Sp

YYa

b

a

ARRAA

KZSD

1

eaItcnAamth

bdfiieaa

0d

Electrochimica Acta 56 (2011) 6517– 6523

Contents lists available at ScienceDirect

Electrochimica Acta

jou rn al hom epa ge: www.elsev ier .com/ locate /e lec tac ta

ynthesis of Zn-doped TiO2 microspheres with enhanced photovoltaicerformance and application for dye-sensitized solar cells

u Zhanga, Lingling Wanga,b, Bingkun Liua, Jiali Zhaia, Haimei Fana, Dejun Wanga,anhong Lina, Tengfeng Xiea,∗

State Key Laboratory of Theoretical and Computational Chemistry, College of Chemistry, Jilin University, Changchun 130023, ChinaState Key Laboratory of Supramolecular Structure and Materials, College of Chemistry, Jilin University, Changchun 130012, China

r t i c l e i n f o

rticle history:eceived 19 November 2010eceived in revised form 21 February 2011ccepted 30 April 2011vailable online 7 May 2011

a b s t r a c t

Zn-doped TiO2 microspheres have been synthesized by introducing a trace amount of zinc nitrate hex-ahydrate to the reaction system. Scanning electron microscope (SEM), field-emission scanning electronmicroscope (FESEM), X-ray diffraction (XRD), and X-ray photoelectron spectroscopy (XPS) have beenutilized to characterize the samples. Both surface photovoltage spectroscopy (SPS) technique based on

eywords:n-doped TiO2

urface photovoltageye-sensitized solar cells

lock-in amplifier and transient photovoltage (TPV) measurement reveal that the slight doping of Zn canpromote the separation of photo-generated charges as well as restrain the recombination due to thestrong interface built-in electric field and the decreasing of surface trap states. The photovoltaic parame-ters of dye-sensitized solar cells (DSSCs) based on Zn-doped TiO2 are significantly better, compared to thatof a cell based on undoped TiO2. The relation between the performance of DSSCs and their photovoltaicproperties is also discussed.

. Introduction

Dye-sensitized solar cells (DSSCs) based on nanoporous TiO2lectrodes have attracted considerable interest because of their rel-tively high efficiency, low cost and facile fabrication process [1–8].n the past two decades, many research efforts have been madeo enhance the power conversion efficiencies for the purpose ofommercialization, such as the improvement of the semiconductoranocrystal anode [9], dye [10], electrolyte [11], and cathode [12].mong all these areas, the nanocrystal TiO2 photoelectrode, whichdsorbs dyes, transmits electrons, and scatters light, is one of theost important components. Thus, the synthesis of TiO2 nanoma-

erials for DSSCs and the studies of their photoelectric propertiesave attracted much attention.

Doping of TiO2 nanomaterials with metals or nonmetals haseen widely studied for more than ten years. Most of the metal-oped (nonmetal-doped) TiO2 nanomaterials have been exploredor photocatalysis. In recent years, some researchers find that dop-ng the TiO2 film with metal ions could be a promising approach tomprove the electrons transfer in the TiO2-based nanostructured

lectrode of DSSCs [13,14]. For example, Lü et al. [15] prepared

series of well-crystallized Nb-doped anatase TiO2 nanoparticlesnd found that the Nb doping leads to a significant increase of

∗ Corresponding author. Tel.: +86 431 85168093.E-mail address: xietf@jlu.edu.cn (T. Xie).

013-4686/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.oi:10.1016/j.electacta.2011.04.118

© 2011 Elsevier Ltd. All rights reserved.

powder conductivity, and a positive shift in the flat-band potential.Similar results were also found when Ta-doped TiO2 was appliedas the photoanodes of DSSCs [16]. Wang and Teng [17] reportedthe synthesis of a Zn-doped TiO2 via Zn2+ introduction into a lay-ered titanate followed by hydrothermal treatment and calcination,which shows its capability for efficient electron transport in DSSCsdue to lower density of empty trap states. Kim et al. [18] pre-pared Cr-doped TiO2 used for the photoanodes of DSSCs, and theenhanced performance is associated with the adjusting of bandstructure of TiO2. Another approach on the optimization of theTiO2 photoelectrode is to improve the light harvesting of the DSSCsby adding scattering layers. It is reported that sub-micrometersized microspheres could enhance the light-harvesting capabilityof the DSSCs because of their comparative particle sizes to opticalwavelengths [19–21]. Although the synthesis of TiO2 with variesnanostructure is necessary for fabricating high efficient DSSCs, tothe best of our knowledge, the synthesis of near-monodispersespherical Zn-doped TiO2 has been scarcely reported.

