Electrochimica Acta 229 (2017) 361–370

Metal-Organic Framework Derived CoNi@CNTs Embedded CarbonNanocages for Efficient Dye-Sensitized Solar Cells

Zhiqiang Xie, Xiaodan Cui, Wangwang Xu, Ying Wang*Department of Mechanical & Industrial Engineering, Louisiana State University, Baton Rouge, LA 70803, USA

A R T I C L E I N F O

Article history:Received 1 December 2016Received in revised form 20 January 2017Accepted 22 January 2017Available online 28 January 2017

Keywords:CoNi alloycarbon nanocagesmetal-organic frameworkcounter electrodedye-sensitized solar cell

A B S T R A C T

The commercialization of dye-sensitized solar cells (DSSCs) has been severely hindered by high cost andscarcity of Pt as counter electrode (CE). Up to date, the design of low-cost Pt-free CE materials with anideal combination of high electrical conductivity, excellent catalytic activity and satisfactory long-termelectrochemical stability still poses challenges to researchers. In this work, we have developed novel CoNialloy@carbon nanotubes embedded carbon nanocages (CoNi@CNTs-C) as the CE of DSSCs for the firsttime, by applying Co/Ni bimetallic metal organic framework (MOF) as the template. The powerconversion efficiency (PCE) of DSSCs can be maximized by simply optimizing the ratio of Ni2+/Co2+

precursor ratio during the synthesis. As a result, the DSSC based on CoNi@CNTs-C-200 CE with a Co/Niatomic ratio of 9:1 exhibits a remarkable PCE of 9.04%, showing a �15% enhancement compared to that ofPt CE (7.88%). Notably, the CoNi@CNTs-C-200 CE also demonstrates excellent long-term electrochemicalstability over 300 cyclic voltammetry cycles, which is much superior to that of conventional Pt CE. Thus,the CoNi@CNTs-C demonstrates a great potential as a low-cost, stable and efficient CE material fornext-generation DSSCs.

© 2017 Elsevier Ltd. All rights reserved.

Contents lists available at ScienceDirect

Electrochimica Acta

journa l home page : www.e l sev ier .com/ loca te /e le cta cta

1. Introduction

Dye-sensitized solar cells (DSSCs) have become one of the mostpromising photovoltaic devices for efficient conversion of solarenergy into electricity, owing to their environmental friendlinessand easy fabrication [1,2]. DSSCs mainly consist of three crucialcomponents: a photoanode, a counter electrode (CE) and aelectrolyte containing iodide-triiodide (I�/I3�) redox shuttle [3].Among them, the CEs play a key role in governing the powerconversion efficiency (PCE) since they mainly transfer electronsfrom the external circuit and perform electrocatalytic reduction ofthe redox couples [4]. Therefore, as an ideal CE, it should possesspromising characteristics, such as excellent electrical conductivity,high electrocatalytic activity and low cost [5]. At present, platinum(Pt) has been widely employed as a high-performance CE for DSSCssince it could satisfy most of the above-mentioned characteristics.Nevertheless, Pt is highly expensive and extremely scarce, therebyseverely impeding large-scale fabrication of DSSCs for practical

* Corresponding author.E-mail address: ywang@lsu.edu (Y. Wang).

http://dx.doi.org/10.1016/j.electacta.2017.01.1450013-4686/© 2017 Elsevier Ltd. All rights reserved.

commercialization. In addition, the DSSCs based on Pt as CEmaterials still show unsatisfactory long-term operation stabilitymainly due to the corrosive iodide/triiodide redox electrolyte [6].

To overcome the aforementioned challenges, it is alwaysdesirable to explore new alternatives to replace Pt as efficientand low-cost CEs for DSSCs. To date, various Pt-free CEs based onconducting polymers [7,8], carbon materials (carbon black [9],graphene [10,11], carbon nanotubes [12,13] etc.) and transitionmetal compounds (selenides [14,15], sulfides [16,17], nitrides [18],oxides [19–21], carbides [22] etc.) have been reported.Unfortunately, these materials are either limited by their lowelectrical conductivities or poor electrocatalytic activities. Theirperformances are still not comparable with the Pt, and most of thereported DSSCs have PCEs generally lower than 7.5% [23,24]. Inaddition, the stability of most new Pt-free CEs remains largelyunknown. Recently, various metal alloys have demonstrated betterelectrocatalytic activity than single metals and most carbonaceousmaterials. For example, Zheng et al. reported that FeNi alloynanoparticles (NPs) encapsulated within carbon nanotubes (CNTs)showed higher PCE than that of Fe NPs encapsulated within CNTs[5]. Motlak et al. reported that the optimal NiCu nanoparticle-decorated graphene as the CE exhibited a PCE of 5.1%, which is

362 Z. Xie et al. / Electrochimica Acta 229 (2017) 361–370

much higher than that of Ni-decorated graphene (2.39%),indicating significantly enhanced electrocatalytic activity [25].CoNi nanoparticles-doped carbon nanofibers prepared byannealing treatment of electrospun nanofibers have demonstrateda PCE of 4.47% [26]. Despite much progress having been made inthe development of alloy-based CEs, the electrocatalytic activityand electrochemical stability are still not comparable to Pt. Suchlimited performance is originated from the much stricterrequirement for the CEs such as chemical composition, particlesize and structural stability. Therefore, more efforts are highlydesired to optimize the alloy-based CEs for more efficient andstable DSSCs.

