Nanoscale Hybrid Amorphous/Graphitic Carbon as Key Towards Next‐Generation Carbon‐Based Oxidative Dehydrogenation Catalysts

Abstract A new strategy affords “non‐nano” carbon materials as dehydrogenation catalysts that perform similarly to nanocarbons. Polymer‐based carbon precursors that combine a soft‐template approach with ion adsorption and catalytic graphitization are key to this synthesis strategy, thus offering control over macroscopic shape, texture, and crystallinity and resulting in a hybrid amorphous/graphitic carbon after pyrolysis. From this intermediate the active carbon catalyst is prepared by removing the amorphous parts of the hybrid carbon materials via selective oxidation. The oxidative dehydrogenation of ethanol was chosen as test reaction, which shows that fine‐tuning the synthesis of the new carbon catalysts allows to obtain a catalytic material with an attractive high selectivity (82 %) similar to a carbon nanotube reference, while achieving 10 times higher space–time yields at 330 °C. This new class of carbon materials is accessible via a technically scalable, reproducible synthetic pathway and exhibits spherical particles with diameters around 100 μm, allowing unproblematic handling similar to classic non‐nano catalysts.


Introduction
Catalysts are key materials in modern society that enable selective transformation of raw materials into valuable products while avoiding waste and saving energy. [1] In case of industrially relevant oxidative dehydrogenation reactions (ODH), most known catalyst systems are based on transition metals (e.g. Fe,V ,M o, Ag). [2] Due to the drawbacks associated with the use of transition metals,s uch as rare occurrence,e nvironmentally harmful mining procedures and toxicity,itishighly interesting that pure carbon was shown to exhibit catalytic activity in these types of reactions and could be as ustainable substitution. [3,4] Ty pical reactions reported with carbon catalysts are the gas-phase oxidative dehydrogenations of ethane, [5] propane, [6] butane, [7] and ethylbenzene, [8] which are dehydrogenated to the corresponding olefins. Beyond alkanes,a lcohols like 1-propanol and ethanol have been oxidized to the corresponding aldehydes using carbon catalysts. [9] Up to date,the development of carbon-based catalysts for oxidative dehydrogenation reactions may be divided into two generations.F irst-generation carbon catalysts were inspired by the discovery of the catalytic activity of coke deposits on metal-based catalysts for the oxidative dehydrogenation of ethylbenzene. [3] In this context, mainly amorphous carbonaceous materials,s uch as activated carbon and carbon black, were investigated. [10,11] Despite showing significant activity and selectivity,these early catalysts suffered from inadequate oxidation stability and were subsequently succeeded by the second generation of carbon-based dehydrogenation catalysts,r epresented by carbon nanomaterials. [11,12] Aw ide variety of carbon nanomaterials,f or example,c arbon nanotubes, [11] carbon nanofibers, [13] onion-like carbon, [14] and fewlayered graphene, [15] have been employed successfully in oxidative dehydrogenation reactions.T he benefit of carbon nanomaterials compared to amorphous first generation catalysts is mainly their crystalline,p redominantly sp 2hybridized microstructure that is responsible for sufficient oxidation resistance and simultaneously enables high redox activities. [16,17] In this context, the presence of large conjugated (graphitic) domains with ah igh density of defect sites (e.g. edges,in-plane defects) seems to be fundamental. These structures enable high redox activity by acting as electron storage for conjugated oxygen surface groups,such as ketonic carbonyl groups,a nchored at edges and defects. [13,15,[17][18][19] As no inner porosity is present in carbon nanomaterials,t hese active centers are situated at the outer surface and are therefore highly accessible.