In addition, the understanding of photovoltaic performancerelated to the charge transfer at the nanostructure is also impor-tant. The surface photovoltage (SPV) method is a well-establishedtechnique for the characterization of semiconductors, which couldprovide direct information about spatial separation (caused by sur-

face states, built-in field and diffusion, etc.) of photo-generatedcharge carriers in comparison with electrochemical methods [22].The surface photovoltage spectroscopy (SPS) gives informationof photovoltage as a function of incident photon energy. And

6 mica A

tomnfmtbcdgbtpd

2

2

rcjepZv1sfabshenw

2e

2

dmIpte

dSbTwbmt

2

nE

518 Y. Zhang et al. / Electrochi

he transient photovoltage (TPV) provides dynamic propertiesf photoinduced charge carriers. So it should be an appropriateethod in studying the photovoltaic properties of semiconductor

anocrystals used in the DSSCs. Herein, we present the methodor the synthesis of near-monodisperse Zn-doped anatase TiO2

icrospheres with a doping level in the range of 0.25–1 at.%. SPSechnique based on lock-in amplfier and TPV measurement wereoth employed to characterize the behavior of photo-generatedharge carriers in the system. Then, we fabricated DSSCs: the Zn-oped TiO2 film serves as the photoanode, an F-doped SnO2 (FTO)lass coated with a Pt layer is the photocathode and the spaceetween the two electrodes is filled with an electrolyte containinghe triiodide/iodide (I3−/I−) redox couple. The relation between thehotovoltage characterization and the efficiency of DSSCs is alsoiscussed.

. Experimental

.1. Preparation of Zn-doped TiO2 microspheres

All the chemicals were of analytical grade and were used aseceived without further purification. In a typical synthesis, aertain amount of zinc nitrate hexahydrate (Zn(NO3)2·6H2O, Tian-in Guangfu Fine Chemical Research Institute) was dissolved inthylene glycol (EG, 25 mL; Beijing Chemical Works) at room tem-erature. Then, tetrabutyl titanate (TTIP, 1 mL, ≥98.0%; Shanghaihanyun Chemical Co., Ltd) was added to the above solution underigorous stirring. The system was bubbled with nitrogen for about0 min to remove the oxygen and water. After that, the system wasealed with parafilm (Pechiney Plastic Packaging) and was stirredor a further 24 h. The samples were synthesized by pouring thebove solution into an acetone (Beijing Chemical Works, ≥99.5%)ath (100 mL) containing ultrapure water (1 mL) under vigoroustirring for 10 min. After aging for 16 h, the white precipitate wasarvested by centrifugation, followed by washing with acetone andthanol five times to remove residual EG from the surface of the tita-ium glycolate particles. After the washing treatment, the samplesere kept in a 50 ◦C oven overnight before characterization.

For the heat treatment, the samples were annealed at 450 ◦C for h with a heating rate of 1 ◦C min−1. The as-prepared samples werexpressed as x% Zn–TiO2, in which x% refers to the Zn/Ti molar ratio.

.2. Preparation of Electrodes and Solar Cells

The TiO2 paste was prepared by mixing Zn-doped TiO2 pow-er (0.2 g) with absolute alcohol (1 mL), followed by ultrasonic andagnetic stirring until the particles was dispersed homogeneously.

n our experiment, the doctor blade technique was used to pre-are the porous Zn-doped TiO2 layer on an FTO glass substrate;he resulted film thickness was about 20 �m. The Zn-doped TiO2lectrodes were then sintered at 450 ◦C for 30 min.

The dye cis-bis(isothiocyanato)bis(2,2′-bipyridyl-4,4′-icarboxylato)-ruthenium(II) (N3) was purchased from SolaronixA. The N3-sensitized semiconductor electrodes were obtainedy soaking the electrodes in a 0.5 mM solution of N3 for 24 h.he working electrode of dye-sensitized semiconductor filmas cohered together with a platinized conducting glass [23]

y epoxy resin. The electrolyte was consisted of 0.6 M butyl-ethylimidazolium iodide (BMII), 0.05 M I2, 0.1 M LiI, and 0.5 M

ert-butylpyridine in 1:1 acetonitrile/valeronitrile [24].

