ARTICLES

PUBLISHED ONLINE: 6 APRIL 2015 | DOI: 10.1038/NNANO.2015.48

A metal-free bifunctional electrocatalyst for

oxygen reduction and oxygen evolution reactions
Jintao Zhang1, Zhenghang Zhao2, Zhenhai Xia2 and Liming Dai1*

The oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) are traditionally carried out with noble metals
(such as Pt) and metal oxides (such as RuO2 and MnO2) as catalysts, respectively. However, these metal-based catalysts
often suffer from multiple disadvantages, including high cost, low selectivity, poor stability and detrimental environmental
effects. Here, we describe a mesoporous carbon foam co-doped with nitrogen and phosphorus that has a large surface
area of ∼1,663 m2 g−1 and good electrocatalytic properties for both ORR and OER. This material was fabricated using a
scalable, one-step process involving the pyrolysis of a polyaniline aerogel synthesized in the presence of phytic acid. We
then tested the suitability of this N,P-doped carbon foam as an air electrode for primary and rechargeable Zn–air
batteries. Primary batteries demonstrated an open-circuit potential of 1.48 V, a specific capacity of 735 mAh gZn −1
(corresponding to an energy density of 835 Wh kgZn −1), a peak power density of 55 mW cm−2, and stable operation
for 240 h after mechanical recharging. Two-electrode rechargeable batteries could be cycled stably for 180 cycles at
2 mA cm−2. We also examine the activity of our carbon foam for both OER and ORR independently, in a three-electrode
configuration, and discuss ways in which the Zn–air battery can be further improved. Finally, our density functional theory
calculations reveal that the N,P co-doping and graphene edge effects are essential for the bifunctional electrocatalytic
activity of our material.

T
hanks to their high energy densities, rechargeable metal–air Here, we report a template-free method for the scalable
batteries have been targeted as a promising technology to fabrication of three-dimensional N and P co-doped mesoporous
meet the energy requirements for future electric vehicles and nanocarbon (NPMC) foams, simply by pyrolysis of polyaniline
other energy-demanding devices1–4. The oxygen reduction reaction (PANi) aerogels synthesized in the presence of phytic acid (see
(ORR) and the oxygen evolution reaction (OER) are at the heart of Methods). The resultant NPMCs show bifunctional catalytic activi-
metal–air batteries: oxygen molecules are reduced by electrons from ties towards ORR and OER. As a result, Zn–air batteries fabricated
the current collector and combine with metal dissolved in the elec- with NPMCs show good performance and long-term stability.
trolyte during discharging, and the reverse process occurs during
charging. Among the metals targeted for this type of battery are Preparation and characterization of electrocatalyst
Li and Zn, both of which are currently under intense scrutiny, but Aniline monomers were polymerized in the presence of phytic acid to
Zn–air batteries have the fundamental advantage of being less produce a PANi hydrogel via a hard template-free gelation process28
costly and safer5–7. One of the major challenges for Zn–air battery (Fig. 1a). To determine the optimum conditions for the formation of
technology is to increase the O2 reduction and evolution efficiencies, the PANi hydrogel, different ratios of aniline and phytic acid were
which will require the development of stable and effective bifunc- examined (for more details see Supplementary Figs 1–4 and associated
tional electrocatalysts, possibly working in aqueous electrolytes discussions in the Supplementary Information). After freeze-drying
with air as the oxygen source6,8. Although precious metals such as the resultant PANi hydrogel into aerogel (Fig. 1b), subsequent pyro-
Pt, Ru and Ir have been used as such electrocatalysts9, their high lysis of the PANi aerogel led to the one-step formation of an NPMC
cost and poor stability are hampering commercialization of the foam (Fig. 1c and Supplementary Fig. 4). As can be seen in the
Zn–air battery technology. insets of Fig. 1b,c, pyrolysis caused a slight shrinkage of the macro-
Recent studies have shown that carbon nanomaterials (carbon porous structure. The individual mesoporous ligaments (Fig. 1d) are
nanotubes, graphene) doped with nitrogen could be an efficient, highly interconnected into a hierarchical porous network. A trans-
low-cost, metal-free alternative to Pt for ORR10–13. Co-doping mission electron microscopy (TEM) image and associated elemental
N-doped carbon nanomaterials with a second heteroatom, such as mapping (Fig. 1e) show the uniform distribution of C, N and P for a
B, S or P, can modulate the electronic properties and surface sample pyrolysed at 1,000 °C (NPMC-1000). Combined thermo-
polarities to further increase ORR activity14–17. In contrast, most gravimetric analysis and mass spectrometry (TGA-MS), X-ray diffrac-
electrocatalysts reported for OER so far are based on transition- tion (XRD), Raman and TEM studies (Supplementary Figs 5–9)
metal oxides18 supported by carbon materials to facilitate electron revealed that the pyrolysis converts most of the thermally stable
transfer19–24. Although the development of metal-based bifunctional domains (such as the benzene rings) into graphitic carbon domains
catalysts has recently attracted considerable attention25,26, the use of that are co-doped with N and P from the PANi and phytic acid,
carbon nanomaterials as bifunctional catalysts has rarely been respectively, accompanied by a release of decomposition gases (CO,
discussed27, and no truly metal-free bifunctional ORR and OER CO2 and so on, Supplementary Table 1). The resulting nanomaterial
catalyst has been reported to date. also possesses a large number of edge-like graphitic structures

1
Center of Advanced Science and Engineering for Carbon (Case4carbon), Department of Macromolecular Science and Engineering, Case Western Reserve
University, 10900 Euclid Avenue, Cleveland, Ohio 44106, USA. 2 Department of Materials Science and Engineering, Department of Chemistry, University of
North Texas, Denton, Texas 76203, USA. * e-mail: [email protected]

444 NATURE NANOTECHNOLOGY | VOL 10 | MAY 2015 | www.nature.com/naturenanotechnology

© 2015 Macmillan Publishers Limited. All rights reserved

NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2015.48 ARTICLES
a

Pyrolysis

n
tio
riza
ly me
Po

(i) (ii) OH (iii) HO

HO OH OH
P OH P
O O P OH O
NH2 HO O HO P O O OH
O O O
O P
P O O H3N
= = HO OH = O O
O P HO
P O
O O O
O OH HO O O P OH
HO P O O O
P P OH
OH OH HO
HO OH

b c d

1 cm

500 nm 500 nm 20 nm

400 nm
C N P

Figure 1 | Preparation of the N and P co-doped porous carbon (NPMC) electrocatalysts. a, Schematic illustration of the preparation process for the NPMC
foams. An aniline (i)–phytic acid (ii) complex (iii) is formed (for clarity, only one of the complexed anilines is shown for an individual phytic acid), followed
by oxidative polymerization into a three-dimensional PANi hydrogel crosslinked with phytic acids. As each phytic acid molecule can complex with up to six
aniline monomers, phytic acid can be used as the crosslinker and protonic dopant to directly form the three-dimensional PANi hydrogel network; for clarity,
only a piece of the two-dimensional network building block is shown in the enlarged view under the three-dimensional PANi hydrogel. The PANi hydrogel is
freeze-dried into an aerogel and pyrolysed in Ar to produce the NPMC (for clarity, only a piece of the two-dimensional NPMC network building block is
shown in the enlarged view under the three-dimensional NPMC). b,c, SEM images of PANi aerogel (b) and NPMC-1000 (c). Inset in c: Digital photo images
of PANi aerogel before (left) and after (right) pyrolysis at 1,000 °C. d,e, High-resolution TEM image (d) and TEM image (e, left), with corresponding
element mapping images of NPMC-1000 (e). The TEM image shows a piece of interconnected network-like scaffold. The element mapping images for C, N
and P show a uniform distribution of the elements.

