www.afm-journal.

de
www.MaterialsViews.com

Bifunctional Nickel Phosphide Nanocatalysts Supported on

FULL PAPER
Carbon Fiber Paper for Highly Efficient and Stable Overall
Water Splitting
Xiaoguang Wang, Wei Li, Dehua Xiong, Dmitri Y. Petrovykh, and Lifeng Liu*

Despite the above mentioned advan-

Self-supported electrodes comprising carbon fiber paper (CP) integrated with tages, water electrolyzers have so far not
bifunctional nickel phosphide (Ni-P) electrocatalysts are fabricated by elec- been widely deployed primarily because
trodeposition of Ni on functionalized CP, followed by a convenient one-step the H2 produced from electrolysis is still
not economically competitive.[5] One of
phosphorization treatment in phosphorus vapor at 500 °C. The as-fabricated
the major approaches to reducing produc-
CP@Ni-P electrode exhibits excellent electrocatalytic performance toward tion costs is to develop highly efficient,
hydrogen evolution in both acidic and alkaline solutions, with only small over- low-cost, and durable electrocatalysts that
potentials of 162 and 250 mV, respectively, attaining a cathodic current density can substantially expedite the kinetics of
of 100 mA cm−2. Furthermore, the CP@Ni-P electrode also exhibits superior hydrogen evolution and oxygen evolu-
catalytic performance toward oxygen evolution reaction (OER). An excep- tion reactions (HER and OER). Presently,
platinum (Pt) and noble metal oxides,
tionally high OER current of 50.4 mA cm−2 is achieved at an overpotential of
including ruthenium oxide (RuO2) and
0.3 V in 1.0 M KOH. The electrode can sustain 10 mA cm−2 for 180 h with only iridium oxide (IrO2), are the state-of-the-
negligible degradation, showing outstanding durability. Detailed microstruc- art electrocatalysts to drive HER and OER,
tural and compositional studies reveal that upon OER in alkaline solution the respectively. Pt, Ru, and Ir are not only
surface Ni-P is transformed to NiO covered with a thin Ni(OH)x layer, forming expensive, but also have limited avail-
ability in the earth’s crust. Therefore, it
a Ni-P/NiO/Ni(OH)x heterojunction, which presumably enhances the elec-
is not practical to deploy electrolyzers
trocatalytic performance for OER. Given the well-defined bifunctionality, a full containing these rare metals on a large
alkaline electrolyzer is constructed using two identical CP@Ni-P electrodes as scale, even if the cost were affordable. In
cathode and anode, respectively, which can realize overall water splitting with this context, research efforts have recently
efficiency as high as 91.0% at 10 mA cm−2 for 100 h. been focused on searching for low-cost,
earth-abundant catalysts as alternatives
to Pt and RuO2/IrO2, for example, transi-
tion metal sulfides,[6–11] selenides,[12–14]
1. Introduction carbides,[15–17] borides,[17,18] nitrides,[19–21] phosphides,[22–38] and
nanocomposites[39,40] for the HER, and transition metals,[41,42]
Hydrogen (H2) is a promising energy carrier and has many oxides,[43–45] hydroxides,[46–49] phosphates,[50–52] and polymeric
advantages over hydrocarbons (e.g., high energy density, no carbon nitride[53,54] for the OER. Compared to the HER, the
harmful products when burnt, and natural abundance) for use OER is more difficult to accomplish as it is a thermodynami-
as a fuel for both stationary power and transportation.[1–3] Water cally and kinetically demanding process involving four sequen-
electrolysis has been proposed to be the cleanest way to produce tial proton-coupled electron transfer steps and the oxygen–
H2, because unlike steam reforming of hydrocarbons this pro- oxygen bonding formation. Ni- and NiFe-based compounds
cess does not emit any pollutants and water is an inexpensive have long been known as active OER catalysts and are pres-
renewable resource. More importantly, H2 produced from water ently being used in commercial alkaline electrolyzers.[41,42] In
electrolysis has high purity and is free of carbon monoxide,[4] recent years, considerable attention has been paid to develop
and therefore can be directly used to feed fuel cells without a advanced NiFe-based nanostructures for improved OER activity
risk of poisoning the anode catalysts. and stability,[46–48] and an OER activity of 20 mA cm−2 has been
achieved with an overpotential as small as 0.3 V.[47,48]
Notwithstanding significant progress, there are very few cat-
Dr. X. G. Wang, Dr. W. Li, Dr. D. H. Xiong, alysts that are able to catalyze both HER and OER in the same
Dr. D. Y. Petrovykh, Dr. L. F. Liu electrolyte.[14,37,48,51,53,55,56] Previously, Pt-free bifunctional elec-
International Iberian Nanotechnology Laboratory (INL) trocatalysts that are simultaneously active for two electrochem-
Av. Mestre Jose Veiga ical reactions such as oxygen reduction reaction and OER have
4715-330 Braga, Portugal
been reported.[57–59] However, the half reactions that they cata-
E-mail: [email protected]
lyze usually take place in two different electrochemical devices.
DOI: 10.1002/adfm.201505509 Developing a bifunctional catalyst that is capable of catalyzing

Adv. Funct. Mater. 2016, 26, 4067–4077 © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim wileyonlinelibrary.com 4067
 www.afm-journal.de
www.MaterialsViews.com
FULL PAPER

