Nanostructured ZnO as a solution-processable transparent electrode material for low-cost photovoltaics

A critical part of thin-film composite solar cells is conduction through the transparent electrode. Currently the material of choice for the window electrode is indium tin oxide (ITO). The natural abundance of indium however is insufficient for highvolume production of solar cells. Moreover, good quality ITO is made by expensive vapor-deposition processes. We propose to develop nanostructured degeneratively doped ZnO dispersed in a solvent as a replacement for vapor-deposited ITO. Nanostructured Aldoped ZnO was grown by solid-source MOCVD or was synthesized via colloidal chemistry. Morphological analysis by SEM indicates the formation of 50-100 nm wires. X-ray diffraction confirmed that the nanowires are crystalline ZnO and energy dispersive x-ray spectroscopy (EDS) showed that Al atoms were incorporated in the wires. Preliminary laser-annealing experiments were conducted with a pulsed excimer laser. ZnO nanowires coalesced under repetitive irradiation and X-ray diffraction of annealed films indicated that no decomposition of ZnO into Zn occurred during laser irradiation. Introduction Solar power generation will certainly play an important role in a low greenhouse gas future. Indeed, the sun provides 120,000 TW of energy while the energy needs of the whole planet are only 13 TW. Thin-film solar cells offer the opportunity to dramatically lower the price of solar energy by using small amounts of materials and low-cost manufacturing technologies. In particular, roll-to-roll deposition technologies for fabrication of solar cells on flexible substrates are game-changing in the photovoltaics market. There has been much recent emphasis on developing new semiconductors for flexible thin-film solar cells. A critical part of a thin-film solar cell however is the transparent electrode, which is commonly made with indium tin oxide (ITO). The natural scarcity of indium and the high processing cost of ITO calls for new materials set to replace ITO in order to meet the challenges of low-cost manufacturing of photovoltaic devices. Zn-based oxides are extremely attractive replacement materials for ITO due to the great natural abundance and low toxicity of Zn. ZnO can be made conductive by doping with heteroatoms or by introducing O vacancies. The most effective dopants are Al, Ga and F. By degeneratively doping ZnO films, carrier concentrations as high as ~2x10 cm can be reached. Electronic-grade ZnO films are currently deposited by vacuum-processing. While ZnO would already introduce substantial cost-savings compared to ITO in terms of raw materials cost, the real game-changing technology is the introduction of roll-to-roll fabrication. In order to make roll-to-roll manufacturing of solar cells cost-competitive vacuum deposition must be abandoned in favor of solution deposition. Indeed, liquidphase deposition of materials will allow the adoption of the low-cost and high-throughput patterning technologies used in the printing industry to solar cell production. Here we propose to take advantage of nanostructuring ZnO as well as laser processing to produce low-resistivity high-transparency ZnO films from solution. Results Synthesis and characterization of ZnO nanowires We pursued two routes for the synthesis of ZnO nanowires degeneratively doped with Al. The first route is by solid-source MOCVD (SS-MOCVD) on a 4” Si wafer. The substrate was held at T>425°C while the metalorganic Zn precursor was flowed on the substrate with a base pressure of oxygen of 3 mTorr. The substrate was covered with colloidal Au particles (30 nm diameter) attached to the substrate with a poly(lysine) coating in order to grown ZnO nanowires rather than a film. We found that the growth morphology depended strongly on the chemistry of the substrate surface prior to the attachment of the Au colloids. If the Au colloids were deposited on a Si surface that had a native SiO2 layer, the ZnO growth process resulted in the formation of small (<50 nm) particles. If the native SiO2 was etched prior to attaching the Au colloids, 3-D growth of ZnO occurred. We verified by x-ray diffraction that the ZnO nanostructures are textured with the <001> crystalline direction perpendicular to the substrate (Figure 1) Figure 1: SS-MOCVD reactor (a). Atomic force micrograph of ZnO nanostructures grown on Si substrates with (b) and without (c) native SiO2. X-ray diffraction pattern of ZnO structures grown by SS-MOCVD showing only the ZnO (0002) peak. A second low-cost synthesis route of ZnO is via colloidal chemistry. A solution of Zn acetate and Al acetate in trioctylamine was refluxed under nitrogen for 15 minutes to 2 hours, depending on the desired nanowire length. The reacted solution was then purified, washed and resuspended in ethanol. This type of chemistry gives rise to doped wires, where the doping level is controlled by the initial Al salt concentration. The significant advantage of this approach compared to vapor-phase doping is that we may avoid the formation of Al2O3, which is detrimental to the conductivity of the film. Suspensions of ZnO nanowires and nanoparticles were cast on a Si substrate for SEM characterization (Figure 2). Figure 2: SEM of ZnO nanowires and nanoparticles grown by colloidal chemistry (a,b,c). Micrographs b and c are taken from nanowire network cast on Si. Micrograph b shows that the growth of nanowires and nanorods originates from ZnO platelets. Micrograph c shows that complete coverage can be obtained by simple solution casting. A high-resolution TEM micrograph of single wires (d) indicates that the wires are crystalline (lattice fringes are indicated by the arrows). EDS spectra taken in the TEM confirm that Al is incorporated in the wires. The amount of Al depends on the initial Al acetate concentration during the synthesis (Figure 3). Laser annealing of ZnO nanowires In order to test the feasibility of laser annealing nanowire networks into continuous thin films, we used commercially available intrinsic ZnO nanowires (Nanolab, Inc.). Nanowire dispersions were cast onto Si and glass substrates and laser annealed with an excimer laser (λ=308 nm, τ=25 ns). The effect of environment (air vs. vacuum), pulse energy and number of pulses was explored (Figure 4). Figure 3: EDS spectra of nanowire samples synthesized with different initial Al acetate concentrations: 1.5 at% (a), 10.5 at% (b) and 19.2 at% (c). The intensity of the Al peak increase accordingly. Figure 4: SEM characterization of excimer annealed ZnO nanowires. All films shown here were annealed in air. Laser parameters were as follows: E=120 mJ, 1 shot (a), E=155 mJ, 10 shots (b), E=165 mJ, 1 shot (c), E=129 mJ, 1000 shots (d). As expected, increased pulse energy generates more violent processes, such as cracking and ablation/redeposition. Increasing the number of pulses at fixed energy allows the formation of a film but also generates damage in the form of laser ablation. The effect of environment is unclear at this point. The results are very promising inasmuch as they show that the anisotropic ZnO nanowires melt and recrystallize forming equiaxed grains. No decomposition of ZnO into Zn was observed (data not shown here). Progress This project aims at developing a a low-cost alternative material and a highlyscalable process to replace sputtered ITO as transparent electrode in solar cells. We have demonstrated that highly-doped ZnO nanowires can be synthesized via colloidal growth and that they can be cast into films on a substrate by solution-processing. We have also demonstrated that laser annealing is a promising way to transform nanowire networks into continuous thin films. This combination of materials synthesis and processing has the potential to become an enabling innovation in a game-changing technology such as high-throughput roll-to-roll fabrication of solar cells. Future plans The colloidal synthesis process will be optimized in order to obtain tight nanowire distributions. The Al content will be quantitatively determined as a function of initial Al acetate concentration. The formation of other phases (e.g. Zn, Al2O3) will be determined by high-resolution TEM. Measurements of electrical conductivity and optical transmission of the films will be conducted as well. In-situ heating will be used to determine at what temperature the Al dopants are activated. Finally, the laser-annealing process will be optimized and its effect on conductivity will be determined.