Metal-organic frameworks (MOFs), built from metalions/clusters as nodes and organic linkers as struts, have attractedmore and more interest in recent years owing to their high surfacearea, large pore volume and diverse structures. Inspired by theseproperties, MOFs have been reported as a new template andprecursor for synthesis of various hierarchical nanostructuredmaterials such as porous carbons [27], metal oxides [28],metal/carbon composites [29] and metal oxides/carboncomposites [30]. These MOF-derived nanostructures can offermany unique advantages: (i) the chemical composition can beeasily tuned by designing MOFs combined with specific thermaltreatment; (ii) MOF-derived nanostructures provide controlledporosity and huge surface area, which can effectively facilitate theaccess of electrolyte into the electrode and ensure largeelectrolyte/electrode contact area; (iii) the charge diffusion lengthscan be largely shortened; (vi) the low cost and ease of synthesisallow MOF-derived nanostructures to be potentially scaled up forindustrial applications. Recently, our group reported CoS2embedded carbon matrix as CEs for DSSCs by sulfurizing MOFs,which displays a comparable PCE to that of Pt [31]. However, tofurther improve the PCE and electrochemical stability ofMOF-derived CEs, a material with a more robust nanostructureand more active catalytic sites are highly desired. Compared tosulfides-based CEs, metal alloy nanoparticles are more environ-mental friendly and have emerged as new candidates for CE inDSSCs, however, the preparation of metal alloy/carbon nano-structures by using bimetallic MOFs as the single precursor hasrarely been reported [32,33]. Compared to other synthesismethods of alloy nanoparticles, one-step pyrolysis of bimetallicMOFs offers a new route to prepare well-dispersed alloy nano-particles embedded conductive carbon matrix without using anysurfactants or toxic reducing agent. The chemical composition andparticle size of alloy can also be optimized by simply varying themixed metal ions ratio during the synthesis.

Herein, novel CoNi alloy@carbon nanotubes embedded carbonnanocages (CoNi@CNTs-C) have been successfully prepared,through one-step pyrolysis of Co/Ni bimetallic metal organicframework (MOF) in Ar atmosphere. Within such MOF-derivednanostructure, the CoNi alloy nanoparticles offer numerous activesites to efficiently catalyze the redox reactions, and meanwhile theCNTs and carbon nanocages serve as conductive network, whichcan prevent CoNi nanoparticles from aggregation and reducecorrosion by acidic electrolyte. As demonstrated in variouselectrochemical measurements, the optimal CoNi@CNTs-C-200CE demonstrates the highest PCE and current density of 9.04% and18.3 mA/cm2, respectively, which are significantly higher than7.88% and 15.0 mA/cm2 from the DSSC based on Pt CE. To ourknowledge, this work is the first report on the application ofCoNi@CNTs-C as counter electrode for efficient DSSCs. We believethat the strategy presented here opens up a new route todevelopment of versatile metal alloy@CNTs-C composites for arich variety of electrochemical applications, such as DSSCs, watersplitting, and so forth.

2. Experimental Section

2.1. Preparation of Co-C and CoNi@CNTs-C

In brief, 2 g cobalt nitrate hexahydrate and 3.06 g 2-methyl-imidazole were separately dissolved in 100 mL methanol,respectively. Afterwards, the two solutions were mixed understirring for 10 min and aged for 24 h at room temperature. Theresultant purple solid was collected from the solution bycentrifugation at 4000 rpm for 10 min and then washed withmethanol at least three times. After washing, the zeoliticimidazolate framework-67 (ZIF-67) powders were dried overnightat 80 �C. Afterwards, cobalt nanoparticles embedded withinN-doped carbon nanocages (Co-C) were prepared by a facilecalcination of ZIF-67 nanocrystals at 900 �C for 2 hours in Aratmosphere.

When introducing different amounts of nickel nitrate hexahy-drate (100, 200, 400 mg, respectively) into the above synthesis ofZIF-67, Ni2+ incorporated ZIF-67 powders were synthesized andthen transformed to CoNi@CNTs-C by a direct pyrolysis at 900 �Cfor 2 hours in Ar atmosphere. For convenience, the as-preparedCoNi@CNTs-C samples were labeled as CoNi@CNTs-C-100, CoN-i@CNTs-C-200 and CoNi@CNTs-C-400, respectively.

2.2. Electrodes preparation

The photoanodes and counter electrodes for DSSCs werefabricated according to our previous reports [16,17]. In brief, forthe preparation of the counter electrodes in DSSCs, 0.2 g of theas-synthesized products were well dispersed in ethanol (2 mL) bycontinuously magnetic stirring. Subsequently, the solutions ofterpineol (0.86 mL) and ethyl celluloses (1.1 mL) in ethanol solventwere dipped into the above mixed solution, respectively, followedby sonication and continuous stirring. Afterwards, the resultedslurry was uniformly coated onto the FTO coated glass (7 V/sqSigma-Aldrich) by simply spin coating method at 4000 rpm for30 s. Afterwards, the annealing treatment of electrodes was carriedout at 450 �C for a half hour in argon atmosphere. For comparisonpurpose, Pt counter electrodes were also prepared by following thestandard preparation method: drop casting 0.5 mMH2PtCl6/isopropanol solution onto the FTO coated glass, followedby post-treatment at 450 �C for 20 min in air.

For the preparation of photoanodes in DSSCs, 1 g TiO2 nano-particles with the size of 25 nm (Sigma Aldrich 99.7%) were welldispersed in ethanol (8 mL), terpineol (4.3 mL) and ethyl celluloses(5.5 mL) in ethanol (10 wt%). Afterwards, the above mixed solutionwas concentrating down at 70 �C due to the gradual solventevaporation. Subsequently, the as-obtained slurry was applied tothe FTO coated glass through the doctor’s blade method. Then, theheat treatment of photoanodes was performed at 125 �C for 6 minin air, 325 �C for 5 min, 375 �C for 5 min, 450 �C for 15 min and500 �C for 15 min sequentially and then the photoanodes cooldown to the room temperature. Afterwards, the TiO2 photoanodeswere soaked in anhydrous ethanol solution containing 0.2 mMN719 dye (Ru[LL’-(NCS)2], L = 2,20-bipyridyl-4,40-dicarboxylic acid,L’ = 2,20-bipyridyl-4,40-ditetrabutyl-ammonium carboxylate,Solaronix Co.) and dye-sensitized for 24 h at room temperature.

