on nanocrystalline TiO2 surfaces

ARTICLE IN PRESS Solar Energy Materials & Solar Cells 94 (2010) 857–864

Contents lists available at ScienceDirect

Solar Energy Materials & Solar Cells journal homepage: www.elsevier.com/locate/solmat

Temperature-dependent infrared spectrum of (Bu4N)2[Ru(dcbpyH)2-(NCS)2] on nanocrystalline TiO2 surfaces Yeonju Park a, Young Mee Jung a,n, Subrata Sarker b, Jae-Joon Lee b,nn, Yunhee Lee c, Kangtaek Lee c, Jung Jin Oh d, Sang-Woo Joo e,nnn a

Department of Chemistry, Kangwon National University, Chunchon 200-701, Republic of Korea Department of Applied Chemistry/Department of Applied Technology Fusion, Konkuk University, Chungju 380-701, Republic of Korea c Department of Chemical Engineering, Yonsei University, Seoul 120-749, Republic of Korea d Department of Chemistry, Sookmyung Women’s University, Seoul 140-742, Republic of Korea e Department of Chemistry, Soongsil University, Seoul 156-743, Republic of Korea b

a r t i c l e in fo

abstract

Article history: Received 6 March 2009 Received in revised form 31 December 2009 Accepted 9 January 2010 Available online 4 February 2010

The thermal degradation behavior of the self-assembled thin films of (Bu4N)2[Ru(dcbpyH)2-(NCS)2] (N719) anchoring on TiO2 surfaces via its carboxylate group has been studied using temperaturedependent diffuse reflectance infrared Fourier transform (DRIFT) spectroscopy. Our analysis indicate that the decomposition temperature of N719 appeared to be at E270 1C on TiO2 surfaces, whereas such decompositions occurred at temperatures higher than E 340 1C in their solid states as line with the thermal analysis data. This change was also found to be irreversible, if heated above 100 1C on TiO2 surfaces. Our 2D correlation spectroscopy and principal component analysis (PCA) applied to the temperature-dependent DRIFT spectra supported that the thermal degradation mechanism for N719 should differ in its solid state and on TiO2 powder surfaces. The NCS stretching vibrational intensities of the neat N719 vibrations were found to shift from 2102 to 1975 cm 1 increase with increase in temperatures, whereas similar vibration changes from 2095 to 2008 cm 1 were observed for N719 attached to TiO2. Referring from the open-circuit potential, short-circuit, fill factor, and efficiency measurements for the N719 dye-loaded photoelectrodes depending on temperature, the NCS stretching band at 2008 cm 1 on TiO2 surfaces appeared to be correlated with the thermal degradation of the DSSCs. & 2010 Elsevier B.V. All rights reserved.

Keywords: Thermal decomposition (Bu4N)2[Ru(dcbpyH)2-(NCS)2] (N719) TiO2 Diffuse reflectance Fourier transform infrared spectroscopy Photovoltaic measurements

1. Introduction Over the past decade, dye-sensitized solar cells (DSSCs) based on the mesoporous TiO2 nanoparticle networks have been extensively investigated as a solar-to-electricity conversion system [1,2]. Ruthenium (II) complexes containing polypyridyl ligands have been widely studied as potential photosensitizers in DSSCs [3]. Metal-to-ligand charge-transfer reactions raise the visible absorption spectra of the dyes and alter their photophysical and redox behavior [4,5]. It is well known that the effect of the anchoring efficiency of conventional dyes to the semiconductor surface is critical in achieving high energy conversion efficiency in DSSCs [5].

n

Corresponding author. Tel.: + 82 33 250 8495; fax: 82 33 253 7582. Corresponding author. Tel.: + 82 43 840 3580; fax: +82 43 851 4169. Corresponding author. Tel./fax: + 82 2 820 0434. E-mail addresses: [email protected] (Y. M. Jung), [email protected] (J.-J. Lee), [email protected] (S.-W. Joo). nn

nnn

0927-0248/$ - see front matter & 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.solmat.2010.01.008

For the most popular dyes such as N719, the carboxylate groups from different bipyridyl ligands must be used for effective anchoring through the bidentate bridging mode [6–8]. The isothiocyanato NCS ligand with low electronegativity is assumed to facilitate the dye regeneration by the redox couple in the electrolyte phase [7]. The stability of the dye at moderately high temperature is one of the main issues to achieve outdoor usages of DSSCs. It has been reported that the N719 dye as a sensitizer exhibited an excellent stability under long-term light soaking at temperatures below 45 1C [8], whereas the performance of the DSSCs degraded severely, if stored at 85 1C in dark conditions [9]. Several infrared spectroscopic reports for dyes and additives on TiO2 surfaces potentially utilized in solar cells have been reported [10–14], and it has been discussed in detail about the role of the molecular interaction between dyes and semiconductor at the surface. However, to the best of our knowledge, there has been no detailed analysis of the mechanism of the thermal degradation of dyes and the corresponding changes in the molecular structure of ruthenium-based dyes on TiO2 surfaces.