However,i nc ase of heterogeneous catalysis,d ue to intrinsic drawbacks such as the large pressure drop and high porosity of fixed nanocarbon catalyst beds,ademanding scaling of synthesis and unclear health hazards,n anocarbon materials are still awaiting industrial application. [20,21] Initial studies with carbide-derived carbons showed that highly crystalline but mesoporous carbon powders seem to exhibit the needed key catalytic features and show similar catalytic properties as carbon nanomaterials. [21] Nevertheless, carbide-derived carbons are currently research model materials,asthe necessary chlorination of carbides at temperatures above 1200 8 8Ch inders application as at echnical catalyst. Inspired by these results,anew generation of carbon-based dehydrogenation catalysts is introduced here,w here the key features for high activity and stability are added to conventional powder carbons by employing as imple and scalable polymer-based synthetic pathway.B esides scalability,p olymers have the advantage that the synthesis is very reproducible,w hile the use of purified monomers enables the synthesis of highly defined carbon precursors that contain aminimum of impurities.Asnano-sized graphite domains are anchoring points for active sites but should be embedded in macroscopic particles,t he approach of this work is to grow crystallites within the carbon matrix by catalytic graphitization during pyrolysis of the polymer precursor.A sac onsequence of this approach, carbon crystallization only occurs in domains which come into contact with the graphitization catalyst, and ah ybrid carbon material, consisting of amorphous and graphitic domains,r esults.F inally,t he active oxidative dehydrogenation catalyst is obtained by creating access to these graphitic domains via selective oxidation of the amorphous parts of the material (Scheme 1).
Key to as uccessful catalyst synthesis following Scheme 1 is the amorphous/graphitic hybrid carbon with heterogeneity in terms of crystallinity at the nanoscale.T oo btain such materials,finely dispersed metal particles (e.g.Ni, Co,Fe) are ap rerequisite and can be obtained through ah omogeneous distribution of metal ions within the precursor polymer. [22][23][24] Such homogeneous distributions can be realized by employing apolymer precursor with ion-exchange properties,which makes up the first requirement for the polymer precursor.As macroscopic powders shall result, the second requirement is good accessibility of the carbon resulting from pyrolysis. Therefore,asuitable template should be incorporated into the precursor polymer to ensure mesoporosity of the carbon material after pyrolysis.
Based on these prerequisites for the new synthesis route, the polycondensation in solution of ap hloroglucinol/formaldehyde reaction system in presence of the soft template Pluronic F127 is chosen to predetermine morphology and ensure am esoporous texture (Scheme 2). [25] To add ion exchange capacity,t he precursor polymer is modified with carboxylic acid groups anchored on the polymer surface.This precursor can subsequently be loaded with metal ions via ion exchange,r esulting in ah omogeneous,a tomic distribution. Pyrolysis of this carbon precursor initially leads to carbonaceous material and then to finely dispersed graphitization catalyst particles through carbothermal reduction of the metal ions.A te levated temperatures (> 700 8 8C), these metal particles migrate through the carbon matrix and graphitize surrounding domains by forming metastable metal carbides, which subsequently decompose to form graphitic carbon. [22,23] In this manner,ahybrid material with heterogeneity at the nanoscale is generated:d omains which did not come into contact with the graphitization catalyst exhibit an amorphous/ turbostratic microstructure,w hile domains that came into contact with the graphitization catalyst are highly crystalline.
Edges and defects of graphite crystallites grown by this approach are initially buried in am atrix of amorphous/ turbostratic carbon.
Thed ifference in oxidation stability between graphitic and amorphous/turbostratic carbon can now be exploited to create access to these structural key features by selective oxidation of non-graphitic domains. Scheme 1. Ah ybrid amorphous/graphitic material is prepared by catalytic graphitizationo fapolymer-based carbon precursor. Removalo f amorphous/turbostratic domains yields the active dehydrogenation catalyst by providingaccess to defect-rich graphitic domains.
