Towards Optimization of Materials for Dye-Sensitized solar cells

Towards Optimization of Materials for Dye-Sensitized solar cells

(Parte 1 de 2)

Towards Optimization of Materials for Dye-Sensitized Solar Cells

By Yanhong Luo, Dongmei Li, and Qingbo Meng*

Humans need energy for almost any type of activity that they perform. It has been estimated that the world energy consumption will double within the next 50 years, whereas the oil reserves in the earth will run out during this century.[1] The increasing energy demands and shortage of fossil fuels have driven people to develop new types of clean and sustainable energy. The supply of energy from the sun to the earth is gigantic and inexhaustible, and would be a good choice to fulfill our needs. Photovoltaic technology is one of the most favorable ways to convert solar energy and is receiving extraordinary attention.

As a new type of photovoltaic technology, dye-sensitized solar cells (DSCs) have attracted widespread interest owing to their potentially low cost of production and relatively high energy conversion efficiency.[2] Typically, a DSC consists of a dye-

sensitized mesoscopic TiO2 photoanode, a Pt counter electrode, and the electrolyte with the I /I3 redox couple. The working principle of the DSC is shown schematically in Figure 1. First, the dye is excited by absorbing the incoming photon (A in Fig. 1) and rapidly injects an electron into the conduction band (CB) of TiO2 (B). Then the electron goes through an external circuit and arrives at the counter electrode, where an I3 ion is reduced (C). Finally the dye is regenerated by I (D). Besides these forward reactions, there also exist some undesirable back reactions, including the decay of the dye excited state (E) and the recombination of the injected electrons with excited dyes or I3 ions (F,G). In fact, the generation of electricity is the result of the dynamic competition between the forward and back reactions.[2d,e] An efficient cell should facilitate the forward reactions and hinder the back reactions.

Careful engineering of all the components, based on a thorough understanding of the physical processes in DSCs, is very important in improving efficiency and designing devices with new concepts. Here, we will discuss some of our recent advances regarding the material and processingtechnique development of DSCs. Considerations on stability, low cost, and environmental-friendliness of the system will be highlighted. Some important contributions from other groups will also be included.

Beijing National Laboratory for Condensed Matter Physics Institute of Physics, Chinese Academy of Sciences Beijing 100190 (P.R. China) E-mail:

DOI: 10.1002/adma.200901078

Dye-sensitized solar cells (DSCs) have received widespread attention owing to their low cost, easy fabrication, and relatively high solar-to-electricity con- version efficiency. Based on the nanocrystalline TiO2 electrode,

Ru-polypyridyl-complex dye, liquid electrolyte with I /I3 redox couple, and Pt counter electrode, DSCs have already exhibited an efficiency above 1% and offer an appealing alternative to conventional solar cells. However, further improvements in the efficiency and stability are still required to drive forward practical application. These improvements require the cooperative optimization of the component materials, structures, and processing techniques. In this Research News article, recent progress in DSCs made by our group are reviewed, including some novel approaches to the synthesis of solid-state and environmentally friendly electrolytes, the fabrication of alternative low-cost nanostructural electrodes, and the control of recombination at the interfaces.

Figure 1. Schematic illustration of the working principle of a dye-sensitized solar cell. Also shown are the different electron transfer processes in the solar cell: A) dye excitation, B) electron injection, C) electron transfer, D) dye regeneration, E) excited dye decay to ground state, F) recombination of the injected electron in TiO2 to the oxidized dye, G) recombination of the injected electron in TiO2 or FTO to I3 . FTO: Fluorine-doped tin oxide.

The dye-sensitized nanocrystalline semiconductor photoanode plays an important role in converting photons into electrical energy. Light harvesting, electron injection and collection, and unwanted electron recombination are all connected with the photoanode. In order to develop DSCs with high efficiency, optimization of the photoanode morphology is highly desirable. The ideal photoanode should have a nanostructured mesoscopic morphology, which can give a high specific surface area for dye adsorption. In addition, other factors, such as optical path length, porosity, and connectivity, should also be carefully considered to optimize the photoanode performance.

