Electrospun carbon nitride supported on poly(vinyl) alcohol as an electrocatalyst for oxygen reduction reactions

Abstract

Electrochemical oxygen reduction reactions (ORRs) via nonprecious catalysts have potential for significant cost reduction in fuel cells. Dense, multi-layered poly(vinyl) alcohol (PVA) nanofibers dispersed with catalytically active carbon nitride (CN_x) nanoparticles were synthesized using an electrospinning process. The size, morphology, elemental composition, and bond structure of the CN_x/PVA nanofibers were analysed using TEM, SEM, FTIR, XPS and Raman spectroscopic studies. Significant improvement in the electrocatalytic activity of the CN_x nanoparticles dispersed in the nanofibers as compared to their native form was observed towards the ORR using voltammetry coupled with FTIR studies. The onset potential and peak current density observed for the CN_x/PVA nanofibers using cyclic voltammetry were comparable to those of a conventional Pt/C (40 : 60% by weight) catalyst. The ORR mechanism was further analysed using a RRDE and in situ FTIR with linear sweep voltammetry studies. RRDE analysis confirmed that the ORR takes place primarily via a 4-electron pathway. The catalytic activity of the CN_x/PVA nanofibers for the ORR was stable over 5000 repetitions in the voltammetric studies coupled with FTIR.

---

1. Introduction

In recent years, the research interest in fuel cells has increased as they are promising energy devices not only due to their wide range of applications but also due to their numerous advantages over conventional energy devices.¹ The oxygen reduction reaction (ORR), being the slowest reaction, is at the heart of studies for the development of cathode catalysts as it dominates the overall performance of proton exchange membrane fuel cells. To date, catalysts that have been proven to be the best for the ORR contain platinum. However, certain factors such as CO poisoning, Pt scarcity and above all, high cost are big obstacles in the commercialisation of fuel cells.^2–4 Hence, recently the focus of the research has shifted to non-platinum catalyst materials.^5–8 In nitrogen rich carbon materials, the ORR occurs via a more efficient one step 4-electron pathway in an acid medium.^4,9–11 The high electron affinity of the adjoining nitrogen atom induces a positive charge density on the carbon atoms. It improves the adsorption of oxygen on the carbon atoms and thus weakens the inherent oxygen bonding facilitating 4-electron transfer in the ORR.¹²

Carbon nitride (CN_x) as a nitrogen rich carbon material is highly stable, oxidation resistant and it can be used as an economically viable catalyst for ORRs, but requires further improvement.^4,9–13 Agglomeration and a low surface area limit its catalytic applications.¹⁴ In order to enhance its activity, CN_x nanoparticles are uniformly dispersed and embedded in polyvinyl alcohol (PVA) nanofibers using an electrospinning technique.^15–17 There are different techniques to prepare CN_x nanofibers and nanotubes using various methods such as template methods^18–21 and direct pyrolysis of the precursor ferrocene/melamine mixtures.¹⁶ The template method has the drawback of making continuous nanofibers.²² Pyrolysis involves difficulty in the preparation of the CN_x structure within the nanofiber matrix.¹⁶ It is evident from recent research studies that electrospinning has unique advantages over the nanofibres formed by other methods.^22–27 In electrospinning, nanofibers are prepared from a polymer solution, e.g., PVA, and the electrostatic field stretches the polymer into fibers as the solvent is evaporated. PVA is chemically and thermally stable, non-toxic and hydrophilic in nature. Its flexibility and toughness characteristics augment the physical properties of the nanofibers.²⁸

Nanofibers composed of CN_x nanoparticles are synthesized using an electrospinning technique and tested for the ORR. PVA is chosen as the supporting polymer because of its elastic properties and as it is conducive to electrospinning as mentioned above. Physical characterization is carried out for both the CN_x nanoparticles and nanofibers to understand the morphology, elemental composition and bond structure. The catalytic activity of the CN_x nanofibers over the native CN_x nanoparticles is observed using voltammetry and compared with the catalytic activity of a conventional Pt/C (40 : 60% by weight) catalyst. The ORR mechanism is analysed using a RRDE and in situ FTIR studies combined with voltammetry.

2. Experimental

2.1. Materials

Ethylene diamine (Merck) and carbon tetrachloride (Merck) were used as precursors for the preparation of carbon nitride. Ethanol (Merck) was used to wash the carbon nitride nanoparticles formed. Poly(vinyl) alcohol (molecular weight: 85 000 to 124 000, Sigma Aldrich) was used as the supporting polymer to enable electrospinning of carbon nitride. Deionized water (18.2 MΩ cm) was used as the solvent for electrospinning. 2-Propanol (Merck) was used to disperse the carbon nitride nanoparticles and Pt/C (Alfa Aesar) to make catalyst ink. A Nafion® dispersion (DE-521, DuPont, USA) was used as a binder. Commercial platinum/carbon (Pt/C) (Pt loading: 40 wt%, Pt on carbon black) was purchased from Alfa Aesar.