Within this contribution this synthesis route towards novel carbon-based dehydrogenation catalysts is studied and the new catalytic materials are tested for their catalytic performance for the oxidative dehydrogenation of ethanol to acetaldehyde.T his reaction is of great practical interest as it represents acatalytic link between bioethanol, which is easily obtainable from renewable resources,a nd an important intermediate in current industrial chemistry. [26] Results and Discussion Am onomer system of phloroglucinol (1,3,5-trihydroxy benzene) and formaldehyde that was subjected to polycondensation in presence of the triblock copolymer soft template Pluronic F127 (according to ap rocedure of Chai et al. [25] ) served as starting material for the catalyst synthesis.T his approach yields spherical polymer particles of about 250 mm diameter in which micelles of the triblock copolymer Pluronic F127 are incorporated in ac ross-linked matrix of phloroglucinol (subsequently denoted as precursor polymer,F igure S1). Aiming to introduce negatively charged groups to the surface of the precursor polymer,u biquitous acidic phenolic hydroxy groups were subjected to etherification with chloroacetic acid (subsequently denoted as carboxy polymer). This surface manipulation only had aminor impact on the pyrolysis behavior compared to the precursor polymer, as indicated by thermogravimetric analysis ( Figure S2), while as ignificantly enhanced ion exchange capacity (from 0.89 to 2.44 mmol g À1 )w as proven by potentiometric titration (Figure S3). Analysis of the carboxy polymer by diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) revealed ab and at 1760 cm À1 that can be assigned to individual (non-dimer) carboxylic acid groups which could neither be detected in the spectra of the precursor polymer nor in the spectra of the main constituents of the polymers ( Figure S4). By exchanging H + for Ni 2+ ,an ickel loading for the polymer of 1.41 AE 0.26 wt %and 1.15 AE 0.07 wt %could be determined by temperature programmed oxidation (TPO) and inductive coupled plasma optical emission spectroscopy (ICP-OES), respectively ( Figure S5).
Pyrolysis of Ni-loaded polymer particles at 1400 and 1500 8 8Cy ields spherical carbon particles of about 100 mm diameter (note:a fter HCl washing, pristine polymer-derived carbons are denoted as PDCXXXX-P,w here XXXX represents the pyrolysis temperature;F igure 1a). Thet extural properties of the pristine PDC1400-P and PDC1500-P materials were characterized using N 2 physisorption (Figure 2a). Thei sotherms show type IV(a) behavior, indicating the presence of meso-and micropores.The hysteresis loops of the isotherms exhibit asteep desorption branch in the relative pressure range of 0.4 < p/p 0 < 0.6 and can be classified as type H2(a). [27] Specific surface areas were determined to be 171 m 2 g À1 and 38 m 2 g À1 for PDC1400-P and PDC1500-P, respectively.R aman spectroscopy of PDC1400-P and PDC1500-P shows distinct first-order D-and G-Bands at 1350 cm À1 and 1580 cm À1 ,r espectively,a sw ell as clearly developed second-order D-Bands at 2700 cm À1 ,i ndicating Scheme 2. Ahybrid amorphous/graphitic material is prepared by catalytic graphitization of ap olymer-based carbon precursor.Removal of amorphous/turbostratic domains yields the active dehydrogenation catalyst by providingaccess to defect-rich graphitic domains.  ad efect-rich graphitic carbon material ( Figure S6). [22] Compared to PDC1500-P,PDC1400-P shows ahigher average I D / I G ratio (obtained from first-order spectra;0 .51 and 1.05, respectively). Furthermore,t he distribution of I D /I G values derived from 10 Raman spectra of ag iven sample is significantly broader in case of PDC1400-P,indicating ahigher degree of inhomogeneity in terms of crystallinity ( Figure S7-S9). TPO analysis shows the occurrence of two carbon species with different oxidation resistances for both PDC1400-P and PDC1500-P (Figure 2b). Them ass fraction of the carbon species with low oxidation resistance (LOR) and high oxidation resistance (HOR) can be estimated from the residual mass at the infliction point of the TPO mass loss curve ( Figure S10). ForP DC1400-P,t he amount of species with LOR is significantly higher (LOR/HOR 71:29), while for PDC1500-P this ratio is reversed and the carbon species with higher oxidation resistance dominates (LOR/HOR 31:69). After thorough washing with HCl, no Ni residue could be detected by TPO.I ts hould be noted at this point that other techniques such as X-ray powder diffraction (XRD) transmission electron microscopy (TEM) as well as ICP-OES analysis failed to provide evidence of relevant Ni residues (Table S1).