NATURE NANOTECHNOLOGY | VOL 10 | MAY 2015 | www.nature.com/naturenanotechnology 445

© 2015 Macmillan Publishers Limited. All rights reserved

ARTICLES NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2015.48

a 1,200 c PANi aerogel d

PANi aerogel
Volume absorbed (cm3 g−1)
QI P1 PANi aerogel
NPMC-900 BA P2
900 NPMC-1000 NC
NPMC-1100
600

N1 NPMC-900
300 P–C NPMC-900
N2
N3 P–O
0 N0

Intensity (a.u.)

Intensity (a.u.)
0.0 0.2 0.4 0.6 0.8 1.0
Relative pressure (P/P0)
NPMC-1000 NPMC-1000
b
0.5 PANi aerogel
dV/dD (cm3 g−1 nm−1)

NPMC-900
0.4 NPMC-1000
NPMC-1100
0.3
NPMC-1100 NPMC-1100
0.2
0.1
0.0

1 10 100 406 404 402 400 398 396 135 132 129
Pore diameter (nm) Binding energy (eV) Binding energy (eV)

Figure 2 | Brunauer–Emmett–Teller characterization and XPS composition analysis. a,b, N2 adsorption–desorption isotherms (a) and corresponding pore
size distributions (b) for PANi aerogel, NPMC-900, NPMC-1000 and NPMC-1100. c,d, High-resolution XPS spectra of N1s (c) and P2p (d) for PANi aerogel,
NPMC-900, NPMC-1000, and NPMC-1100. In c, for the PANi aerogel, the fitted peaks correspond to quinonoid imine (QI), benzenoid amine (BA) and
nitrogen cationic radical (NC). In c, for NPMC samples, the fitted peaks correspond to oxidized pyridinic nitrogen (N0), pyridinic-N (N1), pyrrolic-N (N2) and
graphitic-N (N3). In d, for the PANi aerogel, the fitted peaks correspond to P atoms in phosphate species (P1 and P2) with different binding energies. In d,
for NPMC samples, the fitted peaks correspond to P–C and P–O.

(Supplementary Fig. 7), which play a crucial role in catalytic activity. numerical data in Supplementary Table 2. As expected, the XPS
Both the solution polymerization and the template-free pyrolysis spectra show peaks for C, N and P, as well as an O peak arising
process can be readily scaled up for low-cost mass production. mainly from the phytic acid precursor (Fig. 1a). Nevertheless, the
N2 adsorption–desorption isotherm curves (Fig. 2a) for the possibility of there being physically adsorbed oxygen in NPMCs
NPMC samples exhibit absorbed volumes that are remarkably cannot be ruled out, as the graphitic structure is known to be suscep-
larger than that of the PANi aerogel and that increase significantly tible to oxygen absorption, even at low pressure14. The fitted XPS
with increasing pyrolysis temperature. The specific surface areas peaks for N1s of the PANi aerogel (Fig. 2c) centred at ∼399.5 eV,
(Supplementary Table 2) are also significantly enlarged on increasing 400.3 eV and 401.6 eV are attributable to quinonoid imine, benze-
the pyrolysis temperature of the PANi aerogel (53.5 m2 g−1). In par- noid amine and nitrogen cationic radicals, respectively. The pres-
ticular, there is a significant difference between the samples heated at ence of the NC peak is indicative of the protonic doping of PANi
900 °C (NPMC-900, surface area of 635.6 m2 g−1) and that heated at by phytic acid30. The XPS N1s spectra for NPMC samples can be
1,000 °C (NPMC-1000, surface area of 1,548 m2 g−1). The observed deconvoluted into four different bands at ∼398.6, 400.5, 401.3 and
specific surface areas for NPMC-1000 and NPMC-1100 (1,663 m2 g−1) 402.0 eV, which correspond to pyridinic (N1), pyrrolic (N2), gra-
are also much larger than those of hard template-synthesized porous phitic (N3) and oxidized pyridinic (N0) nitrogen, respectively31.
carbons (∼500–1,200 m2 g−1)11,15,29. The more than doubled surface These various nitrogen species lead to different chemical/electronic
area of NPMC-1000 (compared with that of NPMC-900) agrees well environments for neighbouring carbon atoms and hence different
with the significant weight loss around that temperature revealed by electrocatalytic activities. The curve fitting in Fig. 2c and the corre-
TGA-MS spectra (Supplementary Figs 5 and 6) due, most probably, sponding normalized results (Supplementary Fig. 11) indicate a
to the generation of volatile species (CO, CO2 and so on, also conversion from pyrrolic to graphitic nitrogen with increasing
Supplementary Table 1) from carbonization of the PANi aerogel. temperature, consistent with previous reports on N-doped carbon
The type IV isotherm curves shown in Fig. 2a have an obvious hys- materials32. The P2p spectra (Fig. 2d) of the PANi aerogel were
teresis and confirm the existence of mesopores. The rapid nitrogen deconvoluted into two different bands at ∼132.9 eV (P1) and
uptake (P/P0 > 0.9) might be due to the presence of secondary, 133.8 eV (P2), corresponding to the core levels of P atoms in phos-
much larger pores. Barrett–Joyner–Halenda (BJH) pore size distri- phate species33. Following heat treatment, NPMC-900 showed two
bution curves derived from the N2 desorption branches confirm similar component peaks with a slight binding shift to lower
the presence of mesopores with diameters <10 nm (Fig. 2b) and sig- energy, arising from gradual dehydration and condensation of phos-
nificantly enhanced pore volumes from 0.17 cm3 g−1 for the PANi phoric groups into polyphosphates, and the subsequent charge-
aerogel to 0.50, 1.10 and 1.42 cm3 g−1 for NPMC-900, NPMC- transfer interaction of P with conjugated aromatic C rings to gener-
1000 and NPMC-1100 (Supplementary Table 2), respectively. ate P–C (131.8 eV) and P–O (133.4 eV) bonds in NPMC-1000 and
Overall, our data confirm that the one-step pyrolysis process pro- NPMC-110034,35. This suggests the successful doping of P heteroa-
duces NPMC samples with three-dimensional mesoporous struc- toms into the carbon network through thermal pyrolysis35.
tures with large surface area, high pore volume and a pore size that Further heating caused a gradual loss of the P–C peak due to
is appropriate for electrocatalytic applications. thermal decomposition of heteroatom dopants from the carbon
Typical X-ray photoelectron spectra (XPS) for NPMCs are pro- matrix (Supplementary Table 2), consistent with our TGA and
vided in Supplementary Fig. 10, together with a summary of TGA-MS results (Supplementary Figs 5 and 6). Thus, it is critical

