Available online at ScienceDirect

ScienceDirectJ. Mater. Sci. Technol., 2014, 30(1), 1e7

Plasmon-enhanced Performance of Dye-sensitized Solar Cells Based on

Electrodeposited Ag Nanoparticles

Xinning Luan, Ying Wang*

Department of Mechanical & Industrial Engineering, Louisiana State University, Baton Rouge, LA 70803, USA[Manuscript received June 1, 2013, in revised form June 27, 2013, Available online 13 September 2013]

* Corresp225 5781005-03JournalLimited.http://dx

In the present work, pulse current deposition is used to deposit evenly distributed and uniformly sized Agnanoparticles onto a TiO2 nanotube array as photoelectrode in dye-sensitized solar cells (DSSCs), and the sizeand amount of loading Ag nanoparticles are controlled by the pulse deposition time. Due to the enhanced lightabsorption and electronehole separation caused by plasmon effect, DSSCs based on Ag-modified TiO2

nanotube arrays show higher energy conversion efficiencies than those based on bare nanotubes with thesame tube length. Particularly, DSSC based on nanotubes modified using pulse deposition time 1 s/3 sdelivers the highest energy conversion efficiency of 1.68% and the largest short-circuit current of 4.37 mA/cm2, while DSSC consisting of bare nanotubes exhibits efficiency of 1.20% and short-circuit current of2.27 mA/cm2, which represents a 40% enhancement of cell efficiency in DSSC based on Ag-modified TiO2

nanotubes. It is also noted that overly long pulse deposition time will not further increase DSSC efficiency dueto agglomeration of Ag particles. For example, when the pulse deposition time is increased to 2 s/6 s, DSSCbased on Ag-modified nanotubes exhibits a lower efficiency of 1.42%. Moreover, high-concentration TiCl4treatment on TiO2 nanotube arrays can further increase the energy conversion efficiencies to 3.82% and2.61% for DSSC based on Ag-modified TiO2 nanotubes and DSSC based on bare TiO2 nanotubes,respectively, by significantly creating more surface area for dye loading.

KEY WORDS: TiO2 nanotube; Dye-sensitized solar cell; Pulse current deposition; Ag nanoparticle; TiCl4 treatment

1. Introduction

New generations of solar cells have emerged, as conventionalsilicon solar cells with high energy conversion efficiencies posechallenges of complicated and high-cost processing and pro-duction[1,2]. Dye-sensitized solar cells (DSSCs) have attractedincreasing attention for their easy processing and low cost, sincemesoporous titania with high surface area was used as photo-anode to improved energy conversion efficiency of DSSC asreported by Grätzel et al. in 1991[3e12]. In a DSSC, electrons ofdye molecules are excited by solar energy absorption and thenthe electrons from the photo-excited dye inject into the con-duction band of photo-anode such as titanium dioxide (TiO2).The electrons then diffuse through TiO2 layer to the electrodeand subsequently reach the counter electrode through an externalcircuit. The dye molecules regain electrons from a redox couple

onding author. Prof., Ph.D.; Tel.: þ1 225 578 8577; Fax: þ19162; E-mail address: ywang@lsu.edu (Y. Wang).02/$e see front matter Copyright� 2013, The editorial office ofof Materials Science & Technology. Published by ElsevierAll rights reserved..doi.org/10.1016/j.jmst.2013.09.007

iodide/triiodide (I�/I3�) in an electrolyte. The iodide is regen-erated in turn by the reduction of triiodide at the counter elec-trode, and this process requires a catalytic functionality of Pt onthe cathode surface.TiO2 is commonly used as photo-anode material in current

DSSCs. One-dimensional TiO2 nanostructures such as TiO2

nanotube arrays have attracted much attention as the photo-anode material in DSSCs due to their ordered vertical structurewhich facilitates filling of new sensitizer molecules or non-volatile electrolytes for improved efficiency and stability ofDSSCs, while conventional mesoporous TiO2 film has a disor-dered structure that makes infiltration of larger sensitizer mole-cules or viscous electrolytes difficult and ineffective[13e22].However, DSSC based on TiO2 nanotube array exhibits limitedenergy conversion efficiency owning to its lower surface areathan mesoporous TiO2 film

[23e26].Numerous approaches and strategies have been investigated

so as to enhance photoelectric conversion efficiency of DSSC bysurface modification with noble metal[27e29], using semi-conductor quantum dots[30,31], developing new dyes[32e34], anddesigning new morphology[35e37]. The purposes of thesemethods are to either increase the surface area for dye loading orincrease the light harvesting ability and/or suppress the

