Available online at www.sciencedirect.com

www.elsevier.com/locate/solener

ScienceDirect

Solar Energy 109 (2014) 11–23

Size and concentration effects of gold nanoparticles on opticaland electrical properties of plasmonic dye sensitized solar cells

Nikhil Chander a,⇑, A.F. Khan a,b, Eshwar Thouti a, Sanjay K. Sardana a,P.S. Chandrasekhar a, Viresh Dutta a, Vamsi K. Komarala a

a Photovoltaic Laboratory, Centre for Energy Studies, Indian Institute of Technology Delhi, New Delhi 110 016, Indiab Department of Electronics and Information Technology, Ministry of Communications & Information Technology,

Government of India, New Delhi 110 003, India

Received 31 December 2013; received in revised form 27 June 2014; accepted 6 August 2014

Communicated by: Associate Editor H. Upadhyaya

Abstract

Gold nanoparticles (GNPs) of various sizes (range of 5–85 nm) were synthesized and various concentrations (range of 0.1–0.7 wt%)were blended with TiO2 nanopowder for fabricating conformal TiO2–Au nanocomposite (NC) films. In optical and electrical studies, wehave observed that GNPs of sizes in the range of 15–40 nm, and concentrations in the range of 0.1–0.25 wt% offer the maximum enhance-ment in dye-sensitized solar cell (DSSC) performance due to the enhanced near-field excitation of dye molecules along with incident lightfar-field. The best plasmonic DSSC performance was observed with 0.24 wt% of �36 nm GNPs with an enhancement of 18.44% in pho-tocurrent. Despite the strong absorptance with �5 nm GNPs, only a modest improvement in photovoltaic behavior was observed due toplasmonic heating effects of strongly localized near-fields instead of dye molecules excitation. With �85 nm GNPs, we have observedminimal enhancement in device performance due to large scattering cross-sections, which result in the incident energy to be sent backto the far-field after interacting with GNPs instead of localizing around them. The optimized size and concentration of GNPs were alsoused for fabricating high efficiency DSSCs using commercial TiO2 paste and two different dyes (N719 and N749) in order to study theeffects of apparent extinction coefficients of the dyes as well as device thickness on photocurrent and energy conversion efficiencyenhancements of DSSCs.� 2014 Elsevier Ltd. All rights reserved.

Keywords: Plasmonics; Dye sensitized solar cell; Gold nanoparticles; Optical properties

1. Introduction

Dye-sensitized solar cell (DSSC) is a third generationphotovoltaic technology, which has a potential to signifi-cantly lower the cost for generating electricity from solarenergy (Gratzel, 2001). The sensitized cell configuration is

http://dx.doi.org/10.1016/j.solener.2014.08.011

0038-092X/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Tel.: +91 11 2659 1255; fax: +91 11 2659 1251.E-mail addresses: nikhilphysics@gmail.com (N. Chander), vamsi@ces.

iitd.ac.in (V.K. Komarala).

witnessing a surge in the research activity with the adventof perovskite sensitized solar cells, which are essentiallyan extension of the dye sensitized concept (Im et al.,2011; Lee et al., 2012; Burschka et al., 2013). But the effi-ciencies of DSSCs are still lower than other thin film tech-nologies, and much lower than crystalline silicon solarcells. To improve the efficiency of DSSC, one needs toenhance the light absorption and improve the charge col-lection process. The light absorption can be enhanced byincreasing the thickness of the TiO2 layer so that more

Scheme 1. Schematic depicting the near-field enhancement around thegold nanoparticles in Au–TiO2 plasmonic DSSC.

12 N. Chander et al. / Solar Energy 109 (2014) 11–23

number of dye molecules are available for light harvesting.But this will lead to lower charge collection efficiency as theelectrons have to travel a larger distance to reach the col-lecting electrode.

The researchers are trying to address these problems byutilizing plasmonic properties of metal nanoparticles (NPs)(Atwater and Polman, 2010) as one of the solutions. Thiscan be achieved in two ways, either by enhancing the pho-ton path length in solar cell using the scattering process, orby intensifying the light absorption around NPs, therebyavoiding the need to increase the physical thickness ofthe TiO2 film. Scattering due to metal NPs is generallyemployed in the silicon based solar cells where it is imprac-tical to embed metallic NPs in the active material (Pillaiet al., 2007; Catchpole and Polman, 2008; Thouti et al.,2013). But embedding metal NPs in TiO2 or ZnO layer typ-ically employed in DSSC devices is a viable option forenhancing the light absorption. The commonly used syn-thetic dyes in DSSCs absorb primarily in the visible region,which is the reason why researchers mainly employ Au andAg nanostructures in DSSCs because their surface plasmonresonance (SPR) can be tuned in the visible part of the elec-tromagnetic spectrum.

A careful look at the plasmonic DSSC literature revealsthat a wide range of metal NPs sizes, from �2 nm to�100 nm, have been utilized for improving cell perfor-mance (Brown et al., 2011; Qi et al., 2011; Jeong et al.,2011; Nahm et al., 2011; Deepa et al., 2011; Kawawakiet al., 2013; Li et al., 2013). Gold nano-islands and silvernanoparticles, synthesized by physical vapor depositionand sputtering respectively, have also been reported toenhance photocurrents in DSSC (Ng et al., 2014; Linet al., 2012). It is not clear from these studies as to what sizeof metal NPs should be used for getting optimum perfor-mance from DSSCs. So, there is a need for a systematicstudy of different sizes of metal NPs incorporated in a typ-ical TiO2 mesoporous film employed for fabrication ofDSSC.

Here, in this work we have studied initially the opticalproperties of different sizes and concentrations of goldnanoparticles (GNPs) embedded in a 3D TiO2 mesoporousmatrix to find out the optimum particle size and concentra-tion for obtaining maximum absorption enhancement. Forthis, we have chemically synthesized GNPs of different sizes,blended different concentrations with TiO2 nanopowder tofabricate conformal nanocomposite (NC) films by conven-tional doctor blade method. The TiO2–Au NC films wereused to fabricate efficient DSSC devices. Scheme 1 showsthe plasmonic DSSC architecture with GNPs embeddedinside a TiO2 matrix. The ‘glow’ around the GNPs repre-sents the conversion of incident electromagnetic far-fieldinto near-field around the GNPs due to the SPRs. Totalreflectance, total transmittance and absorptance spectrahave been studied to gain insight into the optical propertiesof NC films containing different sizes and concentrations ofGNPs. The photovoltaic performance of DSSCs containingdifferent sizes and concentrations of GNPs has been

evaluated by quantum efficiency and current density–volt-age measurements. We have tried to correlate and reasonthe observed optical and electrical behavior of plasmonicfilms. After identifying the optimum size and concentrationof GNPs for obtaining maximum efficiency enhancementusing standard N719 dye, the plasmonic effects of GNPshave also been verified with black (N749) dye to look intothe effect of SPRs on extinction coefficients of dyes. Finally,we also investigated the observed effects with commercialTiO2 paste.