XRD analysis of PDC1400-P and PDC1500-P catalytically graphitized carbons reveals pronounced reflections that can be assigned to the graphite lattice planes (002), (100/101), (004), and (110) (Figure 2c). Compared to crystalline graphite-2 H, the reflections of PDC1400-P are shifted to lower 2q values and exhibit broad, to some extent asymmetric profiles characteristic for graphitic phases with ah igh degree of disorder. [23,28] In contrast to the powder diffraction pattern of PDC1400-P,t he (002) and (004) reflections of PDC1500-P exhibit pronounced shoulders at the higher 2q regions of these peaks,s uggesting superposition of two reflections (Figure 2c,d and Figure S11 showing also graphite-2 H). Deconvolution of the data results in two peaks located at 26.178 8 and 26.508 8/2q for the (002) and at 53.748 8 and 54.548 8/2q for the (004) reflection. Asimilar peak splitting was recently described by Alaferdov et al. for sonicated graphite samples and attributed to graphitic phases exhibiting different degrees of order. [29] Ther esults of XRD and Raman spectroscopy point to the coexistence of two graphitic phases:o ne highly disordered and another exhibiting ahigher degree of order. [29] It needs to be noted that polymer samples without Ni loading but pyrolyzed at identical temperatures exhibit merely broad reflections and alow signal-to-noise ratio,suggesting an X-ray amorphous character and demonstrating the importance of the graphitization catalyst to obtain ac arbon material with ananoscale hybrid amorphous/graphitic microstructure (Figure S12).
Supporting results of previous analyses,T EM reveals domains of extended graphitic structures and large amorphous/turbostratic parts for both PDC1400-P and PDC1500-P. At the mesoscopic scale,t hese crystalline/amorphous hybrid materials show graphitic domains exhibiting alarge degree of disorder.Atthe microscopic scale,the catalytic graphitization with Ni seems to produce arather imperfect form of graphite, as various forms of defects,s uch as buckling,m erging and splitting of graphene layers,c an be observed in the graphi-tized parts of PDC1400-P and PDC1500-P (Figure 1b and Figures S13-S14).
Oxidation of the pristine PDC materials in synthetic air at 380 8 8Cd id not lead to an observable degradation of the macrostructure ( Figure S15) but induced distinct textural changes (denoted as after oxidation,P DCXXXX-AO). In case of PDC1400-AO, asignificant rise in N 2 uptake in the low relative pressure range 0 < p/p 0 < 0.05 could be detected, indicating increased microporosity (Figure 2a). Simultaneously,t he specific surface area increases from 171 m 2 g À1 for PDC1400-P to 640 m 2 g À1 for PDC1400-AO. To ac ertain extent, this also applies for PDC1500-AOb ut the most prominent change in the isotherm compared to the pristine material is the rise in N 2 uptake at high relative pressures 0.5 < p/p 0 < 1. No plateau in N 2 uptake at high relative pressures can be detected and the desorption hysteresis changes from type H2(a) to H4. Thedescribed changes of the N 2 isotherm indicate as light increase in microporosity for PDC1500-AO, and the development of pronounced mesoporosity. [27] In case of PDC1500, oxidative treatment increases the specific surface area from 38 m 2 g À1 to 182 m 2 g À1 .I n comparison to the pristine materials,R aman spectroscopy reveals adecrease in the average I D /I G ratio for PDC1400-AO (1.05 to 0.91) and PDC1500-AO(0.51 to 0.39) ( Figure S6). In addition, an arrowing distribution of the I D /I G values over 10 spectra could be observed for both materials.T hese findings indicate that oxidized PDC possesses ah igher ordered microstructure than pristine materials,w hile the inhomogeneity