Nanoscale

COMMUNICATION

Cite this: Nanoscale, 2018, 10, 4194

Received 12th December 2017,Accepted 15th January 2018

DOI: 10.1039/c7nr09260a

rsc.li/nanoscale

Carbon nanotube aerogel–CoS2 hybrid catalyticcounter electrodes for enhanced photovoltaicperformance dye-sensitized solar cells†

Tao Liu,a Xianmin Mai,b Haijun Chen,a Jing Ren,a Zheting Liu,a Yingxiang Li,a

Lina Gao,a Ning Wang,*a,c Jiaoxia Zhang,d,e Hongcai Hea and Zhanhu Guo *e

The carbon nanotube aerogel (CNA) with an ultra-low density,

three-dimensional network nanostructure, superior electronic

conductivity and large surface area is being widely employed as a

catalytic electrode and catalytic support. Impressively, dye-sensi-

tized solar cells (DSSCs) assembled with a CNA counter electrode

(CE) achieved a maximum power conversion efficiency (PCE) of

8.28%, which exceeded that of the conventional platinum

(Pt)-based DSSC (7.20%) under the same conditions. Furthermore,

highly dispersed CoS2 nanoparticles endowed with excellent

intrinsic catalytic activity were hydrothermally incorporated to

form a CNA-supported CoS2 (CNA–CoS2) CE, which was due to the

large number of catalytically active sites and sufficient connections

between CoS2 and the CNA. The electrocatalytic ability and stabi-

lity were systematically evaluated by cyclic voltammetry (CV),

electrochemical impedance spectra (EIS) and Tafel polarization,

which confirmed that the resultant CNA–CoS2 hybrid CE exhibited

a remarkably higher electrocatalytic activity toward I3− reduction,

and faster ion diffusion and electron transfer than the pure CNA

CE. Such cost-effective DSSCs assembled with an optimized CNA–

CoS2 CE yielded an enhanced PCE of 8.92%, comparable to that of

the cell fabricated with the CNA–Pt hybrid CE reported in our pub-

lished literature (9.04%). These results indicate that the CNA–CoS2CE can be considered as a promising candidate for Pt-free CEs

used in low-cost and high-performance DSSCs.

Introduction

Since the pioneering work of Kistler reported in the 1930s,1

aerogels have attracted great attention in many potential appli-cations, covering a wide range of supercapacitors,2 filtrationand separation,3 biomimetic surfaces4 and artificial muscles.5

So far, a variety of materials, such as silica,6 metals,7 carbon-based materials8,9 and polymers,10 have been prepared asultra-flyweight cellular aerogels. Until now, many carbonmaterials (e.g. carbon dots, carbon nanotubes (CNTs), gra-phene etc.) have been widely used in the fields of photocataly-sis, sensors, renewable energy, etc.11–18 Owing to its ultrahighporosity, exceptional electrical conductivity, and extremelyhigh surface area,19 the carbon nanotube aerogel (CNA)material is theoretically regarded as a promising candidate forcatalytic electrodes and catalytic supports. In fact, our previousstudies have discussed its appliance as the counter electrode(CE) of dye-sensitized solar cells (DSSCs).20 As one of thecrucial components of DSSCs, the CE plays a key role in col-lecting electrons from an external circuit and transferringthem from the CE to the electrolyte, followed by catalyzing thereduction of tri-iodide (I3

−) to iodide (I−).21 Our findings showthat the CNA CE exhibits excellent electrocatalytic activity forreducing I3

−, which facilitates continuous dye regeneration,thus resulting in a significantly higher power conversionefficiency (PCE) than that of the conventional noble metalplatinum (Pt)-based device.20 Our study further evidenced thatimmobilization of Pt nanoparticles onto CNA catalyst supportscould offer higher electrocatalytic activity and the corres-ponding PCE is much higher than that of the cells using theCNA CE alone.22 However, the high price and scarcity of Ptgreatly restrain its widespread applications in DSSCs.Moreover, Pt can be easily decomposed into PtI4 in the electro-lyte traditionally containing an I3

−/I− redox couple, which isdetrimental to devices’ long-term stability performance.23

Therefore, decorating a CNA with economical but efficientfunctional nanoparticles could be recognized as a practicalway to further improve the PCE for the commercialization ofCNA-based DSSCs. More recently, extensive studies have

†Electronic supplementary information (ESI) available. See DOI: 10.1039/c7nr09260a

aSchool of Materials and Energy, State Key Laboratory of Electronic Thin Film and

Integrated Devices, University of Electronic Science and Technology of China,

Chengdu 610054, P.R. China. E-mail: wangn02@foxmail.combSchool of Urban Planning and Architecture, Southwest Minzu University,

Chengdu 610041, P.R. ChinacState Key Laboratory of Marine Resource Utilization in South China Sea,

Hainan University, Haikou 570228, P.R. ChinadNational Demonstration Center for Experimental Materials Science and Engineering

Education, Jiangsu University of Science and Technology, Zhenjiang,

212003 P. R. ChinaeIntegrated Composites Lab (ICL), Department of Chemical & Biomolecular

Engineering, University of Tennessee, Knoxville, Tennessee, 37966, USA.

E-mail: zguo10@utk.edu

4194 | Nanoscale, 2018, 10, 4194–4201 This journal is © The Royal Society of Chemistry 2018

Publ

ishe

d on

16

Janu

ary

2018

. Dow

nloa

ded

by U

nive

rsity

of

Ten

ness

ee a

t Kno

xvill

e on

05/

03/2

018

19:3

5:18

.