2.2. Preparation of the CN_x nanoparticles

Carbon nitride was prepared from the chemical precursors ethylene diamine (EDA) and carbon tetrachloride (CCl₄) using the method proposed by Qiu et al.^29,30 First, 110 mL of EDA and 90 mL of CCl₄ were mixed for 12 hours. The mixture was then refluxed at 90 °C in the presence of nitrogen for 6 hours. As a result, a black viscous liquid was formed. The liquid was vacuum dried at 120 °C for 24 hours. After grinding the nanoparticles, calcination was performed at 600 °C for 5 hours under a nitrogen atmosphere. The black lustrous material formed, named as the native CN_x nanoparticles, was washed several times with ethanol and then dried at 100 °C.

2.3. Electrospinning

For the electrospinning of PVA alone, 3 mL of a homogenous solution of 8% (w/v) PVA in water was used. For the CN_x/PVA nanofiber system, firstly 0.15 g of the native CN_x nanoparticles were ultrasonicated in 3 mL of water (5% w/v) for 30 minutes at room temperature. PVA (0.24 g) was added to the CN_x nanoparticle/water dispersion such that a 5 : 8 (wt/wt) ratio of CN_x to PVA was formed. The ratio of 5 : 8 was chosen for the electrospinning because 5 wt% of CN_x nanoparticles (the active species for the ORR) saturates the CN_x nanoparticle/water dispersion and a minimum content of 8 wt% PVA is necessary to prepare stable, uniform nanofibers. Electrospinning (Super ES-2, E-Spin Nanotech Pvt. Ltd, India) of the homogenous solutions was carried out using an infusion syringe pump with a flat tip needle (internal diameter = 0.9 mm) as the spinning head and a 50 kV dual polarity high-voltage power supply (Gamma HV). The syringe pump was used to control the flow rate, whereas the high voltage supply was used to charge the needle and the collector. A 15 cm × 15 cm steel plate was used as the collector, which was connected to the high voltage supply. A circular glassy carbon electrode was placed on the steel plate for collection of the nanofibers for voltammetry. A small amount of Nafion® dispersion was added to improve proton conductivity. The syringe pump was operated at a flow rate of 0.5 mL h⁻¹. Electrospinning was carried out at a voltage of 10 kV and 30 kV for the PVA and CN_x/PVA solutions, respectively. The distance between the needle and the collector was kept at 10 cm. The temperature and relative humidity were maintained at 25 °C and 50% RH for all the experiments.

2.4. Physical characterization

The surface morphology, particle size and diameter of the nanofibers were analysed using SEM (EVO 50, Zeiss UK electron microscope) and TEM (FEI Technai G2 20 electron microscope). A FTIR study (Bruker Vertex 70v) was carried out to understand the bond structure in the CN_x nanoparticles. X-ray photoelectron spectroscopy (XPS) (Axis His-165 Ultra, Kratos Analytical; Shimadzu, Kyoto, Japan) was carried out to study the characteristic absorption peaks of the samples to confirm the presence of CN_x in the nanofibers. XPS was carried out under a basic pressure of 1.7 × 10⁻⁸ Torr, and an anode of mono-Al with a pass energy of 40 (survey scan) was used as the X-ray source. Raman spectroscopy (LabRam HR Ar-ion laser 514 nm, Jobin-Yvon, Longjumeau, France) was carried out for the native CN_x nanoparticles, CN_x/PVA nanofibers and PVA nanofibers to further confirm the presence of CN_x nanoparticles in the CN_x/PVA nanofibers.

2.5. Working electrode preparation

Working electrodes were prepared using the CN_x/PVA nanofibers, PVA nanofibers, native CN_x nanoparticles and Pt/C separately as the catalyst material. 20 mg of the catalyst was ultrasonicated in a solution of 4 mL ethanol and 40 µL Nafion® dispersion for 1 h. 4 µL of the prepared catalyst ink (for the CN_x/PVA nanofibers, native CN_x nanoparticles and PVA nanofibers) was then carefully deposited on a circular glassy carbon electrode of 1.6 mm diameter. For Pt/C, 2.4 µL of the catalyst ink was used. In order to understand the ORR catalytic activity of the as prepared CN_x/PVA nanofibers, the glassy carbon electrode was placed directly on the steel collector during electrospinning. Such CN_x/PVA nanofibers are denoted as CN_x/PVA nanofibers (as prepared) in the study. The CN_x/PVA nanofiber coated glassy carbon electrode was also used for in situ FTIR studies, coupled with cyclic voltammetry. For the RRDE, CN_x/PVA nanofibers and Pt/C were coated on the disk electrode by preparing a catalyst ink.