446 NATURE NANOTECHNOLOGY | VOL 10 | MAY 2015 | www.nature.com/naturenanotechnology

© 2015 Macmillan Publishers Limited. All rights reserved

NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2015.48 ARTICLES
a 1 Pt/C b 400 625 c
0 0.6 V 0.5 V
NPMC-1000 900 1225
0 0.4 V 0.3 V
NPMC-1100 −1 1600 2025 0.4

j −1 (mA−1 cm−2)
0.2 V 0.6 V (Pt/C)
−1 NPMC-900 2500 r.p.m.

j (mA cm−2)
j (mA cm−2)

NMC-1000 −2
−2 NPC-1000 −3 0.3
−3
−4
−4
−5 0.2
−5
−6
−6
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 0.02 0.03 0.04 0.05
E (V versus RHE) E (V versus RHE) ω−1/2 (r.p.m.−1/2)

d 30 e f 16
2 Ring current (×10) Pt/C
12 NPMC-1000 OER
jk (mA cm−2)

Current density
20 0 NPMC-1100
150 μg cm−2

(mA cm−2)
j (mA cm−2)
8 RuO2
−2 450 μg cm−2
4
10
−4
0
−6 ORR
0 −4
Disc current
00
0
0

00

−8
10
00

00
9

10
1

C-
1

1
C-
C-

C-

C-

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.5 1.0 1.5 2.0
M
M
M

NM

NP
NP
NP
NP

E (V versus RHE) Potential (versus RHE)

Figure 3 | Electrocatalytic activity for ORR and OER. a, Linear scan voltammogram (LSV) curves for NPMC-900, NPMC-1000, NPMC-1100, NMC-1000,
NPC-1000 and commercial Pt/C catalyst at an RDE (1,600 r.p.m.) in O2-saturated 0.1 M KOH solution. Scan rate, 5 mV s−1. b, LSV curves of NPMC-1000 in
oxygen-saturated 0.1 M KOH at various rotating speeds. c, K–L plots for NPMC-1000 and Pt/C at various potentials. d, Kinetic current of various samples for
O2 reduction at 0.65 V. e, RRDE measurements (1,600 r.p.m.) of ORR at an NPMC-1000 electrode with different catalyst loadings. f, LSV curves of NPMC-
1000, NPMC-1100, RuO2 and commercial Pt/C catalyst on an RDE (1,600 r.p.m.) in 0.1 M KOH (scan rate, 5 mV s−1), showing the electrocatalytic activities
towards both ORR and OER.

to control the pyrolysis temperature to produce NPMCs with the the corresponding LSV curves for the purely N-doped mesoporous
desired mesopores as well as N and P contents. carbon (NMC-1000; see Methods) and N- and P-doped carbon
(NPC-1000; see Methods). As can be seen, the NPMC-1000 has
Electrochemical evaluation of NPMCs for ORR and OER the highest electrocatalytic activity of all the aforementioned
Cyclic voltammetry (CV) curves (Supplementary Fig. 12) were metal-free catalysts in terms of both the onset potential and limiting
found to exhibit oxygen reduction peaks for all the NPMC electro- current, highlighting the importance of the N,P co-doping and
des in O2-saturated KOH solution, but not the N2-saturated KOH mesoporous structure for ORR. The electron transfer number per
solution. The observed oxygen reduction peak shifted to more oxygen molecule (n) for ORR was determined from the LSV
positive potentials with increasing pyrolysis temperature from curves (Fig. 3b,c and Supplementary Fig. 14) according to the
800 to 1,000 °C, but slightly reversed on further increasing the Koutechy–Levich (K–L) equation (for details see Supplementary
temperature to 1,100 °C. A reduction potential similar to that of Information). The K–L plots (Fig. 3c) show linear relationships
the commercial Pt/C catalyst Pt/XC-72 (20 wt%) was observed at between j −1
k and ω
−1/2
(where jk is the kinetic current and ω is the
the NPMC-1000 electrode, suggesting a high electrocatalytic electrode rotating rate), with a similar slope for the NPMC-1000
activity of the metal-free NPMC catalyst. This is because pyrolysis and Pt/C electrodes, from which n was determined to be ∼4.0,
at a higher temperature normally leads to a higher degree of gra- suggesting a four-electron pathway for ORR11,13. The kinetic
phitization, with higher electrical conductivity (thus a lower current obtained from the intercept of the linearly fitted K–L
charge-transfer resistance in Supplementary Fig. 13) and hence plots at 0.65 V (versus RHE) for the NPMC-1000 electrode is the
better electrocatalytic activity (as from NPMC-900 to NPMC- largest among all the metal-free catalysts investigated in this study
1000). However, overheating (say from 1,000 to 1,100 °C) could (Fig. 3d). Although the NPMC-900 sample has the highest N and
lead to decomposition of the dopants (Supplementary Table 2) P contents (Supplementary Table 2), the relatively low pyrolysis
and hence the observed negative shift of the peak potential from temperature could lead to a high charge-transfer resistance
NPMC-1000 to NPMC-1100. The current density follows the (Supplementary Fig. 13) and hence a relatively poor electrocatalytic
same trend (Supplementary Fig. 12), indicating once again activity. Although the electrical conductivity could be enhanced by
that NPMC-1000 is more electrocatalytically active than both increasing the pyrolysis temperature, the doped heteroatoms would
NPMC-900 and NPMC-1100. be removed (Supplementary Table 2), resulting in a reduction in the
The linear scan voltammogram (LSV) curves in Fig. 3a confirm active sites and overall electrocatalytic activity, as exemplified by
the electrocatalytic performance of NPMC-1000, with a positive NPMC-1100 (Fig. 3a).
onset potential of 0.94 V versus reversible hydrogen electrode To further evaluate the ORR pathways for the NPMC-1000 elec-
(RHE) and a half-wave potential of 0.85 V versus RHE. These trode, we performed rotating ring–disk electrode (RRDE) measure-
values are comparable to those of Pt/C, and outperform most pre- ments. As shown in Fig. 3e, NPMC-1000 electrodes with two
viously reported metal-free ORR catalysts29,36 and even recently different mass loadings (150 and 450 μg cm−2) exhibited high disk
reported carbon-based catalysts with metals27,37,38. Furthermore, current densities ( ∼4 and 6 mA cm−2) for O2 reduction and
the limiting current of the NPMC-1000 electrode is much larger much lower ring current densities ( ∼0.007 and 0.014 mA cm−2)
than those of the NPMC-900 and NPMC-1100 electrodes, and is for peroxide oxidation. Notably, the disk current could be enhanced
comparable to that of Pt/C (Fig. 3a). Also included in Fig. 3a are significantly by increasing the mass loading and even become larger

NATURE NANOTECHNOLOGY | VOL 10 | MAY 2015 | www.nature.com/naturenanotechnology 447