2.3. DSSC fabrication

Typically, the counter electrode and the dye-sensitized TiO2

photoanode were sandwiched together with the hot-melt sealingfilm as the spacer (100 mm in thickness). Afterwards, theelectrolyte containing 0.1 M GTC (guanidine thiocyanate) in a

Z. Xie et al. / Electrochimica Acta 229 (2017) 361–370 363

mixture of acetonitrile and valeronitrile (85:15 vol/vol) (No.ES-0004, IoLiTec Inc., Germany) was injected into the DSSC devices.

2.4. Characterizations and measurements

The crystalline structures of as-obtained products wereidentified by using X-ray diffraction (XRD) technique, and thedata were recorded at a constant scanning rate of 2�/min on aRigaku MiniFlex X-ray diffractometer (Cu Ka radiation). Themicrostructures and morphologies of samples were characterizedby using scanning electron microscope (SEM) and high-resolutiontransmission electron microscope (HRTEM). The SEM imaging andelemental mapping were carried out on a FEI Quanta 3D FEGFIB/SEM with energy-dispersive X-ray spectroscopy (EDS)equipped. TEM and HRTEM imaging was conducted on a JEOLJEM-2010 microscope at 200 kV.

The typical current-voltage (J-V) performance of DSSCs wasmeasured by using a Keithley 2400 source meter, which has a solarlight simulator (model: 67005, Oriel) to simulate sunlight underone sun illumination (100 mW/cm2). A series of electrochemicalmeasurements were also carried out, including the cyclicvoltammetry (CV), electrochemical impedance spectroscopy(EIS) and Tafel polarization. The CV measurements were performedat a scan rate of 20 mV/s by using an electrochemical workstation(CHI 6504C). Specifically, for the CV measurements, an Hg/Hg+

electrode works as reference electrode, the as-fabricated counterelectrode acts as working electrode and a Pt foil works as auxiliaryelectrode. The total exposure area of the counter electrode in anacetonitrile solution containing 0.1 M LiClO4, 10 mM LiI and 1 mMI2 is 1.5 cm2. The EIS and Tafel polarization measurements werecarried out using an electrochemical work station (CHI 6504C). Inbrief, the EIS measurements were conducted by using an appliedvoltage aptitude of 10 mV and bias voltage 0 V in the frequencyrange from 105Hz to 0.01 Hz. The Tafel data were collected in thevoltage range of 1 V to �1 V.

3. Results and discussion

The overall synthesis of MOF-derived CoNi@CNTs-C nano-structures involves two steps. Firstly, the Co/Ni bimetallic metalorganic frameworks (MOFs) are synthesized using various ratio ofCo2+/Ni2+ and 2-methylimidazole at room temperature. During thefollowing pyrolysis process at 900 �C for 2 h, Co2+ and Ni2+ are

Fig.1. (a) XRD patterns and (b) Selected enlarged portion of CoNi@CNTs-C-100, CoNi@CNpure Co and Ni.

simultaneously reduced into metallic Co and Ni by the resultantcarbon from the decomposition of organic linkers and meanwhilemetallic Co and Ni undergo the alloying reaction to form CoNi alloynanoparticles. The resultant CoNi alloy can subsequently functionas catalysts to promote the formation of CNTs. In this process,bimetallic MOFs act not only as the carbon source for the formationof carbon nanotubes with the assistance of CoNi alloy catalysts, butalso as a sacrificial template for the nanocage-like framework [34].Specifically, the carbon nanocages are formed by annealing the as-prepared MOF particles as the template at 900 �C in inert gas, inwhich the organic ligands 2-Methylimidazole (C4H6N2) containingrich nitrogen content serve as the carbon source. During theannealing treatment, the organic ligands within the polyhedron-like MOFs will decompose and transform into the N-doped carbonmatrix, retaining a similar morphology with the MOF template.This formation process of N-doped carbon matrix using MOFs asthe template/precursor has been confirmed by our early work [39]and other previous reports [29,30]. Finally, we obtained theCoNi@CNTs-C with a unique nanostructure, whereas CoNi alloynanoparticles are well encapsulated into the CNTs (CoNi@CNTs),and the CoNi@CNTs are completely embedded into carbonnanocages. It is worth mentioning that the synthesis process ofCoNi nanoparticles embedded carbon nanocages is facile andscalable, since we simply use MOFs as the template and precursorsduring the whole synthesis process. It does not require complexprocedures or precise control of reaction conditions such as pH andconcentration of the solutions.

To identify the crystalline structures and phases of theMOF-derived CoNi@CNTs-C samples, the powder X-ray diffraction(XRD) measurements are first performed. The XRD results showthat all the CoNi@CNTs-C samples are composed of graphiticcarbon and face-centered cubic (fcc) CoNi alloy (Fig. 1a) [26]. It isclearly observed that the strongest (111) reflection peak positionsof all samples lie between the peak position of pure Co (JCPDS#15-0806) and pure Ni (JCPDS# 04-0850) (Fig. 1b), which is wellconsistent with previously reported CoNi alloys [35,36]. The main(111) peaks of CoNi@CNTs-C samples obviously shift to higherangles with the increase of Ni2+ amounts during the synthesis,further confirming the formation of CoNi alloys. No otherimpurities or phases are observed, implying a complete transfor-mation of Ni2+ incorporated ZIF-67 to CoNi@CNTs-C. In addition,the sharp peaks in the XRD spectra of the CoNi alloy indicate thatthe samples are highly crystalline [26].

