Effects of pore structure in nitrogen functionalized mesoporous carbon on oxygen reduction reaction activity of platinum nanoparticles

CARBON

6 0 ( 2 0 1 3 ) 2 8 –4 0

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/carbon

Effects of pore structure in nitrogen functionalized mesoporous carbon on oxygen reduction reaction activity of platinum nanoparticles Sujan Shrestha, Sasha Asheghi, Jeffrey Timbro, William E. Mustain

*

Department of Chemical and Biomolecular Engineering, University of Connecticut, Storrs, CT 06269, United States

A R T I C L E I N F O

A B S T R A C T

Article history:

Nitrogen-doped mesoporous carbons (NMCs) with ordered to disordered pore structures

Received 21 January 2013

were fabricated on SBA-15 modified with different concentrations of tetraethyl orthosili-

Accepted 26 March 2013

cate using pyrrole as a carbon source. The carbonization temperature of NMCs was main-

Available online 6 April 2013

tained at 800 C so that the amount and type of nitrogen functionalities were constant. Pt nanoparticles (NPs) were deposited onto NMCs using a modified polyol process. N2 adsorption isotherms, transmission electron microscopy, X-ray diffraction, Raman spectroscopy, and X-ray photoelectron spectroscopy were used to characterize NMCs and Pt NPs. A significant shift in binding energy was found in Pt NPs deposited on NMC with disordered pore structure compared to Pt NPs deposited on NMC with ordered pore structure. Pt NPs deposited on NMC with the disordered pore structure had the highest intrinsic oxygen reduction reaction (ORR) activity among the Pt/NMC catalysts. It showed that the interaction between Pt NPs and NMCs could be modulated for enhancement of the ORR activity of Pt NPs by changing the pore structure of NMCs.  2013 Elsevier Ltd. All rights reserved.

1.

Introduction

Proton exchange membrane fuel cells (PEMFCs) convert chemical energy of H2 and O2 directly into electrical energy through complementary redox processes that are not limited by Carnot or Rankine heat cycles. Hence, PEMFCs are widely considered to be one of the most promising candidates for a high efficiency and environmentally friendly energy source for the 21st century, and have been explored widely for both stationary and vehicular applications. In a PEMFC, the hydrogen oxidation reaction (HOR) occurs at the anode and the oxygen reduction reaction (ORR) occurs at the cathode. The HOR kinetics are faster than the ORR by several orders of magnitude over most catalysts, such as Pt [1]. Hence, the rate limiting reaction in PEMFCs is the ORR [1]. Pt and Pt-alloys supported on carbon black are the most popular catalyst for the ORR in PEMFCs. Carbon black has high electrical

conductivity, high surface area, chemical stability, and low cost making it nearly ideal as an electrocatalyst support [2]. On the other hand, it is thermodynamically unstable and bonds to catalyst particles via weak van der Waals forces. As a result, there have been considerable studies to replace carbon black or enhance its surface and structural properties. For the latter case, carbon nanotubes (CNTs), carbon nanofibers, graphene, ordered mesoporous carbon (OMC), and hierarchical/macroporous carbons have been explored [3]. Though enhancements in stability have been reported, none of these supports have been shown to increase interaction or bonding with the catalyst support. This is problematic since an improved catalyst-support interaction is required to increase ORR activity through modification of electronic structure of the Pt nanoparticles (NPs) [3]. The most widely accepted descriptor for ORR activity has been the binding energy of O atoms to the catalyst surface,

* Corresponding author: Fax: +1 860 486 2959. E-mail address: [email protected] (W.E. Mustain). 0008-6223/$ - see front matter  2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.carbon.2013.03.053

CARBON

6 0 (2 0 13 ) 2 8–40

which is a linear function of the d-band center [4–6]. Strong Pt–O binding energy eases the breaking of the O@O bond and enhances electron transfer, but results in surface oxidation and high coverage with intermediate species. On the other hand, weak Pt–O bonding leads to an inability to break the O@O bond and poor electron transfer. The d-band center of a catalyst can be modified by electronic and structural effects imparted by electron transfer or physical distortion of the support. Generally, raw and alloys of noble and nonnoble metals in various forms have been widely investigated as supports to adjust the d-band center of Pt NPs following the d-band center model [3]. However, metal ‘‘supports’’ are expensive. As an alternative, heteroatom (N, B) doped carbon has shown potential to manipulate the d-band center of the catalyst through enhanced interaction via surface heteroatoms and is relatively cost effective [7]. In this regard, nitrogen-doped carbon nanostructures have been the most studied. Pt NPs supported on nitrogen-doped carbon structures such as CNTs, graphene, and nanocoils have shown higher ORR activity than Pt NPs on their equivalent undoped counterparts [3]. However, only a few studies have reported Pt NPs on nitrogen doped carbon nanostructures having higher mass activity than that of Pt NPs on carbon black [8]. Relative to Pt/C, Su et al. showed an increase in ORR mass activity of Pt NPs supported on nitrogen doped porous carbon nanospheres (PCNs) [9], while Li et al. showed ORR enhancement on Pt NPs deposited on nitrogen modified carbon composites (NMCCs) [10]. However, these studies do not report specific activities and are unclear regarding the role of nitrogen in increasing ORR activity. Herrmann-Geppert et al. reported that Pt NPs on ammonia treated Black Pearls (BP) carbon have higher specific and mass activities than the Pt NPs on untreated BP carbon; however, the BP carbon used in their study had significant sulfur impurities [11]. Liu et al. have shown that Pt NPs embedded on nitrogen-doped OMC had higher mass and specific activities compared to non-doped OMC and 20 wt.% Pt/Vulcan XC-72R (Johnson-Matthey) [12,13]. In their study, the Pt wt.% (10 wt.%) and N-content (0–3 wt.%) were low. In addition, the effect of the partial embedding of Pt NPs into the carbon on ORR activity was not clear. Nitrogen doping has been shown to change both the electronic and structural properties of carbon nanostructures. For example, nitrogen doping in CNTs leads to a large electron donor state near the Fermi level, improving metallic properties of CNTs [14]. Depending on the synthesis procedure of N-doped CNTs, they can also have novel structures, such as bamboo-like CNTs [15] and vermicular CNT with graphene outgrowths [16]. This is interesting because CNFs doped with nitrogen during growth have more dislocations with increased d-spacing between the graphitic layers, leading to a turbostratic structure [17]. In CNx films, fullerene like distortions have been observed, which was suggested to be a result of graphene plane deformations by nitrogen-containing pentagonal defects [18,19]. Hence, carbon nanostructures doped with nitrogen during synthesis clearly have more structural deformations, distortions or dislocations sometimes resulting in different microstructure compared to their undoped counterparts. Therefore, to understand the effect of nitrogen doped carbon (NC) as a support, it is important to understand how

29

surface chemistry and microstructure are altered as a function of nitrogen doping. This is because changes in the electronic and structural properties of carbon influence its interaction with supported catalyst particles. Furthermore, these properties can have a significant impact on nucleation, growth, distribution and structure of Pt NPs. However, understanding the role of nitrogen in the activity of the nanostructured carbon is difficult since controlling both surface chemistry and microstructure separately presents a huge challenge. The objective of this work is to study the impact of the pore structure of NC supports on dispersion, structure, and ORR activity of Pt NPs. Hence, we have modified the pore structure of NC from ordered to disordered while keeping the amount and type of nitrogen functional groups constant. Although several studies have reported ORR activity of Pt NPs on OMC [20–33], to the best of the author’s knowledge, there have been no reports regarding the impact of pore order of OMC on ORR activity of the supported Pt NPs. To accomplish this, we chose a template-based method [29] to synthesize nitrogen-doped mesoporous carbon. Literature suggests that the surface chemistry of NC depends mainly on carbonization temperature and the precursor [30]. Therefore, the surface chemistry was controlled by maintaining the same carbonization temperature (800 C) and precursor (pyrrole) while the pore structure was varied by template modification. The physicochemical properties and ORR activities of synthesized NCs and supported Pt NPs are presented and discussed below.