2. Experimental methods

2.1. Synthesis of gold nanoparticles

GNPs were synthesized by the well known Turkevichmethod (Daniel and Astruc, 2004). A 30 ml solution of1 mM hydrogen tetrachloroaurate(III) trihydrate in deion-ized water was heated to boil on a hot plate. Differentquantities of 1% w/w trisodium citrate dihydrate aqueoussolution were added to the boiling solution under stirring.Adding different amounts of citrate resulted in formationof GNPs of different sizes. Typically, for a 30 ml gold pre-cursor solution in water, about 1.5 ml, 1 ml and 0.8 ml of1%w/w trisodium citrate aqueous solution was added toobtain particle sizes of �17 nm, �36 nm and �85 nmrespectively (Fig. S1b–S1d, Supplementary material).

A different synthetic route was employed to synthesizeGNPs of size �5 nm (Fig. S1a, Supplementary material).A 100 ml aqueous solution containing 0.25 mM HAuCl4and 0.25 mM tri-sodium citrate was prepared in a glassbeaker. Then �2.4 ml of 0.1 M ice cold NaBH4 solutionwas quickly added while stirring the solution. The colorof solution immediately changed to reddish pink indicatingparticle formation.

About 1.5 ml of a 5% polyvinylpyrrolidone (PVP) aque-ous solution was added to 30 ml of as synthesized GNPssolution and stirred for 8 h at room temperature. After-wards the solution was centrifuged at 12,500 rpm for15 min. The supernatant was removed carefully and theGNPs collected at the bottom of centrifuge tubes werere-dispersed in either ethanol or water. GNPs dispersed

N. Chander et al. / Solar Energy 109 (2014) 11–23 13

in ethanol were generally used for making plasmonic pastesfor DSSC fabrication.

2.2. Fabrication of TiO2–Au films for optical studies

Two pastes were prepared for depositing films. Firstcontained TiO2 nanoparticles (Avg. size <25 nm, SigmaAldrich) mixed with terpineol and ethyl cellulose in a cer-tain ratio. Second paste or the plasmonic paste consistedof GNPs mixed with TiO2 NPs in gold–TiO2 weight ratiosof 0.1%, 0.24%, 0.4% and 0.7%. The two pastes were usedfor making thin films via the doctor blade method on glassmicroscope slides for studying the optical properties. Thefilms prepared on glass slides for study of optical propertieshad a larger area, nearly 3 cm by 2 cm, compared to thefilms fabricated for studying the photovoltaic performance.This was done in order to completely cover the exit andentry port of the integrating sphere while performing totalreflectance and total transmittance measurements. Thethickness of the deposited films was �10 lm, similar tothe ones used for making photo-anodes of DSSCs. Allthe films were annealed at 500 �C in air to remove anyorganic material and to improve adhesion of films on thesubstrates.

2.3. Fabrication of photoanodes and assembly of DSSCs

A slightly different approach was employed to depositfilms for use in DSSCs to get optimum performance fromthe devices. Fluorine doped tin oxide (FTO) glass sub-strates were cleaned using soap water, deionized waterand propan-2-ol in an ultrasonic bath for 10 min each.An ethanolic solution of di-iso-propoxy titanium-bis(ace-tylacetone) was used to deposit thin compact films ofTiO2 on FTO glass substrates by spray pyrolysis at450 �C. The compact film helps in preventing the recombi-nation of electrons and ions present in the electrolyte at theFTO–electrolyte interface. On top of the compact layer amesoporous TiO2 layer of thickness �10 lm was depositedby the doctor blade method. The area of TiO2 films thusprepared was kept constant at 0.3 cm2. After deposition,the films were annealed in air at 500 �C for 30 min. Follow-ing the annealing process a scattering layer of thickness�3 lm, consisting of 150–250 nm TiO2 particles (WER2-O, Dyesol), was deposited on the mesoporous layer. Thefilms were annealed again and subsequently immersed ina 40 mM aqueous TiCl4 solution at 70 �C for 30 min. Thiswas followed by heat treatment up to 500 �C. Thus pre-pared TiO2 films were dipped in 0.3 mM ethanol solutionsof N719 (Di-tetrabutylammonium cis-bis(isothiocyanato)bis(2,20-bipyridyl-4,40-dicarboxylato)ruthenium(II)) dyeand N749 black (triisothiocyanato-(2,20:60,600-terpyridyl-4,40,400-tricarboxylato) ruthenium(II) tris(tetra-butylammo-nium)) dye for 18 h. The counter electrode was made ofsputter deposited �100 nm platinum film on ITO glass.The electrolyte consisted of 0.1 M LiI, 0.05 M I2, 0.5 Mtert-butylpyridine (TBP) and 0.6 M 1-propyl-2,3-dimethyl

imidazolium iodide (DMPII) in acetonitrile. A sandwichconfiguration was made using the photoelectrode and thecounter electrode finally and the cell was sealed using Sur-lyn, which also acts like a spacer layer between the twoelectrodes. To obtain reliable data and ensure repeatabilityfour cells of each size and concentration of GNPs wereassembled and tested.

2.4. Instruments

The total reflectance and total transmittance spectra offilms were recorded using UV–Vis-NIR spectrophotometer(Perkin Elmer Lambda 1050) with a 150 mm integratingsphere. TiO2 film was examined under a scanning electronmicroscope (SEM) at an operating voltage of 30 kV tostudy its morphology. Transmission electron microscopy(TEM) studies were performed by an electron microscope(Tecnai G2 F20, acceleration voltage 200 kV). Thicknessof all the films was measured by Dektak XT (Bruker) sur-face profiler. The testing of solar cells was performed underAM1.5G illumination conditions with an incident lightpower of 100 mW/cm2. Class AAA solar simulator (Sol3A,Oriel Newport, USA) with a 450 W xenon lamp was usedas the light source. A certified Si solar cell was used as a ref-erence for calibrating the light source. A Keithley 2440source-meter was used for the current–voltage measure-ments. Quantum efficiency measurement system (ReRaSpeQuest, Netherlands) was used for recording incidentphoton to current conversion efficiency (IPCE) spectra.The system consisted of a halogen lamp and a 150 W xenonlamp. The incident monochromatic light was chopped at20 Hz and the signals were acquired using a lock-in ampli-fier (Anfatec eLockIn203, Germany). IPCE measurementswere performed with a bias light of 10 mW/cm2 (0.1Sun). All the measurements were performed at roomtemperature.