ARTICLE IN PRESS 858

Y. Park et al. / Solar Energy Materials & Solar Cells 94 (2010) 857–864

Among numerous techniques to study metal (or semiconductor)/ adsorbate interface, vibrational spectroscopy is one of the most useful tools that can provide structural information about chemisorption on the surface of interest [15,16]. Analysis of the temperature dependent spectral changes can provide information on the energetics or phase transition of adsorbates on surfaces of interests [17–19]. In this work, we present the study of thermal degradation of N719 on TiO2 surfaces with a vibrational spectroscopic tool. We used temperature-dependent diffuse reflectance Fourier transform infrared (DRIFT) spectroscopy [20,21] to better understand the structural change of the adsorbed dye on TiO2 surfaces at elevated temperatures. The photovoltaic characterization of the DSSCs as a function of temperature was also performed in parallel to deduce more critical factors about the thermal degradation process of N719 on TiO2 surfaces and therefore the effect on the overall energy conversion efficiency of DSSC at high temperature. This work should be helpful to understand the thermal stability of ruthenium-based dye on TiO2 surfaces and the effect of temperature on the operation of DSSCs.

2. Materials and methods 2.1. Sample preparations The dye cis-diisothiocyanato-bis(2,20 -bipyridyl-4,40 -dicarboxylato) ruthenium (II) bis (tetrabutyl-ammonium), (Bu4N)2 [Ru(dcbpyH)2-(NCS)2] (Bu4N=tetrabutylammonium and dcbpy = dicarboxylbipyridine) N719 (98%) was purchased from Solaronix S.A. (Lausanne, Switzerland) and used without further purification. For the self-assembly of N719 on TiO2 powders (P-25 Degussa, Germany), approximately 0.050 g of TiO2 powder was placed in a vial. Into this was placed a 10 mM stock solution of N719 and self-assembled for 424 h to make sure of high level dye loading in short time period. Although the concentration of N719 as high as 10 mM was used to make sure of efficient selfassembly due to a small sampling volume of TiO2, we have not observed any side effects due to high concentrations under our experimental conditions. We used ethanol as the solvent for the infrared spectroscopy. The sizes of N719 and TiO2 are expected by 1 and 25 nm, respectively, from the previous study [22]. After this, the solution phase was decanted, centrifuged at 1200 rpm, the remaining solid particles were washed 3–4 times with a large quantity of ethanol and left to dry at a desiccator in ambient conditions.

2.3. Fabrication of DSSCs: A viscous paste of nanocrystalline TiO2 was prepared by mixing a pretreated commercial TiO2 (P-25 Degussa, Germany) powder with 50 wt% of hydroxypropyl cellulose (Aldrich) [24]. 6.0 g of TiO2 powder was mixed with 12 mL of distilled water and 0.2 mL of acetyl acetone with constant stirring for 24 h followed by drying at 60 1C and then grinding to make fine powder. 1.0 g of the pretreated TiO2 powder was dispersed in a mixture of 2 mL of distilled water and 0.05 mL of acetyl acetone with constant stirring for 12 h. Then 5 mL of 10% ethanolic solution of hydroxypropyl cellulose was added into the dispersion with an additional amount of ethanol to double the volume. The mixture was stirred vigorously with heating at 50 1C to evaporate the ethanol to get a viscous TiO2 paste in which cellulose was 50% w/ w of TiO2. The paste was doctorbladed on FTO glass substrate (15 O/sq, TEC15, Philkington) and sintered at 500 1C for 30 min in air. These nanocrystalline TiO2 films were then immersed in ethanolic solution of 0.3 mM of cis-diisothiocyanato-bis(2,20 bipyridyl-4,40 -dicarboxylato) ruthenium (II) bis (tetrabutyl-ammonium) dye, namely N719, (ruthenium 535 bis-TBA, Solaronix) overnight to load the sensitizer. The counter electrode was prepared by spin coating 0.6 mM chloroplatinic acid hexahydrate (H2PtCl6 6H2O) followed by sintering at 380 1C for 20 min. Before assembling the DSSCs, the dye-loaded TiO2 photoelectrodes were heated up at 25 (without heat treatment), 140, 200, 240, 280, and 380 1C for 10 min. The heat treated photoelectrode and the platinized counter electrode were sandwiched with 50 mm thick surlyn film as a spacer and sealing agent. A liquid electrolyte (0.6 M DMPII (1, 2-dimethyl-3-propylimidazolium iodide), 0.1 M Li, 0.1 M I2, and 0.5 M TBP (4-tert-butylpyridine) in 3-methoxypropionitrile) was directly injected into the cell through the drilled holes at the counter electrode. 2.4. Measurements and characterizations Photovoltaic performance of the DSSCs was measured by sweeping the potential at a scan rate of 50 mV/s with an electrochemical workstation, Potentiostat/Galvanostat M273A (EG&G Instruments Princeton Applied Research, USA), under the simulated irradiation condition of AM 1.5 (100 mW cm 2 ) solar condition, which was generated by a 300 W xenon lamp (Oriel, 66902) equipped with three air mass filters (Oriel, M6123) used as a light source. The incident light intensity was calibrated by a standard mono-silicon solar cell (PVM108, PV measurement Inc.), which was certified by National Renewable Energy Laboratory (NREL), USA. The effective area of the cell was 0.2 cm2. The thickness of the TiO2 film was measured to be 6 mm by observing the cross section of the photoelectrode by FE-SEM (JSM-6799F, JEOL).