2.

Experimental

2.1.

Sample preparation

First, the SBA-15 template was prepared as reported by Zhao et al. [29]. 6.0 g of Pluronic P 123 (BASF) was dissolved in a solution consisting of 180 mL of 2 M HCl and 45 mL of 18 MX Millipore deionized (DI) water. 13.6 mL of tetraethyl orthosilicate (TEOS) was added to the mixture and the reaction was carried out for 20 h at 45 C and then for 24 h at 150 C. After filtration of the precipitated solid template and washing with copious amounts of acetone and DI water, the template was calcined at 500 C for 3 h. To prepare supports with different order in the pore structure, the pore size of the SBA-15 template was manipulated by additive silicate treatment. To accomplish this, 3.5 g of the SBA-15 was introduced into a fresh solution of 180 mL of 2 M HCl and 45 mL of DI water. Different amounts of TEOS (4, 8, and 12 mL) were added to the mixture at 45 C and stirred for 48 h. The TEOS treated templates (SBA-15-0, SBA-15-4, SBA-15-8 and SBA-15-12 for the 0, 4, 8 and 12 mL additions, respectively) were filtered, washed and calcined at 500 C for 3 h. Fig. 1 presents a simplified illustration of SBA-15 modification through addition of TEOS in an acid media. Nitrogen functionalized mesoporous carbon was formed by a wet pyrrole impregnation, polymerization and pyrolysis procedure [31–33]. 9 mL of pyrrole (C4H5N, 99% Acros Organics) was incorporated into 3.0 g of modified or unmodified SBA-15 template by vacuum impregnation for 24 h. The impregnated pyrrole was polymerized to polypyrrole in

30

CARBON

6 0 ( 2 0 1 3 ) 2 8 –4 0

Fig. 1 – Schematic of hydrolysis and condensation of tetraethyl orthosilicate (TEOS) on the walls of SBA-15.

0.25 M FeCl3 for 24 h. The carbonization was carried out at 800 C for 3 h. Finally, the silica template was dissolved in hot 10 M KOH for 48 h and carbonized polypyrrole (CPPY) was obtained after filtering and washing with DI water. The CPPYs are denoted as CPPY-0, CPPY-4, CPPY-8 and CPPY-12 for the 0, 4, 8 and 12 mL additions of TEOS to the templating procedure, respectively. In this work, Vulcan XC-72R (VC) (Cabot Corp.) was used as a reference material. Pt NPs were deposited onto the carbon supports using a modified polyol process [34]. The pH of 100 mL of ethylene glycol (EG) was raised to 12 using NaOH (Acros Organics) solution. Next, 280 mg of the CPPY-X (X = 0, 4, 8, 12) was introduced to the solution. Hexachloroplatinic acid (H2Cl6PtÆ6H2O, Sigma–Aldrich) was weighed to yield 50 wt.% Pt/carbon, added to the mixture and refluxed at 62 C for 3 h in N2 atmosphere. The resulting material was cooled for 1 h. Then the N2 atmosphere was changed to air and the pH was adjusted to 1.5. After 42 h, the catalyst was filtered and washed with copious amounts of DI water and dried under vacuum overnight at 80 C.

2.2.

Characterization

N2 adsorption isotherms of CPPY-X were obtained at 77 K. The samples were degassed at 300 C for 3 h in inert atmosphere before the isotherms were collected. The Brunauer–Emmett– Teller (BET) specific surface area (SBET) was calculated using the N2 adsorption isotherm between relative pressures (P/Po) of 0.04 and 0.2 [35]. The total pore volume (Vtot) was measured by converting the amount of gaseous nitrogen adsorbed at relative pressure 0.99 (V0.99) to the volume of liquid nitrogen adsorbate (Vtot = 0.0015468 · V0.99). The Kruk–Jaroniec–Sayari correction (KJS), built into the ASAP 2020 V3.04 operating software, was employed to determine the pore size distribution (PSD). The as plot was calibrated against a macroporous silica reference [36] (LiChrospher Si-1000, SBET = 25.1 m2/g, V0.4 = 9.1248 cm3 STP/g) for SBA-15 and nongraphitized

carbon [37] (Cabot BP 280 SBET = 40.2 m2/g, V0.4 = 14.70 cm3 STP/g) for the carbon supports. Transmission electron microscopy was performed with a JEOL 2010 FasTEM. Sample preparation was carried out by dispersing CPPY-X in acetone/ethanol by ultrasonication, followed by deposition and drying on Holey carbon/copper grids. X-ray photoelectron spectroscopy (XPS) was used to investigate the surface composition of the support materials. XPS was performed using a PHI Multiprobe System with twin anode X-ray photoelectron spectrometer using unmonochromatised Al Ka radiation (1486.6 eV) operating at 250 W and 15 kV. The pressure in the analysis chamber was always 108 torr or less. The full survey was taken at 100 eV pass energy with a scan rate of 1 eV/s and the high resolution scans were conducted at 25 eV pass energy with a scan rate of 0.1 eV/s. The spectra were calibrated with respect to the graphitic carbon 1s electron band at 284.6 eV [38]. The background signals were determined using Shirley-type background correction and the curves were fitted with Gaussian/Lorentzian product functions. Raman spectra were taken with a Renishaw 2000 Spectrometer Ramanscope operated by the Wire 2.0 Service pack 5 software. The peak at 521 cm1 from Si was taken as the standard reference. All data were taken using a 514 nm laser. Peak fitting was performed using GRAMS/AI (7.01) software using mixed Gaussian and Lorenztian functions. X-ray diffraction (XRD) was performed on a conventional h-2h diffractometer with a Cu anode and Ni filter. The instrument was a Bruker D8 Advance Diffractometer System. The Ka2:Ka1 intensity ratio was 0.5. The XRD spectra were collected from 5 to 90 in steps of 0.02 with 3s per step. Samples were compressed and leveled in a 0.5 mm deep sample holder. The collected spectra were stripped of Ka2 using the Rachinger method built into DIFFRACplus EVA software. The weight (wt)% of Pt NPs on carbon was found by thermogravimetry with a TA Instruments model TGA Q-500 instrument. All samples put on the Pt pan for analysis were ca.

CARBON

6 0 (2 0 13 ) 2 8–40

7.5 mg. The temperature was increased at 2 C/min from room temperature to 900 C in dry air (20% O2, Airgas) flowing at a rate of 60 mL/min. From 0 to 160 C, the weight loss was assumed to be due to loss of adsorbed water. From 160 to 600 C, the weight loss was attributed to the gasification of carbon. There was less than 3% increase in weight from 600 to 900 C due to Pt oxide formation. Using this method the calculated wt.% of Pt/VC (BASF) matched to that of BASF’s measurement within 0.1 wt.%.

2.3.