3. Results and discussion

3.1. Optical properties of TiO2–Au nanocomposite films

Fig. 1a shows the extinction spectra of GNPs used in thepresent work. The TEM images of GNPs of various sizesare also shown in Fig. 1c–f. In general, the Turkevichmethod leads to formation of particles with a broad sizedistribution, so here we mention only the average size ofparticles. The presence of non-uniform sized particles isnot entirely undesirable from plasmonic enhancementpoint of view. So, we will talk about the size effects keepingin mind the broad distribution and non-uniformity ofparticle sizes.

It can be inferred from Fig. 1a that the spectrum of�85 nm particles is broader than that of �5 nm, �17 nmand �36 nm NPs. The observed wavelength red-shiftsand broadening of SPRs of GNPs with increasing sizeare due to the retardation effects of the dipolar fields acrossthe NPs, which are very receptive to the size of the NP

Fig. 1. (a) Normalized extinction spectra of various sizes of gold nanoparticles in water used in the present study. (b) Scanning electron micrograph of aTiO2 film prepared by doctor blade method. Transmission electron micrographs of gold nanoparticles of average size (c) �85 nm, (d) �36 nm, (e) �17 nmand (f) �5 nm.

14 N. Chander et al. / Solar Energy 109 (2014) 11–23

leading to variations in the far-field radiation patterns.Material extinction (absorption + scattering) function isvery sensitive to the NPs size which determines not onlythe position but also shape of the SPR spectrum. Usuallythe small sized NPs show a dipolar pattern of uniform fieldover all directions after the absorption, compared to thestrong forward/backward scattering by the bigger NPs.The extinction spectrum of �85 nm GNPs exhibits a broadfeature in the red region due to the presence of bigger par-ticles and also possibly due to some aggregation. Due tothis, along with the red-shift of the SPR wavelength, thescattering/absorption cross-sections can also be broadenedbecause of radiation losses. It must be noted here that�5 nm GNPs also show a broad extinction spectrum rela-tive to the �17 and �36 nm GNPs. This widening of thespectrum is because of the surface dispersion effects. Thesurface dispersion or surface scattering of electrons takesplace when their mean free path is comparable to or largerthan the size of the nanoparticle (Noguez, 2007; Link andEl-Sayed, 2003).

Fig. 1b shows the SEM image (200 nm scale) of a TiO2

film prepared by the doctor blade method on glass micro-scope slides. All the films used in the optical studies havea thickness of 10 ± 0.5 lm as measured by the Dektak sur-face profiler. Fig. 2 shows the digital photograph of theTiO2–Au NC films fabricated on glass slides containingvarious sizes and concentrations of GNPs. All the NC filmscontaining �85 nm GNPs exhibit only light coloration ascompared to the brilliant colors shown by NC films fabri-cated using smaller GNPs. It can also be seen that as theconcentration of GNPs in the NC films increases, they

show more intense colors due to the enhanced absorptionin the films.

Total reflectance and total transmittance of TiO2–AuNC films are shown in Fig. 3. Calculated absorptance (A)is shown in Fig. 4. Absorptance or absorption taking placein the films can be measured as a percentage value accord-ing to the relation A = 100 � R � T, where R and T aretotal reflectance and total transmittance expressed in per-centages, respectively. The NC films fabricated using largesized GNPs (�85 nm) show the lowest absorption(Fig. 4d). This observation can also be ascertained by look-ing at the photograph of the films shown in Fig. 2.

In all the NC films, the reflectance spectra show adecline in the 450–650 nm region (Fig. 3). This decrementin reflectance is more pronounced for higher concentra-tions of GNPs. For all NC films, the reflectance values nearthe resonance wavelengths are about 10% less than that ofoff-resonance wavelengths. The minima corresponding toSPRs of GNPs occur at different wavelengths for the fourdifferent sizes of NPs. For small GNPs of mean sizes�5 nm and �17 nm, the SPR is at �560 nm, and SPRwavelengths shift to �570 nm and �580 nm when the par-ticle sizes are �36 nm and �85 nm, respectively. The SPRwavelengths of GNPs in TiO2 are red shifted comparedto the GNPs in aqueous solution (Fig. 1a). The GNPsare embedded in TiO2 matrix (which has higher refractiveindex than that of water); along with the intrinsic/extrinsicsize effects, the chemical interface damping and modifieddielectric function of the effective medium also influencethe SPR peak position. An increase in reflectance wasobserved from the composite films in the red region, which

Fig. 2. Digital photograph of TiO2–Au nanocomposite films containing various sizes and concentrations of gold nanoparticles. TEM images of thecorresponding sizes of gold nanoparticles are also shown.

Fig. 3. Total reflectance and total transmittance (inset) spectra of TiO2–Au films containing different weight ratios of gold nanoparticles of average sizes(a) �5 nm, (b) �17 nm, (c) �36 nm and (d) �85 nm.

N. Chander et al. / Solar Energy 109 (2014) 11–23 15

is more pronounced for higher concentrations of GNPs.The transmittance spectra of NC films show a decline inthe entire measurement range (insets of Fig. 3). The dipsoccurring due to SPRs of GNPs can be clearly seen. Thetransmittance decreases with increasing concentration ofGNPs. But unlike the increment observed in red regionin the case of reflectance spectra, the NC films showreduced transmittance in the entire visible region.

It is of interest to note that NC films containing �85 nmsized GNPs have higher reflectance and transmittance ascompared to the other three samples. Strong optical effectsdue to the particle size are usually observed in the sizeregions which coincide with the quasi-static approxima-tion, in which the cross sections for absorption and scatter-ing for a sphere with diameter ‘d ’ are given by Cabs / d3

and Csca / d 6, respectively (Bohren and Huffman, 1998).

Fig. 4. Calculated absorptance spectra of bare TiO2 and TiO2–Au nanocomposite films containing varying concentrations of gold nanoparticles ofaverage sizes (a) �5 nm, (b) �17 nm, (c) �36 nm and (d) �85 nm. Absorptance has been calculated according to the relation A + R + T = 100, where A, R

and T are absorptance, reflectance and transmittance, respectively.

16 N. Chander et al. / Solar Energy 109 (2014) 11–23

Absorption is proportional to the cube of particle diameter(i.e. the volume or size of particle), but scattering scales asthe sixth power of diameter, so in the larger particles scat-tering dominates over the absorption. Due to this volume(size) dependence of absorption and scattering, we see thatthe composite films containing �85 nm sized GNPs (Fig. 4)have the lowest absorption indicating enhanced scatteringcompared to the other three sizes of NPs. The broadnessof the extinction spectrum of �85 nm GNPs (Fig. 1a) alsosuggests the presence of significant scattering effects.