both HER and OER, two half reactions occurring in the same electrodeposition duration, suggesting an increase in the thick-
electrochemical device (i.e., electrolyzers), is very attractive, ness of the deposited Ni layer.
because this can potentially simplify the fabrication procedure The as-obtained CP@Ni samples were then phosphorized via
of electrolyzers and substantially lower the production costs. a convenient one-step approach we developed recently.[27] They
Nickel phosphide (Ni-P), since initial experimental reports in were exposed to phosphorus vapor carried by high-purity N2 at
2013,[22] has been intensively investigated as a new earth-abun- 500 °C for 1 h. After phosphorization treatment, the sample
dant electrocatalyst that can potentially replace Pt to catalyze color turned to black from silver of the CP@Ni (Figure S3,
HER.[22–28,38] Very recently, Ni2P and Ni5P4 were also demon- Supporting Information). Microstructurally, phosphorization
strated to be active toward OER.[37,56] In this paper, we report results in the formation of a highly dense, vertically aligned
a very convenient and scalable approach to the fabrication of nanosheet (NS) array that uniformly covers the entire surface
self-supported 3D carbon fiber paper (CP) electrode integrated of carbon microfibers (Figure 1c; Figure S4, Supporting Infor-
with vertically aligned Ni-P nanosheets (NSs) as bifunctional mation). The NSs are several micrometers high and <50 nm
catalysts for HER and OER. This is realized by electrodeposi- thick. The NS morphology prevails in all the samples with dif-
tion of Ni on the surface of carbon microfibers, followed by a ferent Ni loadings, however, in CP@Ni(210 min)-P the density
simple one-step phosphorization process in phosphorus vapor. of NSs is lower than that on other samples. We note that both
We demonstrate that without any posttreatment, the self-sup- the Ni electrodeposition and phosphorization processes are
ported 3D CP@Ni-P NS electrode can be directly used as either highly reproducible. For a given Ni deposition duration X, at
a cathode to drive HER or an anode to expedite OER, exhib- least four samples were prepared. They showed no difference
iting excellent catalytic activity and outstanding long-term dura- in morphology under scanning electron microscopy (SEM)
bility. Furthermore, we develop an alkaline electrolyzer using examination. Moreover, the mass loadings of Ni and Ni-P were
two identical CP@Ni-P NS electrodes as cathode and anode, averaged, and only a slight variation was observed (Figure S6,
respectively. We demonstrate that this noble-metal-free electro- Supporting Information).
lyzer can work with an energy efficiency as high as 91.0% at Phase composition of the fabricated CP@Ni-P electrodes was
10 mA cm−2 for at least 100 h without discernible performance examined by XRD. As shown in Figure 1a and Figure S5 (Sup-
degradation. porting Information), the CP@Ni-P consists of Ni5P4 as the
major phase (>80%, International Centre for Diffraction Data
(ICDD) No. 04-014-7901), and Ni2P (ICDD No. 04-003-1863)
2. Results and Discussion and NiP2 (ICDD No. 04-003-2351) as the minor phases. SEM
and energy dispersive X-ray spectroscopy (EDX) analyses show
2.1. Fabrication and Structure of the CP@Ni-P Electrode that the CP@Ni-P sample with less Ni loading (i.e., subjected to
shorter Ni electrodeposition) has a higher P/Ni ratio (Figure S4
CP was chosen as the current collector because it has high and Table S1, Supporting Information). This is also confirmed
electrical conductivity and excellent resistance to corrosion in by quantitative XRD analyses where P-rich phases (i.e., Ni5P4
both acidic and alkaline solutions. The as-received CP con- and NiP2) account for a higher percentage in CP@Ni(30 min)-P
sists of interconnected microfibers with a diameter of ≈10 µm than those in CP@Ni(210 min)-P (94% vs 82%, Figure S5, Sup-
(Figure S1, Supporting Information). Prior to electrodeposition porting Information). Given that phosphorization of the CP@
of Ni, CP was first treated in a mixed solution of sulfuric acid Ni takes place radially starting from the Ni surface toward the
and nitric acid to introduce surface defects and oxyl-groups on C core, the difference in the P content and the percentage of
carbon microfibers as essential sites for the nucleation and P-rich phases among the samples likely arises from the reac-
conformal growth of Ni electrodeposits. The electrodeposition tion kinetics of the phosphorization, taking into account that
of Ni was carried out in a Watts-type electrolyte at a constant CP@Ni(210 min)-P has a much thicker Ni shell (Table S1,
current of 20 mA cm−2 for varying durations of 30, 60, 120, and Supporting Information) and it would need longer time and a
210 min, respectively (as detailed in the Experimental Section). higher concentration of P vapor to form a P-rich phase.
The obtained samples are hereafter denoted as CP@Ni(X), In order to further study the microstructure of the fabri-
where X = electrodeposition duration. After surface functionali- cated Ni-P NSs, we carried out detailed transmission elec-
zation, the reduction of Ni2+ can be readily accomplished on the tron microscopy (TEM) investigations. Figure 1d shows a
surface of carbon microfibers (Figure 1b; Figure S2, Supporting low-magnification TEM image of an individual Ni-P NS from
Information), and the loading of metallic Ni increases with the the CP@Ni(60 min)-P, in which the thin sheet-like structure
increasing electrodeposition time. When the deposition is per- can be clearly distinguished. Extensive selected-area electron
formed for 30 min, carbon surface without Ni deposits is still diffraction (SAED) studies over 20 individual NSs revealed
visible (Figure S2a, Supporting Information). Further increase that most NSs are composed of single-crystalline Ni5P4, and
in the electrodeposition duration leads to a full coverage over only few comprise single-crystalline NiP2. A representative
the entire surface of carbon microfibers. Correspondingly, the SAED pattern taken from an individual Ni5P4 NS is presented
diameter of these fibers is also enlarged (Figure S2b–d and in Figure 1e, showing a well-defined spotted pattern corre-
Table S1, Supporting Information). The increase in Ni loading sponding to the diffraction along the [0001] zone axis of the
was also evidenced by X-ray diffractometry (XRD) examina- Ni5P4 hexagonal structure. High-resolution TEM (HRTEM)
tions. In the XRD patterns of the CP@Ni shown in Figure S2e examination confirms the single-crystalline nature of the NS,
(Supporting Information), the intensity of the C (002) diffrac- and the measured interplanar spacing and angle correspond
tion peak is found to gradually decrease with the increasing well to those of (1100) and (2110) crystal planes of hexagonal

4068 wileyonlinelibrary.com © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Funct. Mater. 2016, 26, 4067–4077
 www.afm-journal.de
www.MaterialsViews.com

FULL PAPER
Figure 1. a) XRD patterns of CP, CP@Ni, and CP@Ni-P electrodes. The standard powder XRD patterns of Ni5P4, Ni2P, and NiP2 are given for refer-
ence. b1) Low- and b2) high-magnification SEM images of CP@Ni. c1) Low- and c2) high-magnification SEM images of CP@Ni-P. d) TEM image of
an individual Ni5P4 NS. e) SAED pattern and f) HRTEM image taken from a single Ni5P4 NS. g–i) HAADF-STEM image and EDX elemental maps of Ni
and P taken from an individual Ni5P4 NS. The CP@Ni(60 min)-P was used for XRD, SEM, and TEM characterization. Scale bars: 10 nm.

Ni5P4. To further investigate the chemical composition of the that the loading mass of Ni-P on CP@Ni(60 min)-P is only
NSs, high-angle annular dark field scanning transmission elec- ≈50% of that of CP@Ni(120 min)-P (Figure S6, Supporting
tron microscopy (HAADF-STEM) was performed. Both Ni and Information), CP@Ni(60 min)-P also possesses much higher
P were found to uniformly distribute over the NS without dis- mass-normalized activity, and was identified as the most active
cernible segregation. self-supported 3D electrode. Therefore, all the subsequent
electrocatalytic tests presented below were made using the
CP@Ni(60 min)-P electrode, unless otherwise specified. For
2.2. Electrocatalytic Performance toward HER simplicity, CP@Ni-P is used in the following to stand for the
CP@Ni(60 min)-P electrode.
To identify the most active CP@Ni-P electrodes, we tested the Electrocatalytic performance of the CP@Ni-P toward
electrocatalytic activity of the electrodes with different Ni-P HER was first evaluated in 0.5 M H2SO4. For comparison,
loadings for HER in 0.5 M H2SO4 solution. When normalized the HER activity of a bare CP and commercially available
to the geometric surface area of the electrodes (i.e., 1 cm2), Pt/C catalysts was also measured. Figure 2a shows the iR-
CP@Ni(60 min)-P and CP@Ni(120 min)-P exhibit much corrected polarization curves of the bare CP, Pt/C, and CP@
better catalytic activity than do CP@Ni(30 min)-P and CP@ Ni-P measured at a scan rate of 5 mV s−1. As expected, Pt/C
Ni(210 min)-P, with an HER onset potential (ηonset, defined exhibits the highest activity for HER, while the bare CP only
as the potential at which the current density is −1 mA cm−2) shows negligible catalytic current even at a high overpoten-
of −18 and −47 mV, and a small overpotential (η100) of 162 tial (e.g., 300 mV). In contrast, the CP@Ni-P electrode has a
and 164 mV, respectively, to reach a cathodic current density ηonset as low as −18 mV. It only needs small overpotentials of
of 100 mA cm−2 (Figure S8, Supporting Information). Given 98 (η10), 116 (η20), and 162 mV (η100) to drive HER at current

