Journal of Materials Chemistry A - DICPpemfc.dicp.ac.cn/201805.pdfcatalysts with high intrinsic oxygen reduction reaction (ORR) activity and enhanced durability. To achieve this, Pt

Journal ofMaterials Chemistry A

PAPER

Publ

ishe

d on

28

Mar

ch 2

018.

Dow

nloa

ded

by D

alia

n In

stitu

te o

f C

hem

ical

Phy

sics

, CA

S on

19/

04/2

018

01:4

5:08

.

View Article OnlineView Journal | View Issue

Nano-engineerin

aFuel Cell System and Engineering Laborato

Chinese Academy of Sciences, Dalian, Liaon

ac.cn; [email protected] of Chinese Academy of Sciences,cAdvanced Electron Microscopy Research Gro

Chinese Academy of Sciences, Dalian, Liaon

† Electronic supplementary informa10.1039/c7ta10901f

‡ The authors contributed equally and ca

Cite this: J. Mater. Chem. A, 2018, 6,6521

Received 13th December 2017Accepted 2nd March 2018

DOI: 10.1039/c7ta10901f

rsc.li/materials-a

This journal is © The Royal Society of C

g of a 3D-ordered membraneelectrode assembly with ultrathin Pt skin on open-walled PdCo nanotube arrays for fuel cells†

Yachao Zeng,‡ab Hongjie Zhang,‡ab Zhiqiang Wang,ab Jia Jia,ab Shu Miao,c Wei Song,a

Yu Xiao,a Hongmei Yu, *a Zhigang Shao *a and Baolian Yia

Proton exchange membrane fuel cells (PEMFCs) suffer from high consumption of Pt and low durability

under harsh operating conditions. To reduce the consumption of Pt and improve the durability, a 3D-

ordered membrane electrode assembly (MEA) based on ultrathin Pt skin on open-walled PdCo nanotube

arrays (NTAs) was nano-engineered for PEMFCs via template-assisted underpotential deposition (UPD)

and galvanic displacement. Benefiting from the advanced nanostructure, a maximum power density of

222.5 kW gPt�1 was achieved with a cathodic Pt loading of 3.5 mg cm�2, which was 13.7 fold higher than

that of a conventional MEA. Accelerated degradation tests demonstrate that the prepared

nanostructured MEA is more stable than the conventional MEA.

Introduction

Energy lies at the core of social development. With theincreasing concern of environmental deterioration, energyconversion devices with environmental benignity, high powerdensity, and high efficiency are desirable. PEMFCs are devicesthat meet all these requirements.1 Through a non-Carnot cycle,PEMFCs convert the chemical energy stored in fuels andoxidants directly into electricity. However, the high consump-tion of Pt and poor durability of PEMFCs impede their wideapplications.

The MEA is the key component of a PEMFC; it is involved inthe whole energy conversion process. Currently, there are twomethods for MEA construction: the catalyst-coated gas diffusionlayer electrode (GDE) method and the catalyst-coatedmembrane (CCM) electrode method. To perform the GDE orCCM method, a homogeneous catalyst ink composed of anelectrocatalyst, ionomer, and solvent is coated onto a substratethrough a decal,2 blade coating,3 spraying,4 or brushing5

process. The most commonly employed electrocatalyst, i.e. XC-72 or BP-2000, in MEAs at present consists of 2–4 nm Ptnanoparticles (NPs) dispersed on a carbon support with a high

ry, Dalian Institute of Chemical Physics,

ing 116023, China. E-mail: hmyu@dicp.

Beijing 100049, China

up, Dalian Institute of Chemical Physics,

ing 116023, China

tion (ESI) available. See DOI:

n been considered as co-rst authors.

hemistry 2018

surface area.6 A typical conguration of the conventional cata-lyst layer is Pt NPs/C wrapped by Naon ionomers packed witheach other, providing tortuous porosity (Scheme 1(a)). Thisconguration has the disadvantage of high Pt utilization. Intortuous porosity, gas molecules tend to be impeded by porewalls and other particles; this leads to reduced diffusive ux.The encapsulation of ionomers on electrocatalyst nanoparticlesdecreases the accessibility of active sites to reactants andreduces electronic conductivity. In addition to the highconsumption of Pt, the degradation of carbon supports andelectrolyte ionomers7 and the Ostwald ripening/aggregation ofPt NPs8 are detrimental to the durability of PEMFCs.

To reduce the high Pt consumption in PEMFCs and also toimprove their durability, one method is to synthesize electro-catalysts with high intrinsic oxygen reduction reaction (ORR)activity and enhanced durability. To achieve this, Pt alloy, core–shell, and nanostructured electrocatalysts have been devel-oped.9 Among these novel catalysts, core–shell electrocatalystswith Pt skin (or monolayer) exposed to their surroundings canmaximize Pt utilization since only the surface atoms areinvolved in ORR. Various approaches, including Cu-mediated Ptdeposition,10,11 chemical reduction,12 spontaneous deposition,13

electrochemical deposition,14,15 and atomic layer deposition16,17

have been employed in the synthesis of Pt skin electrocatalysts.Lattice strain effects and electronic structure modulationinduced by the core materials and core structures are believedto contribute to the activity enhancement in ORR.18 However,the catalyst layers derived from these novel catalysts still facethe same dilemma confronting Pt/C.

Another method is to construct MEAs with rationallydesigned catalyst layers. It has been realized that single cellperformance strongly depends on the catalyst layer structure.

J. Mater. Chem. A, 2018, 6, 6521–6533 | 6521

http://crossmark.crossref.org/dialog/?doi=10.1039/c7ta10901f&domain=pdf&date_stamp=2018-04-13http://orcid.org/0000-0002-0258-7357http://orcid.org/0000-0003-2777-6362http://dx.doi.org/10.1039/c7ta10901fhttp://pubs.rsc.org/en/journals/journal/TAhttp://pubs.rsc.org/en/journals/journal/TA?issueid=TA006015

Scheme 1 Conceptual diagrams of (a) conventional MEA and (b) 3D-ordered MEA based on open-walled Ptskin@PdCo NTAs.

Journal of Materials Chemistry A Paper

Publ

ishe

d on

28

Mar

ch 2

018.

Dow

nloa

ded

by D

alia

n In

stitu

te o

f C

hem

ical

Phy

sics

, CA

S on

19/

04/2

018

01:4

5:08

. View Article Online

Middelman proposed an ideal catalyst layer structure forPEMFCs in which the catalyst layer features ultrathin thicknessand vertically aligned porosity, while the active sites are locatedat the boundary of the electronic and ionic conductors; thisconguration favors the transport of reactants, electrons andprotons.19 Mathematical models predict greatly improved cellperformance of nanostructured MEAs over their conventionalcounterparts,20–25 and experimental studies have veried Mid-delman's hypothesis. Nanostructured materials, includingcarbon,26–28 metals29–32 and metal oxides,33–39 have been carefullyprepared and assembled in MEAs with the aim of improvingmass transport and Pt utilization. Recently, we reporteda nanostructured ultrathin catalyst layer based on open-walledPtCo nanotube arrays for PEMFCs.30 Although pronouncedcell performance with ultralow Pt loading has been achieved,the utilization of Pt has not reached its ultimate limit.

Motivated by all these facts, a 3D-ordered MEA based on Ptskin on PdCo nanotube arrays (NTAs) has been nano-engineered for PEMFCs. As illustrated in Scheme 1(b), thenanostructured MEA is composed of thousands of open-walledNTAs. It is critical that (1) the NTAs are open-walled, (2) theopen-walled NTAs are vertically aligned with gas diffusion layers(GDLs) and membranes, (3) the compact ultrathin Pt skin iscoated on the PdCo NTAs and (4) the catalyst layer is thin. The3D-ordered MEA is constructed through a template-assistedUPD and galvanic displacement method. Co–OH–CO3 nano-wire arrays (NWAs), serving as the template and cobalt source,were synthesized by a hydrothermal method. A nanostructuredPd thin lm was coated on the Co–OH–CO3 NWAs througha PVDmethod. Then, thermal annealing was performed to alloyPd with Co. The highly ordered nanostructure of PdCo NTAswas restructured on GDL through a decal method. During thefabrication process, the PdCo NTAs were open walled; thus,both the interior and exterior surfaces are available to thereactants. Then, Pt skin was deposited on the PdCo NTAsthrough UPD and galvanic displacement. The proposed benets

6522 | J. Mater. Chem. A, 2018, 6, 6521–6533

from these structural advantages are as follows: (1) ultrahigh Ptutilization because both the interior and exterior surfaces areavailable to the reactants, (2) enhanced mass transport due tothe lower tortuosity, (3) increased electronic and ionicconductivity that renders the ionomer unnecessary and (4)ultrahigh durability due to the absence of carbon support. Theproposed MEA construction strategy can effectively address thecost and durability issues confronting PEMFCs and enlightenpossible applications in other electrochemical energy storageand conversion systems.

