A Study on the Electron Transport Properties of TiO2 Electrodes in Dye-Sensitized Solar Cells

(1)

Solar Energy Materials & Solar Cells 91 (2007) 1416–1420

A study on the electron transport properties of TiO

₂

electrodes in

dye-sensitized solar cells

Kun-Mu Lee

a

, Vembu Suryanarayanan

b

, Kuo-Chuan Ho

a,b,

a_{Institute of Polymer Science and Engineering, National Taiwan University, Taipei 10617, Taiwan} b_{Department of Chemical Engineering, National Taiwan University, Taipei 10617, Taiwan}

Available online 9 May 2007

Abstract

The inﬂuences of annealing temperature and different poly (ethylene glycol) (PEG) contents in nano-crystalline TiO2electrodes with

and without N3 dye on the electron transfer in a dye-sensitized solar cell (DSSC) were investigated. It is found that the power conversion efﬁciency increases with the increase in annealing temperature and becomes saturated at 400–500 1C, and further increase lowers the performance which is consistent with the enhancement of the crystalline TiO2particles observed in X-ray diffraction (XRD) patterns and

scanning electron microscopy (SEM) images. Electrochemical impedance spectroscopy (EIS) also confirms this behavior. These results have been further verified by studying the electron lifetimes (te) and electron diffusion coefficients (De) of a bare TiO2 and a

dye-sensitized TiO2ﬁlm using a pulsed laser spectrometer. It is noted that both the electron lifetime and the electron diffusion coefﬁcient

increase with the increase in annealing temperature. However, the evolution of rutile TiO2begins beyond 600 1C and this lowers the dye

absorbance and the electron diffusion coefﬁcients of TiO2electrodes. A similar study was made by varying the content of the PEG in the

TiO2ﬁlms. It is found that with the increase in the PEG content, a decrease in the electron lifetimes and a little hike in the electron

diffusion coefficients are noted, where the cell performance remains almost the same. In addition, the dye adsorption decreases the electron lifetime and increases the electron diffusion coefficient of the TiO2 films regardless of the PEG content and the annealing

temperature.

Keywords: Dye-sensitized solar cells; Nanocrystallined TiO2; Dye adsorption; Electron transfer

1. Introduction

Dye-sensitized solar cells (DSSCs) have been the subject of intense study on account of their high conversion efﬁciency and low cost[1]. These solar cells usually employ liquid electrolytes containing I

/I3redox couple as

support-ing electrolyte in order to reduce the dye cation, generated by the injection of the photo-excited electron. Though the electrons in the nano-TiO2ﬁlm are surrounded by cations, it

has been considered that no large electric ﬁeld gradient is found in the ﬁlm[2,3]. The transport of electrons in the TiO2

electrode has been mainly related with their diffusion

coefficient [4–6]. The electrons must travel a distance in the film to reach the transparent conductive oxide (TCO) layer before charge recombination. Several groups have developed trapping models, which assume that intraband charge trap sites exist in the films and electron travel through the events of trapping and de-trapping[7,8]. Studies on the lifetime of electrons in the TiO2 electrode were

performed in relation to employment of various dyes molecules [9] and electrolytes [10,11]. Further, the effects of TiO2 surface treatment [12], applied bias voltage [13],

TiO2crystalline type [14], nanoparticles size [15]and TiO2

nanoparticles preparation methods and annealing tempera-ture [16] on the performance of the DSSCs were also studied. In this work, we had investigated the inﬂuences of different annealing temperatures and weight percentages of poly (ethylene glycol) (PEG) on the performance of the DSSC as well as lifetime of electrons in the TiO2electrode.

www.elsevier.com/locate/solmat

Corresponding author. Institute of Polymer Science and Engineering, National Taiwan University, Taipei 10617, Taiwan.

Tel.: +886 2 2366 0739; fax: +886 2 2362 3040. E-mail address:[email protected] (K.-C. Ho).

(2)

2. Experimental

Anhydrous LiI, I2, PEG and 4-tertiary butyl pyridine

(TBP) were obtained from Merck and titanium (V) isopropoxide (+98%) and deoxycholic acid (DCA) were purchased from Acros and used as such. CH3CN and

tertiary butanol were purchased from Merck and water molecules were removed by putting molecular sieves (4 A˚) into the solvent. The N3 dye was the commercial product obtained from Solaronix S.A., Aubonne, Switzerland.

The preparation of TiO2 precursor and the electrode

fabrication were carried out based on previous literature

[17] except after autoclave treatment, where the solution was concentrated to 13 wt% and PEG (M.W. 20,000) was added to the TiO2paste to prevent the ﬁlm from cracking

during drying. The TiO2 paste was coated on a

ﬂuorine-doped tin oxide (FTO) glass plate (Rsh¼25 O/square,

Sinonar Corporation, Hsinchu, Taiwan) using glass rod method.

