Cobalt metalloid and polybenzoxazine derived composites for bifunctional oxygen electrocatalysis

Accepted Manuscript Cobalt metalloid and polybenzoxazine derived composites for bifunctional oxygen electrocatalysis Stefan Barwe, Corina Andronescu, Ruben Engels, Felipe Conzuelo, Sabine Seisel, Patrick Wilde, Yen-Ting Chen, Justus Masa, Wolfgang Schuhmann PII:

S0013-4686(18)32745-2

DOI:

https://doi.org/10.1016/j.electacta.2018.12.047

Reference:

EA 33262

To appear in:

Electrochimica Acta

Received Date: 30 September 2018 Revised Date:

29 November 2018

Accepted Date: 8 December 2018

Please cite this article as: S. Barwe, C. Andronescu, R. Engels, F. Conzuelo, S. Seisel, P. Wilde, Y.-T. Chen, J. Masa, W. Schuhmann, Cobalt metalloid and polybenzoxazine derived composites for bifunctional oxygen electrocatalysis, Electrochimica Acta (2019), doi: https://doi.org/10.1016/ j.electacta.2018.12.047. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT

Cobalt metalloid and polybenzoxazine derived composites for bifunctional oxygen electrocatalysis

RI PT

Stefan Barwe, Corina Andronescu*, Ruben Engels, Felipe Conzuelo, Sabine Seisel, Patrick Wilde, Yen-Ting Chen, Justus Masa, Wolfgang Schuhmann*

SC

Analytical Chemistry – Center for Electrochemical Sciences (CES),

Faculty of Chemistry and Biochemistry, Ruhr-Universität Bochum, Universitätsstr. 150;

M AN U

D-44780 Bochum

Abstract

TE D

*Correspondence to: [email protected], [email protected]

EP

The development of bifunctional oxygen electrodes is a key factor for the envisaged application of rechargeable metal-air batteries. In this work, we present a simple procedure

AC C

based on pyrolysis of polybenzoxazine / metal metalloid nanoparticles composites into efficient bifunctional oxygen reduction and oxygen evolution electrocatalysts. This procedure generates nitrogen-doped carbon with embedded metal metalloid nanoparticles exhibiting high activity towards both, oxygen reduction and oxygen evolution, in 0.1 M KOH with a roundtrip voltage of as low as 0.81 V. Koutecký-Levich analysis coupled with scanning electrochemical microscopy reveals that oxygen is preferentially reduced in a 4e− transfer pathway to hydroxide rather than to hydrogen peroxide. Furthermore, the polybenzoxazine derived carbon matrix allows for stable catalyst fixation on the electrode surface, resulting in

ACCEPTED MANUSCRIPT unattenuated activity during continuous alternate polarization between oxygen evolution at 10 mA cm-2 and oxygen reduction at -1.0 mA cm-2.

conversion, metall metalloid compound

1. Introduction

RI PT

Keywords: electrocatalysis, oxygen evolution reaction, oxygen reduction reaction, energy

SC

In the development of rechargeable metal-air batteries, the performance of the air-

M AN U

electrode is a crucial factor since it has to perform both the oxygen reduction reaction (ORR) during discharging and the oxygen evolution reaction (OER) during the charging process.[1,2] Bifunctional oxygen electrocatalysts are needed for this purpose.[3,4] Tremendous research effort is currently devoted towards the development of highly

TE D

active and stable bifunctional catalysts.[5–7] However, the development is hampered by the inherently different nature of the two reactions. Generally, materials exhibiting high activity towards the ORR show only minor activity for the OER and vice versa.[8–

EP

10] Therefore, bifunctional catalysts are often mixtures of two catalysts in which the

AC C

individual catalysts are active for either OER or ORR.[11] Such composite materials are typically platinum group metals or their oxides, however, their scarcity and high costs preclude them from possible large scale application.[11] Thus, cheap and abundant alternatives are much needed. Among them, heteroatom-doped transition metal compounds are known to be bifunctional water splitting catalysts and their activity towards ORR has been improved by supporting them on functionalized carbon structures.[12–15] The doping of metals, metal oxides or metal hydroxides with heteroatoms like phosphorous or boron is known to alter the electronic

ACCEPTED MANUSCRIPT structure of the host material, thus improving its electrocatalytic activity.[16–20] Furthermore, embedding transition metal oxide/hydroxide nanoparticles in nitrogendoped carbon improves electric conductivity, thus facilitating enhanced charge transfer and ultimately increased catalytic activity.[21,22] Additionally, the carbon

RI PT

shell can prevent corrosion and agglomeration of the metallic nanoparticles.[22,23] Recently Wang et al. described Co2P and CoxN nanoparticles embedded in nitrogendoped graphene as promising bifunctional ORR/OER electrocatalysts with decent

SC

cycle life time.[13] Elumeeva et al. activated cobalt boride for the OER and ORR by

as CNT growth catalyst.[24]

M AN U

growing carbon nanotubes (CNT) directly on the catalyst particles, using the particles

Polybenzoxazines are a class of high performance thermosets.[25,26] They arise from polymerisation of benzoxazine (BO) monomers or oligomers and form highly crosslinked

TE D

networks upon thermal treatment. Among others, two unique features make them especially interesting for application as precursors for nitrogen-doped carbon, namely, their near-zero volumetric shrinkage during the polymerisation process, and their ability to

EP

undergo pyrolysis with a high residual char yield.[27–30] Pyrolysis of pBO at temperatures

AC C

higher than 600 °C in the presence of catalytically active moieties or precursors transforms the organic polymer into predominantly graphitic carbon.[31,32] These characteristics allow sequential polymerization and pyrolysis of mixed catalyst/BO films directly on electrode surfaces, yielding a stable nitrogen-doped carbon/catalyst film as we already showed for Prussian blue analogue-based OER catalysts.[33] The high stability of the pBO derived carbon matrix made it possible to investigate the structural changes of a benchmark OER electrocatalyst under drastic conditions of elevated temperature and high electrolyte concentration.[34]

ACCEPTED MANUSCRIPT Herein, we introduce a comparatively simple procedure to incorporate electrocatalysts based on cobalt metalloids (CoxB and CoxP) into a polybenzoxazine (pBO) derived nitrogendoped carbon matrix. The resulting CoxY/NC (Y = B or P, NC = nitrogen-doped carbon) modified electrodes show high bifunctional ORR/OER activity and reduce oxygen nearly

RI PT

exclusively to OH− via the 4e− transfer pathway with negligible hydrogen peroxide formation.

2. Experimental section

SC

2.1 Chemicals

M AN U

Potassium hydroxide (99.5 %) was purchased from Algin (Germany), dichlorobis(triphenylphosphine)cobalt(II) (98 %), ethanol, cobalt(II)chloride hexahydrate, sodium borohydride, bisphenol A, and tetraethylenepentaamine were purchased from Sigma-Aldrich (Germany). Formaldehyde (37 %, stabilised with 10 % methanol) was from Merck (Germany). 1,4-

2.2 CoxB synthesis

TE D

Dioxane, sodium carbonate, and acetic acid were from J.T. Baker (The Netherlands).