© 2015 Macmillan Publishers Limited. All rights reserved

ARTICLES NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2015.48

a b c Utilization of Zn (%)
Pt/C 0 40 80
1.5 NPMC-1000

Power density (mW cm−2)

NPMC-1100 50
NPMC-900 5 mA cm−2
1.2

Potential (V)
1.2

Potential (V)
Zn 0.9
Air
O2 25 mA cm−2
0.6 25
0.9

0.3
Air electrode
0.0 0 0.6
Separator
0 20 40 60 0 328 656
Zn electrode
Current density (mA cm−2) Specific capacity (mAh gZn−1)

d e 1.6 f
5 mA cm−2
1.2 LED
Potential (V) 1.2
Potential (V)

20 mA cm−2
Mechanical recharge
0.8 NPMC-1000 0.8
Pt/C 1 Zn–air battery
NPMC-1000 2
0.4 3
0.4 4

0.0
0 10 20 30 0 20 40 60 80
Time (h) Time (h)

Figure 4 | Performance of a primary Zn–air battery. a, Schematic of the basic configuration of a primary Zn–air battery, in which a carbon paper pre-coated
with NPMC is used as an air cathode and is coupled with a Zn anode and a glassy fibre membrane soaked with aqueous KOH electrolyte as the separator.
The enlarged part illustrates the porous air electrode loaded with electrocatalyst, which is permeable to air. b, Polarization and power density curves of
primary Zn–air batteries using Pt/C, NPMC-900, NPMC-1000, NPMC-1100 as ORR catalyst (mass loading of 0.5 mg cm−2) and 6 M KOH electrolyte (scan
rate, 5 mV s–1). c, Specific capacities of the Zn–air batteries using NPMC-1000 as ORR catalyst are normalized to the mass of the consumed Zn. d,
Discharge curves of the primary Zn–air batteries using Pt/C and NPMC-1000 as ORR catalyst and KOH electrolyte at various current densities (5 and
20 mA cm−2). The observed sudden drop in voltage at 0 h was caused by a sudden increase in current density after resting the battery at the open-circuit
potential without current loading for the testing. e, Long-time durability of the primary Zn–air battery using NPMC-1000 catalyst at a current density of
2 mA cm−2. The battery is mechanically rechargable. The Zn and electrolyte were mechanically replaced at the point where the colour of the curve changes
(the numbers represent the first, second, third and fourth charge cycles). Coloured symbols above the curve are taken from the open-circuit potential by
opening the battery for each mechanical recharge. f, Optical images of an LED (∼2.2 V) before and after being driven by two Zn–air batteries in series.

than that of Pt/C. Supplementary Fig. 15a shows the percentage of Our calculations (shown in Fig. 6, in the following) also reveal
peroxide species with respect to the total oxygen reduction products, that NPMCs possess ORR activities comparable to, or even better
and Supplementary Fig. 15b shows the electron transfer numbers than, that of Pt in alkaline media. Thus, the N,P co-doped carbon
calculated from the RRDE curves (for details see Supplementary could show high catalytic activities in both acidic and
Information). It can be envisioned that oxygen molecules were alkaline environments.
reduced to water via a nearly four-electron pathway (n is over Figure 3f shows the rapidly increased anodic current above
3.85) with a small ratio of peroxide species (less than 8%). Similar ∼1.30 V associated with OER. The good OER catalytic activities
good electrocatalytic activities were also observed for NPMC-1000 for NPMCs were reflected by their lower onset potentials and
in 1 and 6 M KOH electrolytes (Supplementary Figs 16 and 17). higher currents than those of the Pt/C electrode (Fig. 3f ). As a refer-
In comparison with the Pt/C catalyst, the NPMC-1000 electrode ence, we also used a state-of-the-art OER electrode based on RuO2
exhibited better long-term stability, higher resistance to the metha- nanoparticles (see Supplementary Information for synthesis and
nol cross-over effect and CO poisoning effect (Supplementary Supplementary Fig. 21)42, and found that the NPMC-1000 electrode
Fig. 18) in oxygen-saturated 0.1 M KOH, and comparable catalytic exhibited a lower onset potential than the RuO2 nanoparticles,
activity, even in acidic electrolyte (Supplementary Figs 19 and 20). though slightly lower current densities at higher potentials
Because the activation energy of ORR on a catalyst is directly (Fig. 3f, see also Supplementary Fig. 22 in 6 M KOH).
related to its catalytic activity17,39, we used first-principles methods
to calculate the activation energy of ORR elemental steps in acidic NPMC as an air cathode in primary Zn–air battery
and alkaline environments. As depicted in Supplementary Bifunctional catalysts for both ORR and OER are highly desirable
Table 3, of the five barriers in acidic media, H2O formation in the for use in rechargeable Zn–air battery applications. To this end
last step of the ORR has the highest value (0.66 eV) and is therefore we explored the possibility of using NPMC as a metal-free bifunc-
the rate-limiting step for ORR (and is much smaller than that of a tional catalyst. We first constructed a primary Zn–air battery
Pt(111) surface (1.22 eV)40). In alkaline media, our calculated acti- using NPMC as the electrocatalyst (Fig. 4a). The open-circuit poten-
vation energy for the rate-limiting step of the Langmuir– tial (OCP) of this two-electrode primary Zn–air battery is as high as
Hinshelwood mechanism is ∼0.5 eV for N,P co-doped graphene, ∼1.48 V (Supplementary Fig. 23), suggesting a good catalytic per-
while the activation energies for N-doped graphene and on a formance of NPMC-1000, even in this cell configuration.
Pt(111) surface are ∼0.56–0.62 eV and 0.55 eV, respectively40,41. Figure 4b presents polarization and power density curves for

448 NATURE NANOTECHNOLOGY | VOL 10 | MAY 2015 | www.nature.com/naturenanotechnology

© 2015 Macmillan Publishers Limited. All rights reserved

NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2015.48 ARTICLES
a Time (h) b
0 10 20 30

3
Potential (V)

Air Zn O2
2 O2

Air electrode Separator Air electrode

0 60 120 180
Cycle number
c d Time (h)
0 30 60 96 97 98 99 100
4
2.5

3
Potential (V)

Potential (V)
2.0
Pt/C
2
NPMC-1000
NPMC-1100 1.5
1 Pt/C + RuO2

0 1.0
0 50 100 150 0 180 360 576 582 588 594 600
Current density (mA cm−2) Cycle number

Figure 5 | Performance of rechargeable Zn–air batteries. a, Discharge/charge cycling curves of two-electrode rechargeable Zn–air batteries at a current
density of 2 mA cm−2 using the NPMC-1000 air electrode. b, Schematic of the basic configuration of a three-electrode Zn–air battery by coupling the Zn
electrode with two air electrodes to separate ORR and OER. The enlarged parts illustrate the porous structures of the air electrodes, which facilitates gas
exchange. c, Charge and discharge polarization curves of three-electrode Zn–air batteries using NPMC-1000, NPMC-1100 or commercial Pt/C catalyst as
both air electrodes, together with the corresponding curve (that is, Pt/C + RuO2) for the three-electrode Zn–air battery with Pt/C and RuO2 nanoparticles as
each of the air electrodes, respectively. d, Discharge/charge cycling curves of a three-electrode Zn–air battery using NPMC-1000 for the air electrodes
(0.5 mg cm−2 for ORR and 1.5 mg cm−2 for OER) at a current density of 2 mA cm−2.