An active area of 0.25 cm2 was selected from sintered electrode and the electrodes were immersed in 3 104M solution of cis-di(thiocyanato)bis(2,20_{-bipyridyl-4,4}0

-dicar-boxylate)ruthenium (II) (N3 dye) containing acetonitrile and tertiary butanol (in the volume ratio of 171) for 24 h. Pt (100 nm thick) sputtered on FTO was used as the counter electrode and the electrolyte was composed of 0.5 M lithium iodide (LiI)/0.05 M iodine (I2)/0.5 M TBP in CH3CN.

Micrographs were obtained using a Hitachi S-4700 scanning electron microscope (SEM). X-ray diffraction (XRD) patterns were monitored using a Rigaku RAD system with CuKa radiation. The ﬁlm thickness was

measured using profilometer (Sloan Dektak 3030). The photoelectrochemical characterizations of the DSSCs were carried out by using an AM 1.5 simulated light radiation. The light source was a 450 W Xe lamp (Oriel, ]6266) equipped with a water-based IR filter and AM 1.5 filter (Oriel, ]81075).

The photovoltage transients of assembled devices were recorded with a digital oscilloscope (LeCroy, model LT322). Pulsed laser excitation was applied by a fre-quency-doubled Q-switched Nd:YAG laser (Spectra-Phy-sics laser, model Quanta-Ray GCR-3-10) with 2 Hz repetition rate at 355 and 532 nm, respectively, and 7 ns pulse width at half-height. The beam size was larger than 0.25 cm2to cover the area of the device with an incident energy of 1 mJ/cm2. The average lifetime of electron can be estimated approximately by ﬁtting a decay of the open circuit voltage transient with exp(t/te), where t is the time

and te is an average time constant before recombination.

Electron diffusion coefﬁcient was estimated by ﬁtting a decay of the current transient with exp(t/tc) that was

derived from the equation of continuity for electrons in the conduction band [18], where t and tc are the time and

average time constant, respectively. Then, the apparent diffusion coefﬁcient of electron can be estimated by De¼w2=2:35 tc, (1)

where w is the ﬁlm thickness and the factor 2.35 arises from the geometry of the diffusion problem.

The photoelectrochemical characteristics and the AC-impedance measurements of the DSSCs were recorded with a potentiostat/galvanostat (PGSTAT 30, Autolab, Eco-Chemie, the Netherlands) under constant light illumination of 100 mW/cm2. The applied bias voltage and AC amplitude were set at open-circuit voltage of the DSSCs at 10 mV between the FTO-Pt counter electrode and the FTO-TiO2-dye working electrode, respectively, starting

from the short-circuit condition [19]. The impedance spectra were analyzed by an equivalent circuit model interpreting the characteristics of the DSSCs[20].

3. Results and discussion

The effect of different annealing temperatures of the TiO2 electrode on the crystalline nature of the TiO2

particles was investigated using XRD and the results are shown inFig. 1. Only the samples those were autoclaved at a temperature higher than 300 1C exhibits sharp peaks corresponding to anatase phase. However, at 600 1C, the evolution of rutile-TiO2is observed. Furthermore, the

crystalline nature of the TiO2 particles increases with the

increase in annealing temperature. Surface morphology of the TiO2 ﬁlms obtained by SEM for different

annea-ling temperatures is shown in Fig. 2. It reveals a porous structure of the TiO2 particles with an average size of

about 20 nm for sintering from 300 to 500 1C (Fig. 2a–c). However, the SEM image obtained at 600 1C (Fig. 2d) shows the formation of large TiO2 particles and this

may be correlated with the phase transformation of few TiO2 nanoparticles from anatase to rutile [17]. The

results obtained from the XRD data further support this (Fig. 1).

Table 1 shows the performance of the DSSC with the

TiO2electrodes (only 5 mm) thickness annealed at different

temperatures. From this table, it is observed that the

20 30 40 50 60 70 30 °C Anatase Rutile 600 °C 500 °C 400 °C 300 °C Intensity (a.u.) 2 theta (degree)

(3)

photoelectrochemical characteristics such as open circuit voltage (VOC), short circuit current density (JSC), ﬁll factor

(FF) and the conversion efﬁciency (Z) of the DSSC containing un-sintered TiO2electrode is very low; however,

these performance characteristics increase with the increase in annealing temperature and remains stable at 400–500 1C. The cell efﬁciency decreases at the annealing temperature of 600 1C. This may be correlated with the coarsening of TiO2anatase nanoparticles, and that the anatase particle

size could start to increase at 600 1C, leading to a lower surface area. Alternatively, the evolution of rutile TiO2,

which possesses larger size, less surface area and slower electron diffusion rate, has been reported[14]. The low cell

performance at 600 1C can also be related with the decrease in the dye adsorption onto the TiO2 electrode. In

connection with this, we studied the effect of dye adsorption onto the TiO2electrodes at different annealing

temperatures. Our experimental results (not shown here) revealed that the amount of dye adsorption was found to be maximal up to 500 1C and it decreased with further increase in annealing temperature.

