Defective Carbon-Based Materials for the Electrochemical Synthesis

Subscriber access provided by READING UNIV

Article

Defective Carbon-based Materials for the Electrochemical Synthesis of Hydrogen Peroxide Shucheng Chen, Zhihua Chen, Samira Siahrostami, Taeho Roy Kim, Dennis Nordlund, Dimosthenis Sokaras, Stanis#aw Henrik Nowak, John W. F. To, Drew Christopher Higgins, Robert Sinclair, Jens K. Norskov, Thomas F. Jaramillo, and Zhenan Bao ACS Sustainable Chem. Eng., Just Accepted Manuscript • DOI: 10.1021/ acssuschemeng.7b02517 • Publication Date (Web): 30 Oct 2017 Downloaded from http://pubs.acs.org on October 30, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Sustainable Chemistry & Engineering is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sustainable Chemistry & Engineering

Defective Carbon-based Materials for the Electrochemical Synthesis of Hydrogen Peroxide Shucheng Chena†, Zhihua Chena†, Samira Siahrostamia†, Taeho Roy Kimb, Dennis Nordlundc, Dimosthenis Sokarasc, Stanislaw Nowakc , John W. F. Toa, Drew Higginsa, Robert Sinclairb, Jens K. Nørskova,d,*, Thomas F. Jaramilloa,* and Zhenan Baoa,* a

Department of Chemical Engineering, 443 Via Ortega, Stanford, California, 94305, USA Department of Materials Science and Engineering,496 Lomita Mall , Stanford, California 94305, USA c Stanford Synchrotron Radiation Light source, SLAC National Accelerator Laboratory, 2575 Sand Hill Road Menlo, Park, California 94025, USA d SUNCAT Center for Interface Science and Catalysis, SLAC National Accelerator Laboratory, 2575 Sand Hill Road, Menlo Park, California 94025, USA b

Corresponding Author * Email. [email protected], [email protected], [email protected] KEYWORDS: hydrogen peroxide, carbon catalyst, electrocatalysis, oxygen reduction reaction, porosity ABSTRACT: Hydrogen peroxide (H2O2), an important industrial chemical, is currently produced through an energy-intensive anthraquinone process that is limited to large-scale facilities. Small-scale decentralized electrochemical production of H2O2 via a twoelectron oxygen reduction reaction (ORR) offers unique opportunities for sanitization applications and the purification of drinking water. The development of inexpensive, efficient and selective catalysts for this reaction remains a challenge. Herein, we examine two different porous carbon-based electrocatalysts and show that they exhibit high selectivity for H2O2 in alkaline conditions. By rationally varying synthetic methods, we explore the effect of pore size on electrocatalytic performance. Furthermore, by means of density functional calculations, we point out the critical role of carbon defects. Our theory results show that the majority of defects in graphene are naturally selective for the two-electron reduction of O2 to H2O2, and we identify the types of defects with high activity.

INTRODUCTION Hydrogen peroxide (H2O2) is an important industrial chemical with a wide range of applications, including paper and textile manufacturing and environmental protection by means of the detoxification and color removal of wastewater.1 The remarkable oxidation properties of hydrogen peroxide allow it to oxidize various pollutants selectively, while it transforms to water. H2O2 is considered a promising solution for improving access to clean drinking water, an issue that affects many parts of the world. Industrially, H2O2 is produced through the energy intensive anthraquinone oxidation process which is limited to large-scale facilities.1 Moreover, the unstable nature of H2O2 makes long-distance transport challenging and unsafe. These issues make this important chemical largely inaccessible to a large number of people living in remote rural areas who can benefit from it the most. The solution could be smallscale, decentralized electrochemical production of H2O2, enabling use at the point of generation. Recently, electrochemical advanced oxidation processes (EAOPs) have been developed and applied for water purifica-

tion. In these processes, hydrogen peroxide is generated onsite from a two-electron reduction of injected oxygen (eq. 1) and used as an oxidizing agent for water treatment.2 To make the process efficient, an active electrocatalyst that selectively avoids the competing four-electron water formation route (eq. 2) is essential. + 2( + ) → ° = 0.70 . (1) + 4( + ) → 2 ° = 1.23 . (2) The two-electron reduction reaction involves only one intermediate, namely OOH*,3 while the four-electron reduction reaction involves OOH*, O* and OH* intermediates.3–5 Both reactions have OOH* as the first intermediate (* represents an active site on the catalyst surface). A further, single electron reduction of OOH* could result in selective production of H2O2. The further reduction to H2O is suppressed only on catalysts with weak O* binding energies, such as Au surfaces that are capable of preserving the O-O bond.6 It is also known that catalysts with isolated active sites consisting of the reactive center surrounded by element(s) with weak oxygen binding