2 X. Luan and Y. Wang: J. Mater. Sci. Technol., 2014, 30(1), 1e7

recombination of photogenerated electronehole pairs. Amongthese approaches, surface modification by noble metal has beenconsidered as a very effective approach to enhance energyconversion efficiency of DSSC by improving its light harvesting.Surface plasmon induced by noble metals has the potential forenhancing the optical absorption. The surface plasmon effect iscaused by light-driven collective oscillations of conductionelectrons in metallic nanoparticles. The nanoparticles must besmaller than the wavelength of exciting light for electrons tooscillate with the electric field of light. When these conditionsare met, an enhanced electromagnetic field is found nearby thesurface of nanoparticle. The enhanced electromagnetic field ishighly dependent upon the wavelength of incident light, as wellas the shape, size and aggregation state of the nanoparticle.Atwater et al.[38] recently outlined three types of plasmonic light-trapping geometries for incorporating metal nanoparticles: i) Far-field scattering. When light is scattered from noble metal nano-particles in DSSC, light is preferentially scattered and trappedinto the semiconductor by multiple and high-angle scattering,causing an increase in the effective optical path length in the cell.ii) Near-field scattering. When the wavelength of the irradiationsources is correlated with the optical absorption of the surfaceplasmon resonance, an increase in local electromagnetic field isfound nearby metal surface. The photo-excited particles’ near-field causes the creation of electronehole (eeh) pairs in semi-conductor and suppresses the recombination of electrons in dyeand/or electrolyte. iii) Surface plasmon polaritons at the metal/semiconductor interface. A corrugated metallic film on the backof semiconductor can couple sunlight into surface plasmonpolaritons at metal/semiconductor interface. Therefore, utiliza-tion of plasmon effect for enhancement of light is a promisingprocedure for improving the efficiency of DSSC.Till now, surface plasmon arising from metal nanoparticles has

mostly been applied to increase optical absorption and/orphotocurrent in the field of silicon solar cells[39e42] and organicsolar cells[43,44]. Not many examples of this concept have beenintroduced and demonstrated in DSSCs. Considering plasmoneffect of various noble metals, Ag is one of the most suitablecandidates for practical applications due to its easy preparationand lower cost than others. Several methods have been reportedto deposit Ag nanoparticles onto TiO2 such as photo-induceddeposition, calcination reduction, and electrochemical deposi-tion. However, none is able to prevent the aggregation of Agparticles and keep structures uniform during the depositionprocess. Chou et al. synthesized TiO2/Ag composites via drycoating process and investigated their applications in DSSCs[45].Such a nano-metal modified DSSC yielded a photo conversionefficiency of 0.93%, which was higher than DSSC based onunmodified nanotubes (0.71%) with identical film thickness of8.0 mm. However, this efficiency of 0.93% was still low, whichwas ascribed to agglomeration of Ag nanoparticles and a wideparticle size distribution of 20e100 nm. Agarwala et al.[46] alsoprepared Ag nanoparticles onto TiO2 nanotubes by using a photoreduction method, to scatter light and generate surface plasmoneffect. The DSSC efficiency was enhanced by 22% after beingmodified with Ag nanoparticles. However, the particle size waslarger than 70 nm and the nanoparticles were mixed with somerod like structures, resulting in lower dye absorption and lessscattering effect. On the other hand, Lai et al. developed a simplemethod using pulse current (PC) deposition to deposit uniformlydistributed spherical Ag nanoparticles (2e10 nm) on the surfaceof prepared TiO2 nanotube arrays. The generated photocurrent

was found to be 36% higher than that of randomly distributed Agparticles prepared by photocatalytic reduction[47]. PC electrode-position is considered as a facile method for controlling theelectrocrystallization process and hence depositing materialswith unique structures and properties. By tuning current-on time(ton) and current-off time (toff), instantaneous current densityapplied by pulse control can lead to ultrafine-grained structuresand homogeneous surface appearance of the deposit. Therefore,PC electrodeposition can form evenly distributed and uniformlynano-sized Ag. On the other hand, direct current (DC) electro-deposition results in inhomogeneous Ag with rough surface;hence, it is not a suitable method for producing Agnanoparticles[48].As summarized above, size of Ag nanoparticles synthesized

via photo reduction deposition method is too large and uncon-trollable; as a result, these commonly-used methods are notsuitable choices for enhancing plasmon effect in DSSCs. How-ever, the enhancement of plasmon effect can be easily achievedby using pulse current deposition method, which can produceevenly distributed nanoparticles with controllable size distribu-tion by tuning pulse current. However, this method has neverbeen used for depositing metallic particles for plasmon effect inDSSCs. The present work is the first effort to employ pulsecurrent deposition method for enhancing plasmon effect inDSSCs. TiO2 nanotube arrays are prepared via anodization andfollowed by deposition of uniformly distributed spherical Agnanoparticles via pulse current deposition. This process is easy toimplement and can effectively suppress the agglomeration of Agparticles by taking advantage of high nucleation rate of Ag underthe PC condition. DSSCs based on Ag-modified TiO2 nanotubearrays show significantly enhanced performance than DSSCsconsisting of bare TiO2 nanotube arrays. Efficiency of suchDSSCs can also be maximized by tuning pulse deposition timeof Ag nanoparticles. Moreover, efficiency of DSSCs can befurther improved by treating TiO2 nanotubes with high concen-tration TiCl4 prior to Ag deposition.