ExperimentalChemicals

All chemicals were of analytical grade and were used as receivedwithout further purication. Co(NO3)2$6H2O, NH4F, CO(NH2)2,CuSO4$5H2O, H2SO4, and 2-propanol were purchased fromKermel® (Tianjin, China). Naon® 211membrane and Naon®ionomer solution (5 wt%) were purchased from DuPont. Pt/C(40 wt%) was purchased from Johnson Matthey. K2PtCl4(99.95%) was purchased from Borui Co., Ltd (KunMing, China).Pd target (F100 mm, with purity > 99.995%) was purchasedfrom ZhongNuo Advanced Material Technology Co., Ltd (Bei-jing, China).

Synthesis of Co–OH–CO3 NWAs

Co–OH–CO3 NWAs were synthesized on stainless steel plates viaa facile hydrothermal synthesis method reported in ourprevious work.30,39 The solution was prepared by dissolving1.5 mmol of Co(NO3)2$6H2O, 3 mmol of NH4F, and 7.5 mmolCO(NH2)2 in 70 mL of distilled water. Then, the solution wastransferred into Teon-lined stainless steel autoclave liners. Astainless steel plate (2 � 6 cm2 in size) was immersed in thereaction solution. The liner was sealed in a stainless steelautoclave andmaintained at 120 �C for 5 h, then cooled to room

This journal is © The Royal Society of Chemistry 2018

http://dx.doi.org/10.1039/c7ta10901f

Paper Journal of Materials Chemistry A

Publ

ishe

d on

28

Mar

ch 2

018.

Dow

nloa

ded

by D

alia

n In

stitu

te o

f C

hem

ical

Phy

sics

, CA

S on

19/

04/2

018

01:4

5:08


temperature. The Co–OH–CO3 NWAs supported on the stainlesssteel plate were then washed in de-ionized water to removecontaminants.

Fabrication of PdCo NTAs

The PdCo NTAs were fabricated through magnetron sputteringand thermal annealing. Pd was deposited onto Co–OH–CO3NWAs by radio frequency magnetron sputtering as reported inour previous work.30 The magnetron sputtering power densitywas xed at 1.27 W cm�2; the Ar gas pressure was 0.8 Pa. Whenthe magnetron sputtering terminated, the Pd-coated Co–OH–CO3 NWAs were transferred into a tube furnace and annealed inH2/Ar (5 vol% H2) for 1 hour, then cooled to room temperature.To investigate the impact of annealing temperature on singlecell performance, the annealing temperatures were set to300 �C, 400 �C and 500 �C, respectively. Then, the annealedNWAs and non-annealed NWAs were decaled onto GDLs ata pressure of 0.25 MPa. The GDLs were homemade from carbonpaper (Toray®, TGP-H-60) with a micro porous layer (MPL).Aer the decaling, the stainless steel plates were chemicallyremoved by immersing the assemblies in 0.5 M H2SO4 at 80 �C.Aer the removal of the stainless steel plate, the Pd NTAs-basedGDEs were immersed in 0.5 M H2SO4 at 80 �C for another 8 h.Then, the GDEs were cleaned in distilled water several times toremove possible contaminants introduced during the prepara-tion process. The puried GDEs were then dried in air. Forsimplicity, the prepared PdCo NTAs are denoted as PdCo NTAs-300, PdCo NTAs-400, and PdCo NTAs-500 according to theirannealing temperatures.

Preparation of GDEs based on Ptskin@PdCo NTAs

Pt skins were deposited on Pd and PdCo NTAs by galvanicreplacement by Pt of a Cu monolayer (ML) formed by UPD.18,40 Aleak-free reference electrode (Hg/Hg2Cl2) was used, and all thepotentials are given with respect to a reversible hydrogen elec-trode (RHE). A platinum plate served as the counter electrode.Prior to the UPD, the raw GDEs were cycled from 0.0 V to 1.2 V ata scan rate of 50 mV s�1 in 0.5 M H2SO4 for 10 cycles to polishthe surfaces of the nanotube arrays. The cleaned GDEs weretransferred into a 50 mM H2SO4 + 50 mM CuSO4 solution. Thesolution was deaerated with Ar (with purity > 99.99%) duringthe course of UPD. The GDEs were cycled 10 times in the 50 mMH2SO4 + 50 mM CuSO4 solution from 0.25 V to 0.74 V at a scanrate of 20 mV s�1 to determine the terminated potential forUPD. A linear scanning voltammetry (LSV) mode was employedto deposit Cu ML on the nanotube arrays. The potential wasscanned negatively from 0.74 V to the terminated potential ata scan rate of 1 mV s�1. As the UPD terminated, an excess ofdeaerated 10.0 mM K2PtCl4 solution was purged into the solu-tion to replace the Cu ML with Pt. The Pt loadings in the GDEswere determined from the Cu UPD charges.41,42

Preparation of conventional GDE

For comparison, a conventional Pt/C-based GDE was preparedby airbrushing a homogeneous catalyst ink onto GDLs on thehot stage at 60 �C. The catalyst ink was prepared by


ultrasonicating a mixture of Pt/C, Naon® ionomer, and 2-propanol with a mass ratio of 5 : 1 : 200. The Pt loading was setto 50 mg cm�2.

MEA fabrication and single cell tests

MEAs with an active surface area of 2.56 cm2 were prepared bysandwiching Naon® 211 membrane between two GDEs undera pressure of 0.25 MPa at 140 �C for 3 min. The counterpart GDEwith 0.25 mgPt cm

�2 was purchased from Sunrise Power Co.,Ltd. The prepared MEA was sandwiched between two silver-coated end plates with parallel ow elds, and silicon gasketswere positioned between the MEA and each of the silver-coatedend plates.

The single cell performance and durability were evaluated ona homemade fuel cell test stand at a cell temperature of 80 �C.The anode and cathode were purged with humidied H2 and O2with ow rates of 100 mL min�1 and 200 mL min�1, respec-tively. The humidity is 100%. The backpressure for both theanode and cathode was 2 bar. A KFM 2030 Impedance Meter(Kikusui®, Japan) was employed to record the I–V curves. Elec-trochemical impedance spectroscopy (EIS) was performed toinvestigate the electrode kinetics of the prepared MEAs. The EISwas carried out on a Solartron 1287 Electrochemical Interface inconjunction with a Solartron 1260 Frequency Respond Analyzer.The measurements were performed at a certain current densityby applying an ac amplitude of 10 mV over the ac frequencyrange from 1 Hz to 10 kHz.

Durability tests

Accelerating degradation tests (ADTs) were performed on a CHI600C (CH Instruments, USA) by sweeping the potential between0.6 and 1.0 V vs. RHE at a scan rate of 50 mV s�1 to investigatethe durability of the prepared MEAs. During the ADT test, theanode side was purged with H2, while the cathode side waspurged with N2. H2/N2 gases were externally humidied at thedew point temperature of 80 �C. The cell temperature wasmaintained at 80 �C. The gas ow rate of H2/N2 was 50/100mL min�1. Then, the H2 crossover current was measured bylinear scanning voltammetry from 0.0 to 0.7 V at a scan rate of4 mV s�1; the current at 0.4 V served as the H2 crossover current.Aer that, cyclic voltammograms (CVs) were recorded between0.05 V and 1.2 V (vs. RHE) at a scan rate of 100mV s�1 at 30 �C. I–V curves and EIS data were collected every 1000 cycles to diag-nose the single cell performance.

Physical characterizations

The morphologies of the prepared samples were investigatedusing a scanning electronmicroscope (JSM-7800F and IT 300LA,JEOL) equipped with an energy dispersive X-ray spectrometer.Samples were mounted using double-sided adhesive tape.

Samples for scanning transmission electron microscopy(TEM) were prepared by dispersion in ultrahigh purity ethanol.A drop of the suspension was then allowed to evaporate ona carbon-coated copper grid. The structure evolutions of thenanotubes were investigated by high resolution transmissionelectron microscopy (HRTEM, JEM-2100F, JEOL). The nature of

J. Mater. Chem. A, 2018, 6, 6521–6533 | 6523



Publ

ishe

d on

28

Mar

ch 2

018.

Dow

nloa

ded

by D

alia

n In

stitu

te o

f C

hem

ical

Phy

sics

, CA

S on

19/

04/2

018

01:4

5:08


the Pt skin on PdCo NTAs was investigated by aberration-corrected scanning transmission electron microscopy(ACSTEM, ARM200, JEOL) operated at 200 kV.