Thin-film rotating disk measurements

All ex-situ electrochemical measurements were performed with an Autolab PGSTAT302N Potentiostat at 25 C in 0.1 M HClO4 in a custom-built (Adams & Chittenden), thermostated three-compartment electrochemical cell with a Luggin capillary. The counter electrode (CE) was a Pt flag and the reference electrode (RE) was a K2SO4 saturated Hg/Hg2SO4 electrode. Before each experiment, the reference electrode was calibrated against a hydrogen reference electrode (eDAQ Inc., Hydroflex). All data is reported with respect to the reversible hydrogen electrode (RHE). To prepare the working electrodes, approximately 18.5 mg of the Pt/CPPY-X or Pt/VC was dispersed in 25 mL of 24% (by volume) isopropanol solution with 100 lL of Nafion DE-520 dispersion (DuPont) [39]. The mixture was ultrasonicated for 1 h. Immediately after sonication, a 12.6 lL aliquot was pipetted onto a rotating ring disk electrode (RRDE) with a glassy carbon disk (Area = 0.2472 cm2), which was polished to a mirror finish with a 0.05 lm alumina suspension (Buehler). The catalyst ink was dried by rotating the ink covered electrode at 700 RPM in air [40] and the resulting electrode was transferred, covered by a drop of water, to deaerated 0.1 M HClO4. The thin film was subjected to potential cleaning by scanning between 0.05 and 1.4 V at 200 mV/s for 40 cycles. To investigate the oxygen reduction reaction (ORR), the electrolyte was saturated with UHP oxygen (Airgas) for at least 1 h and polarization curves were taken at different rotation rates at a scan rate of 5 mV/s between 0.05 and 1.06 V at 25 C.

3.

Results and discussion

3.1.

Pore characteristics

SBA-15-0 has a honeycomb-like ordered pore structure [29]. In this study, SBA-15-0 was treated with TEOS to modify its pore size and structure. Fig. S1 shows the pore size distributions and the corresponding as plots (inset) for the SBA-15-X templates. The pore size of the SBA-15-0 decreased with increasing TEOS concentration, which indicated that the condensation of silica occurred on the walls of SBA-15-0. The pore area and volume of the templates also decreased with increasing concentration of TEOS, which are shown in Table S1. The loss of pore area and volume was likely due to the filling of primary mesopores, which are defined as the mesopores forming the ordered pore structure [41] This was concluded from the as plot (Fig. S1, inset) in which the volume of N2 adsorbed by the material under analysis was plotted against the standard reduced adsorption, as, for the reference

31

solid. Standard reduced adsorption, as, is the volume of N2 adsorbed, normalized to the volume of N2 adsorbed by the reference solid at relative pressure 0.4 [36]. At as = 1.4 in the as plot (Fig. S1, inset), a step was observed for all of the SBA-15-X templates signifying the filling of primary mesopores. This step was smaller for SBA-15-4 to SBA-15-12 compared to SBA-0 which indicated that modified SBA-15-X did not have as many primary mesopores as that of SBA-15-0. This could have happened due to blockage or filling of the ordered pores in SBA-15-0 during modification by TEOS. There was also a lack of a significant step at low relative pressures (as < 1) in the as plot of SBA-15-X which showed that mesopores were the source of most of their surface area. One interesting observation was that an increase in the TEOS treatment to 12 mL (SBA-15-12) did not result in a decrease in pore diameter compared to SBA-15-8. This was most likely due to the underlying mechanism for the modification of the SBA-15-0, which involves the hydrolysis of TEOS and condensation of silica. Hydrolysis is faster than condensation at low pH [42], so hydrolyzed alkoxysilanes could access pores and condense on the walls of the SBA-15-0. Hydrolysis and condensation also occurred on the outer surface of the template, forming clusters. These clusters were more likely to be prevalent at high TEOS concentrations, which would provide both kinetic and steric hindrance towards the condensation and growth inside of the SBA-15-0 pores. This suggested that there might be a critical concentration above which external clustering and internal wall growth compete, reducing the effectiveness of the silicate additive method as a pure pore modifier. Since SBA-15-12 did not show further pore modification compared to SBA-15-8, it was not extensively studied. One of the attractive features of SBA-15-0 as a template is that the resulting carbon is a faithful inverse replica; in other words, the carbon has the same p6mm symmetry as SBA-15-0 [43,44]. Therefore, the carbon replica of the TEOS-modified SBA-15-0 would be expected to have increasing pore diameters corresponding to the increasing wall thickness of the template. However, the PSD of the carbon replica in Fig. S2 shows little change in the primary pore diameter. In addition, the PSD shows some peak broadening with increasing TEOS concentration. Earlier we hypothesized that the condensation of hydrolyzed silica into clusters took place on the pore entrances based on the lack of change in PSD for SBA-15-12. This phenomenon, though very prominent in SBA-15-12, should occur in all modified templates at different rates depending on the SBA/TEOS ratio. Hence, the increase in widening of the PSD may be due to carbon formed on silicate clusters at the external surface or pore entrance. This also might be the cause for the decrease in BET surface area of CPPY-4 and CPPY-8 compared to CPPY-0 (Table S1) since the SBA-150 was not completely filled with the pyrrole precursor. This resulted in the systematic decrease of the order in the pore structure with increasing TEOS treatment. TEM images of the templates showed that all SBA-15-X templates had an ordered pore structure (Fig. 2). Contrarily, the resulting carbons, CPPY-X, showed various degrees of order in the pore structure until CPPY-8, which showed a disordered pore structure. This could be due to the small pore size of SBA-15-8 and obstruction due to formation of silica clusters

32

CARBON

6 0 ( 2 0 1 3 ) 2 8 –4 0

Fig. 2 – Transmission electron microscope (TEM) pictures of the templates (a) SBA-15-0, (b) SBA-15-4, (c) SBA-15-8 and their respective nitrogen-doped mesoporous carbons (d) CPPY-0, (e) CPPY-4, (f) CPPY-8.

at the pore entrance helped by partial pore filling [45]. In addition, our results suggested that there is a gradual increase in growth of silica structures at the pore entrances with increasing TEOS treatment which could also be seen on TEM images of SBA-15-4 and SBA-15-8. The difficulty in pore filling increases with smaller pore size. Both factors give rise to amorphous structures at the walls of the ordered carbon structure. Thus, the amount of external amorphous carbon increased from CPPY-0 < CPPY-4 < CPPY-8.

3.2.

Carbon microstructure

3.2.1.

Raman analysis

Disordered graphite shows two sharp Raman peaks, one around 1580–1600 cm1 called the ‘‘G peak’’ relating to the main first order band at 1582 cm1 of pure crystalline graph-

ite, and another around 1350 cm1 called the disorder-induced or ‘‘D peak’’ relating to the main first order band at 1332 cm1 of pure crystalline diamond. Hence, in the literature, Raman analysis of the carbon microstructure focuses mainly on these G and D peaks [46–48]. The G and D peaks for the CPPY-X supports, with curve fitting, are shown in Fig. 3. During peak fitting of the G and D peaks, there was a need for a third peak, the ‘‘A peak’’, to capture the amorphous phase of the supports [46]. The G and D peaks were also not clearly separated; however, the separation was more distinct in VC than in CPPY-X. The distances between the G and D peaks were measured and it was found that the distance was largest for VC (Table S2). Among the CPPY-X supports, the separation decreased from CPPY-0  CPPY-4 > CPPY-8. Dresselhaus et al. reported that when point defects were introduced in small amounts on a pure crystalline graphene

CARBON

6 0 (2 0 13 ) 2 8–40

33

is typically indicative of the intensity of disorder. This indicates that CPPY-X did not have the same degree of crystallinity as that of VC. Using the relative intensities of the D and G peak to calculate La (average size of basal planes) [47], it was found that La was larger for VC compared to CPPY-X. The difference in the size of the basal planes suggested differences in microcrystallites in carbons. To further study such microstructural modification, the carbon samples were analyzed with XRD.

3.2.2.

