Journal of Alloys and Compounds 646 (2015) 32–39

Contents lists available at ScienceDirect

Journal of Alloys and Compounds

journal homepage: www.elsevier .com/locate / ja lcom

One pot synthesis of multi-functional tin oxide nanostructuresfor high efficiency dye-sensitized solar cells

http://dx.doi.org/10.1016/j.jallcom.2015.05.1200925-8388/� 2015 Elsevier B.V. All rights reserved.

⇑ Corresponding author.E-mail addresses: rjose@ump.edu.my, joserajan@gmail.com (R. Jose).

Qamar Wali, Azhar Fakharuddin, Amina Yasin, Mohd Hasbi Ab Rahim, Jamil Ismail, Rajan Jose ⇑Nanostructured Renewable Energy Materials Laboratory, Faculty of Industrial Sciences & Technology, Universiti Malaysia Pahang, 26300, Malaysia

a r t i c l e i n f o

Article history:Received 21 January 2015Received in revised form 12 May 2015Accepted 14 May 2015Available online 8 June 2015

Keywords:PhotovoltaicUV–VIS spectroscopy measurementsTiO2–SnO2 composite structureElectron life time and recombination

a b s t r a c t

Photoanode plays a key role in dye sensitized solar cells (DSSCs) as a scaffold for dye molecules, transportmedium for photogenerated electrons, and scatters light for improved absorption. Herein, tin oxidenanostructures unifying the above three characteristics were optimized by a hydrothermal process andused as photoanode in DSSCs. The optimized morphology is a combination of hollow porous nanoparti-cles of size �50 nm and micron sized spheres with BET surface area (up to 29 m2/g) to allow largedye-loading and light scattering as well as high crystallinity to support efficient charge transport. Theoptimized morphology gave the highest photovoltaic conversion efficiency (�7.5%), so far achieved inDSSCs with high open circuit voltage (�700 mV) and short circuit current density (�21 mA/cm2) employ-ing conventional N3 dye and iodide/triiodide electrolyte. The best performing device achieved an incidentphoton to current conversion efficiency of �90%. The performance of the optimized tin oxide nanostruc-tures was comparable to that of conventional titanium based DSSCs fabricated at similar conditions.

� 2015 Elsevier B.V. All rights reserved.

1. Introduction

The photoanode or working electrode (WE) fabrication offerssignificant challenges in achieving high efficiency dye-sensitizedsolar cells (DSSCs). The WE is usually a mesoporous film of wideband gap metal oxide semiconductor (MOS) and has typically threefunctionalities: (i) to anchor large amount of dyes, (ii) to transportof photogenerated electrons to the collecting electrode (FTO), and(iii) to scatter the light additionally to improve the light absorptionby the solar cells [1]. Mesoporous particles (�20–30 nm) of largesurface area (P100 m2/g) is required for loading large amount ofdyes, [2,3] highly crystalline particles or wires with less defectsare preferred for efficient transport, and larger particles (�200–300 nm) are required for light scattering [4–6].

Tin oxide (SnO2) is a promising WE material in DSSCs owing toits wider band gap (�3.6 eV vs. �3.2 eV of TiO2) and larger electronmobility (ln �100–250 cm2 V�1 s�1) than most frequently usedTiO2 (ln < 0.1 cm2 V�1 s�1) [7]. The wider band gap of SnO2

improves device stability; whereas the UV absorption of TiO2

degrades the dye and considerably reduce the operating hours ofDSSCs [8]. The ln of SnO2 is one of the highest in MOS, even innanocrystalline form, in which form the ln sharply decreases byseveral orders of magnitude. The SnO2 NPs have 50% higher ln

(�3.63 � 10�3 cm2 V�1 s�1) than TiO2 (�2.47 � 10�3 cm2 V�1 s�1)[9,10]. Order of magnitude higher ln is reported in SnO2 nanowiresand flowers than NPs [11]. Many wet-chemical methods arereported for synthesis of various morphologies of SnO2 as WE inDSSCs; [12–15] but produced inferior efficiency (g). For instance,Wang et al. [16] synthesized hollow nanospheres (HNS) andreported g < 1% in pristine SnO2. Liu et al. [17] fabricatedcoral-like porous SnO2 HNS and developed DSSCs with g � 1%.The g in the above studies remarkably increased several folds upona TiCl4 treatment. This enhancement is related to the increase inthe Fermi energy of pristine SnO2 upon TiCl4 treatment, whichotherwise occurs at lower energies than that of TiO2 [18]. Thelower Fermi energy of SnO2 increase the energy loss at theSnO2-dye interface and thereby impose a loss-in-potential at thisinterface and subsequently reduces the open circuit voltage (VOC)and recombination resistance in DSSCs. Furthermore, SnO2 has alow iso-electric point (pH � 4–5) than that of TiO2 (pH � 6–7) soresulting in poor dye-loading [19–21] and consequently lowerthe short circuit current density (JSC).