.3. Characterization

The morphologies of the samples were observed on a scan-ing electron microscope (SEM, Shimadzu SSX-550) and an XL 30SEM FEG field-emission scanning electron microscope (FESEM;

cta 56 (2011) 6517– 6523

FEI Company). X-ray diffraction (XRD) data were collected ona Rigaku D/Max 2550 X-ray diffractometer with CuK� radiation(� = 1.5418 A) in the range of 20–70◦ (2�) at a scanning rate of5◦ min−1. X-ray photoelectron spectroscopy (XPS) measurementswere performed using a Thermo VG Scientific Escalab 250 spec-trometer with monochromatized AlK� excitation. The specificsurface areas of the samples were determined by N2-BET methodon an ASAP-2020 instrument (Micromeritics Corporation).

The SPS system was made up of a source of monochro-matic light, a lock-in amplifier (SR830-DSP) with a light chopper(SR540), a sample cell and a computer [22]. A 500 W xenon lamp(CHFXQ500 W, Global xenon lamp power) and a grating monochro-mator (Omni-5007, no. 09010, Zolix) provided monochromaticlight. A low chopping frequency of ∼23 Hz is used. The con-struction of the SPS sample cell is a sandwich-like structure ofFTO–sample–FTO. The FTO electrode has a resistance of 20 �/sq.and a high transmittance of 80% in the visible spectral range(with substrate). Field-induced surface photovoltage spectroscopy(FISPS) measurement is a supplement to the SPS method, which isdeveloped by our group. In FISPS, external electric fields are appliedbetween the two electrodes, which have been described in detailelsewhere [25].

TPV measurements were carried out in a device, which wasdescribed in our previous paper [26]. A sample chamber like aparallel-plate capacitor consisted of the Zn-doped TiO2 electrode,a piece of 10 �m thick mica and a platinum wire gauze electrode. Alaser radiation pulse (wavelength of 355/532 nm and pulse widthof 5 ns) from a third-harmonic Nd:YAG laser (Polaris II, New WaveResearch, Inc.) was used to excite the TPV. The signals were reg-istered by a 500 MHz digital phosphor oscilloscope (TDS 5054,Tektronix) with a preamplifier. The formation of a TPV signal wasdetermined by the factors of light absorption, transport of excesscarriers, structural and electric characteristics of a semiconductingmaterial. Both the TPV and SPS measurements were performed inair atmosphere and at room temperature.

The current density–voltage (I–V) curves were recorded by anelectrochemistry workgroup (CHI 630b, Shanghai). A 500 W xenonlamp was used as the light source. The incident light intensity was100 mW cm−2, which was measured with a radiometer (Photo-electronic Instrument Co., attached to Beijing Normal University,China). The effective area of the cell is 0.2 cm2.

3. Results and discussion

Fig. 1 shows the SEM images of the samples before calcina-tion. As is shown in Fig. 1a, when no zinc nitrate hexahydrate isadded, only irregular particles are observed. However, it is inter-estingly found that relatively small microspheres with unequal sizeare obtained by introducing trace amounts of zinc nitrate hexahy-drate into the reaction system (Fig. 1b); when the Zn/Ti molar ratiois 0.5%, near-monodisperse spheres are obtained (Fig. 1c). With theZn/Ti molar ratio reaching 1%, the microspheres become smallerand irregular again (Fig. 1d).

Fig. 2 shows FESEM images of 0.5% Zn–TiO2 microspheres beforeand after annealing at 450 ◦C. It is clearly seen that after calcination,the sample keeps the original morphology and the average size ofthe particles is slightly reduced.

The formation mechanism for the Zn-doped TiO2 microspheresmay be attributed to a cation-controled hydrolysis of the Ti pre-cursor. Our group has reported the synthesis of Bi-doped TiO2microspheres and confirms that Bi3+ could assist in the formation of

the spherical structure [27]. We assume that the Zn2+ cations playa similar role in the system. After a titanium precursor contain-ing a trace amount of Zn2+ is injected into the acetone bath, Zn2+

adsorbed onto the surface of the nucleus and formed a positively

Y. Zhang et al. / Electrochimica Acta 56 (2011) 6517– 6523 6519

ratios

ctaam

acvatmZaTcXttOut1

Fig. 1. SEM images of samples with different Zn/Ti molar

harged layer to effectively protect against aggregation of particles,hus resulting in the near-monodisperse spheres. When more Zn2+