2. Experimental

Ti foils (0.25 mm thickness, 99.5 wt% purity, Alfa Aesar)were degreased by sonicating in acetone, deionized (DI) water,and ethanol for 15 min, respectively, then rinsed with ethanoland dried in a nitrogen stream. Electrochemical anodization ofthe Ti foils was carried out in a two-electrode cell, with a plat-inum mesh as the counter electrode at room temperature. Theelectrolyte for anodization was prepared with anhydrousethylene glycol (EG) with NH4F (0.3 wt%) and H2O (5 vol%).The voltage was supplied by a DC power supply with a digitaldisplay (Model 1623A, PK Precision). After anodization, thesamples were rinsed in deionized water (DI water) and dried inair, followed by heat treatment at 450 �C in air for 3 h to produceanatase phase. Optional TiCl4 treatment was performed on TiO2

nanotube arrays by soaking annealed nanotube in 100 mL of0.2 mol/L TiCl4 aqueous solution at 70 �C for 30 min and thenannealing at 450 �C in air for 30 min.Pulse-current electrodeposition of Ag nanoparticles was car-

ried out by immersing the as-prepared TiO2 nanotube in anaqueous electrolyte composed of 10 mmol/L AgNO3 and100 mmol/L NaNO3 at room temperature. The TiO2 nanotubearray and a platinum mesh were used as the working and counterelectrodes, respectively. A pulse-current of 10 mA/cm2 withdifferent alternating on-time and off-time (0.5 s/1.5 s, 1 s/3 s, 2 s/

Fig. 1 Schematic showing the pulse current process for electrochemicaldeposition of Ag with a rectangular pulse shape composed ofalternating on-time and off-time for 100 deposition cycles.

Fig. 2 X-ray diffraction pattern of bare TiO2 nanotube array annealed at450 �C in air for 2 h.

X. Luan and Y. Wang: J. Mater. Sci. Technol., 2014, 30(1), 1e7 3

6 s) for 100 cycles were applied. Fig. 1 shows a scheme showingthe pulse current process for electrochemical deposition of Agnanoparticles with a rectangular pulse shape composed of analternating on-time and off-time for 100 deposition cycles. TheAg-deposited samples were then rinsed with DI water and driedin vacuum at 50 �C overnight.To fabricate DSSCs, TiO2 nanotube arrays with or without Ag

modifications were soaked in anhydrous ethanol containing0.2 mmol/L N719 dye (Ru[LL0-(NCS)2], L ¼ 2,20-bipyridyl-4,40-dicarboxylic acid, L0 ¼ 2,20-bipyridyl-4,40-ditetrabutyl-ammonium carboxylate, Solaronix Co.) and sensitized for 24 h atroom temperature. Afterward, these dye-sensitized TiO2 nano-tube arrays with an active area of approximately 0.25 cm2 wererinsed with acetonitrile in order to remove physisorbed N719 dyemolecules. Platinized counter electrode was fabricated by dropcasting 0.5 mmol/L H2PtCl6/isopropanol solution on FTO glasssubstrate that had a hole for electrolyte injection later on, fol-lowed by heating at 400 �C in air for 20 min. The dye-sensitizedTiO2 nanotube arrays with or without Ag modifications weresandwiched together with Pt-coated FTO glass by using a100 mm thick hot-melt sealing film as the spacer (Meltonix 1170-100, Solaronix Co.). DSSCs were sealed by applying heat andpressure with a hot press at 110 �C. A I�/I3� based electrolyte,which contained 0.10 mol/L GTC in a mixture of acetonitrile andvaleronitrile (85:15 v/v) (No. ES-0004, IoLiTec Inc., Germany)was injected through the hole on the Pt-coated FTO into DSSC.Crystal structure of TiO2 nanotube arrays was determined by

X-ray diffraction (XRD) using a Rigaku MiniFlex diffractometerwith Cu Ka radiation operated at 30 kVand 15 mA at a scan rateof 2�/min. A FEI Quanta 3D FEG scanning electron microscope(SEM) at an accelerating voltage of 5e20 kV was used tocharacterize the morphology of TiO2 nanotube arrays. Chemicalcomposition of these nanotubes was studied by energy dispersiveX-ray spectroscopy (EDX) on a Hitachi S-3600N SEM. Surfacecomposition of TiO2 nanotube arrays was analyzed via an X-rayphotoelectron spectroscope (XPS) on the AXIS-165 XPS (KratosAnalytical Ltd.) with an AXIS 165 spectrometer from KratosAnalytical LTD by using a twin-anode Al Ka (1486.6 eV) X-raysource. Transmission electron microscopy (TEM, JEM-1011,JEOL LTD.) with an accelerating voltage of 80 kV was usedto characterize the morphology of Ag-modified TiO2 nanotubes.The currentevoltage (JeV) characteristics of DSSCs wererecorded using a Keithley 2400 source meter. A solar lightsimulator (Model: 67005, Oriel) was used to simulate sunlightunder one sun AM 1.5 G (100 mW/cm2) illumination providedby a 150 W Xenon arc lamp (Model: 6256, Oriel) and calibratedusing a Si solar reference cell (Model: 91150V, Oriel). Backside

illumination mode was used, since the photo-anode consists ofTiO2 nanotube array on Ti substrate which was not transparent tolight.