XRD analysis

The microstructural differences in the CPPY-X supports with decreasing order in the pore structure were observed by dspacing and Lc (average stack height) measured from the (0 0 2) reflection of XRD spectra (Fig. S3) in Table S3. VC had ˚ , which agreed well with the literthe largest d-spacing, 3.63 A ˚ ) [2] while graphite has a d-spacature reported value (3.60 A ˚ . Such large d-spacing in VC shows its ing of 3.354 A turbostratic nature; i.e. there is a lack of 3D order in its microcrystallites. In comparison, CPPY-X had smaller d-spacings than VC. Lc (0 0 2) values measured using the Scherrer equation (Table S3) [2] revealed that the crystalline height decreased from VC to CPPY-X. Hence, Lc from XRD and La from Raman indicated that the carbon crystallites in VC are larger than that of CPPY-X with CPPY-8 likely having the smallest of all the samples. So, there seemed to be a gradual trend in microcrystallinity among CPPY-X supports.

3.3.

Surface chemistry

XPS showed that the C/N atomic% ratio in this work was between 8 and 10, which is consistent with our previous work [33]. The N 1s peaks in CPPYs were deconvoluted into 4 peaks. The peaks at 398.3 eV and 400.7 eV were assigned to pyridinic (N-6) and pyridone nitrogen groups [30]. The peak at 400.7 could also be due to quaternary nitrogen (N-Q) which has also been suggested to be protonated N-6 through formation of Hbridge with nearby hydroxyl or carboxyl groups [30]. The peak at 403.1 eV was likely a result of pyridine-N-oxides and the peak at 405.6 could be NOx [30,50]. The BEs and surface atomic compositions of CPPYs were very similar. This implied that the chemical state and composition of nitrogen functionalities were a function of temperature. The nitrogen groups were comprised mostly of edge structures, N-5 and N-6 which is advantageous as these edge and defect sites can act as active sites for metal nucleation and growth.

3.4. Fig. 3 – Raman spectra of the carbons used in this study. The curves at wavenumber 1000–2000 cm1 are fitted with D, A and G peaks using mixed Gaussian and Lorentzian functions with linear baseline correction.

which had a single G peak, two new peaks arose at 1345 and 1626 cm1 which were identified as the D and D peaks [49]. The D peak appears as a shoulder of the G peak and fuses with both the G and the D peaks as more and more disorder is introduced. Hence, the shift of D peaks towards G peaks

Platinum on carbon

Fig. 4 shows TEM images of 50 wt.% Pt deposited on VC and CPPY-X. It should be mentioned here that both pre-surfaceactivation and post-heat-treatment steps, which lead to better dispersion of NPs [51,52], were avoided during deposition process. Despite the lack of surface treatment, Pt/CPPY-X showed very good dispersion compared to Pt/VC. On VC, Pt NPs were concentrated between the carbon nanoparticles instead of uniformly covering them. On the contrary, Pt NPs were evenly distributed in CPPYs. XRD spectra of Pt NPs on the various supports used in this work are shown in Fig. S4. Different facets of polycrystalline

34

CARBON

6 0 ( 2 0 1 3 ) 2 8 –4 0

Fig. 4 – TEM images of Pt NPs on various carbons used in this study (a) Pt/VC, (b) Pt/CPPY-0, (c) Pt/CPPY-4 and (d) Pt/CPPY-8.

Pt NPs are well-developed for in-house Pt/VC and Pt/VC (BASF) compared to the Pt/CPPY-Xs. This could be due to the small size of Pt NPs supported on the CPPY-Xs. The sizes of Pt NPs were estimated using the Scherrer equation from the Pt (1 1 1) XRD peak (Table S3). The volume averaged size of Pt NPs on VC was significantly larger than that of Pt NPs on CPPYs. High surface area cannot alone ensure good dispersion; surface defects and edges play a significant role in particle dispersion. For example, Pt dispersion is improved in CNTs oxidized using H2SO4–HNO3 mixture to create oxygen functional groups [53]. Thus, the smaller Pt NPs and better dispersion on CPPYs could be linked to their higher surface area and N surface functional groups. N groups on the carbon surface assist in dispersion by introducing electron deficient centers at nearby carbon atoms which can strongly interact with Pt atoms [54–56]. Also, N sites on carbon act as defect sites that also can provide centers for nucleation [56]. When the Pt/VC (in-house) was compared with commercial Pt/VC (50 wt.%) from BASF, the Pt NPs’ size and peak position of (1 1 1) reflection were similar, but the d-spacing of Pt NPs on Pt/VC (BASF) was smaller than that of Pt NPs on Pt/VC (in˚ ). This house), which was closer to that of bulk Pt (2.265 A smaller d-spacing could signify strain in Pt NPs, which could enhance its catalytic activity [57]. Despite the small sizes of Pt NPs on CPPY-X, the d-spacing was slightly higher than that of Pt NPs on VC. This was surprising and could be due to the effect of nitrogen functionalities on CPPY-X. The d-spacing of Pt ˚ , smaller than that NPs on commercial Pt/VC (BASF) was 2.19 A of Pt/VC prepared in-house.

The C1s and N1s XPS peaks of Pt deposited VC and CPPY-X were deconvoluted and compared with similar peaks of bare VC and CPPY-X. The BE and the atomic % of various chemical states of C and N elements were indistinguishable from each other within the error limit. The BEs of the Pto 4f7/2 state in both Pt/VC (in-house) and Pt/VC (BASF) was 71.2 eV, identical to that of bulk Pt 4f7/2 (71.2 eV) [58]. For Pt/CPPY-0 and Pt/ CPPY-4, the BEs of Pt 4f7/2 states were 71.6 eV and 71.8 eV respectively. This agrees with observations by Toyoda et al. who observed a positive shift (ca. 71.2 ! 71.7 eV) in the Pt 4f7/2 core level BE with a decrease in Pt particle size (3.14 ! 0.86 nm) on GCE [59]. However, for Pt/CPPY-8, a negative shift (ca. 0.8 eV) in Pt 4f7/2 core level relative to Pt/ CPPY-0 and Pt/CPPY-4 was found despite similar Pt NPs size (Table S3, Fig. 5), which is discussed below. BE shifts observed using XPS coupled with changes in particle size can be attributed to both initial and final state effects [60,61]. Initial state effects can be attributed to lattice strain, which reduces bond distance and influences d-hybridization. Final state effects are attributed to the relaxation energy due to core–hole screening. As particle size decreases, the BE shifts positive to the bulk value [59]. Specifically, the decrease in bond distance leads to enhanced d-orbital hybridization, strengthening the bonding between the metal atoms, but weakening the bond between metal atoms and adsorbates [60,62]. Therefore, the negative BE shift in Pt/CPPY-8 suggests that the strong d-orbital hybridization between Pt atoms was overcome by the interaction between Pt NPs and CPPY-8. This interaction could arise because of

CARBON

6 0 (2 0 13 ) 2 8–40

35

[54], which creates strong centers of concentrated charge. Very small Pt size creates greater specificity between surface N atoms and catalyst particles. And, if it is assumed that each carbon crystallite was an atom in a crystal structure, amorphous structured CPPY-8 has more low coordination sites than other more ordered CPPY-X; i.e., more free active sites. Indeed, it is the high number of unsaturated sites that makes Ni-based amorphous alloy catalysts more active than their crystalline counterpart [63].

3.5.