Gazing incidence X-ray diffraction (GIXRD) patterns wererecorded on a X-ray diffractometer (PANalytical EMPYREAN)operating at 40 kV and 200 mA with Cu Ka (incident angle ¼0.5�, l ¼ 1.5405 Å, step size ¼ 0.5�) as a radiation source.

X-ray photoelectron spectroscopy (XPS) analyses were per-formed to reveal the electronic structures of the prepared GDEs.The XPS spectra were obtained with an ESCALab250 Xi electronspectrometer using 300 W Al Ka radiation. The base pressurewas about 1 � 10�7 Pa. The binding energies were calibratedto the C1s line at 284.6 eV from adventitious carboncontamination.

The Pd loadings in the GDEs were determined by inductivelycoupled plasma optical emission spectroscopy (ICP-OES, Per-kinElmer 7300DV).

Results and discussionStructural characterization

The fabrication process of the nanostructured GDEs based onPtskin@PdCo NTAs is schematically presented in Scheme S1.†Highly ordered Co–OH–CO3 NWAs served as the template andcobalt source. The growth mechanism of Co–OH–CO3 NWAshas been elucidated in the ref. 30 and 43. Owing to their highlyordered nanostructure and facile preparation method, Co–OH–CO3 NWAs have been widely applied in the synthesis of nano-structured materials for electrochemical energy storage andconversion devices.7,43–45 Herein, Pd has been chosen as the corematerial due to its high ORR activity and low price (its price isca. 1/3 that of Pt). Thermal annealing was employed to provideadditional mechanical strength to Pd@Co–OH–CO3 NWAs.

Fig. 1 shows the typical morphologies of the samplesinvolved in the fabrication. From Fig. 1(a), Co–OH–CO3 NWAs

Fig. 1 (a) In-panel image of Co–OH–CO3 NWAs grown on a stainless ste(c) Cross sectional image of the Co–OH–CO3 NWAs. (d) and (e) In-pancations. (f) Cross sectional image of the PdCo NWAs annealed at 400 �C

6524 | J. Mater. Chem. A, 2018, 6, 6521–6533

present a highly ordered distribution on a large scale; the areanumber density of Co–OH–CO3 NWAs was estimated to be 3 to 4billion cm�2. A more detailed structure is revealed in Fig. 1(b);as can be seen, small tips branching out from the trunk andravines can be identied between smaller nanowires, indicatingthat a single Co–OH–CO3 nanowire may be a bundle of smallernanowires. This unique structure is vital for the formation ofopen-walled PdCo NTAs, as will be discussed later. FromFig. 1(c), the Co–OH–CO3 NWAs grew relatively vertically on thestainless steel plate; the average diameter of the Co–OH–CO3NWAs is ca. 100 nm, and the length is ca. 3.0 mm. Fig. 1(d) and(e) show the in-panel images of the PdCo NWAs annealed at400 �C. From Fig. 1(d), the PdCo NWAs retained an orderednanostructure even aer thermal annealing was conducted.Aer PVD of the Pd thin lm and thermal annealing, thenanowires evolved from needle-like to rodlike shapes.

Then, the nanowire arrays were decaled onto GDL. A GDL isa piece of carbon paper or carbon cloth made from wovencarbon ber. To enhance their water management capabilitiesand also to reduce contact resistance, a thin layer of PTFE-treated carbon powder is mounted onto the carbon paper orcarbon cloth to form a microporous layer (MPL). The existenceof the MPL was vital for the restructuring of highly orderednanostructures in this research. During the decal transfer,nanowire arrays embedded into the MPL. Aer the removal ofthe substrate by acid washing, the PdCo NTAs were integratedwith GDL. The whole process resembles planting rice ina paddy. The appearances of the GDEs prior to and aer thedeposition of Pt skin are presented in Fig. S1.† From Fig. S1(a),†large scale uniform distribution of PdCo NTAs on GDL wasrealized via the decal method. It is noteworthy that the UPD andgalvanic replacement did not alter the uniformity.

Fig. 2 shows the morphology of the GDE based on Ptskin@PdCo NTAs-400. From Fig. 2(a) and (b), Ptskin@PdCo NTAs-400interconnected with each other, forming a network on GDL.

el plate. (b) High resolution SEM image of the Co–OH–CO3 nanowires.el images of the PdCo NWAs annealed at 400 �C at different magnifi-.



Fig. 2 (a) and (b) In-panel SEM images of Ptskin@PdCo NTAs-400 at different magnifications. (c) and (d) Cross sectional images of the GDE basedon Ptskin@PdCo NTAs-400.


Publ

ishe

d on

28

Mar

ch 2

018.

Dow

nloa

ded

by D

alia

n In

stitu

te o

f C

hem

ical

Phy

sics

, CA

S on

19/

04/2

018

01:4

5:08


From Fig. 2(c), Ptskin@PdCo NTAs-400 vertically aligned withGDL, and the tips of Ptskin@PdCo NTAs punctured into theMPL. The thickness of the catalyst layer is estimated to be 3 mm,which is in accordance with the length of the Co–OH–CO3NWAs. From Fig. 2(d), the Ptskin@PdCo NTAs-400 are openwalled; both the interior and exterior surfaces are exposed to thesurroundings.

The formation process of the open-walled nanotube arrayshas been schematically illustrated in Fig. S2.† During thethermal annealing, Co–OH–CO3 NWAs were rstly reduced toCo NWAs by H2. With a further increase in temperature, Cobegan to diffuse into the Pd thin lm, forming a PdCo alloyshell. Aer the decaling of PdCo bimetallic NWAs onto the GDL,purication in acid solution was performed. During the acidwash, Co was etched by acid and H2 bubbles clogged thenanotubes. The H2 bubbles continuously pushed the nanotubewalls, eventually leading to cracking of the nanotubes.

Because it is a powerful technique for probing surfaceproperties, cyclic voltammetry was employed to record thesurface transitions. Fig. S3† shows the CV curves of the Pd-based NTAs prior to and aer the UPD and galvanic replace-ment. Prior to the UPD, the Pd-based NTAs were subjected tointensive CV treatment from 0 to 1.2 V to obtain a refreshedsurface. In Fig. S2,† the classical cyclic voltammogram of Pd-based alloys is presented. Broad peaks of hydrogen adsorptionand desorption on Pd-based NTAs were observed in the rangefrom 0 to 0.3 V. Due to the oxyphilicity of Pd, the adsorption ofoxygenated species begins at a low potential, and the desorptionpeaks of oxygenated species are located at a lower potential.


Aer the deposition of Pt skin on Pd-based NTAs, the peakintensities increased in the hydrogen desorption region, whichserves as indirect evidence for the formation of a core–shellstructure.46 Additionally, a positive shi of oxide (OHads)adsorption and desorption peaks in the region of 0.6 to 1.0 Vwas discovered aer the deposition of Pt skin on Pd-basedNTAs.

To investigate the formation of Pt skin on Pd-based NTAs,ACSTEM-EDS was employed to further investigate the distri-butions of Pt, Pd and Co in the nanotubes. To conrm that boththe interior and exterior surfaces were decorated with Pt skin,ACSTEM-EDS mapping was performed in view of the fact thatboth the interior and exterior surfaces are orientated to theobservation; the results are presented in Fig. 3. In Fig. 3(a), boththe interior (dark) and exterior (bright) surfaces can beobserved. In Fig. 3(b), a typical core–shell structure was ob-tained; Pt skin formed a uniform distribution on PdCo NTAs-400, and no obvious surface aws or pinholes can beobserved. In Fig. 3(c), the uniform distribution of Pt on theinterior and exterior surfaces is conrmed. By means ofACSTEM and EDS analysis, we conrmed that UPD aligned withgalvanic replacement can deposit a compact Pt skin on both theinterior and exterior surfaces. A warning should be emphasizedhere that imperfections in the Pt skin may not have beendetected; further optimization of the fabrication should beperformed to eliminate these aws.

Fig. S4(a)† shows a close-up HAADF-ACSTEM image ofPtskin@PdCo NTAs-400. EDS elemental mappings of Pt(Fig. S4c†), Pd (Fig. S4d†) and Co (Fig. S4e†) in the same region

J. Mater. Chem. A, 2018, 6, 6521–6533 | 6525


Fig. 3 (a) ACSTEM image of an open-walled Ptskin@PdCo nanotube annealed at 400 �C with arrows indicating the interior and exterior surfaces.(b) Overlapping EDS elemental map of Pt, Pd and Co. EDS maps of (c) Pt, (d) Pd and (e) Co in the same area.


Publ

ishe

d on

28

Mar

ch 2

018.