Adv. Funct. Mater. 2016, 26, 4067–4077 © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim wileyonlinelibrary.com 4069
 www.afm-journal.de
www.MaterialsViews.com
FULL PAPER

Figure 2. Electrocatalytic performance of CP@Ni-P for HER in N2-saturated 0.5 M H2SO4. a) Polarization curves of a bare CP, Pt/C, and CP@Ni-P
recorded at a scan rate of 5 mV s−1. b) Tafel plots of Pt/C and CP@Ni-P. c) Polarization curves of the CP@Ni-P electrode recorded before and after
1000 continuous CV cycles in the range of −200 to 60 mV versus RHE at 50 mV s−1. d) Chronopotentiometric curve recorded at −10 mA cm−2.

densities of 10, 20, and 100 mA cm−2, respectively. The HER cathodic current density of 10 mA cm−2 for ≈160 h (Figure 2d),
activity of CP@Ni-P favorably compares to that of most Pt-free during which the overpotential was only increased from 98 to
HER catalysts tested in acidic solutions (Table S2, Supporting 124 mV (≈0.16 mV h−1), demonstrating excellent long-term
Information). Figure 2b presents the Tafel plots, η versus log(j), durability of the electrode in acidic medium.
for Pt/C and CP@Ni-P electrodes. The Tafel slope of 31.6 mV Likewise, the CP@Ni-P electrode also shows outstanding cat-
dec−1 for Pt/C is consistent with the value reported in the lit- alytic performance for HER in alkaline medium. The ηonset for
erature.[22,25] The CP@Ni-P exhibits a Tafel slope of 58.8 mV HER of the CP@Ni-P is only −55 mV in 1.0 M KOH solution
dec−1, larger than that of Pt/C, but superior to that of many (Figure 3a). Further increase in overpotential leads to a rapid rise
earth-abundant HER catalysts reported recently, such as Ni-P in cathodic current, indicative of a high catalytic performance
nanoparticles (87 mV dec−1 at η = 60–140 mV,[23] 81 mV dec−1 at for HER. To attain cathodic current densities of 10, 20, and
η = 150–200 mV[22]), bulk MoS2 (94 mV dec−1),[8] Mo2C nanocrys- 100 mA cm−2, the CP@Ni-P electrode needs overpotentials of
tals (110–235 mV dec−1),[16] and MoSe2/rGO (101 mV dec−1).[13] 117, 150, and 250 mV, respectively, outperforming those of many
The slope falls within the range of 38–116 mV dec−1, sug- nonprecious earth-abundant HER catalysts tested in alkaline
gesting that the HER taking place on the CP@Ni-P surface medium (Table S3, Supporting Information). A Tafel slope of
would follow a Volmer–Heyrovsky mechanism, and that the 85.4 mV dec−1 is obtained for the CP@Ni-P electrode (Figure 3b),
rate of the discharge step would be consistent with that of the which is larger than that measured in 0.5 M H2SO4 solution.
desorption step.[24] By extrapolating the Tafel plot to an overpo- This agrees well with previous literature reports on Ni-P cata-
tential of 0 V, the exchange current density (j0) for HER on the lysts,[28] and might be associated with the drastically decreased
CP@Ni-P is obtained, being 0.24 mA cm−2, outperforming that proton concentration in alkaline medium. Upon ADT for 1000
of many transition metal sulfides and phosphides reported in continuous cycles, the CP@Ni-P electrode only shows a slight
the literature, such as MoS2,[7,8] MoP,[34] and CoP.[36] An accel- decay with an overpotential increase by 38 mV for achieving
erated degradation test (ADT) was made by performing cyclic −100 mA cm−2 (Figure 3c). The long-term durability test revealed
voltammetry (CV) at 50 mV s−1 in the potential range of −200 to that the overpotential required to deliver −10 mA cm−2 was
60 mV versus reversible hydrogen electrode (RHE) to evaluate increased from 118 to 146 mV in the first few minutes, likely
the stability of the CP@Ni-P electrode. As shown in Figure 2c, due to the rebalancing of proton concentration nearby the elec-
the polarization curve recorded after 1000 continuous CV cycles trode. The overpotential then gradually went up to 215 mV after
slightly deviates from the initial one, and the overpotential 96 h with an average rate of 0.72 mV h−1 (Figure 3d). Overall, the
required to achieve current densities of 20 and 100 mA cm−2 electrocatalytic performance of CP@Ni-P toward HER is inferior
merely increases by 9 and 10 mV, respectively, indicative of in alkaline solution than in acidic solution, in agreement with
good stability. The durability test was then performed at a fixed previous reports on Ni2P microparticles.[28]

4070 wileyonlinelibrary.com © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Funct. Mater. 2016, 26, 4067–4077
 www.afm-journal.de
www.MaterialsViews.com

FULL PAPER
Figure 3. Electrocatalytic performance of CP@Ni-P electrodes for HER in N2-saturated 1.0 M KOH. a) Polarization curves of a bare CP, Pt/C, and
CP@Ni-P recorded at a scan rate of 5 mV s−1. b) Tafel plots of Pt/C and CP@Ni-P. c) Polarization curves of the CP@Ni-P electrode recorded before
and after 1000 continuous CV cycles in the range of −200 to 60 mV versus RHE at 50 mV s−1. d) Chronopotentiometric curve recorded at −10 mA cm−2.