Herein, we optimized a hydrothermal process, with slight mod-ifications in that used by Wang et al. [16] that developed a series ofSnO2 nanostructures. A temperature dependent growth processshowed low crystallinity but high surface area for SnO2 NPs syn-thesized at a low temperature (�150 �C) which gave inferior per-formance in DSSCs as is conventionally observed. Increase intemperature (�180 �C) lead growth of HNS (�700–800 nm) along

Q. Wali et al. / Journal of Alloys and Compounds 646 (2015) 32–39 33

with small NPs. The particles synthesized at a higher temperature�200 �C showed an optimum mixture of SnO2 NPs and HNS withimproved crystallinity, desirable light scattering and transportproperties. When these nanostructures were tested as a WE inDSSCs, the SnO2 HNS synthesized at �200 �C showed best perfor-mance among the others with a highest g (�4.0%) achieved usingpristine SnO2 till date. The g was increased to �7.5% upon TiCl4

treatment.

2. Experimental section

2.1. Synthesis of SnO2 nanostructures

The materials were synthesized following Wang et al. [16] but with modifica-tions. In our experiment, we reduce the concentration of the growth solution(16.5 mM) to nearly one half than that reported before (32.7 mM) [16]. In a typicalprocedure, SnCl2�2H2O (0.25 g) was added to a mixture of 1 N HCl (0.6 mL), ethanol(6 mL) and DI water (60 mL) ultrasonicated for 1 h. The resultant transparent solu-tion was then transferred to an autoclave and kept at a pre-heated furnace at�150 �C for �24 h. After cooling to room temperature, the solution was centrifuged,washed three times with DI water and dried it at �60 �C in the oven for overnight.This sample was labelled as ‘‘sample A’’. The same procedure was applied for ‘‘sam-ple B’’ and ‘‘sample C’’; however, the furnace temperature was increased to �180 �Cand �200 �C, respectively. The SnO2 HNS were also synthesized using the reportedprocedure [16] to see its macroscopic particle distribution.

2.2. Characterizations

The annealed nanostructures were characterized for morphology, particulateproperties, and crystal structure. Morphology and microstructure of the materialswere studied by scanning electron microscopy (7800F, FESEM, JEOL, USA). TheBET surface area of the materials were measured using gas adsorption studiesemploying Micromeritics (Tristar 3000, USA) instrument in the nitrogen atmo-sphere. High resolution lattice images and selected area diffraction (SAED) patternswere obtained using transmission electron microscope (TEM) operating at 300 kV(FEI, Titan 80–300 kV). Crystal structure of the material was studied by X-raydiffraction (XRD) technique using Rigaku Miniflex II X-ray diffractometer employ-ing Cu Ka radiation (k = 1.5406 Å).

2.3. SnO2 paste preparation

Three hundred milligram of SnO2 was dispersed in ethanol and addeda-terpinol (18 wt.%) and ethyl cellulose (10 wt.%). The above solution was ultrason-icated for 1 h and heated up to �70 �C to evaporate ethanol until viscous slurry wasformed.

2.4. Solar cell fabrications and testing

The FTO substrates were immersed in 0.1 M aqueous TiCl4 solution at�80 �C for�40 min followed by annealing for �30 min at �450 �C and subsequently cooled toroom temperature. The SnO2 pastes was then coated using Doctor–Blade techniqueon the TiCl4 treated FTO substrates (1.5 cm � 1 cm; sheet resistance �18 X sq�1)and then heated at �450 �C for �30 min. Thickness of the sintered electrodes wasstudied by SEM. The thickness of the films was �8.5 lm and the active area ofthe cells was �0.12 cm2. The sintered electrodes were further treated with aqueousTiCl4 solution (0.2 M) for improving the connectivity between the grains as well assuppressing the electron recombination with the tri-iodide species in the elec-trolyte [22]. This post TiCl4 treatments was done by dipping the electrodes in thesolution at �70 �C for �30 min, washed it with DI water to remove the residualTiCl4 and then sintered at �450 �C for �30 min. The sintered electrodes was thensoaked in RuL2 (NCS)2�2H2O (L = 2,2/-bipyridyl-4,40-dicarboxylic acid (N3 dye,Solaronix) (0.3 mM) for �24 h at room temperature. The unanchoreddye-molecules were removed by washing with ethanol. The DSSCs were sealedusing a �25 lm spacer. A Pt-sputtered FTO glass was used as the counter electrode.The electrolyte was acetonitrile containing 0.1 M lithium iodide, 0.03 M iodine,0.5 M 4-tert-butylpyridine and 0.6 M 1-propyl-2,3-dimethyl imidazolium iodide,which was injected through two small openings at the counter electrode.

Absorption and transmission spectra of the dye anchored electrodes as well asdye desorption test were recorded using a UV–vis NIR spectrometer (UV-2600SHIMADZU, Japan). The current–voltage (I–V) characteristics of the assembledDSSCs were studied using a solar simulator (SOLAR LIGHT, Model 16-S 150)employing single port simulator with power supply (XPS 400) at AM1.5 conditions.The I–V curves were obtained using a potentiostat (Autolab PGSTAT30, Eco ChemieB.V., The Netherlands) employing the NOVA� software. The level of standard irradi-ance (100 mW/cm2) was set with a calibrated c-Si reference solar cell. To avoidstray-light effects, devices were properly masked to expose only the WE area. InDSSCs, two types of I–V characteristics measurement are performed, the normal

