Band-matched transport layers and intrinsically stable perovskite solar cells for application to perovskite Si tandem cells

Abstract

Hybrid perovskite/silicon tandem solar cells offer low-cost alternatives to the commercially established silicon solar cells. In this thesis, we present the material and device optimizations of the subcells that can be used to fabricate tandem cells: Methylammonium Lead Halide (MAPbI3) based perovskite solar cells and Silicon/metal oxide type-II heterojunction based solar cells. Specifically, we focus on (a) improving the Voc of (MAPbI3) solar cells using band-matched polymer hole transporting layers (HTL), (b) improving the intrinsic stability of MAPbI3 by introducing Acetamidinium (AA) cation in the matrix, (c) studying the effect of Magnesium and Bromine substitution in MAPbI3, and (d) developing a Silicon/Cu2O type-II heterojunction solar cell that can act as a bottom cell in the proposed perovskite/silicon tandem solar cell. a) The fermi-splitting (in the absorber) and consequently the Voc of a thin film heterojunction cell depends on the fermi-level of the adjacent transport layers. The most widely used HTL: Spiro-OMeTAD has a HOMO of -5.0 eV, a 0.5 eV valance band maxima (VBM) offset with MAPbI3. In this part, we examine whether a p-type semiconducting polymer: ‘Poly-4-(5-(9,9-dihexyl-7-methyl-9H-fluoren-2-yl)thiophen-2-yl)-5,6-difluoro-7-(5-methylthiophen-2-yl)benzo[c][1,2,5]thiadiazole’ (PF-DTDFBT) having a HOMO level of -5.6 eV: exactly matched to the VBM of MAPbI3 leads to the enhancement of the Voc of the cell. The increased fermi-splitting directly contributed to the improvement of Voc from 1.04 V in standard Spiro-OMeTAD HTL devices to 1.11 V in PF-DTDFBT interlayer devices. In addition, the polymer being hydrophobic leads to an increase in device stability by reducing the seepage of moisture into the active perovskite layer slowing down its degradation. b) One of the major obstacles in the commercialization of perovskites is its instability towards moisture, optical and thermal stimulus. The MAPbI3 structure consists of the organic methyl-ammonium (MA+) cation held in the PbI64- octahedral cage by 3 hydrogen bonds, weakened by the continuous tumbling motion of the C-N spine (in MA+). We report a cation : Acetamidinium, having (i) 4 h-bonds and (ii) restricted C-N bond rotation, that binds more strongly with the PbI64- cage as compared to MA+. Acetamidinium substitution leads to improvement in device performance and stability, which retained 70% of their initial PCE in 480 hours, while the standard cells degrade to ~43 % of their initial PCE in the same time frame. c) The state-of-the-art perovskite devices use the recipe involving weak co-ordination bond between DMSO-PbI2 (solvent-solute binding) broken by an in-situ antisolvent drip (Toluene or Chlorobenzene) to control the nucleation density and grain growth rate. Although for small device areas (~ 1 cm2) this method provides dense compact films with 200-250 nm grain sizes, this method is not suited for depositing films on large area, required for making efficient solar panels. We present the controlled addition of MgCl2/MgI2 in perovskite in the precursor solution that helps in forming films comparable in quality (compactness, grain sizes and carrier lifetime) with the films deposited by anti-solvent drip method. This deposition method neither requires DMSO nor anti-solvent drip making it more commercially attractive. 10% MgCl2/MgI2 addition in MAPbI3 leads to a ~ 80 μs/120 μs increase in carrier lifetime as measured by resonance assisted microwave photoconductance supporting a 150 mV rise in Voc of the devices. d) A silicon/metal oxide heterojunction solar cell is expected to have a lower thermal budget and a higher Jsc as compared to the more popular silicon homojunction cell where a considerable number of incoming photons are lost to the free carrier absorption in the highly doped p++ emitter region. We report the optimization of a type-II n-silicon/Cu2O heterojunction based solar cell. The Si/Cu2O interface is passivated by a 1.3 nm ultra-thin silicon dioxide layer. The passivated devices showed a 200 mV increase in Voc over the unpassivated devices. Further in-situ p-type doping of Cu2O was done by incorporating nitrogen into the Cu2O crystal. This served two purposes : the Voc and FF increased due to enhanced built-in voltage and film conductivity respectively, and the hole fermi-level is now defined by the doped Cu2O layer, providing the relaxation to use lower work function transparent contacts (which otherwise would have decreased the Voc of the cell). The best devices : doped-Cu2O/Si cells with transparent ITO top contact exhibited 5.23% PCE.