2.6. Electrochemical characterization for the oxygen reduction reaction

A three-electrode cell assembly connected to a potentiostat-galvanostat (PGSTAT30, AUTOLAB) was used for the cyclic voltammetry (CV) and linear sweep voltammetry (LSV) studies. An Ag/AgCl (in saturated KCl) electrode and Pt wire were used as the reference and counter electrode respectively. All potential values mentioned are relative to a reversible hydrogen electrode (RHE). The electro-catalyst coated glassy carbon electrode was used as the working electrode with H₂SO₄ (0.5 M) as the electrolyte. CV was carried out on a bare glassy carbon electrode for reference. A voltammetry study was carried out in oxygen purged 0.5 M H₂SO₄ solution for the CN_x/PVA nanofibers to determine the onset potential and peak current densities. The oxygen reduction reaction (ORR) was studied using the prepared native CN_x nanoparticle, CN_x/PVA nanofiber, CN_x/PVA nanofiber (as prepared), PVA nanofiber and Pt/C (40 : 60% by weight) catalysts using a LSV technique, ranging from 1 V to 0.4 V in 0.5 M H₂SO₄ solution and at a scan rate of 20 mV s⁻¹. For the native CN_x nanoparticles, PVA nanofibers and CN_x/PVA nanofibers, a catalyst loading of 1 mg cm⁻² was used, and it was estimated that the CN_x loading is 385 µg cm⁻² (5 : 8) in the CN_x/PVA nanofibers. For Pt/C, a catalyst loading of 600 µg cm⁻² was used. The ORR mechanism was analysed using rotating ring disk electrode (RRDE) voltammetry and an in situ FTIR study combined with cyclic voltammetry. A RRDE study was performed using a glassy carbon electrode as the disk electrode and a platinum ring electrode. The disk electrode potential was swept from 0.8 V to 0.2 V. The ring potential was maintained at 1.2 V throughout the electrochemical measurements in order to oxidize hydrogen peroxide produced via a 2-electron pathway. A RRDE experiment of the CN_x/PVA nanofibers was used for the ORR at rotation speeds of 0, 600, 900, 1600 and 2500 rpm. A RRDE experiment of Pt/C was used for the ORR at 1600 rpm for comparison. An in situ FTIR study combined with LSV was conducted on the CN_x/PVA nanofibers and the corresponding spectra were plotted at different potentials. In situ FTIR studies with voltammetry were conducted over 6000 repetitions to analyse the catalytic activity of the CN_x/PVA nanofibers over a period of time. FTIR spectra, corresponding to the 5^th, 1000^th, 2000^th, 3000^th and 5000^th repetitions, were recorded at 0.7 V. For both the RRDE study and in situ FTIR with LSV studies, Ag/AgCl (in saturated KCl), Pt wire and an oxygen purged 0.5 M H₂SO₄ solution were used as the reference electrode, counter electrode and electrolyte, respectively. TEM was carried out on the CN_x/PVA nanofibers after dispersing them in IPA, before voltammetry and after 2000 and 5000 repetitions of the LSV for the ORR to further understand the catalyst behaviour during voltammetry.

2.7. Scanning electrochemical microscopy

In order to understand the electron conductivity of the CN_x/PVA nanofibers, scanning electrochemical microscopy (SECM) was carried out using a Sensolytics SECM setup attached to a PGSTAT30 (Autolab). The SECM was conducted under feedback mode. For conducting the SECM, the CN_x/PVA nanofibers were coated on carbon paper. SECM was also carried out for the base material (carbon paper) for comparison. A Pt microelectrode (25 µm) was used as the working electrode, Pt foil as the counter electrode, Ag/AgCl (sat KCl) as the reference electrode and oxygen saturated 0.5 M H₂SO₄ solution as the electrolyte. A voltage of 0.5 V was applied at the Pt microelectrode and the current generated was observed.

3. Results and discussion

3.1. SEM and TEM

A TEM micrograph of the native CN_x nanoparticles is shown in Fig. 1(a). The figure indicates that the nanoparticles are formed in irregular shapes with a size of around 20 nm. The CN_x nanoparticles agglomerate quickly as shown in Fig. 1(a). Fig. 1(b) shows a SEM micrograph of the CN_x/PVA nanofibers at a lower magnification which indicates the formation of dense, multi-layered nanofibers. TEM micrographs of the CN_x/PVA nanofibers shown in Fig. 1(c)–(e) confirm the presence of CN_x nanoparticles in the nanofibers. In most of the cases, it was seen that the CN_x nanoparticles are bulging out from the nanofibre (Fig. 1(c)) possibly allowing contact with a glassy carbon electrode or carbon paper. Sometimes the CN_x nanoparticles are completely embedded inside the PVA nanofibers (Fig. 1(d)) preventing contact with an electrode. Overall the diameters of the CN_x/PVA nanofibers lie in the range of 20 nm to 320 nm. The large range in size is attributed to non-uniform dispersion of the conducting CN_x nanoparticles in the PVA solution. The regions with a greater concentration of CN_x nanoparticles have smaller sized nanofibers due to high electrostatic forces. The high electrostatic force dominates the surface tension force leading to breaking of the nanofibers at some points. It was seen from Fig. 1(e) that the CN_x/PVA nanofibers are parallely aligned entrapping the CN_x nanoparticles, which prevents agglomeration and at the same time allows contact with an electrode. When spread onto the glassy carbon electrode using a Nafion® dispersion the CN_x/PVA nanofibre allows both electronic and proton conductivity. If the CN_x nanopowder was completely embedded in the PVA nanofibre, maintaining connectivity for proton and electron transfer would not be possible.

Fig. 1 (a) TEM image of the native CN_x nanoparticles, (b) SEM image of the CN_x/PVA nanofibers, (c) TEM image of the CN_x/PVA nanofibers showing nanofibers of ∼150 nm in diameter, (d) TEM image of the CN_x/PVA nanofibers with CN_x nanoparticles embedded in the CN_x/PVA nanofibers, and (e) TEM image of the CN_x/PVA nanofibers with CN_x nanoparticles entrapped within two aligned nanofibers.