Ts-C-200 and CoNi@CNTs-C-400 samples in comparison with standard XRD peaks of

364 Z. Xie et al. / Electrochimica Acta 229 (2017) 361–370

Scanning electron microscopy (SEM) images (Fig. S1) show thatZIF-67 samples are composed of polyhedron-like particles withsmooth surfaces. However, after one-step pyrolysis process in inertgas atmosphere, the as-synthesized CoNi@CNTs-C-100 andCoNi@CNTs-C-200 samples consist of polyhedron-like carbonnanocages (200�500 nm) with CoNi alloy nanoparticles wellembedded (Fig. 2a-d). However, some of the carbon nanocages inCoNi@CNTs-C-400 sample are broken and one can notice that

Fig. 2. SEM images of (a, b) CoNi@CNTs-C-100, (c, d)

relatively larger CoNi alloy particles are located outside the carbonnanocages (Fig. 2e-f). This observation indicates that the Niamount plays a crucial role in control of the particle size of CoNialloy and thus might affect structural stability of CoNi@CNTs-Csamples in acidic electrolyte when applied as CEs for DSSCs.

Experiments presented further in the work show that DSSCbased on the CoNi@CNTs-C-200 CE demonstrates the highestenergy conversion efficiency, and therefore we perform a detailed

CoNi@CNTs-C-200 and (e, f) CoNi@CNTs-C-400.

Z. Xie et al. / Electrochimica Acta 229 (2017) 361–370 365

transmission electron microscopy (TEM) and high resolution (HR)TEM to better understand its microstructure and nanostructure, asdisplayed in Fig. 3a-d. It can be clearly observed that theCoNi@CNTs-C-200 sample is composed of numerous CoNi alloynanoparticles with the size of 20 � 100 nm (dark dots), which arewell confined in the carbon nanocages (grey framework).Interestingly, one can observe that some short CNTs (marked byred arrow in Fig. 3b) protrude from the surface of carbonnanocages. Further HRTEM images display that individual CoNialloy nanoparticle is completely enwrapped by CNTs (Fig. 3c, d),and the distinct lattice fringe with the measured d-spacing of0.21 nm can be assigned to the (111) plane of CoNi alloy [26]. It isbelieved that metal alloy possesses better catalytic activity forCNTs growth than single metal catalysts, and thus leads to theformation and protrusion of CNTs during pyrolysis of Ni2+

incorporated ZIF-67 at high temperature in Ar atmosphere.The corresponding energy-dispersive X-ray (EDS) mapping

results further confirm the coexistence of C, N, Co and Ni elementswithin the sample, and both the Co and Ni elements showhomogeneous distribution for all the nanoparticles (Fig. 4 andFig. S2), confirming the CoNi alloy phase, which are well consistentwith previous XRD and TEM results. It has been reported thatproper alloying of transition metals is favorable for electronicperturbation of single metals, and thus the alloy such as CoNiusually demonstrates better catalytic activity than pure Co and Nicatalysts [25,26]. In addition, nitrogen doping has been provedeffective in enhancing catalytic activity and electrical conductivityof carbon materials [37–39]. Therefore, we believe thatCoNi@CNTs-C-200 could be a promising CE material fornew-generation DSSCs. It is also worth noting that chemicalcomposition of the as-prepared CoNi alloy in our work can beeasily tuned by varying the weight ratio of cobalt and nickel

Fig. 3. (a, b) TEM images and (c, d) HRTEM ima

precursors during the synthesis of Ni2+ incorporated ZIF-67nanocrystals. The atomic ratio of Co/Ni in the optimal sampleCoNi@CNTs-C-200 is found to be 9:1 through EDS mapping results,indicating a certain amount of Ni ions can be encapsulated withinZIF-67 nanocrystals during the synthesis process in methanol atroom temperature.

Thanks to such a unique hierarchical nanostructure composedof CoNi@CNTs within the carbon nanocages, the as-preparedCoNi@CNTs-C composites are explored as counter electrodematerials for DSSCs for the first time. The classic assembly of aDSSC device is shown in Fig. 5a, which is mainly composed of TiO2

photoanode with N719 dye, I3�/I�redox electrolyte andCoNi@CNTs-C counter electrode. Fig. 5b displays representativeJ-V curves of DSSCs with Pt, CoNi@CNTs-C-100, CoNi@CNTs-C-200,CoNi@CNTs-C-400 and Co-C as CEs, measured under a lightintensity of 100 mW/cm2. The corresponding photovoltaicperformances are summarized in Table 1. It is found that thecorresponding PCEs of these DSSCs are in the order of CoNi@CNTs-C-400 < Co-C < Pt < CoNi@CNTs-C-100 < CoNi@CNTs-C-200.Notably, the DSSC with conventional Pt as CE yields PCE = 7.88%,Jsc = 15.0 mA/cm2 and Voc = 0.73 V, while the one with CoNi@CNTs-C-200 as CE demonstrates much higher photovoltaic parameterswith PCE = 9.04%, Jsc = 18.30 mA/cm2 and Voc = 0.76 V, showing�15% overall PCE enhancement compared to the Pt counterpart.Such remarkable PCE from the DSSC consisting of CoNi@CNTs-C-200 demonstrates an ideal Pt-free CE for next-generation DSSCs.

To scrutinize the remarkable difference in photovoltaicperformances of these DSSCs, a series of electrochemical measure-ments are carried out, including cyclic voltammetry (CV),electrochemical impedance spectroscopy (EIS) and Tafel polariza-tion. Fig. 6 displays CV curves of different CEs based on Pt,CoNi@CNTs-C-100, CoNi@CNTs-C-200, CoNi@CNTs-C-400 and

ges of CoNi@CNTs-C-200 (optimal sample).

Fig. 4. (a) SEM image of CoNi@CNTs-C-200 (optimal sample) and its corresponding EDS elemental mapping of (b) C, (c) Co and (d) Ni, respectively.