2.2. Instrumental measurements A portion of the N719-assembled powdered sample was transferred onto a diffuse IR heated chamber (Pike Technologies) equipped with a temperature controller to control temperatures up to 500 1C. The infrared spectra were obtained using an FT-IR spectrometer with a maximum resolution of 0.09 cm 1 (Thermo Nicolet 6700). A total of 32 scans were measured in the range 1000–3500 cm 1 with a resolution of 4 cm 1. At each temperature, the sample remained heated for 3 min. Data processing was carried out using the OMNIC v7.3 software. We performed NMR and mass spectrometry experiments of the N719 dye using a Bruker DPX 400 MHz FT-NMR spectrometer and an Applied Biosystems Voyager-DE STR, respectively. Their spectra appeared to be analogous to the previous report [23] Themal analysis is performed by TA instruments DSC Q2000/SDT Q600.

3. Results and discussion 3.1. Infrared spectra of N719 Fig. 1(a) shows an external reflection absorption spectrum of N719 in its solid state. The vibrational bands at 1717 cm 1 can be ascribed to the C =O stretching. Due to the antisymmetric and symmetric n (COO ) modes, the two bands were found at 1545 and 1375 cm 1, respectively. These results clearly indicate that N719 should adsorb on TiO2 powder surfaces in its carboxylate form. The n(NCS) bands were strongly observed at 2102 and 2095 cm 1 for the solid state and TiO2 surfaces. On TiO2 surfaces, the vibrational bands at 1614, 1472, 1436, and 1233 cm 1 could be ascribed to the bipyridine’s ring modes. Their peak positions

ARTICLE IN PRESS Y. Park et al. / Solar Energy Materials & Solar Cells 94 (2010) 857–864

2008

1764

(b)

1000

2102

1717

1298 1368 1405 1436 1466 1543

1617

1230 1247 1143

1021

0.2

0.0

(a)

1500 Wavenumber (cm-1)

2000

Fig. 1. Infrared spectra of N719 in (a) its solid state and (b) after the assembly on TiO2 powder surfaces.

Table 1 Spectral data and vibrational assignments for N719.a Assignmentb

N719 OR (solid)

DRIFT on TiO2

2102 1717 1617 1543 1466 1436 1405 1368 1298 1247 1230 1143 1021 a b

n(NCS) n(NCS) n(C = O) n(C = C) (bpy) nas(COO ) n(C = C) (bpy) n(C = N) (bpy) n(C = N) (bpy) n(C = N) (bpy) ns(COO )

2095 2008 1764 1614 1545 1545 1472 1436 1436 1375

d(C–H)

n(C = N) (bpy) n(C–N) (benzene ring) n(C= C) intern-ring (bpy) nskeletal (benzene ring) n(C = C) ring breathing

1288 1233

Unit in cm 1. Based on Ref. [14].

NCS

SCN

N

Ru

N

(Bu4N) OOC

N

O

COO (Bu4N)

N

C

C O

O

powder surfaces is depicted in Fig. 2 as consistent with the previous study [7].

3.2. Temperature-dependent infrared spectra of N719 on TiO2 powder surfaces

2095

1614

1375 1436 1472 1545

1233 1288

Absorbance (Arbitrary Unit)

are listed in Table 1, along with the appropriate vibrational assignments. Our assignment is mainly based on previous literature [14]. A plausible binding structure of N719 on TiO2

859

O

TiO2 Fig. 2. Plausible binding structure of (Bu4N)2[Ru(dcbpyH)2-(NCS)2] (N719) on TiO2 powder surfaces.