Zn–air batteries based on the NPMC air cathodes. The NPMC-1000 circuits. As exemplified in Supplementary Fig. 23, two Zn–air
catalyst showed a current density of ∼70 mA cm−2 and a peak power button batteries were connected in series to generate a sufficiently
density of ∼55 mW cm−2, comparable to those of a Pt/C catalyst high OCP of ∼2.8 V to power different light-emitting diodes
(∼60 mA cm−2 and 50 mW cm−2). The good performances of our (LEDs) (Fig. 4f and Supplementary Fig. 26).
NPMC-1000 foam derive from its porous structure, which facilitates
an efficient diffusion of O2 gas and electrolyte to the active sites. NPMC as an air cathode in rechargeable Zn–air battery
When normalized to the mass of consumed Zn, the specific capacity We also developed a rechargeable Zn–air battery1,4. In a recharge-
of our battery was over 735 mAh g−1 Zn (corresponding to an energy able Zn–air battery, the kinetics is mainly limited by the cathode
density of ∼835 Wh kg−1 Zn) at a current density of 5 mA cm ,
−2
reaction5:
which corresponds to ∼89.6% utilization of the theoretical capacity ORR
−→
(∼820 mAh g−1 5
Zn) (Fig. 4c). When the current density was increased O2 + 2H2 O + 4e− ←−4OH−
to 25 mA cm−2, the specific capacity of the battery was ∼689 mAh g−1Zn
OER

(corresponding to an energy density of ∼695 Wh kg−1 Zn). These A two-electrode rechargeable Zn–air battery with NPMC-1000 as its
values are higher than those of Zn–O2 batteries that use a metal- bifunctional catalyst (Supplementary Fig. 27) showed good rechar-
oxide-based CoO/CNT catalyst (∼570 Wh kg−1 −2 6
Zn at 10 mA cm ) . geability, as evidenced by 180 discharge/charge cycles over a period
Notably, the potential of the battery using NPMC-1000 (∼1.26 V) of 30 h (Fig. 5a), which is better than the performance of a Zn–air
at the current density of 5 mA cm−2 is also higher than batteries battery using core-corona-structured bifunctional catalyst with
using Pt/C (∼1.16 V). Furthermore, no significant potential drop lanthanum nickelate centres supporting N-doped carbon nanotubes
was observed under galvanostatic discharge for 30 h at 5 mA cm−2 (75 cycles for 12.5 h)8 and a MnO2/Co3O4 hybrid (60 cycles for 14 h)
and 14 h at 20 mA cm−2 (Fig. 4d), indicating a good catalytic stab- catalyst43. With the relatively poor catalytic activity of Pt/C for
ility for ORR. Although Zn is gradually consumed during the dis- OER in mind, for comparison we also fabricated a reference two-
charging process, the battery can be mechanically recovered by electrode Zn–air battery using mixed Pt/C and RuO2 as a bifunc-
refilling the Zn plate and KOH electrolyte. Typically, our battery tional electrocatalyst for ORR and OER (Supplementary Fig. 28).
can continue to work over 240 h with almost no potential decrease Compared with Fig. 5a, the results in Supplementary Fig. 28 show
during two cycles (Supplementary Fig. 24), comparing favourably a lower charging potential, indicating a relatively high catalytic
with most other recently reported primary Zn–air batteries activity towards OER.
(Supplementary Table 4). The battery can be mechanically Although NPMC-1000 accelerates both ORR and OER, a
charged for many cycles (Fig. 4e). Our Zn–air battery can also be certain degree of irreversibility is unavoidable due to the different
operated in KOH electrolyte at a lower concentration (1.0 M catalytic activities of the same catalyst toward the ORR and OER
KOH) with excellent durability (Supplementary Fig. 25). To meet reactions. Consequently, a deteriorating performance was observed
specific energy and/or power needs for various practical appli- for the two-electrode rechargeable Zn–air battery during a long-
cations, multiple Zn–air batteries can be integrated into series term cycling test (Supplementary Fig. 29)8. However, the catalytic

NATURE NANOTECHNOLOGY | VOL 10 | MAY 2015 | www.nature.com/naturenanotechnology 449

© 2015 Macmillan Publishers Limited. All rights reserved

ARTICLES NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2015.48

a N,P coupled c e
P doped
−0.5 N doped
Pt

−1.0
−ηORR (V)

−1.5

−2.0
−1 0 1 2 3 4
Free energy of adsorption OH* (eV)

b N,P coupled d f
−0.5
P doped
N doped
RuO2
−1.0
−ηOER (V)

−1.5 RuO2
−0.4

−0.5
−2.0
−0.6
1.50 1.65
−2.5
0.0 0.5 1.0 1.5 2.0
Difference between ΔG for O* and OH* (eV) C N P O H

g h O2(g) + 2H2O + 4e−

OOH* + H2O + OH− + 3e− OOH* + H2O + OH− + 3e−
2 U = −0.0797 V
U=0V 0
O* + H2O + 2OH− + 2e−
U=0V
O2(g) + 2H2O + 4e−
Free energy (eV)
Free energy (eV)

1 OH* + 3OH− + e−
O* H2O + 2OH− + 2e−

4OH− −1
0
U = 0.402 V
OH* + 3OH− + e−

−1 U = 0.402 V 4OH−
−2
U = 0.797 V
Reaction coordinate Reaction coordinate

Figure 6 | Mechanistic study of bifunctionality for ORR and OER. a,b, ORR (a) and OER (b) volcano plots of overpotential η versus adsorption energy of O*
and the difference between the adsorption energy of O* and OH*, respectively, for N-doped, P-doped and N,P-doped graphene. c–f, Initial structure (c) and
structures after the adsorption of hydroxyl OH* (d), oxyl O* (e) and peroxyl OOH* (f) intermediates on N and P coupled graphene. O*, OH* and OOH* are
adsorbed intermediates. The overpotentials of the best catalysts predicted theoretically for ORR (Pt) (ref. 47) and OER (RuO2) (ref. 42) are also plotted in a
and b, respectively. Inset in b: Detail of the volcano top in the main panel. g,h, Schematic energy profiles for the OER pathway (g) and the ORR pathway (h)
on N,P co-doped graphene in alkaline media.