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

can also maintain the O-O bond and work reasonably well for this reaction.7–10 To this end, the two-electron reduction of oxygen to H2O2 has been the subject of substantial studies in recent years.7–9,11–15 Many prospective electrocatalysts have been suggested, including noble metals,3,16,17 metal alloys7,9,16 and carbon-based materials.11,18,19 It has been shown that the two-electron reduction dominates on metal alloys such as Pt/Pd-Hg8,9 and Pd-Au7; however, the noble metal reliance and high price impedes large-scale applicability. Carbon-based materials are particularly interesting alternatives to precious metals as H2O2 electrochemical synthesis catalysts. They are inexpensive, earth abundant and have surface and structural properties that can be tuned to induce favorable electrochemical properties. Chemical doping20–22 and the ease of defect incorporation23 make it possible to modify the electrochemical properties of carbon-based materials. Among the large variety of carbon-based materials, porous carbon structures have shown great promise for electrocatalysis, due to their high surface area, large pore/volume ratio and high electrical conductivity.11 Several reports describe nitrogen-doped carbon materials, which showed moderate activity and selectivity for the electrosynthesis of H2O2 via the 2e- reduction process.11,18,19 Specifically, quaternary nitrogen and pyrrolic nitrogen11 or the synergistic effect of N-doped structures and the surface oxygen-containing functional groups19 were postulated to be the active sites. There is still limited understanding in identifying the active sites in these catalysts Regardless, the majority of these recent works found that nitrogen dopants consistently demonstrate improved activity towards the 4e- reduction in alkaline conditions, selectivity that is more amenable for fuel cell applications.24–27 Undoped carbon has also shown potential for the 2e- mechanism, though limited to very few studies. Hierarchical porous carbons (HPC) derived from metal organic framework carbonization under H2 were reported to selectively reduce O2 to H2O2 at pH 1-7.15 It was suggested that certain defects and sp3C bonds in the HPC constitute the active sites. Particularly, vacancy and edge type are typically the intrinsic defects that exist in various HPC materials, depending on the synthesis procedures and/or chemical treatments employed.15,23 In another study, a different result was found in that creating edge defects in graphene with plasma treatment, the 4-electron ORR route can be greatly enhanced.28 Understanding the important role of defect sites in carbon based materials in ORR electrocatalysis, given the intrinsic high surface and the great variety of defect sites co-existing in the materials, motivates further investigation. In addition to the different varieties of surface defect sites and their site densities, pore size is an important physical property that can also affect the overall catalytic performance. Specifically, in a recent study a mesopore-dominant nitrogendoped carbon was compared to a micropore-dominant nitrogen-doped carbon, and it was suggested that a mesoporous structure enhanced mass transport and promoted the formation of H2O2.29,11 This arose due to the fact that with larger pore sizes, H2O2 can transport more rapidly out of the catalyst layer, reducing its residency time and the likelihood that it can be further reduced to form H2O. Thus, it would also be of great scientific interest to selectively synthesize defect containing carbons with different, but well-defined pore sizes, and to observe how this affects the electro-synthesis of H2O2. In this work, both the defect and pore-size effect will be investigated,

Page 2 of 9

providing further understanding on the role these key properties play on electrochemical H2O2 synthesis. Herein we examine two synthesized porous carbon catalysts with well-defined pore size regimes. Particularly, predominantly microporous (MicroC) and predominantly mesoporous (MesoC) carbons resulted from polymerization of phenol or its derivative with formaldehyde and subsequent carbonization at 850 °C. These carbon materials provide high selectivity and activity for the electrochemical synthesis of hydrogen peroxide in alkaline conditions. Through this systematic investigation, we show that pore size plays a significant role affecting mass transport and stability of the carbon materials for this important electrochemical reaction. Catalyst evaluation for different pH environments also reveals that there is a pH effect on performance. Physical and chemical characterization techniques were correlated with results from electrochemical measurements to elucidate the unique features of these carbon materials that contribute to their H2O2 production activity. Density functional theory (DFT) is also used to investigate the effect of defects in tailoring the local electronic structure and show that they could in principle be detrimental for the catalytic properties of the carbon catalysts investigated herein. We consider a wide variety of different defect types to assess their activity for the electrochemical reduction of O2. We show that majority of defects in graphene model systems are naturally selective for the two-electron reduction of O2 to H2O2. By constructing the trends in activity, we map out the most active defect sites for the two-electron reduction route. RESULTS AND DISCUSSION Catalyst synthesis. Two different defect containing carbons (MesoC and MicroC) were prepared from methods reported previously.30,31 These methods employ similar precursors, albeit modified synthetic procedures to achieve distinctly different pore sizes. Both carbons were prepared using phenol or a phenol derivative (phloroglucinol), and formaldehyde as precursors. MesoC used Pluronic F-127 act as the soft template while MicroC was prepared from an aerogel without adding any templates.

Figure 1. (a) Nitrogen adsorption and desorption isotherms at 77K and pore size distributions of MesoC and MicroC. (b) Raman spectra of MesoC and MicroC. Id/IG ratio for each sample is also indicated. (c) XPS survey scan for MesoC and MicroC and the content of O is indicated. High resolution scan of C (1s) peak is shown in the inset. (d) NEXAFS spectra of C-K Edge for HOPG, MesoC and MicroC.