Fig. 3(a) shows the temperature-dependent infrared spectra of N719 in its neat state, in a temperature range between 25 and 380 1C. It is noteworthy that the n(NCS) vibrational bands exhibited significant spectral changes. For N719, the NCS band at 2102 cm 1 weakened as the temperature increased and the band at 1975 cm 1 became more prominent and was comparable in intensity to that at 2102 cm 1. In the previous femtosecond transient infrared experiment [25], the three bands were found at 1900, 1950, and 2000 cm 1 for the ruthenium dye system with the NCS bonds. Referring to this work, the two bands at 1975 and 1875 cm 1 could be ascribed to the NCS vibration of the N719-TiO2 system. The lines at 1975 and 1875 cm 1 should be due to the NCS stretching lines, since either aromatic rings or carboxylate groups do not appear in this wavenumber region. Considering the high temperatures above 260 and 280 1C for the solid and the surface adsorbate, respectively, the bands became strongly intensified, the N719 dye is supposed to be charred and scorched, resulting in such a redshift of the NCS stretching band. The temperature difference of 100 cm 1 at 1975 and 1875 cm 1 may also indicate that these bands should be related to the NCS stretching bands at 2102 and 2008 cm 1 with a difference of 94 cm 1 for the solid state and the surface-adsorbed state, respectively, at room temperature. Fig. 3(b) shows the temperature-dependent infrared spectra of N719 on TiO2 powder surfaces. It was found that most bipyridine ring bands, including the ns(COO ) peak at 1375 cm 1, exhibited less spectral changes in the temperature range between 25 and 380 1C. For N719 on TiO2, the NCS band at 2095 cm 1 weakened as the temperature increased, and the band at 2008 cm 1 grew more prominent. At 120–140 1C, the NCS band at 2008 cm 1 was observed to be comparable to or stronger than that at 2095 cm 1. The NCS band appeared to be changed quite sensitively up to 140 1C, as consistent with the previous report [10]. At 240 1C, the NCS band at 2095 cm 1 almost disappeared and quite a weak feature was observed at 2125 cm 1. At temperatures above 280 1C, a new feature was subsequently observed at 1875 cm 1 but it almost disappeared at 360 1C or higher. All these results indicate that the decomposition of the NCS band in N719 should occur at much lower temperature on TiO2 powder surfaces than in its solid state. The red shift vibrational band from 2102 to 1975 cm 1 may indicate that the bond length of the NCS stretching mode should increase resulting in a reduction of its vibrational frequency. This red shift was found to be less from 2095 to 2008 cm 1 for N719 on TiO2, suggesting that the structural change for the NCS mode was less significant, presumably due to the interaction between the NCS stretching vibration and TiO2 surface. Also it is possible that the electron transfer from N719 in the antibonding state to TiO2 may increase the vibrational frequency. It seems conclusive that the NCS mode in N719 should interact with TiO2 substantially, as manifested in the temperaturedependent IR spectra. Also it is expected that the chemical links between bipyridine rings and Ru or between NCS and Ru is weakening as the temperature increases. Since Ru is highly electron withdrawing in the ground state of N719, the weakening of the chemical link with Ru induced the higher electron density in NCS and COO (through bipyridine ring), which results in the red-shift of the vibration frequencies. It is



2102

1975

1717

0.5

380oC 360 340 320

0.4

300

0.3

280 260 240 220 200

1143


0.6

1617

1543

1466

0.7

0.2 140 0.1

120 80 25

0.0 1000

1500

2000

Wavenumber (cm-1) 380oC 360 340 320 300 280 2095 2125

1875

1764

1614

1545

1375

1436

0.3

1360


0.4

0.2

260 240 220 200 140

0.1

2008

120

0.0 1000

1500

80 25

2000

Wavenumber (cm-1) Fig. 3. (a) Infrared spectra of N719 in its neat state and DRIFT spectra of N719 on TiO2 powders, taken by increasing temperatures, at 25–380 1C.

consistent with the previous observation of red-shift of both n(NCS) and ns(COO ) bands [26]. The ns(COO ) peaks at 1375 cm 1 for room temperature exhibited a red-shift to 1360 cm 1 as the temperature increased. The ns(COO ) peaks were still observable at a temperature as high as 380 1C. Nevertheless, the ns(COO ) peaks at 1375 cm 1 (at 25 1C) were still observable up to the temperature as high as 380 1C although it exhibited a red-shift to 1360 cm 1 as the temperature increased. It is likely that the adsorption of N719 on TiO2 through carboxylate linkage is quite strong and it survived at high temperature along with the bipyridine (bpy) ring structure remaining. To achieve a deeper insight into the temperature-dependent infrared spectra of N719 on TiO2 powder surfaces, we performed

2D correlation spectroscopy. 2D correlation spectroscopy is a well-established analytical technique to interpret the spectral data obtained under some external perturbation (e.g., temperature in this study) that provides considerable utility and benefit in many fields of spectroscopic studies [27–29]. The details of 2D correlation spectroscopy were described previously. Fig. 4(a) shows synchronous 2D correlation spectrum derived from the temperature-dependent DRIFT spectra of N719. A power spectrum extracted along the diagonal line in the synchronous 2D correlation spectrum is also shown at the top of Fig. 4(a). In the synchronous 2D correlation spectrum, we mainly observed that intensities of bands at 2102, 1717, 1617, 1543, 1466, and 1368 cm 1 decreased as the temperature increased, while those at 1975, 1436, 1299, 1276, 1143, and


2000

2000

1800 1600 1400 1200

2095

2008 2200

1878

1378 1438 1473 1550 1618 1657

1976

2200

Wavenumber(cm-1)

Wavenumber (cm-1)

1100

1276 1350 1434 1471 1545 1621 1675 1722

2101


1800 1600 1400 1200

1000 1000 1200 1400 1600 1800 2000 2200 Wavenumber (cm-1)

1000 1200 1400 1600 1800 2000 2200 Wavenumber(cm-1)

Fig. 4. Synchronous 2D correlation spectra obtained from the temperature-dependent DRIFT spectra of (a) N719 and (b) N719 on TiO2 powders. Solid and dashed lines represent positive and negative cross-peaks, respectively.

