Volume 5 Issue 7 - September 12, 2008
Assembly of CdS Quantum Dots onto Mesoscopic TiO2 Films for Quantum Dot-Sensitized Solar Cell Applications
Yu-Jen Shen , Yuh-Lang Lee*

Department of Chemical Engineering, National Cheng Kung University
*Email: yllee@mail.ncku.edu.tw

Nanotechnology, 19, 045602 (2008)

Font Normal   Font Enlarge

Dye sensitization of mesoporous TiO2 electrodes have been extensively studied recently due to the light-harvesting properties of the dye attached on the surface of the TiO2 nanocrystallite. Energy conversion efficiencies up to 11% have been achieved using ruthenium complex as the sensitizer.1 As an alternative, semiconductor quantum dots (QDs) which adsorb light in the visible region, such as CdS,2 CdSe, 3,4 PbS, 5 InAs, 6 and InP,7 have also been used as sensitizers of DSSCs. It is also possible to utilize hot electrons to generate multiple electron-hole pairs per photon through the impact ionization effect.8 Another advantage of the QD-sensitizers over conventional dyes is their high extinction coefficient, which is known to reduce the dark current and increase the overall efficiency of a solar cell.

Although there are studies devoted to the synthesis and application of QDs for light harvesting in a DSSC cell, the photophysics and photochemistry of QDs are only poorly understood and the energy conversion efficiency is still low (less than 0.5%).5 One of the reasons leading to the poor efficiency of a QDs-sensitized DSSC is the difficulty of assembling the QDs into the mesoporous TiO2 matrix to obtain a well-covered monolayer of QDs on the TiO2 crystalline surface. Bifunctional linker molecules such as mercaptopropionin acid (MPA)3 and 3-mercaptopropyl trimethyoxysilane (MPTMS)3, 9-13 are commonly used as surface modifiers of TiO2 for linking with QDs. An incident photon-to-current conversion efficiency (IPCE) of 12 % was reported for a CdSe-sensitized DSSC.3

The carboxylic acid groups in N3 and “black dyes” molecules are known to play an important role in the adsorption of such dyes onto the TiO2 surface and in the transferring of excited electrons into the TiO2 matrix.14,15 Based on this characteristic, the adsorption principle of the N3 dye was mimicked to assemble the QDs in this work. CdS QDs were surface modified by mercaptosuccinic acid (MSA) to render a surface with carboxylic acid groups. The MSA-modified CdS QDs (MSA-CdS QDs) thus behave like a N3 dye in terms of the adsorption legend attaching to the TiO2 surface (Figure 1a). The two carboxylic acid groups present in a MSA molecule are believed to provide well adsorption of MSA-CdS QDs onto the bare TiO2 surface. As an alternative, TiO2 surface can be modified first using bifunctional linker molecules to render a surface with amine (-NH2) or thiol (-SH) groups. The interactions of the amine group with the carboxylic acid group of MSA-CdS QDs (Figure 1c) and of the thiol group with the CdS surface (Figure 1b) are promising methods to incorporate the CdS QDs onto the TiO2 surface. These assembly strategies are used to prepare CdS QDs-sensitized solar cells and the effects of various methods on the incorporated amount of QDs and on the energy conversion efficiency of the CdS-sensitized DSSC cell are studied in this work.
Figure 1. Schematic model illustrating various bifunctional surface modifiers employed to link CdS-QDs onto the TiO2 surface. The MSA-CdS-QDs adsorbed directly on the bare TiO2 surface using carboxylic acid/TiO2 interaction (a). TiO2 surfaces were modified by MPTMS (b) or APMDS (c) using, respectively, the thiol/CdS (b) or amine/carboxylic acid (c) interaction to incorporate the QDs.

