Transition metal oxide absorbers for photovoltaics

Abstract

Several oxides are non-toxic, stable, and can be deposited using inexpensive techniques on a variety of substrates. There is a compelling case for use of oxide in photovoltaics, referred to as “all-oxide photovoltaics”. Traditional photovoltaics technologies often use oxides, e.g. as transparent electrodes (SnO2: F. In2O3: SnO2, etc.) or carrier-selective contacts (TiO2, MoOx, etc.). However, oxides are rarely used as light-absorbers because oxides tend to have large bandgap (> 2 eV), low hole mobility (< 10-4 cm2/Vs), and low carrier diffusion lengths (<100 nm). Despite considerable effort, few oxides have been demonstrated as efficient solar absorbers. One class of oxides still poorly investigated, are the multi-cation transition oxides. This is a very large library, so it is conceivable that good solar absorbers are waiting to be discovered. In this work, we present a study of three multi-cation transition metal oxides for their application in solar cells: Zn2Mo3O8, Mn2V2O7, and Ag2CrO4.

Zn2Mo3O8 is a non-centrosymmetric oxide with a low-bandgap of 2.1 eV. Polycrystalline films deposited at room temperature are n-type with Hall mobility of 0.6 - 0.7 cm2V-1s-1. DFT calculations suggest that the valence band is composed of both O 2p and Mo 3d orbitals – which could lead to higher hole mobility than that typical oxides. The valence band composition was experimentally determined using resonant photoelectron spectroscopy, which confirms this assertion. Unfortunately, DFT calculations also show that ZMO has high energy Frenkel excitons (0.78 eV), which will not dissociate at room-temperature, leading to reduced voltage in solar cells. Au/ZMO/TiO2 Schottky diodes with ZMO films deposited at room temperature show photoresponse but no photovoltage. This shows that ZMO is a carrier conducting semiconductor that can be used in a photodetector but not as a solar absorber.

Mn2V2O7(MVO) was previously demonstrated as a photocatalyst with a low bandgap of 1.6 eV. MVO also has a d-contribution to the valence band which leads to high hole mobility. MVO has a low melting point of ≈1080 oC, allowing deposition by PLD at high homologous temperatures which result in films with a large lateral grain size >10 μm on SiO2 and ~ 1 μm on SrRuO3. The conductivity of MVO changes across five orders of magnitude by changing the deposition pressure. The ability to deposit ‘conductive’ MVO allows us to make rectifying devices, but the photovoltaic performance of solar cells is poor. The film deposited at higher temperatures is p-type with high Hall mobility (>200 cm2/V-s). An interdigitated back- contacted cell with carrier-selective TiO2 contact shows improved performance with an open- circuit voltage of 0.33 V and a short circuit current of 0.2 mA/cm2, with an efficiency of 0.017%. N2 -annealing is shown to reduce the Urbach energy from 264 meV to 76 meV, comparable to CIGS which is a commercial solar absorber. The bandgap and crystallinity of MVO are unaltered. Annealed MVO film shows significant solar cell performance with a short circuit current of 0.46 mA/cm2 and an open-circuit voltage of 0.21 V with an efficiency of 0.024% in lateral top contacted solar cells. The annealed film also shows photoluminescence(PL) at low temperatures – another indicator of enhanced electronic quality. The PL is a doublet with a low FWHM of just 100 μeV and separation of 380 μeV at 4.6 K. Analysis of the temperature-dependent PL reveals a pair of shallow dopant levels: a donor at 5.6 meV and acceptor at 48.9 meV from the corresponding band edges, which led to p-type intrinsic doping.

Ag2CrO4(ACO) is a known photocatalyst with low bandgap (1.8 eV) and a low melting point (660 oC). The latter leads to highly crystalline films, even when deposition temperatures are low. DFT shows that ACO has a valence band composed of O 2p and Ag 4d, which could lead to higher hole mobility. Schottky solar cells using ACO show typical photovoltaic behavior, with a short circuit current of 12 μA/cm2 and open-circuit voltage of 0.2 V. However, the power conversion efficiency is just 0.0007%, like other emerging oxide absorbers. An investigation into the possible reasons shows that the Urbach tail at the bandgap is very large (885 meV), a sign of significant electronic disorder in the film.

Just like oxide-absorbers, p-type oxides are also quite uncommon. There are few known p-type wide-bandgap oxides, like Cu2O, but band-alignment and interfaces are often the limiting factors for their integration in devices. Here we successfully integrated-type Cu2O with a silicon absorber to make hole-selective contact. Photoelectron spectroscopy measurements reveal that the band-alignment between Cu2O and Si blocks the flow of electrons from silicon to Cu2O but allows the passage of holes. Interface recombination was reduced by integrating an ultra-thin SiO2 layer between Cu2O and Si. The p-Cu2O/n-Si heterojunction with the passivating SiO2 interlayer showed an open-circuit voltage of 0.53 V, which at the time was a record among cells without back surface passivation.

In summary, the properties of transition metal oxide thin films for were investigated for application in photovoltaics, out of which MVO and ACO proved to be potential candidates as oxide absorbers, and Cu2O a candidate for a hole-blocking layer. This work is expected to contribute to the development of efficient all-oxide-solar cells.