Physical Vapor Deposition of Yb-Doped Cesium Lead Halide Perovskites

Recently, ytterbium-doped CsPb(Cl_1-xBr_x)₃ (x&lt;0.35) perovskite showed efficient quantum cutting, a process wherein photons absorbed at high energies (<i>e.g.</i>, &gt; 2.5 eV) generate two photons with energies (~ 1.25 eV) close to the silicon bandgap energy (1.1 eV). The mechanism for this quantum cutting process is not well understood, but the leading theories suggest that Pb vacancy defects in the perovskite efficiently trap excitons and transfer energy to two nearby Yb³⁺ ions, which subsequently emit from the Yb³⁺ ²F_5/2→ ²F_7/2 transition.¹ A thin layer of this material on a silicon solar cell can convert blue photons to two near-infrared (NIR) photons, decreasing energy losses due to the relaxation of the high-energy charge carriers to the band edges. In this way, silicon solar cell efficiencies can surpass the Quessier limit. The majority of reports on ytterbium-doped perovskites rely on nanocrystal or solution-based synthesis to make the materials.^2,3<br/><br/>We are exploring physical vapor deposition, a solvent-free technique, to synthesize halide-perovskite thin films. Specifically, we deposited CsPb(Cl_1-xBr_x)₃ and Yb-doped CsPb(Cl_1-xBr_x)₃ films by co-evaporating CsCl, PbCl₂, PbBr₂ and YbCl₃ and controlling the flux of each using quartz crystal microbalances. Films were characterized using x-ray diffraction, scanning electron microscopy, optical absorbance, x-ray photoelectron spectroscopy, photoluminescence, and photoluminescence quantum yield measurements. Specifically, we investigated the effects of deposition temperature, Yb-concentration, stoichiometry, post-deposition annealing, and annealing environment on the structure, morphology, and photoluminescence quantum yield of Yb-doped CsPbCl₃ films formed by PVD and using a combinatorial high throughput approach.<br/><br/>Yb-doped CsPbCl₃ films annealed at 350 ^oC for 2 hours emitted NIR photoluminescence at ~990 nm (Yb³⁺ ²F_5/2→ ²F_7/2 transition, 1.25 eV) with quantum yields (PLQY) exceeding 70%, when excited with photons with energies above the CsPbCl₃ bandgap (<i>e.g.</i>, &gt; 3.0 eV). PLQY was maximized at the lowest Yb concentrations and decreased as Yb concentration increased from 2% to 10%. This indicates that not all Yb are optically active. The PLQY is especially sensitive to the annealing environment. For some film compositions, identical films annealed in the air have higher PLQY than films annealed in an N₂-filled glovebox, and films annealed in an N₂-filled glovebox followed by annealing in the air have even higher PLQY. We hypothesize that grain growth in the glove box followed by oxygen passivation of remaining defects in the air is responsible for the high PLQYs in these films. This PLQY trend with annealing environment, however, is sensitive to the Cs and Yb concentrations as well as the halide composition. For instance, Yb-doped orthorhombic CsPb(Cl_1-xBr_x)₃ films with x ~ 0.35 behave differently than Yb-doped cubic CsPbCl₃ films, with annealing in N₂-filled glovebox having the highest PLQY. SEM images show that annealing in air leads to smaller grains. Grain growth is also inhibited when as little as 2% Yb (relative to lead) is incorporated in the film. Surface analysis of Yb-doped CsPbCl₃ films with XPS shows that Yb is detected on the surface of films annealed in a glovebox but missing from the surface of films annealed only in the air.<br/> <br/>1. C. S. Erickson, M. J. Crane, T. J. Milstein, D. R. Gamelin, <i>J. Phys. Chem. C</i>, 2019, <b>123</b>, 12474-12484.<br/>2. D. M. Kroupa, J. Y. Roh, T. J. Milstein, S. E. Creutz, D. R. Gamelin, <i>ACS Energy Lett.</i>, 2018, <b>3</b>, 2390-2395.<br/>3. T. J. Milstein, D. M. Kroupa, D. R. Gamelin, <i>Nano Lett.</i>, 2018, <b>18</b>, 3792-3799.