View Article OnlineView Journal | View Issue

focused on using transition metal compounds, includingalloys,24,25 nitrides,26,27 sulfides,28,29 phosphides,30,31 car-bides,32,33 and oxides,34,35 as a new class of alternative cata-lytic materials owing to their abundant feedstock, low cost,excellent intrinsic catalytic activity for I3

− reduction and thevast room in morphological tuning.36 Among them, cobaltsulfides (e.g. CoS,37–39 CoS2,

40–42 etc.) have won a prominentposition due to their facile synthesis, nontoxicity, outstand-ing electrocatalytic capability and long-term stability. To date,despite the attractive advantages of and intensive effortsdevoted to cobalt sulfide CEs, their relatively poor electricalconductivity limits their photovoltaic performance.43 In thepresent study, with a facile hydrothermal route, functionalCoS2 nanoparticles are uniformly dispersed on the CNA cata-lytic support (CNA–CoS2), which is developed as the potentialPt-free CE material in DSSCs. The CNA serves as a conductivescaffold supporting the CoS2 catalyst, while preventing itfrom aggregating. Our studies have revealed that the utiliz-ation of CoS2 nanoparticles exhibits great advantages inenhancing the catalytic activity and hence an impressive PCEof cells (8.92%), which is comparable to that of our previouslyreported device based on Pt nanoparticles supported on theCNA (CNA-Pt) CE (9.04%). Reports concerning the fabricationof CoS2 embedded in supporting carbon materials and theirapplication as the CE in the DSSCs are rare, e.g. CoS2–gra-phene composite CE36 and CoS2 embedded carbon nanocageCE.43 The achieved PCE in both these publications, however,is significantly lower than that given in the present work.Therefore, the as-synthesized CNA–CoS2 hybrid CE endowedwith abundant electrocatalytic activity sites and a fast elec-tron transfer pathway has been demonstrated as a promisingcandidate for Pt-free CEs.

Experimental sectionSynthesis of a CNA

A CNA is prepared through sol–gel and freeze drying pro-cesses, involving resorcinol and formaldehyde for the necess-ary condensation as well as sodium carbonate (Na2CO3) asthe catalyst of polymerization reactions.20,22 Briefly, 0.05 gpristine CNTs (Nanocyl NC 7000, offered by Nanocyl,Belgium) were uniformly dispersed in 4.5 mL distilled watervia sonication for 4 h. 0.4 g resorcinol, 0.6 g formaldehydeand 0.003 g Na2CO3 were added to 1.5 mL distilled water toform a homogeneous solution. The well-dispersed CNT sus-pension was poured into the above aqueous solution andthen stirred with a magnetic bar for 1 h. Subsequently, thewet gel was further aged at 80 °C for 72 h to form a chemi-cally cross-linked CNT, followed by freeze-drying to remove allthe water from it. The obtained dry gel is carbonized at1100 °C for 1 h under an Ar atmosphere. After cooling natu-rally, the resultant black ultra-light-weight CNA was washedwith deionized water repeatedly, and dried in a vacuum ovenat 60 °C overnight. The overall shape of the CNA monolithwas controlled by the mold used for gelation.

Preparation of a CNA/CoS2 CE

Typically, 1 mmol CoCl2·6H2O and 20 mmol thiourea were dis-solved in deionized water (50–80 mL). Subsequently, thepristine CNA bulk was transferred into the above solution. Thefinal CoS2 content in the composites was under the control ofthe amount of deionized water. Next, the mixture was sealed in100 mL Teflon-lined stainless steel autoclaves and heated at180 °C for 12 h. After being cooled to room temperature, theobtained black products were washed repeatedly with distilledwater and vacuum dried at 60 °C. The hydrothermal synthesisprocess allowed CoS2 nanoparticles to be immobilized homo-geneously onto the CNA owing to the strong interactioninduced by the large specific surface energy. The fillingcontent of CoS2 in the composite was calculated by thermalgravimetric analysis. Finally, a thin slice of CNA/CoS2 with∼0.6 cm on each side and ∼0.2 cm thick was transferred onprecleaned fluorine-doped tin oxide (FTO)-coated glass (pur-chased from Nippon Sheet Glass Co., Ltd, Japan). The conven-tional Pt/FTO CE via the thermal decomposition of H2PtCl6(30 mM in isopropanol) was also fabricated as a controlexperiment.

Fabrication of dye adsorbed photoanodes

FTO glass with a dimension of 1.5 cm × 1.5 cm was sequen-tially cleaned with soapy water, deionized water, ethanol andacetone by sonication for 20 min. TiO2 photoanodes were fab-ricated on cleaned FTO glass by the traditional doctor-bladetechnique. Typically, solution A (350 mL isopropyl titanate in2.5 mL anhydrous ethanol) was slowly added into solution B(35 mL, 2 M HCl in 2.5 mL anhydrous ethanol) to obtain TiO2

precursor solution, which was then cast onto FTO glass byspin coating at a rotating speed of 2000 rpm for 60 s. Afterannealing at 500 °C for 30 min in air, a 40 nm thick TiO2