Fig. 5. (a) Scheme of a DSSC configuration using CoNi@CNTs-C as the counter electrode. (b) Photocurrent density-voltage (J-V) curves of DSSCs with Pt, CoNi@CNTs-C-100,CoNi@CNTs-C-200, CoNi@CNTs-C-400 and Co-C, measured under a light intensity of 100 mW/cm2.

366 Z. Xie et al. / Electrochimica Acta 229 (2017) 361–370

Co-NC, recorded at a scan rate of 20 mV/s in acetonitrile solutioncontaining 10.0 mM LiI, 1.0 mM I2 and 0.1 M LiClO4. Two pairs ofoxidation/reduction peaks (Ox-1/Red-1, Ox-2/Red-2) can beobserved for the CV curves of all the CEs, which can be assignedto the following reaction (1) and (2), respectively[40].

I3� + 2e ! 3I� (1)

3I2 + 2e ! 2I3� (2)

The characteristics of Ox-1/Red-1 pair are our focus of analysisas the CEs mainly perform electrocatalytic reduction of the redoxcouples (I3�/I�) [41,42]. The second redox peak in the CV curve is

Table 1Photovoltaic performance of DSSCs with various CEs.

Samples Jsc (mA/cm2) Voc (V) FF h (%)

Co-C 15.5 0.75 0.66 7.63CoNi@CNTs-C-100 18.0 0.73 0.62 8.10CoNi@CNTs-C-200 18.3 0.76 0.65 9.04CoNi@CNTs-C-400 11.8 0.75 0.72 6.36Pt 15.0 0.73 0.72 7.88

Fig. 6. Cyclic voltammetry (CV) curves of the CEs with Pt, CoNi@CNTs-C-100, CoNi@CNTs-C-200, CoNi@CNTs-C-400 and Co-C obtained at a scan rate of 20 mV/s in acetonitrilesolution containing 10.0 mM LiI, 1.0 mM I2 and 0.1 M LiClO4.

Z. Xie et al. / Electrochimica Acta 229 (2017) 361–370 367

attributed to the reduction of molecular iodine (I2) [26,37]. Ingeneral, the peak-to-peak separation (Epp) and peak currentdensity are utilized to study catalytic activities of CEs in DSSCs.The small Epp and high peak current density are two crucialindicators for high electrocatalytic activity [37]. Table 2summarizes the detailed parameters obtained from these CVcurves. Close examination shows that the Epp value of theCoNi@CNTs-C-200 CE is only 267 mV, which is much smaller thanthat of conventional Pt CE (320 mV). In addition, the Epp values ofall the CEs are in the order of CoNi@CNTs-C-200 < CoNi@CNTs-C-100 < Pt < Co-C < CoNi@CNTs-C-400, while the correspondingpeak current densities are in the reversed order: CoNi@CNTs-C-200 > CoNi@CNTs-C-100 > Pt > Co-C > CoNi@CNTs-C-400. There-fore, the above results indicate that CoNi@CNTs-C-200 andCoNi@CNTs-C-100 demonstrate superior electrocatalytic activitiescompared to other CEs. The large differences in electrocatalyticactivities of three CoNi@CNTs-Cs are likely ascribed to the differentchemical composition and particle size of alloy. As observed fromSEM images in Fig. 2, the particle size of alloy in CoNi@CNTs-C-400

Table 2Parameters derived from the Nyquist plots and CV curves.

Samples Rs (V) Rct (V) ZN (V) Epp (mV)

Co-C 16.34 1.34 2.13 301CoNi@CNTs-C-100 15.72 1.52 2.41 275CoNi@CNTs-C-200 15.69 1.38 1.94 267CoNi@CNTs-C-400 16.86 1.81 2.30 305Pt 15.91 2.09 2.89 320

is much larger than CoNi@CNTs-C-100 and CoNi@CNTs-C-200,thereby reducing the surface area and active sites for catalyticreaction. In addition, the breakage of carbon nanocages inCoNi@CNTs-C-400 sample may lead to poor chemical stabilityCoNi alloy in acidic electrolyte compared to the other twoCoNi@CNTs-C samples. Therefore, CNTs-C-400 demonstrates theworst electrocatalytic activities. It is also found that bothCoNi@CNTs-C-100 and CoNi@CNTs-C-200 display better electro-catalytic activities than Co-C, indicating the superiority of alloy

compared to single metal catalysts.To further elucidate the catalytic activity differences of all the

CEs, electrochemical impedance spectroscopy (EIS) measurementsare also carried out to study intrinsic charge transfer and transportproperties by using symmetrical dummy cells. The EIS spectra(Nyquist plots) obtained from varied CEs are shown in Fig. 7a. It canbe seen that Nyquist plots are composed of two semicircles indifferent frequency ranges. The first semicircle (left) in the high-frequency range corresponds to the interfacial charge transferresistance (Rct) between the electrode and electrolyte. The secondsemicircle (right) in the low-frequency range indicates the well-known Nernst mass diffusion impedance (ZN) within the electro-lyte. According to previous reports [36,37], smaller Rct and ZNusually indicate higher catalytic activity and faster mass diffusionof CEs within the electrolyte, respectively. The overall seriesresistances (Rs) of symmetrical dummy cells fabricated with variedCEs is also considered, which is generally attributed to the sum ofresistance of FTO substrate, CE materials as well as electric contact[16,37]. The Nyquist plots are also fitted with an equivalent circuitinset in Fig. 7a and the derived EIS parameters are summarized inTable 2. By comparison, it is found that both Rct and ZN ofCoNi@CNTs-C-200 are smaller than other CEs, indicating the bestcatalytic activity among all the CEs, which is in good accordancewith the CV observation. It has been widely reported that pristinecarbon materials or single Co or Ni metallic nanoparticles usuallydemonstrate limited available catalytic active sites, therebyresulting in unsatisfactory power conversion efficiency of DSSCdevices. However, our work shows that CoNi alloy nanoparticlesenwrapped by CNTs within the unique carbon nanocages can

Fig. 7. (a) Nyquist plots for symmetric cells fabricated with aforementioned samples. (b) The corresponding Tafel polarization curves of symmetric cells.