Dow

nloa

ded

by D

alia

n In

stitu

te o

f C

hem

ical

Phy

sics

, CA

S on

19/

04/2

018

01:4

5:08


are also shown. To present the relative positions of theelemental distributions, an overlapping EDS map (Fig. S4b†)and a line-scan EDS (Fig. S4f†) have been co-presented. FromFig. S4(b),† a brighter line emerges along the rim, indicatingthat Pt was enriched on the surface of the PdCo NTAs. Thethickness of the Pt skin was determined from the full width athalf-maximum (FWHM) of the Pt M peak.42,47,48 The thickness isca. 1.29 nm. The estimated thickness is much greater than thetheoretical value of Pt monolayer (2.78 Å). The discrepancy inthickness may originate from the uneven surface of Ptskin@PdCo NTAs-400. In the case of a rough surface, even if Ptmonolayer was formed, the Pt signals may overlap along therim, leading to the illusion that a thicker Pt skin rather than a Ptmonolayer was deposited. In Fig. S4(f),† a notable difference inthe elemental distributions of Pd and Co can be identied. Theprole of Co K indicates a depletion of Co near the surface of thePdCo nanotube wall. The depletion of Co may be caused bythe electrochemical dealloying induced by CV treatment prior tothe UPD. This treatment is important to ensure that the Pt skinwas deposited through galvanic replacement of Cu ML ratherthan Co.

GIXRD was performed to analyze the crystal structure ofPtskin@PdCo NTAs. Fig. 4 presents the XRD patterns of theGDEs. Standard JCPDS cards (no. 01-089-8489 C, no. 01-089-8490 C, no. 00-004-0802 Pt, no. 01-87-0643 Pd, no. 00-050-1437Pd2Co) served as references. Graphite carbon peaks can be

6526 | J. Mater. Chem. A, 2018, 6, 6521–6533

identied in all the samples, which is ascribed to the GDL. Face-centered cubic crystal phase was detected for all of thePtskin@PdCo NTAs-based GDEs. The parameters derived fromthe diffraction peaks at (220) are summarized in Table S1.† Thediffraction peaks shied to higher diffraction angles with thethermal annealing temperature, indicating that latticecontraction resulted from the thermal annealing. No distinctdiffraction patterns corresponding to Pt crystal can be identiedin Ptskin@PdCo NTAs, indicating that Pt did not form largeparticles or islands on the Pd-based NTAs. The diffractionpatterns of Ptskin@PdCo NTAs-500 correspond well with thoseof Pd2Co (JCPDS PDF 00-05-1437), indicating that thermalannealing at higher temperature leads to the formation of PdCoalloy with higher Co content. The Co amounts in Ptskin@PdCoNTAs were determined through EDS. Fig. S5† presents themorphologies and compositions of the nanostructured GDEs.

From the ESD spectra, the Co content increased with thethermal annealing temperature (the ratios of Pd : Co are sum-marised in Table S2†). The EDS analysis indicates that the Cocontent can be modulated through ne control over theannealing temperature.

From Fig. S5(a–d),† thermal annealing also has a greatimpact on the morphology. When the temperature increased to500 �C, agglomerates formed on the GDE. Annealing at highertemperature leads to sintering. Fig. S6† illustrates themorphologies of the single Ptskin@Pd and Ptskin@PdCo



Fig. 4 XRD patterns of the nanostructured GDEs based on Ptskin@Pdand Ptskin@PdCo NTAs annealed at different temperatures with thestandard JCPDS cards listed below.


Publ

ishe

d on

28

Mar

ch 2

018.

Dow

nloa

ded

by D

alia

n In

stitu

te o

f C

hem

ical

Phy

sics

, CA

S on

19/

04/2

018

01:4

5:08


nanotubes. As shown in Fig. S6(a),† nanowhiskerettes wereoriented uniformly on the surface of the Ptskin@Pd nanotubes.With increasing temperature, the surfaces of the PdCo nano-tubes coarsened and the nanowhiskerettes merged into largegrains. As shown in Fig. S6(d),† nanopores were formed duringthe thermal annealing and acid wash treatment. Thus, Co–OH–CO3 NWAs as both the template and cobalt source weresuccessfully employed in modulating the morphology andcomposition of PdCo NTAs.

Performance of Ptskin@PdCo NTAs-based MEAs

The Ptskin@PdCo NTAs-based GDEs were tested as cathodes andanodes in a single cell conguration. Cyclic voltammetry andEIS were employed as diagnostic tools to obtain better knowl-edge of the kinetics of the electrode process. The cathodeperformance is presented in Fig. 5, with the conventional Pt/Cand Pd NTAs-based GDEs serving as baselines. Key parame-ters are also summarized in Table 1 for comparison.

From Table 1, the Pt loading in the nanostructured GDEsranged from 3.5 to 5.6 mg cm�2, which is about one order ofmagnitude lower than that of the conventional GDE. Consid-ering the fact that Pd also has ORR electroactivity, the cellperformance was then normalized into platinum group metal(PGM) loading per geometric active surface area (Fig. 5(b)).

From Fig. 5(a), comparable cell performance to that ofa conventional GDE was achieved by Ptskin@PdCo NTAs-400,


although the Pt loading in Ptskin@PdCo NTAs-400 is one orderof magnitude lower than in the conventional GDE. The Pd NTAspresented the lowest performance. Due to its more negativeoxidizing potential and stronger adsorption of oxygenatedspecies, Pd is expected to be less active for ORR. By overlayinga Pt skin on the Pd and PdCo NTAs, a notable improvement incell performance was achieved. More importantly, the cellperformance varied with the thermal annealing temperature.With increasing thermal annealing temperature, the cellperformance initially increased; the maximum performancewas achieved by Ptskin@PdCo NTAs-400 (776.7 mW cm

�2), aerwhich the performance decreased.

Generally, the I–V curve for a fuel cell can be divided intothree regions, i.e., the activation polarization region, the ohmicpolarization region and the mass transport polarization region.The activation polarization occurs at a low current density (1 A cm�2).

In the activation polarization region, EIS (@100 mA cm�2)was conducted to quantify the ORR electroactivity and decouplethe factors that affect the single cell performance. The resultsare presented in Fig. 5(c) and Table 1. R0, which is derived fromthe intersection of the EIS spectrum with the real axis, indicatesthe comprehensive ohmic resistance of a single cell, includinginterfacial contact resistances, bulk resistances and ionicresistances in the catalyst layers. Considering that all the vari-ables other than the cathode catalyst layers were exactly thesame in this research, R0 is a scale of the difference in thecathode catalyst layers. In Table 1, the R0 values of the studiedMEAs are signicantly lower than that of the conventional MEA,demonstrating that the electronic conductivity in the Ptskin@PdCo NTAs is improved compared to the conventional catalystlayer. The decrease in the electronic resistance of Ptskin@PdCoNTAs may be due to (1) the higher electronic conductivity of thePtskin@PdCo NTAs compared to carbon NPs, (2) the highlyordered express highway for electrons in Ptskin@PdCo NTAscompared with the arbitrary construction in the conventionalcatalyst layer, and (3) the absence of an electronic insulatingionomer in the nanostructured catalyst layers.

Although no proton conductive ionomer was introduced inPtskin@PdCo NTAs-400, a comparable performance to that ofa conventional MEA was achieved with Pt loading an order ofmagnitude lower. Research on ionomer-free catalyst layers havebeen previously reported in the ref. 49–52. The transport ofprotons may be mainly due to the surface adsorbed water and/or oxygenated species. Litster et al. proposed that a similarproton conduction mechanism may be adopted in liquid satu-rated Naon® and bulk water on a Pt surface.53 Also, the protonconductivity on the liquid saturated Pt surface was estimated tobe�1 Sm�1, which is 1 to 2 orders of magnitude lower than thatof the Naon® surface.52 Beneting from the greatly reducedcatalyst layer thickness, the poor intrinsic proton conductivityon the Pt skin was compensated.

Rct, which is usually derived from a tting method, describesa global reacting resistance towards specic reactions (e.g.,ORR, HOR). From Table 1, the values of Rct varied in the order of

J. Mater. Chem. A, 2018, 6, 6521–6533 | 6527


Fig. 5 (a) Single cell performance of Pd NTAs, Ptskin@Pd NTAs, Ptskin@PdCo NTAs-300, Ptskin@PdCo NTAs-400 and Ptskin@PdCo NTAs-500 ascathodes. (Testing conditions: active area 2.56 cm2, 80 �C; the gas flow rate of H2/O2was 100/200mLmin

�1, and the gas humidity was 100%). (b)Single cell performance of the prepared MEAs with the current density normalized to the cathode PGM loadings. (c) EIS data obtained at anoperating current density of 100 mA cm�2. (d) Characteristic CVs of the prepared GDEs recorded at 30 �C with a scan rate of 50 mV s�1.