CoxB was synthesised as previously described elsewhere.[20,35] Briefly, aqueous CoCl2•6

EP

H2O solution (20 mL, 0.5 M) was deaerated by means of Schlenk vacuum technique, then

AC C

flushed with argon and kept at 0 °C using an ice bath. NaBH4 (1.0 M) in NaOH (0.1 M), separately deaerated and flushed with argon, were slowly added to the CoCl2•6 H2O solution by means of a syringe. A dark precipitate formed immediately. The obtained precipitate was collected by filtration and washed with tri-distilled water and ethanol. The washed precipitate underwent thermal treatment at 500 °C for 2h in an argon atmosphere. 2.3 CoxP synthesis

ACCEPTED MANUSCRIPT Cobalt phosphide nanoparticles were prepared by reductive thermal decomposition of dichlorobis (triphenyl-phosphine) cobalt(II) [(C6H5)3P]2CoCl2 for 2 h at 400 °C in an argon atmosphere containing 25 % H2 in accordance with a procedure published elsewhere [36].

RI PT

2.4 BA-tepa synthesis

The benzoxazine oligomer was synthesised following a procedure published elsewhere. [37,38] Bisphenol A (BA), tetraethylenepentaamine (tepa) and formaldehyde were used in a

SC

1:1:4 ratio. BA (4 g, 0.0175 mol) was dissolved in 1,4-dioxane (10 mL) and the temperature

M AN U

was maintained below 5 °C using an ice bath. Formaldehyde (5.69 g, 0.07 mol, 37 %w/w in water) was added followed by addition of tepa (3.49 g, 0.0175 mol). The resulting reaction mixture was stirred for 1 h at room temperature. BA-tepa was precipitated using aqueous NaHCO3 (250 mL, 0.1 M) and washed with tri-distilled water (250 mL). The BA-tepa

2.5 Electrode preparation

TE D

oligomers were dried under vacuum at room temperature.

EP

Catalyst modified glassy carbon electrodes (GC-RDE, ø = 3.15 mm, Ageom = 0.078 cm2, HTW,

AC C

Germany) were prepared according to the following procedure: a glassy carbon rod was polished on polishing paper (3M, USA) with grades of 3 µm and 1 µm in order to get a mirror-like surface. The polished electrodes were drop-coated with the catalyst ink (7.8 µL) containing the catalytically active powder (5 mg mL−1) in a solution of BA-tepa (5 mg mL−1 in 1 %v/v acetic acid). The catalyst inks were sonicated for at least 20 min prior to drop-coating in order to get a homogeneous suspension. Catalyst film modified electrodes underwent subsequent polymerisation and pyrolysis by heat treatment in an Ar atmosphere (180 °C for 2 h, 200 °C for 2 h, and T °C for 2 h, where T = 500, 600, 650 ot 700 °C; the heating ramp was

ACCEPTED MANUSCRIPT always 10 K min−1). The areal loading with respect to the amount of CoxY is 0.5 mg cm-2 prior to pyrolysis and is assumed unchanged during the thermal treatment. The prepared electrodes were sealed by placing a PTFE tube tightly around the glassy carbon rod in such a manner that only the disk shaped surface (ø = 3.15 mm, Ageom = 0.078 cm2) was exposed to

RI PT

the electrolyte solution.

2.6 Electrochemical measurements

SC

Electrochemical measurements were performed in a three-electrode configuration in

M AN U

0.1 M KOH with the catalyst modified and sealed GC-RDE as working electrode (WE), a platinum mesh in a separated compartment as counter electrode and a Ag/AgCl/3 M KCl as reference electrode. The electrodes were first conditioned by means of potential cycling between 0 and −1 V vs. Ag/AgCl/3 M KCl in aqueous Ar saturated

TE D

KOH solution until reproducible voltammograms were obtained. Afterwards, the Ar background was measured by linear sweep voltammetry (LSV) (5 mV s−1) in the same potential window. The ORR activity was determined by means of linear sweep

EP

voltammetry (5 mV s−1) in O2-saturated 0.1 M KOH in the aforementioned potential

AC C

window at various rotation speeds (100, 400, 900, and 1600 rpm). The voltammograms were corrected for the Ar background. The OER activity was determined at a rotation speed of 1600 rpm in the potential window from 0 to 1 V vs. Ag/AgCl 3 M KCl (5 mV s−1). All potentials were corrected for the ohmic drop according to Ecorr = Emeas – i • Rsol. The solution resistance (Rsol) was determined by means of electrochemical impedance spectroscopy (EIS) in the frequency range from 100 kHz to 10 Hz with an ac perturbation voltage amplitude of 5 mV (RMS). The corrected potential was converted to the

ACCEPTED MANUSCRIPT reversible hydrogen electrode (RHE) scale according to ERHE = Ecorr + E0Ag/AgCl3MKCl + 0.059• pH. The pH value of the 0.1 M KOH electrolyte solution was measured to be 12.89 using a pH electrode for high alkaline solutions (Dr. Kornder Anlagen- und

RI PT

Messtechnik, Germany).

Koutecký-Levich (KL) equation:





= + 





  




(1) (2)

M AN U

 = .   ( ) 

SC

The number of transferred electrons per oxygen molecule was calculated using the

with J being the measured current, Jk the kinetic current, ω the angular velocity (rpm), B the Levich constant, n the number of transferred electrons, F the Faraday constant,

TE D

D0 the diffusion coefficient, C0 the concentration of O2, and  the kinematic viscosity. Scanning electrochemical microscopy (SECM) experiments were performed in a four-electrode configuration with the catalyst modified glassy carbon electrode as sample (WE1, ø =

EP

5 mm), a Pt mesh as counter electrode, a Ag/AgCl/3 M KCl as reference electrode and a Pt microelectrode as SECM tip (WE2, ø = 25 µm). The working distance of the tip electrode was

AC C

12.5 µm. The collection efficiency (CE) of the tip towards produced hydrogen peroxide at the sample was determined using a Hg sample electrode to ensure a 2 electron process during oxygen reduction.[39] A constant potential of −0.4 V vs. Ag/AgCl 3 M KCl was applied at the Hg electrode and a constant potential of 0.7 V vs. Ag/AgCl 3 M KCl at the SECM tip. The CE was calculated according to CE = −itip / isample with itip as the tip current and isample as the current measured at the sample. The CE was determined to be 3.7•10−4.

ACCEPTED MANUSCRIPT The number of transferred electrons (n) was determined by applying a LSV in the potential window used before for the ORR (5 mV s−1) at the sample electrode and a constant potential of 0.7 V vs. Ag/AgCl 3 M KCl at the tip electrode. The measured currents, isample and itip, were used for calculation according to equation 3.[40,41] 

 

RI PT

=

   !