These observations suggest that the surface plasmonicabsorption of only a definite size of GNPs can generatestrong near-fields with evanescent waves around the NPs.These results are in agreement with earlier published theo-retical studies (Jain et al., 2006; Temple et al., 2009;Tanabe, 2007), which show that the intermediate sizedNPs (�50 nm) are best suited for absorption applicationsand larger particles exhibit significant amounts of scatter-ing which may be utilized for light trapping applications(Atwater and Polman, 2010; Thouti et al., 2013). The lightscattering by GNPs before excitation of the dye moleculescannot enhance the photocurrent significantly (Kawawakiet al., 2013). In our study the composite films containingsmall GNPs (�5, �17, �36 nm) were found to exhibit high

absorption. While fabricating DSSC devices, we used vary-ing w/w ratio of Au–TiO2 to observe surface plasmonenhancement effects.

Fig. 5 shows the reflectance spectra of TiO2–Au 36 nmcomposite film of 0.24 wt%, and pure TiO2 film (inset) atdifferent steps of the fabrication process. After the firstannealing step of 500 �C all the organic binders used formaking films got evaporated and we see an increase inreflectance near the band edge (�400 nm) for both thefilms. For the NC film containing 36 nm GNPs, the dipcorresponding to SPR of GNPs becomes more pronouncedand red shifts to �570 nm. This red shift occurs because theincreased compactness of mesoporous film can result in aslight increase of refractive index (modified effective dielec-tric medium around the particle) due to the heating pro-cess. Following this step the films were dipped in 40 mMTiCl4 aqueous solution for 30 min and subsequently rinsedwith de-ionized water. This post-treatment using TiCl4solution results in the formation of a thin shell of TiO2

on NPs present in the mesoporous matrix (Ito et al.,2008; Sommeling et al., 2006). Sommeling et al. (2006)performed a detailed study and proved the formation ofthis thin TiO2 shell using TEM, BET (Brunauer–Emmett–Teller) surface area analysis and weight analysis.

Fig. 5. Total reflectance spectra of TiO2–Au 36 nm 0.24 wt% film, andTiO2 film (inset) showing the effect of TiCl4 treatment. The dipcorresponding to surface plasmon resonances of gold nanoparticles,which is red-shifted upon heating and with TiCl4 treatment to �570 nmand �580 nm, respectively is also shown.

N. Chander et al. / Solar Energy 109 (2014) 11–23 17

The TiCl4 treatment passivates the surface of DSSC filmswhere the electrons do not leak to the surface; thisimproves the electron lifetime and transport process andan improvement in photocurrent is obtained (Fuke et al.,2009; O’Regan et al., 2007). Due to these reasons, the TiCl4treatment has become a standard procedure in the fabrica-tion of DSSCs. The reflectance of both the films slightlyincreases following this treatment indicating that the filmshave become richer in TiO2 content. The valley corre-sponding to the SPR of GNPs further red shifts to�581 nm indicating that the surrounding medium of GNPshas changed. This red-shift of more than 10 nm means thatthe surrounding medium experienced by the GNPs has fur-ther modified because of the TiO2 shell formation. Themetallic NP’s polarizability (dipole moment) depends ondielectric function of the metal as well as surrounding med-ium apart from the frequency of incident light. Due to this,the finite geometry with modified surface will generateretarded stationary surface charge density wave, which islocalized at the particle. The electrical behavior of deviceswith and without TiCl4 treatment is also discussed later.

3.2. Plasmonic DSSCs

Four batches of seventeen DSSCs, corresponding tofour different particle sizes and four different concentra-tions and also the control device, were fabricated. The pho-tovoltaic parameters of these 68 cells were recorded andcompared to obtain accurate and reliable data. Fig. 6shows the average values and standard deviations ofshort-circuit current density (Jsc), open circuit voltage(Voc), fill factor (FF) and power conversion efficiency(PCE) of the sixty-four plasmonic DSSC devices and the

control device fabricated in the present study. Enhance-ment of photocurrent for certain weight percentages ofGNPs is clearly evident from Fig. 6. The J–V curves ofplasmonic devices showing highest photocurrent for a par-ticular size are plotted in Fig. 7a. IPCE spectra for the cor-responding devices are also shown for comparison inFig. 7b. In general, small sized NPs (40 nm and smaller)exhibit better performance than the bigger NPs (85 nm).It must be mentioned here that all the devices fabricatedin the present study were tested immediately after assemblyand no long term stability studies were performed.

3.2.1. Correlation between optical and electrical properties of

TiO2–Au plasmonic DSSCs

The plasmonic devices fabricated using 0.1 and0.24 wt% of �5 nm GNPs show 4.6% and 8.13% enhance-ment in photocurrent, respectively compared to the controldevice. Despite the fact that absorptance of NC films con-taining �5 nm GNPs is similar to the absorptance values of�17 and �36 nm GNPs, the enhancement in current isquite low compared to the enhancement shown by devicesfabricated using �17 and �36 nm GNPs. This apparentanomaly seen in the optical and electrical behavior of�5 nm GNPs indicates that the increased absorption oflight is not resulting in enhanced charge carrier generationin DSSC. This could be due to undesirable phenomenonlike dominant parasitic absorption, so that the maximumenergy is dissipated as heat within the nanoparticle withoutoptical excitation of surrounding dye molecules. It shouldbe noted that the IPCE spectrum (Fig. 7b) of the best plas-monic DSSC (0.24 wt%) fabricated using �5 nm GNPsshows only a small enhancement near the 550 nm wave-length region. It means that the enhancement in photocur-rent due to increased light harvesting is not significant. Thehigh absorptance shown by the NC film containing0.24 wt% of �5 nm GNPs along with the small IPCEenhancement confirms that the parasitic absorption orplasmonic heating effects are dominant and therefore wedo not see a great improvement in photovoltaic perfor-mance for this device.