scan mode where the voltage is changing stepwise from the short circuit (V = 0 orcurrent = ISC) to the open circuit (V = VOC or current = 0), and the reverse scan modeoperate where the voltage sweep from open circuit state (V = VOC) to the short cir-cuit current state (V = 0). Usually reverse scan mode delivered better results thanthat of normal scan mode when the delay time is shorter than the time requiredfor a cell to acquire its equilibrium state. This discrepancy between normal andreveres scan mode could be reduced by taking the delay time an order of magnitudelonger than that required for the silicon based solar cells. During the I–V measure-ment the delay time is necessary after each step of sweep voltage in order to stabi-lize the device. The most obvious characteristics of DSSCs I–V measurement is thetemporal response which is much low as compared to the silicon based solar cells.Therefore, sufficient time is required for DSSCs transit time from short circuit toopen circuit [23,24]. In our experiment, the DSSCs were illuminated for 10 minunder solar cells simulator at room temperature�25 �C to stabilize the temperaturebefore measuring the I–V curve. The applied bias voltage source was swept fromshort-circuit current (ISC, at V = 0) to open circuit (V = VOC, at ISC = 0). Following arethe I–V measurement parameters: current range �0–1 mA, maximum time for opencircuit potential determination �180 s, wait time �5 s, potential range ��0.200–0.900 V, step potential �0.00244 s, scan rate �0.02 V/s and delay time �0.122 s.Measurements were repeated for five times for each DSSCs.

The incident photon to current conversion efficiency (IPCE) measurements wasdone using the Bukoh Keiki (CEP-2000) instrument, Japan. Five sets of devices werefabricated using each SnO2 nanostructures and the measurements were repeatedfor 10 times to assure the consistency in the values.

3. Results and discussions

3.1. Morphological properties

Fig. 1 shows the SEM images of the SnO2 nanostructures syn-thesized at �150, �180, and �200 �C, respectively. They arelabelled as samples A, B, and C, respectively. The SEM image ofparticles synthesized using the increased precursor concentration,as reported in the reference [16] (32.7 mM), is also shown in Fig. 1.More SEM images are shown in Supporting Information (SI)(Fig. S1, SI). Sample A consists of uniform particles of size�100 nm; a closer examination revels that the particles are aggre-gates of <10 nm sized grains. At �180 �C (sample B), a wider dis-tribution of aggregates were observed with size up to �800 nm;however, basic building blocks of each particle were remainedsame (�10 nm). Bigger particles further grow >1 lm at 200 �Cwith relatively large size distribution (Sample C). On the otherhand, the particles synthesized following the increased precursorconcentration showed coarsening (Fig. 1d). Size distribution andproperties of the aggregates were closely analyzed using TEM.Fig. 2a shows typical TEM images of the sample C. Spherical aggre-gates of large size distribution in the range of �50 nm–1 lm wasobserved; particles in each aggregate remain practically constantat �10 nm. All aggregates including the larger ones showed partialtransparency to the electron beam; from which we infer that theaggregates are hollow. Hollow nature of the aggregates could alsobe observed from SEM (Fig. 1c). Crystallinity of the sample wasjudged from the high resolution transmission electron microscopic(HRTEM) lattice images and selected area electron diffraction(SAED) patterns, which are shown in Fig. 2b. More TEM andHRTEM images are presented in SI (Fig. S2, SI). The HRTEM imagesshowed aggregates of defect free nanograins and SAED patternshowed diffraction spots oriented along a circle. These observa-tions show that the particles are of high crystallinity. In theX-ray diffraction (XRD) pattern, sharp and intense peaks of thesample C reveal that they are highly crystalline as observed fromthe HRTEM images and SAED patterns. Smaller particles withsuperior crystallinity are recommended for efficient charge trans-port in DSSCs while anchoring large amount of dyes [25]. Thespecific surface area, pore size and volume distribution were stud-ied by Brunauer–Emmett–Teller (BET) method in nitrogen adsorp-tion and desorption environment. The particulate properties of thesamples from the above study, such as surface area, pore size andvolume distributions are listed in Table 1. The BET surface area were�50,�45, and�29 m2 g�1, respectively for samples A, B, and C. The

Fig. 1. (a) Illustrates SEM images of the synthesized sample A, (b) depicts sample B, while (c and d) illustrate sample C and the reference, respectively.

Fig. 2. (a) TEM images of the nanostructure (sample C) at different magnificationlevel while, (b) show the HRTEM and SAED pattern, respectively. Table 1

BET surface area, pore sizes and volume distribution for the respective synthesizedSample A, B and C.

Sample BET surface area(m2 g�1)

Pore size(nm)

Pore volume(cm3 g�1)

Sample A 50 15.1 0.19Sample B 45 13.1 0.15Sample C 29 11.7 0.09

34 Q. Wali et al. / Journal of Alloys and Compounds 646 (2015) 32–39

lowering of surface area could be due to the formation of aggregatesat high temperature for samples B & C. The high surface area andvarying pore size are beneficial for the DSSCs as they help in largedye loading and help improving permeations of the electrolyte.Adsorption–desorption isotherm of the three samples correspondsto type IV isotherm (Fig. S3, SI). The area under hysteresis loops

increased with processing temperature thereby indicating increas-ing the pore size distribution. The XRD patterns (Fig. 3) show thatall the samples have the similar crystal structure (cassiterite phase,tetragonal crystal system, space group P42/mnm, JCPDS file card #41–1445). The lattice parameters calculated from the XRD patternswere a = 4.7380 Å and c = 3.1865 Å. The sample C has remarkably dif-ferent crystallinity than the others, which was judged through alower value of the full width at half maximum of the diffractionpeaks. High crystallinity is suggested to improve the conductivityof metal oxide semiconductors and improve the photovoltaic prop-erties of DSSCs fabricated using them [25].