compact layer can be obtained. The commercial TiO2 paste(Dyesol, 18NR-T) was diluted with ethanol in a weight ratio of2 : 7. Thereafter, the diluted TiO2 paste was covered onto theas-prepared TiO2 compact layer by the traditional doctor-blademethod and then subjected to the calcination process at125 °C for 5 min, at 325 °C for 5 min, at 375 °C for 5 min, at450 °C for 5 min, and at 500 °C for 15 min in air sequentially.The final TiO2 photoanodes were further immersed in 0.02 MTiCl4 solution at 70 °C for 30 min, and then heated to 500 °Cfor 30 min in air. After the temperature was cooled down toabout 100 °C, they were immersed in a 0.5 mM ethanol solu-tion of the N719 dye ([cis-di(thiocyanato)-N,N′-bis(2,2′-bipyri-dyl-4-carboxylic acid)-4-tetrabutylammonium carboxylate])(Solaronix Co., Ltd, Switzerland) overnight. Finally, the as-prepared dye sensitized photoanodes were scratched into0.16 cm2 and then rinsed with ethanol to remove physically-adsorbed dye molecules.

Construction of DSSC devices

The electrolyte containing 0.05 M I2, 0.5 M LiI, and 0.5 M4-tert-butylpyridine in 8 mL acetonitrile was injected into thespacer between the dye sensitized TiO2 photoanode and CEs

Nanoscale Communication

This journal is © The Royal Society of Chemistry 2018 Nanoscale, 2018, 10, 4194–4201 | 4195

Publ

ishe

d on

16

Janu

ary

2018

. Dow

nloa

ded

by U

nive

rsity

of

Ten

ness

ee a

t Kno

xvill

e on

05/

03/2

018

19:3

5:18

. View Article Online

through the capillary effect. The symmetrical dummy cell forelectrochemical measurement was fabricated by using twoidentical CEs and filled with the electrolyte, similar to thatused in the assembled DSSCs.

Characterization

The aerogel morphology and CNT microstructure were charac-terized by field-emission scanning electron microscopy(FESEM, JSM-7500F, JEOL) equipped with an energy-dispersiveX-ray spectrometer (EDS, NORAN System 7, Thermo Fisher).Further structural analyses were carried out by using a trans-mission electron microscope (TEM, JEOL-2010, JEOL). TheX-ray diffraction (XRD) patterns were collected by using a diffr-actometer (XRD-7000, SHIMADZU) with Cu Kα radiation (λ =1.54 Å) operated at 40 kV and 40 mA. The porosity of theaerogel was quantitatively characterized by nitrogen adsorp-tion/desorption experiments, which was carried out on aQuadrasorb EVO Model QDS-30. The thermal gravimetriccurves were determined with a NETZSCH STA449C. The cyclicvoltammetry (CV) measurements were performed in a threeelectrode test system, which includes Pt wire as the auxiliaryelectrode and Ag/AgCl as the reference electrode dipped in anacetonitrile solution of 10 mM LiI, 1 mM I2 and 0.1 M LiClO4

at a scanning rate of 50 mV s−1 using an electrochemical work-station (CHI660c, CHI Instruments). Tafel polarization curvesof the dummy cells were determined by the same electro-chemical workstation with a scan rate of 10 mV s−1.Photocurrent density–voltage ( J–V) curves of the assembledDSSCs were recorded with a Keithley 2401 source meter underthe illumination of AM 1.5G (100 mW cm−2) solar lightcoming from a AAA solar simulator (Sun 3000 SolarSimulators, ABET Technologies, USA), which was calibrated

using a certified reference Si solar cell (9115DV, Oriel) beforeeach measurement. The incident photon-to-current conversionefficiency (IPCE) was measured as a function of wavelengthfrom 400 to 800 nm using a solar cell quantum efficiencymeasurement system (Oriel QE-PV-SI, Newport).Electrochemical impedance spectroscopy (EIS) measurementswere carried out on symmetric cells at an open circuit voltagebias with an AC modulation amplitude of 10 mV in a frequencyrange from 0.1 MHz to 100 mHz.

Results and discussion

Fig. 1 presents the morphology and structure of CNTs, CNAand CNA–CoS2. SEM characterization shows the masses ofcurled CNTs with no apparent difference in their morphologyand size, Fig. 1A. TEM shown in Fig. 1B reveals an individualCNT with an outer diameter of about 12 nm and an inner dia-meter of ∼6 nm. Fig. 1C shows a digital photograph for a partof the as-prepared CNA with a bulk density of about 40mg cm−3 and a large porosity of >99%. The adsorption/desorp-tion hysteresis loop represents a typical characteristic of three-dimensional porous materials, Fig. S1A.† By fitting the iso-therms to the Brunauer–Emmett–Teller (BET) model, thesurface area of the CNA was estimated to be 482 m2 g−1. Thelight-weight characteristic enables the CNA to stand on thefeelers of Setaria viridis.20 The pore size distribution calculatedby the Barrett–Joyner–Halenda (BJH) method, depicted anarrow peak centered at ∼5.5 nm, Fig. S1B.† From the low-magnification SEM image (Fig. 1D), a small fragment of theCNA exhibits relatively even and smooth surfaces. However,the detailed observations at high magnification indicate that

Fig. 1 The morphology and structure of CNTs, CNA and CNA–CoS2. (A) SEM images of CNTs, (B) TEM image of a single CNT, (C) photograph of thesynthesized CNA monolith, (D–E) SEM images of the CNA with different magnifications, (F) porous architecture inside the CNA bulk materials, (G–H)surface morphology of the broken CNA–CoS2 hybrid at different magnifications. A HRTEM image of CoS2 nanoparticles (inset in (H), scale bar10 nm).