Electrochemical behavior

Cyclic voltammograms (CVs) for Pt supported on VC and CPPY-X in N2 saturated HClO4 are shown in Fig. 6. The CVs were well defined and clearly showed all of the characteristics of polycrystalline Pt [64]. The region (0.05–0.3 V) before the double layer (DL) was associated with Pt–H adsorption/ desorption while the region (0.45–1.2 V) after the DL was consistent with Pt-oxide formation/reduction. The H-adsorption/ desorption region featured multiple peaks indicative of Hadsorption/desorption on different crystal facets of cubooctahedral Pt crystallites [65]. The Pt-oxide formation started at ca. 0.6 V during the forward scan (0.05 ! 1.2 V) and Pt-oxide and the oxygen reduction peak occurred at ca. 0.7 V during the reverse scan (1.2 ! 0.05 V). The CV of Pt/CPPY-X mirrored the CV of Pt/VC in shape and position of different peaks except for the marked difference in the DL region (0.3–0.45 V) where the DL current of Pt/CPPY was significantly higher than that of Pt/VC. Pt/CPPY-X has significantly higher DL capacitance than VC due to the higher surface area of CPPY-X and pseudocapacitance or electron delocalization due to surface N atoms [33,66,67]. Another piece of information that can be obtained from Fig. 6 is the electrochemically active specific surface area of the Pt NPs (APt). Assuming one platinum atom adsorbs one hydrogen atom and that the current due to hydrogen evolution makes up for the charge required to complete partial monolayer coverage, APt can be calculated from the area under the Pt–H oxidation/formation peaks (0.05–0.4 V) after the double layer region correction using Eq. (1) [64].

Fig. 5 – Pt 4f XPS spectra of Pt NPs deposited on VC and CPPY-X. The peaks were fitted with doublet of of Pt0, Pt2+ and Pt4+ peaks.

the synergistic effect from small Pt NPs size, high N content, and amorphous carbon nanostructure. High N content likely facilitates aggregation of N sites, as shown by DFT studies

Fig. 6 – Cyclic voltammograms of Pt NPs deposited on VC and CPPY-X recorded in N2 saturated 0.1 M HClO4 at 25 C with a scan rate of 20 mV/s.

36

CARBON

6 0 ( 2 0 1 3 ) 2 8 –4 0

Table 1 – The Pt wt.% in the catalyst (WPt), and parameters from the thin-film RDE experiment, where LPt is the Pt loading in the thin-film, APt is the electrochemically active specific surface area, Is is the specific activity at 0.9 V, Im is the mass-specific activity at 0.9 V and n is the number of electrons in an oxygen reduction reaction. Sample

WPt (wt.%)

LPt (lg/cm2)

Apt (m2/gPt)

Is (lA/cm2Pt)

Im (A/mgPt)

n

Pt/VC (BASF) Pt/CPPY-0 Pt/CPPY-4 Pt/CPPY-8

48.1 42.6 44 44.2

18.2 16.0 16.3 16.7

21.4 83.0 55.1 30.5

297 123 208 226

0.055 0.102 0.115 0.069

3.7 3.7 3.5 3.5

APt ¼

QH 210lC=cm2  LPt

ð1Þ

where QH is the charge under the Pt–H oxidation/formation peaks and LPt is the Pt loading on the working electrode. From the APt values listed in Table 1, it was clear that Pt/CPPY catalysts had a higher active surface area than Pt/VC (BASF) despite similar Pt loadings. Pt/CPPY (83 m2/gPt) showed the highest improvement in APt, almost 4 times that of Pt/VC (BASF) (21.4 m2/gPt). This remarkable enhancement was due to nominal agglomeration and uniform dispersion of Pt NPs on CPPY-X, which could be subscribed to high surface area and N surface atoms on CPPY as discussed earlier. Among the Pt/CPPY-X catalysts, Pt/CPPY-0 had the highest APt and Pt/CPPY-8 had the lowest APt. Here, CPPY-0 had the ordered pore structure while CPPY-8 had the disordered pore structure. In addition, Pt NPs on CPPY-X were less than 2 nm (Table S3). Due to these reasons, Pt NPs on CPPY-8 were less accessible than those on CPPY-0 and CPPY-4. Hence, the dissimilarity in APt of the Pt/CPPY-X catalysts could be attributed to this difference in accessibility.

3.6.

Oxygen reduction activity

Positive (0.1 ! 1.1 V) going linear sweep voltammograms of Pt/CPPYs and Pt/VC (BASF) at 1600 RPM in 0.1 M HClO4 at 25 C with a scan rate of 5 mV/s are shown in Fig. 7. These polarization curves were corrected for ohmic resistance

Fig. 7 – The oxygen reduction reaction (ORR) polarization curves (positive sweep) at 1600 RPM in O2 saturated 0.1 M HClO4 at 25 C with a scan rate of 5 mV/s and Tafel plot calculated after mass-transfer correction of the catalysts used in this study (inset).

which was measured through potentiostatic impedance; the x-intercept of the Nyquist plot was taken as the ohmic resistance. To extract the true kinetic current density (jk), the total current density (j) was corrected using the mass-transfer limited current density (jd) using Eq. (2). jk ¼

j  jd jd  j

ð2Þ

The mass-transfer limited current density can be taken as the current density at 0.4 V. However, one must first verify that this jd is within the theoretical boundary-layer diffusion limited current density at 1600 RPM, which can be calculated from the Levich Equation [68]; jd ¼ 0:62nFDO2=3 x1=2 m1=6 CO

ð3Þ

where F is the Faraday constant, DO is the O2 diffusivity, x is the rotation rate, m is the kinematic viscosity, and CO is the bulk O2 concentration. At room temperature, the theoretical diffusion limited current density is 6.04 mA/cm2 assuming a 4e ORR [69].The mass-transfer limited region (<0.6 V) in Fig. 7 was within 10% of this theoretical value which points out that the ORR occurred primarily through a 4e process. To confirm this, polarization curves were taken at different rotation rates and the current densities at 0.4 V normalized 1=6  CO at different rotation rates were plotted by 0:62D2=3 O m 1/2 against x as in Fig. S5 for Pt catalysts. The linearity of the plot meant that the diffusion limited current can be modeled after the Levich equation. The slopes were used to find n = 3.5–3.7 for Pt/CPPY-X (Table 1). Next, the log of jk was plotted against the electrode potential to make Tafel plots, which are shown in the inset of Fig. 7. The Tafel plots of the Pt/VC and Pt/CPPY-X almost overlap each other signifying a similar ORR mechanism and kinetics for Pt regardless of the support. The experimental plots were fitted linearly in the low overpotential region (>0.9 V) and high overpotential region (<0.85 V). The Tafel slopes in both regimes agreed well with literature values for bulk Pt and Pt/ C: 60 mV/dec and 120 mV/dec at low overpotential and high overpotential, respectively in HClO4 [70]. Thus, the rate determining step (RDS) was the first electron transfer step notwithstanding the support. After normalizing the current density to APt and LPt, jk at 0.9 V gives specific (Is) and mass activity (Im) of Pt (Table 1). The Is and Im of Pt/VC (BASF) were 297 lA/cm2Pt and 0.055 A/ mgPt, respectively. Gasteiger et al. have reported 190–200 lA/ cm2Pt and 0.062–0.069 A/mgPt for 40 wt.% Pt/VC (ETEK) which were measured in O2-saturated 0.1 M HClO4 at 60 C and a 5 mV/s scan rate [71]. Although both catalysts have identical supports, there were not enough details about the Pt

