Hot carrier solar cells require mechanisms to dramatically reduce the rate at which carriers thermalize in semiconductors. Now, side-valley trapping of hot carriers with long decay lifetimes is shown to increase the chance of extraction of carriers while they are still hot.
While in a conventional solar cell the excess carrier energy above the conduction and below the valence band edges is lost as heat, in a hot carrier solar cell this energy can contribute to extracted electric power. This enables power conversion efficiency higher than conventional single-junction solar cells1,2. However, such extraction of hot carrier energy is extremely challenging as hot carriers rapidly lose their excess energy as they tend towards thermal equilibrium with the lattice.
To enable hot carrier extraction, the solar absorber material should be able to significantly slow the rate of carrier cooling by at least an order of magnitude3. One potential approach is the scattering of hot electrons into the upper L and X side valleys in the conduction band (as illustrated for InGaAs in Fig. 1a). This intervalley scattering has been observed in various materials such as InP and GaAs4. Once carriers are scattered into side valleys, they require interaction with a longitudinal optical (LO) phonon of large momentum to scatter back into the main Γ valley. As a consequence, the back-scattering process has a long time constant of many picoseconds that results in a significant slowing of carrier thermalization. Side valleys have therefore been suggested as a means of storing hot carriers and potentially allowing them to be extracted whilst still hot5.
While intervalley scattering of electrons coupled with LO phonons has been investigated theoretically for several semiconductors6, practical demonstration and application in hot carrier solar cells has not yet been realized. Writing in Nature Energy, Ian Sellers and colleagues now report trapping of hot carriers in the side valleys of InGaAs and extraction of a small proportion of hot carriers at voltages higher than the absorber band gap — a key requirement for a hot carrier cell7.
The researchers sandwich a 250 nm n-type In0.53Ga0.47As absorber layer between p-type and n+-type Al0.48In0.52As layers, all lattice-matched to the InP substrate. The InGaAs direct band gap is 1.45 eV with the L side valley in the conduction band 0.5 eV above the Γ central valley (Fig. 1a). This L side valley is well aligned with the Γ conduction band edge of the AlInAs emitter, thus facilitating charge transport across the heterojunction.
Sellers and colleagues show that carriers excited by photons with energy above the L valley energy are scattered into the L side valleys of the InGaAs absorber (Fig. 1b). In this side valley, radiative recombination is exponentially reduced and LO phonon emission is linearly reduced. Both effects contribute to slowing the rate of carrier cooling. The researchers also demonstrate that a small number of these hot carriers trapped in the InGaAs L valley are injected into the Γ valley of the AlInAs conduction band and are collected to produce a voltage of 1.25 V. This voltage is significantly higher than the band gap energy of the InGaAs absorber.
To exceed the efficiency limit of conventional solar cells, however, it is also necessary to harvest carriers that are excited by photons with energy below the side-valley energy. To that end, Sellers and colleagues show that carriers with energies below the L valley (Fig. 1c) are accelerated into it and across the junction — where they are extracted from the cell — by built-in electric fields. There are two fields, generated by an external bias of –0.3 V: one of 42 kV cm–1 in the active region and another of 100 kV cm–1 at the heterojunction interfaces. This overall extraction process is similar to the Gunn effect.
Under an operating voltage of 1.5 V, which flattens the bands, the efficiency of carrier extraction is reduced because the electric fields are now only present at the heterojunction interfaces. This could partially explain why current density–voltage curves shown in the study show poor fill factor and hence no gain in efficiency. Sellers and colleagues suggest that collection efficiency could also be limited by poor alignment of momentum for the InGaAs L valley and the AlInAs Γ valley and the need for carriers to tunnel through this heterojunction interface.
The study represents an important step forwards in hot carrier solar cells by demonstrating side-valley trapping as an effective means of extracting hot carriers at a high external voltage. It also identifies limiting factors such as contacting materials and extraction of low-energy carriers that need to be addressed to realize practical hot carrier cells with truly enhanced efficiency.
Ross, R. T. & Nozik, A. J. J. Appl. Phys. 53, 3813–3818 (1982).
Green, M. A. Third Generation Photovoltaics (Springer-Verlag, Berlin/Heidelberg, 2006).
Conibeer, G. J. et al. Sol. Energy. Mater. Sol. Cells 135, 124–129 (2015).
Saeta, P. N. et al. Appl. Phys. Lett. 60, 1477–1479 (1992).
Clady, R. et al. Prog. Photovolt. 20, 82–92 (2012).
Luque, A. & Marti, A. Sol. Energy Mater. Sol. Cells 94, 287–296 (2010).
Esmaielpour, H. et al. Nat. Energy https://doi.org/10.1038/s41560-020-0602-0 (2020).
About this article
Parasitic photon process versus productive photon process: a theoretical study of free-carrier absorption in conventional and hot-carrier solar cells
Journal of Physics D: Applied Physics (2020)