Page 3 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Physical, structural and chemical characterization. The pore structures of MesoC and MicroC were characterized by gas sorption. MicroC shows significant N2 uptake at a relative pressure below 0.01, which is a typical characteristic of micropores (Fig. 1a). In contrast, MesoC shows the majority of N2 uptake in the mesopore range. Pore size distributions calculated by nonlinear density functional theory (NLDFT) reveals that MicroC consists of a micropore volume of 0.29 cm3/g and a mesopore volume of 0.07 cm3/g, with an average pore size of less than 0.5 nm. MesoC consists of a micropore volume of 0.03 cm3/g and a mesopore volume of 0.69 cm3/g, with an average pore size of ca. 4 nm. Transmission electron microscopy (TEM) imaging confirms the mesoporous features of MesoC (Fig.S1a). The measured BET surface areas of MesoC and MicroC are 425 m2/g and 740 m2/g, respectively. Raman spectroscopy showed that MesoC and MicroC have intense G-band and D-band contributions (Fig. 1b). The Dband to G-band intensity ratio (ID/IG) was calculated to be 1.04 and 1.00 for MesoC and MicroC, respectively, indicating a similar content of defects. High-resolution TEM (HRTEM) further confirms the existence of defects ((Fig. S1b,d), whereby both carbon materials contain a disordered and isotropic structure that consists of tightly curled carbon layers. Some of the distances between the layers are larger than the lattice spacing of graphite, 0.335 nm, which may indicate the presence of structural defects within each layer. The surface composition of these carbon materials was analyzed by X-ray photoelectron spectroscopy (XPS). Only carbon and oxygen were detected, with MicroC possessing a higher oxygen content (4.3 vs 3.8 at.%). The nearly overlapping C1s feature shown as the inset of Fig. 1d demonstrates the similar chemical identity of carbon surface atoms on both MicroC and MesoC. Deconvolution of the C 1s line has been used previously to quantify the different types of C in carbon materials, such as C=C, C-C, C-OH, C=O and O=C-OH.15,28 Nevertheless, distinguishing between sp2 and sp3-bonded by means of XPS remains a challenge due to similarities in chemical shifts. To gain more detailed insights into the local electronic structure, in particular to more accurately characterize the sp2 and sp3 like character of the carbon structures, we employed near-edge x-ray absorption fine structure spectroscopy (NEXAFS). Fig. 1d shows the NEXAFS spectra of MesoC and MicroC along with a reference Highly Ordered Pyrolytic Graphic (HOPG) sample. The HOPG shows the characteristic π* resonance from sp2-like carbon at 285.3eV, the core-exciton peak at 291.65eV associated with long-range sp2 order, and σ and extended oscillations from a well-ordered sample. Both MesoC and MicroC show significant intensity in the π* region, evidencing a primarily graphitic network, and both have significant intensity in the region between 286 and 290 eV which can be attributed to electronegative functional groups (e.g. the peak near 288eV is associated with carboxyl like carbonyls, C=O). The broad π* feature for MesoC and MicroC compared to HOPG suggests that these samples have a larger number of sp2 carbon defect sites. In particular, as has been observed in aromatic compounds, lower energy states can be associated with more unstable aromatic functionalities, supporting the prevalence of defect states in an sp2-dominated matrix.32 The absence of core-hole exciton (~291.8eV) in these two carbon materials indicates the long-range sp2 network has been broken.33 This is consistent with the HRTEM images shown in

Fig.S1b,d, a reasonable expectation given the method of synthesis. ORR activity. Fig. 2a shows the linear sweep voltammograms collected at 1600 rpm in O2-saturated electrolyte, along with the collected ring currents adjusted for collection efficiency. Both carbon materials exhibit activity and selectivity for the 2e- ORR, with an onset potential very close to the thermodynamic equilibrium potential of 0.7 V (vs. RHE) and able to reach the mass transport limiting current around 0.5 V vs. RHE. At most potentials investigated, the 2e- ORR mechanism is favored, in some cases reaching > 70% selectivity (Fig.2b). At the more positive electrode potentials, closer to 0.7 V vs. RHE, the 4e- ORR mechanism plays a more prominent role. It is worth noting that the onset potential for the 2eprocess is slightly more positive than 0.7 V (vs. RHE), which may be attributed in part to the 4e- ORR, thermodynamically allowed at those potentials, and a pH effect discussed below. Compared with MesoC and MicroC, HOPG is quite inert and produces insignificant current. This is consistent with NEXAFs observations, showing that HOPG is highly ordered with negligible defective sites that are able to actively reduce oxygen.

Figure 2. (a) RRDE voltammograms at 1,600 r.p.m. in O2saturated 0.1M KOH electrolyte with the disc current density, ring current density adjusted by collection efficiency. (b) H2O2 selectivity as a function of the applied potential. Koutecky-Levich plots of (c) MesoC. (d) MicroC at different potentials. The theoretical lines for n = 2 and n = 4 are shown for comparison. The number of electrons transferred, n, and H2O2 selectivity can be correlated as follows: % H2O2 = (2 - n/2)*100%, e.g. n=2.40 yields 80 % H2O2.