re added (e.g. 1% Zn–TiO2), more of the nucleus can be kept stablet the early stage of the nucleation process, and the smaller sizedicrospheres were obtained.Typical XRD patterns of the samples doped with different

mount of Zn after annealing at 450 ◦C for 2 h are shown in Fig. 3a. Itan be seen that after calcination, titanium glycolate particles con-ert into the anatase, and the average crystallite size is 15–25 nm,s calculated by the Scherer equation. Due to the low Zn content,he possible impurities such as zinc oxide were not detected. XPS

easurements were used to further investigate the states of dopedn in the samples. Fig. 3b–d shows the XPS spectra of Ti 2p, O 1s,nd Zn 2p. In Fig. 3b, the binding energy peaks corresponding toi 2p3/2 of the four samples are located at 458.4–458.6 eV, indi-ating that Ti is in the Ti4+ form in TiO2 [28]. Fig. 3c shows thePS spectra of the O1s region. The strong peak at 529.7 eV and

he shoulder at 531.8 eV are ascribed to bulk O2− from TiO2 andhe –OH, respectively [29,30]. A weak positive shift of Ti 2p and1s energy peaks for all the Zn-doped samples relative to that of

ndoped anatase sample is observed, which indicates the interac-ions between Zn and TiO2. In Fig. 3d, Zn 2p3/2 is observed at around021.9 eV. According to the experimental conditions (the Zn-doped

Fig. 2. FESEM images of 0.5% Zn–TiO2 microspheres (a) before an

of (a) 0%, (b) 0.25%, (c) 0.5% and (d) 1%. Scale bar: 1 �m.

TiO2 microspheres have been annealed at 450 ◦C) and XPS results,the Zn element should be in 2+ state bonding with oxygen [31].

The lock-in-based SPS measurements were first carried out toexplore the photovoltaic properties of the samples. The resultsare shown in Fig. 4. In Fig. 4, the black line represents the SPSresponse of undoped TiO2 sample and the colored lines repre-sent that of Zn-doped TiO2 with different Zn/Ti molar ratios. Fromthe results, we can find that all the four samples have obviousresponses at ultraviolet region. However, the undoped TiO2 haveanother response area at visible region under 525 nm (Fig. 4). Withan optical band gap about 3.2 eV, the SPS response of TiO2 (anatase)at visible region could be attributed to a sub-band-gap transitionof photo-generated charges. According to previous reports, TiO2nanoparticles always have surface trap states located between thevalence band and conduction band, which could cause sub-band-gap transition [32,33]. Therefore the disappearance of photovoltaicresponses of sub-band-gap transition in Zn-doped TiO2 means thatthe slight doping of Zn could reduce the amount of trap states.In the DSSCs, some empty surface trap states could capture theinjected electrons from conduction band of TiO2 when electrons

transfer through grain boundaries among nanocrystals, which areconsidered as one of the key processes that reduces the efficiency.

Fig. 5 shows the FISPS and corresponding phase spectra ofundoped TiO2 and Zn-doped TiO2. When positive electric fields are

d (b) after calcination at 450 ◦C for 2 h. Scale bar: 500 nm.

6520 Y. Zhang et al. / Electrochimica Acta 56 (2011) 6517– 6523

spher

amuptwwabtt

F4

Fig. 3. XRD patterns and XPS spectra of Zn–TiO2 micro

dded, the FISPS intensities of all the four samples increase dra-atically. However, when negative electric fields are applied, only

ndoped TiO2 exhibits enhanced photovoltaic responses, while thehotovoltaic responses of Zn-doped TiO2 become weak. Accordingo the reports of Wei et al. [34], when nanoporous TiO2 is contactedith FTO, a built-in electric field with a direction from FTO to TiO2ill be formed at their interface due to the Fermi level difference,

nd photo-generated electrons and holes will be firstly separated

y the built-in electric field. For our experiment, the construction ofhe photovoltaic cell is a sandwich-like structure, which means theop electrode (an FTO glass) is contacted with the sample. There-

ig. 4. SPS of Zn–TiO2 microspheres with different Zn/Ti molar ratios calcined at50 ◦C. Inset: schematic of configuration in SPS and FISPS measurements.

es with different Zn/Ti molar ratios calcined at 450 ◦C.