368 Z. Xie et al. / Electrochimica Acta 229 (2017) 361–370

efficiently promote the charge transfer from metallic nanoparticlesto the 3D conductive carbon frameworks, which may reduce thesurface work function of carbon framework, thereby improvingcatalytic activity of carbon nanocages. It is also demonstrated thatCoNi alloy with proper chemical composition demonstratesobviously enhanced catalytic activity than bare metallic Co, whichis well consistent with previous report [26]. Therefore, chemicalcomposition of CoNi alloy plays a crucial role in the catalyticactivity of a CE.

Tafel polarization is another powerful technique, which isusually utilized to investigate catalytic activity of varied CEs forDSSCs. Therefore, the Tafel polarization curves of all the CEs areobtained by using the symmetrical dummy cells, as displayed inFig. 7b. It is clearly observed that the slopes of both the cathodicand anodic branches of CEs are in the order of CoNi@CNTs-C-200 > CoNi@CNTs-C-100 > Pt > Co-C > CoNi@CNTs-C-400. This re-sult suggests that CoNi@CNTs-C-200 exhibits the highest exchangecurrent density (J0), indicating the best catalytic activity among allthe CEs.

Fig. 8. CV curves of the CEs based on (a) CoNi@CNTs-C-200 and (b) conventional Pt for 30containing 10.0 mM LiI, 1.0 mM I2 and 0.1 M LiClO4.

Taking the above results of CV, EIS and Tafel polarization intoconsideration, CoNi@CNTs-C-200 demonstrates the best catalyticactivity among all the CEs. Therefore, compared to the other CEsincluding conventional Pt, the DSSC based on CoNi@CNTs-C-200CE exhibits the highest PCE of 9.04%, Jsc of 18.3 mA/cm2, and amoderate fill factor (FF) of 0.65, proving as a very promising Pt-freeCE for application in DSSCs. Notably, among iodine-mediatedDSSCs using carbon materials and transition metal compounds[5,25,26,36], the DSSC based on such unique hybrid of CoNi alloyand N-doped carbon network (CoNi@CNTs-C-200) delivers thehighest PCE and Jsc.

To further explore practical application of CoNi@CNTs-C-200 asan alternative Pt-free CE for new-generation DSSCs, its electro-chemical stability is another crucial factor and needs to beseriously considered. Therefore, we fabricate symmetrical cellsusing CoNi@CNTs-C-200 and Pt photo-electrode, respectively, andtest them for over 300 electrochemical cycles via CV characteriza-tion at a scan rate of 50 mV/s (Fig. 8). It is found that both Epp valueand current density of CoNi@CNTs-C-200 show subtle change,

0 cycles. The CV curves are recorded at a scan rate of 50 mV/s in acetonitrile solution

Z. Xie et al. / Electrochimica Acta 229 (2017) 361–370 369

however, the Pt CE displays obvious reduction of peak currentdensity and increased Epp. Therefore, CoNi@CNTs-C-200demonstrates outstanding electrochemical stability compared tocommonly used Pt for DSSCs. The excellent stability could beascribed to the unique hybrid structure of CoNi alloy and N-dopedcarbon framework, whereas both CNTs and carbon framework canefficiently protect the embedded CoNi alloy nanoparticles fromcorrosion in the acidic electrolyte during long-term operation.

4. Conclusions

In summary, we report, for the first time, a facile synthesis ofCoNi alloy@CNTs within carbon nanocages by direct pyrolysis ofNi2+ incorporated ZIF-67 in inert gas atmosphere. It is found thatthe CoNi@CNTs-C-200 CE shows the best catalytic activity towardsthe reduction of I3�, thereby leading to an impressive PCE of 9.04%as a low-cost Pt-free CE for application in next-generation DSSCs,showing �15% overall PCE enhancement compared to theexpensive Pt counterpart (7.88%). More importantly, theCoNi@CNTs-C-200 CE also exhibits much better electrochemicalstability than conventional Pt CE during long-term electrochemicalcycling, which is crucial for practical application inhigh-performance DSSCs. This work may open up a new routefor further study of various metal organic frameworks as templatesin design and development of low-cost, environmentally friendly,highly stable and efficient counter electrodes for DSSCs. Due to itsunique structure and superior properties, the as-preparedCoNi@CNTs-C-200 may also find wide applications in watersplitting, lithium ion batteries, supercapacitors and other fields.

Acknowledgements

The financial support by Research Enhancement Awards (REA)and Research Awards Program (RAP) sponsored by LaSPACE, LSUEconomic Development Assistantship and Chevron InnovativeResearch Fund (CIRS) are greatly acknowledged. We performed thematerials characterizations by using shared instrumentationfacility (SIF) at Louisiana State University (LSU).

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.electacta.2017.01.145.

References

[1] A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo, H. Pettersson, Dye-sensitized solarcells, Chem. Rev. 110 (2010) 6595–6663.

[2] U. Bach, D. Lupo, P. Comte, J.E. Moser, F. Weissortel, J. Salbeck, H. Spreitzer, M.Gratzel, Solid-state dye-sensitized mesoporous TiO2 solar cells with highphoton-to-electron conversion efficiencies, Nature 395 (1998) 583–585.

[3] H.N. Tian, Z. Yu, A. Hagfeldt, L. Kloo, L. Sun, Organic redox couples and organiccounter electrode for efficient organic dye-sensitized solar cells, J. Am. Chem.Soc. 133 (2011) 9413–9422.

[4] M.J. Ju, J.C. Kim, H.J. Choi, I.T. Choi, S.G. Kim, K. Lim, J. Ko, J.J. Lee, I.Y. Jeon, J.B.Baek, H.K. Kim, N-doped graphene nanoplatelets as superior metal-freecounter electrodes for organic dye-sensitized solar cells, ACS Nano 7 (2013)5243–5250.

