1 Introduction
Silicon solar cells with passivation contact structures utilize hydrogenated amorphous silicon (a-Si:H) or silicon oxide (SiOx) as a highly effective passivation layer[1,2]. This layer is inserted between the crystalline silicon (c-Si) wafer and the carrier transport layer, resulting in drastically reduced interface recombination losses and higher efficiency when compared to PERC (Passivated Emitter and Rear Cell) solar cell. An exemplary type of solar cells using passivation contacts is the silicon heterojunction (SHJ) solar cell, which is known for its high conversion efficiency, simple preparation process and low process temperature. Recently, the power conversion efficiency (PCE) of SHJ solar cell has reached 26.8%[3]. However, the short-circuit current density (Jsc) of SHJ solar cell tends to be lower than that of PERC cell, which is mostly attributable to the parasitic absorptions of intrinsic a-Si:H (a-Si:H(i)), highly-doped nano-crystalline silicon (nc-Si:H), and transparent conductive oxides (TCO) layer[4,5]. Compared to the passivation intrinsic a-Si:H layer, the parasitic absorption loss induced by the highly-doped nc-Si:H layer is more pronounced, as a certain thickness is required to form an effective built-in potential. In addition, highly-doped nc-Si:H also faces problems such as Auger recombination loss and complex doping techniques using toxic gases.
In order to replace highly-doped nc-Si:H, some dopant-free passivation contact structures have been explored. Compound semiconductors that feature a wide bandgap were used as the carrier transport layer (contact layer) in creating a silicon/compound heterojunction (SCH) solar cell with a c-Si wafer. Theoretically, it has the potential to mitigate energy loss associated with heavy doping, such as Auger recombination, bandgap narrowing, and free carrier absorption[6,7]. Furthermore, the range of work function in compound semiconductors is broader compared to that of doped a-Si:H or nc-Si:H. This characteristic is advantageous in designing and producing devices with higher built-in potentials in principle.
Many oxides and fluorides have been adopted as carrier transport layers. High work function transition metal oxides (TMO), such as MoOx[8,9], WOx[10], and V2Ox[11], are utilized as hole transport layers (HTL) in SCH solar cells. The ones with low work function, such as SnO2[12], TiO2[13,14] and LiF[15,16], are used as electron transport layers (ETL). These ETL and HTL compounds can be prepared using some simple techniques, such as thermal evaporation, atomic layer deposition (ALD)[17], electron beam evaporation[18], and spin-coating[19], which have the potential for low-cost compared to plasma enhance chemical vapor deposition (PECVD) commonly used for a-Si:H and nc-Si:H.
Nowadays, SHJ solar cells are mostly based on n-type c-Si wafers and commonly use a front-contact back-junction structure, where n-type nc-Si:H (nc-Si:H(n+)) layer and p-type a-Si:H (or nc-Si:H) emitter are respectively located on the front and rear side of the cell. The main reason for adopting this structure is that p-type doping of a-Si:H or nc-Si:H has lower efficiency than n-type doping, thus requiring a larger thickness, which leads to more severe parasitic absorption when located on front side of the cell. Therefore, finding new emitter materials with higher transmittance and larger work function to replace p-type a-Si:H or nc-Si:H is more urgent compared to replacing n-type nc-Si:H. MoOx with a wide bandgap and high work function has been considered the most probable alternative to p-type a-Si:H or nc-Si:H and the MoOx/c-Si(n) SCH solar cells have received more and more attention [8,9,20]. A PCE of 23.85% has been achieved by Cao and Paul Procel et al. recently[21].
However, the thermal stability of MoOx/c-Si(n) SCH solar cells is an important issue and it was found that the device performance decreases significantly when annealed above 130 oC[8, 22]. The device degradation could mostly be attributed to the formation of a SiOx barrier layer at the interface between MoOx and a-Si:H[14,23,24]. The large area metallization process (screen printing Ag grid) commonly used in SHJ solar cells is not suitable for SCH solar cells, since a solidification temperature of about 200 oC is required. The metal electrode preparation techniques reported in the literature for the SCH solar cells are mainly electroplating[8] or thermal evaporation [14]. Up to now, limited by the metallization process, all the reported MoOx/c-Si(n) SCH solar cells, no matter front-junction or back-junction, are small-area devices. Besides the metallization process, deposition of MoOx film is also an important factor hindering the preparation of large-area SCH solar cell. So far, the MoOx HTL in most of the reported high-efficiency SCH solar cells[20,21] was prepared using thermal evaporation which is not conducive for continuous deposition of uniform large-area films. Exploring large-area SCH solar cell processes is of great significance for future mass production.
In this paper, studies on large-area MoOx/c-Si(n) SCH solar cells featuring a front-contact back-junction architecture were conducted. Large-area MoOx films were deposited using hot-wire oxidation-sublimation deposition (HWOSD) technique[25]. By optimizing the device fabrication process flow including screen printing of the silver grids, the MoOx layer can avoid suffering from high-temperature process. By using ICO/Ag as the back reflector, a power conversion efficiency (PCE) of 21.59% was achieved on the champion SCH solar cell with the size of 274.15 cm2 (M6).