Experimental Section

Synthesis of MSA-modified CdS particles 16,17 CdS QDs were prepared using a water-in-oil reverse micelle process.18 CTAB was used as an emulsifier to prepare a quaternary water-in-oil microemulsion. Colloidal CdS QDs were then prepared by mixing equal volumes of two micellar solutions containing, respectively, 0.5M Cd(NO3)2 and 0.5M Na2S, introduced in the microemulsions as aqueous solutions. CdS QDs formed immediately after mixing the micellar solutions, and the following surface modification procedure was carried out by adding MSA directly into the micellar solution. About 24h of continuous stirring was allowed for the surface reaction and then the solution was evaporated at 50 ˚C to remove hexane. The residue were washed with an ethanol/water mixture (8/2 by volume) to remove the dissolve salts and water-soluble compounds. The powder was then dried under a N2 stream and stored in a dark location. The MSA capped CdS was redispersed in chloroform containing 2000 ppm of CTAB.

Preparation of nanoporous TiO2 photoelectrode. Commercially available indium-tin oxide (ITO) were used to prepare TiO2 photoelectrode. TiO2 paste was spin-coated on the ITO and sintered at 450 ˚C for 30 min. Surface modification of the TiO2 substrate was performed by immersing TiO2 substrate into a 1 wt% MPTMS/toluene solutions or 1 wt% APMDS/chloroform solution for about 5 min. CdS QDs were assembled onto the TiO2 electrodes (bare TiO2, MPTMS-modified TiO2, and APMDS-modified TiO2) by immersing the electrodes in a MSA-CdS QDs/chloroform solution for 16h, followed by rinsing with ethanol to remove the surfactant and weakly adsorbed CdS.

Assembling the photoelectrochemical cell. A 30-nm-thick Pt film was deposited on an ITO substrate used as a counter electrode. The CdS-sensitized TiO2 electrode and a Pt-coated counter electrode were sandwiched using a 60μm thick sealing material  under pressure at about 100 ˚C. An MPN (3-Methoxypropionitrile) solution consisting of 0.1 M lithium iodide, 0.05 M iodine, 0.6M DMPII (1-propyl-2,3-dimethylimidazolium iodide) and 0.5M TBP (4-tert butylpyridine) was used as the redox electrolyte of the DSSC cells. The active area of the cell was 0.16 cm2.

Results and Discussion

The MSA modified CdS QDs were dispersed in chloroform and a well dispersed condition was confirmed by the transparency of its yellow solution and by the TEM image. The UV-vis absorption spectrum of the CTAB-MSA-CdS QDs dispersed in chloroform was shown in Figure 2. The average diameter estimated from the TEM image is 5.8 nm. The adsorption amount of the MSA-CdS QDs can be evaluated using the absorbance at UV-vis excitonic peaks. Figure 3 shown the results indicate the different adsorption rates and equilibrium adsorption amounts of MSA-CdS QDs on various TiO2 surfaces. For the bare TiO2 surface, the absorbance increases slowly with adsorption time and reaches an equilibrium value after about 16 h. For the MPTMS-modified surface, the absorbance increases more quickly in the early adsorption period with an equilibrium value obtained after 14 h. The equilibrium absorbance obtained for the MPTMS-modified surface is more than twice as high as that for the bare TiO2. These results indicate that an MPTMS-modified surface has ahigher ability of incorporating the CdS QDs, attributed to the specific thiol/CdS interaction. On the contrary, the interaction of carboxylic acid with the TiO2 is moderate, leading to a slower adsorption rate and lower incorporation amount of MSA-CdS QDs on the bare TiO2 surface. For the APMDS-modified surface, the rapidly increasing rate of absorbance in the early adsorption period indicates that the reaction between amine (–NH2) and carboxylic acid groups also has an advantage of incorporating the MSA-CdS QDs. Compared with the MSA-CdS spectrum in solution, the onset of absorption shifts slightly towards the longer wavelengths, especially for TiO2-MPTMS-MSA-CdS electrode. A red shift of the spectrum indicates the growth and coalescence of CdS QDs during the assembly process. This results imply that aggregation and coalescence of CdS QDs probably occurred on TiO2-MPTMS-MSA-CdS electrode.
Figure 3. The absorption spectra of the MSA-CdS QDs adsorbed on TiO2 films surface.
Figure 2. UV-vis spectra and TEM image of MSA-CdS QDs dispersed in chloroform at the presence of 2000 ppm CTAB.