Communication Nanoscale

4196 | Nanoscale, 2018, 10, 4194–4201 This journal is © The Royal Society of Chemistry 2018

Publ

ishe

d on

16

Janu

ary

2018

. Dow

nloa

ded

by U

nive

rsity

of

Ten

ness

ee a

t Kno

xvill

e on

05/

03/2

018

19:3

5:18

. View Article Online

there is a rough surface with an irregular, hierarchical nano-porous structure on the CNA surface, Fig. 1E. The inner struc-ture of the CNA exhibits a self-assembled, interconnected,porous three-dimensional framework with a continuous porediameter ranging from several to tens of nanometers. At theselected zones marked by an ellipse (Fig. 1F), CNTs in theinner part of the CNA are tightly adhered to each other due tothe decomposition of organic solutes at high temperature.Meanwhile, the abundant nanoscale voids not only serve as ahighly conductive scaffold but also provide an extremely highsurface area for electrochemical catalysis. This case is quitedifferent from the pristine CNT samples (Fig. 1A), which ismainly attributed to the dehydration after freeze-drying treat-ment and the carbonization of residual organic species underhigh-temperature heat treatment. The low-resolution SEMimage of the as-prepared CNA–CoS2 hybrid clearly demon-strates that CoS2 nanoparticles are homogeneously decoratedon the surface of the CNA, Fig. 1G. In the high magnificationSEM image of CNA–CoS2 (Fig. 1H), the uniform CoS2 nano-particles without aggregation are highly distributed inside andoutside the CNA by a facile hydrothermal route. The high-

resolution TEM (inset of Fig. 1H) shows clear and uniformlattice fringes with a fixed crystallographic orientation with alattice spacing of 0.27 nm, ascribing to the (2 0 0) planarspacing of the cubic phase CoS2. By using a two-probemeasurement, the electrical resistivity of the CNA is about1.3 × 10−2 Ωm, which is smaller than that of the one beforecarbonization (5.3 × 10−2 Ωm). This could be caused by theabundance of the existing C–C bonding between the carbo-nized CNTs. In addition, compared with the CNA, CNA–CoS2has a larger electrical resistivity (4.1 × 10−2 Ωm) due to the rela-tively high resistance of CoS2.

The typical X-ray diffraction (XRD) pattern is presented inFig. 2. For comparison, the XRD patterns of the CNTs and CNAare given in Fig. 2a and b, respectively. There is no consider-able difference in the peak positions of the CNTs and CNA pro-ducts (see the vertical dashed lines), which implies that theCNT structure is nearly unchanged even after the freeze dryingand high temperature heat treatment. Fig. 2c shows the XRDpattern of the CNA–CoS2 hybrid. Except for the peaks of theCNA, other diffraction peak signals match well with cubic CoS2(JCPDS no. 89-1492), in which the diffraction peaks at 2θ =27.5°, 32.4°, 36.1°, 39.9°, 46.5°, 55.0°, 60.2° and 62.8° can bereadily indexed to the (111), (200), (210), (211), (220), (311),(023) and (321) planes, respectively. No XRD patterns of otherphases such as Co1−xS or cobalt oxides are observed. Theseresults further confirm that CoS2 nanoparticles are success-fully deposited onto CNA bulk materials. EDS analysis(Fig. S2†) reveals that the atomic ratio of Co to S is calculatedto be 1 : 1.98, which coincides with the results revealed by XRDand TEM.

To identify the electrocatalytic activity towards thereduction of I3

− for the as-prepared CEs, CV experiments havebeen carried out on the CNA CE (Fig. 3A(b)) and CNA–CoS2 CE(Fig. 3A(c)) in an iodide-based electrolyte. The CV of the con-ventional Pt CE is also measured as a reference under thesame conditions, Fig. 3A(a). The relatively negative pair corres-ponds to the redox reaction of I3

−/I−. The electrocatalyticactivity for the reduction of I3

− to I− can be visualized by theFig. 2 XRD patterns of (a) CNTs, (b) CNA and (c) the CNA–CoS2 hybrid.

Fig. 3 (A) Cyclic voltammetry (CV) curves of the I3−/I− redox couple for the different CEs, (B) 20 continuous cycle CV scans of the CNA–CoS2

cathode with a scan rate of 50 mV s−1. (a) Pt, (b) CNA, and (c) CNA–CoS2.

Nanoscale Communication

This journal is © The Royal Society of Chemistry 2018 Nanoscale, 2018, 10, 4194–4201 | 4197

Publ

ishe

d on

16

Janu

ary

2018

. Dow

nloa

ded

by U

nive

rsity

of

Ten

ness

ee a

t Kno

xvill

e on

05/

03/2

018

19:3

5:18

. View Article Online

current density of the reduction peak ( Jred) and the oxidationpeak ( Jox) as well as peak to peak splitting (Epp).