The samples were sintered onto transparent conducting glasssubstrates (FTO) duly spin coated with a thin (�500 nm) TiO2 layerfor fabrication of solar cells (Fig. S4, SI). The DSSCs fabricated usingsamples A, B, and C were termed as ADSSCs, BDSSCs, and CDSSCs,respectively. One set of electrodes using the commercial TiO2 pastewas also fabricated and used as a reference cell (PDSSCs).

Fig. 4 shows a typical SEM image (Sample C) showing thecross-section of the electrode of thickness �8.5 lm. The particleswere well sintered onto FTO; a closer examination shows sporadicdistribution of larger particles as well as their shell structure(Fig. 4d). The hollow structures retained their initial morphologyeven after extensive mechanical agitation during the paste makingprocedure and subsequent thermal annealing. No agglomerationwas found in the WE film; ensuring high porosity for electrolytepermeation [26].

3.2. Optical properties of the dye-anchored electrodes

The light harvesting properties of the dye-anchored WE werestudied by UV–Vis absorption spectroscopy. Dye-loading was

10 20 30 40 50 60 70 80

(321

)

(202

)(301

)(1

12)

(310

)(0

02)

(220

)(2

11)

(200

)(1

01)

(110

) Sample CSample BSample A

Inte

nsity

(a.u

)

2θ (deg)

Fig. 3. XRD patterns of the respective three synthesized samples A, B and C.

300 400 500 600 700 800 9000.0

0.2

0.4

0.6

0.8

1.0

Abs

. (a.

u)

λ (nm)

A T R (Electrode A) A T R (Electrode B) A T R (Electrode C)

0

20

40

60

80

100

Tra.

Ref

(%)

Fig. 5. Normalized absorbance, transmittance and reflectance curves of theelectrodes A, B and C.

Q. Wali et al. / Journal of Alloys and Compounds 646 (2015) 32–39 35

measured using desorption test (Fig. S5, SI), which were �278,�203, and �131 nmol/cm2 for electrodes based on sample A, Band C, respectively. Difference in dye-loading is attributed to thedifference in their surface area. The light scattering properties ofthe electrodes were studied by recording their absorbance(Fig. S5a, SI), transmittance (Fig. S5b, SI) and reflectance spectra(Fig. S5c, SI); Fig. 5 compares normalized absorbance, transmit-tance and reflectance of the electrodes A, B and C, respectively.Electrodes of samples A and B showed similar absorbance althoughslightly improved absorbance was observed for the later despite ofits lower dye-loading. Light scattering by the larger particles isresponsible for this increment in absorbance. The electrode of sam-ple C showed larger absorption cross-section, i.e., the area underthe absorbance curve, due to the presence of larger particlesdespite its inferior dye-loading, than that of the other two.

Superior light harvesting property of the electrode C is moreobvious in the transmittance spectra (Fig. S5b in SI and Fig. 5). Asthe size of the particles in sample C corresponds to the wavelengthof the visible (�360–700 nm) and near-infrared (�700 nm–�2.5 lm) regions a strong light scattering could be expected.

Fig. 4. FESEM cross section of the fabricated DSSCs

Moreover, the presence of micron and mesoporous sized NPswould increase the reflection of light and eventually enhance theoptical path length for incident photons [27]. The transmittanceof the electrodes were �34, �30, and <10% at the dye’s absorptionwavelength range for electrodes A, B, and C, respectively. Thus,�90% of the incident light is absorbed by the electrode C, whereasconsiderable portion of the incoming light is transmitted in theother electrodes. In a diffuse reflection (Figs. 5 and S5c), incidentphotons are reflected in all directions by the photoanode therebyincreasing the light harvesting efficiency. If the internal diffusereflection is taken into consideration, electrode C has a higherreflectance >700 nm owing to its comparable size together withmacro and mesopores, which is expected to increase the incidentlight reflected and scattered inside the photoanode. However, ifwe take external reflectance into account, the electrode C pos-sesses lower reflectance in the 350–700 nm range than electrodesA and B. This low reflectance of the electrode C reveals high absor-bance; and therefore, leads to high photovoltaic parameters.

based on sample C at various magnifications.