fore we can explain this phenomenon by analyzing the strengthand direction of the built-in electric field at the interface betweenTiO2 and FTO. For undoped TiO2, both positive and negative exter-nal electric fields can produce enhanced photovoltaic responses,indicating the separation of photo-generated charge carrier can becontrolled easily by external electric field and the built-in electricfield of this sample is very weak. However, for Zn-doped TiO2, thenegative external electric fields do not affect the photovoltage effi-ciently (or slightly weakened the photovoltage), which indicatesthat strong built-in electric fields are formed in their interfaces andcontrol the charge separation process. When the direction of exter-nal electric field is consistent with that of built-in electric field,the photovoltaic responses should be enhanced by external elec-tric field. In contrary, the photovoltaic responses will be weakened.Therefore we can judge the directions of built-in electric fields inthese Zn-doped TiO2 samples, which should be from FTO to Zn-doped TiO2. According to the phase spectra, the phase curves ofZn-doped TiO2 are close when varies external electric fields (from+1 V to −0.5 V) are applied across the sample, indicating the direc-tions of photo-generated charges movement in Zn-doped TiO2 arenot changed in most cases, which further proved that strong built-in electric fields should be formed in these samples. Our FISPSresults indicate that the DSSCs based on Zn-doped TiO2 could havehigher electron collection efficiency at TiO2/FTO interface becauseelectrons diffusing to the interface could be efficiently swept to theFTO by the strong interface built-in electric fields with directionsfrom FTO to TiO2. It can be observed that the FISPS are split intotwo peaks for undoped TiO2 under positive biases. The reason forthis is not yet clear and needs further study.

The TPV measurements were carried out to investigate the

transport properties of photo-generated charge carriers in the sam-ples. In these measurements, a piece of mica is inserted between thefilm and top electrode. So the TPV responses only reflect the sep-aration of photo-generated charge carriers in the nanostructure.

Y. Zhang et al. / Electrochimica Acta 56 (2011) 6517– 6523 6521

F ith diZ

FilroahdtfZ

tsatAi

Faw

ig. 5. FISPS and corresponding phase spectra (insets) of Zn–TiO2 microspheres wn–TiO2, (d) 1.0% Zn–TiO2.

ig. 6 shows the TPV responses of nanoporous TiO2 films contain-ng different amounts of Zn with the film surface exposed to theaser pulse before and after dye sensitizing. According to the TPVesults, the following features can be found: (1) The TPV responsesf all samples have the positive sign and show pronounced apexest about ∼10−7 s. (2) The TPV intensities of Zn-doped TiO2 areigher than that of undoped TiO2. (3) The TPV maximums (tmax) areistributed between 10−5 s and 10−3 s. After dye sensitizing withhe Zn/Ti molar ratios increased from 0.0% to 1%, the tmax reachesrom 2 × 10−5 s to 1 × 10−3 s and apparent retardation of tmax forn-doped TiO2 are observed.

It has been reported that the fast carrier separation (<10−7 s)akes place within one nanoparticle, and then inter-particle diffu-ion of photo-generated charge carriers would be dominant for TPV

−7

t long timescale [34]. So the apex at ∼10 s can be considered ashe transformation point of the two charge separation processes.ccording to previous reports [26,34,35], a positive TPV response

mplies some mechanisms may exist such as that negative charges

ig. 6. TPV curves of Zn–TiO2 films with different Zn/Ti molar ratios: (a) the samples arend pulse width of 5 ns before dye sensitizing. Inset: schematic of TPV configuration with

avelength of 532 nm and pulse width of 5 ns after dye sensitizing. Inset: schematic of TP

fferent Zn/Ti molar ratios calcined at 450 ◦C: (a) TiO2, (b) 0.25% Zn–TiO2, (c) 0.5%

transfer towards the bottom electrode and positive charges accu-mulate at the surface area nearby. For films without dye sensitizing,both electrons and holes would diffuse from surface to bottom dueto the concentration gradient, but the diffusion coefficient of elec-trons is larger than that of holes [36], resulting in positive TPVresponses (as shown in Fig. 6a). When the films are sensitized withN3 dye and exited by a laser pulse of 532 nm, the dye moleculeswill absorb light and inject electrons into the TiO2 conductionband, the electrons then diffuse to the bottom electrode, which alsoresults in a positive TPV (as shown in Fig. 6b). The enhanced TPVof Zn-doped TiO2 film means the separation of photo-generatedelectrons and holes is much more pronounced in these films. Withthe increasing amount of Zn, the retardation of tmax at longer timesmeans the charge recombination time is gradually delayed. When

electrons transport in the nanostructure, the empty trap states inthe nanocrystals could capture them and hinder the separationof electron and holes. As we have discussed above, the doping ofZn significantly decreased the amount of trap states, hence the

excited with a laser radiation pulse with a power of 50 �J, wavelength of 355 nmTiO2, (b) the samples are excited with a laser radiation pulse with a power of 50 �J,V configuration with dye-sensitized TiO2.