3. Results and Discussion

XRD analysis is utilized to examine the crystallinity of TiO2

nanotubes. Fig. 2 shows XRD patterns of bare TiO2 nanotubearrays annealed at 450 �C in air for 2 h. The peaks at 2q ¼ 25.4�

and 2q ¼ 48.0� correspond to (101) and (200) in TiO2 anatasephase, and all of the other dominant peaks correspond to thereflection of Ti foil substrate. This result suggests that the as-prepared anodic TiO2 nanotubes convert to anatase phase afterbeing heated at high temperatures, which can be used as photo-anodes in DSSCs.Fig. 3 shows SEM images and EDX spectrum of TiO2

nanotube arrays modified with Ag nanoparticles (NPs) via pulsedeposition time of 1 s/3 s. The cross-section SEM image inFig. 3(a) shows that the structure of TiO2 nanotube arrays pre-serves well after the PC deposition. Thickness of the TiO2

nanotube arrays is estimated to be w6.1 mm, and the averagediameter of these nanotubes is w120 nm. It is important to notethat Ag NPs are uniformly distributed and anchored along thesurface of the tube structure. It is observed from Fig. 3(b) thatmost of Ag nanoparticles have a size in a range of 10e40 nm.The top-view SEM image in Fig. 3(c) shows Ag NPs depositedon the open entrances of TiO2 nanotube arrays. It can be seenthat Ag NPs deposited on the top open entrances of TiO2

nanotubes (showing a particle size up to 60 nm) are larger thanAg NPs on the tube walls. Meanwhile, Ag NPs deposited on thetop open entrances of TiO2 nanotubes are denser (without anyobvious agglomeration) than Ag NPs on the tube walls. Such alarger particle size and denser distribution of Ag NPs on the topof TiO2 nanotubes are possibly due to the top open entrances ofTiO2 nanotubes being more exposed to electrochemical deposi-tion. Fig. 3(d) displays the EDX spectrum of these TiO2 nano-tubes arrays modified with Ag NPs. A small amount of Agelement is detected, as evidenced by its weak but discerniblepeaks in the EDX spectrum.To further study the details about Ag NPs deposited on TiO2

nanotube arrays via pulse deposition time of 1 s/3 s, its TEMimage is obtained as shown in Fig. 4. The average diameter ofAg NPs shown in the TEM image is approximately 20 nm,which is consistent with the particle size shown in the SEM

Fig. 3 (a, b) Cross-sectional, (c) top-view SEM images, and (d) EDS spectrum of TiO2 nanotube arrays modified with Ag nanoparticles via pulsedeposition time of 1 s/3 s.

4 X. Luan and Y. Wang: J. Mater. Sci. Technol., 2014, 30(1), 1e7

images earlier. This homogeneous dispersion of Ag NPs onTiO2 nanotube arrays is desirable for efficient light harvest-ing, yielding high plasmon-induced activity for electrontransportation.To further confirm the formation of Ag NPs from the pulse

deposition, XPS is also employed to study surface compositionof Ag-modified TiO2 nanotubes arrays. The XPS spectrum taken

Fig. 4 TEM image of a TiO2 nanotube modified with Ag nanoparticlesvia pulse deposition time of 1 s/3 s.

from TiO2 nanotubes arrays modified with Ag NPs via pulsedeposition time of 1 s/3 s exhibits the Ag 3d peak in Fig. 5(a). InFig. 5(b), the peaks observed at 368.1 eV and 374.2 eV areascribed to metallic Ag 3d5/2 and Ag 3d3/2, respectively, con-firming metallic Ag existence on the surface of TiO2 nanotubearrays. These findings reveal that Ag-modified TiO2 nanotubearrays are successfully achieved by electrochemical deposition ofAg in AgNO3-containing solution.Performances of DSSCs based on bare and Ag-modified TiO2

nanotubes electrodes are examined under 1 sun AM 1.5 simu-lated sunlight. Fig. 6 presents JeV curves of DSSCs based on6.1-mm-long TiO2 nanotubes arrays modified with Ag NPs usingdifferent pulse deposition time (0.5 s/1.5 s, 1 s/3 s, and 2 s/6 s).Photovoltaic characteristics of these DSSCs, such as short-circuitcurrent (Jsc), open-circuit voltage (Voc), fill factor (FF), and ef-ficiency (h) are summarized in Table 1. It is evident that theshort-circuit currents of all Ag-modified TiO2 nanotubes basedDSSCs (0.5 s/1.5 s, 1 s/3 s, and 2 s/6 s) are larger than that ofbare TiO2 nanotubes based DSSCs, since the Ag NPs on sidewalls enhance absorption and charge separation, while there areno substantial differences in the photovoltages. DSSC based onAg@TiO2 (Ag-modified TiO2 nanotubes) synthesized via theshorter pulse deposition time of 0.5 s/1.5 s, delivers an energyconversion efficiency of 1.28%, which is just slightly higher thanthe efficiency (1.20%) of DSSC based on bare nanotube,possibly because the amount of Ag NPs loading is too low toshow any plasmon effect. The DSSC based on Ag@TiO2 1 s/3 s(Ag-modified TiO2 via pulse deposition time of 1 s/3 s) deliversthe highest energy conversion efficiency (1.68%) exhibiting Voc

Fig. 5 (a) XPS spectra and (b) Ag 3d XPS peak of TiO2 nanotubesmodified with Ag nanoparticles via pulse deposition time of 1 s/3 s.