Table 1 Characteristic parameters of fuel cell performance with different GDEs as cathodes

CathodesPt loading/mg cm�2

PGM loading/mg cm�2

Power density

R0/mU cm2 Rct/mU cm

2 ECSA/m2 gPGM�1mW cm�2 kW gPt

�1 kW gPGM�1

Conventional GDE 50.0 50.0 812.8 16.3 16.3 90.6 450.9 73.5Pd NTAs — 32.3 334.8 — 10.4 61.1 871.2 42.3a

Ptskin@Pd NTAs 5.6 37.9 494.7 87.7 13.1 77.8 858.9 82.1Ptskin@PdCo NTAs-300 4.1 36.4 718.2 175.2 19.8 54.6 552.5 70.0Ptskin@PdCo NTAs-400 3.5 35.7 776.7 222.5 21.7 58.2 531.8 82.7Ptskin@PdCo NTAs-500 5.1 37.4 563.5 109.6 15.1 70.4 630.8 80.2

a The ECSA of the Pd NTAs was determined by the deposition charge of Cu ML with a factor of 407 mC cm�2.55


Publ

ishe

d on

28

Mar

ch 2

018.

Dow

nloa

ded

by D

alia

n In

stitu

te o

f C

hem

ical

Phy

sics

, CA

S on

19/

04/2

018

01:4

5:08


conventional GDE < Ptskin@PdCo NTAs-400 < Ptskin@PdCoNTAs-300 < Ptskin@PdCo NTAs-500 < Ptskin@Pd NTAs < PdNTAs. This tendency is in accordance with the I–V curves. It iswell known that a reaction activity is a product of the intrinsicactivity and available active sites. The intrinsic ORR activity canbe tuned through the electronic and geometric properties of theelectrocatalyst.

XPS, a surface sensitive detection method, was performed toprobe the electronic properties of the prepared GDEs. Fig. S7†shows the high-resolution Pt 4f spectra of the nanostructuredGDEs and the conventional Pt/C GDE. The most intense peak ofPt 4f is assigned to metallic Pt (Pt0), and the other two peaks areassigned to PtII and PtIV, respectively. Notably, the Pt 4f bindingenergy of Ptskin@Pd NTAs showed a signicant negative shi

6528 | J. Mater. Chem. A, 2018, 6, 6521–6533

(0.24 eV) compared to that of Pt/C; this may be caused by thestrong electron interaction between Pt and Pd because theelectronegativity of Pt (2.28 eV) is greater than that of Pd (2.20eV).54 However, for Ptskin@PdCo NTAs, the tendency shi of thePt 4f binding energy decreased in different degrees and wasmost pronounced when the heat treatment temperatures forPdCo NTAs were 300 �C and 400 �C. The EDS results showedthat the Co content in PdCo NTAs increased with increasingannealing temperature. Aer alloying Pd with Co, the interac-tion between Pt and PdCo became stronger, and the muchsmaller Pd and Co atoms induced a lattice contraction in the Ptskin; this induced a negative shi of the d-band-center of Pt.56

This change in the electronic structure of Pt may be the cause ofthe positive displacement of the binding energy and is




Publ

ishe

d on

28

Mar

ch 2

018.

Dow

nloa

ded

by D

alia

n In

stitu

te o

f C

hem

ical

Phy

sics

, CA

S on

19/

04/2

018

01:4

5:08


benecial to improve the catalytic activity of Pt for ORR.57 Aerelectrochemical scanning treatment for Ptskin@PdCo NTAs, thetransition metal Co in the shallow layer dissolves in the acidicenvironment and the surface of the NTAs is covered by a Pt-richskin structure. It has been reported that alloy catalysts with thinPtskin-surface layer structures can show higher ORR activity andcyclic stability due to their effective electronmodulation and theprotection of the Pt skin.58–60 Table S6† summarizes theparameters derived from XPS; the metallic Pt content of thePtskin@PdCo NTAs-400 is obviously higher than that underother conditions, which may also be the reason why Ptskin@-PdCo NTAs-400 showed the best performance during ourexperiments.

The available active sites were determined using thehydrogen desorption areas from the cyclic voltammograms(Fig. 5(d)) and then by calculating the effects of factors such asPGM loading in the electrode and the charge required to oxidizea monolayer of H. From Table 1, the ECSAs of Ptskin@Pd NTAsand Ptskin@PdCo NTAs increased by nearly 2 times that of thePd NTAs, indicating that the ECSAs are mainly contributed bythe Pt skin. Although the Pt loadings in the nanostructuredGDEs are nearly one order of magnitude lower, a similar ECSA(normalized to PGM loading) was obtained, demonstrating thevery high dispersion of Pt atoms on Pd and the PdCo NTAs.

In the mass transport region (>1200 mA cm�2), a potentialdrop was observed in the nanostructured GDEs. The drop maybe ascribed to the hydrophilic properties of the Pt skin. Fig. S8†compares the contact angles of the nanostructured GDEs andthe conventional GDE. The contact angle of the conventionalGDE is signicantly higher than those of the nanostructuredGDEs, manifesting their much improved capabilities in expel-ling liquid water. However, the contact angles of the nano-structured GDEs fell near or below 90�, indicating that watertends to reside in the NTAs.

The hydrophilic properties impeded the nanostructuredGDEs from achieving higher performance. However, when theperformances were normalized into PGM loadings, the struc-tural advantages of the open-walled Ptskin@PdCo NTAs pre-vailed. In Fig. 5(b), the difference in cell performance can beclearly identied. Maximum mass specic power densities of222.5 kW gPt

�1 and 21.7 kW gPGM�1 were achieved by Ptskin@

PdCo NTAs-400 with 3.5 mgPt cm�2 and 32.3 mgPd cm

�2, respec-tively, which is 13.7-fold higher than the conventional GDE with50 mgPt cm

�2. When the cell performance is normalized to the Ptloading, the power densities of Pt skin-based MEAs are muchgreater than those of the conventional GDE, revealing theultrahigh Pt utilization in the nanostructured GDEs.

The remarkable cell performance and the ultrahigh Pt utili-zation arise from the unique structure of the Ptskin@PdCoNTAs. In the cathode of a fuel cell, the electrocatalysis of ORR iscoupled withmultiphase mass transport. As shown in Fig. 2 and3, the open-walled Ptskin@PdCo NTAs are vertically aligned withthe GDL and membrane. Due to the open-walled structure, theinterior walls of the Ptskin@PdCo NTAs can be easily accessed bythe reactants. Meanwhile, the water produced from the ORR canbe easily propelled from the active sites. The reduced thicknessin the Ptskin@PdCo NTAs shortened the transport lengths of


electrons, reactants (i.e., H+, O2) and products (i.e., H2O).Compared with an arbitrary packing conguration withtorturous porosity, the straight channels in Ptskin@PdCo NTAswill enhance the mass transport by improving the effectivetransport coefficients. In a conventional catalyst layer, Naonionomers are frequently employed. On one hand, Naon ion-omers provide additional dimensional stability to the catalystlayer. On the other hand, Naon ionomers act as protonconductors. However, the insulating Naon ionomers willincrease the ohmic resistance in a catalyst layer. Recently, it wasclaimed that local oxygen transport resistance near the activesite is caused by Naon ionomers.61–63 Hence, it is vital todevelop novel proton conductors to further enhance theperformance of PEMFCs. Herein, Ptskin@PdCo NTAs are directlyaligned with GDLs without any binder. Also, water thin lm is indemand as a native proton conductor due to its relatively highproton conductivity and high O2 solubility.64,65

Fig. S9† presents the cell performance of Ptskin@PdCo NTAsserving as anodes. Due to the fast kinetics of HOR on Pt, it ishighly possible to reduce to the Pt loading in PEMFCs from theanode side. The anode performance is summarized in TableS4.† A maximum power density was achieved by Ptskin@PdCoNTAs-500 (1698.1 mW cm�2) with a Pt loading of 5.1 mg cm�2,which is 13.6 fold higher than that of the conventional GDE.The much improved performance also benets from theadvantageous open-walled vertical nanotubular structure, inwhich protons and H2 can be easily delivered.

Durability tests

Durability is another critical parameter that strongly impactsthe application of PEMFCs. The durabilities of the Pt skin-basedand conventional Pt/C MEAs were evaluated through an ADTprotocol suggested by DOE.66 In Fig. 6(a), a notable performancedrop of Ptskin@PdCo NTAs-400 MEA can be observed aer therst 1000 ADT cycles; thereaer, the performance of Ptskin@PdCo NTAs-400 MEA appears to stabilize, and no distinctiveperformance loss is subsequently observed. However, in thecase of the conventional Pt/C MEA, progressive performanceloss can be seen during the ADT cycles. To gain insight into thesingle cell performance evolution, EIS (@100 mA cm�2 and 1 Acm�2), H2 crossover currents and CV curves were recorded every1000 ADT cycles. The key parameters are summarized inTable 2.