It is worth mentioning that after the extended stability and electron transfer steps and the oxygen–oxygen bonding forma-
long-term durability tests (i.e., ≈160 h in 0.5 H2SO4 or 96 h in tion. Therefore, OER is deemed as a major bottleneck for the
1.0 M KOH) under HER conditions, the morphology of the CP@ reaction rate of water electrolysis. While transition metal phos-
Ni-P electrodes remains unchanged, namely, the densely packed phides have been intensively studied in the past three years
Ni-P NSs still cover the whole surface of carbon microfibers as for use to catalyze HER, their usage as OER catalysts is rarely
they did after the phosphorization (Figure S9, Supporting Infor- reported.[37,55,56] We have investigated the OER performance of
mation). TEM investigations confirm that the single-crystalline the CP@Ni-P electrode in 1.0 M KOH solution, and compared
nature of the NSs is preserved, and Ni and P are evenly distrib- it with that of a bare CP and the state-of-the-art RuO2 nanop-
uted over the NSs. Nevertheless, a thin layer was observed on owders. As shown from the polarization curves in Figure 4a,
the surface of the NSs (Figure S10, Supporting Information). the bare CP is inactive for OER up to 1.67 V, and the RuO2/Ti
X-ray photoelectron spectroscopy (XPS) analysis (detailed in electrode shows obvious anodic OER current at ≈1.45 V, con-
the Supporting Information) reveals that after HER, the sur- sistent with the previous report.[52] In comparison, the CP@
face chemistry of CP@Ni-P electrodes is different from that of Ni-P electrode exhibits remarkably improved catalytic activity
the as-fabricated CP@Ni-P: the P signal becomes significantly for OER with an anodic current density of 20 mA cm−2 at an
lower and the surface layer (3–5 nm) is dominated by Ni(OH)x overpotential as small as 0.18 V. Experimentally, we observed
(Figure S7, Supporting Information). We further examined the that O2 gas bubbles started to appear on the surface of the CP@
tested samples using XRD and performed quantitative analysis Ni-P at 1.45 V, and the O2 evolution became intensive as the
of the phase composition. The results show that after extended overpotential increased. Correspondingly, the anodic current
testing, the CP@Ni-P electrode is still composed of Ni5P4 as the density rapidly went up, reaching 50.4 mA cm−2 at an overpo-
major phase and Ni2P and NiP2 as the minor phases. Moreover, tential of 0.3 V. The OER performance of the CP@Ni-P elec-
the content of each phase varies only slightly (Figure S11, Sup- trode is found to outperform that of many nonprecious OER
porting Information), indicating excellent stability of the CP@ catalysts reported so far, including CoPi,[50] layered double
Ni-P in both acidic and alkaline media under HER conditions. hydroxides,[46–49] and recently reported transition metal phos-
No diffraction peaks from Ni(OH)x were detected, suggesting phides (Table S4, Supporting Information).[52,55] The Tafel slope
that the thin surface Ni(OH)x layer is amorphous. of the CP@Ni-P electrode is 73 mV dec−1 in the overpotential
range of 0.1–0.2 V, slightly higher than that of the RuO2/Ti
2.3. Electrocatalytic Performance toward OER electrode (i.e., 70 mV dec−1, Figure 4b). The Tafel slope in the
overpotential range of 0–0.1 V is much larger (242 mV dec−1),
The OER is a thermodynamically and kinetically demanding which might correspond to the slow adsorption of oxygenated
process because it involves four sequential proton-coupled species. The CP@Ni-P electrode exhibits very good stability

Adv. Funct. Mater. 2016, 26, 4067–4077 © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim wileyonlinelibrary.com 4071
 www.afm-journal.de
www.MaterialsViews.com
FULL PAPER

Figure 4. Electrocatalytic performance of CP@Ni-P for OER in 1.0 M KOH. a) Polarization curves of a bare CP, RuO2, and CP@Ni-P recorded at a scan
rate of 5 mV s−1. b) Tafel plots of RuO2 and CP@Ni-P. c) Polarization curves of the CP@Ni-P recorded before and after 700 continuous CV cycles in
the range of 1.1–1.5 V versus RHE at 50 mV s−1. d) Chronopotentiometric curve of the CP@Ni-P electrode recorded at 10 mA cm−2. Inset: zoomed
view of the curve in the beginning 0.4 h.

upon continuous cyclic voltammetric scans between 1.1 and is not observed on the polarization curve (Figure 4a),[37,43]
1.5 V at 50 mV s−1. As shown in Figure 4c, the polarization extensive morphological, microstructural, and compositional
curve only slightly drifts after extended scanning of 700 cycles. analyses indicate that structural and compositional changes
Moreover, the anodic current achieved at the overpotential of indeed took place in the CP@Ni-P electrode during the OER
0.3 V does not decrease but increases to 53.8 mA cm−2. This catalysis. Figure S12 (Supporting Information) shows the
could be attributed to the activation resulting from the struc- morphological evolution of the CP@Ni-P electrode over time
tural and compositional transformation of the Ni-P catalysts, as during the chronopotentiometric test at 10 mA cm−2 (Sup-
will be discussed in detail below. After the stability test, we con- porting Information). After the OER for 20 min, the sheet-like
tinued to examine the durability of the CP@Ni-P electrode at morphology is still retained except that the NSs become thicker
a fixed current density of 10 mA cm−2. As shown in Figure 4d, (Figure S12a, Supporting Information). The thickness of the
the initial potential of the electrode is ≈1.35 V, in good agree- NSs keeps increasing as the electrolysis proceeds. As a result,
ment with that observed in the polarization curve (Figure 4a). the NSs stick to each other evolving into a seemingly compact
As the electrolysis proceeds, the electrode potential gradually film with some cracks on its surface (Figure S12b, Supporting
increases to ≈1.5 V and gets stabilized around that value for Information). Further prolonging the OER test duration (e.g.,
up to 180 h, showing superior long-term durability. The evo- 6 h, 24 h) results in a dusted electrode having many particles
lution of O2 gas bubbles was clearly observed throughout the of various sizes dispersed on the surface of carbon microfibers
whole chronopotentiometric test. The increase of the electrode (Figure S12c,d, Supporting Information). As will be discussed
potential in the beginning of the electrolysis likely stems from later, these particles may result from the structural transforma-
the increased electrode resistance due to the formation of more tion of Ni5P4 NSs into NiO under the OER conditions. After
resistive nickel oxide/hydroxide during the electrolysis. extended OER electrocatalysis for 180 h, a coarsened electrode
Very recently, Ni2P nanowires/nanoparticles have been inves- surface was observed (Figure S12e, Supporting Information).
tigated for use as OER catalysts.[37] It was reported that upon the Extensive TEM analyses of the surface NSs/particles provide
application of an anodic potential Ni2P was oxidized, generating detailed information about the local structural and composi-
an oxygenated surface layer over time (i.e., nickel oxide NiOx, tional transformation of the Ni-P catalysts (Figures S13–S15,
or nickel oxyhydroxide, NiOOH). This resulted in an Ni2P/NiOx Supporting Information). The results show: (i) P content drasti-
(or NiOOH) heterojunction that was believed to be the major cally decreases with the increasing OER test duration, indicating
contributor to the observed high OER performance.[37] Although that P leaches out from the Ni-P NSs very fast during OER. For
the oxidation peak characteristic of Ni2+ → Ni3+ transformation the CP@Ni-P electrode subjected to OER for 1 h, oxygen has

4072 wileyonlinelibrary.com © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Funct. Mater. 2016, 26, 4067–4077
 www.afm-journal.de
www.MaterialsViews.com

electrocatalytic activity for OER, as pointed out previously.[37]