0.15

0.15

group II

280

0.10 0.05

PC2

PC2

0.05

240

300

220 320

0.00

200 180 25 4060 140 160 100 120 80

340 380

-0.05 400

-0.10 -0.3

0.10

260

360

-0.2

group III -0.1

group I 0.0

0.1

380 340 320 280 400 360 300

0.2

200 180

0.00

160

-0.05

140 120 100 80 60 25 40

-0.10 -0.15

PC1

260 240 220

group II

-0.2

-0.1

0.0 0.1 PC1

group I

0.2

0.3

Fig. 5. Scores on first two PCs for the temperature-dependent DRIFT spectra of (a) N719 in its solid state and (b) N719 on TiO2.

1100 cm 1 increased. The band intensity at 2102 cm 1 assigned to the NCS stretching mode changed significantly as the temperature varied. Fig. 4(b) shows synchronous 2D correlation spectrum obtained from the temperature-dependent infrared spectra of N719 on TiO2 powder surfaces. A power spectrum extracted along the diagonal line in the synchronous 2D correlation spectrum is also shown at the top of Fig. 4(b). 2D correlation spectra for N719 on TiO2 powder surfaces are found to be completely different from those for N719 (Fig. 4 (a)). Interestingly the band intensity at 1472 cm 1 for N719 on TiO2 powder surfaces increased as the temperature increased, while that at 1436 cm 1 decreased. However, the change was observed oppositely for N719 in its solid state. Although not absolutely certain, this may be related to the fact that the band at 1436 cm 1 locates near the broad carboxylate ns(COO ) band at 1375 cm 1 on TiO2 surfaces,

whereas the carboxylate band should be absent for the solid state. This 2D correlation analysis result indicates that the surface decomposition mechanism on TiO2 powder surfaces should not be the same as that of its solid state of N719. To better understand such discrepancies in a greater detail, principal component analysis (PCA) [30–33] was applied to the temperature-dependent DRIFT spectra of N719 in its solid state and on TiO2 powder surfaces. Fig. 5(a) and (b) shows the PCA score plot based on the temperature-dependent DRIFT spectra of N719 in its solid state and on TiO2 powder surfaces, respectively. From the PCA results shown in Fig. 6(a), the N719 spectra in its solid state can be separated into three groups for lower temperature (25–200 1C), intermediate temperature (220– 320 1C) and higher temperature (340–400 1C) regions. The spectra of N719 on TiO2 powder surfaces can be separated into two groups by considering the PC1 and PC2 scores as shown in



0

50

100

150

200

250

300

350

400

0

50

100

150 200 250 Temperature (oC )

300

350 400

η (%)

4.0 2.0 0.0

FF (%)

60 45 30

Jsc (mA/cm2)

10.0

5.0

0.0

Voc (V)

0.8 0.6 0.4 0.2 0.0

Fig. 6. Open circuit potential, VOC; short-circuit current, JSC; fill factor, FF and efficiency, Z of the DSSCs with heat treated N719 dye-loaded TiO2 photoelectrodes. The dye-loaded TiO2 electrodes were heated up at different temperatures for 10 min.

Fig. 5(b). They are the spectra comprising the lower temperatures (below 260 1C) and the higher temperatures (above 260 1C). Guided by this analysis, 2D correlation spectroscopy was employed to the three sets of N719 spectra. Although not shown here, the synchronous 2D correlation spectra can be divided for lower, intermediate, and higher temperature regions, respectively. The synchronous 2D correlation spectra for three temperature regimes are found to be completely different. For the lower temperature region, the band intensity near 2096 cm 1 changed greatly. All band intensities increased concomitantly with temperature, while that at 2096 cm 1 decreased. A new band at 1973 cm 1, which was not found at lower temperatures, was observed in intermediate temperatures. Interestingly the band intensities at 2096 and 1723 cm 1 decreased as the temperature increased, while that at 1973 cm 1 increased. The synchronous 2D correlation spectrum at the higher temperature region is clearly different from those below 340 1C. All band intensities near 1150 cm 1 increased greatly with increase in the temperature, but those at 1977 and 1575 cm 1 decreased. In contrast, the DRIFT spectra of N719 on TiO2 can be analyzed by the synchronous 2D correlation spectra from the first set of temperature-dependent spectra (25–260 1C) and the second one (280–400 1C). The synchronous 2D correlation spectrum at lower temperatures is clearly different from that at higher temperatures (see supporting information). Below 260 1C, the band intensity at 2008 cm 1 increased significantly as the temperature in-