Plasmonic DSSCs containing �17 and �36 nm GNPsshow the best overall device performance. The DSSC with0.24 wt% of �17 nm GNPs exhibited an increment of16.52% in photocurrent and 17.83% in overall power con-version efficiency, while these enhancement values are18.44% and 19.12% for the DSSC containing �36 nmGNPs (Table 1). The IPCE spectra for these devices showmaximum values in the 500–550 nm region indicating thatthese enhancements are indeed because of improved lightharvesting. At resonance condition, the enhanced near-fieldeffects are most prominent and therefore the spectral fea-tures of IPCE spectra (Fig. 7b) match with the SPRs ofGNPs observed in the absorptance spectra (Fig. 4). Theamounts of dye loadings for DSSCs containing 0.24 wt%17 and 36 nm GNPs are similar to that of bare TiO2 DSSC(supporting Fig. S2), indicating that the current enhance-ment is only because of plasmonic effects and not because

Fig. 6. Concentration dependence of device parameters of plasmonic DSSCs sensitized with N719 dye and containing gold nanoparticles of average sizes�5 nm, �17 nm, �36 nm and �85 nm. Averaged values of the parameters along with the standard deviations are shown.

Fig. 7. (a) Current density versus voltage curves, (b) incident photon to current conversion efficiency (IPCE) spectra of control device and plasmonicdevices. Only the concentrations giving highest photocurrent for a particular size are shown in the graphs. Inset (b) shows relative enhancement of IPCEfor control and plasmonic device containing 0.24 wt% of �36 nm gold nanoparticles.

18 N. Chander et al. / Solar Energy 109 (2014) 11–23

of any changes in the dye loading. The electronic excita-tions of dye molecules are more efficient under the back-ground of localized electric fields of GNPs. The enhancedphotocurrent is due to the GNPs near-field excitationsalong with the incident light far-fields (Brown et al., 2011).

From Fig. 4, we can see that the NC films containingsmall sized GNPs show higher absorptance values thanthe NC films containing 85 nm GNPs. This means one

would expect relatively poor performance from the plas-monic DSSCs fabricated using 80 nm GNPs. Indeed thishas been observed in the photovoltaic parameters, whichare shown in Fig. 6. All concentrations of 85 nm GNPshave low current values compared to DSSCs fabricatedusing other particle sizes. Also, the current falls drasticallyas the concentration of GNPs is increased to 0.4 and0.7 wt%. Apart from the lower absorptance some other fac-

Table 1Table displaying current density, open circuit voltage, power conversion efficiency, percent increment in current and efficiency for the control andplasmonic DSSCs sensitized with N719 dye. Only the data for devices fabricated with optimum concentration for a particular size are tabulated.

Type of DSSC Voc (mV) Jsc (mA/cm2) Fill factor (%) Efficiency (%)

N719 dye

Control DSSC 653 ± 4 14.1 ± 0.32 67 ± 1 6.17 ± 0.15Au 5 nm – 0.24 wt% 670 ± 3.5 15.28 ± 0.18 66.5 ± 1 6.85 ± 0.14Au 17 nm – 0.24 wt% 656 ± 3.2 16.43 ± 0.1 67.5 ± 1 7.27 ± 0.1Au 36 nm – 0.24 wt% 644 ± 5 16.7 ± 0.23 68.5 ± 1 7.35 ± 0.1Au 85 nm – 0.1 wt% 639 ± 3 14.8 ± 0.15 69 ± 2 6.52 ± 0.14

N. Chander et al. / Solar Energy 109 (2014) 11–23 19

tors are also responsible for this performance degradation.By utilizing larger nanoparticles, the surface area of the NCfilms is reduced leading to a reduction in dye loading (seeFig. S2). So any enhancement obtained via increasedabsorption is negated by a concomitant decrease in thenumber of available dye molecules in the mesoporous film.So, we see that the plasmonic devices containing 0.4 and0.7 wt% of 85 nm GNPs show lower efficiency than thecontrol device. In fact, with 85 nm GNPs we see animprovement in photocurrent and overall efficiency onlyfor 0.1 and 0.24 wt% DSSCs. The enhancement in photo-current is very small, 5% and 2.5% for best devices contain-ing 0.1 and 0.24 wt% GNPs respectively. IPCE spectrum(Fig. 8b) for these larger particles shows enhancement nearthe 550 nm region which provides evidence that this littleenhancement is due to increased light absorption. But plas-monic enhancement effects of these larger NPs offer only atiny improvement in photocurrent. Such a minute incre-ment in current makes the larger nanoparticles a poorchoice for fabricating high efficiency plasmonic devicesdue to their large scattering cross-sections, which resultsin a large portion of the incident energy to be sent backto the far-field instead of localizing around GNPs.

In general, lower concentrations (0.1 and 0.24 wt%) ofGNPs of sizes in the range 15–40 nm show the largestenhancement in current and energy conversion efficiency.Higher concentrations do not show improved performance.

Fig. 8. Extinction spectra of N719 dye and N749 (black) dye.

This could happen because of the saturation effect of local-ized surface plasmon induced enhancement, and also dueto the increase in impurities and recombination lossesbrought by incorporation of GNPs in TiO2 matrix. Thelower photocurrent values observed for higher concentra-tions provide support to this argument (see Fig. 6). Higherconcentrations would increase the parasitic absorption,therefore the concentration of GNPs has to be optimizedto obtain current enhancement. It is very much similar tocontrolling the surface coverage of metal NPs on siliconin the case of plasmonic silicon solar cells where an increasein surface coverage would degrade performance due toincrease in parasitic absorption (Pillai et al., 2007; Thoutiet al., 2013). It must also be noted here that fill factor val-ues of plasmonic DSSC devices are similar or slightly betterthan the fill factor of control device. The power conversionefficiency follows the trend of Jsc (Fig. 6), which means thatthe Jsc values are the dominant factors in determining theoverall device efficiency.

The values of Jsc calculated by integrating the IPCEspectra were about 5–6 mA lower than the values obtainedusing the J–V measurement from the solar simulator. Thishappened because the IPCE spectra were recorded at a rel-atively high chopping frequency of 20 Hz. Due to somehardware limitations the chopping frequency of the inci-dent light could not be brought down to 1–2 Hz or DCmode, which is the preferred chopping frequency settingfor measuring the IPCE spectra of DSSCs. An increase inchopping frequency above the optimum value reduces theIPCE spectrum and has been well documented in the scien-tific literature (Guo et al., 2010; Xue et al., 2012). But therelative enhancement in IPCE spectra is still evident andvalid because all the devices were measured at the samechopping frequency of 20 Hz.