The high selectivity towards H2O2 for these catalysts is further confirmed with the Koutecky-Levich plots (Fig. 2c,d). The number of electrons transferred (n) calculated for both MesoC and MicroC are close to 2, consistent with the high selectivity observed by the Rotation Ring Disk Electrode (RRDE) measurements. From both techniques, the MesoC shows a slightly higher selectivity than MicroC. The activity of MesoC is also greater, as evidenced by the approximate 60 mV difference between the two curves in the onset region. The high activity and selectivity for H2O2 for both MesoC and MicroC could reflect the performance of defect sites, as both the Raman and NEXAFS spectra indicate a significant presence of defects in both materials. Further below, we describe DFT studies aimed at elucidating the role of carbon defects in graphene for the 2 e- ORR.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Both MesoC and Micro C exhibit favorable H2O2 selectivity with maxima > 70 %, as measured by direct sampling at the rotating ring of the RRDE (Fig. 2a,b). The selectivity of MesoC is consistently higher across most of the potential range investigated, a difference that might be explained by differences in pore-sizes between the two materials. Any H2O2 generated deep within these porous carbon catalysts must be able to diffuse out through the pores without decomposing for it to be detected at the ring. This poses a challenge for both MesoC and MicroC catalysts given their high degree of porosity, particularly for MicroC with its characteristic pore size of less than 1 nm. Koutecky-Levich analysis is another means to assess selectivity. By this metric both catalysts exhibit selectivities in the range of 80-100 %, somewhat higher than that measured by the RRDE method. The chemical disproportionation of H2O2 or HO2- into O2 and H2O is a well-known process34 that impacts the electrochemical measurements in several ways. For one, the decomposition of the product would naturally lower amount of H2O2 reaching the rotating ring of the RRDE compared to that expected by Koutecky-Levich analysis, consistent with our data. Secondly, the O2 bubbles generated by this process could also be trapped inside the pore structures, blocking access to catalytic sites and ultimately lowering the total current density drawn; under these conditions, a full 4 epathway would exhibit a current density of approximately 5.7 mA/cm2, while a standard 2 e- pathway would be half of that.35 A further complication to analysis is the subsequent electrochemical reduction of any O2 generated locally by chemical disproportionation. Reducing this O2 back to peroxide represents a cycle that could in principle repeat over and over, trending n (# of electrons transferred) towards higher values than that anticipated for a given intrinsic selectivity for H2O2 vs. H2O.36At this point there is insufficient information on how chemical disproportionation occurs in MesoC or MicroC, or on the fate of the disproportionation products within these two materials, that would allow for precise quantitative comparisons on their intrinsic selectivity toward H2O2 production based on diffusion-limited current densities or KouteckyLevich analysis. For this reason, we report selectivity data based on the most conservative measurement, that of direct sampling of H2O2 at the ring of the RRDE. This measurement provides a minimum value for intrinsic selectivity towards H 2O 2. It is preferable to conduct the 2e- ORR in neutral or slightly acidic environments considering the field application since most of the stabilizer are weak acids.37 As such, MicroC and MesoC were also examined at other pHs (Fig. S2). Under pHneutral conditions (pH=7), the high selectivity for H2O2 (>70%) is preserved for MesoC while that of MicroC is somewhat lower. There is a 0.2V delay in the onset potential region for both catalysts (Fig. S2a,b). Moreover, an electrolyte pH effect was also observed, whereby activity is greatly reduced under acidic conditions (Fig.S2c,d). Based on a previous study on gold surfaces in basic media,38 the rate determining step for the reaction could involve a proton/electron decoupled election transfer, with HO2- generated through the disproportionation of the first O2- intermediate. One hypothesis for the change upon different pH is that the ratedetermining step of this reaction is pH independent. However, further investigations into mechanistic aspects will be needed before strict conclusions can be drawn.

Page 4 of 9

ORR stability. Stability is a key catalyst performance metric to consider, longer operational lifetimes would benefit the techno-economics of this technology. As the primary product of the process, H2O2, is a highly reactive oxidizing agent, a commercial device for this process would be subject to a harsh chemical environment. To examine catalyst stability we performed accelerated durability testing (ADT) by measuring 11,000 cyclic voltammograms between 0.2 and 1.1 V vs. RHE at a scan rate of 200 mV/s, over 20 hours of continuous cycling. Periodic evaluation of ORR activity throughout ADT revealed that the MesoC remained stable during the whole process while MicroC exhibited some decay in both onset potential and current density (Fig. 3a,b). a

b

c

d

Figure 3. Stability performance of MesoC (a) and MicroC (b) in base. (c) Capacitance measurements under N2 atmosphere for MesoC and MicroC before and after stability tests. (d) Capacitance measurement under N2 atmosphere for MicroC with different mass loadings.

To investigate possible changes to the materials themselves before and after ADT, we measured capacitance curves before and after the 11,000 cycle ADT, collected in the absence of O2 by saturating the electrolyte with N2 (Fig. 3c). The capacitance behavior of MesoC is nearly identical before and after ADT, however that of MicroC changes significantly, an indication of changes to the material itself. Subsequent studies at various loadings of MicroC (Fig. 3d) reveal that the shape of capacitance curves is not strongly dependent on loading, and hence the observed changes in capacitance curves before and after ADT are not likely due to a mechanical loss of catalyst (detachment). Considering the similar chemical compositions of MesoC and MicroC, the differences in their response to ADT could be related to differences in their characteristic pore sizes, an area for further investigation. Interestingly, the capacitance measurements in Fig. 3c indicate that MesoC has a higher electrochemically active surface area (ECSA) than MicroC, despite MicroC’s higher BET surface area. This suggests that the microporous nature of MicroC is generally less accessible to electrolyte than the mesoporous structure of MesoC, consistent with expectations for mass transport in pores < 1 nm, and in line with the discussions presented earlier regarding the activity and selectivity comparisons between the two catalysts. To further confirm the difference in accessibility between MesoC and MicroC, ECSA of these two materials were determined (Fig. S3). MesoC gives a ECSA of 673.25 cm2/mg, much higher than that of MicroC, 406.75 cm2/mg. The electrochemical current were also plotted with normalization by ECSA. As seen (Fig. S3c), MesoC and MicroC gives similar H2O2 current density (vs. ECSA). This is consistent with experimental findings: since both carbon mate-