creased, while that at 2096 cm 1 decreased. However, a new band at 1879 cm 1 appeared above 260 1C. The band intensities at 2010, 1879, and 1595 cm 1 decreased concomitantly as the temperature further increased, while that at 1364 cm 1 increased. The band intensity at 1595 cm 1 at higher temperatures decreased much more substantially than the case at lower temperatures. In the previous temperature-dependent experiment of 10,12pentacosadiynoic acid, the temperature for the phase transitions on metal substrates appeared to be lower than those in the solid state [34]. The driving force of the reorganization from the disordered to the ordered state from the substantial van der Waals interaction between organic molecules appeared to be lower on the metal substrates. In the case of N719 after the adsorption on TiO2, the interaction between adsorbates could become lower, which may lead to a lower transition temperature. The lower transition temperature for the adsorbed state may indicate that the NCS group of the N719 dye should also interact with TiO2 nanoparticle surfaces. Considering the sizes of N719 and TiO2 nanoparticles [22], the NCS group of N719 may be affected by the other TiO2 nanoparticles, resulting in degradation at lower temperature. On the other hand, increase in heating time from 10 to 20 min resulted in faster degradation at lower temperature. Our experiments suggest that the duration time for heating also kinetically affects the thermal degradation. Fig. 6 shows the effect of temperature on the photovoltaic properties of DSSCs as a function of temperature. Both opencircuit voltage (VOC) and short-circuit current (JSC) decreased slowly and monotonically with the increase in temperature for heat treatment up to ca. 120–140 1C, where the NCS band at 2008 cm 1 started to become comparable to or stronger than that at 2095 cm 1. The fill factor (FF) and the overall efficiency (Z) were decreased accordingly. All these parameters decreased significantly when the dye-loaded TiO2 photoelectrode was treated over 140 1C and there was almost no photocurrent above 200 1C of heat treatment due to decomposition of the NCS group and this result seems consistent with the temperature-dependent infrared spectra of N719 on TiO2 powder surfaces. All these results indicated that the decomposition of the NCS band in N719, and therefore the degradation of the dye functionality, should occur at much lower temperature when it is adsorbed on TiO2 surfaces than in its solid state. It is admitted, however, that such a temperature as above 200 1C should be too high to provide information on the practical usages of DSSCs. We attempted to check the irreversibility of the NCS stretching band by increasing the temperature from 20 to 100 1C and then lowering back the temperature to 20 1C as shown in Fig. 7. It was found that the intensified feature at 2000 cm 1 in the NCS stretching region as the temperature increased would not come back to the original value if the sample was heated as high as at 100 1C. Our result indicates that the NCS stretching band can provide useful information on the thermal degradation of the N719 sample in DSSCs. We obtained thermogravimetric data of N719 in solid and adsorbed states. It seems the overall weight loss for N719/TiO2 system occurred in wider temperature range but it can be attributed to the constant dehydration or solvent drying of TiO2 nanoparticles and the overall pattern of weight loss was essentially similar in both cases. A dramatic change was observed at 270 and 340 1C for N719 in solid and adsorbed states, respectively, indicating the thermal decomposition of the dye. These results indicate that the decomposition of the NCS band in N719, and therefore the degradation of the dye functionality, keeps occurring at the temperature above 250 1C and it is very likely that these new infrared bands, at 1975 and 1875 cm 1, should be originated from a decomposed N719. Fig. 8 illustrates

ARTICLE IN PRESS

0.3 20oC 40 60 80 100 80 60 40 20

0.2

0.1 1000

1500

2000 -1)

2008


0.4 2008



0.10

863

20oC 40 60 80 100 80 60 40 20

0.05 1900 2000 2100 2200 Wavenumber (cm-1)

Wavenumber (cm

100

100

80

99 98

60

97 40 N719 dye 20

Weight (%)

Weight (%)

Fig. 7. Irreversibility of NCS stretching bands by increasing from 20 to 100 1C and then lowering down to 20 1C. The spectra region between (a) 1000–2200 and 1900– 2200 cm 1.

96

N719 dye on TiO2 95 125

225 325 Temperature (°C)

425

Heat flow endo down (mW)

19

19.0

N719 dye N719 dye on TiO2

20

19.5

21

20.0

22 20.5 23

Heat flow endo down (mW)

25

21.0 24 50

100

150 200 Temperature (°C)

250

300

Fig. 8. (a) Thermogravimetric curve for N719 and N719 on TiO2. Please note that the y-axis scales are different for the two samples. (b) Differential scanning calorimetry curve for the two samples indicate a phase change at E200 1C and above E300 1C for N719 and N719 on TiO2, respectively. The thermal behavior above 300 1C had to be inferred from the thermogravimetric curve.

its solid state and N719 on TiO2 surfaces, respectively. Our 2D correlation spectroscopy and principal component analysis (PCA) applied to the temperature-dependent DRIFT spectra of N719 in its solid state and on TiO2 powder surfaces indicated that the N719 spectra in its solid state can be separated into three groups for lower temperature (25–200 1C), intermediate temperature (220–320 1C) and higher temperature (340–400 1C) regions, while the spectra of N719 on TiO2 powder surfaces can be separated into the two groups for the lower temperatures (below 260 1C) and the second comprising spectra for the higher temperatures (above 260 1C). Despite our attempt to perform NMR and mass spectrometry experiments of the N719 dye on TiO2 surfaces, we could not obtain a good quality of spectra. We examined another ruthenium dye such as the ruthenium 505 dye (cis-dicyanobis(2,20 -bipyridyl-4,40 -dicarboxylic acid)), as in the case of N719, the decomposition temperature on TiO2 was found to be lower than in its solid state. Our work should be useful to understand the thermal stability and structural change for one of the most widely used ruthenium dyes for the dye-sensitized solar cells.