As mentioned earlier the TiCl4 treatment of mesoporousfilms leads to the formation of a thin TiO2 shell on GNPsresulting in some kind of core–shell structure. We fabri-cated thin plasmonic devices (�4 lm) with and withoutTiCl4 treatment and studied their J–V characteristics. Itwas observed that plasmonic device containing 0.24 wt%�36 nm GNPs, which was not subjected to the TiCl4procedure, exhibited reduced photocurrent, fill factor,photovoltage and efficiency (supporting Fig. S3). This hap-pens probably because the bare GNPs act as recombina-tion centers; although plasmonic enhancement ofphotocurrent with bare gold and silver NPs has been

20 N. Chander et al. / Solar Energy 109 (2014) 11–23

reported earlier (Deepa et al., 2011; Lin et al., 2012; Liuet al., 2013; Muduli et al., 2012). Our observations indicatethat TiO2 shell formation due to TiCl4 treatment is pre-venting the GNPs from coming in direct contact with theiodide–triiodide redox species, thus avoiding corrosion.But, as mentioned earlier, no long term stability studieswere performed; so we cannot comment on the stabilityof this thin titania shell in the long run compared to propercore–shell nanostructures. A similar chemical post-treat-ment involving the use of titanium isopropoxide has alsobeen reported to result in a TiO2 shell formation and pre-vent the metal NPs from corrosion (Jeong et al., 2011;Ng et al., 2014).

3.2.2. Size effects of gold nanoparticles on dye sensitizedsolar cell performance

The results obtained in the present work emphasize theimportance of using an optimum size of Au NPs for gettingenhanced photocurrents from DSSCs. The main observa-tion is that the intermediate sized Au NPs (�36 nm) pro-vide more enhancement than larger or smaller Au NPs. Itis also worth noting that enhanced absorptance itself isnot useful if a corresponding enhancement in current isnot obtained (as is the case with �5 nm and �85 nm AuNPs).

The conduction electrons of Au NPs oscillate and reso-nate with the electric field of incident electromagnetic radi-ation giving rise to optical near-field (oscillating electricfield) localized around the nanoparticles. This near-fieldcan excite the dye molecules more efficiently than the inci-dent far-field light, thus providing photocurrent enhance-ment. The Au NPs practically function as light capturingantennae for the dye molecules. The enhancement of pho-tocurrent depends on the intensity of the near-field, whichin turn depends on the size/shape of nanoparticle, and thesurrounding medium. It has been reported earlier thatintermediate sized Au NPs (diameter �40–50 nm) exhibitthe strongest localized electric field (near-field) (Jainet al., 2006; Kelly et al., 2003; Kawawaki et al., 2013). Thisfact makes the mid-sized Au NPs an ideal candidate forabsorption enhancement applications, e.g. in DSSCs.

Mie theory predicts that larger metal NPs (diameter>50 nm) have significant scattering components in theirextinction spectra and small NPs exhibit little or no scatter-ing (Fig. S4, Supplementary material). Although scatteringby large sized metal NPs has been shown to improve per-formance of silicon solar cells (Sardana et al., 2014), onlysmall sized metal NPs (diameter <50 nm) have been gener-ally used for plasmonic enhancement of photocurrent bynear-field effects in DSSCs (Brown et al., 2011; Qi et al.,2011; Jeong et al., 2011; Li et al., 2013). Unlike the caseof silicon solar cells, light scattering by employing largenanoparticles is not very useful in DSSCs. As mentionedearlier in Section 3.1, scattering of light before theexcitation of dye molecules by near-fields cannot providephotocurrent enhancement. This is the reason why largesized Au NPs provide little or no enhancement in the pho-

tocurrent because they possess a large scattering efficiency(Fig. S4). For very small particles (diameter <15 nm), theelectric-field intensity is suppressed due to surface scatter-ing of electrons, and also the observed absorption enhance-ment (Fig. 4a) is primarily parasitic in nature, which doesnot provide any significant improvement in photocurrent(Figs. 6 and 7).

Thus, we are left with only the mid-sized Au NPs (diam-eter �15–40 nm). We have obtained the best performancewith 0.24 wt% of �36 nm Au NPs. Our systematic experi-ments, with respect to the Au NP size, revealed that theintermediate sized Au NPs in the size range of 15–40 nmare ideal for obtaining enhanced photocurrents (by near-field effects) from the DSSCs. These observations are insync with the already published theoretical and experimen-tal studies.

3.2.3. High efficiency devices

State of the art DSSCs employ a 12–14 lm mesoporousTiO2 layer and 5–6 lm of scattering layer (Ito et al., 2008),which are relatively thicker than the devices fabricated inour study to observe plasmonic enhancement effects, whichhave a �10 lm mesoporous layer and �3 lm of scatteringlayer. After identifying the optimum size and concentrationof GNPs from earlier study, we fabricated high efficiencycontrol and plasmonic DSSCs with �13 lm mesoporousTiO2 layer and �5 lm of scattering layer using commer-cially available titania paste (Dyesol 18NR-T). Apart fromusing N719 dye, another commonly used dye namely N749(black) dye was also used for sensitizing the device. Theblack dye has a wider absorption range which extends upto �800 nm compared to that of N719 dye, which absorbsonly up to �650 nm (see Fig. 8). Also the extinctioncoefficient of black dye (6766 M�1 cm�1 at 600 nm)(Nazeeruddin et al., 1999) is lower than that of N719 dye(14,700 M�1 cm�1 at 535 nm) (Yang et al., 2011). TheJ–V curves of high efficiency control and plasmonic DSSCscontaining 0.24 wt% of�36 nm GNPs, which are sensitizedwith N719 and black dyes are shown in Fig. 9. Theenhancements in photocurrent and efficiency for N719dye DSSC are �13% and �15%, respectively; for blackdye the enhancements are �9% and 13.75%, respectively(see Table 2).

The major observations with this study are (from Tables1 and 2); (1) We have seen with our configuration that thebest plasmonic device efficiency (�7.46%) is nearly thesame as that of control high efficiency device (�7.59%),which means that even with �10 lm mesoporous TiO2

layer with reduced dye loading, we are able to fabricatethe same high efficiency device using GNPs by having ‘opti-cally thick’ photanode with better photon management. (2)The observed difference in enhancement of photocurrents(�13% in case of N719 dye and �9% in case of blackdye) with two dyes can be attributed to the difference inextinction coefficients as well as absorption band regions.Under the back ground of near-fields the apparent extinc-tion coefficient of any absorbing molecule can be enhanced

Fig. 9. Current density versus voltage curves and incident photon to current conversion efficiency spectra of high efficiency devices fabricated usingcommercial titania paste mixed with 0.24 wt% of �36 nm gold nanoparticles and sensitized with, (a) N719 and (b) N749 black dyes.

Table 2Table displaying current density, open circuit voltage, power conversion efficiency, percent increment in current and efficiency for the high efficiencycontrol and plasmonic DSSCs sensitized with N719 and N749 (black) dyes and fabricated using a commercial titania paste. Data for only the best devicesis shown in the table.

