Doping-free Silicon Heterojunction Solar Cells with Metal-oxide or Silicide-based Carrier Selective Contacts

Needs for clean and sustainable energy is gaining more attention owing to severe climate change and increased electricity consumption for daily lives. To fulfill the demand, energy production by renewable energy, especially, by photovoltaic (PV) energy is increased. While PV market is expanding, there is intense competition for developing new technologies aiming at high efficiency and low-cost solar cells and modules. The major PV modules in the market are based on homojunction crystalline silicon solar cells. In the fabrication process, typically, a doping process is used to form a heavily doped region to induce field-effect passivation and lower the contact resistivity (ρc). However, this high doping concentration also leads to optoelectrical losses, such as Schockley-Read-Hall-, Auger recombination and parasitic absorption of free-carriers. However, achieving low ρc with low recombination current density (J0) at the contacts is very challenging without heavily doped region or layers. Achieving a low ρc is more difficult when thermionic emission becomes a major conduction mechanism for carrier collection. It is because the Schottky energy barrier and band bending at the metal-semiconductor (MS) contacts are not easily modified by the metal work function due to a large number of surface states causing metal Fermi level (EF,metal) pinning at the c-Si surface. To resolve these issues, metal-insulator-semiconductor (MIS) contact structure would be an alternative to depin EF,metal on the c-Si surface by blocking the penetration metal wave function which generates a significant density of surface states. Moreover, the large bandgap of the metal oxide may open a new opportunity for reducing parasitic light absorption at the contact. To address high J0,metal and high ρc of doping-free contacts, we have developed a unique electron contact featuring i-a-Si:H/TiOx/low work function metal (ATOM). i-a-Si:H was used for excellent surface passivation, however, i-a-Si:H also resulted in EF,metal pinning when the metal is deposited directly on a-Si:H layers. Then, TiOx was inserted between i-a-Si:H and the metal layer for depinning EF,metal. Low work function metals, such as calcium (work function (Φ) =2.9 eV) and Yb (Φ =2.5 eV), were used to enhance band bending at the contact region. As another approach, a low ρc was achieved by applying low work function metal silicide, such as ytterbium silicide (YbSix). YbSix was grown by annealing an i-a-Si:H/Yb stack at low temperatures below 200 °C. Metal silicides have a relatively weak EF,metal silicide pinning, therefore, the work function effect on the Schottky energy barrier is relatively higher with respect to MS contact. A doping-free hole contact was developed using MoOx which has a high work function. Depending on the process temperature and thickness of MoOx, the performance of the hole contact can be significantly changed. The high band-gap of MoOx gained in short-circuit current density (JSC) compared to the silicon heterojunction (SHJ) solar cells with p-a-Si:H. According to TEM image, interfacial SiOx was observed between MoOx and i-a-Si:H, and the SiOx thickness did not depend on MoOx thickness in the range (6 - 10 nm). However, the relatively weak thermal stability of MoOx needs to be improved. Then, doping-free cells were prepared by combining MoOx based hole- and two kinds of electron contacts, MoOx with i-a-Si:H/TiOx/Yb/Ag (MolyATOM) and MoOx with YbSix (MolYSili). The thermal stability of MolYSili and MolyATOM were improved by an anneal treatment of i-a-Si:H before the MoOx deposition. Improved thermal stability of doping-free cells was confirmed by an annealing test at module level as well as at cell level. Champion MolyATOM and MolYSili cell achieved an efficiency, respectively, of 17.6 % and 17 % at an active cell area of 16 cm2. Silicon solar cells were fabricated without the needs of a doping process and conventional pn junction formation. The induced junction could effectively split the photo-generated carriers and collected well by band bending engineering. This study will be very helpful to understand challenges and opportunity of doping-free cells.status: publishe