Fig. 3a shows electrochemical impedance spectroscopic (EIS) analysis of the DSSC associated with different annealing temperatures of the TiO2 electrodes and the

equivalent circuit is shown inFig. 3b. In general, the EIS spectrum of the DSSC containing liquid electrolyte shows three semicircles in the measured frequency range of 10 mHz–65 kHz. The ohmic serial resistance (Rs)

corre-sponds to the electrolyte and the FTO resistance and the resistances Rct1, Rct2 and Rdiff correspond to the charge

transfer process occurring at the Pt counterelectrode, the TiO2/dye/electrolyte interface and the Warburg diffusion

process of I/I3

in the electrolyte, respectively. High interfacial charge transfer resistance (Rct2) resulting from

large semi-circle is observed for the DSSC containing un-sintered TiO2 electrode and the Rct2 decreases with

increase in annealing temperature from 300 to 500 1C (Fig. 3a). On the other hand, an increase in the Rct2is noted

at 600 1C (Fig. 3a). Presumably, this is due to the partial

Fig. 2. SEM images of the TiO2ﬁlm annealed at different temperatures.

Table 1

Cell performances of the DSSCs with TiO2 ﬁlms annealed at different

temperatures Samplea no. Annealing temperature ( 1C) JSC(mA/cm2) VOC(V) Z (%) FF 1 — 1.76 0.645 0.57 0.498 2 300 9.80 0.781 5.01 0.655 3 400 11.76 0.783 5.72 0.622 4 500 11.20 0.763 5.35 0.626 5 600 7.72 0.824 4.23 0.666

a_{The thickness of TiO}

(4)

transformation of anatase TiO2 to form rutile TiO2 and

results in a larger particle size and thus a slower electron diffusion rate.

Laser-pulse induced photovoltage transients were re-corded for bare and dyed TiO2 electrode at different

annealing temperatures (from 30 to 600 1C). All experi-ments were repeated for three times at each annealing temperature. The error bar represents the range of variation of the three data collected. The resultant electron lifetime (te) and electron diffusion coefﬁcients (De) were plotted as a

function of increase in the annealing temperature (Fig. 4a and b). From Fig. 4a, it has been observed that the te of

bare TiO2electrode is higher than that of dyed TiO2and the

tein both cases increase from 300 to 600 1C. On the other

hand, Deof dyed TiO2electrode is higher than that of bare

TiO2electrode and the Deincreases from 30 to 400 1C and

then decreases at a high temperature of 600 1C (Fig. 4b). This is once again correlated with the formation of the rutile TiO2 particles that has low electron diffusion

coefﬁcient. Further, this result also implies that the crystal structures of the particles in both bare and dyed TiO2

electrodes inﬂuence the Deobviously.

We had also measured the value of Deof the TiO2ﬁlms

adsorbed with DCA, instead of N3 dye, and found that the value of De was higher for the former (2.31 104cm2/s)

when compared to the latter (0.47 104cm2/s). This indicates that adsorption of carboxyl group increases the value of the Dein the TiO2electrode.

It is well known that the addition of PEG to the TiO2

colloidal solution, after autoclaving, inﬂuences the porosity

Fig. 3. (a) EIS spectra of the DSSCs with TiO2ﬁlm annealed at different

temperatures; (b) equivalent circuit for (a).

Fig. 5. Plots of (a) the electron lifetimes and (b) electron diffusion coefﬁcients of bare and dyed TiO2ﬁlms containing different contents of PEG.

(5)

of the TiO2electrode[21]. Herein, we have investigated the

inﬂuence of lifetime of electrons in the TiO2 electrode as

well as the performance of the DSSC by varying the amount of PEG content in the TiO2 electrode (the

thickness of the TiO2 electrode is kept at 5 mm and

the annealing temperature is maintained at 500 1C). Once again, all experiments were repeated for three times at each PEG content. The error bar represents the range of variation of the three data collected.Fig. 5a and bbrings the correlation between the electron lifetime and its diffusion coefﬁcient in the TiO2electrode with and without

dye for different PEG contents (30%, 60% and 90%) and

Table 2shows the performances of the DSSCs containing

the TiO2 electrodes with the above PEG contents. The

electron lifetime (De) decreases and the value of the De

increases with the increase in PEG content in both cases (Fig. 5); the conversion efﬁciency of the DSSC changes slightly (5–6%) (Table 2) and this shows that there is not much change in the electron diffusion lengths in hiking the PEG content.