36 Q. Wali et al. / Journal of Alloys and Compounds 646 (2015) 32–39

3.3. Photovoltaic characteristics of SnO2 DSSCs

Fig. 6 shows the photocurrent density (J)–photovoltage (V)characteristics of the best performing devices; a statistics of thePV parameters of all tested devices are listed in Table 2. The g ofthe DSSCs increased in the order gADSSCs < gBDSSCs < gCDSSCs <gPDSSCs. The superior g of the PDSSCs arises from its �80% higherVOC and �50% higher fill factor (FF) (Table 2) which are due tolower loss-in-potential at the dye-TiO2 and dye-electrolyte inter-faces of PDSSCs [28] compared to the other devices. However, byusing perylene dyes [29] with lower unoccupied molecular orbitalenergies and cobalt based electrolyte [30] the loss-in-potential atthe dye-SnO2 and dye-electrolyte interfaces could be reduced con-siderably thereby achieving high VOC and FF in SnO2 DSSCs. TheCDSSCs showed the highest JSC compared to A & BDCSSs whichcould be attributed to the increased light scattering of the elec-trode C. The CDSSCs showed the highest g � 4.0% (JSC � 16mA cm�2, VOC = 491 mV and FF � 0.50) so far achieved using pris-tine SnO2 nanostructures. The performance of the CDSSCs was fur-ther improved by TiCl4 treatment (TCDSSCs); which showed �80%higher g (�7.5%) than the parent device. Although the TCDSSCsgave a higher JSC (�17%) than the PDSSCs lower FF (�20%) of theformer restricts its performance beyond that of PDSSCs. However,compared with A & BDSSCs, the CDSSCs offered improved FF, whichcould be attributed to the widening of its pore-size distribution asreported by Chen et al. [31].

A systematic increase in VOC (VOC(CDSSCs) > VOC(BDSSCs) >VOC(ADSSCs)) was observed in the three set of devices despite thechemical similarity of the WE material. This increase in VOC couldbe due to increased light scattering achieved using varying particlesize and crystallinity. The high VOC and the suppression of electronrecombination by TiCl4 blocking layer were evaluated from thecharge transport parameters by electrochemical impedance spec-troscopy (EIS). The EIS is a powerful tool used to analyze the chargetransport through the WE material, electron transfer and recombi-nation at the dye anchored WE/electrolyte interface, charge trans-fer at the counter electrode, and the ions diffusion of theelectrolyte in DSSCs [32]. The EIS curves of the three devices, viz.CDSSCs, TCDSSCs and PDSSCs, and the descriptive model to elabo-rate the frequency response in the mentioned processes of DSSCsare presented in the SI (Fig. S6, SI). The transmission line showsparallel channels representing the transport of electrons throughthe MOS and redox species in the electrolyte. The charge recombi-nation occurs at the electrode–electrolyte interface. If L is thethickness of MOS film, the electron transport resistance, RT = rtL,

0

5

10

15

20

J (m

A/c

m2 )

TCDSSCs PDSSCs CDSSCs BDSSCs ADSSCs

0.0 0.2 0.4 0.6 0.8

Potential (V)

Fig. 6. J–V characteristics curve of five devices ADSSCs, BDSSCs, CDSSCs, TCDSSCsand PDSSCs under 1 sun condition, respectively.

interfacial charge recombination resistance, RCT = rct/L where allthese parameters are defined in the reference [32]. The Nyquistplots of DSSCs display three regions. The first small circle ofNyquist plot represents the charge transport resistance at thehigher frequency regime (>1000 Hz), whilst the secondsemi-circle at the middle frequency (1000 Hz < f < 1 Hz) symbolizethe electrons recombination at the dye anchored WE/electrolyteinterface. The low frequency (<1 Hz) reveals the diffusion of ionsin the electrolyte which reflects in the third semi-circle. TheNyquist plot for TCDSSCs shows larger diameter than the othertwo as shown in Fig. 7, thereby, assures the suppression of elec-trons recombination resistance and consequently, long electron lifetime (sn). To further identify the high performance of TCDSSCs, weextracted charge transport parameters using Z-view software ofBisquert transmission line model [32]. TCDSSCs exhibit outstand-ing performance which obeys the ideal condition RCT� RT of anyhigh performing DSSCs. The calculated RCT were �19, �145 and�70 O, while RT (�5, �10 and �43 O), observed for CDSSCs,TCDSSCs and PDSSCs at 0.7 V, respectively. The large semi-circlefor TCDSSCs showed increase in RCT and consequently increasedsn according to the relation, sn = RCT � Cl, where Cl is the chemicalcapacitance [32]. The sn was calculated for the respective DSSCsusing mid frequency of the Bode-Phase plots using sn = 1/2pfo

where fo is the maximum frequency at the mid peak. The calcu-lated values of fo were �21.98, �4.14 and �16.26 Hz, with corre-sponding sn �7, �38 and �10 ms for CDSSCs, TCDSSCs andPDSSCs, respectively. Long sn implies that [33] electrons could sur-vive for long time before recombination, therefore, leads to high JSC

and VOC, respectively.