3.2. FTIR

The FTIR spectrum of the native CN_x nanoparticles is shown in Fig. 2. The three major bands centred at around 1100, 1626 and 3418 cm⁻¹ confirm the presence of the aromatic C–N single bond stretching mode, deformation modes and the stretching mode of N–H groups respectively.³⁰ Moreover, the absence of C≡N is confirmed as no peak at 2200 cm⁻¹ was found. The bands around 2922 and 2854 cm⁻¹ correspond to terminal CH₃ groups. The band around 1384 cm⁻¹ and the shoulder around 1224 cm⁻¹ correspond to sp³ C–C bonds.³⁰ The data further confirm the presence of nitrogen in pyridinic and pyrrolic forms. The FTIR spectra of CN_x/PVA show some notable differences from the PVA spectra. There is a major band for the CN_x/PVA nanofibers at around 1100 cm⁻¹, as in the spectrum of the native CN_x nanoparticles, which is expected due to the presence of the C–N single bond stretching mode of CN_x in the nanofibers. Shifting of the peak from 3430 cm⁻¹ for PVA (due to the OH stretch mode of alcohols) and 3420 cm⁻¹ for the native CN_x nanoparticles to 3310 cm⁻¹ for the CN_x/PVA nanofibers shows that there are some interactions between the CN_x nanoparticles and PVA, leading to certain structural modifications, which may be responsible for the enhanced ORR activity and conductivity. The band at 1650 cm⁻¹ for the CN_x/PVA nanofibers confirms the presence of the N–H bending mode, as observed using XPS. Since there is no significant change in the percentage of the pyridine form of N in the CN_x/PVA nanofibers compared to the native CN_x nanoparticles (as confirmed using XPS), it is inferred that a percentage of the pyrrolic form of N is converted to an amino N form. This is possible as the electrospinning of the solution with the nanoparticles required a high voltage (30 kV) and some percentage of the pyrrolic N of CN_x could absorb this energy and get converted to amino N form. Since the amino N form contributes to the ORR and pyrrolic N does not,³¹ the conversion to an amino N form contributes to the increase in ORR activity with the CN_x/PVA nanofibers. The band at 1730 cm⁻¹ corresponds to the C=O of aldehydes, due to modification of the PVA structure. It can be inferred that the alcohol groups in the PVA get converted to an aldehyde. The resulting H species, along with the high energy due to the electrospinning process, aid in the conversion of the pyrrole groups to the more active amino groups.

Fig. 2 FTIR spectra of the native CN_x nanoparticles, PVA nanofibers and CN_x/PVA nanofibers.

3.3. XPS

X-ray photoelectron spectroscopy (XPS) results for the native CN_x nanoparticles and the CN_x/PVA nanofibers are shown in Fig. 3. The C 1s peak of the nanoparticles and nanofibers, corresponding to a binding energy of 284.6 eV and 285.7 eV, matches with the binding energy of pure graphitic sites and of sp² C atoms bonded to N inside an aromatic structure respectively. The small peak at 288.8 eV corresponds to sp² hybridised carbon atoms in an aromatic ring attached to a NH₂ group.³⁰ N 1s peak analysis for the CN_x nanoparticles shows the presence of peaks at 400.7 eV and 398.9 eV, which correspond to pyrrolic N and pyridinic N, respectively.³¹ Nitrogen was observed to be present in a very small quantity in the nanofibers (2%) as compared to the nanoparticles (29%). This lower than expected nitrogen content is because of the huge carbon content present in both CN_x and PVA.¹² A new peak for the nanofibers at 399.7 eV was noticed, which corresponds to amino N. A small amount of pyrrolic N, being least stable,³¹ gets converted to amino N on application of the high potential during electrospinning, as explained in the FTIR section. The percentage of pyridinic N, mainly responsible for the ORR, in the CN_x/PVA nanofibers (68%) is similar to that in the native CN_x nanoparticles (65%). It is worth noticing that in spite of the fact that the nitrogen content is reduced considerably in the nanofibers, the activity of the nanofibers is increased for the ORR, which is attributed to the increase in surface area per unit volume for the CN_x nanoparticles in the nanofibers as compared to their agglomerate form. The conversion of inactive pyrrolic N to active amino N also contributes to the increase in the ORR activity.

Fig. 3 XPS spectra of the CN_x nanoparticles, (a) C 1s peaks and (b) N 1s peaks; XPS spectra of the CN_x/PVA nanofibers, (c) C 1s and (d) N 1s peaks.

3.4. Raman spectroscopy

Fig. 4 shows the Raman spectra for the native CN_x nanoparticles, CN_x/PVA nanofibers and PVA nanofibers kept under the same conditions during measurement. The D and G bands around 1350 and 1590 cm⁻¹ are present in both the native CN_x nanoparticles and CN_x/PVA nanofibers. The G band corresponds to sp³ hybridised (graphitic) carbon and the D band corresponds to structural defects at the edges of sp² hybridised carbon (attributed to pyridinic N in the structure), which further confirms the results of the XPS analysis. The intensity ratio of the D and G bands (I_D/I_G) was calculated to determine the extent of defects induced as a result of the presence of pyridinic N in both the native CN_x nanoparticles and CN_x/PVA nanofibers.³¹ For both these cases, the I_D/I_G ratio was calculated to be 0.3, which shows that the pyridinic character of the CN_x nanoparticles survived under the high voltage and electrostatic force of the electrospinning process, as confirmed using XPS as well. The decrease in the intensities of the peaks for the nanofibers is attributed to the lower than expected content of CN_x nanoparticles due to a higher PVA content in the nanofibers. It may be noted that no significant peak was observed in the Raman spectrum of the PVA nanofibers in the range of 1300 to 1700 cm⁻¹.