(3)

 

•


%

(4)

M AN U

" #  $ =

SC

The amount of detected hydrogen peroxide was calculated according to:

2.7 X-ray photoelectron spectroscopy (XPS)

XPS measurements were recorded in an ultrahigh vacuum (UHV) set-up having a monochromatic Al Kα X-ray source (1486.6 eV; anode operating at 14.5 kV and 30.5 mA) and a highresolution Gammadata-Scienta SES 2002 analyser. A flood gun was used to compensate for

TE D

the charging effects. The base pressure in the measurement chamber was ∼10−10 mbar. For

EP

high resolution spectra, a pass energy of 200 eV was applied.

2.8 X-ray diffraction (XRD)

AC C

X-Ray diffraction (XRD) data was obtained using a Panalytical X'PERT Pro MPD X-ray diffractometer with a Cu Kα radiation source (λ = 1.5418 Å) in the range 2θ = 20–90°. Data treatment has been performed using the Highscore Plus Software package.

2.9 Scanning electron microscopy (SEM) SEM images were taken using a Quanta ED FEG scanning electron microscope (FEI) with an acceleration voltage of 20 kV.

ACCEPTED MANUSCRIPT 2.10 Transmission electron microscopy TEM was performed on a JEOL microscope (JEM-2800) with a Schottky-type emission source working at 200 kV. Energy dispersive X-ray spectroscopy (EDS) mapping was performed with the equipped double SDD detectors at a solid angle of 0.98 steradians with a detection area

RI PT

of 100 mm2. Due to its paramagnetic nature, during collection of the TEM images the sample was embedded in a carbon-based matrix in order to prevent its detachment from the TEM grid inside the high-vacuum chamber. For this purpose, a suspension containing 3 mg of

SC

sample material, 20 µL Leit-C conductive carbon cement (Plano) and 100 µL toluene was

M AN U

prepared. 50 µL of this suspension were spin-coated on a copper TEM grid covered with a 610 nm thick carbon film (Plano) using a Laurell WS-650-23 spin-coater at 6000 rpm.

3 Results and discussion

TE D

Electrodes modified with bifunctional oxygen electrocatalysts were prepared by a comparatively simple three-step procedure. The electrodes were first modified with a film of CoxY (Y = B or P) homogeneously suspended in a solution of the benzoxazine (BO) oligomer

EP

BA-tepa. BA-tepa is obtained as the Mannich type reaction product of bisphenol A (BA),

AC C

tetraethylenpentamine (tepa) and formaldehyde.[37,38] BA-tepa was deliberately chosen because its resulting polymer shows a high thermal stability towards pyrolysis temperatures of up to 800 °C,[42] while the five amino groups in tepa introduce nitrogen into the formed pyrolytic carbon. The modified electrodes were subjected to thermal treatment for subsequent polymerisation and pyrolysis of the BA-tepa film in argon atmosphere. In the second step, the deposited mixture of BA-tepa and CoxY particles was polymerized forming an electrically insulating organic polymer film embedding the catalyst particles. In order to create an electrically conductive environment and to form a nitrogen-doped carbon matrix,

ACCEPTED MANUSCRIPT in the third step, the composite film was pyrolysed at temperatures between 500 – 700 °C. The resulting composite consists of the CoxY nanoparticles embedded in the BA-tepa derived nitrogen-doped carbon (NC) matrix. The NC matrix does not only activate the CoxY/NC composite (Figure S1 & S2) for the ORR, but also serves as a stable catalyst fixation matrix,

RI PT

which is essential during OER, as reported in our earlier work.[33]

3.1 Influence of pyrolysis temperature

SC

The pyrolysis temperature of the CoxY/pBO composites can have a significant influence on

M AN U

the activity of the resulting catalysts. CoxB embedded in a pBO film was therefore pyrolysed at various temperatures between 500 °C and 700 °C to find the optimal conditions for the highest bifunctional activity. Comparisons of the necessary potential to achieve current densities of −1 mA cm−2 (for the ORR) and +10 mA cm−2 (for the OER) reveal a substantial

TE D

difference in the ORR activity in the temperature range between 500 °C and 700 °C (Figure 1a&b). The potential needed to attain a current density of −1 mA cm−2 during ORR was 0.54 V vs. RHE for samples pyrolysed at 500 °C and 0.81 V vs. RHE for samples pyrolysed at

EP

650 °C (Figure 1a). A further increased of the temperature to 700 °C leads to a decrease in

AC C

ORR activity. Similar tendencies are also reflected in the current densities attained at 0.2 V vs. RHE, indicating different preferential ORR pathways (Figure S2). In contrast to the ORR, the pyrolysis temperature had only a minor influence on the OER activity. For example, at a current density of +10 mA cm−2, a difference of only 60 mV was observed between the least and most active material formed at 500 °C and 650 °C, respectively (Figure 1b & Figure S2). Similarly, the Tafel slopes for the OER of the various materials (Figure S2c) represented the same trend in activity as the potential to achieve a current density of +10 mA cm-2, with the exception of the material prepared at 600 °C which shows the lowest Tafel slope (66 mV dec-

ACCEPTED MANUSCRIPT 1

). Despite pyrolysis at 650 °C did not lead to the lowest Tafel slope, the material shows the

highest bifunctional activity. This is represented by a sum overpotential of as low as 0.8 V between the ORR at −1 mA cm−2 and the OER at +10 mA cm−2 and an OER Tafel slope of 75 mV dec-1 (Figure 1c & Figure S2c). Examination of the surface state of the CoxB/NC catalysts

RI PT

prepared at the various temperatures using XPS revealed differences between the samples. All samples had dominant boron, nitrogen and cobalt signals, however, the surface oxidation states of the acquired signals varied substantially (Figure 1d-f). Elemental boron and Co2B

SC

can be observed in all the samples regardless of the preparation temperature at binding

M AN U

energies of 187.4 eV and 778.4 eV, respectively (Figure 1d&f).[20] The observed binding energy of 778.4 eV for Co2B is in good agreement with previously reported values.[20,43] Pyrolysis at 500 °C led to a material with a high degree of surface oxidation, as evidenced by the dominant peaks of boron-oxo species at 191.9 eV (Figure 1f) and the Co 2p3/2 and Co

TE D

2p1/2 peaks at 781.5 eV and 797.3 eV, respectively, indicating the formation of oxidised Co species (Figure 1d). Together with surface oxidation, this led to comparably poor performance towards the ORR as shown by the LSV of the 500 °C sample (Figure S2a). For

EP

the pyrolysis temperatures 600 °C, 650 °C and 700 °C, not only the peak for boron-oxo