[5] X.J. Zheng, J. Deng, N. Wang, D.H. Deng, W.H. Zhang, X.H. Bao, C. Li, Podlike N-Doped Carbon Nanotubes Encapsulating FeNi Alloy Nanoparticles: High-performance counter electrode materials for dye-sensitized solar cells, Angew.Chem. Int. Ed. 53 (2014) 7023–7027.

[6] E. Olsen, G. Hagen, S.E. Lindquist, Dissolution of platinum in methoxypropionitrile containing LiI/I2, Sol. Energy Mater. Sol. Cells 63 (2000) 267–273.

[7] H.N. Tsao, J. Burschka, C. Yi, F. Kessler, M.K. Nazeeruddin, M. Gratzel, Influenceof the interfacial charge-transfer resistance at the counter electrode in dye-sensitized solar cells employing cobalt redox shuttles, Energy Environ. Sci. 4(2011) 4921–4924.

[8] H. Tian, Z. Yu, A. Hagfeldt, L. Kloo, L. Sun, Organic redox couples and organiccounter electrode for efficient organic dye-sensitized solar cells, J. Am. Chem.Soc. 133 (2011) 9413–9422.

[9] T.N. Murakami, S. Ito, Q. Wang, M.K. Nazeeruddin, T. Bessho, I. Cesar, P. Liska, R.Humphry-Baker, P. Comte, P. Pechy, M. Grätzel, Highly efficient dye-sensitizedsolar cells based on carbon black counter electrodes, J. Electrochem. Soc. 153(2006) A2255–A2261.

[10] Y. Xue, J. Liu, H. Chen, R. Wang, D. Li, J. Qu, L. Dai, Nitrogen-doped graphenefoams as metal-free counter electrodes in high-performance dye-sensitizedsolar cells, Angew. Chem. 51 (2012) 12124–12127.

[11] S. Hou, X. Cai, H. Wu, X. Yu, M. Peng, K. Yan, D. Zou, Nitrogen-doped graphenefor dye-sensitized solar cells and the role of nitrogen states in triiodidereduction, Energy Environ. Sci. 6 (2013) 3356–3362.

[12] M. Wu, X. Lin, T. Wang, J. Qiu, T. Ma, Low-cost dye-sensitized solar cell based onnine kinds of carbon counter electrodes, Energy Environ. Sci. 4 (2011)2308–2315.

[13] J. Han, H. Kim, D.Y. Kim, S.M. Jo, S.Y. Jang, Water-soluble polyelectrolyte-graftedmultiwalled carbon nanotube thin films for efficient counter electrode of dye-sensitized solar cells, ACS Nano 4 (2010) 3503–3509.

[14] F. Gong, H. Wang, X. Xu, G. Zhou, Z.S. Wang, In situ growth of Co0.85Se andNi0.85Se on conductive substrates as high-performance counter electrodes fordye-sensitized solar cells, J. Am. Chem. Soc. 134 (2012) 10953–10958.

[15] F. Gong, X. Xu, Z. Li, G. Zhou, Z.S. Wang, NiSe2 as an efficient electrocatalyst fora Pt-free counter electrode of dye-sensitized solar cells, Chem. Commun. 49(2013) 1437–1439.

[16] X.D. Cui, W.W. Xu, Z.Q. Xie, Y. Wang, High-performance dye-sensitized solarcells based on Ag-doped SnS2 counter electrodes, J. Mater. Chem. A 4 (2016)1908–1914.

[17] X.D. Cui, W.W. Xu, Z.Q. Xie, Y. Wang, Hierarchical SnO2@ SnS2 counterelectrodes for remarkable high-efficiency dye-sensitized solar cells,Electrochim. Acta 186 (2015) 125–132.

[18] M. Wu, X. Lin, Y. Wang, L. Wang, W. Guo, D. Qi, X. Peng, A. Hagfeldt, M. Grätzel,T. Ma, Economical Pt-free catalysts for counter electrodes of dye-sensitizedsolar cells, J. Am. Chem. Soc. 134 (2012) 3419–3428.

[19] Y. Hou, D. Wang, X.H. Yang, W.Q. Fang, B. Zhang, H.F. Wang, G.Z. Lu, P. Hu, H.J.Zhao, H.G. Yang, Rational screening low-cost counter electrodes for dye-sensitized solar cells, Nat. Commun. 4 (2013) 1583.

[20] L. Wang, Y.T. Shi, H. Zhang, X.G. Bai, Y.X. Wang, T.L. Ma, Iron oxidenanostructures as highly efficient heterogeneous catalysts for mesoscopicphotovoltaics, J. Mater. Chem. A 2 (2014) 15279–15283.

[21] L. Wang, Y.T. Shi, Y.X. Wang, H. Zhang, H.W. Zhou, Y. Wei, S.Y. Tao, T.L. Ma,Composite catalyst of rosin carbon/Fe3O4: highly efficient counter electrodefor dye-sensitized solar cells, Chem. Commun. 50 (2014) 1701–1703.

[22] M.X. Wu, X. Lin, A. Hagfeldt, T.L. Ma, Low-cost molybdenum carbide andtungsten carbide counter electrodes for dye-sensitized solar cells, Angew.Chem. 50 (2011) 3520–3524.

[23] Q.H. Li, J.H. Wu, Q.W. Tang, Z. Lan, P.J. Li, J.M. Lin, L.Q. Fan, Application ofmicroporous polyaniline counter electrode for dye-sensitized solar cells,Electrochem. Commun. 10 (2008) 1299–1302.