Page 5 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


rials have similar defect content, the inherent activity should be close between them, and it is the pore structures that affect the overall produced current. The change in capacitance behavior of MicroC before and after ADT could potentially be related to greater sensitivity of microporous surfaces to chemical and physical phenomena during reaction, including oxidation induced by H2O2 and the chemical disproportionation process that forms chemically reactive radicals as well as O2 bubbles that can mechanically impact the material. Effects such as these have been investigated in studies involving the oxidation of carbon supports in polymer electrolyte membrane (PEM) fuel cells,39,40 in particular when catalysts are involved that are selective towards the 2e- pathway. Ultimately, the activity, selectivity, and stability measurements on MicroC and MesoC suggest that a mesoporous structure is more favorable for H2O2 production, along with long-term operational stability. The above results have implied highly defective sites and large pore sizes are desired for ideal ORR catalysts to produce H2O2 with high activity, selectivity and stability. To further confirm this finding, a commercial mesoporous carbon, CMK3, has been tested. As reported in our recent work41, CMK-3 has a much higher surface area and higher mesopore volume than MesoC. NEXAFS spectra suggests it also has a higher content of sp2 carbon defect sites (Fig. S4). Consistently, CMK-3 exhibits high stability, and gives a much earlier onset potential a higher current and improved H2O2 selectivity (>90% at 0.6 V vs RHE). Theoretical results. During the synthesis of nanostructured carbons, a wide variety of defect configurations may be formed due to chemical treatment and pyrolysis step(s).23 It is well known that defects can tailor the electronic structure of carbon materials and significantly modulate their oxygen reduction activity.23,42 For example, pentagonal defects have been recently reported to have a significant contribution to oxygen reduction activity in carbon-based metal free catalysts.42 Herein, we use computational approaches to investigate a wide variety of defect configurations in order to understand their role in oxygen reduction. We study twodimensional graphene sheets as model systems and introduce different types of defects (Fig. S4), including pentagon edges (SI), single vacancies (SV) as well as double vacancies (DV) such as 555-777, 5555-6-7777, 555-6-777, and Stone Wales 55-77. We also consider three different line defects including 55-8-55, 55-77 and 555-777. Here the 5, 6, 7 and 8 refer to the pentagon, hexagon, heptagon and octagon in the defect configuration. Our calculations indicate that the pentagon edge, SV, and 585 DV are too reactive to contribute to the oxygen reduction catalysis, while other DVs are able to catalyze ORR (SI). Figure 4a displays the most active defect configurations with the outlined active site located at the five member rings.

Figure 4. (a) Different defect type configurations examined in this study. D1 to D8 are double vacancy defect types with nonhexagonal ring members. D1 and D6 display different active sites in 555-6-777 double vacancies. D2 and D4 are 555-777 line defects, and D2 is 55-8-55 line defects. D3, D7 and D8 are 555-777, 55-77 and 5555-6-7777 double vacancy defects, respectively. The numbers 5, 6, 7 and 8 refer to pentagon, hexagon, heptagon and octagon. (b) Two- (red) and four-electron (black) ORR related volcano plots for the electro-reduction of O2 to hydrogen peroxide and water, respectively are displayed with the limiting potential plotted as a function of ∆ . Both volcano plots are based on the RHE scale. The equilibrium potentials for both two- and fourelectron ORR are shown as dashed red and black lines, respectively. Blue squares display the activities of Pt(111) surface adapted from Ref. [5], PtHg4 and Pd/Au alloys adapted from Ref. Ref. [8].

We start by modeling the adsorption energies of ORR intermediates, namely O*, OH* and OOH*, for all the considered defect configurations. Our calculations indicate scaling relationships between the adsorption energies of the OH* and OOH* (Fig. S3). These scaling relationships are very similar to those observed for transition metals,41 transition metal oxides,42 and other two-dimensional material surfaces.43 Since the energies of all adsorbates scale with each other, we can use one of them to underline the activity of the carbon based materials towards both two- and four-electron reduction reactions.4,8 We use ∆ as a descriptor and plot the activity volcano limiting potential, !" , is a metric of activity, defined as the lowest potential at which all the reaction steps are downhill in free energy. Fig.4b displays the calculated limiting potential



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

as a function of ∆ for both two-electron and four-electron reduction reactions. As can be seen from Fig.4b, the right side of the twoelectron volcano (weak binding side) partially overlaps with the right side of the four-electron volcano. Far to the right side, the oxygenated species bind too weakly to the surface, and therefore the O-O bond is preserved. This results in increased selectivity towards H2O2 formation, however it comes at the cost of large overpotentials for the reaction. The theoretical overpotential, defined as the maximum difference between the limiting potential and equilibrium potential for the twoelectron path is governed by the binding of OOH*. This is furthermore directly related to the binding energy of OH* due to the scaling relations that exist between oxygenated intermediates. Therefore, controlling the overpotential is a matter of tuning the free energy of OH*, whereby the overpotential will be zero if the free energy diagram is flat at the equilibrium potential of 0.70 V vs RHE. 8 It is important to note that all of the studied defects are located on the right side of the two-electron volcano, indicating that they bind oxygenated species weakly. As a result, successfully adsorbing these intermediates to the surface is a limiting factor for ORR activity. However, some of the studied defects, such as 555-6-777 (D1), 555-777 line defect (D2, D4) and 555-777 point defects (D3), bind oxygenated species at very favorable energies for the electrochemical synthesis of H2O2. The activities of these types of defects are expected to be close to, or even higher than that of the state of the art PtHg4 for the two-electron oxygen reduction to H2O2. Consistent with this, our experimental characterizations reveal the existence of a variety of sp2-type carbons in MicroC and MescoC, with both carbon materials exhibiting good activity and selectivity towards the 2e- ORR with an onset that is close to the thermodynamic equilibrium potential. While it remains challenging to experimentally determine exact active sites, the modeling here suggests that a promising direction is to further tune these active defects, where even greater activity and selectivity could potentially be achieved. CONCLUSION In this work, two different porous carbon materials have been synthesized and investigated as electrocatalysts for the synthesis of H2O2 by means of 2e- oxygen reduction reaction (ORR). Tailoring the synthetic methods allowed for control over the pore structures, resulting in either a predominantly microporous carbon (MicroC) or a predominantly mesoporous carbon (MesoC) structure as measured by BET, with other physical and chemical properties quite similar between the two materials as measured by Raman spectroscopy, X-ray photoelectron spectroscopy (XPS), and X-ray absorption spectroscopy (XAS). Electrochemical investigations show that both carbons exhibit excellent performance in alkaline conditions, both high activity as noted with an early onset potential that is close to the thermodynamic equilibrium potential (0.7V vs. RHE), and a high selectivity over 70% for H2O2. While both catalysts performed well for the reaction, MesoC demonstrated greater electrochemical accessibility and better activity, selectivity, and stability than MicroC, characteristics that are likely attributed to the different pore sizes of the two materials, given their similar characteristics otherwise. An electrolyte pH effect on the reactivity is also observed.