Type of DSSC Voc (mV) Jsc (mA/cm2) Fill factor (%) Efficiency (%)

N719 dye

Control DSSC (commercial paste) 677 ± 2 16.1 ± 0.1 69.7 ± 0.5 7.6 ± 0.1Au 36 nm – 0.24 wt% (commercial paste) 672 ± 2 18.13 ± 0.13 71.4 ± 1 8.7 ± 0.2

N749/black dye

Control DSSC (commercial paste) 628 ± 2 15.14 ± 0.14 68 ± 1 6.47 ± 0.2Au 36 nm – 0.24 wt% (commercial paste) 619 ± 3 16.51 ± 0.1 72 ± 0.5 7.36 ± 0.15

N. Chander et al. / Solar Energy 109 (2014) 11–23 21

due to the intense electromagnetic field fluctuation, whichis the most significant feature of localized SPRs of metalNPs. But, the black dye absorption peak is off-resonantwith the SPRs of GNPs compared to the N719 dye. Conse-quently, the observed plasmonic enhancement of photocur-rent is lower in the case of black dye. (3) The otherdifference in photocurrent enhancement seen between thethicker high efficiency and thinner normal devices. The nor-mal DSSC containing 0.24 wt% of 36 nm GNPs exhibited�18% higher photocurrent compared to the control device,however, for the case of high efficiency plasmonic devicethe photocurrent enhancement is �13%. So, the deviceswith thicker films have somewhat lower plasmonicenhancement of photocurrent. But even this increment of�13% observed in photocurrent for the case of high effi-ciency devices is significant. Researchers have also reportedsimilar enhancements in current (10–15%) for high effi-ciency plasmonic DSSCs (g � 10%) even though their base-line efficiency was already quite high (�8–9%) (Choi et al.,2012; Dang et al., 2013).

4. Conclusions

We have presented a systematic study on the influence ofsize and concentration of spherical GNPs on the perfor-mance of plasmonic DSSCs. We have shown using opticaland electrical studies, that particles of sizes in the range of15–40 nm and concentrations in the range of 0.1–0.25%

offer the maximum enhancement in DSSC performance.Incorporation of 17 and 36 nm sized GNPs resulted inenhancement of photocurrent and energy conversion effi-ciency providing the evidence of plasmonic effects in thedevice due to the enhanced near-fields excitation of dyemolecules along with incident light far-fields. The best plas-monic DSSC device with 0.24 wt% of 36 nm GNPs exhib-ited an enhancement of 18.44% in photocurrent and19.12% in overall power conversion efficiency. Small GNPsof size �5 nm showed little improvement in photovoltaicbehavior despite the high absorptance leading us to con-clude that the enhanced near-field is wasted as plasmonicheating and not resulting in optical excitation of surround-ing dye molecules. With �85 nm GNPs, we have observedminimal enhancement in the device performance due tolarge scattering cross-section. The enhancements in photo-current and conversion efficiency were also observed withhigh efficiency devices after incorporating optimized sizeand concentration of GNPs in the device structure. Ourmethodology can be easily adopted for other technologieslike organic solar cells and perovskite solar cells to enhancethe efficiency.

Acknowledgements

Authors would like to acknowledge the partial supportfrom the Dept. of Sci. & Tech. (DST), India under theSolar Energy Enabling Research Grant Number

22 N. Chander et al. / Solar Energy 109 (2014) 11–23

RP02468. Authors would like to thank the NanoscaleResearch Facility (NRF) of IIT Delhi for optical character-ization of samples. One of the authors (A.F.K.) gratefullyacknowledges DST INSPIRE Faculty Award (IFA-CH-27) for research fellowship.

Appendix A. Supplementary material

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.solener.2014.08.011.

References

Atwater, H.A., Polman, A., 2010. Plasmonics for improved photovoltaicdevices. Nat. Mater. 9, 205–213.

Bohren, C.F., Huffman, D.R., 1998. Absorption and Scattering of Lightby Small Particles. Wiley, New York, pp. 136–141.

Brown, M.D., Suteewong, T., Santosh, R.S., D’Innocenzo, V., Petrozza,A., Lee, M.M., Wiesner, U., Snaith, H.J., 2011. Plasmonic dye-sensitized solar cells using core–shell metal-insulator nanoparticles.Nano Lett. 11, 438–445.

Burschka, J., Pellet, N., Moon, S.J., Baker, R.H., Gao, P., Nazeeruddin,M.K., Gratzel, M., 2013. Sequential deposition as a route to high-performance perovskite-sensitized solar cells. Nature 499, 316–319.

Catchpole, K.R., Polman, A., 2008. Design principles for particle plasmonenhanced solar cells. Appl. Phys. Lett. 93, 191113.

Choi, H., Chen, W.T., Kamat, P.V., 2012. Know thy nano neighbor

plasmonic versus electron charging effects of metal nanoparticles indye-sensitized solar cells. ACS Nano 6, 4418–4427.

Dang, X., Qi, J., Klug, M.T., Chen, P.Y., Yun, D.S., Fang, N.X.,Hammond, P.T., Belcher, A.M., 2013. Tunable localized surfaceplasmon-enabled broadband light-harvesting enhancement for high-efficiency panchromatic dye-sensitized solar cells. Nano Lett. 13, 637–642.

Daniel, M.C., Astruc, D., 2004. Gold nanoparticles: assembly, supramo-lecular chemistry, quantum-size-related properties, and applicationstoward biology, catalysis, and nanotechnology. Chem. Rev. 104, 293–346.

Deepa, K.G., Lekha, P., Sindhu, S., 2011. Efficiency enhancement inDSSC using metal nanoparticles: a size dependent study. Solar Energy86, 326–330.

Fuke, N., Katoh, R., Islam, A., Kasuya, M., Furube, A., Fukui, A.,Chiba, Y., Komiya, R., Yamanaka, R., Han, L., Harima, H., 2009.Influence of TiCl4 treatment on back contact dye-sensitized solar cellssensitized with black dye. Energy Environ. Sci. 2, 1205–1209.

Gratzel, M., 2001. Photoelectrochemical cells. Nature 414, 338–344.Guo, X.Z., Luo, Y.H., Zhang, Y.D., Huang, X.C., Li, D.M., Meng, Q.B.,

2010. Study on the effect of measuring methods on incident photon-to-electron conversion efficiency of dye-sensitized solar cells by home-made setup. Rev. Sci. Inst. 81, 103106.

Im, J.H., Lee, C.R., Lee, J.W., Park, S.W., Park, N.G., 2011. 6.5%efficient perovskite quantum-dot-sensitized solar cell. Nanoscale 3,4088–4093.