activity of NPMC can be improved by optimizing the pore struc- Zn–air battery using Pt/C and RuO2 as the ORR and OER cata-
ture, the heteroatom doping site, electrode surface chemistry and lysts, respectively (Supplementary Fig. 30). As shown in
cell configuration. Indeed, the NPMC battery performance was sig- Supplementary Table 5, our battery is comparable to, or even
nificantly enhanced by using an optimized three-electrode con- better than, most of the recently reported rechargeable Zn–air
figuration (Fig. 5b), which prevents the bifunctional catalyst from batteries based on metal/metal-oxide electrodes6,45,46.
coming into contact with the oxidative (or reductive) potential
during ORR (or OER). In this case, the activities toward ORR Mechanistic study on ORR and OER of bifunctional NPMCs
and OER could be independently regulated by adjusting the cata- To gain further insights into the ORR and OER catalytic mechan-
lyst mass loading on each of the two air electrodes6,44, and a isms of NPMC, we performed first-principles calculations using
balanced reversible transfer between oxygen reduction and evol- density functional theory (DFT) methods to determine the elec-
ution was readily achieved. Figure 5c shows the discharge and tronic structures and catalytic reactions for the N,P co-doped
charge polarization curves for the three-electrode batteries with carbon structures (Fig. 6). In the co-doped structure, N and P
various air electrodes. The three-electrode rechargeable Zn–air may exist on the graphene in different forms, such as the isolated
battery using NPMC-1000 for the air electrodes showed no N-dopant, the isolated P-dopant and/or as N,P-coupled dopants
obvious voltage change over 600 discharge/charge cycles (for (in which N and P are close to each other). To study the
100 h, Fig. 5d), which is comparable to that of a three-electrode ORR/OER catalytic activities of these structures, all possible types

450 NATURE NANOTECHNOLOGY | VOL 10 | MAY 2015 | www.nature.com/naturenanotechnology

© 2015 Macmillan Publishers Limited. All rights reserved

NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2015.48 ARTICLES
of doping structure were built, as shown in Supplementary Fig. 31. on the NPMC metal-free air electrode operating in ambient air with
Furthermore, the doping positions in each of the structures were an aqueous KOH electrolyte exhibited an open-circuit potential of
changed with respect to the graphene edge to reveal the effect of ∼1.48 V, an energy density of ∼835 Wh kg−1 Zn and a peak power
doping sites. The possible ORR and OER pathways on N,P co- density of ∼55 mW cm−2, as well as excellent durability (over 240 h
doped graphene are listed in Supplementary equations (1) to (13) of operation after two mechanical recharges). A three-electrode
(Supplementary Section 4). As the overpotential η of ORR/OER is rechargeable battery using two NPMC metal-free air electrodes to
an important measure of the catalytic activities of a catalyst47, we separate ORR and OER also showed good stability (600 cycles for
calculated the overpotential for each active site on the doped struc- 100 h of operation). First-principles simulations revealed that the
tures and determined the minimum overpotential for ORR and N and P co-doping and the highly porous network of the carbon
OER. An ideal catalyst should be able to facilitate ORR and OER foam are crucial to generating bifunctional activity towards both
just above the equilibrium potential, with zero overpotential. ORR and OER. We anticipate that our nanomaterial will also be
However, the ideal case cannot be achieved because the binding useful for other electrocatalytic applications.
energies of the intermediates are correlated47. Thus, thermodynami-
cally, the lower overpotential indicates the better catalyst. Figure 6a,b Methods
Preparation of NPMCs. PANi aerogel was prepared by oxidative polymerization in
presents volcano plots, that is, the overpotential versus descriptors the presence of phytic acid according to a published procedure28. Typically, 5 ml of
for various reaction sites on N,P co-doped graphene structures in aniline monomer was added into 20 ml phytic acid solution (16%, wt∕wt in water).
alkaline environments. From this theoretical analysis, the isolated Ammonium persulphate (APS, 0.96 g) was dissolved in 10 ml deionized water under
N-doped, the isolated P-doped and the N,P-coupled structures stirring. After cooling to 4 °C, the two solutions were mixed together and kept
were identified to have minimum ORR overpotentials of 0.44, 0.47 overnight without stirring. The resultant hydrogel was washed by immersion in
deionized water for two days then freeze-dried for 24 h to produce polyaniline
and 0.47 V, respectively, while the lowest OER overpotentials for aerogel for pyrolysis. To prepare N,P co-doped mesoporous carbon (NPMC) foams,
these structures were 0.41, 0.49 and 0.39 V, respectively. Note that the PANi aerogel was calcined at the desired temperatures (900, 1,000 and 1,100 °C)
the N,P-coupled doping gives the best OER performance, while for 2 h under argon. The obtained samples were designated NPMC-900, NPMC-
the ORR overpotential could reach as low as 0.44 V based on the 1000 and NPMC-1100, respectively. For comparison, phytic acid was removed from
estimate of the volcano plot (Fig. 6a). Overall, the minimum the PANi hydrogel by a de-doping process against NH3·H2O washing. The pure
N-doped mesoporous carbon foam was then prepared by annealing the de-doped
overpotentials of N,P co-doped graphene for ORR and OER are PANi aerogel at 1,000 °C (designated NMC-1000). N and P co-doped carbon