Page 6 of 9

Spectroscopic characterizations reveal that both MicroC and MesoC contain sp2-type defects that may serve as the active sites for the 2e- ORR. To explore this possibility further, density functional theory (DFT) calculations were employed to investigate a wide variety of possible defect configurations. DFT results show that the many types of defects in model graphene systems are expected to show selectivity for the 2eORR. Several of the defect configurations modeled were identified as having activities as high as, or even higher than, the state of the art PtHg4 for the 2e- reduction process. Based on these observations we assign the high activity of the experimentally examined carbon-based materials to the presence of defects. Further work to tune the types and concentrations of defects could lead to further improvements in catalyst activity and selectivity.

ASSOCIATED CONTENT Supporting Information The Supporting Information is available free of charge on the ACS Publications website. Experimental and computation details (PDF)

AUTHOR INFORMATION Corresponding Author * Email. [email protected], [email protected], [email protected]

Author Contributions S.C., Z.C., Z.B. and T.F.J. conceived and designed the experiments. S.C. carried out catalysts design, synthesis and characterizations. Z.C. performed electrochemical tests and characterizations. T.R.K. and R.S. performed TEM experiments. D.N., S.D. and S.N. performed NEXAFS tests. S.S. and J.K.N. conceived and designed DFT calculations. S.S. performed the DFT calculations. All authors discussed the results and co-wrote the paper. All authors have given approval to the final version of the manuscript. †These authors contributed equally to this work.

ACKNOWLEDGMENT This work was supported by the US Department of Energy, Office of Science, and Office of Basic Energy Sciences under Award Number DE-SC0008685 and by the SUNCAT Center for Interface Science and Catalysis, a partnership between SLAC National Accelerator Laboratory and the Department of Chemical Engineering at Stanford University. Use of Stanford Synchrotron Radiation Lightsource, SLAC National Accelerator Laboratory, is supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Science under Contract No. DE-AC0276SF00515.

REFERENCES (1) Campos-Martin, J. M.; Blanco-Brieva, G.; Fierro, J. L. G. hydrogen peroxide synthesis: an outlook beyond the anthraquinone process. Angew. Chem. Int. Ed. Engl. 2006, 45, 6962–6984. (2) Martínez-Huitle, C. A.; Ferro, S. Electrochemical oxidation of organic pollutants for the wastewater treatment:direct and indirect processes. Chem. Soc. Rev. 2006, 35, 1324–1340. (3) Viswanathan, V.; Hansen, H. A.; Rossmeisl, J.; Nørskov, J. K. Unifying the 2e- and 4e- reduction of oxygen on metal surfaces. J. Phys. Chem. Lett. 2012, 3, 2948–2951.