44 Comparedwith the Pt film electrode, the dramatically increased catalyticsurface area of the CNA CE provides efficient transfer channelsfor electrons and electrolytes, giving rise to apparentlyincreased current densities. With regard to the CNA–CoS2 CE,the high density and homogeneous distribution of CoS2 nano-particles onto the CNA ensure plenty of catalytically active sitesand the sufficient connection between CoS2 and the CNA con-ductive network. The CNA–CoS2 CE provides multiple advan-tages in driving ion diffusions and collecting electrons fromthe external circuit as well as accelerating electron transfer tocatalytic CoS2 sites, thereby achieving high efficiency electro-catalysts for the reduction of I3

−.36 Compared with the CNACE, the high electrocatalytic capability of the CNA–CoS2 CE isalso reflected in a smaller Epp, suggesting a lower overpotentialfor I3

− reduction43 and hence better redox reaction reversibil-ity.45 In addition, a multi-cycle successive CV test (20 cycles)was conducted to evaluate the electrochemical stability of theCNA–CoS2 electrode in the I3

−/I− electrolyte. No significantchange in the CV curves between the first and the 20th CVcurves was observed (Fig. 3B), illustrating the excellent long-term electrochemical stability of the CNA–CoS2 CE in the I3

−/I−

electrolyte system.The J–V curves of the DSSCs using CEs based on the Pt,

CNA and CNA–CoS2 have been collected in Fig. 4. It should benoted that the thickness of the applied TiO2 photoanode isnearly the same (∼20 μm), as shown in Fig. S3.† Table 1 liststhe main photovoltaic parameters, including short-circuitcurrent density ( Jsc), open-circuit voltage (Voc), fill factor (FF),the maximum value (ηmax) and the average value (ηavg.) of PCE.

They obey the following relationships: PCE ¼ Jsc � Voc � FFPin

,

where Pin is the power of incident light (100 mW cm−2). Thechampion DSSC assembled with the optimized CNA–CoS2 CEexhibits a Jsc of 16.51 mA cm−2, a Voc of 763 mV and a FF of0.71, yielding a ηmax of 8.92%, much higher than those of the

cells fabricated with the individual CNA (8.28%) and Pt CE(7.32%). Compared with the Pt CE, such an enhancement inthe PCE for the CNA–CoS2 one can be mainly attributed to thehigher Jsc, in accordance with the IPCE results (Fig. S4†). Thehigh electrocatalytic capability of CNA–CoS2 affords more I−

through the reduction of I3− to I−, which in turn boosts the

regeneration of dyes and hence the light capturing and photo-electron injecting, thus resulting in the highest Jsc among thethree samples. More significantly, the photovoltaic parametersof the CNA–CoS2 based device are very close to those of ourpublished work on the DSSC based on the CNA-Pt CE ( Jsc =16.57 mA cm−2, Voc = 779 mV, FF = 0.70, ηmax = 9.04%).22 Itshould be noted that the photoanode, dye and electrolytes arethe same in both cases, thus ensuring perfectly comparableDSSCs in the literature to the DSSCs in this study. Hence, theresultant CNA–CoS2 CE can be considered as a promising Pt-free CE for I3

−/I− based DSSCs.It is well-known that the interfacial charge transfer pro-

cesses at the CE/electrolyte interface can also be used to evalu-ate the electrocatalytic behavior of the CEs. Hence, the EISmeasurement is performed using a symmetric cell, which iscomposed of two identical electrodes. Fig. 5A shows the resul-tant Nyquist plots of the dummy cells and the equivalentcircuit is illustrated in the figure as an inset. The semicircle inthe high frequency represents the electron transfer resistance(Rct) at the CE/electrolyte interface, while the semicircle at thelow frequency corresponds to the Nernst diffusion impedance(ZW) of the I−/I3

− redox species. The intercept of the real axis atthe high frequency is the ohmic series resistance (Rs), consist-ing of the external wire resistance, the contact resistance, con-ductivity of CEs and FTO glass.46 The values of Rs and Rct areevaluated by fitting the impedance spectra with the equivalentcircuit, and are listed in Table 2. The Rs value of CNA–CoS2 is alittle larger than that of CNA, in good agreement with the resis-tivity measurements. However, the Rct value of CNA–CoS2 issmaller than that of the CNA, which reflects that the additionof CoS2 speeds up the electron transfer processes that occur atthe interface of the CE and electrolyte, resulting in theincreased reduction rates of I3

−/I− redox couples and FF as aresult, in good agreement with the CV and J–V analysis results.The diffusion coefficient of I3

− is inversely proportional to theNernst diffusion impedance (ZW), which can be achieved con-sidering the diameter of the low-frequency semicircle.Compared with the CNA CE, the CNA–CoS2 one shows asmaller ZW value, which also contributes to the high electro-

Fig. 4 J–V characteristics of the DSSCs assembled with different CEsunder AM 1.5G one sun illumination (100 mW cm−2). (a) Pt, (b) CNA, and(c) CNA–CoS2.

Table 1 Comparison of photovoltaic parameters derived from the J–Vcharacteristics of DSSCs assembled with different CEs

CEs Jsc (mA cm−2) Voc (mV) FF ηmax. (%) ηavg.a (%)

Pt 14.67 750 0.67 7.32 7.12 ± 0.15CNA 16.08 742 0.69 8.28 8.04 ± 0.17CNA–CoS2 16.51 763 0.71 8.92 8.64 ± 0.25CNA-Pt22 16.57 779 0.70 9.04 8.77 ± 0.27

a The average values of the PCE (ηavg.) were obtained from five cells.