0.44 V and 0.39 V respectively, lower than those of the best catalysts (NPC-1000) and RuO2 nanoparticles were also synthesized as references (for
identified theoretically (∼0.45 V for ORR on Pt47 and ∼0.42 V for preparation processes see Supplementary Information).
OER on RuO242), indicating that the bifunctional N,P co-doped gra- Electrocatalytic activity evaluation. All electrochemical measurements were
phene catalyst could outperform its metal/metal-oxide counter- conducted in a three-electrode configuration at room temperature (∼25 °C). The
parts. Clearly, the N and P co-doping generates synergistic effects potential, measured against an Ag/AgCl electrode, was converted to potential versus
to improve the electrocatalytic activities towards both OER and RHE according to Evs RHE = Evs Ag/AgCl + Eø Ag/AgCl + 0.059 pH. To prepare the
ORR. working electrode, 5 mg of NPMC samples were dispersed in an aqueous solution
containing 0.95 ml deionized water and 0.05 ml of 5 wt% Nafion under sonication.
For OER, the most active structure was identified to be the N,P- The obtained homogeneous catalyst ink (6 µl) was dropped onto a mirror-polished
coupled graphene shown in Fig. 6c, with the active site located at the glassy carbon electrode. The mass loading was 0.15 mg cm−2 unless otherwise
edge of the graphene. The elementary reactions of OER over gra- noted. The Pt/C (20 wt%, ETEK) electrode was prepared using the same procedure.
phene in an alkaline environment are shown in Fig. 6d–f. For the A 0.1 M KOH aqueous solution saturated with oxygen was used as the
active sites on this graphene, the OER is uphill when the electrode electrolyte unless otherwise stated.
For the Zn–air battery test, the air electrode was prepared by uniformly coating
potential is 0 V, but when the potential increases to 0.797 V the as-prepared catalyst ink onto carbon paper (SPECTRACARB 2040-A, Fuel Cell
(0.395 V in overpotential), all the elementary reaction steps store) then drying it at 80 °C for 2 h. The mass loading was 0.5 mg cm−2 unless
become downhill and OER occurs spontaneously over 0.797 V otherwise noted. A Zn plate was used as the anode. Both electrodes were assembled
(Fig. 6g). Because the OER overpotential is reduced by the co- into a home-made Zn–air battery, and 6 M KOH aqueous solution was used as
doping, the OER is facilitated overall by the N,P co-doped graphene. the electrolyte unless otherwise stated. The same procedure was used to prepare
air electrodes with a catalyst mass ratio of 1:3 as the ORR and OER electrodes,
Similar phenomena are observed for ORR, but the reaction active in a three-electrode rechargeable Zn–air battery.
sites are different from those for OER, although they are also
located near the graphene edges48. The most active site was ident- Received 30 April 2014; accepted 16 February 2015;
ified to be N-dopant (elementary reaction pathways have been published online 6 April 2015
described for N-doped graphene in our previous papers39,49). For
ORR, the most active site is also located at the edge and the elemen- References
1. Dunn, B., Kamath, H. & Tarascon, J-M. Electrical energy storage for the grid: a
tary reactions and energy change are shown in Fig. 6h. OOH for-
battery of choices. Science 334, 928–935 (2011).
mation is the rate-determining step of ORR, and is endothermic 2. Peng, Z., Freunberger, S. A., Chen, Y. & Bruce, P. G. A reversible and
in alkaline media under a voltage of 0.402 V. At zero external poten- higher-rate Li–O2 battery. Science 337, 563–566 (2012).
tial all reactions are downhill, except for O2 adsorption, which has a 3. Park, M., Sun, H., Lee, H., Lee, J. & Cho, J. Lithium-air batteries: survey on
small energy barrier that disappeared at −0.0797 V. Overall, the dis- the current status and perspectives towards automotive applications from a
battery industry standpoint. Adv. Energy Mater. 2, 780–800 (2012).
tance of the doping sites from the graphene edge seems critical to 4. Goodenough, J. B. Evolution of strategies for modern rechargeable batteries.
the adsorption of chemical species and catalytic activity. In most Acc. Chem. Res. 46, 1053–1061 (2013).
cases, OER and ORR usually occur near the edge of the graphene, 5. Kraytsberg, A. & Ein-Eli, Y. The impact of nano-scaled materials on advanced
but at different sites. metal–air battery systems. Nano Energy 2, 468–480 (2013).
6. Li, Y. et al. Advanced zinc–air batteries based on high-performance hybrid
electrocatalysts. Nature Commun. 4, 1805 (2013).
Conclusions 7. Lee, J-S. et al. Metal–air batteries with high energy density: Li–air versus Zn–air.
We have developed a low-cost and scalable approach to prepare Adv. Energy Mater. 1, 34–50 (2011).
three-dimensional mesoporous carbon foams co-doped with N 8. Chen, Z. et al. Highly active and durable core–corona structured bifunctional
and P (NPMCs), simply by pyrolysing polyaniline aerogels obtained catalyst for rechargeable metal–air battery application. Nano Lett. 12,
from a template-free polymerization of aniline in the presence of 1946–1952 (2012).
9. Morales, L. & Fernandez, A. M. Unsupported PtxRuyIrz and PtxIry as
phytic acid. The resultant NPMCs show efficient catalytic activities bi-functional catalyst for oxygen reduction and oxygen evolution reactions in
for both ORR and OER as bifunctional air electrodes in primary and acid media, for unitized regenerative fuel cell. Int. J. Electrochem. Sci. 8,
rechargeable Zn–air batteries. Typically, our primary battery based 12692–12706 (2013).