Page 7 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(4) Viswanathan, V.; Hansen, H. A.; Rossmeisl, J.; Nørskov, J. K. Universality in oxygen reduction electrocatalysis on metal surfaces. ACS Catal. 2012, 2, 1654–1660. (5) Nørskov, J. K.; Rossmeisl, J.; Logadottir, A.; Lindqvist, L.; Kitchin, J. R.; Bligaard, T.; Jónsson, H. Origin of the overpotential for oxygen reduction at a fuel-cell cathode. J. Phys. Chem. B 2004, 108, 17886–17892. (6) Siahrostami, S.; Verdaguer-Casdevall, A.; Karamad, M.; Chorkendorff, I.; Stephens, I. E. L.; Rossmeisl, J. Activity and selectivity for O2 reduction to H2O2 on transition metal surfaces. ECS Trans. 2013, 58, 53–62. (7) Jirkovský, J. S.; Panas, I.; Ahlberg, E.; Halasa, M.; Romani, S.; Schiffrin, D. J. Single atom hot-spots at Au-Pd nanoalloys for electrocatalytic H2O2 production. J. Am. Chem. Soc. 2011, 133, 19432–19441. (8) Siahrostami, S.; Verdaguer-Casadevall, A.; Karamad, M.; Deiana, D.; Malacrida, P.; Wickman, B.; Escudero-Escribano, M.; Paoli, E. a; Frydendal, R.; Hansen, T. W.; Chorkendorff, I.; Stephens, I. E. L. S.; Stephens, I. E.; Rossmeisl, J. Enabling direct H2O2 production through rational electrocatalyst design. Nat. Mater. 2013, 12, 1137–1143. (9) Verdaguer-Casadevall, A.; Deiana, D.; Karamad, M.; Siahrostami, S.; Malacrida, P.; Hansen, T. W.; Rossmeisl, J.; Chorkendorff, I.; Stephens, I. E. L. Trends in the electrochemical synthesis of H2O2: Enhancing activity and selectivity by electrocatalytic site engineering. Nano Lett. 2014, 14, 1603–1608. (10) Siahrostami, S.; Björketun, M. E.; Strasser, P.; Greeley, J.; Rossmeisl, J. Tandem cathode for proton exchange membrane fuel cells. Phys. Chem. Chem. Phys. 2013, 15, 9326–9334. (11) Fellinger, T.-P.; Hasché, F.; Strasser, P.; Antonietti, M. Mesoporous nitrogen-doped carbon for the electrocatalytic synthesis of hydrogen peroxide. J. Am. Chem. Soc. 2012, 134, 4072–4075. (12) Choi, C. H.; Kwon, H. C.; Yook, S.; Shin, H.; Kim, H.; Choi, M. J. Phys. Chem. C 2014, 118, 30063–30070. (13) von Weber, A.; Baxter, E. T.; White, H. S.; Anderson, S. L. Hydrogen peroxide synthesis via enhanced two-electron oxygen reduction pathway on carbon-coated Pt surface. J. Phys. Chem. C 2015, 119, 11160–11170. (14) Choi, C. H.; Kim, M.; Kwon, H. C.; Cho, S. J.; Yun, S.; Kim, H.-T.; Mayrhofer, K. J. J.; Kim, H.; Choi, M. Tuning selectivity of electrochemical reactions by atomically dispersed platinum catalyst. Nat. Commun. 2016, 7, 10922. (15) Liu, Y.; Quan, X.; Fan, X.; Wang, H.; Chen, S. High-yield electrosynthesis of hydrogen peroxide from oxygen reduction by hierarchically porous carbon. Angew. Chemie 2015, 127, 6941–6945. (16) Siahrostami, S.; Verdaguer-Casadevall, A.; Karamad, M.; Deiana, D.; Malacrida, P.; Wickman, B.; Escudero-Escribano, M.; Paoli, E. A.; Frydendal, R.; Hansen, T. W.; others. Enabling direct H2O2 production through rational electrocatalyst design. Nat. Mater. 2013, 12, 1137-1143. (17) Marković, N. M.; Adić, R. R.; Vešović, V. B. Structural effects in electrocatalysis: oxygen reduction on the Au(100) single crystal electrode. J. Electroanal. Chem. Interfacial Electrochem. 1984, 165, 121–133. (18) Park, J.; Nabae, Y.; Hayakawa, T.; Kakimoto, M. Highly selective two-electron oxygen reduction catalyzed by mesoporous nitrogen-doped carbon. ACS Catal. 2014, 4, 3749-3754. (19) Lee, Y.-H.; Li, F.; Chang, K.-H.; Hu, C.-C.; Ohsaka, T. Novel synthesis of N-doped porous carbons from collagen for electrocatalytic production of H2O2. Appl. Catal. B Environ. 2012, 126, 208– 214. (20) Dai, L. Functionalization of graphene for efficient energy conversion and storage. Acc. Chem. Res. 2013, 46, 31–42. (21) Liu, J.; Song, P.; Ning, Z.; Xu, W. Recent advances in heteroatom-doped metal-free electrocatalysts for highly efficient oxygen reduction reaction. Electrocatalysis 2015, 6, 132–147. (22) Wong, W. Y.; Daud, W. R. W.; Mohamad, A. B.; Kadhum, A. A. H.; Loh, K. S.; Majlan, E. H. Recent progress in nitrogen-doped carbon and its composites as electrocatalysts for fuel cell applications. Int. J. Hydrogen Energy. 2013, 38, 9370–9386.

