(657b) Growth of ZnO Nanowires for Dye-Sensitized Solar Cells

Nanostructured mesoporous TiO₂ films are used as photoanodes in novel solar-to-electric energy conversion devices such as dye-sensitized solar cells (DSSCs) [1]. In a DSSC, a mesoporous TiO₂ film, made up of 5-10 nm diameter nanoparticles is deposited on a transparent conducting oxide (TCO) and a monolayer of a dye is adsorbed onto the nanoparticle surfaces to form a photosensitized anode. The porous space between the particles is filled with an electrolyte containing the I₃^-/I^- redox couple and the cell is completed by sandwiching this nanostructured TiO₂-dye-electrolyte interface between the TCO and a photocathode. The incident photons excite electrons from the HOMO to LUMO levels in the dye and these photoexcited electrons are injected into the TiO₂ nanoparticles. The charged dye is reduced through an electrochemical reaction with I^- in the electrolyte. The electrons hop from particle to particle, flow through the external load and reduce I₃^- at the cathode to complete the circuit. Electron transport through the nanoparticle film is via random walk with multiple trapping and detrapping events. This hopping mechanism increases transport time across the mesoporous film and puts stringent limits on the electron recombination kinetics at the TiO₂ electrolyte interface; essentially, only electrolytes or hole conducting media with extraordinarily slow electron recombination at the TiO₂ surface can be used to achieve high electron collection efficiencies. Thus, high efficiency DSSCs have been limited to liquid electrolytes containing I₃^-/I^-, the only known hole conducting media with electron recombination much slower than electron transport. This inflexibility with the hole conducting medium has limited the DSSC open circuit voltages, which is determined by the difference between the I₃^-/I^- redox level and conduction band edge of the TiO₂. Slow electron transport also limits the film thicknesses to less than the electron diffusion length and prevents further increase of the film's optical density in the infrared region of the spectrum where dye absorption cross sections are smaller.

Recently, there has been increased interest in replacing the nanoparticles with one dimensional nanostructures such as nanowires, nanotubes or chains of nanoparticles [2-4]. This interest is based on the conjecture that such one dimensional nanostructures would provide direct transport from the point of injection to the anode without particle-to-particle hopping. This may result in faster electron transport and/or slower recombination allowing thicker photoanodes to be assembled with hole conducting media other than the ubiquitous I₃^-/I^- electrolyte.

The two candidate materials that can be used as the wide band gap semiconductor in DSSCs are ZnO and TiO₂. The first nanowire based DSSCs were made from ZnO nanowires due to the relative ease of their synthesis compared to TiO₂ nanowires. So far, the highest overall efficiency of a ZnO nanowire DSSC is ~1.7% [5]. This is a factor of six lower than that achieved using TiO₂ nanoparticle DSSCs. The lower overall energy conversion efficiency is due to the significantly lower light harvesting efficiency of the ZnO nanowire DSSCs as compared to TiO₂ nanoparticles DSSCs. The light harvesting efficiency of the ZnO nanowire DSSCs is low because the surface area of the nanowires is an order of magnitude less than that of the particle films. Improving the ZnO nanowire solution synthesis process to make nanowire arrays with surface areas similar to that of TiO₂ nanoparticles films would greatly improve the efficiency of ZnO nanowire DSSCs.

The solution synthesis of zinc oxide nanowires is a two step process. First, a conducting oxide substrate is coated with a ZnO seed layer and the nanowires are grown from the seed layer in an aqueous solution of zinc nitrate and methenamine. The nanowires do not all grow perpendicular to the substrate and as they grow longer they run into neighboring wires and stop growing. As a result the areal nanowire number density decreases as the height of the nanowire film is increased. This results in lower surface areas for dye adsorption and, consequently, lower light harvesting and overall energy conversion efficiencies. To make nanowire films with higher surface area one must (i) increase the density of nanowires, (ii) orient them perpendicular to the substrate and (iii) reduce their radial growth rate so that they do not grow into each other to form a continuous film. In short, one must form a very dense carpet of thin (~10-100 nm), long (10-100 microns), nanowires oriented perpendicular to the surface. Understanding the nanowire growth mechanism may help elucidate ways to engineer the solution synthesis process such that thinner, longer more dense nanowire arrays can be synthesized. Additionally, most of the reactants are lost to homogenous growth during the reaction and do not contribute to nanowire growth. Understanding of the growth mechanism may also lead to more efficient synthesis conditions and strategies.

In this presentation we describe our work exploring the growth mechanism of ZnO nanowires from zinc nitrate and methenamine. Using EDTA titration of zinc ions we found that the depletion rate of the zinc in the nanowire synthesis reaction is first order with respect to zinc ion concentration and with respect to the initial methenamine concentration. Proton NMR was used to monitor the methenamine concentration during the reaction. The methenamine concentration decreases during the first 3 hours of the reaction and then remains constant. The nanowires reach higher overall lengths when grown in solutions with lower methenamine concentrations but the growth rate slows over time.

While all hydrothermal growth processes are intended to be isothermal, in reality they rarely are and it can take many minutes or hours for the growth vessel and the growth solution to reach the steady state temperature. Indeed, in situ measurements of the solution temperature evolution with time in ZnO nanowire growth showed that the thermal inertia of the beaker and the solution determine the transient temperature rise during the synthesis reaction and it may take as long as 1.5 hours for the solution to reach the steady state growth temperature. This temporal evolution of the solution temperature strongly affects the nanowire growth. When the synthesis is done isothermally at 90°C by mixing solutions preheated to this temperature, the pH remains constant at ~5.8. If the solution is allowed to heat up to 90°C over a 1.5 hour time period, the pH starts at ~6.8 and then drops to 5.8 over the 1.5 hour time period that it takes the solution to heat to 90°C, and then remains at 5.8. Nanowires grown in an isothermal solution at 90°C are shorter than those that are grown in a solution that is gradually heated to 90°C. This difference is due to the competition between homogeneous growth which yields particles in solution and heterogeneous growth which yields nanowires on the substrate. At high temperatures the ZnO precursor supersaturation is higher causing homogenous growth to dominate and deplete the precursor too fast to allow for long periods of heterogeneous nanowire growth. In contrast, when heating slowly, the substrate is exposed to lower precursor supersaturation for longer times and heterogeneous growth can occur without significant depletion of the ZnO precursor via homogeneous growth. This information on ZnO nanowire growth may allow us to engineer the temporal evolution of the temperature and reactant concentration profiles to yield longer, thinner more dense nanowires.

References:

[1] O'Regan B, Grätzel M Nature 1991, 353, 737-40

[2] Baxter J B, Aydil E S Appl. Phys. Lett 2005, 86, 053114

[3] Zhu K, Neale N R, Miedaner A, Frank A J Nano Lett 2007, 7, 69-74

[4] Adachi M, Murata Y, Takao J, Jiu J, Sakamoto M and Wang F, J. Am. Chem. Soc. 2004, 126, 14943-49

[5] Gao Y, Nagai M, Chang T, Shyue J, Cryst. Growth Des., 2007, 7, 2467-2471