Due to the difficulty in imaging the the QDs in the mesopores, glass plates were used as substrates to incorporate the MSA-CdS QDs using the present assembling methods. The surface morphology of glass substrates and the MSA-CdS-modified glass substrates were investiaged using AFM and the results are shown in Figure 4. Figure 4a shows the AFM image of a bare glass surface. The glass substrate has a very smooth topography with a surface root-mean-square (RMS) roughness of about 0.24 nm. For the MPTMS-modified glass surface (Figure 4b), the surface is also smooth. After incorporation the MSA-CdS, outstanding clusters were observed as shown in Figure 4c. The heights of the majority clusters rang between 4 and 7 nm, corresponding to the sizes of the MSA-CdS QDs. The MPTMS-modified glass After adsorption the MSA-CdS QDs for 6h, the surface of the glass substrate becomes very rough as shown in Figure 4d. Most of the clusters are higher than the size of a single QD, indicating the significant aggregation of MSA-CdS on the MPTMS-modified surface.
Figure 4 AFM images of a bare glass surface (a), and a glass surface modified by MPTMS (b).MSA-CdS QDs on a bare glass substrate for 16h (c), MSA-CdS QDs on a MPTMS-modified glass substrate for 6h (d).

Figure 5. The IPCE of various CdS-sensitized TiO2 electrodes as a function of wavelength measured under illumination of an xenon arc lamp (Oriel 300 W).
The QDs-sensitized TiO2 electrodes were used to fabricate DSSC cells using I-/I3- as the redox couple. Figure 5 shows the IPCE of the cells at different wavelengths measured from the short circuit photocurrents. The TiO2-MSA-CdS electrode exhibits the best IPCE performance among the three electrodes and the maximum IPCE value of 20 % was observed at around 400 nm. For the other electrodes with higher CdS incorporation amounts, the IPCE value at 400 nm is about 6 % for the (TiO2-APMDS-MSA-CdS) electrode and about 13 % for the (TiO2-MPTMS-MSA-CdS) one. The photocurrent-voltage (I-V) curves for the DSSC cells were measured at an illumination of 100% sun (AM 1.5, 100 mW/cm2) and the results are shown in Figure 6. Fill factors higher than 0.7 were obtained for these cells.
Figure 6. The I-V characteristics of the CdS-sensitized TiO2 electrodes measured under illumination AM1.5, 100 mW/cm2, and under dark conditions (shown in the lower part).
The open circuit voltage (VOC) and short circuit current density (ISC) measured for the (TiO2-MSA-CdS) electrode are, respectively, 573 mV and 0.71 mA/cm2, and the overall efficiency (η) is 0.30 %. For the other two electrodes, the overall efficiencies are nearly identical (0.19%) with lower VOC and ISC values compared to the TiO2-MSA-CdS electrode. Comparing the results of the UV-vis absorbance and the photochemical response, it is clear that the energy conversion efficiency of the QDs-sensitized DSSCs is not directly proportional to the incorporation amount of QDs on the TiO2 electrode. Therefore, the TiO2/CdS interfacial structure, including the arrangement of the QDs in the mesoporous matrix and the linker molecules between TiO2 and QDs, may play an important role in carrier transport, including the injection of the excited electron to TiO2 and the scavenging of the photogenerated holes, which, thereby, affect the efficiency of a DSSC cell.

The dark current measurement was also performed on the CdS-modified TiO2 electrodes and the result is shown in the lower part of Figure 6. The applied voltage required to drive the electrons across the CdS-modifed photoelectrodes increases in the order: (TiO2-MPTMS-MSA-CdS) < (TiO2-APMDS-MSA-CdS) < (TiO2-MSA- CdS), which also implies the increasing energy barrier of these electrodes for an injected electron in the TiO2 to recombine with the oxidized species in the electrolyte.20,21 This result also shows that the (TiO2-MSA-CdS) electrode is superior to the others in retarding the interfacial recombination of injected electrons from TiO2 to the electrolyte. The superior property of the (TiO2-MSA-CdS) electrode in inhibiting charge recombination is attributed to the better coverage of QDs on the TiO2 surface, which is also responsible for its higher energy conversion efficiency. In contrast, the higher dark currents obtained for the MPTMS and APMDS-modified electrodes imply a lower coverage of QDs on the TiO2 surfaces.