Fig. 4 Raman spectra at a 514.5 nm wavelength of the native CN_x nanoparticles, CN_x/PVA nanofibers and PVA nanofibers.

3.5. Voltammetry

Fig. 5(a) and (b) show linear sweep voltammetry plots for the native CN_x nanoparticles, CN_x/PVA nanofibers, CN_x/PVA nanofibers (as prepared), PVA nanofibers, Pt/C and bare glassy carbon electrode. The potential window for the CV is shown from 0.2 V to 1.2 V because no characteristic peak was observed by extending the potential range beyond 1.2 V. The plots reveal significant improvement in the electrocatalytic activity of the CN_x nanoparticles dispersed in PVA nanofibers as compared to their native form. The onset potentials for the catalysts, determined from the plots, where the slope changes prior to the reduction peak, and the peak current densities are indicated in Table 1. Since PVA does not exhibit any ORR activity, as shown in the expanded portion from I of Fig. 5(a) in Fig. 5(b), the catalytic activity of the CN_x/PVA nanofibers is attributed to active CN_x nanoparticles. The observed onset potential of 0.88 V and peak current density of 5.48 mA cm⁻² for the CN_x/PVA nanofibers are enhanced as compared to the onset potential (0.75 V) and peak current density (0.12 mA cm⁻²) of the CN_x nanoparticles in their native form. The enhanced onset potential substantiates that the CN_x/PVA nanofibers offer less resistance to the ORR. The CN_x/PVA nanofibers exhibit a similar peak current density for the ORR when employed as prepared and as a catalyst ink. However, the CN_x/PVA nanofibers (catalyst ink) gives a broad voltage–current density peak, whereas the CN_x/PVA nanofibers (as prepared) gives a relatively sharper peak. The catalytic activity of the CN_x/PVA nanofibers is comparable to the Pt/C catalyst (onset potential: 0.97 V and peak current density: 6.41 mA cm⁻²). The improvement in the mass activity (16 A g⁻¹) of the CN_x nanoparticles in the CN_x/PVA nanofibers as compared to 0.12 A g⁻¹ for the native form is attributed to the increased number of active sites in the entrapped nanofibers due to dispersion of the CN_x nanoparticles in the PVA nanofibers, as compared to the agglomerated native CN_x nanoparticles, conversion of some of the pyrrolic N to amino N during electrospinning, and structural modification of the CN_x/PVA interface. The ORR activity of the CN_x/PVA nanofibers is improved as compared to previously reported bulk CN_x supported on carbon black (onset potential: 0.76 V and peak current density of 2.21 mA cm⁻²)¹⁴ and bulk CN_x pyrolysed at 1000 °C (onset potential 0.85 V).¹²

Fig. 5 (a) Linear sweep voltammograms of the CN_x/PVA nanofibers (prepared through a catalyst ink and as prepared), native CN_x nanoparticles, Pt/C, PVA nanofibers and bare glassy carbon electrode in 0.5 M H₂SO₄ solution at a scan rate of 20 mV s⁻¹ for the ORR. (b) Expanded portion (I) of the voltammograms of the native CN_x nanoparticles, PVA nanofibers and bare glassy carbon electrode.

Table 1 Comparison of the onset potential and peak current density of the CN_x/PVA nanofibers, native CN_x nanoparticles and Pt/C

Catalyst	Onset potential (V)	Peak current density (mA cm^−2)
CNx/PVA nanofibers	0.88	−5.48
Native CNx nanoparticles	0.75	−0.12
Pt/C	0.97	−6.41

---

The disk current and ring current plots for the RRDE experiment using the CN_x/PVA nanofibers for the ORR at 0, 600, 900, 1600 and 2500 rpm are shown in Fig. 6(a) and (b). The disk current and ring current plots for the RRDE experiment using Pt/C for the ORR at 1600 rpm are also shown in Fig. 6(a) and (b). It can be seen that the ORR activity of the CN_x/PVA nanofibers is comparable to that of Pt/C. The number of electrons (n) transferred for the CN_x/PVA nanofibers is calculated from the formula for the ORR mechanism:¹²

where, n is the number of electrons transferred, I_r is the ring current (A), I_d is the disk current (A) and N is the collection efficiency of the ring (0.424). The values of n in the potential range of 0.2 to 0.6 V were plotted for the CN_x nanoparticles and CN_x/PVA nanofibers (Fig. 6(c)). Average values of n equal to 3.5 and 3.7 were obtained for the CN_x nanoparticles and CN_x/PVA nanofibers, respectively. This indicates that the ORR on the CN_x nanoparticles predominantly takes place via a highly efficient 4-electron pathway. The ‘n’ value was further confirmed using the Koutecky–Levich equation:³¹