of the material appears to decrease with oxidative treatment (Figures S8 and S9). TPO measurements show as hift in the ratio between oxidation-stable and -unstable carbon species:t he mass fraction of the less oxidation-resistant carbon species decreases during oxidation of the pristine PDC materials (Figure 2b). Despite removing some of the carbon material of low oxidation resistance of PDC1400-P,P DC1400-AOs till contains significant amounts thereof (LOR/HOR 67:33). In contrast, oxidative treatment removed the greatest part of the less oxidation-resistant species for PDC1500-AO(LOR/HOR 15:85). XRD suggests that the oxidation procedure did not provoke any decline in crystallinity,asposition and shape of the observed reflexes did not change significantly for PDC1400-AOo rP DC1500-AO (Figure 2c,d). TEM analysis of PDC1500-AOs howed al oss of amorphous/turbostratic domains after oxidation in synthetic air.While the graphitized parts of the material did not seem to be affected by the oxidation, most of the amorphous matrix appears to be eliminated (Figure 1c,F igure S14).
After thorough characterization of the new generation of carbonaceous materials,t he gas-phase oxidation of ethanol (EtOH) to acetaldehyde (AcH) at 330 8 8Cisemployed as atest reaction to assess the catalytic activity of the new PDC materials.M ulti-walled carbon nanotubes (CNTs) as typical 2 nd generation catalysts serve as benchmark (Figures 3a nd S16-S17, studies were exclusively carried out within the regime of constant reaction rates under exclusion of film diffusion limitations (see side note S1)). Besides acetaldehyde,ethyl acetate (EtOAc) as well as carbon monoxide (CO) and carbon dioxide (CO 2 )were detected as side products for the CNT catalyst. As electivity towards AcH of 79 %w as obtained (at an EtOH conversion of 38 %), whereas the selectivity to the side products EtOAc,C O, and CO 2 amounted to 9%,3 %, and 9%,r espectively.I nterestingly, at similar conversion PDC1500-AOa lso shows an excellent selectivity towards AcH of 82 % (Figure 3b,c). In contrast, the resulting EtOH conversion and selectivity of PDC1400-AO remains inferior compared to the CNT benchmark. At the expense of selectivity towards AcH (54 %), PDC1400-AO displays ah igh selectivity towards the side products EtOAc (26 %), CO (11 %), and CO 2 (9 %) (Figures 3c,d and S18).
Regarding activity comparison, the technically relevant volume-based space-time yield is employed. In this context, the great advantage of the new generation non-nanocarbon catalysts compared to the carbon nanotube benchmark is revealed. Due to the pronounced higher catalyst bed density of the PDC catalysts compared to the CNTs,t he space-time yield of acetaldehyde at 330 8 8Cw as found to be nearly one order of magnitude higher for PDC1500-AO(171 kg m À3 h À1 ) compared to CNTs (19 kg m À3 h À1 ) ( Figures 3d,S 19). For PDC1400-AOt he production rate is lower in comparison to PDC1500-AObyafactor of 3.5.
In addition to the superior activity of PDC1500-AO, the new materials must also exhibit the high stability that is characteristic for 2 nd generation nano-carbon catalysts.A n experiment with 50 ho ns tream (carbon balance at 99.70 AE 3%), without observable loss of activity and change in selectivity was carried out and highlights the stability of the new generation PDC1500-AOc atalyst (Figure 3b). SEM analysis of the used PDC catalysts (denoted as after reaction,P DCXXXX-AR) showed that even extended periods (> 50 h) on stream did not induce observable degradation of the macrostructure (Figures S15). Further [c] Selectivity towards acetaldehyde, ethyl acetate, CO, and CO 2 for PDC1400-AO, PDC1500-AO,a nd the CNT benchmark catalyst.