[24] J.H. Wu, Q.H. Li, L.Q. Fan, Z. Lan, P.J. Li, J.M. Lin, S.C. Hao, High-performancepolypyrrole nanoparticles counter electrode for dye-sensitized solar cells, J.Power Sources 181 (2008) 172–176.

[25] M. Motlak, N.A.M. Barakat, A.G. Ei-Deen, A.M. Hamza, M. Obaid, O.B. Yang, M.S.Akhtar, K.A. Khalil, NiCu bimetallic nanoparticle-decorated graphene as noveland cost-effective counter electrode for dye-sensitized solar cells andelectrocatalyst for methanol oxidation, Appl. Catal. A 501 (2015) 41–47.

[26] M. Motlak, N.A.M. Barakat, M.S. Akhtar, A.M. Hamza, B.S. Kim, C.S. Kim, K.A.Khalil, A.A. Almajid, High performance of NiCo nanoparticles-doped carbonnanofibers as counter electrode for dye-sensitized solar cells, Electrochim.Acta 160 (2015) 1–6.

[27] M. Hu, J. Reboul, S. Furukawa, L. Radhakrishnan, Y. Zhang, P. Srinivasu, H. Iwai,H. Wang, Y. Nemoto, N. Suzuki, S. Kitagawa, Y. Yamauchi, Direct synthesis ofnanoporous carbon nitride fibers using Al-based porous coordinationpolymers (Al-PCPs), Chem. Commun. 47 (2011) 8124–8126.

[28] W. Cho, Y.H. Lee, H.J. Lee, M. Oh, Systematic transformation of coordinationpolymer particles to hollow and non-hollow In2O3 with pre-definedmorphology, Chem. Commun. 31 (2009) 4756–4758.

[29] W. Xia, R.Q. Zou, L. An, D.G. Xia, S.J. Guo, A metal–organic framework route to insitu encapsulation of Co@Co3O4@C core@bishell nanoparticles into a highlyordered porous carbon matrix for oxygen reduction, Energy Environ. Sci. 8(2015) 568–576.

[30] Y. Hou, J.Y. Li, Z.H. Wen, S.M. Cui, C. Yuan, J.H. Chen, Co3O4 nanoparticlesembedded in nitrogen-doped porous carbon dodecahedrons with enhancedelectrochemical properties for lithium storage and water splitting, NanoEnergy 12 (2015) 1–8.

[31] X.D. Cui, Z.Q. Xie, Y. Wang, Novel CoS2 embedded carbon nanocages by directsulfurizing metal–organic frameworks for dye-sensitized solar cells,Nanoscale 8 (2016) 11984–11992.

[32] L.T. Yu, J. Liu, X.J. Xu, L.G. Zhang, R.Z. Hu, J.W. Liu, L.C. Yang, M. Zhu, Metal-organic framework-derived NiSb alloy embedded in carbon hollow spheres assuperior lithium-ion battery anodes, ACS Appl. Mater. Interfaces (2016), doi:http://dx.doi.org/10.1021/acsami.6b14233.

[33] X.N. Li, A.I. Rykov, B. Zhang, Y.J. Zhang, J.H. Wang, Graphene encapsulatedFexCoy nanocages derived from metal-organic frameworks as efficientactivators for peroxymonosulfate, Catal. Sci. Technol. 6 (2016) 7486–7494.

370 Z. Xie et al. / Electrochimica Acta 229 (2017) 361–370

[34] J.K. Sun, Q. Xu, Functional materials derived from open framework templates/precursors: synthesis and applications, Energy Environ. Sci. 7 (2014)2071–2100.

[35] M.J. Hu, B. Lin, S.H. Yu, Magnetic field-induced solvothermal synthesis of one-dimensional assemblies of Ni-Co alloy microstructures, Nano Res. 1 (2008)303–313.

[36] M.J. Hu, Y. Lu, S. Zhang, S.R. Guo, B. Lin, M. Zheng, S.H. Yu, High yield synthesisof bracelet-like hydrophilic Ni-Co magnetic alloy flux-closure nanorings, J. Am.Chem. Soc. 130 (2008) 11606–11607.

[37] D.H. Deng, L. Yu, X.Q. Chen, G.X. Wang, L. Jin, X.L. Pan, G.Q. Deng, X.H. Bao, Ironencapsulated within pod-like carbon nanotubes for oxygen reduction reaction,Angew. Chem. 52 (2013) 371–375.

[38] J. Deng, L. Yu, D.H. Deng, X.Q. Chen, F. Yang, X.H. Bao, Highly active reduction ofoxygen on a FeCo alloy catalyst encapsulated in pod-like carbon nanotubeswith fewer walls, J. Mater. Chem. A 1 (2013) 14868–14873.

[39] Z.Q. Xie, Z.Y. He, X.H. Feng, W.W. Xu, X.D. Cui, J.H. Zhang, C. Yan, M.A. Carreon, Z.Liu, Y. Wang, Hierarchical sandwich-like structure of ultrafine N-rich porouscarbon nanospheres grown on graphene sheets as superior lithium-ion batteryanodes, ACS Appl. Mater. Interfaces 8 (2016) 10324–10333.

[40] S.N. Yun, A. Hagfeldt, T.L. Ma, Pt-free counter electrode for dye-sensitized solarcells with high efficiency, Adv. Mater. 26 (2014) 6210–6237.

[41] T.Y. Hsieh, T.C. Wei, P. Zhai, S.P. Feng, M. Ikegami, T. Miyasaka, A room-temperature process for fabricating a nano-Pt counter electrode on a plasticsubstrate for efficient dye-sensitized cells, J. Power Sources 283 (2015)351–357.

[42] M.X. Wu, Y.D. Wang, X. Lin, N. Yu, L. Wang, L.L. Wang, A. Hagfeldt, T.L. Ma,Economical and effective sulfide catalysts for dye-sensitized solar cells ascounter electrodes, Phys. Chem. Chem. Phys. 13 (2011) 19298–19301.