B = 0.2nFAC_O2D_O2^2/3ν^−1/6 (3)

where, I_k is the kinetic current, ω is the rotation per minute, n is the number of electrons transferred, F is Faraday's constant (96 500), A is the electrode area (0.1256 cm²), C_O2 is the concentration of O₂ in 0.5 M sulphuric acid (1.1 × 10⁻⁶ mol cm⁻³), D_O2 is the diffusion coefficient (1.4 × 10⁻⁵ cm² s⁻¹), and ν is the kinematic viscosity (0.01 cm s⁻¹). The Koutecky–Levich (I_d⁻¹ vs. ω^−1/2) plots at different potentials shown in Fig. 6(d) exhibit close linearity. The slopes of the lines closely match with that of the reference plot for n = 4 and an n value of 3.7 is obtained, which further confirms the dominance of the 4-electron pathway mechanism. The hydrogen peroxide yield was calculated from the RRDE data for the CN_x/PVA nanofibers using the formula³²
and the hydrogen peroxide yield was calculated to be ∼22%. The same percentage of hydrogen peroxide yield was previously reported¹² for bulk CN_x pyrolised at 1000 °C. The H₂O₂ formation can be attributed to the ORR occurring on amino N (21% of the present N in the CN_x/PVA nanofibers) sites present in the CN_x/PVA nanofibers as it has already been established in the literature that ORRs on amino N sites take place via a combination of 2-electron and 4-electron reactions.³³ The ORR on pyridinic N (68% of the present N in the CN_x/PVA nanofibers) sites occurs via a 4-electron pathway towards water formation with little or no hydrogen peroxide formation.³¹

Fig. 6 Linear sweep voltammograms of the CN_x/PVA nanofibers for the ORR showing the variation of current density with rotation speed and that of Pt/C at 1600 rpm on a rotating ring disc glassy carbon electrode in 0.5 M H₂SO₄ solution at scan rate of 20 mV s⁻¹: (a) disk current; (b) ring current; (c) number of electrons exchanged versus voltage for the CN_x/PVA nanofibers (eqn (1)) and (d) Koutecky–Levich plots for the ORR for the CN_x/PVA nanofibers (eqn (2)).

Fig. 7(a) shows the in situ FTIR results along with cyclic voltammetry (Fig. 7(b)). The cyclic voltammetry results show a quinone–hydroquinone type redox couple which does not take part in the ORR, confirmed by its presence under nitrogen atmosphere.¹⁴ CV revealed that the ORR under an oxygen atmosphere initiates at 0.88 V and continues until the reduction peak of the quinone–hydroquinone type redox couple is reached, depicted by an increase in current density at peak potential under an oxygen atmosphere as compared to a nitrogen atmosphere. The same is confirmed by the in situ FTIR results as shown by the three bands (Fig. 7(c)–(e)). The band X (3410 cm⁻¹) in Fig. 7(c) is attributed to a water ν(OH) band formed as a result of the end product of the ORR.³³ The bands Y (2590 cm⁻¹) and Z (1980 cm⁻¹) in Fig. 7(d) and (e) are attributed to hydroquinone and quinone type structures associated with the reduction and oxidation peaks of the redox couple respectively. The band X is observed to increase with a decrease in potential to 0.5 V, confirming the occurrence of the ORR throughout the range. A slight decrease in band X observed between 0.9 V and 0.8 V implies the initiation of the ORR. If the water formation had not started, the dip in the peak would have been significant, as can be observed for bands Y and Z in the potential region 0.9 V to 0.8 V, where the redox couple is not initiated yet. The band Y peak is slightly shifted in the range of 0.8 V to 0.7 V, attributed to the significant changes in bond behaviour due to simultaneous occurrence of the ORR and initiation of the reduction of quinone type structures at the electrode. The band Y peak increases on reducing the potential further to 0.5 V, corresponding to a reduction peak in CV. Similarly, the band Z peak increases with the decrease in potential in the range 0.8 V to 0.5 V, corresponding to an oxidation peak in CV.

Fig. 7 (a) In situ FTIR spectra at the potentials 0.9 V, 0.8 V, 0.7 V, 0.6 V and 0.5 V with (b) cyclic voltammetry of the CN_x/PVA nanofibers in nitrogen and oxygen purged 0.5 M H₂SO₄ for the ORR. Expanded portions of the FTIR bands: (c) X at 3410 cm⁻¹, (d) Y at 2590 cm⁻¹, and (e) Z at 1980 cm⁻¹.

In order to understand the ORR activity over a period of time, the CN_x/PVA nanofibers were subjected to 6000 repetitions of the reduction under in situ FTIR and voltammetry studies. The ORR linear sweep voltammograms at different LSV repetitions are shown in Fig. 8(a). Corresponding FTIR spectra at 0.7 V for different LSV repetitions are shown in Fig. 8(b). It was noticed that the ORR activity of the CN_x/PVA nanofibers improves with the increasing number of repetitions. However the ORR activity becomes stable after 5000 repetitions. There is no appreciable change in the voltammogram noticed for the 6000^th repetition. For the FTIR, major changes in four peaks were studied, namely, 3410 cm⁻¹ (K), 2590 cm⁻¹ (L), 1980 cm⁻¹ (M) and 2350 cm⁻¹ (N).