Aer 5000 ADT cycles, the power density retention ofPtskin@PdCo NTAs-400 MEA (63.5%) is signicantly higher thanthat of the conventional MEA (17.5%); additionally, the ECSAretention of Ptskin@PdCo NTAs-400 (60.7%) also exceeds that ofthe conventional GDE (37.6%). The R0 of Ptskin@PdCo NTAs-400MEA remained nearly unchanged; however, the values forconventional MEA increased twofold aer the ADT. The muchincreased ohmic resistance in the conventional MEA indicateda deterioration of the proton conduction pathway in theconventional catalyst layer.67 The EIS (@100 mA cm�2) forPtskin@PdCo NTAs-400 and conventional MEAs increasedsimultaneously, indicating a decrease in ORR activity. EIS ob-tained at a high current density can serve as a measure of mass

J. Mater. Chem. A, 2018, 6, 6521–6533 | 6529


Fig. 6 Single cell performance of (a) Ptskin@PdCo NTAs-400 and (b) conventional Pt/C GDE during accelerated degradation tests. EIS ofPtskin@PdCo NTAs-400 and conventional Pt/C GDE recorded at (c) 100 mA cm

�2 and (d) 1 A cm�2 during the ADT. (e) ECSAs retention ofPtskin@PdCo NTAs-400 and conventional Pt/C GDE during the ADT. (f) H2 crossover current densities of Ptskin@PdCo NTAs-400 andconventional Pt/C MEA during the ADT.

Table 2 Summary of single cell performance degradation during the accelerated degradation tests

Cathode

Power density/mW cm�2 ECSA/m2 gPt

�1 R0/mU cm2

Rct/mU cm2

(@100 mA cm�2)

InitialAer 5000CV cycles Initial

Aer 5000CV cycles Initial

Aer 5000CV cycles Initial

Aer 5000CV cycles

Conventional GDE 809.6 142.0 73.5 27.6 91.0 195.0 45.0 572.0Ptskin@PdCo NTAs-400 776.7 493.0 82.7 50.2 56.0 88.0 53.0 68.0


Publ

ishe

d on

28

Mar

ch 2

018.

Dow

nloa

ded

by D

alia

n In

stitu

te o

f C

hem

ical

Phy

sics

, CA

S on

19/

04/2

018

01:4

5:08


transport resistance. Fig. 6(d) presents the EIS of Ptskin@PdCoNTAs-400 MEA recorded at 1 A cm�2. Sole semi-circles withresembled diameters are presented, manifesting that the mass

6530 | J. Mater. Chem. A, 2018, 6, 6521–6533

transport resistance is steady aer 5000 ADT cycles. The highlystable mass transport resistance may be due to the mechanicrobustness of Ptskin@PdCo NTAs. However, EIS (@1 A cm

�2) for




Publ

ishe

d on

28

Mar

ch 2

018.

Dow

nloa

ded

by D

alia

n In

stitu

te o

f C

hem

ical

Phy

sics

, CA

S on

19/

04/2

018

01:4

5:08


the conventional GDE could be barely obtained aer 5000 ADTcycles, due to degradation of the catalyst layer structure inducedfrom the ADT. Carbon material will experience severe corrosionunder the dynamic operating conditions of PEMFCs; thecorrosion of the carbon support will eventually lead to a struc-tural collapse of the cathode electrode, further worsening themass transport in the catalyst layer. The performance drop ofthe preparedMEAs was diagnosed bymonitoring the ECSAs andH2 crossover current densities. From Fig. 6(e), the ECSAs ofPtskin@PdCo NTAs-400 MEA remained constant aer 3000 ADTcycles, indicating that the active sites in Ptskin@PdCo NTAs-400were stable aer 3000 ADT cycles. Meanwhile, a durativedecrease in the ECSAs was observed for the conventional GDE.In the case of Pt/C, the Pt NPs tended to aggolomerate into largeparticles through Ostwald ripening/aggregation, and thecorrosion of the carbon support led to detachment of the PtNPs. In contrast, in the case of Ptskin@PdCo NTAs-400, a self-healing mechanism may contribute to the high stability of Ptskin during the ADT.10,40,47,68 Among the three constituents ofthis core–shell electrocatalyst, Co is the most reactive metal,followed by Pd, with the lowest standard electrode potential.During the potential sweeping, Co and Pd leached out from thecore. The dissolution of Co and Pd inhibits the progress of thepotentials to high values, hence preventing the dissolution ofthe Pt skin. As the dissolution of Pd and Co proceeds from thecore, the excess of Pt atoms from the shell can form a partial bi/multilayer due to the Kirkendall effect.68 Fig. S10† presentsa linear EDS performed on a cross section of the Ptskin@PdCoNTAs-400-based MEA aer 5000 ADT cycles. It can be seen thatPt was mainly concentrated in the cathode and anode catalystlayers, while Co was distributed in the cathode and anode GDLs.Pd was deposited in the membrane and anode via reduction ofPd2+ by H2 diffusing from the anode. As shown in Fig. 6(f) , theH2 crossover current increased synchronously with the currentdensity; the H2 crossover current of Ptskin@PdCo NTAs-400 washigher than that of the conventional Pt/C GDE. The mildincrease of H2 crossover current in Ptskin@PdCo NTAs-400veries that the Pd dissolution did not damage the electrolytemembrane during the ADT.

Intriguingly, Co is present only in the GDLs and the cathode.This result is in accord with the reported research.48,69,70

Generally, the dissolution of transition metals will contaminatethe Naon ionomers,71,72 causing limited proton transport and/or reduced O2 permeability.73 The absence of Co cation inNaonmembrane and anode catalyst layers gives a positive signthat Ptskin@PdCo NTAs are promising for PEMFCs. However,the dissolution of Pd and Co should be avoided to meet therequirements for application of Ptskin@PdCo NTAs in fuel cells.It has been suggested that by alloying Pd with Au, the durabilityof Pt skin-based electrocatalysts can be greatly enhanced.40,42,74

Inspired by this, the durability of the Ptskin@PdCo NTAs can begreatly enhanced by adopting the same strategy.

Conclusions

In summary, a 3D-ordered MEA with ultralow Pt loading andenhanced durability was designed and constructed for PEMFCs.


Using Co–OH–CO3 NWAs as the template and cobalt source,open-walled PdCo NTAs were synthesized through a PVD andthermal annealing method. The open-walled PdCo NTAs wererestructured on GDL through a facile decal method. To endowthe PdCo NTAs with ORR activity and also durability, ultrathinPt skin was deposited on the PdCo NTAs through UPD andgalvanic displacement. The Ptskin@PdCo NTAs were employedas the cathode and anode for a PEMFC. Maximum powerdensities of 222.5 kW gPt

�1 (cathode) and 330.4 kW gPt�1

(anode) were accomplished. The much improved single cellperformance may be due to the advantageous structure of theopen-walled Ptskin@PdCo NTAs, whose reduced thickness andstraight porosity enhance mass transport. More importantly,the open-walled Ptskin@PdCo NTAs exhibited impressivestability during durability tests. A self-healing mechanism maybe the reason for the high stability of Ptskin@PdCo NTAs.

The nano-engineering strategy reported in this work is notonly limited to Co–OH–CO3 NWAs as the template and Pd ascore material; diverse nanostructured templates and electro-chemical active materials can also be incorporated in thisframework, which will greatly promote the development ofelectrochemical energy conversion and storage devices with theintent of reducing noble metal consumption.

Conflicts of interest

There are no conicts to declare.

Acknowledgements

We gratefully acknowledge the nancial support from theNational Key Research and Development Program of China (No.2016YFB0101205), the National Natural Science Foundation ofChina (No. U1508202, No. U1664259, No. 91434131 and No.21576257).

Notes and references

1 F. Barbir, PEM Fuel Cells : Theory and Practice, ElsevierAcademic Press, Waltham, MA, 2nd edn, 2013.

2 M. S. Saha, D. K. Paul, B. A. Peppley and K. Karan,Electrochem. Commun., 2010, 12, 410–413.

3 I.-S. Park, W. Li and A. Manthiram, J. Power Sources, 2010,195, 7078–7082.

4 K.-H. Kim, K.-Y. Lee, H.-J. Kim, E. Cho, S.-Y. Lee, T.-H. Lim,S. P. Yoon, I. C. Hwang and J. H. Jang, Int. J. Hydrogen Energy,2010, 35, 2119–2126.

5 J. Xie, K. L. More, T. A. Zawodzinski and W. H. Smith,J. Electrochem. Soc., 2004, 151, A1841–A1846.

6 H. A. Gasteiger, S. S. Kocha, B. Sompalli and F. T. Wagner,Appl. Catal., B, 2005, 56, 9–35.

7 P. Malacrida, M. Escudero-Escribano, A. Verdaguer-Casadevall, I. E. L. Stephens and I. Chorkendorff, J. Mater.Chem. A, 2014, 2, 4234–4243.