[d] Space-time yield of acetaldehyde and conversion for PDC1400-AO, PDC1500-AO, and CNT catalysts and pictures of the fixed catalyst beds of PDC1500-AO and CNT (see Figure S19). Catalytic testing was conducted using 90 mg of catalyst at 330 8 8Cinatubular fixed bed reactor with 4.3 vol %EtOH, 10 vol %O 2 at atotal volume flow of 20 mL min À1 (STP) with He as balance.
characterization of the spent catalysts yielded as light decrease in specific surface area for PDC1400 (from 640 m 2 g À1 to 540 m 2 g À1 )a nd as light increase for PDC1500 (from 182 m 2 g À1 to 209 m 2 g À1 ) ( Figure 2a). Them icrostructure of the PDC materials,p robed by TPO (Figure 2b), Raman spectroscopy ( Figures S6-S9), XRD (Figures 2c,d), and TEM (Figures 1d,S14) did not appear to be influenced significantly by the chosen reaction conditions.F urthermore,t he state of hybridization (sp 2 fraction) as determined by electron energy loss spectroscopy (EELS) mapping (Figures S20 and S21) and X-ray photoelectron spectroscopy (XPS, Figures S22 and S23) is very similar for the used catalysts PDC1400-AR and PDC1500-AR (see side note S2 and Table S2 in SI for detailed discussion on the characterization of the spent PDC catalysts).
Besides structural features,t he in situ formed surface functionalization is believed to play amajor role and ketonic/ quinoidic carbonyl and phenolic surface groups are understood to be the relevant groups for ODH reactions on carbon catalysts. [3,7,11,19] Te mperature-programmed desorption (TPD) of the polymer-derived carbons reveals that PDC1400-AOand PDC1400-AR possess asignificantly higher overall amount of oxygen surface groups compared to PDC1500-AOa nd PDC1500-AR ( Figure S24). However, analysis of the TPD CO emission profile of PDC1400-AR and PDC1500-AR shows that PDC1500-AR exhibits higher concentrations (229 mmol g À1 for PDC1500-AR vs. 173 mmol g À1 for PDC1400-AR) of ahigh-temperature-stable (emission maximum ca. 980 8 8C) CO emitting surface species that is associated with quinones ( Figure S25). [30,31] This finding is supported by the analysis of the XPS O1s region which, in case of PDC1500-AR, shows ah igher contribution of an oxygen species which exhibits ab inding energy of around 530 eV compared to PDC1400-AR (2.2 at %f or PDC1400-AR vs.2.6 at %for PDC1500-AR, Figure S26). Theregion of this binding energy is associated with the presence of surface ketones/quinones. [31] However,itshould be noted at this point that the observed differences in the surface oxide profiles between PDC1400-AR and PDC1500-AR are by far not as significant as the differences in catalytic performance,w hich indicates the presence of ah igh number of spectator species that do not contribute to the catalytic activity of the carbon material.