Ito, S., Murakami, T.N., Comte, P., Liska, P., Gratzel, C., Nazeeruddin,M.K., Gratzel, M., 2008. Fabrication of thin film dye sensitized solarcells with solar to electric power conversion efficiency over 10%. ThinSolid Films 516, 4613–4619.

Jain, P.K., Lee, K.S., El-Sayed, I.H., El-Sayed, M.A., 2006. Calculatedabsorption and scattering properties of gold nanoparticles of differentsize, shape, and composition: applications in biological imaging andbiomedicine. J. Phys. Chem. B 110, 7238–7248.

Jeong, N.C., Prasittichai, C., Hupp, J.T., 2011. Photocurrent enhancementby surface plasmon resonance of silver nanoparticles in highly porousdye-sensitized solar cells. Langmuir 27, 14609–14614.

Kawawaki, T., Takahashi, Y., Tatsuma, T., 2013. Enhancement of dye-sensitized photocurrents by gold nanoparticles: effects of plasmoncoupling. J. Phys. Chem. C 117, 5901–5907.

Kelly, K.L., Coronado, E., Zhao, L.L., Schatz, G.C., 2003. The opticalproperties of metal nanoparticles: the influence of size, shape, anddielectric environment. J. Phys. Chem. B 107, 668–677.

Lee, M.M., Teuscher, J., Miyasaka, T., Murakami, T.N., Snaith, H.J.,2012. Efficient hybrid solar cells based on meso-superstructuredorganometal halide perovskites. Science 338, 643–647.

Li, Y., Wang, H., Feng, Q., Zhou, G., Wang, Z.S., 2013. Goldnanoparticles inlaid TiO2 photoanodes: a superior candidate forhigh-efficiency dye-sensitized solar cells. Energy Environ. Sci. 6, 2156–2165.

Lin, S.J., Lee, K.C., Wu, J.L., Wu, J.Y., 2012. Plasmon-enhancedphotocurrent in dye-sensitized solar cells. Solar Energy 86, 2600–2605.

Link, S., El-Sayed, M.A., 2003. Optical properties and ultrafast dynamicsof metallic nanocrystals. Annu. Rev. Phys. Chem. 54, 331–366.

Liu, W.L., Lin, F.C., Yang, Y.C., Huang, C.S., Gwo, S., Huang, M.H.,Huang, J.S., 2013. The influence of shell thickness of Au@TiO2 core–shell nanoparticles on the plasmonic enhancement effect in dye-sensitized solar cells. Nanoscale 5, 7953–7962.

Muduli, S., Game, O., Dhas, V., Vijayamohanan, K., Bogle, K.A.,Valanoor, N., Ogale, S.B., 2012. Solar Energy 86, 1428–1434.

Nahm, C., Choi, H., Kim, J., Jung, D.R., Kim, C., Moon, J., Lee, B.,Park, B., 2011. The effects of 100 nm-diameter Au nanoparticles ondye-sensitized solar cells. Appl. Phys. Lett. 99, 253107.

Nazeeruddin, M.K., Zakeeruddin, S.M., Humphry-Baker, R., Jirousek,M., Liska, P., Vlachopoulos, N., Shklover, V., Fischer, C.H., Gratzel,M., 1999. Acid–base equilibria of (2,20-bipyridyl-4,40-dicarboxylicacid)ruthenium(II) complexes and the effect of protonation oncharge-transfer sensitization of nanocrystalline titania. Inorg. Chem.38, 6298–6305.

Ng, S.P., Lu, X.Q., Ding, N., Wu, C.M.L., Lee, C.S., 2014. Plasmonicenhanced dye-sensitized solar cells with self-assembly gold–TiO2@-core–shell nanoislands. Solar Energy 99, 115–125.

Noguez, C., 2007. Surface plasmons on metal nanoparticles: theinfluence of shape and physical environment. J. Phys. Chem. C111, 3806–3819.

O’Regan, B.C., Durrant, J.R., Sommeling, P.M., Bakker, N.J., 2007.Influence of the TiCl4 treatment on nanocrystalline TiO2 films in dye-sensitized solar cells. 2. Charge density, band edge shifts, andquantification of recombination losses at short circuit. J. Phys. Chem.C 111, 14001–14010.

Pillai, S., Catchpole, K.R., Trupke, T., Green, M.A., 2007. Surfaceplasmon enhanced silicon solar cells. J. Appl. Phys. 101, 093105.

Qi, J., Dang, X., Hammond, P.T., Belcher, A.M., 2011. Highly efficientplasmon-enhanced dye-sensitized solar cells through metal@oxidecore–shell nanostructure. ACS Nano 5, 7108–7116.

Sardana, S.K., Chava, V.S.N., Thouti, E., Chander, N., Kumar, S.,Reddy, S.R., Komarala, V.K., 2014. Influence of surface plasmonresonances of silver nanoparticles on optical and electrical propertiesof textured silicon solar cell. Appl. Phys. Lett. 104, 073903.

Sommeling, P.M., O’Regan, B.C., Haswell, R.R., Smit, H.J.P., Bakker,N.J., Smits, J.J.T., Kroon, J.M., van Roosmalen, J.A.M., 2006.Influence of a TiCl4 post-treatment on nanocrystalline TiO2 films indye-sensitized solar cells. J. Phys. Chem. B 110, 19191–19197.

Tanabe, K., 2007. Optical radiation efficiencies of metal nanoparticles foroptoelectronic applications. Mater. Lett. 61, 4573–4575.

Temple, T.L., Mahanama, G.D.K., Reehal, H.S., Bagnall, D.M., 2009.Influence of localized surface plasmon excitation in silver nanoparticleson the performance of silicon solar cells. Sol. Energy Mater. Sol. Cells93, 1978–1985.

N. Chander et al. / Solar Energy 109 (2014) 11–23 23

Thouti, E., Chander, N., Dutta, V., Komarala, V.K., 2013. Opticalproperties of Ag nanoparticle layers deposited on silicon substrates. J.Opt. 15, 035005.

Xue, G., Yu, X., Yu, T., Bao, C., Zhang, J., Guan, J., Huang, H., Tang,Z., Zou, Z., 2012. Understanding of the chopping frequency effect onIPCE measurements for dye-sensitized solar cells: from the viewpoint

of electron transport and extinction spectrum. J. Phys. D: Appl. Phys.45, 425104.

Yang, S.H., Wu, K.L., Chi, Y., Cheng, Y.M., Chou, P.T., 2011.Tris(thiocyanate) ruthenium(II) sensitizers with functionalized dicarb-oxyterpyridine for dye-sensitized solar cells. Angew. Chem. Int. Ed. 50,8270–8274.