2.1. Optical Design of the Photoanode

According to the Lambert–Beer law, the DSC’s light harvesting efficiency is determined by the dye extinction coefficient, the attached dye concentration, and optical path length within the film. Engineering the structure to increase the optical path length in the photoanode can extend the photoresponse of any dye. This length can be larger than the thickness of the film if light is scattered within it or reflected at the back of the cell. In order to investigate the influence of scattering on the light-harvesting and photovoltaic performance of DSCs, several models have been developed.[3] These models are mainly based on Mie theory or the four-flux model. Recently, we have developed a model that correlates the optical process and the electrochemical process.[3d] In this model, the four-flux model has been adopted to describe the optical absorption and scattering process, while a onedimensional electrical model is used to depict the electrochemical process in a DSC. It can be used to quantitatively correlate the photovoltaic parameters with the optical properties of DSCs. The calculated results indicate that a diffuse reflecting plane at the counter electrode or a suitable mixture of small and large particles in the photoanode can greatly enhance the diffusion light flux and thus significantly increase the short-circuit current density JSC and final conversion efficiency in a DSC. This calculation result has also been verified by several experimental studies by different groups.[4]

2.2. Hindering Recombination in the Photoanode

Controlling recombination in DSCs is another important way to enhance the conversion efficiency. It has been demonstrated that recombination at the fluorine-doped tin oxide (FTO)/electrolyte interface must be taken into account, especially for a system with solid-state electrolytes or organic dyes.[5] Normally a thin TiO2 underlayer between the FTO and the porous TiO2 layer has been used to reduce this kind of recombination.[6] The underlayer should completely cover the FTO surface for effective separation and also be sufficiently thin to collect electrons efficiently.

There are several methods of preparing the TiO2 underlayer, for example, spray pyrolysis, spin coating, and dip coating.[5c,6b,6c]

Recently, we have developed a feasible and low-cost method of fabricating an underlayer using screen printing.[6a] The thickness of this underlayer can be readily controlled by changing the composition of the paste. By introducing an underlayer of about 30nm thickness into the DSC, we achieved an efficiency of 8.5%, which is much higher than the efficiency of 7.4% for the cells without a compact layer.

2.3. Low-Temperature Fabrication of the Photoanode

Lightweight and flexible DSCs with a plastic substrate have attracted much attention because they are suitable for roll-to-roll mass production.[7] However, typical plastic substrates cannot withstand the sintering process up to 4508C that is usually used in fabricating high-efficiency DSCs with FTO glass substrate. Thus developing a low-temperature fabrication route for the photoanode is a challenge. Interconnection between the nanoparticles and surface activation are two crucial factors for this low-temperature process. Using aqueous solution for lowtemperature coating is a promising strategy, which has the advantages of low cost, nontoxicity, and easy disposal of production waste. Recently, we have developed a feasible method of fabricating ZnO photoanodes at room temperature.[7a] An aqueous ZnO paste is prepared by milling the suspension of nanoparticle ZnO in acetic acid solution. Acetic acid can etch the surface of ZnO particles to form zinc acetate, which will act as a binder to connect the ZnO particles together during film preparation. In addition, an effective surface activation can be achieved by an ammonia treatment, and DSCs with 4.5% efficiency have been obtained. Flexible DSCs with an ITO-PET (indium tin oxide–coated poly(ethylene terephthalate)) substrate and a gel electrolyte have been fabricated with the same process and presented an efficiency of 3.8% with the following photovoltaic characteristics: open circuit voltage VOC ¼0.556V, JSC¼9.9mA cm 2, and fill factor FF¼0.68.

3. Electrolytes

In DSCs, the electrolyte is responsible for charge transport between the photoanode and the counter electrode. Most highly efficient DSCs have been constructed with liquid electrolytes with the I /I3 redox couple and volatile organic solvents. The unique performance of I /I3 liquid electrolytes is mainly attributed to, first, the favorable penetration into the porous film and, second, the fast dye regeneration combined with exceedingly slow recombination. However, the evaporation of toxic solvents and the leakage of liquid electrolytes as well as the sealing problem will significantly limit their practical application. Thus, attempts have been made to solve the above problems by replacing liquid electrolytes with (quasi) solid-state electrolytes, such as hole transport materials (HTMs), ionic liquids, or solid-state composite electrolytes.