AC C

species is significantly less intense, but also the Co 2p signals for oxidised Co species at 781.5 eV and 797.3 eV. A drastic decrease of the contribution of oxidised species is observed for the sample pyrolysed above 650 °C in argon. The decreased oxidation of the 650 °C sample led to a much higher activity towards the ORR (Figure 1a). In contrast, the pyrolysis temperature does not similarly influence the OER (Figure 1b&S2b) because the composition and nature of the OER active sites differs from those involved in the ORR. As described elsewhere,[20] a Co(OH)2/Co2B core-shell configuration which is formed prior to the OER represents the active form of the OER catalyst. CoxB/NC and CoxP/NC were further prepared

ACCEPTED MANUSCRIPT by pyrolysis at 650 °C for in-depth physical and electrochemical investigations. After pyrolysis, the structure and morphology of the prepared materials were investigated by XRD, XPS, SEM, TEM and EDS. 3.2 Characterisation of optimised CoxB/NC and CoxP/NC

RI PT

Powder diffraction patterns of both composite materials show distinct signals evoked by reflections at the (111), (200), (220) and (100), (002), (101), (103) planes of cubic (fcc, pdf 00015-0806) and hexagonal (hcp, pdf 01-089-7373) cobalt, respectively (Figure 2a & Figure

SC

S3a). Additionally, reflections of amorphous cobalt boride and Co2B at 2 θ angles of 45° and

M AN U

42.7°, 45.9°, respectively, indicate that cobalt boride is present in the amorphous as well as in the crystalline phase in the CoxB/NC composite (Figure 2a).[44] Similarly, cobalt phosphide is present in the amorphous as well as in the crystalline phase in CoxP/NC (Figure S3a). The broad feature centred around 2 θ = 45° indicates the presence of amorphous cobalt

TE D

phosphide[45] and a small peak at 2 θ = 41.2° can be assigned to reflections at the (111) plane of crystalline Co2P (pdf 01-072-9563). The second intense reflection at the (201) plane of Co2P occurs at 44.8° and is superimposed by the strong reflection of cobalt. Small

EP

reflections at 2 θ = 42.5° and 62.7° may be due to traces of cobalt oxide in the sample. The

AC C

XRD patterns of both materials show additional reflections at 2 θ values around 26°, which can be assigned to the (002) reflection of graphitic carbon.[31] Other reflections of graphitic carbon e.g. at the (101), (004) or (110) planes are too low to be detected in the current samples, which hampers conclusions about the final structure of the carbon matrix. However, previously reported results of thermally treated pBO as catalyst immobilisation matrix concluded a mixture of amorphous and partially graphitised carbon to be most likely.[33] The XP spectrum of the cobalt core-level region of CoxB/NC shows a main binding energy for the Co 2p3/2 peak at 778.4 eV, which is in good agreement with the binding

ACCEPTED MANUSCRIPT energy of 778.4 eV for crystalline Co2B (Figure 2b). Deconvolution of the spectrum reveals the presence of Co2+ species as in CoO at a binding energy of 779.3 eV. Surface oxidation is a well-known process often occurring at this class of materials when exposed to atmospheric conditions.[20,36] Similarly, the main peak of the Co 2p3/2 region reveals oxidation of

RI PT

CoxP/NC with the main peaks at the binding energies of 781.4 eV and 780.4 eV corresponding to Co(OH)2 and CoO, respectively (Figure S3b). In addition to the XRD results, the presence of metallic cobalt and cobalt phosphide is indicated by the peak at a binding

SC

energy of 778.7 eV. The XPS spectra of the non-metal moieties indicate a partial oxidation

M AN U

for both B as well as P, but more importantly, both elements interact intimately with cobalt (Figure 2c & Figure S3c). The B 1s spectrum of CoxB/NC reveals the presence of two distinct boron species at binding energies of 187.4 eV and 192.0 eV (Figure 2c). The former one points towards boron in interaction with cobalt while the latter can be assigned to boron-

TE D

oxo species. Deconvolution of the B 1s core level spectrum showed a third peak at binding energies of 189.7 eV correlating to B-N species.[46] The P 2p core-level spectrum was deconvoluted into two doublets with P 2p3/2 signals at binding energies at 123.0 eV and 133.1 eV

EP

(Figure S3c). The former signal matches well phosphorus in interaction with cobalt as in Co2P

AC C

and CoP while the latter indicates oxidised phosphorus as in phosphates and pyrophosphates. Generally, both materials exhibit a certain degree of surface oxidation with non-oxidised cores as it is expected and already reported for metal-metalloid materials.[20,36] The N 1s spectra of both composites exhibit signals for pyridinic, pyrrolic and graphitic nitrogen at binding energies of 398.9 eV, 400.7 eV and 404.0 eV, respectively (Figure 2d & Figure S3d). Scanning electron microscopy reveals the occurrence of nanoparticles embedded in the carbon matrix for both composite materials (Figure 3 and Figure S4). For purpose of characterisation, the composites where prepared at a larger scale

ACCEPTED MANUSCRIPT in a quartz glass boat and ground afterwards to a powder. The obtained composite particles consist of nanometre sized particles embedded in a carbon matrix. Energy dispersive X-ray spectroscopy (EDS) mapping reveals a homogeneous distribution of the expected elements Co, C and P within the composite (Figure 3c-f). Due to the low weight of boron, its signal is

RI PT

not detectable with EDS mapping (Figure S4c-e), however, its presence and chemical state was already demonstrated by XRD and XPS. Scanning transmission electron microscopy (STEM) imaging confirms the embedded particles to be nanometre sized (Figure 4a & Figure

SC

S5-S8). EDS mapping for CoxP/NC reveals that the sample is characterised by very small

M AN U

particles in combination with comparatively larger nanoparticles with different compositions (Figure 4 & Figure S6-S8). The very small nanoparticles are cobalt rich, thus most likely cobalt nanoparticles, while the larger particles consist of cobalt and phosphorus (Figure 4 & Figure S9). Islands with cobalt only particles are also present (Figure S7). STEM images of CoxB/NC

TE D

revealed the presence of cobalt-based nanoparticles (Figure S5), however, the low weight of boron prevented its mapping. Although characterisation by STEM, SEM, EDS, XRD and XPS suggests the presence of elemental cobalt, however, the presence of cobalt as solid solution

EP

cannot be ruled out completely.