8 S. Zhang, X. Z. Yuan, J. N. C. Hin, H. Wang, K. A. Friedrichand M. Schulze, J. Power Sources, 2009, 194, 588–600.

J. Mater. Chem. A, 2018, 6, 6521–6533 | 6531



Publ

ishe

d on

28

Mar

ch 2

018.

Dow

nloa

ded

by D

alia

n In

stitu

te o

f C

hem

ical

Phy

sics

, CA

S on

19/

04/2

018

01:4

5:08


9 M. Shao, Q. Chang, J.-P. Dodelet and R. Chenitz, Chem. Rev.,2016, 116, 3594–3657.

10 R. R. Adzic, Electrocatalysis, 2012, 3, 163–169.11 R. R. Adzic, J. Zhang, K. Sasaki, M. B. Vukmirovic, M. Shao,

J. X. Wang, A. U. Nilekar, M. Mavrikakis, J. A. Valerio andF. Uribe, Top. Catal., 2007, 46, 249–262.

12 Y. Zhang, Y.-C. Hsieh, V. Volkov, D. Su, W. An, R. Si, Y. Zhu,P. Liu, J. X. Wang and R. R. Adzic, ACS Catal., 2014, 4, 738–742.

13 Y. Dai and S. Chen, ACS Appl. Mater. Interfaces, 2015, 7, 823–829.

14 S. H. Ahn, H. Tan, M. Haensch, Y. Liu, L. A. Bendersky andT. P. Moffat, Energy Environ. Sci., 2015, 8, 3557–3562.

15 Y. Liu, D. Gokcen, U. Bertocci and T. P. Moffat, Science, 2012,338, 1327–1330.

16 L. Su,W. Jia, C.-M. Li and Y. Lei, ChemSusChem, 2014, 7, 361–378.

17 H. Yang, Angew. Chem., Int. Ed., 2011, 50, 2674–2676.18 J. Zhang, M. B. Vukmirovic, Y. Xu, M. Mavrikakis and

R. R. Adzic, Angew. Chem., Int. Ed., 2005, 44, 2132–2135.19 E. Middelman, Fuel Cells Bulletin, 2002, 2002, 9–12.20 C. Y. Du, X. Q. Cheng, T. Yang, G. P. Yin and P. F. Shi,

Electrochem. Commun., 2005, 7, 1411–1416.21 M. Chisaka and H. Daiguji, Electrochem. Commun., 2006, 8,

1304–1308.22 C. Y. Du, T. Yang, P. F. Shi, G. P. Yin and X. Q. Cheng,

Electrochim. Acta, 2006, 51, 4934–4941.23 S. M. Rao and Y. Xing, J. Power Sources, 2008, 185, 1094–1100.24 M. M. Hussain, D. Song, Z. S. Liu and Z. Xie, J. Power Sources,

2011, 196, 4533–4544.25 S. Sun, H. Zhang and M. Pan, Fuel Cells, 2015, 15, 456–462.26 S. Murata, M. Imanishi, S. Hasegawa and R. Namba, J. Power

Sources, 2014, 253, 104–113.27 Z. Q. Tian, S. H. Lim, C. K. Poh, Z. Tang, Z. Xia, Z. Luo,

P. K. Shen, D. Chua, Y. P. Feng, Z. Shen and J. Lin, Adv.Energy Mater., 2011, 1, 1205–1214.

28 M. D. Gasda, G. A. Eisman and D. Gall, J. Electrochem. Soc.,2010, 157, B437–B440.

29 S. Galbiati, A. Morin and N. Pauc, Electrochim. Acta, 2014,125, 107–116.

30 Y. Zeng, Z. Shao, H. Zhang, Z. Wang, S. Hong, H. Yu andB. Yi, Nano Energy, 2017, 34, 344–355.

31 O.-H. Kim, Y.-H. Cho, S. H. Kang, H.-Y. Park, M. Kim,J. W. Lim, D. Y. Chung, M. J. Lee, H. Choe and Y.-E. Sung,Nat. Commun., 2013, 4, 2473.

32 D. F. van der Vliet, C. Wang, D. Tripkovic, D. Strmcnik,X. F. Zhang, M. K. Debe, R. T. Atanasoski, N. M. Markovicand V. R. Stamenkovic, Nat. Mater., 2012, 11, 1051–1058.

33 C. K. Zhang, H. M. Yu, Y. K. Li, W. Song, B. L. Yi andZ. G. Shao, Electrochim. Acta, 2012, 80, 1–6.

34 C. K. Zhang, H. M. Yu, Y. K. Li, Y. Gao, Y. Zhao, W. Song,Z. G. Shao and B. L. Yi, Chemsuschem, 2013, 6, 659–666.

35 M. A. Garćıa-Contreras, S. M. Fernández-Valverde,J. R. Vargas-Garćıa, M. A. Cortés-Jácome, J. A. Toledo-Antonio and C. Ángeles-Chavez, Int. J. Hydrogen Energy,2008, 33, 6672–6680.

6532 | J. Mater. Chem. A, 2018, 6, 6521–6533

36 M. S. Saha, M. N. Banis, Y. Zhang, R. Y. Li, X. L. Sun, M. Caiand F. T. Wagner, J. Power Sources, 2009, 192, 330–335.

37 M. S. Saha, R. Li, M. Cai and X. Sun, J. Power Sources, 2008,185, 1079–1085.

38 M. S. Saha, R. Y. Li, M. Cai and X. L. Sun, Electrochem. Solid-State Lett., 2007, 10, B130–B133.

39 Y. Zeng, X. Guo, Z. Wang, J. Geng, H. Zhang, W. Song, H. Yu,Z.-G. Shao and B. Yi, Nanoscale, 2017, 9, 6910–6919.

40 K. Sasaki, H. Naohara, Y. Choi, Y. Cai, W.-F. Chen, P. Liu andR. R. Adzic, Nat. Commun., 2012, 3, 1–9.

41 J. X. Wang, H. Inada, L. Wu, Y. Zhu, Y. Choi, P. Liu,W.-P. Zhou and R. R. Adzic, J. Am. Chem. Soc., 2009, 131,17298–17302.

42 K. A. Kuttiyiel, K. Sasaki, D. Su, M. B. Vukmirovic,N. S. Marinkovic and R. R. Adzic, Electrochim. Acta, 2013,110, 267–272.

43 X. Xia, J. Tu, Y. Zhang, X. Wang, C. Gu, X.-b. Zhao andH. J. Fan, ACS Nano, 2012, 6, 5531–5538.

44 Q. Yang, Z. Y. Lu, T. Li, X. M. Sun and J. F. Liu, Nano Energy,2014, 7, 170–178.

45 C. Zhang, H. Yu, L. Fu, Y. Gao, J. Jia, S. Jiang, B. Yi andZ. Shao, RSC Adv., 2014, 4, 58591–58595.

46 D. Dang, H. Zou, Z. a. Xiong, S. Hou, T. Shu, H. Nan, X. Zeng,J. Zeng and S. Liao, ACS Catal., 2015, 5, 4318–4324.

47 A. Kongkanand, N. P. Subramanian, Y. Yu, Z. Liu, H. Igarashiand D. A. Muller, ACS Catal., 2016, 6, 1578–1583.

48 P. Yu, M. Pemberton and P. Plasse, J. Power Sources, 2005,144, 11–20.

49 S. J. An and S. Litster, in Polymer Electrolyte Fuel Cells 13, ed.H. A. Gasteiger, A. Weber, K. Shinohara, H. Uchida, S.Mitsushima, T. J. Schmidt, S. R. Narayanan, V. Ramani, T.Fuller, M. Edmundson, P. Strasser, R. Mantz, J. Fenton,F. N. Buchi, D. C. Hansen, D. L. Jones, C. Coutanceau, K.SwiderLyons and K. A. Perry, 2013, vol 58, pp. 831–839.

50 M. Cimenti, D. Bessarabov, M. Tam and J. Stumper, inElectrode Processes Relevant to Fuel Cell Technology, ed. V.Birss, W. Mustain, D. Wilkinson, P. Kulesza and K. Ota,2010, vol 28, pp. 147–157.

51 J. W. Lim, Y.-H. Cho, M. Ahn, D. Y. Chung, Y.-H. Cho,N. Jung, Y. S. Kang, O.-H. Kim, M. J. Lee, M. Kim andY.-E. Sung, J. Electrochem. Soc., 2012, 159, B378–B384.

52 E. L. Thompson and D. R. Baker, in Polymer Electrolyte FuelCells 11, ed. H. A. Gasteiger, A. Weber, S. R. Narayanan,D. Jones, P. Strasser, K. SwiderLyons, F. N. Buchi,P. Shirvanian, H. Nakagawa, H. Uchida, S. Mukerjee,T. J. Schmidt, V. Ramani, T. Fuller, M. Edmundson,C. Lamy and R. Mantz, 2011, vol 41, pp. 709–720.