3.1. Solid-State Electrolytes Based on HTMs

Solid-state DSCs employing solid HTMs are currently under intense investigation.[8,9] In principle, if a material with p-type semiconducting behavior can accept holes from the dye cation and has no absorption in the visible spectrum region, it can

2 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Mater. 2009, 21, 1–5 Final page numbers not assigned

replace the liquid electrolyte of DSCs. For practical fabrication, the filling of the pores with HTMs and the interface contact between HTMs and the dye/TiO2 film are the main challenges for solid-state DSCs.

The most representative HTMs are CuI and spiro-MeOTAD (2,20,7,70-tetrakis(N,N-di-p-methoxyphenyl-amine)-9,90-spiro-bifluorene). In a first attempt, the efficiencies of the solid-state DSCs based on pure CuI or spiro-MeOTAD were less than 1% because of incomplete filling and bad contact with the dye/TiO2 film.[8b,9a] As shown in Figure 2A, large crystals are formed when CuIis directly deposited on the photoanode. Incorporation of a crystal growth inhibitor into the coating solution has been demonstrated to overcome this problem.[8a,c] In this case, the crystal size of CuI decreases dramatically to the order of nanometers (Fig. 2B) and the adsorbed crystal growth inhibitor layer permits good contact between the CuI crystals themselves as well as with the TiO2 with attached dye. The efficiency increased signifi- cantly, up to 3.8%. In DSCs with spiro- MeOTAD, the photovoltaic performance has also been improved to 5.1% by adding

4-tert-butylpyridine, doping with Li[CF3SO2]2N, and using an amphiphilic ruthenium dye.[9a,b]

3.2. Quasi-Solid-State Electrolytes Based on Ionic Liquids

Ionic liquids possess several advantages, such as high chemical and thermal stability, nonflammability, and wide electrochemical win- dow, and have been regarded as ideal substitutes for volatile liquid electrolytes to avoid the known sealing and stability issues in DSCs.[10] They function not only as the iodide source but also as the solvent. Ionic liquids based on 1,3-dialkylimidazolium iodides have already been widely used in DSCs. However, their relatively high viscosity appears to limit the mass transportation and lead to unsatisfactory device performance. Certain mixtures of ionic liquids containing different anions have been found to decrease viscosity.[10a,b] In the meantime, organic gelators or inorganic additives have been used to solidify ionic-liquid-based electrolytes.[10c,d] Conversion efficiencies of about 6–7% have been achieved.[10d]

Guanidinium molten salts are another kind of ionic liquid that can be prepared in a straightforward manner and have good solubility and thermal properties. We have recently synthesized a new cyclic guanidinium ionic liquid OGI (1,3-dimethyl-2- N00-methyl-N00-octyl imidazo guanidinium iodide) as an electrolyte for DSCs.[1] Initial results have shown that high conversion efficiencies of 5.41% and 6.38% are achieved with light intensity of 100 and 9.81 mW cm 2, respectively. As the viscosity of the OGI ionic liquid electrolyte is about 630 cP (0.63Pa s), further efficiency increase could be expected by mixing a low-viscosity ionic liquid with it.