AC C

3.3 Electrochemical investigation of CoxB/NC and CoxP/NC Rotating disk electrode (RDE) LSVs at various rotation speeds indicate a mass transport limited current response with nearly equidistant current/potential transients at 100, 400, 900 and 1600 rpm for both materials (Figure 5a&c). The potentials needed to yield a current density of −1 mA cm−2 were as high as 0.83 V and 0.82 V vs. RHE for CoxB/NC and CoxP/NC, respectively, placing them among the lowest reported for cobalt-based ORR catalysts [12,13,47–58], however, they do not outperform the solely ORR catalyst Pt/C (Figure S10). In general, the ORR may proceed either in a 4 e− transfer pathway to hydroxide ions or via a 2

ACCEPTED MANUSCRIPT e− transfer pathway to hydrogen peroxide, resulting in a different number of transferred electrons per oxygen molecule (Figure 5f).[59,60] The formed hydrogen peroxide can either disproportionate in solution or undergo a further 2 e− reduction to hydroxide. The formation of hydrogen peroxide is undesired because of its corrosive and oxidising nature. In order to

RI PT

determine the mechanism of the ORR at the two catalysts, the number of electrons transferred per oxygen molecule was derived by Koutecký-Levich (KL) analysis (Figure 5b&d). According to the KL analysis, CoxB/NC transfers approximately 3.2 electrons per oxygen

SC

molecule while CoxP/NC preferentially reduces oxygen via the 4 electron transfer

M AN U

mechanism. However, KL analysis of the ORR on electrodes modified with catalyst powders can only give an apparent number of transferred electrons. Conventionally, the actual amount of hydrogen peroxide is measured by rotating ring disk electrode (RRDE) experiments, at which oxygen is reduced at the disk and the eventually formed hydrogen

TE D

peroxide is detected at the ring electrode by its electrochemical oxidation. This value in turn is used for calculation of the number of transferred electrons during ORR. The here applied methodology of electrode preparation, however, does not allow the use of the very precisely

EP

manufactured RRDEs because they would be damaged during the pyrolysis procedure.

AC C

Therefore, scanning electrochemical microscopy (SECM) was applied to validate the number of transferred electrons derived by the KL analysis and to detect and quantify the amount of formed hydrogen peroxide (Figure 5e).[41] Working in the sample generation/tip collection mode of SECM, a 25 µm diameter platinum microelectrode tip was positioned closely above the catalyst modified electrode at a working distance of one tip radius (12.5 µm). The potential of the catalyst modified sample electrode was swept from 1.0 to 0.0 V vs. RHE, similar to the potential window used for RDE voltammetry, while the tip potential was kept constant at 0.7 V vs. Ag/AgCl/3 M KCl, a potential sufficiently high to oxidise formed

ACCEPTED MANUSCRIPT hydrogen peroxide. The recorded tip and sample currents (Figure S11) were used to calculate the number of transferred electrons and the amount of hydrogen peroxide produced (see Experimental Section). Contrary to the results from the KL analysis, CoxB/NC reduced oxygen over the entire potential window via a 4 electron mechanism without

RI PT

significant hydrogen peroxide production (Figure 5e), while between 3.7 and 4.0 electrons were transferred during the ORR on CoxP/NC with a maximum of 15 % hydrogen peroxide production at potentials around 0.15 vs. RHE (Figure 5e). The results obtained from SECM

SC

therefore confirm that both catalysts reduce O2 preferentially via the 4 electron transfer

M AN U

pathway. The high electrocatalytic activity towards complete oxygen reduction in a 4 electron transfer mechanism is presumably due to the presence of pyridinic and pyrrolic nitrogen, as well as the boron and phosphorus moieties linked to the cobalt atoms and/or ions. Linear sweep voltammetry revealed higher OER activity of the carbon containing

TE D

composites, CoxB/NC and CoxP/NC, both of which required 0.05 V less overpotential to drive a current density of +10 mA cm−2, compared to their pristine CoxB and CoxP counterparts (Figure 6a). Both materials are similarly active as the tested RuO2 reference sample and

EP

outperform the IrO2 reference samples, which underlines their high OER activity (Figure

AC C

S10). After the thermal treatment process, which resulted in embedment of the cobalt/cobalt metalloids inside the NC matrix, the Tafel slope changed substantially, from 103 mV dec−1 for the pristine material (CoxB) to 80 mV dec−1 for CoxB/NC. For the phosphidebased materials, a Tafel slope of 81 mV dec−1 was determined for both the composite and the pristine material (Figure 6b). A change in the Tafel slope generally indicates a change in the nature of OER active sites, and thus the rate determining step.[61] Different active sites would also explain the substantial change in overpotential. Therefore, the thermal treatment that transforms the insulating polybenzoxazine matrix into the conductive

ACCEPTED MANUSCRIPT nitrogen-doped carbon matrix, simultaneously enhances the OER activity of the composite. The feasibility of the prepared boride- and phosphide-based materials embedded in a stabilising nitrogen-doped carbon matrix for their possible application as catalysts for bifunctional oxygen electrodes in rechargeable metal-air batteries was investigated by pulse

RI PT

chronopotentiometry (Figure 7), alternating between the OER and the ORR. In rechargeable metal-air batteries, the OER takes place during the charging process while the ORR occurs upon discharging. These processes were simulated by current steps of 10 mA cm−2 and

SC

−1 mA cm−2 for the OER and ORR, respecttively, for 15 min each (Figure 7). For both CoxB/NC

M AN U

and CoxP/NC, the OER response did not show any significant change in the measured potential window, indicating highly stable OER active centres as well as a stable immobilisation of the powder materials on the electrode surfaces. However, the round-trip potential, expressed as the difference between the potentials measured at current densities

TE D

of +10 mA cm−2 and −1 mA cm−2, changed from 0.84 V to 1.1 V for CoxB/NC and from 0.8 V to 1.03 V for CoxP/NC. The initial bifunctional activities of the presented materials are comparable to the activities of bifunctional metal metalloid catalysts.[13,52] The increase in

EP

the round-trip voltage arises upon subjection of the ORR active sites to high anodic

AC C

potentials during OER and is a notable problem for bifunctional oxygen catalysts.[48,62] Here, in the specific case of CoxB- and CoxP-based materials, polarisation to high anodic potentials irreversibly oxidises the N-coordinated cobalt species, forming a Co(OH)2 shell around a metal-metalloid core.[20,36] Nevertheless, the potential response reached a steady state after the fifth alternating ORR/OER cycle. The deactivation process of ORR active sites has to be thoroughly investigated and addressed in order to achieve better cycling stability of bifunctional ORR/OER electrocatalysts.

ACCEPTED MANUSCRIPT 4. Conclusion A facile procedure to activate the bifunctional water splitting catalysts, CoxB and CoxP, for the ORR is presented by embedding them in a nitrogen-doped carbon matrix. The carbon matrix additionally acts as a catalyst powder stabilisation matrix. By pyrolysis of CoxB and

RI PT

CoxP polymer composites directly on an electrode surface, a conducting and stable N-doped carbon matrix is formed with the catalyst particles embedded in it, obviating the use of nonconductive organic binders like Nafion. The prepared materials were investigated for their

SC

electrochemical activity towards the ORR, the OER, and as bifunctional OER/ORR catalysts.

M AN U

CoxB/NC and CoxP/NC showed good performance for both reactions preferentially reducing O2 to OH− and exhibiting low overpotentials for both ORR and OER. The round-trip voltage between the ORR at −1 mA cm−2 and the OER at 10 mA cm−2 was only 0.81 V. The prepared catalyst films show high stability on the electrode surface. However, the cycle lifetime of the

active sites.