Communication Nanoscale

4198 | Nanoscale, 2018, 10, 4194–4201 This journal is © The Royal Society of Chemistry 2018

Publ

ishe

d on

16

Janu

ary

2018

. Dow

nloa

ded

by U

nive

rsity

of

Ten

ness

ee a

t Kno

xvill

e on

05/

03/2

018

19:3

5:18

. View Article Online

catalytic activity. It is also noted that the CNA–CoS2 has alarger CPE than the CNA, which indicates that the added CoS2increases the catalytic surface area, thereby improving the cata-lytic activity. The time constant (τ, also named electron life-time) can be expressed as τ = Rct×CPE/S,

47 where S is the areaof CEs (0.36 cm2). Hence, the τ of CNA–CoS2 is 1.9 × 10−2 s,larger than that of the CNA (1.4 × 10−3 s). To provide furtherevidence on the electrocatalytic capabilities of the CEs, Fig. 5Bshows the Tafel polarization profiles of the dummy cells fabri-cated with two identical CNA and CNA–CoS2 CEs. Theexchange current density ( J0) and limiting diffusion currentdensity ( Jlim) are closely related to the catalytic activities of the

CEs.22 Jlim can be obtained from the current density at lowslopes and high potential regions. Apparently, Fig. 5B(b)shows a higher Jlim than that in Fig. 5B(a), demonstrating thatthe CNA–CoS2 CE has a higher diffusion rate (D) for the I3

− inthe electrolyte than the CNA CE according to eqn (1).13

D ¼ l2nFC

Jlim ð1Þ

where n (n = 2) is the number of electrons involved in thereduction of I3

− at the electrode, l is the spacer thickness, F isthe Faraday constant and C is the I3

− concentration in theredox electrolyte. The calculated D value for CNA–CoS2 and theCNA is 9.3 × 10−6 and 4.7 × 10−6 cm2 s−1, respectively. It isknown that J0 can directly denote the electrocatalytic perform-ance and J0 increases with the decreasing Rct. The anodic orcathodic branches in Fig. 5B(b) exhibit an obvious larger slopethan that in Fig. 5B(a), indicating that the CNA–CoS2 CE pos-sesses a higher J0 than the CNA CE. Overall, the CNA–CoS2 CEexhibits superior electrocatalytic activity on the I3

−/I− redoxreaction, which is in accordance with the CV and EIS results.

Fig. 5 (A) Nyquist plots of the dummy cell fabricated with two identical CEs. Rs: ohmic series resistance; Rct: charge-transfer resistance; CPE:double layer capacitance; ZW: Nernst diffusion impedance. The inset illustrates an equivalent circuit for fitting Nyquist plots. (B) Tafel curves of thesymmetrical cells fabricated with two identical CEs. (a) CNA and (b) CNA–CoS2 CEs.

Table 2 The simulated EIS parameters of the dummy cells fabricatedwith the CNA CE and CNA–CoS2 CE, respectively

CE Rs (Ω cm2) Rct (Ω cm2) ZW (Ω cm2) CPE (F)

CNA 9.31 9.30 3.24 5.43 × 10−5

CNA–CoS2 9.54 6.44 2.27 1.08 × 10−3

Fig. 6 (A) Average pore diameter and pore volume as a function of the CoS2 filler content in CNA–CoS2 CEs. (B) PCE as a function of the CoS2 fillercontent in CNA–CoS2 CEs.

Nanoscale Communication

This journal is © The Royal Society of Chemistry 2018 Nanoscale, 2018, 10, 4194–4201 | 4199

Publ

ishe

d on

16

Janu

ary

2018

. Dow

nloa

ded

by U

nive

rsity

of

Ten

ness

ee a

t Kno

xvill

e on

05/

03/2

018

19:3

5:18

. View Article Online

It is inevitable that the evident increase in the CoS2 fillercontent consequently leads to their aggregation. As a result, aportion of CoS2 nanoparticles cannot contact with the CNA,thereby blocking the electron transfer to the catalytic surfacesites of CoS2. Meanwhile, the aggregation of CoS2 also resultsin an obviously smaller interspace, leading to the reduced poresize and pore volume (Fig. 6A), which is not favorable for theelectrolyte diffusion channel. Nevertheless, the low density ofCoS2 nanoparticles is also not favorable because of thereduced number of catalytically active sites for I3

− reduction.Hence, the content of CoS2 should be well controlled toachieve an efficient electrolyte diffusion channel and highcatalytic activity. As expected, the PCE of the DSSCs with CNA–CoS2 CEs initially increases with the increasing mass percen-tages of CoS2 and then reaches a maximum efficiency of 8.92%when the mass percentage is 11.8% (Fig. S5†), followed by acontinuous decrease of the PCE due to the electrocatalyticactivity loss, as displayed in Fig. 6B.