6522 Y. Zhang et al. / Electrochimica A

FT

ebtcm

DiZmqtpDeelsEtbaab

cZai�btos[o

TP

ig. 7. I–V characteristics of the N3-sensitized TiO2 cell and N3-sensitized Zn-dopediO2 cells.

lectron transport will be freer and charge recombination coulde reduced. Another reason for the delayed recombination may behe built-in electric field between the bottom FTO and TiO2, whichould also promote the injection of electrons to FTO when electronsove to the bottom electrode.Fig. 7 shows current–voltage (I–V) characteristic curves of four

SSCs. We summarize the corresponding photovoltaic parametersn Table 1. The results show that with the increasing amount ofn, the open-circuit voltage (VOC) is increased. The VOC of DSSCsay be affected by some factors. Upon illumination, the electron

uasi-Fermi level (E∗Fn) of the TiO2 film is shifted and close to

he conduction band minimum, while the energy position of thelatinized counter electrode maintains the Eredox. Photovoltage ofSSCs is thus determined by the energy difference between thelectron E∗

Fn and Eredox [37]. When electrons diffuse in the film,mpty trap states with low energy levels will capture electrons toower the E∗

Fn. As we have found that undoped TiO2 have more traptates compared with other samples, it is rational to conclude the∗Fn is low in this cell, resulting in a small VOC. The other reason forhe variation of VOC may be the recombination rate. A lower recom-ination rate generally leads to a larger VOC. With the increasedmount of Zn, the tmax at which the TPV begin to decrease is gradu-lly retarded as shown in TPV results, indicating the recombinationecome more slowly and the VOC would be increased.

Both short-circuit current (JSC) and power conversion efficien-ies (�) have a maximum value for the cell constructed with 0.5%n–TiO2. The sample (0.5% Zn–TiO2) has the strongest photovolt-ge as shown in Fig. 6, indicating the charge separation efficiencyn this sample is the best. It may be one of the reasons for its high

and JSC. When the Zn/Ti molar ratio reaches 1%, the JSC and �ecome relatively lower again. According to Fig. 6b, it is noted thathe TPV uphill (10−6 to 10−4 s) of 1% Zn–TiO2 is weaker than thatf the other two dye-sensitized Zn-doped samples, indicating its

eparation process is weakened by some factors. Wang and Teng17] figure that the Zn-doping will enhanced the band bendingf nanoparticles, which may increase the barrier of grain bound-

able 1hotovoltaic parameters of DSSCs constructed with Zn–TiO2 electrodes.

Sample VOC (V) JSC (mA/cm2) FF �(%) SBET (m2/g)

TiO2 0.63 2.00 0.46 0.58 19.320.25% Zn–TiO2 0.64 12.41 0.42 3.32 30.820.5% Zn–TiO2 0.73 14.58 0.44 4.63 48.691.0% Zn–TiO2 0.75 7.14 0.49 2.60 52.96

cta 56 (2011) 6517– 6523

ary. Although the barriers could inhibit the carrier recombination,they will also limit the transport of electrons in the nanostruc-ture. Therefore, excess Zn-doping would weaken the performanceof DSSC.

We should also consider the particle size effect because theirmorphologies and sizes are different. As is mentioned above,the diameters of these microspheres are about several hundrednanometers. Taking into account the density of anatase and assum-ing compact spherical particle, the calculated surface area forcompact microspheres with 200 nm diameter should be smallerthan 10 m2/g. However, as summarized in Table 1, the specific sur-face areas such as 0.25% and 0.5% Zn–TiO2 measured by N2-BETmethod are 30.82 and 48.69 m2/g, which indicate the microspheresshould have a porous structure. These microspheres consist ofpacked nanocrystallites, and the intervals of nanocrystallites pro-vide the added surface area. The specific surface area of undopedTiO2 (only 19.32 m2/g) is significantly smaller than that of Zn-dopedTiO2. Thus, the amount of chemisorptions of dye molecules onundoped TiO2 could be lower than that of Zn-doped TiO2, whichshould be one of the reasons for its very low efficiency. The photo-voltaic performance of DSSCs based on Zn–TiO2 does not match theBET results, indicating the major factor that could improve photo-voltaic performance is the separation of photo-generated chargesinduced by Zn-doping. The microspheres also have light scatteringeffect due to their sub-micrometer sizes. The sample consisting oflarge microspheres possesses a large light scattering efficiency [19].So the light scattering effect is not the reason for the poor perfor-mance of standard DSSC. The fill factors (FF) of all the four cells arepoor. One possible reason may be that the microspheres are con-structed by densely packed nanoparticles. Thus, the electrolyte isdifficult to penetrate into its interior, resulting in a decrease in theFF [38].