FULL PAPER
already become dominant in the NSs (Figure S13, Supporting
Information). After OER for 24 h, only trace amount of P can To elucidate the role of Ni-P/NiO/Ni(OH)x heterojunctions, we
be detected in the NSs (Figure S14, Supporting Information). prepared a control sample where vertically aligned NiO NSs
XPS investigation on this sample also confirms that no P signal were grown directly on the carbon microfibers of CP using a
can be detected on the electrode surface, and the outermost hydrothermal method reported previously (denoted as CP@
surface (3–5 nm) is dominated by Ni(OH)x (Figure S7c, Sup- NiO).[60] These NiO NSs show morphology similar to that of
porting Information). (ii) The original flat and smooth Ni5P4 Ni-P NSs (Figure S18a, Supporting Information). According
NSs become nanoporous as the OER test proceeds, which can to previous reports,[42,61] a Ni(OH)x layer would form on the
be seen from the high-magnification TEM and HAADF-STEM surface of NiO under alkaline conditions. Therefore, a NiO/
images in Figures S13–S15 (Supporting Information). This Ni(OH)x interface analogous to that of Ni-P/NiO/Ni(OH)x
structural transformation may result from the P leach-out and could be expected under the OER conditions we used, justifying
was observed even in the sample subjected to OER for only the comparison of CP@NiO to CP@Ni-P. Our electrochemical
1 h. (iii) The HRTEM study showed that the produced nano- measurements reveal that the CP@NiO exhibits much poorer
porous structure is polycrystalline with a small crystallite size. OER performance than the CP@Ni-P electrode, only showing
Fast Fourier transformation electron diffraction reveals a ring- a small anodic current of 0.72 mA cm−2 at the overpotential
like pattern typical of a face-centered cubic structure. In con- of 0.3 V and requiring a large overpotential of 0.59 V to attain
junction with lattice distance calculation and the EDX results, 20 mA cm−2 (Figure S18b, Supporting Information). Although
it can be identified that these nanocrystallites are composed of it is as stable and durable as the CP@Ni-P electrode during the
NiO. This agrees with the previous report on Ni2P nanoparti- long-term chronopotentiometric test at 10 mA cm−2, its OER
cles/nanowire.[37] No diffractions from Ni(OH)x were observed, activity is much lower (Figure S18c, Supporting Information).
implying that the surface Ni(OH)x layer is amorphous. This control experiment clearly demonstrates that the forma-
We further examined the evolution of macroscopic phase tion of Ni-P/NiO/Ni(OH)x heterojunctions under OER condi-
composition of the CP@Ni-P electrode over time under OER tions and the underlying Ni-P have played a critical role in the
conditions at 10 mA cm−2. Our XRD results show that the achievement of the excellent OER performance of CP@Ni-P
P-rich phases (i.e., NiP2 and Ni5P4) rapidly diminish as the OER electrodes.
proceeds (Figure S16, Supporting Information). After OER for
24 h, only the Ni2P phase can be resolved. The Ni2P phase was
able to sustain extended OER electrolysis up to 180 h at the 2.4. Overall Electrochemical Water Splitting Using the CP@Ni-P
time a trace amount of Ni2O3 phase (ICDD No. 00-014-0481) Bifunctional Electrodes
started to appear. It is noted that no diffraction peaks from NiO
or Ni(OH)x were observed in the electrodes subjected to OER According to the above analyses, the self-supported CP@Ni-P
for 1, 6, and 24 h, though NiO crystallites were found to already can serve as both a cathode for HER and an anode for OER,
form locally in these samples according to our TEM investiga- exhibiting a bifunctional characteristic. This bifunctionality is
tions. We speculate that this is likely due, on the one hand, to shown more clearly in Figure 5a. In 1.0 M KOH solution, if a
the small crystallite size and poor crystallinity of the produced cathodic potential is applied (i.e., E < E0(H2/H2O)), the elec-
NiO; on the other hand, to the low percentage of NiO phase trode shows superior HER activity; while if an anodic poten-
compared to that of Ni2P in the composite, such that it cannot tial exceeding the thermodynamic threshold for O2 evolution
be resolved using XRD. (i.e., E > E0(O2/H2O)) is applied, the CP@Ni-P electrode will
Combining the SEM and TEM investigations with our XRD trigger an efficient OER activity. Based on this finding, we
analysis, we hypothesize that under OER electrolysis conditions have made an alkaline electrolyzer using the self-supported
(i.e., 10 mA cm−2 in 1.0 M KOH), the structural and composi- CP@Ni-P electrode paired with another identical one. Elec-
tional change first occurs on the surface of the NSs because trochemical tests in 1.0 M KOH solution reveal that this noble-
the NSs have much higher contact area with the alkaline solu- metal-free water electrolyzer starts to show catalytic current
tion (Figure S17, Supporting Information). This results in the at 1.50 V (Figure 5b). Experimentally, appreciable gas bubbles
leach-out of P, formation of nanoporosity, and transformation can be observed to evolve on the surface of cathode (i.e., H2) at
of Ni5P4 into NiO. Pulverization of the NSs also takes place in ≈1.55 V and anode (i.e., O2) at ≈1.68 V. The appearance of H2
this process, as evidenced by the formation of particles on the and O2 is not synchronized, because on the one hand, the evo-
electrode surface (Figure S12, Supporting Information). Con- lution of O2 involves multiple proton-coupled charge transfer
sequently, some active catalysts will lose physical contact with steps and therefore requires a higher potential to drive; on the
the electrode upon the evolution of O2 gas bubbles. Experimen- other hand, compared to the generated H2, more O2 initially
tally, we indeed observed the detachment of some fine powders generated is likely dissolved in the electrolyte given its relatively
from the electrode during the extended OER electrolysis (Movie large solubility in water (≈27 times higher than that of H2 at
S1, Supporting Information). Nevertheless, the durability of 20 °C).[62] Therefore, O2 bubbles could not be seen until the
the electrode was not affected (Figure 4d). We believe that on generated O2 accumulates to a certain amount, resulting in a
the one hand the increased active surface area arising from the lag in observing O2 gas bubbles.
formation of nanopores compensates the activity degradation The cell’s capability of splitting water was examined at var-
resulting from the loss of catalysts; on the other hand, an Ni-P/ ious current densities from 10 to 100 mA cm−2 (Figure 5c).
NiO/Ni(OH)x heterojunction may have formed on the surface Upon increasing the current density, the cell voltage increases
of the remaining Ni2P layer, which presumably enhances the accordingly but rapidly gets stabilized, showing outstanding

Adv. Funct. Mater. 2016, 26, 4067–4077 © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim wileyonlinelibrary.com 4073
 www.afm-journal.de
www.MaterialsViews.com
FULL PAPER

Figure 5. a) Polarization curve recorded in a three-electrode configuration in a wide potential window (−0.24 to 1.47 V versus RHE) showing the
bifunctionality of CP@Ni-P toward both HER and OER. b) Polarization curve recorded at a scan rate of 5 mV s−1 in a two-electrode configuration.
c) Chronopotentiometric curve of water electrolysis at several different current densities. d) Energy efficiency of the CP@Ni-P electrolyzer as a function
of current density. e) Long-term durability tests made at 10 and 20 mA cm−2. f) A digital photograph showing the evolution of H2 and O2 gas from the
electrodes at 20 mA cm−2. The electrolyte used in all tests was 1.0 M KOH aqueous solution.

stability. When the applied current density comes back to 20 of transition metal layered double hydroxides[47,48] and close to
from 100 mA cm−2, the cell voltage is restored only with a slight that of the state-of-the-art Pt/IrO2 electrolyzer measured at the
increase in overpotential, possibly resulting from the structural same current density.[52] The electrolysis efficiency reduces as
and compositional change in the anode. Nevertheless, the cell the current density increases (Figure 5d), however, an efficiency
voltage remains quite stable for ≈25 h. Volumes of the gener- of 69.0% is still retained even at 100 mA cm−2.
ated H2 and O2 gas at 10 mA cm−2 were quantitatively measured Long-term durability of the CP@Ni-P electrolyzer was tested
using a home-made graded H-cell, and compared with the the- in two-electrode configuration at 10 and 20 mA cm−2. Cell
oretical values calculated from the passed charges (Figure S19, voltages of 1.63 and 1.73 V are required to maintain 10 and
Supporting Information). The measured volumes match the 20 mA cm−2, respectively. As seen in Figure 5e, the constant
calculated ones well, and the volume ratio of H2 and O2 is very electrolysis can be sustained for up to 100 h with very little
close to 2:1, indicating a virtually 100% Faradaic efficiency for increase in overpotential. Phenomenally, H2 gas bubbles appear
both anode and cathode. The energy efficiency of the CP@Ni-P on the surface of cathode as soon as a current is applied;
electrolyzer is estimated by dividing the “thermoneutral poten- while O2 gas bubbles become visible on the anode surface in
tial” of water electrolysis, i.e., 1.48 V, by the cell’s voltage (see ≈5 min. Afterwards, both H2 and O2 evolve on the electrode
details in the Supporting Information).[63] An efficiency as high surface in a very stable manner throughout the durability test
as 91.0% is achieved at 10 mA cm−2, which is higher than that (Figure 5f; Movie S1, Supporting Information). We examined