3.4. Incident photon to current conversion efficiency

The difference in collection efficiencies of the DSSCs were stud-ied from their incident photon to current conversion efficiency(IPCE) spectra also called external quantum efficiency (EQE). TheIPCE could be defined as the number of electrons generated inthe external circuit by light divided by the number of photonshit the cell. Fig. 8a compares the IPCE spectra of the CDSSCs,TCDSSCs and PDSSCs, whereas Fig. 8b compares the action spectrawith absorbed photons to converted electrons (APCE) or internalquantum efficiency (IQE) and reflectance of electrode C, respec-tively. The improved PV performance and long sn in TCDSSC is alsoreflected from its IQE (APCE) spectrum, which is defined as theratio between number of photoelectrons collected in a solar celland the number of photons absorbed. The IQE (APCE) calculatedusing the relation EQE = IQE � R � T, and the spectrum showsnearly unity quantum yield in TCDSSCs for generated and collectedcharge carriers at �520 nm wavelength. The spectra displayed sig-nificant enhancement in the IPCE of the TCDSSCs devices comparedto the other two. The highest IPCE for the devices were �90%,�78%, and �72% at k � 520 nm near the peak absorbance of theN3 sensitizer for all types of DSSCs. The improvement in TCDSSCscan be attributed to the enhanced light scattering efficiency ofthe particles composing variable particle sizes despite its relativelylow surface area and subsequent inferior dye-loading.Furthermore, the J–V data was validated from the IPCE measure-ments of the DSSCs by calculating the JSC by the relation

IPCE ð%Þ ¼ JSCðmA=cm2ÞPðmW=cm2Þ �

1240kðnmÞ � 100%

The integrated IPCE over the entire wavelength (k � 300–800 nm) was used to calculate the JSC as depicted in Fig. 8a. The cal-culated JSC of the CDSSCs, TCDSSCs and PDSSCs were �14, �16 and�15 mA/cm2, respectively, which agree with their measured JSC

(Table 2). Integrated IPCE curves of C, TC and PDSSCs, from whichthe JSC is evaluated, are depicted (Fig. S7, SI).

Table 2The photovoltaic characteristics (JSC, VOC, FF and g) along with dye loading of the five respective A & BDSSCs, CDSSCs, TCDSSCs and PDSSCs.

Electrode JSC (mA/cm2) VOC (V) FF (%) g (%) Dye loading n-mole/cm2

Mean Best Mean Best Mean Best Mean Best

ADSSCs 6.7 7.2 0.35 0.37 0.36 0.38 0.85 1.02 278BDSSCs 10.1 11.1 0.39 0.40 0.34 0.37 1.34 1.62 203CDSSCs 14.7 16.3 0.46 0.49 0.48 0.50 3.34 4.0 131TCDSSCs 19.7 21.3 0.69 0.71 0.48 0.50 6.56 7.5 –PDSSCs 16.4 17.5 0.72 0.731 0.61 0.63 7.21 8.04 259

Fig. 7. Nyquist plot illustration shown in (a) the normalized Bode Phase diagram depicts in (b) while the Bode Phase illustrate in (c) for three respective devices CDSSCs,TCDSSCs and PDSSCs.

0

400 500 600 700 8000

20

40

60

80

100

IPC

E (%

)

λ (nm)λ

CDSSCsTCDSSCsPDSSCs

Fig. 8a. Incident photon to current conversion efficiency of the three devicesCDSSCs, TCDSSCs and PDSSCs, respectively.

400 500 600 700 8000

20

40

60

80

100

IQE-

-EQ

E--R

efle

ctan

ce (%

)

λ (nm)

Electrode C (IPCE)Electrode C (APCE)Electrode C (R)

Fig. 8b. IPCE, absorbed photon to current converted electrons (APCE) andreflectance of the electrode C.

Q. Wali et al. / Journal of Alloys and Compounds 646 (2015) 32–39 37

0.0 0.2 0.4 0.6 0.80.01

0.1

1

10

100CDSSCsTCDSSCsPDSSCs

τ n (s)

Voltage (V)

Fig. 9. Electrons life time from OCVD measurement for the respective, CDSSCs,TCDSSCs and PDSSCs.

38 Q. Wali et al. / Journal of Alloys and Compounds 646 (2015) 32–39

3.5. Open circuit voltage decay measurement

In order to investigate and validate the long sn of the TCDSSCs,open circuit voltage decay (OCVD) measurements were performed[34]. The OCVD measures the temporal decay of VOC upon remov-ing the illumination source in DSSC operating at steady statethereby providing a real time measurement of the charge transportparameters through the device [35,36]. As OCVD measurement isperformed in the dark; no electrons could recombine with oxidizeddye molecules, therefore, electrons solely recombine with thetri-iodide species of the electrolyte. The recombination kineticsof electrons can also verified by plotting sn against the VOC asshown in Fig. 9 by the following equation [34].

sn ¼ �kBT

e

� �dVOC

dt

� ��1

where kBT is the thermal energy e is the elementary positive chargeand dVOC/dt represent the decaying of VOC with respect to time.Measured OCVD curves of the three types of devices are in SI(Fig. S8). Fig. 9 compares the sn of the devices measured from theOCVD curves. Interestingly, TCDSSCs exhibit remarkably increasedsn over the entire voltage range as compared to the other twodevices, thereby providing a real time measurement evidence forincreased carrier lifetime in the TCDSSCs. At a typical voltage of�300 mV, the calculated sn for the CDSSCs, TCDSSCs and PDSSCswere �0.7 s, �15 s and �1.5 s, respectively, which clearly validatedthe high performance of the TCDSSCs device.

