Pd supported on carbon containing nickel, nitrogen and sulfur for ethanol electrooxidation

Carbon material containing nickel, nitrogen and sulfur (Ni-NSC) has been synthesized using metal-organic frameworks (MOFs) as precursor by annealing treatment with a size from 200 to 300 nm. Pd nanoparticles supported on the Ni-NSC (Pd/Ni-NSC) are used as electrocatalysts for ethanol oxidation in alkaline media. Due to the synergistic effect between Pd and Ni, S, N, free OH radicals can form on the surface of Ni, N and S atoms at lower potentials, which react with CH3CO intermediate species on the Pd surface to produce CH3COO− and release the active sites. On the other hand, the stronger binding force between Pd and co-doped N and S is responsible for enhancing dispersion and preventing agglomeration of the Pd nanoparticles. The Pd(20 wt%)/Ni-NSC shows better electrochemical performance of ethanol oxidation than the traditional commercial Pd(20 wt%)/C catalyst. Onset potential on the Pd(20 wt%)/Ni-NSC electrode is 36 mV more negative compared with that on the commercial Pd(20 wt%)/C electrode. The Pd(20 wt%)/Ni-NSC in this paper demonstrates to have excellent electrocatalytic properties and is considered as a promising catalyst in alkaline direct ethanol fuel cells.

sulfur (Ni-NSC), which can be easily synthesized. Through carbonization and doping steps, the Pd nanoparticles supported on the Ni-NSC (Pd/Ni-NSC) can be successfully obtained. The as-prepared porous Pd/Ni-NSC materials will be used as catalysts for ethanol electrooxidation. Zhong et al. have studied the PdNi/C nanoparticles by X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS) and revealed the presence of metallic Ni and the oxide phases NiO, Ni(OH) 2 , NiOOH in the catalyst matrix 21 . So PdNi materials have been used as electrocatalysts for the ethanol oxidation reaction in alkaline media 14,26,27 . Niu et al. 28 have reported that small palladium nanoparticles (Pd NPs) inside sulfur-doped carbon microsphere (S-CMS) show both high electrocatalytic activity and long durability for methanol oxidation reaction (MOR) in the DMFCs. Wei et al. have studied the interaction between nitrogen in the carbon nanotubes (CNTs) and Pd in the catalysts 29 . The nitrogen-doped CNTs-supported Pd catalysts exhibit superior electrochemical activity for ethanol oxidation relative to the pristine CNTs. Zhang et al. 30 have reported that the introduced nitrogen and sulfur co-doping could generate abundant active sites on the graphene surface supported Pd nanoparticles (Pd/NS-rGO) and the Pd/NS-rGO catalyst also reveals superior anti-poisoning ability and amazing stability.  Fig. 1b. The characteristic D and G bands of graphitic carbon can be seen at 1340 and 1589 cm −1 in both two materials, which could be assigned to the defective and graphitic structure of the carbon materials 31 . D band arises from disordered sp 3 defect sites and is also associated with defects and disorder in carbon structure 32 . The degree of graphitization of carbon materials can be quantified by the intensity ratio of D to G bands. In this case, those defects will in turn act as active sites and reinforce the interaction between Pd particles and support, and thus increase the catalysts loading and particles dispersion. The I D /I G ratio of Pd(20 wt%)/Ni-NSC is 1.05, which is smaller than that of Ni-NSC(1.11). It may be attributable to the supporting Pd, which can decrease defect sites and disorder in carbon structure.