CARBON

6 0 (2 0 13 ) 2 8–40

deposition method and Pt NP characteristics to make a true comparison possible. In addition, the Pt loading and reaction temperatures for their experiments are different from those presented in this work. Nevertheless, Gasteiger et al. showed that the mass activity of Pt/VC (ETEK) decreased from ca. 0.14 A/mgPt to ca. 0.065 A/mgPt when the Pt wt.% increases from 20 to 40, which makes 0.055 A/mgPt for 50 wt.% Pt/VC (BASF) a reasonable result. Compared to Pt/VC (BASF), Pt/CPPY-X had lower specific activities but higher mass activities. The lower specific activities of Pt/CPPY-X relative to Pt/VC (BASF) could be assigned partly to the particle size effect; Pt NPs with size >3 nm have been shown to have higher ORR specific activities than Pt NPs with size <2 nm [59,71–73]. The higher mass activities of Pt/ CPPY-X compared to Pt/VC (BASF) might be the result of better Pt NPs dispersion. Among the Pt/CPPY-X catalysts, Pt/CPPY-4 had the highest Im (0.115 A/mgPt). The difference in mass

37

activities among the Pt/CPPY-X catalysts was likely due to different specific surface areas and specific activities. An increasing order from Pt/CPPY-0 (123 lA/cm2Pt) to Pt/CPPY-8 (226 lA/cm2Pt) was found with regard to specific activities. The difference between these catalysts was the order in the pore structure among CPPY-X supports. Also, XPS showed strong interaction between the catalyst and support in Pt/ CPPY-8 which could be attributed to amorphous nanostructure, large N content, and small size of Pt NPs as discussed earlier. These results implied that the enhanced interaction between the catalyst and the support could be achieved through the structural disorder in nitrogen doped mesoporous carbon. Very high specific activity of Pt/CPPY-8 among the Pt/CPPY-X catalysts could be the result of such interaction between Pt and CPPY-8. Likewise, it could be deduced that Pt/ CPPY-4 had the second highest activity as CPPY-4 had structural disorder in-between those of CPPY-0 and CPPY-8.

Fig. 8 – Cyclic voltammograms of (a) Pt/VC (BASF) and (b) Pt/CPPY-8 in N2 saturated 0.1 M HClO4 at 25 C with a scan rate of 50 mV/s for 1000 cycles for accelerated degradation test (ADT).

38

3.7.

CARBON

6 0 ( 2 0 1 3 ) 2 8 –4 0

Platinum stability

The stability of Pt NPs was investigated using an accelerated degradation test (ADT) where the catalyst was cycled for 1000 cycles between 0 and 1.4 V vs. NHE in 0.1 M HClO4 at 25 C with a scan rate of 50 mV/s. Pt/CPPY-8 and Pt/VC were subjected to ADTs and the CVs after every 100 scans are presented in Fig. 8. A decrease in the areas of Pt–H adsorption/ desorption peaks of both catalysts could be observed as the number of scans increased. The APt decreased as a result of potential cycling. For Pt/VC (BASF), the APt decreased by 32.9% while for Pt/CPPY-8, the APt decreased by 70.3%. The final APt of Pt/CPPY-8 was 48% of the final APt of Pt/VC (BASF). This was unexpected as stronger interaction with the support was expected to increase electrochemical stability of Pt NPs. Fig. S6 shows the TEM image of Pt/CPPY-8 after ADT. The average size of Pt NPs was ca. 3.6 nm, and although some agglomeration was observed, it was very mild. The loss of active surface area in Pt/VC (BASF) is attributed to coarsening of Pt NPs [74]. However, there can be several mechanisms behind the coarsening of Pt NPs, such as Ostwald ripening, Pt migration, detachment from the carbon support and Pt dissolution and redeposition [74]. It also has been shown that Pt catalyzes carbon corrosion through a water–gas type reaction [75], which will be more severe with high Pt loading and high support surface area as in the CPPY-X supports used in this study. In addition, Raman studies suggested that CPPY-X was not as graphitic as VC. Hence, the loss of APt could be primarily due to Pt catalyzed carbon corrosion. This was supported by a decrease of DL current, which was directly proportional to surface area of charging and discharging, by 43.0% in Pt/CPPY-8 compared to 15.5% in Pt/VC (Fig. 8). In addition, the particle size plays a significant role in Pt stability [52,74]. Well dispersed smaller NPs have greater tendency to dissolve than agglomerated larger NPs due to reduction in surface energy [76,77]. Therefore, Pt/ VC (BASF) having 3.7 nm Pt NPs was more stable than Pt/ CPPY-8 having 1.5 nm. However, to fully elucidate the mechanism more work is needed.

4.

Conclusion

The goal of this work was to further our understanding of the structural effects in NMC which affect the ORR activity of the supported catalyst. To achieve this goal, we synthesized NMC with different order in the pore structure by controlling the pore filling in the SBA-15 template, which was accomplished by reducing the pore size of the template by TEOS addition. This resulted in the CPPY-X supports with different order in the pore structure, from ordered (CPPY-0) to disordered (CPPY-8) as revealed by TEM analysis. However, by keeping the carbonization temperature constant, the atomic percentage of surface N and both the types and distribution of functional groups were equivalent in all of the CPPY-X samples. Thus, the surface chemistry was controlled and any changes in interaction between Pt and CPPY-X could only have been due to the structural differences. Raman and XRD indicated that the microcrystallinity was also different in CPPY-X, with CPPY-8, perhaps having smallest carbon crystallites.

Therefore, we could conclude that template based carbon synthesis also allowed control of the microcrystalline nature of the carbon. The effect of this order in the pore structure among the CPPY-X supports on interaction with Pt NPs was underscored by the shift in BE of Pt NPs in Pt/CPPY-8. Hence, the structure based interaction between catalyst and support was established. Also, small Pt NPs size and high N content were thought to be critical in highlighting this pore structure-based interaction as small Pt NPs size increased specificity with the support and the high nitrogen content introduced highly active centers through aggregation of nitrogen functionalities. Because of this enhanced interaction, Pt/CPPY-8 had the highest ORR activity among the Pt/CPPY-X catalysts. This provides a way to tailor nitrogen-doped carbon supports for higher ORR activity on Pt NPs.

Acknowledgment The authors would like to thank the U.S. Department of Energy, Office of Basic Energy Sciences who supported this work under Grant No. DE-FG02-10ER16200.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.carbon. 2013.03.053.

R E F E R E N C E S

[1] Yuan XZ, Wang H. PEM fuel cell fundamentals. In: Zhang J, editor. PEM fuel cell electrocatalysts and catalyst layers: fundamental and applications. London: Springer; 2008. p. 1–79. [2] Kinoshita K. Carbon: electrochemical and physicochemical properties. New York: John Wiley & Sons; 1988. p. 1–2. [3] Shrestha S, Liu Y, Mustain WE. Electrocatalytic activity and stability of Pt clusters on state-of-the-art supports: a review. Catal Rev Sci Eng 2012;53(3):256–336. [4] Kitchin JR, Nørskov JK, Barteau MA, Chen JG. Modification of the surface electronic and chemical properties of Pt(1 1 1) by subsurface 3d transition metals. J Chem Phys 2004;120(21):10240–6. [5] Hammer B, Nørskov JK. Theoretical surface science and catalysis-calculations and concepts. Adv Catal 2000;45:71–129. [6] Greeley J, Norskov JK, Mavrikakis M. Electronic structure and catalysis on metal surfaces. Annu Rev Phys Chem 2002;53:319–48. [7] Zhou Y, Holme T, Berry J, Ohno T, Ginley D, O’Hayre R. Dopant-induced electronic structure modification of of HOPG surfaces: implications for high activity fuel cell catalysts. J Phys Chem C 2010;114(1):506–15. [8] Zhou Y, Neyerlin K, Olson TS, Pylypenko S, Bult J, Dinh HN, et al. Enhancement of Pt and Pt-alloy fuel cell catalyst activity and durability via nitrogen-modified carbon supports. Energy Environ Sci 2010;3(10):1437–46. [9] Su F, Tian Z, Poh CK, Wang Z, Lim SH, Liu Z, et al. Pt nanoparticles supported on nitrogen-doped porous carbon