4. Conclusions

In conclusion, three SnO2 morphologies were synthesized at dif-ferent temperatures (�150, �180, and �200 �C) by hydrothermalmethod and employed as photoanodes in DSSCs. The material syn-thesized at �200 �C unify three functionalities, viz. moderate sur-face area for dye-loading, presence of larger particles for lightscattering, and high crystallinity for efficient charge transport toenable fabrication of high performance DSSCs. The smaller NPs(�50 nm) in the sample synthesized at �200 �C enabled loadingof large amount of dyes and presence of scattering particles (upto 1 lm) in it resulted high absorbance (�90%) by the photoan-odes. These properties enabled collection of �90% of photogener-ated electrons from �8.5 lm thick photoanode film. Highphotovoltaic conversion efficiency of �7.5% is obtained using theSnO2 nanostructures with a VOC � 700 mV and JSC � 21 mA/cm2.

The DSSCs fabricated using the optimized SnO2 morphologyexcelled the performance of TiO2 based devices despite the latter’ssuperior dye-loading. These achievements are particularly inter-esting using the N3 dye and iodide/triiodide electrolyte as theyare not good choice for SnO2 DSSCs because conduction band andvalence band energies of SnO2 are much lower than that of TiO2

which increases the over potential and charge recombination.

Appendix A. Supplementary material

‘‘More transmission electron microscopic images, electrodediffraction pattern, FESEM images showing the cross-section ofthe working electrode films, adsorption-desorption plots, absorp-tion spectra of the desorbed dye-solution, EIS data and OCVDcurve’’ these material are available free of charge via the Internetat http://www.elsevier.com.

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.jallcom.2015.05.120.

References

[1] M. Grätzel, Solar energy conversion by dye-sensitized photovoltaic cells, Inorg.Chem. 44 (2005) 6841–6851.

[2] Q. Zhang, G. Cao, Nanostructured photoelectrodes for dye-sensitized solarcells, Nano Today 6 (2011) 91–109.

[3] J.-H. Yum, R. Humphry-Baker, S.M. Zakeeruddin, M.K. Nazeeruddin, M. Grätzel,Effect of heat and light on the performance of dye-sensitized solar cells basedon organic sensitizers and nanostructured TiO2, Nano Today 5 (2010) 91–98.

[4] R. Jose, V. Thavasi, S. Ramakrishna, Metal oxides for dye-sensitized solar cells, J.Am. Ceram. Soc. 92 (2009) 289–301.

[5] I.G. Yu, Y.J. Kim, H.J. Kim, C. Lee, W.I. Lee, Size-dependent light-scatteringeffects of nanoporous TiO2 spheres in dye-sensitized solar cells, J. Mater. Chem.21 (2011) 532–538.

[6] C.S.D. Qifeng Zhang, Park Kwangsuk, Liu Dawei, Zhou Xiaoyuan, Jeong Yoon-Ha, Caoa Guozhong, Light scattering with oxide nanocrystallite aggregates fordye-sensitized solar cell application, Nanophotonics 4 (2010) 041540.

[7] G. Shang, J. Wu, S. Tang, L. Liu, X. Zhang, Enhancement of photovoltaicperformance of dye-sensitized solar cells by modifying tin oxide nanorodswith titanium oxide layer, J. Phys. Chem. C 117 (2013) 4345–4350.

[8] E. Ramasamy, J. Lee, Ordered mesoporous SnO2-based photoanodes for high-performance dye-sensitized solar cells, J. Phys. Chem. C 114 (2010) 22032–22037.

[9] N.K. Elumalai, R. Jose, P.S. Archana, V. Chellappan, S. Ramakrishna, Chargetransport through electrospun SnO2 nanoflowers and nanofibers: role ofsurface trap density on electron transport dynamics, J. Phys. Chem. C 116(2012) 22112–22120.

[10] S. Gubbala, H.B. Russell, H. Shah, B. Deb, J. Jasinski, H. Rypkema, M.K. Sunkara,Surface properties of SnO2 nanowires for enhanced performance with dye-sensitized solar cells, Energy Environ. Sci. 2 (2009) 1302–1309.

[11] E.N. Kumar, R. Jose, P.S. Archana, C. Vijila, M.M. Yusoff, S. Ramakrishna, Highperformance dye-sensitized solar cells with record open circuit voltage usingtin oxide nanoflowers developed by electrospinning, Energy Environ. Sci. 5(2012) 5401–5407.

[12] V. Ganapathy, E.-H. Kong, Y.-C. Park, H.M. Jang, S.-W. Rhee, Cauliflower-likeSnO2 hollow microspheres as anode and carbon fiber as cathode for highperformance quantum dot and dye-sensitized solar cells, Nanoscale 6 (2014)3296–3301.

[13] H. Song, K.-H. Lee, H. Jeong, S.H. Um, G.-S. Han, H.S. Jung, G.Y. Jung, A simpleself-assembly route to single crystalline SnO2 nanorod growth by orientedattachment for dye sensitized solar cells, Nanoscale 5 (2013) 1188–1194.

[14] C. Gao, X. Li, B. Lu, L. Chen, Y. Wang, F. Teng, J. Wang, Z. Zhang, X. Pan, E. Xie, Afacile method to prepare SnO2 nanotubes for use in efficient SnO2–TiO2 core–shell dye-sensitized solar cells, Nanoscale 4 (2012) 3475–3481.