Considering the far higher specific surface area (540 vs. 209 m 2 g À1 )a nd higher number of oxygen surface groups of PDC1400-AR, the only differences that speak for PDC1500-AR as the "better" dehydrogenation catalyst are found to be the presence of agraphitic phase of higher stacking order or, to be more exact, of as tacking distance closer to ideal graphite,and small deviations in the surface oxide profile.In light of the evidence for the presence of ah igh number of spectator species,wewant to propose the hypothesis that the carbon backbone of as uitable surface oxide plays ac rucial role in the redox activity of the functional group,a nd differentiates an active site from as pectator.H ence,e ven the presence of ah igh number of the "right" oxygen surface functionality may not translate into catalytic activity if these groups are located on an unsuitable carbon backbone (see side note S3 for further discussion). Strikingly,P DC1400 and PDC1500 exhibit fundamental differences even though the synthesis temperature only differs by 100 8 8C. As this behavior could be reproduced by several PDC batches,our current hypothesis relies on aphase transition of the Ni graphitization catalyst in the range between 1400 and 1500 8 8C. Macroscopically,N if eatures am elting point at 1455 8 8C, which falls directly in between the utilized synthesis temperatures.M olten Ni might exhibit different graphitization characteristics compared to conventional Ni particles,l eading to the observed differences in carbon microstructure as well as texture and finally to superior catalytic performance.T his phenomenon is not limited to Ni as graphitization catalyst, but also applies to Co.U nderscoring the presented phase-transition hypothesis, Co also features am elting point between 1400 and 1500 8 8C, which translates into graphitic domains of stacking distances close to ideal graphite when Co is employed to graphitize PDC at 1500 8 8C( Figure S27a). Furthermore,PDC graphitized with Co at 1500 8 8Cs howed as imilarly high catalytic performance as Ni-graphitized PDC1500-AO( Figure S27b).
In order to obtain further catalytic insights for PDC1500-AO and the CNT benchmark, ad etailed (macro-)kinetic study was carried out, varying oxygen concentration, ethanol concentration, and temperature during steady-state experiments ( Figure S28). It needs to be noted that influences of pore diffusion limitation could be ruled out by consulting the Prater-Weisz criterion (see Equations S1-3, Table S3). Kinetics of consumption of oxygen and ethanol were found to follow approximately ap ower-law approach within the investigated concentration ranges.I nl ight of similar surface oxide profiles,t he reaction orders for the consumption of oxygen were found to be 0.26 for both PDC1500-AOand for the CNT catalyst, hinting at similar mechanisms for O 2 activation ( Figures S26 and S29). Reaction orders for ethanol were determined to be 0.59 and 0.39, while the apparent activation energies for acetaldehyde formation were found to be 65.4 kJ mol À1 and 89.2 kJ mol À1 for PDC1500-AOa nd the CNT catalyst, respectively ( Figure 4). Then on-integer reaction orders hint at acomplex network of elementary reactions being responsible for the observed kinetic behavior, which needs to be clarified in further studies.

Conclusion
We propose asynthesis route towards anew generation of carbon-based dehydrogenation catalysts based on amorphous/graphitic hybrid materials obtained by catalytic graphitization of ap olymer precursor.T he active catalyst is prepared by removing the amorphous matrix of the hybrid materials by mild oxidation, thereby creating access to the previously grown graphitic domains.These graphitic domains are disordered on am esoscopic scale and rich in defects, providing asuitable carbon backbone for highly redox-active oxygen surface groups.Areaction temperature above 1500 8 8C appears to be crucial to the synthesis of an active,s elective, and stable PDC catalyst. While showing the same excellent selectivity at similar conversion in the ODH of ethanol as carbon nanotubes,p olymer-derived carbons synthesized at 1500 8 8Co utperform the CNT benchmark in terms of spacetime yield by nearly one order of magnitude.I na ddition to superior catalytic performance,t he next-generation PDC materials are accessible via ascalable synthetic pathway and exhibit spherical particles with diameters around 100 mm. There are several conceivable strategies that could broaden the future impact of these new carbon catalysts.B eyond Ni, several transition metals (eg. Co,F e) can serve as graphitization catalyst, and the influence of the nature and loading of the graphitization catalyst on microstructure and catalytic performance could be studied. Furthermore,d oping with heteroatoms such as N, S, and Pc ould be employed to tailor the electronic properties (redox activity) and surface chemistry of PDC materials.O nt he one hand, the polymer-based synthesis strategy enables ad irect, controlled heteroatom doping by copolymerisation while,onthe other hand, simple post-synthesis doping methods promise to be feasible as well. Finally,t he scope of applications for the new PDC catalysts could be extended to include the oxidative dehydrogenation of various other relevant substrates such as alkanes and alcohols beyond ethanol, as well as to electro-and photocatalysis.