Table 1 Photovoltaic characteristics of dye-sensitized solar cells shownin Fig. 6

Jsc (mA/cm2) Voc (V) FFa hb (%)

Ag@TiO2 0.5 s/1.5 s 2.56 0.77 0.65 1.28Ag@TiO2 1 s/3 s 4.37 0.79 0.49 1.68Ag@TiO2 2 s/6 s 3.71 0.82 0.47 1.42

Bare-TiO2 2.27 0.78 0.68 1.20

a Fill-factor (FF) ¼ Pmax/(Isc � Voc).b Power conversion efficiency (h) ¼ (Isc (mA/cm2) � Voc(V) � FF/(100 (mW/cm2))) � 100%.

X. Luan and Y. Wang: J. Mater. Sci. Technol., 2014, 30(1), 1e7 5

of 0.79 V, Jsc of 4.37 mA/cm2, and FF of 0.49, which representsa 40.0% increase in cell efficiency compared to DSSC consistingof bare TiO2 nanotube array (1.20%). In theory, the maximumvoltage (Vmax) generated by DSSC is the difference between the(quasi-)Fermi level of the TiO2 and the redox potential of theelectrolyte, about 0.7 V under solar illumination conditions.Since the loaded Ag nanoparticles do not affect the Fermi levelof TiO2 and redox potential of electrolyte much, Ag depositionon TiO2 would not improve Voc. The improved Jsc is resulted

Fig. 6 JeV characteristics of DSSCs based on different TiO2 nanotubearrays (Ag@TiO2 0.5 s/1.5 s: TiO2 nanotubes modified with Agnanoparticles via pulse deposition time of 0.5 s/1.5 s; Ag@TiO2

1 s/3 s: TiO2 nanotubes modified with Ag nanoparticles via pulsedeposition time of 1 s/3 s; Ag@TiO2 2 s/6 s: TiO2 nanotubesmodified with Ag nanoparticles via pulse deposition time of 2 s/6 s; bare TiO2: unmodified TiO2 nanotubes).

from higher plasmon resonance effect caused by coverage of AgNPs on the TiO2 nanotube surface, which increases the numberof photoelectrons generated and improves the effective separa-tion of electronehole pairs. Therefore, the loaded Ag nano-particles can improve short-circuit current rather than the open-circuit voltage in these DSSCs. Interestingly, the efficiency ofDSSC based on Ag-modified TiO2 nanotube array decreasesagain, as the pulse deposition time used for depositing Agnanoparticles is increased to 2 s/6 s, due to agglomeration of AgNPs and formation of some particles with larger size resultedfrom longer pulse deposition time as shown in Fig. S1. Suchagglomeration of Ag NPs on TiO2 nanotubes results in lesssurface area for dye loading as well as ineffective light absorp-tion and charge separation.To further improve energy conversion efficiency of DSSC

based on TiO2 nanotubes, treatment of TiO2 nanotubes with highconcentration TiCl4 is carried out in this work. It has been re-ported that TiCl4 treatment can enhance dye absorption on thesurface, improve the electron transport properties in nano-crystalline TiO2 films, and increase the charge separation effi-ciency[49e51]. After being treated in TiCl4 aqueous solution, TiO2

nanotube arrays are sensitized with N719 dye and then integratedinto DSSCs. Fig. 7 compares the JeV characteristics of DSSCsbased on TiCl4 treated bare and Ag-modified TiO2 nanotubes asa function of different pulse deposition time (0.5 s/1.5 s, 1 s/3 s,and 2 s/6 s). Detailed photovoltaic performance parameters arepresented in Table 2. Compared to untreated bare TiO2 nano-tubes, the energy conversion efficiency of DSSC based on TiCl4treated bare TiO2 nanotubes is increased to 2.61%, and suchenhancement is mainly originated from the increased short-circuit current density. Likewise, the efficiency of DSSC basedon Ag@TiO2eTiCl4 0.5 s/1.5 s (TiCl4-treated TiO2 nanotubesmodified with Ag NPs via pulse deposition time of 0.5 s/1.5 s)increase to 2.56% presenting a 100% enhancement of efficiencycompared to DSSC based on Ag@TiO2 0.5 s/1.5 s (1.28%). Theshort-circuit current density of DSSC based on Ag@TiO2eTiCl40.5 s/1.5 s shows slight enhancement than that of DSSC based onbare TiO2eTiCl4 and the enhancement is not significant due tosmall amount of Ag loading resulted from short pulse depositiontime of Ag. In particular, the effect of Ag modification toenhance energy conversion efficiency is more significant forDSSC based on Ag@TiO2eTiCl4 1 s/3 s (TiCl4 treated TiO2

modified with Ag NPs via pulse deposition time of 1 s/3 s),which delivers the highest energy conversion efficiency (3.82%)with Voc of 0.75 V, Jsc of 9.27 mA/cm2 and FF of 0.55 (Table 2),which represents a 46.4% increase in cell efficiency compared tothat of DSSC based on bare TiO2eTiCl4 (TiCl4-treated TiO2

nanotubes) nanotube array (2.61%). These results confirm thatAg NPs loaded on TiO2 nanotube surface significantly increasethe performance of DSSCs.