(23) Terrones, H.; Lv, R.; Terrones, M.; Dresselhaus, M. S. The role of defects and doping in 2D graphene sheets and 1D nanoribbons. Rep. Prog. Phys. 2012, 75, 62501. (24) Guo, D.; Shibuya, R.; Akiba, C.; Saji, S.; Kondo, T.; Nakamura, J. Active sites of nitrogen-doped carbon materials for oxygen reduction reaction clarified using model catalysts. Science. 2016, 351, 361-365. (25) Yang, H. Bin; Miao, J.; Hung, S.-F.; Chen, J.; Tao, H. B.; Wang, X.; Zhang, L.; Chen, R.; Gao, J.; Chen, H. M.; Dai, L.; Liu, B. Identification of catalystic sites for oxygen reduction and oxygen evolution in N-doped graphene materials: Development of highly efficient metal-free bifunctional electrocatalyst. Sci. Adv. 2016, 2, 4. (26) Gong, K.; Du, F.; Xia, Z.; Durstock, M.; Dai, L. Nitrogendoped carbon nanotube arrays with high electrocatalytic activity for oxygen reduction. Science 2009, 323, 760–764. (27) Liang, H.-W.; Zhuang, X.; Brüller, S.; Feng, X.; Müllen, K. Hierarchically porous carbons with optimized nitrogen doping as highly active electrocatalysts for oxygen reduction. Nat. Commun. 2014, 5, 4973. (28) Tao, L.; Wang, Q.; Dou, S.; Ma, Z.; Huo, J.; Wang, S.; Dai, L. Edge-rich and dopant-free graphene as a highly efficient metal-free electrocatalyst for the oxygen reduction reaction. Chem. Commun. 2016, 52, 2764–2767. (29) Park, J.; Nabae, Y.; Hayakawa, T.; Kakimoto, M. Highly selective two-electron oxygen reduction catalyzed by mesoporous nitrogen-doped carbon. ACS Catal. 2014, 4, 3749–3754. (30) Liang, C.; Dai, S. Synthesis of mesoporous carbon materials via enhanced hydrogen-bonding interaction. J. Am. Chem. Soc. 2006, 128, 5316–5317. (31) Scherdel, C.; Reichenauer, G. Carbon xerogels synthesized via phenol-formaldehyde gels. Microporous Mesoporous Mater. 2009, 126, 133–142. (32) Gordon, M. L.; Tulumello, D.; Cooper, G.; Hitchcock, A. P.; Glatzel, P.; Mullins, O. C.; Cramer, S. P.; Bergmann, U. Innershell excitation spectroscopy of fused-ring aromatic molecules by electron energy loss and X-ray Raman techniques. J. Phys. Chem. A 2003, 107, 8512–8520. (33) Brühwiler, P. A.; Maxwell, A. J.; Puglia, C.; Nilsson, A.; Andersson, S.; Mårtensson, N. π* and σ* excitions in C1s absorption of graphite. Phys. Rev. Lett. 1995, 74, 614–617. (34) Wiggins-Camacho, J. D.; Stevenson, K. J. Mechanistic discussion of the oxygen reduction reaction at nitrogen-doped carbon nanotubes. J. Phys. Chem. C 2011, 115, 20002–20010. (35) Kocha, S.; Garsany, Y.; Myers, D. Testing Oxygen Reduction Reaction Activity with the Rotating Disc Electrode Technique | Department of Energy http://energy.gov/eere/fuelcells/downloads/testing-oxygen-reductionreaction-activity-rotating-disc-electrode. 2013. (36) Maldonado, J.; Stevenson, K. J. Influence of nitrogen doping on oxygen reduction electrocatalysis at carbon nanofiber electrode. J. Phys. Chem. B 2005, 109, 4707–4716. (37) Jones, C. W.; Jones, C. W. Applications of Hydrogen Peroxide and Derivatives; RSC Clean Technology Monographs; Royal Society of Chemistry: Cambridge, 1999. (38) Kim, J.; Gewirth*, A. A. Mechanism of oxygen electroreduction on gold surfaces in basic media. J. Phys. Chem. B. 2006, 110(6), 2565-2571. (39) Maass, S.; Finsterwalder, F.; Frank, G.; Hartmann, R.; Merten, C. Carbon support oxidation in PEM fuel cell cathodes. J. Power Sources 2008, 176, 444–451. (40) Qiao, J.; Saito, M.; Hayamizu, K.; Okada, T. Degradation of perfluorinated ionomer membranes for PEM fuel cells during processing with H2O2. J. Electrochem. Soc. 2006, 153, A967-A974. (41) Chen, Z.; Chen, S.; Siahrostami, S.; Chakthranont, P.; Hahn, C.; Nordlund, D.; Dimosthenis, S.; Nørskov, J.K.; Bao, Z.; Jaramillo, T.F. Development of a reactor with carbon catalysts for modular-scale, low-cost electrochemical generation of H2O2. React. Chem. Eng. 2017, 2, 239-245. (42) Jiang, Y.; Yang, L.; Sun, T.; Zhao, J.; Lyu, Z.; Zhuo, O.; Wang, X.; Wu, Q.; Ma, J.; Hu, Z. Significant contribution of intrinsic



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

carbon defects to oxygen reduction activity. ACS Catal. 2015, 5, 6707- 6712. (43) Abild-Pedersen, F.; Greeley, J.; Studt, F.; Rossmeisl, J.; Munter, T. R.; Moses, P. G.; Skúlason, E.; Bligaard, T.; Nørskov, J. K. Scaling properties of adsorption energies for hydrogen-containing molecules on transition-metal surfaces. Phys. Rev. Lett. 2007, 99, 016105. (44) Fernández, E. M.; Moses, P. G.; Toftelund, A.; Hansen, H. A.; Martínez, J. I.; Abild-Pedersen, F.; Kleis, J.; Hinnemann, B.; Rossmeisl, J.; Bligaard, T.; Nørskov, J. K. Scaling relationships for

Page 8 of 9

adsorption energies on transition metal oxide, sulfide and nitride surfaces. Angew. Chem. Int. Ed. Engl. 2008, 47, 4683–4686. (45) Siahrostami, S.; Tsai, C.; Karamad, M.; Koitz, R.; GarcíaMelchor, M.; Bajdich, M.; Vojvodic, A.; Abild-Pedersen, F.; Nørskov, J. K.; Studt, F. Two-dimensional materials as catalysts for energy conversion. Catal. Letters. 2016, 146,1917-1921.


Page 9 of 9 SYNOPSIS TOC

Insert Table of Contents artwork here Carbon defects have been shown as active sites for 2e- ORR process. The electrochemical production of H2O2 through ORR with cheap, stable and earth abundant carbon catalysts provides a clean and sustainable way to replace the current energy-intensive industrial anthraquinone process.

2

Current Density (mA/cm )

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Ring (H2O2 Current, adjusted by detection efficiency)

2 0 Disk (ORR Current)

-2 MesoC MicroC

-4 0.2

0.4 0.6 0.8 Potential (V vs. RHE)

1.0


9

Defective Carbon-Based Materials for the Electrochemical Synthesis

Recommend Documents