EP

Conflicts of interest

TE D

catalysts needs further improvement by optimisation of the oxidative stability of the ORR

AC C

There are no conflicts to declare Acknowledgements

Mrs. Sandra Schmidt is acknowledged for recording the SEM images. We thank Prof. Markus Winterer (University Duisburg-Essen) for providing access to his XRD instrument. The authors are grateful to the Deutsche Forschungsgemeinschaft (DFG) in the framework of the CRC/TRR 247 (project A2) and the Bundesministerium für Bildung und Forschung (BMBF) in the framework of the project “Mangan” (FKZ

ACCEPTED MANUSCRIPT 03EK3548). P.W. is grateful to the Association of the Chemical Industry (VCI) for funding of the PhD fellowship.

AC C

EP

TE D

M AN U

SC

RI PT

References [1] F. Cheng, J. Chen, Metal-air batteries: from oxygen reduction electrochemistry to cathode catalysts, Chem. Soc. Rev. 41 (2012) 2172–2192. [2] J.-S. Lee, S. Tai Kim, R. Cao, N.-S. Choi, M. Liu, K.T. Lee, J. Cho, Metal-Air Batteries, Adv. Energy Mater. 1 (2011) 2. [3] J. Wang, K. Wang, F.-B. Wang, X.-H. Xia, Bioinspired copper catalyst effective for both reduction and evolution of oxygen, Nat. Commun. 5 (2014) 5285. [4] J.-I. Jung, H.Y. Jeong, J.-S. Lee, M.G. Kim, J. Cho, A bifunctional perovskite catalyst for oxygen reduction and evolution, Angew. Chem. Int. Ed. 53 (2014) 4582–4586. [5] S. Chao, M. Geng, 3,5-Diamino-1,2,4-triazole as a Nitrogen precursor to synthesize highly efficient Co-N/C non-precious metal bifunctional catalyst for oxygen reduction reaction and oxygen evolution reaction, Int. J. Hydrogen Energy 41 (2016) 12995–13004. [6] J. Su, G. Xia, R. Li, Y. Yang, J. Chen, R. Shi, P. Jiang, Q. Chen, Co 3 ZnC/Co nano heterojunctions encapsulated in N-doped graphene layers derived from PBAs as highly efficient bi-functional OER and ORR electrocatalysts, J. Mater. Chem. A 4 (2016) 9204–9212. [7] X. Qiao, S. Liao, R. Zheng, Y. Deng, H. Song, L. Du, Cobalt and nitrogen codoped graphene with inserted carbon nanospheres as an efficient bifunctional electrocatalyst for oxygen reduction and evolution, ACS Sustain. Chem. Eng. 4 (2016) 4131–4136. [8] V. Nikolova, P. Iliev, K. Petrov, T. Vitanov, E. Zhecheva, R. Stoyanova, I. Valov, D. Stoychev, Electrocatalysts for bifunctional oxygen/air electrodes, J. Power Sources 185 (2008) 727–733. [9] Y. Shao, S. Park, J. Xiao, J.-G. Zhang, Y. Wang, J. Liu, Electrocatalysts for Nonaqueous Lithium–Air Batteries, ACS Catal. 2 (2012) 844–857. [10] J. Christensen, P. Albertus, R.S. Sanchez-Carrera, T. Lohmann, B. Kozinsky, R. Liedtke, J. Ahmed, A. Kojic, A Critical Review of Li∕Air Batteries, J. Electrochem. Soc. 159 (2012) R1. [11] F.-D. Kong, S. Zhang, G.-P. Yin, Z.-B. Wang, C.-Y. Du, G.-Y. Chen, N. Zhang, Electrochemical studies of Pt/Ir–IrO2 electrocatalyst as a bifunctional oxygen electrode, Int. J. Hydrogen Energy 37 (2012) 59–67. [12] K. Chen, X. Huang, C. Wan, H. Liu, Efficient oxygen reduction catalysts formed of cobalt phosphide nanoparticle decorated heteroatom-doped mesoporous carbon nanotubes, Chem. Commun. 51 (2015) 7891–7894. [13] X. Zhong, Y. Jiang, X. Chen, L. Wang, G. Zhuang, X. Li, J.-g. Wang, J. Mater. Chem. A 4 (2016) 10575–10584. [14] Y. Lin, L. Yang, Y. Zhang, H. Jiang, Z. Xiao, C. Wu, G. Zhang, J. Jiang, L. Song, Defective carbon-CoP nanoparticles hybrids with interfacial charges polarization for efficient bifunctional oxygen electrocatalysis, Adv. Energy Mater. 8 (2018) 1703623. [15] Z. Zhou, N. Mahmood, Y. Zhang, L. Pan, L. Wang, X. Zhang, J.-J. Zou, CoP nanoparticles embedded in P and N co-doped carbon as efficient bifunctional electrocatalyst for water splitting, J. Energy Chem. 26 (2017) 1223–1230. [16] S. Anantharaj, S.R. Ede, K. Sakthikumar, K. Karthick, S. Mishra, S. Kundu, Recent trends and perspectives in electrochemical water splitting with an emphasis on sulfide, selenide, and phosphide catalysts of Fe, Co, and Ni, ACS Catal. (2016) 8069–8097. [17] Y.