53 S. J. An and S. Litster, in Polymer Electrolyte Fuel Cells 13, ed.H. A. Gasteiger, A. Weber, K. Shinohara, et al., 2013, vol. 58,pp. 831–839.

54 J. H. Yang, J. Yang and J. Y. Ying, ACS Nano, 2012, 6, 9373–9382.

55 L. Fang, Q. Tao, M. Li, L. Liao, D. Chen and Y. Chen, Chin. J.Chem. Phys., 2010, 23, 543–548.

56 J. L. Zhang, M. B. Vukmirovic, Y. Xu, M. Mavrikakis andR. R. Adzic, Angew. Chem., Int. Ed., 2005, 44, 2132–2135.




Publ

ishe

d on

28

Mar

ch 2

018.

Dow

nloa

ded

by D

alia

n In

stitu

te o

f C

hem

ical

Phy

sics

, CA

S on

19/

04/2

018

01:4

5:08


57 Y. Nie, L. Li and Z. D. Wei, Chem. Soc. Rev., 2015, 44, 2168–2201.

58 W. He, H. J. Jiang, Y. Zhou, S. D. Yang, X. Z. Xue, Z. Q. Zou,X. G. Zhang, D. L. Akins andH. Yang,Carbon, 2012, 50, 265–274.

59 L. Liu, G. Samjeske, S. Nagamatsu, O. Sekizawa, K. Nagasawa,S. Takao, Y. Imaizumi, T. Yamamoto, T. Uruga andY. Iwasawa, J. Phys. Chem. C, 2012, 116, 23453–23464.

60 H. Q. Li, G. Q. Sun, N. Li, S. G. Sun, D. S. Su and Q. Xin,J. Phys. Chem. C, 2007, 111, 5605–5617.

61 T. A. Greszler, D. Caulk and P. Sinha, J. Electrochem. Soc.,2012, 159, F831–F840.

62 R. Jinnouchi, K. Kudo, N. Kitano and Y. Morimoto,Electrochim. Acta, 2016, 188, 767–776.

63 A. Kongkanand and M. F. Mathias, J. Phys. Chem. Lett., 2016,7, 1127–1137.

64 I. V. Zenyuk and S. Litster, Electrochim. Acta, 2014, 146, 194–206.

65 Q. Wang, M. Eikerling, D. Song and Z. S. Liu, J. Electrochem.Soc., 2007, 154, F95–F101.

66 R. M. Mukundan, Accelerated Testing Validation, https://www.hydrogen.energy.gov/pdfs/review14/fc016_mukundan_2014_o.pdf).


67 S. Park, Y. Y. Shao, H. Y. Wan, V. V. Viswanathan,S. A. Towne, P. C. Rieke, J. Liu and Y. Wang, J. Phys. Chem.C, 2011, 115, 22633–22639.

68 K. Sasaki, H. Naohara, Y. Cai, Y. M. Choi, P. Liu,M. B. Vukmirovic, J. X. Wang and R. R. Adzic, Angew.Chem., Int. Ed., 2010, 49, 8602–8607.

69 Q. Jia, J. Li, K. Caldwell, D. E. Ramaker, J. M. Ziegelbauer,R. S. Kukreja, A. Kongkanand and S. Mukerjee, ACS Catal.,2016, 6, 928–938.

70 X. Peng, S. Zhao, T. J. Omasta, J. M. Roller andW. E. Mustain, Appl. Catal., B, 2017, 203, 927–935.

71 X. Jie, Z.-G. Shao and B. Yi, Electrochem. Commun., 2010, 12,700–702.

72 X. Wang, L. Zhang, G. Li, G. Zhang, Z.-G. Shao and B. Yi,Electrochim. Acta, 2015, 158, 253–257.

73 D. Myers, N. Kariuki, R. Ahluwalia, X. Wang and J.-K. Peng,Rationally Designed Catalyst Layers for PEMFCPerformance Optimization, http://www.hydrogen.energy.gov/pdfs/review15/fc106_myers_2015_o.pdf.).

74 C. Koenigsmann, M. E. Scoeld, H. Q. Liu and S. S. Wong,J. Phys. Chem. Lett., 2012, 3, 3385–3398.

J. Mater. Chem. A, 2018, 6, 6521–6533 | 6533


Nano-engineering of a 3D-ordered membrane electrode assembly with ultrathin Pt skin on open-walled PdCo nanotube arrays for fuel cellsElectronic supplementary information (ESI) available. See DOI: 10.1039/c7ta10901fNano-engineering of a 3D-ordered membrane electrode assembly with ultrathin Pt skin on open-walled PdCo nanotube arrays for fuel cellsElectronic supplementary information (ESI) available. See DOI: 10.1039/c7ta10901fNano-engineering of a 3D-ordered membrane electrode assembly with ultrathin Pt skin on open-walled PdCo nanotube arrays for fuel cellsElectronic supplementary information (ESI) available. See DOI: 10.1039/c7ta10901fNano-engineering of a 3D-ordered membrane electrode assembly with ultrathin Pt skin on open-walled PdCo nanotube arrays for fuel cellsElectronic supplementary information (ESI) available. See DOI: 10.1039/c7ta10901fNano-engineering of a 3D-ordered membrane electrode assembly with ultrathin Pt skin on open-walled PdCo nanotube arrays for fuel cellsElectronic supplementary information (ESI) available. See DOI: 10.1039/c7ta10901fNano-engineering of a 3D-ordered membrane electrode assembly with ultrathin Pt skin on open-walled PdCo nanotube arrays for fuel cellsElectronic supplementary information (ESI) available. See DOI: 10.1039/c7ta10901fNano-engineering of a 3D-ordered membrane electrode assembly with ultrathin Pt skin on open-walled PdCo nanotube arrays for fuel cellsElectronic supplementary information (ESI) available. See DOI: 10.1039/c7ta10901fNano-engineering of a 3D-ordered membrane electrode assembly with ultrathin Pt skin on open-walled PdCo nanotube arrays for fuel cellsElectronic supplementary information (ESI) available. See DOI: 10.1039/c7ta10901fNano-engineering of a 3D-ordered membrane electrode assembly with ultrathin Pt skin on open-walled PdCo nanotube arrays for fuel cellsElectronic supplementary information (ESI) available. See DOI: 10.1039/c7ta10901fNano-engineering of a 3D-ordered membrane electrode assembly with ultrathin Pt skin on open-walled PdCo nanotube arrays for fuel cellsElectronic supplementary information (ESI) available. See DOI: 10.1039/c7ta10901fNano-engineering of a 3D-ordered membrane electrode assembly with ultrathin Pt skin on open-walled PdCo nanotube arrays for fuel cellsElectronic supplementary information (ESI) available. See DOI: 10.1039/c7ta10901f

Nano-engineering of a 3D-ordered membrane electrode assembly with ultrathin Pt skin on open-walled PdCo nanotube arrays for fuel cellsElectronic supplementary information (ESI) available. See DOI: 10.1039/c7ta10901fNano-engineering of a 3D-ordered membrane electrode assembly with ultrathin Pt skin on open-walled PdCo nanotube arrays for fuel cellsElectronic supplementary information (ESI) available. See DOI: 10.1039/c7ta10901fNano-engineering of a 3D-ordered membrane electrode assembly with ultrathin Pt skin on open-walled PdCo nanotube arrays for fuel cellsElectronic supplementary information (ESI) available. See DOI: 10.1039/c7ta10901fNano-engineering of a 3D-ordered membrane electrode assembly with ultrathin Pt skin on open-walled PdCo nanotube arrays for fuel cellsElectronic supplementary information (ESI) available. See DOI: 10.1039/c7ta10901f

Nano-engineering of a 3D-ordered membrane electrode assembly with ultrathin Pt skin on open-walled PdCo nanotube arrays for fuel cellsElectronic supplementary information (ESI) available. See DOI: 10.1039/c7ta10901fNano-engineering of a 3D-ordered membrane electrode assembly with ultrathin Pt skin on open-walled PdCo nanotube arrays for fuel cellsElectronic supplementary information (ESI) available. See DOI: 10.1039/c7ta10901fNano-engineering of a 3D-ordered membrane electrode assembly with ultrathin Pt skin on open-walled PdCo nanotube arrays for fuel cellsElectronic supplementary information (ESI) available. See DOI: 10.1039/c7ta10901f

Documents

Journal of Materials Chemistry A - DICPpemfc.dicp.ac.cn/201805.pdfcatalysts with high intrinsic oxygen reduction reaction (ORR) activity and enhanced durability. To achieve this, Pt