No abstract

Electron-beam-induced damages in methylammonium lead triiodide (MAPbI3) perovskite thin films were studied by cathodo­luminescence (CL) spectroscopy. We find that high-energy electron beams can significantly alter perovskite properties through two distinct mechanisms: (1) defect formation caused by irradiation damage and (2) phase transformation induced by electron-beam heating. The former mechanism causes quenching and broadening of the excitonic peaks in CL spectra, whereas the latter results in new peaks with higher emission photon energy. The electron-beam damage strongly depends on the electron-beam irradiation conditions. Although CL is a powerful technique for investigating the electronic properties of perovskite materials, irradiation conditions should be carefully controlled to avoid any significant beam damage. In general, reducing acceleration voltage and probing current, coupled with low-temperature cooling, is more favorable for CL characterization and potentially for other scanning electron-beam-based techniques as well. We have also shown that the stability of perovskite materials under electron-beam irradiation can be improved by reducing defects in the original thin films. In addition, we investigated effects of electron-beam irradiation on formamidinium lead triiodide (FAPbI3) and CsPbI3 thin films. FAPbI3 shows similar behavior as MAPbI3, whereas CsPbI3 displays higher resistance to electron-beam damage than its organic–inorganic hybrid counterparts. Using CsPbI3 as a model material, we observed nonuniform luminescence in different grains of perovskite thin films. We also discovered that black-to-yellow phase transformation of CsPbI3 tends to start from the junctions at grain boundaries.

Thin-film solar cells based on polycrystalline Cu(In,Ga)Se2 (CIGS) and CdTe photovoltaic semiconductors have reached remarkable laboratory efficiencies. It is surprising that these thin-film polycrystalline solar cells can reach such high efficiencies despite containing a high density of grain boundaries (GBs), which would seem likely to be nonradiative recombination centers for photo-generated carriers. In this paper, we review our atomistic theoretical understanding of the physics of grain boundaries in CIGS and CdTe absorbers. We show that intrinsic GBs with dislocation cores exhibit deep gap states in both CIGS and CdTe. However, in each solar cell device, the GBs can be chemically modified to improve their photovoltaic properties. In CIGS cells, GBs are found to be Cu-rich and contain O impurities. Density-functional theory calculations reveal that such chemical changes within GBs can remove most of the unwanted gap states. In CdTe cells, GBs are found to contain a high concentration of Cl atoms. Cl atoms donate electrons, creating n-type GBs between p-type CdTe grains, forming local p-n-p junctions along GBs. This leads to enhanced current collections. Therefore, chemical modification of GBs allows for high efficiency polycrystalline CIGS and CdTe thin-film solar cells.

We report on the microscopic structure of the SiOx layer and the transport mechanism in polycrystalline Si (poly-Si) passivated contacts, which enable high-efficiency crystalline Si (c-Si) solar cells. Using electron beam induced current (EBIC) measurements, we accurately map nanoscale conduction-enabling pinholes in 2.2 nm thick SiOx layers in a poly-Si/SiOx/c-Si stack. These conduction enabling pinholes appear as bright spots in EBIC maps due to carrier transport and collection limitations introduced by the insulating 2.2 nm SiOx layer. Performing high-resolution transmission electron microscopy at a bright spot identified with EBIC reveals that conduction pinholes in SiOx can be regions of thin tunneling SiOx rather than a geometric pinhole. Additionally, selectively etching the underlying poly-Si layer in contacts with 1.5 and 2.2 nm thick SiOx layers using tetramethylammonium hydroxide results in pinhole-like etch features in both contacts. However, EBIC measurements for a contact with a thinner, 1.5 nm SiOx layer do not reveal pinholes, which is consistent with uniform tunneling transport through the 1.5 nm SiOx layer. Finally, we theoretically show that reducing the metal to the c-Si contact size from microns, like in the p-type passivated emitter rear contact, to tens of nanometers, like in poly-Si contacts, allows lowering of the unpassivated contact area by several orders of magnitude, thus resulting in excellent passivation, as has been demonstrated for these contacts.

No abstract