CARBON

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19] [20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

6 0 (2 0 13 ) 2 8–40

nanospheres as an electrocatalyst for fuel cells. Chem Mater 2010;22(3):832–9. Li X, Park S, Popov BN. Highly stable Pt and PtPd hybrid catalysts supported on a nitrogen-modified carbon composite for fuel cell application. J Power Sources 2010;195(2):445–52. Herrmann-Geppert I, Radnik J, Kramm UI, Fiechter S, Bogdanoff P. Improved platinum electrocatalyst for the oxygen reduction reaction using nitrogen-modified carbon support. ECS Trans 2011;41(1):1161–71. Liu SH, Chen SC, Sie WH. Heat-treated platinum nanoparticles embedded in nitrogen-doped ordered mesoporous carbons: synthesis, characterization and their electrocatalytic properties toward methanol-tolerant oxygen reduction. Int J Hydrogen Energy 2011;36(23):15060–7. Liu SH, Wu MT, Lai YH, Chiang CC, Yu N, Liu SB. Fabrication and electrocatalytic performance of highly stable and active platinum nanoparticles supported on nitrogen-doped ordered mesoporous carbon for oxygen reduction reaction. J Mater Chem 2011;21(33):12489–96. Czerw R, Terrones M, Charlier JC, Blase´ X, Foley B, Kamalakaran R, et al. Identification of electron donor states in N-doped carbon nanotubes. Nano Lett 2001;1(9):457–60. Ma X, Wang E, Zhou W, Jefferson DA, Chen J, Deng S, et al. Polymerized carbon nanobells and their field-emission properties. Appl Phys Lett 1999;75(20):3105–7. Kurt R, Karimi A. Influence of nitrogen on the growth mechanism of decorated C:N nanotubes. ChemPhysChem 2001;2(6):388–92. Maldonado S, Stevenson KJ. Influence of nitrogen doping on oxygen reduction electrocatalysis at carbon nanofiber electrodes. J Phys Chem B 2005;109(10):4707–16. Sjo¨stro¨m H, Stafstro¨m S, Boman M, Sundgren JE. Superhard and elastic carbon nitride thin films having fullerenelike microstructure. Phys Rev Lett 1995;75(7):1336–9. Stafstro¨m S. Reactivity of curved and planar carbon–nitride structures. Appl Phys Lett 2000;77(24):3941–3. Joo SH, Choi SJ, Oh I, Kwak J, Liu Z, Terasaki O, et al. Ordered nanoporous arrays of carbon supporting high dispersions of platinum nanoparticles. Nature 2001;412:169–72. Joo SH, Pak C, You DJ, Lee SA, Lee HI, Kim JM, et al. Ordered mesoporous carbons (OMC) as supports of electrocatalysts for direct methanol fuel cells (DMFC): effect of carbon precursors of OMC on DMFC performances. Electrochim Acta 2006;52(4):1618–26. Lin ML, Huang CC, Lo MY, Mou CY. Well-ordered mesoporous carbon thin film with perpendicular channels: application to direct methanol fuel cell. J Phys Chem C 2008;112(3):867–73. Joo SH, Kwon K, You DJ, Pak C, Chang H, Kim JM. Preparation of high loading Pt nanoparticles on ordered mesoporous carbon with a controlled Pt size and its effects on oxygen reduction and methanol oxidation reactions. Electrochim Acta 2009;54(24):5746–53. Gupta G, Slanac DA, Kumar P, Wiggins-Camacho JD, Kim J, Ryoo R, et al. Highly stable Pt/ordered graphitic mesoporous carbon electrocatalysts for oxygen reduction. J Phys Chem C 2010;114(24):10796–805. Song S, Liang Y, Li Z, Wang Y, Fu R, Wu D, et al. Effect of pore morphology of mesoporous carbons on the electrocatalytic activity of Pt nanoparticles for fuel cell reactions. Appl Catal B 2010;98(3–4):132–7. Tintula KK, Sahu AK, Shahid A, Pitchumani S, Sridhar P, Shukla AK. Directed synthesis of MC-PEDOT composite catalyst-supports for durable PEFCs. J Electrochem Soc 2011;158(6):B622–31. Wu Z, Lv Y, Xia Y, Webley PA, Zhao D. Ordered mesoporous platinum@graphitic carbon embedded nanophase as a highly

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

[41]

[42] [43] [44]

[45]

39

active, stable, and methanol-tolerant oxygen reduction electrocatalyst. J Am Chem Soc 2012;134(4):2236–45. Zeng J, Francia C, Dumitrescu MA, Videla AHAM, Ijeri VS, Specchia S, et al. Electrochemical performance of Pt-based catalysts supported on different ordered mesoporous carbons (Pt/OMCs) for oxygen reduction reaction. Ind Eng Chem Res 2012;51(22):7500–9. Zhao D, Feng J, Huo Q, Melosh N, Fredrickson GH, Chmelka BF, et al. Triblock copolymer syntheses of mesoporous silica with periodic 50 to 300 angstorm pores. Science 1998;279(5350):548–52. Pels JR, Kaptejin F, Moulijn JA, Zhu Q, Thomas KM. Evolution of nitrogen functionalities in carbonaceous materials during pyrolysis. Carbon 1995;33(11):1641–53. Yang CM, Weidenthaler C, Spliethoff B, Mayanna M, Schu¨th F. Facile template synthesis of ordered mesoporous carbon with polypyrrole as carbon precursor. Chem Mater 2005;17(2):355–8. Fuertes AB, Centeno TA. Mesoporous carbon with graphitic structures fabricated by using porous silica materials as templates and iron-impregnated polypyrrole as precursor. J Mater Chem 2005;15(10):1079–83. Shrestha S, Mustain WE. Properties of nitrogenfunctionalized ordered mesoporous carbon prepared using polypyrrole precursor. J Electrochem Soc 2010;157(11):B1665–72. Oh HS, Oh JG, Hong YG, Kim H. Investigation of carbonsupported Pt nanocatalyst preparation by the polyol process for fuel cell applications. Electrochim Acta 2007;52(25):7278–85. Kruk M, Jaroniec M, Sakamoto Y, Terasaki O, Ryoo R, Ko CH. Determination of pore size and pore wall structure of MCM41 by using nitrogen adsorption, transmission electron microscopy, and X-ray diffraction. J Phys Chem B 2000;104(2):292–301. Jaroniec M, Kruk M, Oliver JP. Standard nitrogen adsorption data for characterization of nanoporous silicas. Langmuir 1999;15(16):5410–3. Kruk M, Jaroniec M, Gadkaree KP. Nitrogen adsorption studies of novel synthetic active carbons. J Colloid Interface Sci 1997;192(1):250–6. Xie Y, Sherwood PMA. X-ray photoelectron-spectroscopic studies of carbon fiber surfaces. 11Differences in the surface chemistry and bulk structure of different carbon fibers based on poly(acrylonitrile) and pitch and comparison with various graphite samples. Chem Mater 1990;2(3):293–9. Garsany Y, Baturina OA, Swinder-Lyons KE, Kocha SS. Experimental methods of quantifying the activity of platinum electrocatalysts for the oxygen reduction reaction. Anal Chem 2010;82(15):6321–8. Garsany Y, Singer IL, Swinder-Lyons KE. Impact of film-drying procedures on RDE characterization of Pt/VC electrocatalysts. J Electroanal Chem 2011;662(2):396–406. Jaroniec M, Kruk M, Jaroniec CP, Sayari A. Modification of surface and structural properties of ordered mesoporous silicates. Adsorption 1999;5(1):39–45. Brinker CJ. Hydrolysis and condensation of silicates: effects on structure. J Non-Cryst Solids 1988;100(1–3):31–50. Ryoo R, Joo SH, Kruk M, Jaroniec M. Ordered mesoporous carbons. Adv Mater 2001;13(9):677–81. Jun S, Joo SH, Ryoo R, Kruk M, Jaroniec M, Liu Z, et al. Synthesis of new, nanoporous carbon with hexagonally ordered mesostructure. J Am Chem Soc 2000;122(43): 10712–3. Fuertes AB, Nevskaia DM. Control of mesoporous structure of carbons synthesized using a mesostructured silica as template. Microporous Mesoporous Mater 2003;62(3):177–90.