4. Summary and conclusions The surface-induced thermal degradation mechanism of the self-assembled thin films of (Bu4N)2[Ru(dcbpyH)2-(NCS)2] (N719) anchoring on TiO2 surfaces via its carboxylate group has been examined using diffuse reflectance infrared Fourier transform (DRIFT) spectroscopy. The thermal decomposition of the NCS group of N719 appeared to occur at a lower temperature on TiO2 surfaces ( E270 1C), whereas such decomposition occurred at temperatures much higher than E340 1C in their solid state. Several vibrational features related to the NCS peaks were found to vary quite sensitively between 1850 and 2150 cm 1 by heating the sample.

Acknowledgements

thermogravimetric and differential scanning calorimetry data, which appeared to be consistent with the previous report [35]. Our thermal analysis data supported that the major decomposition should occur at E270 and E340 1C for N719 in

This work was supported by the National Research Foundation of Korea (NRF) Grant funded by the Korea government (MEST) (no. 2009-0065428), the National Research Foundation of Korea (NRF) grant funded by the Korea government (MEST) (No. 20080062166), and the National Research Foundation (2009-0074905).



References ¨ [1] M. Gratzel, Photoelectrochemical cells, Nature 414 (2001) 338–344. [2] J. Liu, W. Wang, H. Yu, Z. Wu, J. Peng, Y. Cao, Surface ligand effects in MEH-PPV/ TiO2 hybrid solar cells, Sol. Energy Mater. Sol. Cells 92 (2008) 1403–1409. ¨ [3] A. Hagfeldt, M. Gratzel, Molecular photovoltaics, Acc. Chem. Res. 33 (2000) 269–277. [4] N.A. Anderson, T. Lian, Ultrafast electron transfer at the molecule–semiconductor nanoparticle interface, Annu. Rev. Phys. Chem. 56 (2005) 491–519. [5] K.J. Sreeram, C.P. Aby, B.U. Nair, T. Ramasami, Colored cool colorants based on rare earth metal ions, Sol. Energy Mater. Sol. Cells 92 (2008) 1462–1467. [6] H. Park, E. Bae, J.-J. Lee, J. Park, W. Choi, Effect of the anchoring group in Rubipyridyl sensitizers on the photoelectrochemical behavior of dye-sensitized TiO2 electrodes: carboxylate versus phosphonate linkages, J. Phys. Chem. B 110 (2006) 8740–8749. [7] (a) T. Stergiopoulos, M.-C. Bernard, A.H.-L. Goff, P. Falaras, Resonance microRaman spectrophotoelectrochemistry on nanocrystalline TiO2 thin film electrodes sensitized, Coord. Chem. Rev. 248 (2004) 1407–1420; (b) K. Kilsa, E.I. Mayo, B.S. Brunschwig, H.B. Gray, N.S. Lewis, J.R. Winkler, Anchoring group and auxiliary ligand effects on the binding of ruthenium complexes to nanocrystalline TiO2 photoelectrodes, J. Phys. Chem. B 108 (2004) 15640–15651. [8] A. Hinsch, J.M. Kroon, R. Kem, I. Uhlendorf, J. Holzbock, A. Meyer, J. Ferber, Long-term stability of dye-sensitized solar cells, Prog. Photovoltaics Res. Appl. 9 (2001) 425–438. ¨ [9] P.M. Sommeling, M. Spath, H.J.P. Smit, N.J. Bakker, J.M. Kroon, Long-term stability testing of dye-sensitized solar cells, J. Photochem. Photobiol. A: Chem. 164 (2004) 137–144. [10] H.G. Agrell, J. Lindgren, A. Hagfeldt, Degradation mechanisms in a dyesensitized solar cell studied by UV–VIS and IR spectroscopy, Sol. Energy 75 (2003) 169–180. [11] H. Kusama, H. Arakawa, Influence of alkylaminopyridine additives in electrolytes on dye-sensitized solar cell performance, Sol. Energy Mater. Sol. Cells 81 (2004) 87–99. [12] C. Pe´rez Leoń, L. Kador, B. Peng, M. Thelakkat, Influence of the solvent on the surface-enhanced Raman spectra of ruthenium (II) bipyridyl complexes, J. Phys. Chem. B 109 (2005) 5783–5789. ¨ [13] Z. Zhang, S.M. Zakeeruddin, B.C. O’Regan, R. Humphry-Baker, M. Gratzel, Influence of 4-guanidinobutyric acid as coadsorbent in reducing recombination in dye-sensitized solar cells, J. Phys. Chem. B 109 (2005) 21818–21824. [14] K.S. Finnie, J.R. Bartlett, J.L. Woolfrey, Vibrational spectroscopic study of the coordination of (2,20 -bipyridyl-4,40 -dicarboxylic acid) ruthenium (II) complexes to the surface of nanocrystalline titania, Langmuir 14 (1998) 2744–2749. [15] J.C. Love, L.A. Estroff, J.K. Kriebel, R.G. Nuzzo, G.M. Whitesides, Self-assembled monolayers of thiolates on metals as a form of nanotechnology, Chem. Rev. 105 (2005) 1103–1169. [16] J.T. Yates, T.E. Madey, Vibrational Spectroscopy of Molecules on Surfaces, Plenum Press, Plenum, New York, 1987. [17] S. Kim, K. Ihm, T.-H. Kang, S. Hwang, S.-W. Joo, Binding property and structure of aromatic isocyanide self-assembly monolayers on Ag and Au surfaces, Surf. Interface Anal. 37 (2005) 294–299.