Fig. 7 JeV characteristics of DSSCs based on different TiCl4-treatedTiO2 nanotube arrays (Ag@TiO2eTiCl4 0.5 s/1.5 s: TiO2

nanotubes modified with Ag nanoparticles via pulse depositiontime of 0.5 s/1.5 s; Ag@TiO2eTiCl4 1 s/3 s: TiO2 nanotubesmodified with Ag nanoparticles via pulse deposition time of 1 s/3 s; Ag@TiO2eTiCl4 2 s/6 s: TiO2 nanotubes modified with Agnanoparticles via pulse deposition time of 2 s/6 s; bare TiO2e

TiCl4: unmodified TiO2 nanotubes).

6 X. Luan and Y. Wang: J. Mater. Sci. Technol., 2014, 30(1), 1e7

It is, however, noteworthy that DSSC based on Ag@TiO2eTiCl4 2 s/6 s shows a large decrease in efficiency compared toDSSC based on TiCl4 treated bare TiO2 nanotubes. DSSC basedon Ag@TiO2eTiCl4 2 s/6 s (TiCl4 treated TiO2 nanotubesmodified with Ag NPs via pulse deposition time of 2 s/6 s) de-livers a low energy conversion efficiency of 1.55% with Voc of0.63 V, Jsc of 3.66 mA/cm2, FF of 0.67, as summarized inTable 2. It can be observed that the drastic efficiency drop ofDSSCs based on Ag@TiO2eTiCl4 2 s/6 s is mainly originatedfrom the decreased short-circuit current. The decrease of short-circuit current may be attributed to the following reasons.Long pulse deposition time (2 s/6 s) may cause agglomeration ofAg NPs on the TiO2 nanotubes surface and this effect is moredistinct for TiO2 nanotubes treated with TiCl4. TiCl4 treatmentyields very tiny TiO2 nanoparticles on the surface of TiO2

nanotubes (Fig. S2(a and b)), some of which may partially plugthe top entrances of TiO2 nanotubes (Fig. S2(c)), resulting inmore sites for Ag deposition. When the deposition time is verylong and since the tops of TiO2 nanotubes are partially pluggedand are much more exposed to electrochemical deposition thantube walls, large agglomerates of Ag NPs would form on the topof TiO2 nanotubes without much deposition of Ag NPs alongtube walls (Fig. S2(def)), leading to reduced plasmon effect anddecreased surface area for dye loading. In the case of surfaceplasmon, metallic nanoparticles must be smaller than thewavelength of the exciting light in order for electrons to oscillatetogether in response to the electric field of the excitation light.Agglomerates of Ag NPs as well as those with larger particle size(more than 200 nm) are much more difficult to be excited byradical light. Another issue is that Ag agglomerates or larger Ag

Table 2 Photovoltaic characteristics of dye-sensitized solar cells shownin Fig. 7

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

Ag@TiO2eTiCl4 0.5 s/1.5 s 5.41 0.78 0.61 2.56Ag@TiO2eTiCl4 1 s/3 s 9.27 0.75 0.55 3.82Ag@TiO2eTiCl4 2 s/6 s 3.66 0.63 0.67 1.55

Bare TiO2eTiCl4 5.39 0.75 0.64 2.61

particles have more severe plasmon decay. The intensity ofelectrons transferred to TiO2 depends on the distance that plas-mon travel from radical light impact position to TiO2. Vesseuret al.[52] pointed out that plasmon excited by radical light, whichare emitted radially from the impact position, are lost throughabsorption unless they are converted to a target. Therefore,plasmon excited on the surface of TiO2 decays when transferredto the bottom of nanotube. Hence, plasmon decay is more severein the case of Ag@TiO2eTiCl4 2 s/6 s, which has large Agagglomerates or particles on the tops of TiO2 nanotubes and fewAg nanoparticles along the walls.It can be seen by comparing Fig. 6 with Fig. 7 that, DSSCs

based on TiCl4 treated TiO2 nanotubes show higher energyconversion efficiency than DSSCs based on untreated TiO2

nanotube, due to larger surface area of TiO2 nanotube resultedfrom TiCl4 treatment. It can also be observed that the effectof Ag NPs to improve energy conversion efficiency andshort-circuit current is more distinct for DSSCs based onAg@TiO2eTiCl4 1 s/3 s than DSSCs consisting of TiO2 nano-tubes modified with Ag via other shorter or longer pulsedeposition time, because this pulse deposition time (1 s/3 s)yields evenly-distributed and uniformly-sized Ag nanoparticles,leading to enhanced surface plasmon resonance and subsequentimproved optical absorption. The particle size of Ag NPs syn-thesized via commonly-used methods such as photo reductiondeposition methods is too large and uncontrollable, while pulsecurrent deposition can provide uniformly distributed nano-particles which are necessary for plasmon effect in DSSC.