Conclusions

A novel CNA–CoS2 hybrid as an attractive Pt-free CE materialfor high-efficiency DSSCs was investigated. The CNA endowedwith high electrical conductivity, high porosity, and a largesurface area allows the fast diffusion of an electrolyte into thepores and rapid collection of electrons from the externalcircuit, thereby resulting in a higher PCE (8.28%) than the con-ventional Pt-based device (7.32%). Furthermore, the CNAserves as an ideal electrocatalytic support material in the as-synthesized CNA–CoS2 CE via a facile hydrothermal route.Extensive electrochemical analyses demonstrate that the CNA–CoS2 CE achieves better catalytic ability, and faster iondiffusion and electron transport than the CNA CE. This isbecause the conductive CNA network ensures sufficientcontact with the surface of highly dispersed CoS2, which pro-vides more active catalytic sites. Owing to the encouragingelectrocatalytic activity on the I3

−/I− redox reaction and theexceptional electron-transport pathway, our fabricated DSSCbased on the optimized CNA–CoS2 CE produces an impressivePCE of 8.92%, close to that of the reported CNA-Pt CE basedcell in our published work (9.04%). The CNA–CoS2 also showsan excellent long-term electrochemical stability in the I3

−/I−

electrolyte system as the CEs of DSSCs. Overall, this work mayprovide a feasible way in pursuit of the low-cost and facile syn-thesis of CE materials for the practical applications of DSSCs.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors gratefully acknowledge financial support of thisresearch by the National Natural Science Foundation of China

(No. 51702036, 51472043, 61761016, 51508484 and 51775152),the National Key R&D Plan (2016YFE0126900), SichuanScience and Technology Planning Project (2017GZ0135), theFundamental Research Funds for the Central Universities (No.ZYGX2015KYQD039, ZYGX2015J028, ZYGX2015J032 andZYGX2015Z010) and the Sichuan Thousand TalentedProfessorship Program.

Notes and references

1 S. S. Kistler, Nature, 1931, 127, 741.2 J. Li, X. Wang, Q. Huang, S. Gamboa and P. J. Sebatian,

J. Power Sources, 2006, 158, 784–788.3 M. A. Shannon, P. W. Bohn, M. Elimelech, J. G. Georgiadis,

B. J. Mariñas and A. M. Mayes, Nature, 2008, 452, 301–310.4 L. Qu, L. Dai, M. Stone, Z. Xia and Z. L. Wang, Science,

2008, 322, 238–242.5 A. E. Aliev, J. Oh, M. E. Kozlov, A. A. Kuznetsov, S. Fang,

A. F. Fonseca, R. Ovalle, M. D. Lima, M. H. Haque,Y. N. Gartstein, M. Zhang, A. A. Zakhidov andR. H. Baughman, Science, 2009, 323, 1575–1578.

6 S. S. Prakash, C. J. Brinker, A. J. Hurd and S. M. Rao,Nature, 1995, 374, 439–443.

7 T. A. Schaedler, A. J. Jacobsen, A. Torrents, A. E. Sorensen,J. Lian, J. R. Greer, L. Valdevit and W. B. Carter, Science,2011, 334, 962–965.

8 X. Gui, J. Wei, K. Wang, A. Cao, H. Zhu, Y. Jia, Q. Shu andD. Wu, Adv. Mater., 2010, 22, 617–621.

9 H. Sun, Z. Xu and C. Gao, Adv. Mater., 2013, 25, 2554–2560.10 H. Hu, Z. Zhao, W. Wan, Y. Gogotsi and J. Qiu, ACS Appl.

Mater. Interfaces, 2014, 6, 3242–3249.11 B. Song, T. Wang, H. Sun, Q. Shao, J. Zhao, K. Song,

L. Hao, L. Wang and Z. Guo, Dalton Trans., 2017, 46,15769–15777.

12 C. Lin, L. Hu, C. Cheng, K. Sun, X. Guo, Q. Shao, J. Li,N. Wang and Z. Guo, Electrochim. Acta, 2018, 260, 65–72.

13 T. Liu, K. Yu, L. Gao, H. Chen, N. Wang, L. Hao, T. Li,H. He and Z. Guo, J. Mater. Chem. A, 2017, 5, 17848–17855.

14 Y. Li, B. Zhou, G. Zheng, X. Liu, T. Li, C. Yan, C. Cheng,K. Dai, C. Liu, C. Shen and Z. Guo, J. Mater. Chem. C, 2018,DOI: 10.1039/C7TC04959E, in press.

15 J. Huang, Y. Cao, Q. Shao, X. Peng and Z. Guo, Ind. Eng.Chem. Res., 2017, 56, 10689–10701.

16 C. Wang, M. Zhao, J. Li, J. Yu, S. Sun, S. Ge, X. Guo, F. Xie,B. Jiang, E. Wujcik, Y. Huang, N. Wang and Z. Guo,Polymer, 2017, 131, 263–271.

17 C. Cheng, R. Fan, Z. Wang, Q. Shao, X. Guo, P. Xie, Y. Yin,Y. Zhang, L. An, Y. Lei, J. Ryu, A. Shankar and Z. Guo,Carbon, 2017, 125, 103–112.

18 K. Zhang, G. Li, L. Feng, N. Wang, J. Guo, K. Sun, K. Yu,J. Zeng, T. Li, Z. Guo and M. Wang, J. Mater. Chem. C, 2017,5, 9359–9369.

19 J. Zou, J. Liu, A. S. Karakoti, A. Kumar, D. Joung, Q. Li,S. I. Khondaker, S. Seal and L. Zhai, ACS Nano, 2010, 4,7293–7302.

