cost.(2) Water-soluble salt template and ice template can be removed easily but their controllability is a little bit of low.

It is highly believed that overcoming the abovementioned issues could promote the development of low-cost hierarchically porous carbons for a wealth of applications.In addition, the elaborate regulation of the process could contribute to highly tunable structures, leading to an excellent platform to investigate the fundamental processes for energy storage and conversion.By rational combination of the aforementioned two aspects, the customized preparation of porous carbon for desired application at low cost could come true and promote the essential progress of this rising field.

Acknowledgements

This work was financially supported by National Natural Science Foundation of China (22138013,22179145), Shandong Provincial Natural Science Foundation (ZR2020ZD08), and the startup support grant from China University of Petroleum (East China).