4. Conclusion

In summary, TiO2 nanotube arrays are synthesized by anodizingTi foils in EG electrolytes with a small amount of NH4F and H2Ounder alternating-voltage conditions, and subsequent heat treat-ment transforms them to anatase phase for modification with Agnanoparticles via electrochemical deposition, for enhanced effi-ciencies of DSSCs. Ag nanoparticles with a size distribution up to10e40 nm are formed on the side walls of TiO2 nanotubes.Metallic Ag formation is confirmed by EDS and XPS analyses. Inaddition, the loading amount of Ag nanoparticle can be easilyoptimized by adjusting pulse deposition time. Most of DSSCsbased on Ag-modified TiO2 nanotube arrays show significantlyhigher energy conversion efficiencies than DSSC consisting ofbare nanotube array of the same length. Such enhancement isascribed to the surface plasmon resonance induced by uniformlydeposited Ag NPs. Increasing the Ag deposition content is able tofurther improve the plasmon effect, but superfluous Ag contentyields not as good as or worse photovoltaic performance, andDSSC based on TiO2 nanotubes modified with Ag via pulsedeposition time of 1 s/3 s display the highest energy conversionefficiency. TiCl4 treatment for TiO2 nanotubes can further enhanceenergy conversion efficiency of DSSCs by increasing surface areaof TiO2 nanotubes for more dye loading. Our results confirm thatmodifying anodic TiO2 nanotube arrays with noble metal nano-particles is an effective approach for maximizing efficiency ofDSSCs and will advance the DSSC technology.

AcknowledgmentsThis work is supported by LABOR e RCS grant and BP e

Gulf of Mexico Research Initiative (GRI) grant. The authorsacknowledge Materials Characterization Center at LSU for usingXRD and SEM. X.N. Luan acknowledges LSU Graduate SchoolEnrichment Award.

X. Luan and Y. Wang: J. Mater. Sci. Technol., 2014, 30(1), 1e7 7

Appendix A. Supplementary data

Supplementary data related to this article can be found athttp://dx.doi.org/10.1016/j.jmst.2013.09.007.

REFERENCES

[1] J. Preat, D. Jacquemin, E.A. Perpete, Energy Environ. Sci. 3 (2010)891e904.

[2] Z. Ning, Y. Fu, H. Tian, Energy Environ. Sci. 3 (2010) 1170e1181.[3] B. Oregan, M. Grätzel, Nature 353 (1991) 737e740.[4] M. Grätzel, Coord. Chem. Rev. 111 (1991) 167e174.[5] A. Kay, M. Grätzel, J. Phys. Chem. 97 (1993) 6272e6277.[6] M.K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphrybaker,

E. Muller, P. Liska, N. Vlachopoulos, M. Grätzel, J. Am.Chem. Soc. 115 (1993) 6382e6390.

[7] M. Grätzel, K. Kalyanasundaram, Curr. Sci. 66 (1994) 706e714.[8] M. Grätzel, Renew. Energy 5 (1994) 118e133.[9] W. Jia, S. Dang, H. Liu, Z. Zhang, T. Li, X. Liu, B. Xu, J. Mater.

Sci. Technol. 29 (2013) 415e418.[10] S.L. Chung, C.M. Wang, J. Mater. Sci. Technol. 28 (2012) 713e722.[11] Y. Wang, X. Cui, Y. Zhang, X. Gao, Y. Sun, J. Mater. Sci. Technol.

29 (2013) 123e127.[12] R. Ranjusha, P. Lekha, K.R.V. Subramanian, V. Nair Shantikumar,

A. Balakrishnan, J. Mater. Sci. Technol. 27 (2011) 961e966.[13] D. Guan, C. Cai, Y. Wang, J. Nanosci. Nanotechnol. 11 (2011)

3641e3650.[14] M. Law, L.E. Greene, J.C. Johnson, R. Saykally, P.D. Yang, Nat.

Mater. 4 (2005) 455e459.[15] J.Y. Liao, B.X. Lei, Y.F. Wang, J.M. Liu, C.Y. Su, D.B. Kuang,

Chem. Eur. J. 17 (2011) 1352e1357.[16] B.X. Lei, J.Y. Liao, R. Zhang, J. Wang, C.Y. Su, D.B. Kuang, J.