Fig. 8 (a) Polarization curves of the CN_x/PVA nanofibers for the ORR at different LSV repetitions and (b) FTIR spectra obtained at 700 mV for different LSV repetitions.

Initially, the ORR and hydroquinone–quinone type redox reaction are initiated as observed from the presence of bands K, L and M during the initial 5 repetitions. During 5 to 1000 LSV repetitions, a sharp increase in band L and band M suggests the formation of hydroquinone and quinone type structures subjugating the water band (K) peak at the electrode. During 1000 to 2000 LSV repetitions, no significant change in bands L and M was observed. However improvement in the reduction peaks suggests that hydroquinone–quinone type redox reactions are dominant during this period. The reversible nature of the reaction contributes to the no change in the peaks of bands L and M. For the LSV repetitions 2000 to 3000, the ORR reaches a steady state with a slight improvement in the L and M bands. It may be noticed that water formation due to the ORR begins to dominate resulting in a sharp increase in the band K peak. After 3000 LSV repetitions, the ORR dominates at the electrode resulting in a decrease in the band L and M peaks and an increase in the band K peak. The continuous decrease of band N is assigned to evolution of CO₂ in the electrolyte surrounding the electrode, which may be due to degeneration of the glassy carbon electrode or PVA. This may require further investigation. The overall dominance of the ORR after 5000 LSV repetitions suggests the CN_x/PVA is catalytically stable for the ORR.

In order to understand the possible reason for the ORR catalytic activity of the CN_x/PVA nanofibers, TEM was carried out for the CN_x/PVA nanofibers after dispersing them in isopropyl alcohol (IPA) before the ORR voltammetry (Fig. 9(a)), after 2000 repetitions of LSV (Fig. 9(b)) and after 5000 repetitions of LSV (Fig. 9(c)). It was seen from Fig. 9(a) that the CN_x nanoparticles are present on the surface and in the bulk of the nanofiber. The CN_x/PVA nanofiber becomes lighter colored in nature, as seen in Fig. 9(b), after 2000 repetitions of LSV. The PVA may be dissolving out of the fiber exposing the CN_x nanoparticles on the surface of the fiber. After 5000 repetitions of LSV, it can be noticed that the PVA nanofibers have disintegrated (Fig. 9(c)) but the CN_x nanoparticles still remain attached to the disintegrated form of PVA with the formation of agglomerates. Retention of the CN_x nanoparticles on the glassy carbon electrode or carbon paper is attributed to the presence of the Nafion® dispersion and the binding properties of PVA, although not present in nanofiber form.

Fig. 9 TEM images of the CN_x/PVA nanofibers dispersed in isopropyl alcohol (a) before the ORR voltammetry, (b) after 2000 LSV repetitions for the ORR and (c) after 5000 LSV repetitions for the ORR.

In order to understand the electron conductivity of the CN_x/PVA nanofibers, SECM was carried out for the CN_x/PVA nanofibers coated on carbon paper and for uncoated carbon paper. An increased current generated for the CN_x/PVA nanofibers, as compared to the carbon paper, at the microelectrode during the approach curve (Fig. 10(a)) confirms the conductive nature of the CN_x/PVA nanofibers. A SECM plot of the current variation with electrode area surface (XY), keeping the height (Z) constant, was plotted (Fig. 10(b)). Z is the electrochemical active region of the CN_x/PVA nanofibers coated and uncoated carbon paper very close to the surface, decided using the approach curve experiment. It was seen that an increase in current for the CN_x/PVA nanofibers is observed as compared to the uncoated carbon paper, which further confirms the conductive nature of the CN_x/PVA nanofibers.

Fig. 10 (a) Current–distance (approach) curve obtained when the Pt microelectrode (25 µm) was approaching the carbon paper (CN_x/PVA nanofiber coated and uncoated) surface, (b) SECM plot showing the variation of current with electrode area surface (XY) keeping the height (Z) constant.

4. Conclusion

The electrocatalytic activity of CN_x nanoparticles is improved many folds by dispersing them in PVA nanofibers, as compared to their agglomerated native form. The PVA nanofibre not only helps in dispersing and entrapping the CN_x nanoparticles, but is also responsible for the improved ORR activity as shown by XPS and Raman spectra confirming the presence of pyridine N and the conversion of a small percentage of pyrrole N to active amino N during application of a high voltage in electrospinning. An improved onset potential and peak current density of the CN_x/PVA nanofibers for the ORR are observed using voltammetry, which are comparable to a platinum based catalyst. RRDE studies confirmed that the CN_x/PVA nanofibre follows a 4-electron ORR pathway. The active sites of the CN_x/PVA nanofibers were found to be stable on exposing them to 5000 voltammetric repetitions observed using in situ FTIR.

5. Acknowledgement

The authors would like to acknowledge the Department of Information Technology, Government of India and RCUK-DST for funding the project.