Communication Nanoscale

4200 | Nanoscale, 2018, 10, 4194–4201 This journal is © The Royal Society of Chemistry 2018

Publ

ishe

d on

16

Janu

ary

2018

. Dow

nloa

ded

by U

nive

rsity

of

Ten

ness

ee a

t Kno

xvill

e on

05/

03/2

018

19:3

5:18

. View Article Online

20 H. Chen, T. Liu, N. Wang, L. Zhao, Q. Zhao, J. Ren, H. Heand H. Lin, Electrochim. Acta, 2015, 174, 871–874.

21 M. Wu, X. Lin, Y. Wang, L. Wang, W. Guo, D. Qi, X. Peng,A. Hagfeldt, M. Grätzel and T. Ma, J. Am. Chem. Soc., 2012,134, 3419–3428.

22 H. Chen, T. Liu, J. Ren, H. He, Y. Cao, N. Wang and Z. Guo,J. Mater. Chem. A, 2016, 4, 3238–3244.

23 J. H. Guo, Y. T. Shi, Y. T. Chu and T. L. Ma, Chem.Commun., 2013, 49, 10157–10159.

24 X. Chen, Q. Tang, B. He, L. Lin and L. Yu, Angew. Chem.,Int. Ed., 2014, 53, 10799–10803.

25 Y. Duan, Q. Tang, J. Liu, B. He and L. Yu, Angew. Chem.,Int. Ed., 2014, 53, 14569–14574.

26 J. Balamurugan, T. D. Thanh, N. H. Kim and J. H. Lee, Adv.Mater. Interfaces, 2016, 3, 1500348.

27 G. R. Li, J. Song, G. L. Pan and X. P. Gao, Energy Environ.Sci., 2011, 4, 1680–1683.

28 S. Shukla, N. H. Loc, P. P. Boix, T. M. Koh, R. R. Prabhakar,H. K. Mulmudi, J. Zhang, S. Chen, C. F. Ng, C. H. A. Huan,N. Mathews, T. Sritharan and Q. Xiong, ACS Nano, 2014, 8,10597–10605.

29 G. Yue, J. Wu, Y. Xiao, M. Huang, J. Lin and J. Y. Lin,J. Mater. Chem. A, 2013, 1, 1495–1501.

30 M. X. Wu, J. Bai, Y. D. Wang, A. J. Wang, X. Lin, L. Wang,Y. Shen, Z. Wang, A. Hagfeldt and T. Ma, J. Mater. Chem.,2012, 22, 11121–11127.

31 M. S. Wu and J. F. Wu, Chem. Commun., 2013, 49, 10971–10973.

32 M. X. Wu, X. Lin, A. Hagfeldt and T. L. Ma, Angew. Chem.,2011, 123, 3582–3586.

33 J. S. Jang, D. J. Ham, E. Ramasamy, J. Lee and J. S. Lee,Chem. Commun., 2010, 46, 8600–8602.

34 Y. Hou, D. Wang, X. H. Yang, W. Q. Fang, B. Zhang,H. F. Wang, G. Z. Lu, P. Hu, H. J. Zhao and H. G. Yang, Nat.Commun., 2013, 4, 1583.

35 X. Lin, M. Wu, Y. Wang, A. Hagfeldt and T. Ma, Chem.Commun., 2011, 47, 11489–11491.

36 X. Duan, Z. Gao, J. Chang, D. Wu, P. Ma, J. He, F. Xu,S. Gao and K. Jiang, Electrochim. Acta, 2013, 114, 173–179.

37 M. Wang, A. M. Anghel, B. Marsan, N.-L. C. Ha,N. Pootrakulchote, S. M. Zakeeruddin and M. Grätzel,J. Am. Chem. Soc., 2009, 131, 15976–15977.

38 S. Peng, T. Zhang, L. Li, C. Shen, F. Cheng, M. Srinivasan,Q. Yan, S. Ramakrishna and J. Chen, Nano Energy, 2015,16, 163–172.

39 C.-W. Kung, H.-W. Chen, C.-Y. Lin, K.-C. Huang, R. Vittaland K.-C. Ho, ACS Nano, 2012, 6, 7016–7025.

40 J.-C. Tsai, M.-H. Hon and I.-C. Leu, Chem. – Asian J., 2015,10, 1932–1939.

41 J. Jin, X. Zhang and T. He, J. Phys. Chem. C, 2014, 118,24877–24883.

42 J.-C. Tsai, M.-H. Hon and I.-C. Leu, RSC Adv., 2015, 5,4328–4333.

43 X. Cui, Z. Xie and Y. Wang, Nanoscale, 2016, 8, 11984–11992.

44 Q. D. Tai, B. L. Chen, F. Guo, S. Xu, H. Hu, B. Sebo andX. Z. Zhao, ACS Nano, 2011, 5, 3795–3799.

45 X. Miao, K. Pan, G. Wang, Y. Liao, L. Wang, W. Zhou,B. Jiang, Q. Pan and G. Tian, Chem. – Eur. J., 2014, 20, 474–482.

46 A. Hauch and A. Georg, Electrochim. Acta, 2001, 46, 3457–3466.

47 Q. Jiang, Y. P. Yeh, N. Lu, H. W. Kuo, M. Lesslie and T. Xu,J. Renewable Sustainable Energy, 2016, 8, 013701.

Nanoscale Communication

This journal is © The Royal Society of Chemistry 2018 Nanoscale, 2018, 10, 4194–4201 | 4201

Publ

ishe

d on

16

Janu

ary

2018

. Dow

nloa

ded

by U

nive

rsity

of

Ten

ness

ee a

t Kno

xvill

e on

05/

03/2

018

19:3

5:18

. View Article Online