Phys. Chem. C 114 (2010) 15228e15233.[17] T.P. Chou, Q. Zhang, G. Cao, J. Phys. Chem. C 111 (2007) 18804e

18811.[18] G.K. Mor, K. Shankar, M. Paulose, O.K. Varghese, C.A. Grimes,

Nano Lett. 6 (2006) 215e218.[19] K. Zhu, T.B. Vinzant, N.R. Neale, A.J. Frank, Nano Lett. 7 (2007)

3739e3746.[20] P. Roy, D. Kim, K. Lee, E. Spiecker, P. Schmuki, Nanoscale 2

(2010) 45e59.[21] D. Guan, Y. Wang, Nanoscale 4 (2012) 2968e2977.[22] M. Chang, L. Wu, X. Li, W. Xu, J. Mater. Sci. Technol. 28 (2012)

594e598.[23] W. Shockley, J. Appl. Phys. 32 (1961) 1402.[24] H.J. Snaith, Adv. Funct. Mater. 20 (2010) 13e19.[25] F. Zhuge, J. Qiu, X. Li, X. Gao, X. Gan, W. Yu, Adv. Mater. 23

(2011) 1330e1334.[26] X. Wang, S. Karanjit, L. Zhang, H. Fong, Q. Qiao, Z. Zhu, Appl.

Phys. Lett. 98 (2011) 082114.

[27] T. Kume, S. Hayashi, H. Ohkuma, K. Yamamoto, Jpn. J. Appl.Phys. 34 (Part 1) (1995) 6448e6451.

[28] M. Westphalen, U. Kreibig, J. Rostalski, H. Luth, D. Meissner, Sol.Energy Mater. Sol. Cells 61 (2000) 97e105.

[29] C. Hagglund, M. Zach, B. Kasemo, Appl. Phys. Lett. 92 (2008)053110.

[30] I. Mora-Sero, J. Bisquert, J. Phys. Chem. Lett. 1 (2010) 3046e3052.

[31] S.J.B. Sambur, T. Novet, B.A. Parkinson, Science 330 (2010) 63e66.

[32] M. Grätzel, Acc. Chem. Res. 42 (2009) 1788e1798.[33] M.K. Nazeeruddin, T. Bessho, L. Cevey, S. Ito, C. Klein, F. De

Angelis, S. Fantacci, P. Comte, P. Liska, H. Imai, M. Grätzel,J. Photochem. Photobiol. A 185 (2007) 331e337.

[34] P. Wang, S.M. Zakeeruddin, J.E. Moser, M.K. Nazeeruddin,T. Sekiguchi, M. Grätzel, Nat. Mater. 2 (2003) 402e407.

[35] X. Luan, D. Guan, Y. Wang, J. Phys. Chem. C 116 (2012)14257e14263.

[36] J. Zhang, C. Fu, X. Yang, W. Cao, J. Inorg. Organomet. Polym.Mater. 21 (2011) 43e49.

[37] Z. Han, J. Zhang, X. Yang, H. Zhu, W. Cao, Sol. Energy Mater. Sol.Cells 94 (2010) 755e760.

[38] H.A. Atwater, A. Polman, Nat. Mater. 9 (2010) 205e213.[39] H.R. Stuart, D.G. Hall, Appl. Phys. Lett. 73 (1998) 3815e3817.[40] D.M. Schaadt, B. Feng, E.T. Yu, Appl. Phys. Lett. 86 (2005)

063106.[41] D. Derkacs, S.H. Lim, P. Matheu, W. Mar, E.T. Yu, Appl. Phys.

Lett. 89 (2006) 093103.[42] S. Pillai, K.R. Catchpole, T. Trupke, M.A. Green, J. Appl. Phys.

101 (2007) 093105.[43] J.L. Wu, F.C. Chen, Y.S. Hsiao, F.C. Chien, P. Chen, C.H. Kuo,

M.H. Huang, C.S. Hsu, ACS Nano 5 (2011) 959e967.[44] B.P. Rand, P. Peumans, S.R. Forrest, J. Appl. Phys. 96 (2004)

7519e7526.[45] C.S. Chou, R.Y. Yang, C.K. Yeh, Y.J. Lin, Powder Technol. 194

(2009) 95e105.[46] S. Agarwala, G.W. Ho, J. Solid State Chem. 189 (2012) 101e107.[47] Y. Lai, H. Zhuang, K. Xie, D. Gong, Y. Tang, L. Sun, C. Lin,

Z. Chen, New J. Chem. 34 (2010) 1335e1340.[48] J.Cl. Puippe, N. Ibl, Plat. Surf. Finish. 67 (1980) 68e72.[49] L.Y. Zeng, S.Y. Dai, K.J. Wang, X. Pan, C.W. Shi, L. Guo, Chin.

Phys. Lett. 21 (2004) 1835e1837.[50] C.J. Barbe, F. Arendse, P. Comte, M. Jirousek, F. Lenzmann,

V. Shklover, M. Grätzel, J. Am. Ceram. Soc. 80 (1997) 3157e3171.

[51] P.M. Sommeling, B.C. O’Regan, R.R. Haswell, H.J.P. Smit,N.J. Bakker, J.J.T. Smits, J.M. Kroon, J.A.M. van Roosmalen,J. Phys. Chem. B 110 (2006) 19191e19197.

[52] E.J.R. Vesseur, J. Aizpurua, T. Coenen, A. Reyes-Coronado,P.E. Batson, A. Polman, MRS Bull. 37 (2012) 752e760.