Results and Discussion
The Fourier transform infrared (FTIR) spectra in Fig. 2 confirm various surface groups of Ni-NSC. The peaks at 3570 cm −1 in the spectra should be considered as the O-H stretching vibration. The peak at 1755 cm −1 is attributed to the C=C and C=N stretching vibration in the conjugated structure 33 . In addition, two obvious peaks at 2414 and 2389 cm −1 in the spectrum of Ni-NSC should be ascribed to the -N=C=O and -N-C-S stretching vibration, respectively. It indicates that the N and S atoms have been doped to the Ni-NSC. The absorbance from 970 to 1200 cm −1 is attributed to the stretching vibrations of the C-O-C and C-S-C bond 34 . Besides, the bending vibration of the C-S bond also shows a weak peak at 646 cm −1 , which confirms the successful doping of sulfur element. Moreover, surface groups of Pd(20 wt%)/Ni-NSC compared with that of Ni-NSC in FTIR spectra have a weaker peak, which is due to Pd nanoparticles growing on the Ni-NSC surface and weakening the absorption intensity.
Chemical bonding states in the Pd(20 wt%)/Ni-NSC were analyzed by XPS (Fig. 3). From the compared survey of Ni-NSC and Pd(20 wt%)/Ni-NSC shown in Fig. 3a, the peaks are corresponding to existence of Pd 3d, Ni 2p, C 1 s, N 1 s and S 2p, respectively. Ni 2p spectrum presents a complicated structure with two strong satellite peaks adjacent to the main peaks as shown in Fig. 3b, which is due to multi-electron excitation. The presence of multiple-valence nickel element in the Pd(20 wt%)/Ni-NSC is apt to occur redox reaction due to electron transfer. The binding energy values of XPS spectrum of Pd 3d display a doublet that is composed of a low-energy band (Pd 3d 5/2 ) and a high-energy band (Pd 3d 3/2 ) at 335.4 and 340.7 eV (Fig. 3c), which are similar to the typical Pd 0 specie as previous report 35 . These above data show that Pd specie attached to the surface of Ni-NSC exists in the form of Pd 0 . From the high resolution peak split in Fig. 3d, the C 1 s spectrum is deconvoluted into four peaks, sp 2 -hybridized graphite-like carbon C=C at 284.6 eV, sp 3   be resolved into four sub-peaks due to the spin-orbit coupling, including pyridinic-N (398.5 eV), pyrrolic-N (399.7 eV), graphitic-N (400.7 eV) and pyridine-N-oxide groups (401.7 eV) 37 . Besides, the S 2p spectrum can be deconvoluted to one pair of spin-orbit doublet at 161.0 (S 2p 3/2 ) and 162.9 eV (S 2p 5/2 ) in Fig. 3f, indicating the existence of metal sulfide and formation of C-S-C at 164.1 eV 37 . These results of the existence of mentioned-above chemical bonds illustrate that Ni, N and S atoms are doped well into the carbon material to form Ni-NSC, in which Pd nanoparticles are inserted.
The as-synthesized microporous Ni-NSC and Pd(20 wt%)/Ni-NSC materials were characterized using scanning electron microscopy (SEM) (Fig. 4a,b). It is found that the size of the Ni-NSC and Pd(20 wt%)/Ni-NSC particles is both ranged from 200 to 300 nm. The typical transmission electron microscopy (TEM) and high-resolution TEM (HRTEM) images for Pd(20 wt%)/Ni-NSC are shown in Fig. 4c,d. It is clearly observed that deep shadow particles on microporous Ni-NSC known as Pd nanoparticles exhibit a spherical-like shape with well-dispersion. As shown in Fig. 4e, it clearly reveals a lattice spacing of 0.218 and 0.221 nm, which correspond to the distance of the (111) crystal plane of Ni 4 N and the (204) one of NiC x . Besides, one lattice fringe measured from the image can be found around 0.226 nm, which corresponds to the distance of the (111) crystal plane of Pd. Figure 4f Fig. 5, the Ni-NSC shows poor performance for ethanol oxidation. Compared with the CV in the absence of ethanol, an ethanol oxidation peak can be clearly observed in the CV curve on the Pd(20 wt%)/Ni-NSC electrode in the presence of 1.0 mol L −1 ethanol. It is obviously that the activity of ethanol oxidation on the Pd(20 wt%)/Ni-NSC electrode is much higher than that on the commercial Pd(20 wt%)/C electrode. The onset potential (E s ) is −0.628 V on the Pd(20 wt%)/Ni-NSC electrode, which is 36 mV more negative compared with that on the commercial Pd(20 wt%)/C electrode (−0.592 V). The lower value of E s shows easier electrooxidation of ethanol. The current for ethanol electrooxidation on the Pd(20 wt%)/Ni-NSC electrode begins to rise much more sharply at more negative potential than that on the commercial Pd(20 wt%)/C electrode. It demonstrates that ethanol can be more easily electrochemically oxidized on the Pd(20 wt%)/Ni-NSC electrode than that on the commercial Pd(20 wt%)/C electrode. The peak current density (j p ) is 110.3 mA cm −2 (−0.293 V) on the Pd(20 wt%)/Ni-NSC electrode, which is 2.6 times as high as that on the commercial Pd(20 wt%)/C electrode, which is 41.8 mA cm −2 (−0.316 V). Current density at the potential of 0.3 V (j −0.3V ) is 108.1 mA cm −2 on the Pd(20 wt%)/Ni-NSC electrode, which is 2.8 times as high as that on the commercial Pd(20 wt%)/C electrode, which is 38.0 mA cm −2 .
In addition, the compared Tafel plots of ethanol electrooxidation, calculated from the quasi-steady-state curve obtained in 1.0 mol L −1 KOH containing 1.0 mol L −1 ethanol solution with a sweep rate of 2 mV s -1 at a Pd loading of 0.10 mg cm −2 , show linear region with the respective Tafel slope of 190.24 mV dec -1 on the Pd(20 wt%)/ Ni-NSC electrode and 244.21 mV dec −1 on the commercial Pd(20 wt%)/C electrode as shown in Fig. 6. The Tafel value on the Pd(20 wt%)/Ni-NSC electrode is lower than that on the commercial Pd(20 wt%)/C electrode, which shows that ethanol oxidation occurs favourably on the Pd(20 wt%)/Ni-NSC electrode. In ethanol electrooxidation controlled by the adsorption of OH ads , the expected Tafel slope is 120 mV dec -1 and a positive deviation may be explained by the formation of an inactive oxide layer on the palladium surface and by the 2D structure of the catalytic layer 38 .    Fig. 7. It is well-known that the intermediate species such as CO-like species during alcohol oxidation will block the electrode surface to poison the catalyst, then the current of alcohol oxidation will decrease 39 . Nevertheless, at the end of the test,   the oxidation current density is 2.5 mA cm −2 on the Pd(20 wt%)/Ni-NSC electrode which is larger than that on the commercial Pd(20 wt%)/C electrode (1.8 mA cm −2 ). The good stability of Pd/Ni-NSC catalyst could be relevant with well-dispersed Pd particles, the stronger binding force between Pd and co-doped N and S to inhibit agglomeration of particles 29,30 . It is generally considered that co-doped two-types N and S are responsible for the interaction with metal catalysts and prevent them from aggregation together 30,40 .
The ethanol electrooxidation on Pd-based electrocatalysts is as following step 26 The reaction (4) is the rate-determining step. The (CH 3 CO) ads is adsorbed onto the surface of the Pd and it can block the active sites and slows the reaction kinetics. The ethanol oxidation on the Pd/NS-rGO may be explained as a bi-functional mechanism 10,41 . Density functional theory calculations reveal that the improved activity and stability stems from the promoted production of free OH radicals (OH ads , on Ni active sites) which facilitate the oxidative removal of carbonaceous poison and combination with CH 3 CO radicals on adjacent Pd active sites 14 . The electron-accepting N and S species can impart a relatively high positive charge density on neighboring OH − in the alkaline media, the OH ads species can be more easy formed on the surface of N or S atoms than that on the surface of Pd at lower potentials as following reactions.