4. Conclusion

In this work, a series of Zn-doped sub-micrometer spheres wereprepared and characterized. Then we have qualitatively investi-gated the photovoltaic properties of the samples by SPS and TPVtechniques. It is obvious that the Zn-doping eliminate the sub-band-gap photovoltage due to the reducing of trap states. FISPSmeasurements indicate that the Zn-doped TiO2/FTO interface hasa strong built-in electric field, which may help the collection ofphoto-generated electrons. The TPV responses further reveal thatZn-doped TiO2 not only promote the charge separation but alsodelay the charge recombination in these films. The DSSCs basedon Zn-doped TiO2 show improved performance. Some factors thataffect the performance are discussed with their photovoltaic prop-erties. Our experimental results prove that SPS and TPV techniquescan be very useful methods in studying the photovoltaic propertiesof DSSCs.

Acknowledgements

For financial support, we are grateful to the National BasicResearch Program of China (973 Program) (No. 2007CB613303),the National Natural Science Foundation of China (No. 20703020,20873053) and the Science and Technology Developing Funding ofJilin Province (No. 201115012).

References

[1] B. O’Regan, M. Grätzel, Nature 353 (1991) 737.

[2] M. Grätzel, Nature 414 (2001) 338.[3] M.K. Nazeeruddin, A. Kay, I. Rodicio, R. Humpbry-Baker, E. Müller, P. Liska, N.

Vlachopoulos, M. Grätzel, J. Am. Chem. Soc. 115 (1993) 6382.[4] P. Wang, S.M. Zakeeruddin, J.E. Moser, M.K. Nazeeruddin, T. Sekiguchi, M.

Grätzel, Nat. Mater. 2 (2003) 402.

mica A

[

[[

[[

[

[[[[

[

[

[

[

[

[

[

[[[[

[

[[

[

[

[

Y. Zhang et al. / Electrochi

[5] D.P. Hagberg, J.-H. Yum, H. Lee, F.D. Angelis, T. Marinado, K.M. Karlsson, R.Humphry-Baker, L.C. Sun, A. Hagfeldt, M. Grätzel, Md.K. Nazeeruddin, J. Am.Chem. Soc. 130 (2008) 6259.

[6] Y. Bai, Y.M. Cao, J. Zhang, M.K. Wang, R.Z. Li, P. Wang, S.M. Zakeeruddin, M.Grätzel, Nat. Mater. 7 (2008) 626.

[7] Z.P. Tian, H.M. Tian, X.Y. Wang, S.K. Yuan, J.Y. Zhang, X.B. Zhang, T. Yu, Z.G. Zou,Appl. Phys. Lett. 94 (2009) 031905.

[8] F. Sauvage, D.H. Chen, P. Comte, F.Z. Huang, L.-P. Heiniger, Y.B. Cheng, R.A.Caruso, M. Grätzel, ACS Nano 4 (2010) 4420.

[9] Y.C. Qiu, W. Chen, S.H. Yang, Angew. Chem. Int. Ed. 49 (2010) 3675.10] M.K. Wang, S.-J. Moon, D.F. Zhou, F.L. Formal, N.-L. Cevey-Ha, R. Humphry-

Baker, C. Grätzel, P. Wang, S.M. Zakeeruddin, M. Grätzel, Adv. Funct. Mater. 20(2010) 1821.

11] S. Yanagida, Y.H. Yu, K. Manseki, Acc. Chem. Res. 42 (2009) 1827.12] M.K. Wang, A.M. Anghel, B. Marsan, N.-L.C. Ha, N. Pootrakulchote, S.M.