4074 wileyonlinelibrary.com © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Funct. Mater. 2016, 26, 4067–4077
 www.afm-journal.de
www.MaterialsViews.com

FULL PAPER
the morphology and composition changes of both cathode and carried out in galvanostatic mode with a constant electroplating current
anode after the durability test (Figures S20 and S21, Supporting density of 20 mA cm−2 for varying durations of 30, 60, 120, and 210 min,
Information). We found that the morphology of the cathode did respectively, using a Keithley 2400 source meter as the power supply.
After deposition, the CP@Ni composite was thoroughly rinsed with
not change, as expected; while on the anode surface the array deionized (DI) water and absolute ethanol for several times, and then
of Ni-P NSs disappeared and was transformed into NiO parti- dried in air. The phosphorization treatment was subsequently performed
cles with various sizes, but the underlying Ni-P layer remained. using the one-step approach we recently developed.[26] Briefly, the CP@
This observation agrees well with that made for the tested elec- Ni was loaded in a ceramic boat, and ≈1 g of red phosphorus (P, Sigma-
trodes under HER and OER conditions, respectively (Figures S9 Aldrich) was placed 2 cm away from the CP@Ni at the upstream side.
and S12, Supporting Information). It is worth noting that we Afterwards, the boat was put into a tube furnace (Carbolite). The furnace
was purged with high-purity nitrogen (N2, 99.999%) at a flow rate of
have fabricated several alkaline electrolyzers using the CP@
800 SCCM for 30 min, heated to 500 °C at 5 °C min−1, and kept at this
Ni-P electrodes and observed similar water splitting perfor- temperature for 1 h. The furnace was then cooled down to 250 °C at
mance, illustrating high reproducibility toward water splitting 5 °C min−1 and maintained at this temperature for 3 h. (CAUTION:
using our CP@Ni-P electrolyzers. This step is very important to convert the residual toxic and pyrophoric
yellow P to inflammable red P. Without doing this, smoke or a flame
will be observed when opening the worktube.) Finally, the furnace was
naturally cooled down to room temperature. The worktube was flushed
3. Conclusions with N2 for 10 min before taking out the sample. The resultant CP@
Ni-P was washed sequentially with DI water, ethanol, and acetone, and
In summary, we have developed a self-supported 3D carbon then dried in an N2 flow. The weight of the sample was measured before
fiber paper electrode integrated with bifunctional Ni-P electro- electrodeposition of Ni and after phosphorization using a high-precision
catalysts. The electrode is fabricated by electrodeposition of Ni electronic balance (Sartorius CPA225D) in order to calculate the loading
on functionalized carbon fiber paper, followed by a convenient mass of the grown Ni-P. At least four samples prepared under the same
one-step phosphorization process in phosphorus vapor. The conditions were measured and their weights were averaged to minimize
the error. The average Ni-P mass loadings for the CP@Ni-P obtained
as-fabricated electrode exhibits superior catalytic performance from CP@Ni subjected to the Ni electrodeposition for 30, 60, 120, and
toward hydrogen evolution reaction in both acidic and alkaline 210 min (denoted as CP@Ni(X)-P, X = electrodeposition duration) are
solutions. Moreover, it also shows outstanding catalytic perfor- 14.2, 25.8, 49.6, and 83.1 mg cm−2.
mance toward oxygen evolution reaction in alkaline media. The To clarify the role of Ni-P/NiO/Ni(OH)x heterojunctions that are
high performance originates, on the one hand, from the intrin- formed upon OER, a control sample comprising CP covered with
sically high electrocatalytic activity of Ni-P for both HER and vertically aligned NiO NSs was prepared using a hydrothermal method
adapted according to a previous report.[60]
OER and the formation of Ni-P/NiO/Ni(OH)x under the OER
Characterization: SEM investigation was performed using a FEI
conditions; and on the other hand, from the macroporous con- Quanta 650 FEG microscope equipped with INCA 350 EDX (Oxford
figuration of the electrode that facilitates the mass transport of Instruments). TEM, HRTEM, SAED, and EDX mapping studies were
reactants/products as well as the release of H2/O2 gas from the carried out on a probe-corrected transmission electron microscope
electrode surface. A noble-metal-free alkaline water electrolyzer operating at 200 kV (FEI Titan ChemiSTEM 80-200). Powder X-ray
has been constructed using two identical self-supported bifunc- diffraction (XRD) experiments were conducted on a X’Pert PRO
diffractometer (PANalytical) working at 45 kV and 40 mA with the Cu
tional CP@Ni-P electrodes as cathode and anode, respectively.
Kα radiation (λ = 1.541874 Å) and a PIXcel detector. Data were collected
An electrocatalysis efficiency as high as 91% has been achieved using the Bragg–Brentano configuration in the range of 20°–80°
at 10 mA cm−2. The electrolyzer also shows very good long-term with a scan speed of 0.01° s−1. The XRD patterns were indexed and
durability for up to 100 h. Given the commercial availability of quantitatively analyzed according to the ICDD PDF-4+ database using
carbon fiber paper, easy fabrication procedure, and remarkable the HighScore software package (PANalytical). The XPS measurements
bifunctionality of Ni-P, the use of CP@Ni-P electrodes may sig- were performed using an ESCALAB 250 Xi system (Thermo Scientific)
nificantly simplify the fabrication of alkaline water electrolyzers equipped with a monochromated microfocused Al Kα X-ray source that
defined an analysis spot of ≈650 × 400 µm2. All spectra were acquired
and substantially lower the production costs, and therefore pro- in normal emission with an effective analyzer collection angle of ≈30°.
mote the widespread deployment of electrolyzers in the future. Uniform charge neutralization was provided by beams of low-energy
(≤10 eV) Ar+ ions and electrons guided by a magnetic lens; consistent
charge neutralization was verified by observing adventitious C 1s peak at
4. Experimental Section ≈284.8 eV for all samples.
Electrochemical Measurements: All electrochemical measurements
Fabrication of CP@Ni-P and CP@NiO Electrodes: Prior to were conducted at room temperature (≈25 °C) in a typical three-
electrodeposition, the CP (FuelCell Store) was first cut into small electrode or two-electrode configuration using a Biologic VMP-3
pieces with a size of 1 cm × 2 cm, and then was treated in a mixed potentiostat/galvanostat. The HER performance was evaluated in
solution of sulfuric acid and nitric acid (v (98% H2SO4):v (70% HNO3):v 0.5 M H2SO4 (pH = 0.28) or 1.0 M KOH (pH = 14) solution using the
(H2O) = 1:1:1) at 60 °C for 24 h under vigorous magnetic stirring. as-fabricated CP@Ni-P as the working electrode, a graphite plate as
The acid treatment introduces oxyl-groups and surface defects on the the counter electrode, and a saturated calomel electrode (SCE) as the
surface of carbon microfibers, which is essential for the nucleation reference. For comparison, the HER performance of a bare CP and
and conformal growth of Ni electrodeposits. The electrodeposition of commercially available Pt/C catalysts (20% Pt, FuelCell Store) supported
Ni was performed in a Watts-type electrolyte consisting of 200 g L−1 on a Ti sheet (loading mass: ≈0.3 mg cm−2) was also measured. Before
NiSO4·6H2O + 40 g L−1 NiCl2·6H2O + 40 g L−1 H3BO3 in a two-electrode each HER test, the electrolyte was deaerated by N2 bubbling for 30 min,
configuration at room temperature. A piece of CP was used as the and the N2 bubbling was then maintained during the measurement.
working electrode with an area of 1 cm2 exposed to the electrolyte, and The OER performance was evaluated in the same three-electrode
a piece of Ni foam was used as the counter electrode, which was folded configuration in 1.0 M KOH solution except that a Pt plate was used as
into a U-shape surrounding the working electrode. The deposition was the counter electrode. In this case, a bare CP and commercially available