The reaction (6) happens more easily than reaction (4). Pd acts as main catalyst for catalysing the dehydrogenation of ethanol during the oxidation reaction and free OH radicals (OH ads ) can form on the N or S atoms surface at lower potentials. These oxygen containing species react with (CH 3 CO) ads intermediate species on the Pd surface to produce CH 3 COO − and release the active sites. On the other hand, the introduction of N and S atoms significantly changes the electronic structure of carbon materials and the supported Pd nanoparticle. Therefore, due to the electrondonating effects of N and S atoms, the electron cloud density of Pd may increase, which can stabilize Pd, and the N and S groups impart a basic nature to the carbon surface and bind strongly to Pd, enhancing the Pd nanoparticls dispersion and preventing agglomeration of the Pd particles, thereby improving the electrochemical activity and stability of the Pd-based catalysts 29,42 .
In summary, highly active Pd/Ni-NSC materials obtained by heating method and chemical reduction from vanillic thiosemicarbazone ligand (L = C 9 H 10 N 3 O 2 S) and its coordination compound NiL 2 as precursors to obtain porous carbon materials have been successfully developed and demonstrated as excellent catalysts for efficient ethanol electrooxidation in alkaline media. Thanks to the synergistic effect between Pd and Ni, S, N, free OH radicals can form on the surface of Ni, N and S atoms at lower potentials, which react with (CH 3 CO) ads intermediate species on the Pd surface to produce CH 3 COO − and release the active sites. On the other hand, the electrondonating effects of N and S atoms enhance dispersion and prevent agglomeration of the Pd nanoparticls. The Pd/Ni-NSC shows extraordinary catalytic activity and stability for electrochemical oxidation of ethanol, more competitive than those of the traditional commercial Pd/C catalyst. This work provides new insights into using transition metal oxide and renewable carbon materials as high performance electrocatalyst for alkaline DEFCs.

Methods
Materials synthesis. All reagents were of commercially available and analytical grade (AR). All chemicals were purchased from Aladdin and used as received. Vanillin thiosemicarbazone ligands, Ni-vanillic thiosemicarbazone were synthesized according to the literature 40 . Ni-NSC materials were prepared by pyrolyzing the obtained NiL 2 composite. In a typical procedure, a ceramic boat containing Ni-NSC (500 mg) was placed in a quartz tube, heated under N 2 atmosphere at a ramping rate of 3 °C min −1 , and kept at 500 °C for 2 h. Pd/Ni-NSC was prepared by reduction Pd(NH 3 ) 4 Cl 2 on the Ni-NSC powders using an excess 0.01 mol L −1 NaBH 4 solution. The ratio of Pd and Ni-NSC was controlled by stoichiometric calculation with the same weight ratio of commercial Pd/C as 20 to 80.
Characteriazation. XRD was carried out using a X'Pert powder X-ray diffractometer with Cu K α radiation (λ = 0.15418 nm). Raman spectra were obtained using laser confocal micro-Raman spectroscopy (LabRAM HR800, Horiba Jobin Yvon). FTIR spectra were collected using a Tensor27 (Bruker) spectrometer. XPS measurements were performed in an ESCALAB 250 spectrometer under vacuum (about 2 × 10 −9 mbar). Field-emission SEM images were conducted on a Quanta 400 FEG microscope (FEI Company). TEM images corresponding element energy-dispersive X-ray spectrometer (EDS) were carried out on a JEOL JEM-2010 (JEOL Ltd.). All electrochemical measurements were tested in a three-electrode cell using a CHI 760e electrochemical work station at 25 °C. Solutions were freshly prepared before each experiment. A platinum foil (3.0 cm 2 ) was used as counter electrode. All the potentials were measured versus a saturated calomel electrode (SCE, 0.241 V vs. NHE) electrode. CV data were recorded between −0.95 to 0.40 V with a scan rate at 50 mV s −1 .