Zakeeruddin, M. Grätzel, J. Am. Chem. Soc. 131 (2009) 15976.13] K.H. Ko, Y.C. Lee, Y.J. Jung, J. Colloid Interface Sci. 283 (2005) 482.14] H. Imahori, S. Hayashi, T. Umeyama, S. Eu, A. Oguro, S. Kang, Y. Matano, T.

Shishido, S. Ngamsinlapasathian, S. Yoshikawa, Langmuir 22 (2006) 11405.15] X.J. Lü, X.L. Mou, J.J. Wu, D.W. Zhang, L.L. Zhang, F.Q. Huang, F.F. Xu, S.M. Huang,

Adv. Funct. Mater. 20 (2010) 1.16] J. Liu, H.T. Yang, W.W. Tan, X.W. Zhou, Y. Lin, Electrochim. Acta 56 (2010) 396.17] K.-P. Wang, H.S. Teng, Phys. Chem. Chem. Phys. 11 (2009) 9489.18] C. Kim, K.-S. Kim, H.Y. Kim, Y.S. Han, J. Mater. Chem. 18 (2008) 5809.19] Q.F. Zhang, T.P. Chou, B. Russo, S.A. Jenekhe, G.Z. Cao, Adv. Funct. Mater. 18

(2008) 1654.20] F.Z. Huang, D.H. Chen, X.L. Zhang, R.A. Caruso, Y.-B. Cheng, Adv. Funct. Mater.

20 (2010) 1301.21] Y.J. Kim, M.H. Lee, H.J. Kim, G. Lim, Y.S. Choi, N.-G. Park, K. Kim, W.I. Lee, Adv.

Mater. 21 (2009) 3668.

[

[

cta 56 (2011) 6517– 6523 6523

22] Q.D. Zhao, T.F. Xie, L.L. Peng, Y.H. Lin, P. Wang, L. Peng, D.J. Wang, J. Phys. Chem.C 111 (2007) 17136.

23] N. Papageorgiou, W.F. Maier, M. Grätzel, J. Electrochem. Soc. 144 (1997)876.

24] C. Klein, Md.K. Nazeeruddin, D.D. Censo, P. Liska, M. Grätzel, Inorg. Chem. 43(2004) 4216.

25] J. Zhang, D.J. Wang, Y.M. Chen, T.J. Li, H.F. Mao, H.J. Tian, Q.F. Zhou, H.J. Xu, ThinSolid Films 300 (1997) 208.

26] Q.L. Zhang, D.J. Wang, X. Wei, T.F. Xie, Z.H. Li, Y.H. Lin, M. Yang, Thin Solid Films491 (2005) 242.

27] H.Y. Li, D.J. Wang, P. Wang, H.M. Fan, T.F. Xie, Chem. Eur. J. 15 (2009) 12521.28] Y. Masuda, T. Ohji, K. Kato, Cryst. Growth Des. 10 (2010) 913.29] H.-X. Wu, T.-J. Wang, Y. Jin, Ind. Eng. Chem. Res. 45 (2006) 1337.30] Y. Xie, X.J. Zhao, H.Z. Tao, Q.N. Zhao, B.S. Liu, Q.H. Yuan, Chem. Phys. Lett. 457

(2008) 148.31] L.Q. Jing, Z.L. Xu, J. Shang, X.J. Sun, W.M. Cai, H.C. Guo, Mater. Sci. Eng. A 332

(2002) 356.32] L. Kronik, Y. Shapira, Surf. Interface Anal. 31 (2001) 954.33] J.H. Yang, D.S. Warren, K.C. Gordon, A.J. McQuillan, J. Appl. Phys. 101 (2007)

023714.34] X. Wei, T.F. Xie, D. Xu, Q.D. Zhao, S. Pang, D.J. Wang, Nanotechnology 19 (2008)

275707.35] B. Mahrov, G. Boschloo, A. Hagfeldt, L. Dloczik, Th. Dittrich, Appl. Phys. Lett. 84

(2004) 5455.36] V. Duzhko, V.Yu. Timoshenko, F. Koch, Th. Dittrich, Phys. Rev. B 64 (2001)

075204.37] D. Cahen, G. Hodes, M. Grätzel, J.F. Guillemoles, I. Riess, J. Phys. Chem. B 104

(2000) 2053.38] Q. Shen, J. Kobayashi, L.J. Diguna, T. Toyoda, J. Appl. Phys. 103 (2008)

084304.