40

CARBON

6 0 ( 2 0 1 3 ) 2 8 –4 0

[46] Jawhari T, Roid A, Casado J. Raman spectroscopic characterization of some commercially available carbon black materials. Carbon 1995;33(11):1561–5. [47] Ferrari AC, Robertson J. Interpretation of Raman spectra of disordered and amorphous carbon. Phys Rev B 2000;61(20):14095–107. [48] Knight DS, White WB. Characterization of diamond films by Raman spectroscopy. J Mater Res 1989;4(2):385–93. [49] Dresselhaus MS, Jorio A, Hofmann M, Dresselhaus G, Saito R. Perspectives on carbon nanotubes and graphene Raman spectroscopy. Nano Lett 2010;10(3):751–8. [50] Biniak S, Szyman´ski G, Siedlewski J, S´wia˛tkowski A. The characterization of activated carbons with oxygen and nitrogen surface groups. Carbon 1997;35(12):1799–810. [51] Yu X, Yu S. Recent advances in activity and durability enhancement of Pt/C catalytic cathode in PEMFC:part I. Physico-chemical and electronic interaction between Pt and carbon support, and activity enhancement of Pt/C catalyst. J Power Sources 2007;172(1):133–44. [52] Bezerra CWB, Zhang L, Liu H, Lee K, Marques ALB, Marques EP, et al. A review of heat-treatment effects on stability of PEM fuel cell catalysts for oxygen reduction reaction. J Power Sources 2007;173(2):891–908. [53] Yu R, Chen L, Liu Q, Lin J, Tan K-L, Ng SC, et al. Platinum deposition on carbon nanotubes via chemical modification. Chem Mater 1998;10(3):718–22. [54] Groves MN, Chan ASW, Malardier-Jugroot C, Jugroot M. Improving platinum catalyst binding energy to graphene through nitrogen doping. Chem Phys Lett 2009;481(4– 6):214–9. [55] Gong K, Du F, Xia Z, Durstock M, Dai L. Nitrogen-doped carbon nanotube arrays with high electrocatalytic activity for oxygen reduction. Science 2009;323(5915):760–4. [56] Holme T, Zhou Y, Pasquarelli R, O’Hayre R. First principles study of doped carbon supports for enhanced platinum catalysts. Phys Chem Chem Phys 2010;12(32):9461–8. [57] Mavrikakis M, Hammer B, Nørskov JK. Effect of strain on reactivity of metal surfaces. Phys Rev Lett 1998;81(13):2819–22. [58] Eberhardt W, Fayet P, Cox DM, Fu Z, Kaldor A, Sherwood R, et al. Photoemission from mass-selected monodispersed Pt clusters. Phys Rev Lett 1990;64(7):780–3. [59] Toyoda E, Jinnouchi R, Hatanaka T, Morimoto Y, Mitsuhara K, Visikovskiy A, et al. The d-band structure of Pt nanoclusters correlated with the catalytic activity for an oxygen reduction reaction. J Phys Chem C 2011;115(43):21236–40. [60] Richter B, Kuhlenbeck H, Freund HJ, Bagus PS. Cluster corelevel binding-energy shifts: the role of lattice strain. Phys Rev Lett 2004;93(2):026805. [61] Corcoran J, Tavassol H, Rigsby MA, Bagus PS, Wieckowski A. Application of XPS to study electrocatalysts for fuel cells. J Power Sources 2010;195(24):7856–79. [62] Zhou WP, Lewera A, Larsen R, Masel RI, Bagus PS, Wieckowski A. Size effects in electronic and catalytic properties of

[63] [64]

[65]

[66] [67]

[68]

[69]

[70]

[71]

[72]

[73]

[74]

[75]

[76]

[77]

unsupported palladium nanoparticles in electroxidation of formic acid. J Phys Chem B 2006;110(27):13393–8. Deng JF, Li H, Wang W. Progress in design of new amorphous alloy catalysts. Catal Today 1999;51(1):113–25. Hamann CH, Hamnett A, Vielstich W. Electrochemistry. Weinheim: Germany: WILEY-VCH Verlag GmbH & Co. KGaA; 2007. Schmidt TJ, Gasteiger HA, Sta¨b GD, Urban PM, Kolb DM, Behm RJ. Characterization of high-surface-area electrocatalysts using a rotating disk electrode configuration. J Electrochem Soc 1998;145(7):2354–8. Koh M, Nakajima T. Adsorption of aromatic compounds on CxN-coated activated carbon. Carbon 2000;38(14):1947–54. Be´guin F, Szostak K, Lota G, Frackowiak E. A self-supporting electrode for supercapacitors prepared by one-step pyrolysis of carbon nanotube/polyacrylonitrile blends. Adv Mater 2005;17(19):2380–4. Bard AJ, Faulkner LR. Electrochemical methods: fundamentals and applications. New York: John Wiley & Sons; 1988. Markovic´ NM, Gasteiger HA, Grgur BN, Ross PN. Oxygen reduction reaction on Pt (1 1 1): effect of bromide. J Electroanal Chem 1999;467(1–2):157–63. Paulus UA, Schmidt TJ, Gasteiger HA, Behm RJ. Oxygen reduction on a high-surface area Pt/Vulcan carbon catalyst: a thin-film rotating ring-disk electrode study. J Electroanal Chem 2001;495(2):134–45. Gasteiger HA, Kocha SS, Sompalli B, Wagner FT. Activity benchmarks and requirements for Pt, Pt-alloy, and non-Pt oxygen reduction catalysts for PEMFCs. Appl Catal B 2005;56(1–2):9–35. Mayrhofer KJJ, Blizanac BB, Arenz M, Stamenkovic VR, Ross PN, Markovic NM. The impact of geometric and surface electronic properties of Pt-catalysts on the particle size effect in electrocatalysis. J Phys Chem B 2005;109(30):14433. Shao M, Peles A, Shoemaker K. Electrocatalysis on platinum nanoparticles: particle size effect on oxygen reduction reaction activity. Nano Lett 2011;11(9):3714–9. Shao-Horn Y, Sheng WC, Chen S, Ferreira PJ, Holby EF, Morgan D. Instability of supported platinum nanoparticles in lowtemperature fuel cells. Top Catal 2007;46(3–4):285–305. Stevens DA, Hicks MT, Haugen GM, Dahn JR. Ex situ and in situ stability studies of PEMFC catalysts. J Electrochem Soc 2005;152(12):A2309–15. Makharia R, Kocha SS, Yu PT, Sweikart MA, Gu W, Wagner FT, et al. Durable PEM fuel cell electrode materials: requirements and benchmarking methodologies. ECS Trans 2006;1(8): 3–18. Holby EF, Sheng W, Shao-Horn Y, Morgan D. Pt nanoparticle stability in PEM fuel cells; influence of particle size distribution and crossover hydrogen. Energy Environ Sci 2009;2(8):865–87.