-P. Zhu, Y.-P. Liu, T.-Z. Ren, Z.-Y. Yuan, Self-Supported Cobalt Phosphide Mesoporous Nanorod Arrays, Adv. Funct. Mater. 25 (2015) 7337–7347. [18] Y. Liang, X. Sun, A.M. Asiri, Y. He, Amorphous Ni-B alloy nanoparticle film on Ni foam: rapid alternately dipping deposition for efficient overall water splitting, Nanotechnology 27 (2016) 12. [19] M. Liu, J. Li, Cobalt phosphide hollow polyhedron as efficient bifunctional electrocatalysts for the evolution reaction of hydrogen and oxygen, ACS Appl. Mater. Interfaces 8 (2016) 2158–2165. [20] J. Masa, P. Weide, D. Peeters, I. Sinev, W. Xia, Z. Sun, C. Somsen, M. Muhler, W. Schuhmann, Amorphous Cobalt Boride (Co2B) as a Highly Efficient Nonprecious Catalyst for Electrochemical Water Splitting: Oxygen and Hydrogen Evolution, Adv. Energy Mater. 6 (2016) 1502313-1 – 1502313-10. [21] X. Zhu, C. Tang, H.-F. Wang, Q. Zhang, C. Yang, F. Wei, Dual-sized NiFe layered double hydroxides in situ grown on oxygen-decorated self-dispersal nanocarbon as enhanced water oxidation catalysts, J. Mater. Chem. A 3 (2015) 24540–24546. [22] H.T. Chung, J.H. Won, P. Zelenay, Active and stable carbon nanotube/nanoparticle composite electrocatalyst for oxygen reduction, Nat. Commun. 4 (2013) 1922. [23] Y. Liu, G. Yu, G.-D. Li, Y. Sun, T. Asefa, W. Chen, X. Zou, Coupling Mo2 C with Nitrogen-Rich Nanocarbon Leads to Efficient Hydrogen-Evolution Electrocatalytic Sites, Angew. Chem. Int. Ed. 54 (2015) 10752–10757. [24] K. Elumeeva, J. Masa, D. Medina, E. Ventosa, S. Seisel, Y.U. Kayran, A. Genç, T. Bobrowski, P. Weide, J. Arbiol, M. Muhler, W. Schuhmann, Cobalt boride modified with N-doped carbon nanotubes as a high-performance bifunctional oxygen electrocatalyst, J. Mater. Chem. A 5 (2017) 21122–21129. [25] N.N. Ghosh, B. Kiskan, Y. Yagci, Polybenzoxazines‑New high performance thermosetting resins: Synthesis and properties, Prog. Polym. Sci. 32 (2007) 1344–1391. [26] B. Kiskan, N.N. Ghosh, Y. Yagci, Polybenzoxazine-based composites as high-performance materials, Polym. Int. 60 (2011) 167–177. [27] H. Yee Low, H. Ishida, Structural effects of phenols on the thermal and thermo-oxidative degradation of polybenzoxazines, Polymer 40 (1999) 4365– 4376. [28] C.H. Lin, S.X. Cai, T.S. Leu, T.Y. Hwang, H.H. Lee, Synthesis and properties of flame-retardant benzoxazines by three approaches, J. Polym. Sci. A Polym. Chem. 44 (2006) 3454–3468. [29] H. Zhang, W. Gu, R. Zhu, Q. Ran, Y. Gu, Nitrogen configuration of polybenzoxazine carbide, High Temp. Mater. Processes 34 (2015) 245–250. [30] K. Hemvichian, H. Ishida, Thermal decomposition processes in aromatic amine-based polybenzoxazines investigated by TGA and GC–MS, Polymer 43 (2002) 4391–4402. [31] M. Sevilla, A.B. Fuertes, Fabrication of porous carbon monoliths with a graphitic framework, Carbon 56 (2013) 155–166. [32] L. Wan, J. Wang, Y. Sun, C. Feng, K. Li, Polybenzoxazine-based nitrogen-containing porous carbons for high-performance supercapacitor electrodes and carbon dioxide capture, RSC Adv. 5 (2015) 5331–5342. [33] S. Barwe, C. Andronescu, J. Masa, E. Ventosa, S. Klink, A. Genç, J. Arbiol, W. Schuhmann, Polybenzoxazine-derived N-doped carbon as matrix for powderbased electrocatalysts, ChemSusChem 10 (2017) 2653–2659. [34] C. Andronescu, S. Barwe, E. Ventosa, J. Masa, E. Vasile, B. Konkena, S. Möller, W. Schuhmann, Powder catalyst fixation for post-electrolysis structural characterisation of NiFe layered double hydroxide based oxygen evolution reaction electrocatalysts, Angew. Chem. Int. Ed. (2017) 11258–11262. [35] B. Ganem, J.O. Osby, Synthetically useful reactions with metal boride and aluminide catalysts, Chem. Rev. 86 (1986) 763–780. [36] J. Masa, S. Barwe, C. Andronescu, I. Sinev, A. Ruff, K. Jayaramulu, K. Elumeeva, B. Konkena, B. Roldan Cuenya, W. Schuhmann, Low overpotential water splitting using cobalt–cobalt phosphide nanoparticles supported on nickel foam, ACS Energy Lett. (2016) 1192–1198.