[15] S. Fujihara, T. Maeda, H. Ohgi, E. Hosono, H. Imai, S.-H. Kim, Hydrothermalroutes to prepare nanocrystalline mesoporous SnO2 having high thermalstability, Langmuir 20 (2004) 6476–6481.

[16] H. Wang, B. Li, J. Gao, M. Tang, H. Feng, J. Li, L. Guo, SnO2 hollow nanospheresenclosed by single crystalline nanoparticles for highly efficient dye-sensitizedsolar cells, CrystEngComm 14 (2012) 5177–5181.

[17] J. Liu, T. Luo, T.S. Mouli, F. Meng, B. Sun, M. Li, J. Liu, A novel coral-like porousSnO2 hollow architecture: biomimetic swallowing growth mechanism andenhanced photovoltaic property for dye-sensitized solar cell application,Chem. Commun. 46 (2010) 472–474.

[18] S. Chappel, S.-G. Chen, A. Zaban, TiO2-coated nanoporous SnO2 electrodes fordye-sensitized solar cells, Langmuir 18 (2002) 3336–3342.

[19] H.P. Yi-Zhou Zhang, Sun Yanqiu, Lai Wen-Yong, Wei Ang, Huang Wei, Poroustin oxide nanoplatelets as excellent-efficiency photoelectrodes and gassensors, Int. J. Electrochem. Sci. 8 (2013) 3371.

Q. Wali et al. / Journal of Alloys and Compounds 646 (2015) 32–39 39

[20] J. Gong, J. Liang, K. Sumathy, Review on dye-sensitized solar cells (DSSCs):fundamental concepts and novel materials, Renew. Sust. Energy Rev. 16 (2012)5848–5860.

[21] H. Zheng, Y. Tachibana, K. Kalantar-zadeh, Dye-sensitized solar cells based onWO3, Langmuir 26 (2010) 19148–19152.

[22] S. Ito, T.N. Murakami, P. Comte, P. Liska, C. Grätzel, M.K. Nazeeruddin, M.Grätzel, Fabrication of thin film dye sensitized solar cells with solar to electricpower conversion efficiency over 10%, Thin Solid Films 516 (2008) 4613–4619.

[23] X. Yang, M. Yanagida, L. Han, Energy Environ. Sci. 6 (1) (2013) 54–66.[24] Y. Hishikawa, M. Yanagida, N. Koide, Presented at the Photovoltaic Specialists

Conference, 2005. Conference Record of the Thirty-first IEEE, 2005(unpublished).

[25] P.S. Archana, E. Naveen Kumar, C. Vijila, S. Ramakrishna, M.M. Yusoff, R. Jose,Random nanowires of nickel doped TiO2 with high surface area and electronmobility for high efficiency dye-sensitized solar cells, Dalton Trans. 42 (2013)1024–1032.

[26] Q. Wali, A. Fakharuddin, I. Ahmed, M.H. Ab Rahim, J. Ismail, R. Jose,Multiporous nanofibers of SnO2 by electrospinning for high efficiency dye-sensitized solar cells, J. Mater. Chem. A 2 (2014) 17427–17434.

[27] M. Grätzel, Recent advances in sensitized mesoscopic solar cells, Acc. Chem.Res. 42 (2009) 1788–1798.

[28] B.E. Hardin, H.J. Snaith, M.D. McGehee, The renaissance of dye-sensitized solarcells, Nat. Photon. 6 (2012) 162–169.

[29] S. Ferrere, A. Zaban, B.A. Gregg, Dye sensitization of nanocrystalline tin oxideby perylene derivatives, J. Phys. Chem. B 101 (1997) 4490–4493.

[30] A. Yella, H.-W. Lee, H.N. Tsao, C. Yi, A.K. Chandiran, M.K. Nazeeruddin, E.W.-G.Diau, C.-Y. Yeh, S.M. Zakeeruddin, M. Grätzel, Porphyrin-sensitized solar cellswith cobalt (II/III)–based redox electrolyte exceed 12 percent efficiency,Science 334 (2011) 629–634.

[31] D. Chen, F. Huang, Y.-B. Cheng, R.A. Caruso, Mesoporous anatase TiO2 beadswith high surface areas and controllable pore sizes: a superior candidate forhigh-performance dye-sensitized solar cells, Adv. Mater. 21 (2009) 2206–2210.

[32] J. Bisquert, Theory of the impedance of electron diffusion and recombination ina thin layer, J. Phys. Chem. B 106 (2002) 325–333.

[33] A. Fakharuddin, P.S. Archana, Z. Kalidin, M.M. Yusoff, R. Jose, Standardization ofphotoelectrode area of dye-sensitized solar cells, RSC Adv. 3 (2013) 2683–2689.

[34] A. Zaban, M. Greenshtein, J. Bisquert, Determination of the electron lifetime innanocrystalline dye solar cells by open-circuit voltage decay measurements,ChemPhysChem 4 (2003) 859–864.

[35] P.S. Archana, A. Gupta, M.M. Yusoff, R. Jose, Tungsten doped titanium dioxidenanowires for high efficiency dye-sensitized solar cells, Phys. Chem. Chem.Phys. 16 (16) (2014) 7448–7454.

[36] J. Bisquert, Physical electrochemistry of nanostructured devices, Phys. Chem.Chem. Phys. 10 (1) (2008) 49–72.