Adv. Funct. Mater. 2016, 26, 4067–4077 © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim wileyonlinelibrary.com 4075
 www.afm-journal.de
www.MaterialsViews.com
FULL PAPER

RuO2 powders (99.9%, Alfa Aesar) supported on a Ti sheet (loading [11] D. Voiry, H. Yamaguchi, J. W. Li, R. Silva, D. C. B. Alves, T. Fujita,
mass: 0.5 mg cm−2) were employed as reference. The full electrolyzer M. W. Chen, T. Asefa, V. B. Shenoy, G. Eda, M. Chhowalla, Nat.
cell was assembled using two identical CP@Ni-P electrodes and Mater. 2013, 12, 850.
measured in a two-electrode mode in 1.0 M KOH. The SCE electrode was [12] A. I. Carim, F. H. Saadi, M. P. Soriaga, N. S. Lewis, J. Mater. Chem.
calibrated before each measurement in Ar/H2-saturated 0.5 M H2SO4 A 2014, 2, 13835.
solution using a clean Pt wire as the working electrode. All linear scan [13] H. Tang, K. Dou, C. C. Kaun, Q. Kuang, S. Yang, J. Mater. Chem. A
voltammograms measured in a three-electrode configuration are iR- 2014, 2, 360.
corrected. The correction was done according to the following equation [14] C. Tang, N. Y. Cheng, Z. H. Pu, W. Xing, X. P. Sun, Angew. Chem.,
Int. Ed. 2015, 54, 9351.
E corr = E mea − iRs (1)
[15] C. Wan, Y. N. Regmi, B. M. Leonard, Angew. Chem., Int. Ed. 2014,
126, 6525.
where Ecorr is iR-corrected potential, Emea experimentally measured [16] N. S. Alhajri, D. H. Anjum, K. Takanabe, J. Mater. Chem. A 2014, 2,
potential, and Rs the equivalent series resistance extracted from 10548.
the electrochemical impedance spectroscopy measurement. Unless
[17] H. Vrubel, X. L. Hu, Angew. Chem., Int. Ed. 2012, 51, 12703.
otherwise specified, all potentials are reported versus RHE by converting
[18] S. Gupta, N. Patel, A. Miotello, D. C. Kothari, J. Power Sources 2015,
the potentials measured versus SCE according to the following formula
279, 620.
[19] W. F. Chen, K. Sasaki, C. Ma, A. I. Frenkel, N. Marinkovic,
E (RHE ) = E ( SCE ) + 0.241+ 0.059pH (2) J. T. Muckerman, Y. M. Zhu, R. R. Adzic, Angew. Chem., Int. Ed.
2012, 51, 6131.
[20] J. F. Xie, S. Li, X. D. Zhang, J. J. Zhang, R. X. Wang, H. Zhang,
B. C. Pan, Y. Xie, Chem. Sci. 2014, 5, 4615.
Supporting Information [21] L. Ma, L. R. L. Ting, V. Molinari, C. Giordano, B. S. Yeo, J. Mater.
Chem. A 2015, 3, 8361.
Supporting Information is available from the Wiley Online Library or [22] E. J. Popczun, J. R. McKone, C. G. Read, A. J. Biacchi, A. M. Wiltrout,
from the author. N. S. Lewis, R. E. Schaak, J. Am. Chem. Soc. 2013, 135, 9267.
[23] L. G. Feng, H. Vrubel, M. Bensimon, X. L. Hu, Phys. Chem. Chem.
Phys. 2014, 16, 5917.
[24] Z. P. Huang, Z. Z. Chen, Z. B. Chen, C. C. Lv, M. G. Humphrey,
Acknowledgements C. Zhang, Nano Energy 2014, 9, 373.
L. F. Liu acknowledges the support of FCT Investigator grant (Ref. [25] Z. H. Pu, Q. Liu, C. Tang, A. M. Asiri, X. P. Sun, Nanoscale 2014, 6,
No. IF/01595/2014) from the Foundation of Science & Technology, 11031.
Portugal (FCT). X. G. Wang, W. Li, and D. H. Xiong thank the [26] X. G. Wang, Y. V. Kolen’ko, L. F. Liu, Chem. Commun. 2015, 51,
Marie Curie Programme (NanoTrainforGrowth) for COFUND 6738.
postdoctoral fellowships (Grant Agreement No. 600375). [27] X. G. Wang, Y. V. Kolen’ko, X. Q. Bao, K. Kirnov, L. F. Liu, Angew.
Chem., Int. Ed. 2015, 54, 8188.
Received: December 21, 2015 [28] A. B. Laursen, K. R. Patraju, M. J. Whitaker, M. Retuerto, T. Sarkar,
Revised: February 17, 2016 N. Yao, K. V. Ramanujachary, M. Greenblat, G. C. Dismukes, Energy
Published online: April 13, 2016 Environ. Sci. 2015, 8, 1027.
[29] E. J. Popczun, C. G. Read, C. W. Roske, N. S. Lewis, R. E. Schaak,
Angew. Chem., Int. Ed. 2014, 126, 5531.
[30] J. Q. Tian, Q. Liu, A. M. Asiri, X. P. Sun, J. Am. Chem. Soc. 2014,
[1] J. Rifkin, The Hydrogen Economy, Tarcher Penguin, New York 2012. 136, 7587.
[2] R. L. Busby, Hydrogen and Fuel Cells, Pennwell Corporation, [31] P. Jiang, Q. Liu, Y. H. Liang, J. Q. Tian, A. M. Asiri, X. P. Sun, Angew.
Tulsa 2005. Chem., Int. Ed. 2014, 53, 1.
[3] J. S. Zhang, X. C. Wang, Angew. Chem., Int. Ed. 2015, 54, 7230. [32] J. F. Callejas, J. M. McEnaney, C. G. Read, J. C. Crompton,
[4] A. Manabe, M. Kashiwase, T. Hashimoto, T. Hayashida, A. Kato, A. J. Biacchi, E. J. Popczun, T. R. Gordon, N. S. Lewis, R. E. Schaak,
K. Hirao, I. Shimomura, I. Nagashima, Electrochim. Acta 2013, ACS Nano 2014, 8, 11101.
100, 249. [33] Y. Xu, R. Wu, J. F. Zhang, Y. M. Shi, B. Zhang, Chem. Commun.
[5] T. Lipman, An overview of hydrogen production and storage sys- 2013, 49, 6656.
tems with renewable hydrogen case studies 2011 http://www. [34] P. Xiao, M. A. Sk, L. Thia, X. M. Ge, R. J. Lim, J. Y. Wang, K. H. Lim,
cesa.org/projects/hydrogen-and-fuel-cells/hydrogen-and-fuel- X. Wang, Energy Environ. Sci. 2014, 7, 2624.
cell-resource-library/resouorce/an-overview-of-hydrogen-produc- [35] J. M. EcEnaney, J. C. Crompton, J. F. Callejas, E. J. Popczun,
tion-and-storage-system-with-renewable-hydrogen-case-studies A. J. Biacchi, N. S. Lewis, R. E. Schaak, Chem. Mater. 2014, 26, 4826.
(accessed: February 2016). [36] Q. Liu, J. Q. Liu, W. Cui, P. Jiang, N. Y. Cheng, A. M. Asiri, X. P. Sun,
[6] Y. G. Li, H. L. Wang, L. M. Xie, Y. Y. Liang, G. S. Hong, H. J. Dai, Angew. Chem., Int. Ed. 2014, 53, 6710.
J. Am. Chem. Soc. 2011, 133, 7296. [37] L. A. Stern, L. Feng, F. Song, X. L. Hu, Energy Environ. Sci. 2015, 8,
[7] J. Xie, H. Zhang, S. Li, R. Wang, X. Sun, M. Zhou, J. Zhou, 2347.
X. W. Lou, Y. Xie, Adv. Mater. 2013, 25, 5807. [38] J. L. Lado, X. G. Wang, E. Paz, E. Carbo-Argibay, N. Guldris,
[8] T. Y. Wang, L. Liu, Z. W. Zhu, P. Papakonstantinou, J. B. Hu, M. Li, C. Rodriguez-Abreu, L. F. Liu, K. Kovnir, Y. V. Kolen’ko, ACS Catal.
Energy Environ. Sci. 2013, 6, 625. 2015, 5, 6503.
[9] L. Liao, S. N. Wang, J. J. Xiao, X. J. Bian, Y. H. Zhang, M. D. Scanlon, [39] H. D. Hou, A. B. Laursen, J. S. Zhang, G. G. Zhang, Y. S. Zhu,
X. L. Hu, Y. Tang, B. H. Liu, H. H. Girault, Energy Environ. Sci. 2014, X. C. Wang, S. Dahl, I. Chorkendorff, Angew. Chem., Int. Ed. 2013,
7, 387. 52, 3621.
[10] D. Merki, S. Fierro, H. Vrubel, X. L. Hu, Chem. Sci. 2011, 2, 1262. [40] G. G. Zhang, G. S. Li, X. C. Wang, ChemCatChem 2015, 7, 2864.