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

[37] C. Andronescu, S. Pöller, W. Schuhmann, Electrochemically induced deposition of poly(benzoxazine) precursors as immobilization matrix for enzymes, Electrochem. Commun. 41 (2014) 12–15. [38] A.A. Alhwaige, T. Agag, H. Ishida, S. Qutubuddin, Biobased chitosan/polybenzoxazine cross-linked films: preparation in aqueous media and synergistic improvements in thermal and mechanical properties, Biomacromolecules 14 (2013) 1806–1815. [39] C.J. van Velzen, M. Sluyters-Rehbach, A.G. Remijnse, G.J. Brug, J.H. Sluyters, The electrochemical reduction of oxygen to hydrogen peroxide at the dropping mercury electrode, J. Electroanal. Chem. Interfacial Electrochem. 134 (1982) 87–100. [40] C.M. Sanchez-Sanchez, A.J. Bard, Hydrogen peroxide production in the oxygen reduction reaction at different electrocatalysts as quantified by scanning electrochemical microscopy, Anal. Chem. 81 (2009) 8094–8100. [41] C.M. Sanchez-Sanchez, J. Rodriguez-Lopez, A.J. Bard, Scanning electrochemical microscopy. 60. Quantitative calibration of the SECM substrate generation/tip collection mode and its use for the study of the oxygen reduction mechanism, Anal. Chem. 80 (2008) 3254–3260. [42] P. Lorjai, S. Wongkasemjit, T. Chaisuwan, A.M. Jamieson, Significant enhancement of thermal stability in the non-oxidative thermal degradation of bisphenol-A/aniline based polybenzoxazine aerogel, Polym. Degrad. Stabil. 96 (2011) 708–718. [43] G. Mavel, J. Escard, P. Costa, J. Castaing, ESCA surface study of metal borides, Surf. Sci. 35 (1973) 109–116. [44] U.B. Demirci, P. Miele, Cobalt in NaBH4 hydrolysis, Phys. Chem. Chem. Phys. 12 (2010) 14651–14665. [45] N. Lu, J. Cai, L. Li, Dependence of interfacial adhesion of Co–P film on its microstructure, Surf. Coat. Technol. 206 (2012) 4822–4827. [46] F. Yang, Y.Z. Li, W. Chu, C. Li, D.G. Tong, Mesoporous Co–B–N–H nanowires, Catal. Sci. Technol. 4 (2014) 3168. [47] J. Masa, W. Xia, I. Sinev, A. Zhao, Z. Sun, S. Grützke, P. Weide, M. Muhler, W. Schuhmann, MnxOy/NC and CoxOy/NC nanoparticles embedded in a nitrogen-doped carbon matrix for high-performance bifunctional oxygen electrodes, Angew. Chem. Int. Ed. 53 (2014) 8508–8512. [48] D.M. Morales, J. Masa, C. Andronescu, Y.U. Kayran, Z. Sun, W. Schuhmann, Few-layer graphene modified with nitrogen-rich metallo-macrocyclic complexes as precursor for bifunctional oxygen electrocatalysts, Electrochim. Acta 222 (2016) 1191–1199. [49] S. Zhao, B. Rasimick, W. Mustain, H. Xu, Highly durable and active Co3O4 nanocrystals supported on carbon nanotubes as bifunctional electrocatalysts in alkaline media, Appl. Catal. B-Environ. 203 (2017) 138–145. [50] Y. Sun, Z. Shen, S. Xin, L. Ma, C. Xiao, S. Ding, F. Li, G. Gao, Ultrafine Co-doped ZnO nanoparticles on reduced graphene oxide as an efficient electrocatalyst for oxygen reduction reaction, Electrochim. Acta 224 (2017) 561–570. [51] A. Aijaz, J. Masa, C. Rosler, W. Xia, P. Weide, A.J.R. Botz, R.A. Fischer, W. Schuhmann, M. Muhler, Co@Co3O4 encapsulated in carbon nanotube-grafted nitrogen-doped carbon polyhedra as an advanced bifunctional oxygen electrode, Angew. Chem. Int. Ed. 55 (2016) 4087–4091. [52] H. Jiang, C. Li, H. Shen, Y. Liu, W. Li, J. Li, Supramolecular gel-assisted synthesis Co2P particles anchored in multielement co-doped graphene as efficient bifunctional electrocatalysts for oxygen reduction and evolution, Electrochim. Acta 231 (2017) 344–353. [53] T. Zhao, S. Gadipelli, G. He, M.J. Ward, D. Do, P. Zhang, Z. Guo, Tunable Bifunctional Activity of Mnx Co3-x O4 Nanocrystals Decorated on Carbon Nanotubes for Oxygen Electrocatalysis, ChemSusChem 11 (2018) 1295–1304. [54] K. Elumeeva, M.A. Kazakova, D.M. Morales, D. Medina, A. Selyutin, G. Golubtsov, Y. Ivanov, V. Kuznetzov, A. Chuvilin, H. Antoni, M. Muhler, W. Schuhmann, J. Masa, Bifunctional Oxygen Reduction/Oxygen Evolution Activity of Mixed Fe/Co Oxide Nanoparticles with Variable Fe/Co Ratios Supported on Multiwalled Carbon Nanotubes, ChemSusChem 11 (2018) 1204–1214. [55] T. Liu, L. Zhang, Y. Tian, Earthworm-like N, S-Doped carbon tube-encapsulated Co 9 S 8 nanocomposites derived from nanoscaled metal–organic frameworks for highly efficient bifunctional oxygen catalysis, J. Mater. Chem. A 6 (2018) 5935–5943. [56] T. Li, Y. Lu, S. Zhao, Z.-D. Gao, Y.-Y. Song, Co 3 O 4 -doped Co/CoFe nanoparticles encapsulated in carbon shells as bifunctional electrocatalysts for rechargeable Zn–Air batteries, J. Mater. Chem. A 6 (2018) 3730–3737. [57] K. Chakrapani, G. Bendt, H. Hajiyani, T. Lunkenbein, M.T. Greiner, L. Masliuk, S. Salamon, J. Landers, R. Schlögl, H. Wende, R. Pentcheva, S. Schulz, M. Behrens, The Role of Composition of Uniform and Highly Dispersed Cobalt Vanadium Iron Spinel Nanocrystals for Oxygen Electrocatalysis, ACS Catal. 8 (2018) 1259–1267. [58] X. Li, N. Xu, H. Li, M. Wang, L. Zhang, J. Qiao, 3D hollow sphere Co 3 O 4 /MnO 2 -CNTs, Green Energy & Environment 2 (2017) 316–328. [59] C.F. Zinola, A.M. Castro Luna, W.E. Triaca, A.J. Arvia, Kinetics and mechanism of the electrochemical reduction of molecular oxygen on platinum in KOH, J. Appl. Electrochem. 24 (1994) 531–541. [60] Y. Gorlin, C.-J. Chung, D. Nordlund, B.M. Clemens, T.F. Jaramillo, Mn 3 O 4 Supported on Glassy Carbon, ACS Catal. 2 (2012) 2687–2694. [61] T. Shinagawa, A.T. Garcia-Esparza, K. Takanabe, Insight on Tafel slopes from a microkinetic analysis of aqueous electrocatalysis for energy conversion, Sci. Rep. 5 (2015) 13801 (1-21). [62] S. Dresp, F. Luo, R. Schmack, S. Kühl, M. Gliech, P. Strasser, An efficient bifunctional two-component catalyst for oxygen reduction and oxygen evolution in reversible fuel cells, electrolyzers and rechargeable air electrodes, Energy Environ. Sci. 9 (2016) 2020–2024.

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

TE D

Figure 1. Averaged potentials at −1 mA cm−2 (a), 10 mA cm−2 (b), potential difference (c), for electrodes pyrolysed at various temperatures. Error bars represent the standard deviation between measurements (N = 3). High resolution XPS spectra for CoxB/NC in the Co 2p (d), N

AC C

EP

1s (e) and B 1s (f) regions.

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

TE D

Figure 2. XRD pattern of CoxB/NC (a). XPS spectra of the Co 2p3/2 (b), B 1s (c) and the N 1s (d)

AC C

EP

core-level regions of CoxB/NC.

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 3. SEM micrographs of CoxP/NC at 650x (a) and 150 000x (b) magnification. EDS

TE D

elemental mapping of the area shown in (c). The map shown in blue in (d) represents the distribution of cobalt, the red map (e) the distribution of carbon and the green map (f) the

AC C

EP

distribution of phosphorus (scale bar in c-f: 1 µm).

Figure 4. STEM image of CoxP/NC (a) and EDS mapping. Superimposed elemental maps of Co and P (b). Single elemental maps for Co (c) and P (d).

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 5. ORR activity of CoxY/NC (Y = B or P). LSVs at various rotation speeds for CoxB/NC (a) and CoxP/NC (c) (O2-saturated 0.1 M KOH, 5 mV s−1) with the corresponding KL analysis (b and d). Transferred number of electrons and hydrogen peroxide detection as a function of the applied potential by means of SECM (e). Simplified schematic representation of the different ORR pathways (f).

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 6. LSVs for the OER on CoxB/NC, CoxP/NC, CoxB and CoxP (1600 rpm, O2-saturated

AC C

0.1 M KOH, 5 mV s−1) (a) and the corresponding Tafel plots (b).

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 7. Alternating pulse chronopotentiometry for bifunctional stability test. Each cycle consists of 15 min OER at 10 mA cm−2 and 15 min ORR at −1 mA cm−2 (1600 rpm, O2-

TE D

saturated 0.1 M KOH). For each cycle, the potential values at 7.5 min for both reactions are

AC C

EP

presented. Error bars represent the standard deviation between measurements (N = 3).

ACCEPTED MANUSCRIPT Bifunctional water splitting catalysts were activated for the oxygen reduction reaction by embedding in nitrogen-doped carbon.

•

Direct preparation of CoxB/NC and CoxP/NC on the electrode surface obviates the use of non-conductive organic binders.

•

CoxB/NC and CoxP/NC showed good performance for both reactions preferentially reducing O2 to OH− and exhibiting low overpotentials for both ORR and OER.

AC C

EP

TE D

M AN U

SC

RI PT

•