3.3. Solid-State Composite Electrolytes Based on Addition Compounds

Recently, our group has employed solid-state composite electro- lytes based on LiI addition compounds, such as [Li(HPN)x]I (where HPN is 3-hydroxylproponitrile), LiI-CH3OH, and

LiI-C2H5OH, to replace the liquid electrolytes.[12–14] [Li(HPN)x]I is derived from the reaction between LiI and HPN in different molar ratios.[12a] The phase diagram (Fig. 3A) of the LiI/HPN binary system shows that a single compound [Li(HPN)2]I can be obtained when the molar ratio of LiI/HPN is 1:2.[12a–c] X-Ray crystallographic analysis indicates that three-dimensional (3D) diffusion tunnels for I– ions are available, in view of the geometrical structure along the a-, b-, and c-axes.[12b,c] Obviously, this spatial configuration will benefit I and I3 ion transportation. Theoretical ab initio calculations have been used to further reveal the ion diffusion dynamics. As shown in Figure 3B, the

Figure 2. Scanning electron microscopy (SEM) images of CuI crystals deposited on the TiO2 porous film with added dye: (A) CuI without molten salt. B) CuI with 1-methyl-3-ethylimidazolium thiocyanate. Reproduced with permission from [8a]. Copyright 2003 American Chemical Society.

Figure 3. A) Phase diagram of the LiI-HPN system. (Reproduced with permission from [12a]. Copyright 2004 the Electrochemical Society.) B) Energy barriers for the hopping of I and Li to their neighboring sites along c direction. Inset: X-ray structure plot for [Li(HPN)2]I viewed along the c-axis. calculated activation energy for the hopping of I (0.73eV) to its neighboring sites along the c direction is much lower than that of Liþ (8.39eV), which indicates this addition compound is actually an I– ion conductor.[12e]

A series of solid-state electrolytes based on LiI(HPN)x (2 x 4) addition compounds have been obtained. Among them, on [Li(HPN)4]I gave only 1.8% efficiency. Just as discussed above, the fast crystallization of [Li(HPN)4]I makes it incompletely filled into the TiO2 porous films. In order to inhibit the crystallization,

SiO2 nanoparticles have been introduced to control the crystalline growth rate of the addition compounds. As a consequence, good filling and better interfacial contact between the electrolyte and the electrodes can be obtained. In addition, the homogeneous dispersion of nanoparticles into the electrolyte can increase the ionic conductivity of the electrolyte.[12c,d] Using this [Li(HPN)4]I/

SiO2 composite electrolyte, the efficiency of the DSCs has been dramatically improved to 5.48%.[12c]

Otheraddition compounds derivedfrom thereactions between

LiI and other small organic molecules such as methanol and ethanol have also been investigated.[13,14] Reaction of LiI and

CH3OH affords [Li(CH3OH)4]I, the crystal structure of which has been reported by Rabenau and coworkers.[13c] With

[Li(CH3OH)4]I/SiO2 solid-state composite electrolyte, the DSC only gave 2.7% conversion efficiency. Further addition of triethylamine hydrothiocyanate (THT) ionic liquid obviously improves the cell performance. DSCs based on [Li(CH3OH)4]I/

THT/SiO2 electrolyte exhibit VOC¼0.68V, JSC¼9.1mA cm 2, FF¼0.6, and total efficiency of 4.43%. An environmentally friendly quasi-solid-state composite electrolyte has further been synthesized through the reaction of LiI and ethanol. DSCs assembled with Li(C2H5OH)xI/SiO2 composite electrolyte

environmentally benign electrolyte based on AlI3–ethanol has also been synthesized in situ by our group.[15] A 5.9% power conversion efficiency has been obtained by using the AlI3–ethanol electrolyte containing an appropriate amount of 4-tert- butylpyridine and iodine. Further solidification of AlI3 electrolyte is under way.

4. Carbon Counter Electrodes

The counter electrode (CE) has the function to transfer electrons arriving from the external circuit to the electrolyte including the I /I3 redox couple. An effective CE should have the following properties: 1) good catalytic activity for the reaction I3 þ2e!3I ; 2) chemical/electrochemical stability; and 3) mechanical stability and robustness. In order to minimize the charge-transfer overpotential, highly catalytic material has been used to speed up the reaction I3 þ2e!3I . Pt is the most commonly used catalytic material. However, the noble metal Pt is extremely expensive and may be corroded by the iodide solution.[16] A low-cost electrode with high electrochemical activity and chemical stability is an important requirement to enhance the practical application of DSCs.

(Parte 1 de 2)