[18] D.S. Kang, K.-S. Kwon, S.I. Kim, M.-S. Gong, S.S.A. Seo, T.W. Noh, S.-W. Joo, Temperature-dependent Raman spectroscopic study of the nematic liquid crystal 4-n-pentyl-40 -cyanobiphenyl, Appl. Spectrosc. 59 (2005) 1136–1140. [19] J.K. Lim, B.K. Yoo, W. Yi, S. Hong, H.-j Paik, K. Chun, S.K. Kim, S.-W. Joo, Photodetachment of aryl moieties from covalently functionalized singlewalled carbon nanotubes by UV laser irradiation, J. Mater. Chem. 16 (2006) 2374–2379. [20] L. Lucarelli, V. Nadtochenko, J. Kiwi, Environmental photochemistry: quantitative adsorption and FTIR studies during the TiO2-photocatalyzed degradation of Orange I, Langmuir 16 (2000) 1102–1108. [21] R. Amadelli, A. Maldotti, L. Samiolo, Adsorption and photo-oxidation of 3, 4-dihydroxy-cinnamic acid on TiO2 films, Catal. Today 144 (2009) 149–153. [22] A. Sasahara, C.L. Pang, H. Onishi, STM observation of a ruthenium dye adsorbed on a TiO2 (1 1 0) surface, J. Phys. Chem. B 110 (2006) 4751–4755. [23] H.T. Nguyen, H.M. Ta, T. Lund, Thermal thiocyanate ligand substitution kinetics of the solar cell dye N719 by acetonitrile, 3-methoxypropionitrile, and 4-tert-butylpyridine, Sol. Energy Mater. Sol. Cells 91 (2007) 1934–1942. [24] D.-W. Seo, S. Sarker, N.C.D. Nath, S.-W. Choi, A.J.S. Ahammad, J.-J. Lee, W.-G. Kim, Synthesis of a novel imidazolium-based electrolytes and application for dye-sensitized solar cells, Electrochim. Acta 55 (2010) 1483–1488. [25] H.N. Ghosh, J.B. Asbury, T. Lian, Ultrafast electron transfer dynamics in sensitized TiO2 nanoparticles, PINSA 66 (2000) 177–197. [26] S.-W. Joo, Electric field-induced charge transfer of (Bu4N)2[Ru(dcbpyH)2(NCS)2] on gold, silver, and copper electrode surfaces investigated by means of surface-enhanced Raman scattering, Bull. Korean Chem. Soc. 28 (2007) 1405–1409. [27] Y.M. Jung, I. Noda, New approaches to generalized two-dimensional correlation spectroscopy and its applications, Appl. Spectrosc. Rev. 41 (2006) 515–547. [28] I. Noda, Y. Ozaki, Two-Dimensional Correlation Spectroscopy: Applications in Vibrational Spectroscopy, John Wiley & Sons, Inc., New York, 2004. [29] I. Noda, Generalized two-dimensional correlation method applicable to infrared, Raman and other types of spectroscopy, Appl. Spectrosc. 47 (1993) 1329–1336. [30] The program can be downloaded from the homepage of Prof. Yukihiro Ozaki of Kwansei Gakuin University, Japan. /http://science.kwansei.ac.jp/ ozaki/S. [31] E.R. Malinowski, Factor Analysis in Chemistry, 2nd ed., Wiley, New York, 1991. [32] H. Martens, T. Næs, Multivariate Calibration, John Wiley & Sons, New York, 1991. [33] B.G.M. Vandeginste, D.L. Massart, L.M.C. Buydens, S. De Jong, P.J. Lewi, J. Smeyers-Verbeke, Handbook of Chemometrics and Qualimetrics: Part B, Elsevier Science B.V., Amsterdam, The Netherlands, 1998. [34] J.K. Lim, Y. Lee, K. Lee, M. Gong, S.-W. Joo, Reversible thermochromic change of molecular architecture for A diacetylene derivative 10,12-pentacosadiynoic acid thin films on Ag surfaces, Chem. Lett. 36 (2007) 1226–1228. [35] Md.K. Nazeeruddin, M. Amirnasr, P. Comte, J.R. Mackay, A.J. McQuillan, R. ¨ Houriet, M. Gratzel, Adsorption studies of counterions carried by the sensitizer cis-dithiocyanato(2,20 -bipyridyl-4,40 -dicarboxylate) ruthenium(II) on nanocrystalline TiO2 Films, Langmuir 16 (2000) 8525–8528.