We report the room-temperature photoluminescence spectra of nanocrystalline TiO 2 in the anatase and rutile phases and in mixed-phase samples obtained commercially (Degussa P25) and by thermal treatment of nanocrystalline anatase. The photoluminescence spectrum of anatase spans a broad range of visible wavelengths, while the much more intense rutile emission is found in the near-infrared. Photoluminescence spectra as a function of contacting fluid provide insight into the microscopic nature of the luminescence, the basis for its breadth, and the influence of solvent on inter-and intraparticle electron transfer. Anatase photoluminescence results from at least two spatially isolated trap-state distributions, one of which is absent or quenched in P25 and in the presence of hole scavengers. TiO 2 nanocrystalline films containing a small amount of rutile show solvent-dependent relative intensities of the anatase and rutile photoluminescence that reveal carrier transport between the two phases. Photoluminescence spectroscopy is shown to be a useful approach for determining the energetic distribution of midband gap states.

The defect photoluminescence from TiO 2 nanoparticles in the anatase phase is reported for nanosheets which expose predominantly (001) surfaces and compared to that from conventional anatase nanoparticles which expose mostly (101) surfaces. Also reported is the weak defect photoluminescence of TiO 2 nanotubes, which we find using electron backscattered diffraction to consist of walls which expose ( 110) and ( 100) facets. The nanotubes exhibit photoluminescence that is blue-shifted and much weaker than that from conventional TiO 2 nanoparticles. Despite the preponderance of (001) surfaces in the nanosheet samples, they exhibit photoluminescence similar to that of conventional nanoparticles. We assign the broad visible photoluminescence of anatase nanoparticles to two overlapping distributions: hole trap emission associated with oxygen vacancies on (101) exposed surfaces, which peaks in the green, and a broader emission extending into the red which results from electron traps on undercoordinated titanium atoms, which are prevalent on (001) facets. The results of this study suggest how morphology of TiO 2 nanoparticles could be optimized to control the distribution and activity of surface traps. Our results also shed light on the mechanism by which the TiCl 4 surface treatment heals traps on anatase and mixed-phase TiO 2 films and reveals distinct differences in the trap-state distributions of TiO 2 nanoparticles and nanotubes. The molecular basis for electron and hole traps and their spatial separation on different facets is discussed.

Using X-ray and ultraviolet photoelectron spectroscopy, the surface band positions of solution-processed CH3NH3PbI3 perovskite thin films deposited on an insulating substrate (Al2O3), various n-type (TiO2, ZrO2, ZnO, and F:SnO2 (FTO)) substrates, and various p-type (PEDOT:PSS, NiO, and Cu2O) substrates are studied. Many-body GW calculations of the valence band density of states, with spin-orbit interactions included, show a clear correspondence with our experimental spectra and are used to confirm our assignment of the valence band maximum. These surface-sensitive photoelectron spectroscopy measurements result in shifting of the CH3NH3PbI3 valence band position relative to the Fermi energy as a function of substrate type, where the valence band to Fermi energy difference reflects the substrate type (insulating-, n-, or p-type). Specifically, the insulating- and n-type substrates increase the CH3NH3PbI3 valence band to Fermi energy difference to the extent of pinning the conduction band to the Fermi level; whereas, the p-type substrates decrease the valence band to Fermi energy difference. This observation implies that the substrate's properties enable control over the band alignment of CH3NH3PbI3 perovskite thin-film devices, potentially allowing for new device architectures as well as more efficient devices.

We report on the effect of humidity on the structural, optical, and electrical properties of formamidinium lead halide perovskite (FAPbI 3 ; prepared by a solvent engineering method) and the device characteristics of planar FAPbI 3 solar cells. The relative humidity strongly affects the perovskite film morphology, which changes from a uniform, fully covered FAPbI 3 film at low relative humidity (e.g., ~2%) to an inhomogeneous film consisting of many voids (or pinholes) at high humidity (30%-40%). This morphological deterioration with increasing humidity is also accompanied by a reduction of the film crystallinity, decay of optical property, and shorter carrier lifetime. The device based on a planar FAPbI 3 film shows the best conversion efficiency of 16.6% (with the stabilized output efficiency of 16.4%) at a low humidity (~2%). Higher humidity leads to lower device performance, mainly due to the loss of open-circuit voltage and fill factor, which is consistent with the decrease of recombination resistance.

For nanomaterials, surface chemistry can dictate fundamental material properties, including charge-carrier lifetimes, doping levels, and electrical mobilities. In devices, surface defects are usually the key limiting factor for performance, particularly in solar-energy applications. Here, we develop a strategy to uniformly and selectively passivate defect sites in semiconductor nanomaterials using a vapor-phase process termed targeted atomic deposition (TAD). Because defects often consist of atomic vacancies and dangling bonds with heightened reactivity, we observe-for the widely used p-type cathode nickel oxide-that a volatile precursor such as trimethylaluminum can undergo a kinetically limited selective reaction with these sites. The TAD process eliminates all measurable defects in NiO, leading to a nearly 3-fold improvement in the performance of dye-sensitized solar cells. Our results suggest that TAD could be implemented with a range of vapor-phase precursors and be developed into a general strategy to passivate defects in zero-, one-, and two-dimensional nanomaterials.