NATURE NANOTECHNOLOGY | VOL 10 | MAY 2015 | www.nature.com/naturenanotechnology 451

© 2015 Macmillan Publishers Limited. All rights reserved

ARTICLES NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2015.48

10. Gong, K., Du, F., Xia, Z., Durstock, M. & Dai, L. Nitrogen-doped carbon 33. Cui, X. et al. Microstructure and corrosion resistance of phytic acid conversion
nanotube arrays with high electrocatalytic activity for oxygen reduction. coatings for magnesium alloy. Appl. Surf. Sci. 255, 2098–2103 (2008).
Science 323, 760–764 (2009). 34. Gorham, J., Torres, J., Wolfe, G., d’Agostino, A. & Fairbrother, D. H. Surface
11. Liu, R., Wu, D., Feng, X. & Müllen, K. Nitrogen-doped ordered mesoporous reactions of molecular and atomic oxygen with carbon phosphide films.
graphitic arrays with high electrocatalytic activity for oxygen reduction. J. Phys. Chem. B 109, 20379–20386 (2005).
Angew. Chem. Int. Ed. 49, 2565–2569 (2010). 35. Puziy, A. M., Poddubnaya, O. I., Socha, R. P., Gurgul, J. & Wisniewski, M.
12. Wang, S. et al. BCN graphene as efficient metal-free electrocatalyst for the XPS and NMR studies of phosphoric acid activated carbons. Carbon
oxygen reduction reaction. Angew. Chem. Int. Ed. 51, 4209–4212 (2012). 46, 2113–2123 (2008).
13. Zheng, Y., Jiao, Y., Ge, L., Jaroniec, M. & Qiao, S. Z. Two-step boron and 36. Yang, S., Feng, X., Wang, X. & Müllen, K. Graphene-based carbon nitride
nitrogen doping in graphene for enhanced synergistic catalysis. Angew. Chem. nanosheets as efficient metal-free electrocatalysts for oxygen reduction reactions.
Int. Ed. 52, 3110–3116 (2013). Angew. Chem. Int. Ed. 50, 5339–5343 (2011).
14. Wang, S. et al. Vertically aligned BCN nanotubes as efficient metal-free 37. Zhao, Y., Watanabe, K. & Hashimoto, K. Self-supporting oxygen reduction
electrocatalysts for the oxygen reduction reaction: a synergetic effect by electrocatalysts made from a nitrogen-rich network polymer. J. Am. Chem. Soc.
co-doping with boron and nitrogen. Angew. Chem. Int. Ed. 50, 134, 19528–19531 (2012).
11756–11760 (2011). 38. Xiang, Z. et al. Highly efficient electrocatalysts for oxygen reduction based on
15. Liang, J., Jiao, Y., Jaroniec, M. & Qiao, S. Z. Sulfur and nitrogen dual-doped 2D covalent organic polymers complexed with non-precious metals. Angew.
mesoporous graphene electrocatalyst for oxygen reduction with synergistically Chem. Int. Ed. 53, 2433–2437 (2014).
enhanced performance. Angew. Chem. Int. Ed. 51, 11496–11500 (2012). 39. Li, M., Zhang, L., Xu, Q., Niu, J. & Xia, Z. N-doped graphene as catalysts for
16. Xue, Y. et al. Three-dimensional B,N-doped graphene foam as a metal-free oxygen reduction and oxygen evolution reactions: theoretical considerations.
catalyst for oxygen reduction reaction. Phys. Chem. Chem. Phys. 15, J. Catal. 314, 66–72 (2014).
12220–12226 (2013). 40. Sha, Y., Yu, T. H., Liu, Y., Merinov, B. V. & Goddard, W. A. Theoretical study of
17. Jiao, Y., Zheng, Y., Jaroniec, M. & Qiao, S. Z. Origin of the electrocatalytic oxygen solvent effects on the platinum-catalyzed oxygen reduction reaction. J. Phys.
reduction activity of graphene-based catalysts: a roadmap to achieve the Chem. Lett. 1, 856–861 (2010).
best performance. J. Am. Chem. Soc. 136, 4394–4403 (2014). 41. Yu, L., Pan, X., Cao, X., Hu, P. & Bao, X. Oxygen reduction reaction
18. Grimaud, A. et al. Double perovskites as a family of highly active catalysts for mechanism on nitrogen-doped graphene: a density functional theory study.
oxygen evolution in alkaline solution. Nature Commun. 4, 2439 (2013). J. Catal. 282, 183–190 (2011).
19. Chen, S., Duan, J., Jaroniec, M. & Qiao, S. Z. Three-dimensional N-doped 42. Man, I. C. et al. Universality in oxygen evolution electrocatalysis on oxide
graphene hydrogel/NiCo double hydroxide electrocatalysts for highly efficient surfaces. ChemCatChem 3, 1159–1165 (2011).
oxygen evolution. Angew. Chem. Int. Ed. 52, 13567–13570 (2013). 43. Du, G. et al. Co3O4 nanoparticle-modified MnO2 nanotube bifunctional
20. Zhao, Y., Nakamura, R., Kamiya, K., Nakanishi, S. & Hashimoto, K. Nitrogen- oxygen cathode catalysts for rechargeable zinc–air batteries. Nanoscale
doped carbon nanomaterials as non-metal electrocatalysts for water oxidation. 5, 4657–4661 (2013).
Nature Commun. 4, 2390 (2013). 44. Toussaint, G., Stevens, P., Akrour, L., Rouget, R. & Fourgeot, F. Development
21. Tian, J., Liu, Q., Asiri, A. M., Alamry, K. A. & Sun, X. Ultrathin graphitic C3N4 of a rechargeable zinc–air battery. ECS Trans. 28, 25–34 (2010).
nanosheets/graphene composites: efficient organic electrocatalyst for oxygen 45. Lee, D. U., Choi, J-Y., Feng, K., Park, H. W. & Chen, Z. Advanced extremely
evolution reaction. ChemSusChem 7, 2125–2130 (2014). durable 3D bifunctional air electrodes for rechargeable zinc–air batteries.
22. Ng, J. W. D., Tang, M. & Jaramillo, T. F. A carbon-free, precious-metal-free, Adv. Energy Mater. 4, 1301089 (2014).
high-performance O2 electrode for regenerative fuel cells and metal–air batteries. 46. Chen, S., Duan, J., Ran, J., Jaroniec, M. & Qiao, S. Z. N-doped graphene film-
Energy Environ. Sci. 7, 2017–2024 (2014). confined nickel nanoparticles as a highly efficient three-dimensional oxygen
23. Chen, S., Duan, J., Jaroniec, M. & Qiao, S-Z. Nitrogen and oxygen dual-doped evolution electrocatalyst. Energy Environ. Sci. 6, 3693–3699 (2013).
carbon hydrogel film as a substrate-free electrode for highly efficient 47. Nørskov, J. K. et al. Origin of the overpotential for oxygen reduction at a
oxygen evolution reaction. Adv. Mater. 26, 2925–2930 (2014). fuel-cell cathode. J. Phys. Chem. B 108, 17886–17892 (2004).
24. Ma, T. Y., Dai, S., Jaroniec, M. & Qiao, S. Z. Graphitic carbon ntride nanosheet– 48. Bao, X. et al. A first-principles study of the role of quaternary-N doping on the
carbon nanotube three-dimensional porous composites as high-performance oxygen reduction reaction activity and selectivity of graphene edge sites.
oxygen evolution electrocatalysts. Angew. Chem. Int. Ed. 53, 7281–7285 (2014). Top. Catal. 56, 1623–1633 (2013).
25. Prabu, M., Ramakrishnan, P. & Shanmugam, S. CoMn2O4 nanoparticles 49. Zhang, L. & Xia, Z. Mechanisms of oxygen reduction reaction on nitrogen-
anchored on nitrogen-doped graphene nanosheets as bifunctional electrocatalyst doped graphene for fuel cells. J. Phys. Chem. C 115, 11170–11176 (2011).
for rechargeable zinc–air battery. Electrochem. Commun. 41, 59–63 (2014).
26. Sun, C. et al. Graphene–Co3O4 nanocomposite as an efficient bifunctional Acknowledgements
catalyst for lithium–air batteries. J. Mater. Chem. A 2, 7188–7196 (2014). This work is supported financially by the Air Force Office of Scientific Research (AFOSR;
27. Tian, G-L. et al. Nitrogen-doped graphene/carbon nanotube hybrids: in situ FA-9550-12-1-0069 and FA9550-12-1-0037) and the National Science Foundation (NSF-
formation on bifunctional catalysts and their superior electrocatalytic activity AIR-IIP-1343270 and NSF-CMMI-1363123). The authors thank Z. Xiang, H. Pan, Y. Xue,
for oxygen evolution/reduction reaction. Small 10, 2251–2259 (2014). Y. Yu and L. Zhang for discussions and help.
28. Pan, L. et al. Hierarchical nanostructured conducting polymer hydrogel with
high electrochemical activity. Proc. Natl Acad. Sci. USA 109, 9287–9292 (2012). Author contributions
29. Yang, D-S., Bhattacharjya, D., Inamdar, S., Park, J. & Yu, J-S. Phosphorus-doped J.Z. and L.D. conceived and designed the experiments. J.Z. performed sample
ordered mesoporous carbons with different lengths as efficient metal-free characterization and electrochemical studies. Z.Z. and Z.X. performed the theoretical
electrocatalysts for oxygen reduction reaction in alkaline media. J. Am. Chem. calculation. J.Z. and L.D. analysed experimental data and wrote the paper. All authors
Soc. 134, 16127–16130 (2012). discussed the results and commented on the manuscript.
30. Zhang, J., Jiang, J., Li, H. & Zhao, X. S. A high-performance asymmetric
supercapacitor fabricated with graphene-based electrodes. Energy Environ. Sci. Additional information
4, 4009–4015 (2011). Supplementary information is available in the online version of the paper. Reprints and
31. Ding, W. et al. Space-confinement-induced synthesis of pyridinic- and pyrrolic- permission information is available online at www.nature.com/reprints. Correspondence and
nitrogen-doped graphene for the catalysis of oxygen reduction. Angew. Chem. requests for materials should be addressed to L.D.
Int. Ed. 52, 11755–11759 (2013).
32. Su, F. et al. Nitrogen-containing microporous carbon nanospheres with Competing financial interests
improved capacitive properties. Energy Environ. Sci. 4, 717–724 (2011). The authors declare no competing financial interests.

452 NATURE NANOTECHNOLOGY | VOL 10 | MAY 2015 | www.nature.com/naturenanotechnology

© 2015 Macmillan Publishers Limited. All rights reserved