References

1. Recent Trends in Fuel Cell Science and Technology, ed. S. Basu, Springer, New York, 2007.
2. N. M. Markovic, H. a. Gasteiger, B. N. Grgur and P. N. Ross, J. Electroanal. Chem., 1999, 467, 157–163.
3. H. Naohara, S. Ye and K. Uosaki, Electrochim. Acta, 2000, 45, 3305–3309.
4. L. Zhang and Z. Xia, J. Phys. Chem. C, 2011, 115, 11170–11176.
5. Y. Li, W. Zhou, H. Wang, L. Xie, Y. Liang, F. Wei, J.-C. Idrobo, S. J. Pennycook and H. Dai, Nat. Nanotechnol., 2012, 7, 394–400.
6. J. Liu, D. Takeshi, D. Orejon, K. Sasaki and S. M. Lyth, J. Electrochem. Soc., 2014, 161, F544–F550.
7. R. Wang, T. Zhou, H. Li, H. Wang, H. Feng, J. Goh and S. Ji, J. Power Sources, 2014, 261, 238–244.
8. D. Singh, J. Tian, K. Mamtani, J. King, J. T. Miller and U. S. Ozkan, J. Catal., 2014, 317, 30–43.
9. V. Nallathambi, N. Leonard, R. Kothandaraman and S. C. Barton, Electrochem. Solid-State Lett., 2011, 14, B55–B58.
10. H. Li, H. Liu, Z. Jong, W. Qu, D. Geng, X. Sun and H. Wang, Int. J. Hydrogen Energy, 2011, 36, 2258–2265.
11. S. Chen, J. Bi, Y. Zhao, L. Yang, C. Zhang, Y. Ma, Q. Wu, X. Wang and Z. Hu, Adv. Mater., 2012, 24, 5593–5597.
12. S. M. Lyth, Y. Nabae, N. M. Islam, S. Kuroki, M. Kakimoto and S. Miyata, J. Electrochem. Soc., 2011, 158, B194–B201.
13. Y. Zheng, J. Liu, J. Liang, M. Jaroniec and S. Z. Qiao, Energy Environ. Sci., 2012, 5, 6717–6731.
14. S. M. Lyth, Y. Nabae, S. Moriya, S. Kuroki, M. A. Kakimoto, J. I. Ozaki and S. Miyata, J. Phys. Chem. C, 2009, 113, 20148–20151.
15. Y. Miyamoto, M. L. Cohen and S. G. Louie, Solid State Commun., 1997, 102, 605–608.
16. M. Terrones, H. Terrones, N. Grobert, W. K. Hsu, Y. Q. Zhu, J. P. Hare, H. W. Kroto, D. R. M. Walton, P. Kohler-Redlich, M. Ruhle, J. P. Zhang and a. K. Cheetham, Appl. Phys. Lett., 1999, 3932, 25–28.
17. P. K. Panda, Ceram. Int., 2013, 39, 4523–4527.
18. S. L. Sung, S. H. Tsai, C. H. Tseng, F. K. Chiang, X. W. Liu and H. C. Shih, Appl. Phys. Lett., 1999, 74, 197–199.
19. M. Terrones, N. Grobert, J. Olivares, J. P. Zhang, H. Terrones, K. Kardatos, W. K. Hsu, J. P. Hare, P. D. Townsend, K. Prassides, A. K. Cheetham, H. W. Kroto and D. R. M. Walton, Nature, 1997, 388, 52–55.
20. M. Terrones, N. Grobert, J. P. Zhang, H. Terrones, J. Olivares, W. K. Hsu, J. P. Hare, a. K. Cheetham, H. W. Kroto and D. R. M. Walton, Chem. Phys. Lett., 1998, 285, 299–305.
21. M. Terrones, P. Redlich, N. Grobert, S. Trasobares, W. Hsu, H. Terrones, Y. Zhu, J. P. Hare, C. L. Reeves, A. K. Cheetham, M. Rühle, H. W. Kroto and D. R. M. Walton, Adv. Mater., 1999, 11, 655–658.
22. C. R. Martin, Chem. Mater., 1996, 8, 1739–1746.
23. T. Ondarçuhu and C. Joachim, Europhys. Lett., 1998, 42, 215–220.
24. P. X. Ma and R. Zhang, J. Biomed. Mater. Res., 1999, 46, 60–72.
25. G. Liu, J. Ding, L. Qiao, A. Guo, B. P. Dymov, J. T. Gleeson, T. Hashimoto and K. Saijo, Chem.–Eur. J., 1999, 5, 2740–2749.
26. Z. M. Huang, Y. Z. Zhang, M. Kotaki and S. Ramakrishna, Compos. Sci. Technol., 2003, 63, 2223–2253.
27. D. H. Reneker and A. L. Yarin, Polymer, 2008, 49, 2387–2425.
28. M. S. Islam and M. R. Karim, Colloids Surf., A, 2010, 366, 135–140.
29. Y. Qiu and L. Gao, Chem. Commun., 2003, 2378–2379.
30. A. Vinu, K. Ariga, T. Mori, T. Nakanishi, S. Hishita, D. Golberg and Y. Bando, Adv. Mater., 2005, 17, 1648–1652.
31. N. Daems, X. Sheng, I. F. J. Vankelecom and P. P. Pescarmona, J. Mater. Chem. A, 2014, 2, 4085–4110.
32. X. Yang, L. Gan, C. Zhu, B. Lou, L. Han, J. Wang and E. Wang, Chem. Commun., 2014, 50, 234–236.
33. K. Kunimatsu, T. Yoda, D. A. Tryk, H. Uchida and M. Watanabe, Phys. Chem. Chem. Phys., 2010, 12, 621–629.

---