By these results, a surface that has a fast adsorption rate to the QDs (e.g., MPTMS- or APMDS-modified TiO2 film), the pore size will decrease quickly due to the anchoring of QDs on the pore surface. Therefore, the bottleneck of a mesopore is easily blocked due to the quick adsorption rate and higher adsorption amount of QDs on the surface. In such a case, the inner region of a pore cannot be well covered by the QDs due to the difficulty of diffusing across the bottleneck for the QDs. For the adsorption of MSA-CdS QDs on a bare TiO2 film, the adsorption rate is much slower and a longer time is required to immobilize a QD on the pore surface. Therefore, the QDs to transfer to the inner region of a pore and form a better-covered CdS layer. The different arrangements of QDs in the TiO2 matrix caused by the various assembly methods, as well as the charge transfer resistance of the surface linkers, are responsible for the different conversion efficiencies of the QD-sensitized DSSCs.


In summary, adsorption of MSA-modified CdS QDs onto bare TiO2 films using the carboxylic acid/TiO2 interaction appears to be an efficient method to obtain a better-covered QDs monolayer in the mesoporous matrix, attributable to the moderate interaction and adsorption rate of the MSA-CdS QDs on the TiO2 surface. The strong thiol/CdS or amino/carboxylic acid interaction caused the clogging of QDs at the bottleneck of a mesopore, obstructing the inner diffusion of QDs. The QDs coverage in the pores was thus decreased. Furthermore, MSA was also proved to be a surface modifier with a lower charge transport resistance comapred to that of MPTMS and APMDS linkers. The arrangements of QDs in the TiO2 matrix, as well as the charge transfer resistance of a surface linker, are responsible for the conversion efficiencies of the QD-sensitized DSSCs.

  • Nazeeruddin M K , De Angelis F , Fantacci S , Selloni A , Viscard G. , Liska P , Ito S , Takeru B and Gratzel M 2005 J. Am. Chem. Soc. 127 16835
  • Vogel R , Hoyer P and Weller H 1994 J. Phys. Chem. 98 3183
  • Robel I , Subramanian V , Kuno M and Kamat P V 2006 J. Am. Chem. Soc. 128 2385
  • Liu D and Kamat P V 1993 J. Phys. Chem. 97 10769
  • Plass R , Pelet S , Krueger J , Gratzel M and Bach U 2002 J. Phys. Chem. B 106 7578
  • Yu P , Zhu K , Norman A G , Ferrere S , Frank A J and Nozik A J 2006 J. Phys. Chem. B 110 25451
  • Zaban A , Micici O I , Gregg B A and Nozik A J 1998 Langmuir 14 3153
  • Nozik A J 2002 Physica E 14 115
  • Sato H , Ohtsu T and Komasawa I 2000 J. Colloid Interface Sci. 230 200
  • Schmitt J , Mächtle P , Eck D , Möhwald H and Helm C A 1999 Langmuir 15 3256
  • Shipway A N , Katz E and Willner I 2000 Chemphyschem 1 18
  • Granot E , Patolsky F and Willner I 2004 J. Phys. Chem. B 108 5875
  • Zayats M , Kharitonov A B , Pogorelova S P , Lioubashevski O , Katz E and Willner I 2003 J. Am. Chem. Soc. 125 16006.
  • Mwaura J K , Zhao X , Jiang H , Schanze K S and Reynolds J R 2006 Chem. Mater. 18 6109
  • Stathatos E , Lianos P , Laschewsky A , Ouari O and Van Cleuvenbergen P 2001 Chem. Mater. 13 3888
  • Shen Y J , Lee Y L and Yang Y M 2006 J. Phys. Chem. B 110 9556
  • Henglein A 1989 Chem. Rev. 89 1861
  • Veinot J G C , Galloro J , Pugliese L , Pestrin R and Pietro W J 1999 Chem. Mater. 11 642
  • Agostiano A , Catalano M , Curri M L , Della Monica M , Manna L and Vasanelli L 2002 Micron 31 253.
  • Krüger J , Plass R and Grätzel M 2002 Appl. Phys. Lett. 81 367
  • Krüger J , Plass R , Cevey L , Piccirelli M and Grätzel M 2001 Appl. Phys. Lett. 79 2085
< previousnext >
Copyright National Cheng Kung University