4076 wileyonlinelibrary.com © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Funct. Mater. 2016, 26, 4067–4077
 www.afm-journal.de
www.MaterialsViews.com

FULL PAPER
[41] T. Smolinka, in Encyclopedia of Electrochemical Power Sources [52] N. Jiang, B. You, M. L. Sheng, Y. J. Sun, Angew. Chem., Int. Ed. 2015,
(Ed: J. Garche), Elsevier, Amsterdam, Netherlands, 2009, p394. 54, 6251.
[42] M. Gong, H. J. Dai, Nano Res. 2015, 8, 23. [53] M. Grzelczak, J. S. Zhang, J. Pfrommer, J. Hartmann, M. Driess,
[43] C. C. L. McCrory, S. H. Jung, J. C. Peters, T. F. Jaramillo, J. Am. M. Antonietti, X. C. Wang, ACS Catal. 2013, 3, 383.
Chem. Soc. 2013, 135, 16977. [54] M. W. Zhang, Z. S. Luo, M. Zhou, C. J. Huang, X. C. Wang,
[44] L. A. Stern, X. L. Hu, Faraday Discuss. 2014, 176, 363. Sci. China Mater. 2015, 58, 867.
[45] J. S. Zhang, M. Grzelczak, Y. D. Hou, K. Maeda, K. Domen, X. Z. Fu, [55] Y. Yang, H. L. Fei, G. Ruan, J. M. Tour, Adv. Mater. 2015, 27, 3175.
M. Antonietti, X. C. Wang, Chem. Sci. 2012, 3, 443. [56] M. Ledendecker, S. K. Calderon, C. Papp, H. P. Steinruck,
[46] M. Gong, Y. G. Li, H. L. Wang, Y. Y. Liang, J. Z. Wu, J. G. Zhou, M. Antonietti, M. Shalom, Angew. Chem., Int. Ed. 2015, 54, 12361.
J. Wang, T. Regier, F. Wei, H. J. Dai, J. Am. Chem. Soc. 2013, 135, [57] Y. Gorlin, T. F. Jaramillo, J. Am. Chem. Soc. 2010, 132, 13612.
8452. [58] S. Mao, Z. H. Wen, T. Z. Huang, Y. Hou, J. H. Chen, Energy Environ.
[47] F. Song, X. L. Hu, Nat. Commun. 2014, 5, 4477. Sci. 2014, 7, 609.
[48] J. S. Luo, J. H. Im, M. T. Mayer, M. Schreier, M. K. Nazeeruddin, [59] J. T. Zhang, Z. H. Zhao, Z. H. Xia, L. M. Dai, Nat. Nanotechnol.
N. G. Park, S. D. Tilley, H. J. Fan, M. Gratzel, Science 2014, 345, 2015, 10, 444.
1593. [60] H. Long, T. L. Shi, H. Hu, S. L. Jiang, S. Xi, Z. R. Tang, Sci. Rep.
[49] J. W. Nai, H. J. Yin, T. T. You, L. R. Zheng, J. Zhang, P. X. Wang, 2014, 4, 7413.
Z. Jin, Y. Tian, J. Z. Liu, Z. Y. Tang, L. Guo, Adv. Energy Mater. 2015, [61] L. A. Stern, X. L. Hu, Phys. Chem. Chem. Phys. 2014, 176, 363.
5, 1401880. [62] N. I. Kolev, Multiphase Flow Dynamics, Vol. 4, Springer, Heidelberg,
[50] M. W. Kanan, D. G. Nocera, Science 2008, 321, 1072. Germany, 2009, Chapter 11, pp. 209–239.
[51] S. Cobo, J. Heidkamp, P. A. Jacques, J. Fize, V. Fourmond, [63] J. O’M Bockris, B. E. Conway, E. Yeager, R. E. White, Comprehen-
L. Guetza, B. Jousselme, V. Ivanova, H. Dau, S. Palacin, sive Treatise of Electrochemistry: Electrochemical Processing, Vol. 2,
M. Fontecave, V. Artero, Nat. Mater. 2012, 11, 802. Plenum Press, New York, NY, USA 1981, p. 14.

Adv. Funct. Mater. 2016, 26, 4067–4077 © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim wileyonlinelibrary.com 4077