4. Conclusions

From the above results, it is concluded that the conversion efﬁciency of the DSSC increases with the increase in annealing temperature of the TiO2and remains

stable at 400–500 1C, however, further increase lowers the cell performance and this is correlated with the evolution of rutile phase of TiO2 particles, as conﬁrmed from SEM

images, XRD patterns of the TiO2 electrode and EIS

studies of the DSSC. Further, our investigations on the electron lifetime and the electron diffusion coefficient also confirm these results. Moreover, it is also observed that with the increase in the PEG content of the TiO2films, the

electron lifetime decreases, while the electron diffusion coefﬁcient increases and changing the PEG content results in little variation on the performance of the cell.

Acknowledgments

This work was ﬁnancially supported by the Academia Sinica, Taipei, Taiwan, the Republic of China, under Grant AS-94-TP-A02. Helpful discussions with Professor Tien-Yau Luh, of Department of Chemistry, National Taiwan University, are appreciated. This work was partially supported by the Photovoltaics Technology Center, Industrial Technology Research Institute (ITRI), Chutung, Hsinchu, Taiwan. We also want to thank Professor King-Chuen Lin and his research group mem-bers, of Department of Chemistry, National Taiwan University, for the help in making the pulsed laser apparatus available to us.

References

[1] B.O. Reagen, M. Gra¨tzel, Nature 353 (1991) 373.

[2] K. Schwarzburg, F. Willing, J. Phys. Chem. B103 (1999) 5743. [3] J. van de Lagemaat, N.-G. Park, A.J. Frank, J. Phys. Chem. B104

(2000) 2044.

[4] S. So¨dergren, A. Hagfeldt, J. Olsson, S.-E. Lindquist, J. Phys Chem. 98 (1994) 5552.

[5] F. Cao, G. Oskam, G.J. Meyer, P.C. Searson, J. Phys. Chem. 100 (1996) 17021.

[6] A. Solbrand, H. Lindstrom, H. Resmo, A. Hagfeldt, S.-E. Lindquist, J. Phys. Chem. B101 (1997) 2514.

[7] A.C. Fisher, L.M. Peter, E.A. Ponomarev, A.B. Walker, K.G.U. Wijayantha, J. Phys. Chem. B104 (2000) 949.

[8] G. Nelson, Phys. Rev. B59 (1999) 15374.

[9] Y. Tachibana, S.A. Haque, I.P. Mercer, J.R. Durrant, D.R. Klug, J. Phys. Chem. B104 (2000) 1198.

[10] S. Pelet, J.-E. Moser, M. Gra¨tzel, J. Phys. Chem. B104 (2000) 1791.

[11] J. Nelson, S.A. Haque, D.R. Klug, J.R. Durrant, Phys. Rev. B63 (2001) 205321.

[12] G. Schlichtho¨rl, S.Y. Huang, J. Sprague, A.J. Frank, J. Phys. Chem. B101 (1997) 8141.

[13] S.A. Haque, Y. Tachibana, R.L. Willis, J.-E. Moser, M. Gra¨tzel, D.R. Klug, J.R. Durrant, J. Phys. Chem. B104 (2000) 538. [14] N.-G. Park, J. van deLagemaat, A.J. Frank, J. Phys. Chem. B104

(2000) 8989.

[15] S. Nakade, Y. Saito, W. Kubo, T. Kitamura, Y. Wada, S. Yanagida, J. Phys. Chem. B107 (2003) 8607.

[16] S. Nakade, M. Matsuda, S. Kambe, Y. Saito, T. Kitamura, T. Sakata, Y. Wada, H. Mori, S. Yanagida, J. Phys. Chem. B106 (2002) 10004.

[17] C.J. Barbe, F. Arendse, P. Comte, M. Jirousek, F. Lenzmann, V. Shklover, M. Gra¨tzel, J. Am. Ceram. Soc. 80 (1997) 3157. [18] J. van de Lagemaat, A.J. Frank, J. Phys. Chem. B105 (2001)

11194.

[19] C. Longo, J. Freitas, M.A. De Paoli, J. Photochem. Photobiol. 159 (2003) 33.

[20] M.C. Bernard, H. Cachet, P. Falaras, A. Hugot-Le Goff, M. Kalbac, I. Lukes, N.T. Oanh, T. Stergiopoulos, I. Arabatzis, J. Electrochem. Soc. 150 (2003) E155.

[21] K.-M. Lee, V. Suryanarayanan, K.-C. Ho, Sol. Energy. Mater. Sol. Cells 90 (2006) 2398.

Table 2

Cell performances of the DSSCs with TiO2 ﬁlms annealed at 500 1C

containing different PEG contents Samplea no. PEG contents (%) JSC(mA/cm 2 ) VOC(V) Z (%) FF 4–1 30 11.00 0.763 5.25 0.626 4–2 60 11.68 0.775 5.63 0.622 4–3 90 11.60 0.781 5.81 0.642

a_{The thickness of TiO}