Optical Simulation-Aided Design and Engineering of Monolithic Perovskite/Silicon Tandem Solar Cells.

Monolithic perovskite/c-Si tandem solar cells have attracted enormous research attention and have achieved efficiencies above 30%. This work describes the development of monolithic tandem solar cells based on silicon heterojunction (SHJ) bottom- and perovskite top-cells and highlights light management techniques assisted by optical simulation. We first engineered (i)a-Si:H passivating layers for (100)-oriented flat c-Si surfaces and combined them with various (n)a-Si:H, (n)nc-Si:H, and (n)nc-SiOx:H interfacial layers for SHJ bottom-cells. In a symmetrical configuration, a long minority carrier lifetime of 16.9 ms was achieved when combining (i)a-Si:H bilayers with (n)nc-Si:H (extracted at the minority carrier density of 1015 cm-3). The perovskite sub-cell uses a photostable mixed-halide composition and surface passivation strategies to minimize energetic losses at charge-transport interfaces. This allows tandem efficiencies above 23% (a maximum of 24.6%) to be achieved using all three types of (n)-layers. Observations from experimentally prepared devices and optical simulations indicate that both (n)nc-SiOx:H and (n)nc-Si:H are promising for use in high-efficiency tandem solar cells. This is possible due to minimized reflection at the interfaces between the perovskite and SHJ sub-cells by optimized interference effects, demonstrating the applicability of such light management techniques to various tandem structures.

IMAL-Net: Interpretable multi-task attention learning network for invasive lung adenocarcinoma screening in CT images.

PURPOSE: Feature maps created from deep convolutional neural networks (DCNNs) have been widely used for visual explanation of DCNN-based classification tasks. However, many clinical applications such as benign-malignant classification of lung nodules normally require quantitative and objective interpretability, rather than just visualization. In this paper, we propose a novel interpretable multi-task attention learning network named IMAL-Net for early invasive adenocarcinoma screening in chest computed tomography images, which takes advantage of segmentation prior to assist interpretable classification. METHODS: Two sub-ResNets are firstly integrated together via a prior-attention mechanism for simultaneous nodule segmentation and invasiveness classification. Then, numerous radiomic features from the segmentation results are concatenated with high-level semantic features from the classification subnetwork by FC layers to achieve superior performance. Meanwhile, an end-to-end feature selection mechanism (named FSM) is designed to quantify crucial radiomic features greatly affecting the prediction of each sample, and thus it can provide clinically applicable interpretability to the prediction result. RESULTS: Nodule samples from a total of 1626 patients were collected from two grade-A hospitals for large-scale verification. Five-fold cross validation demonstrated that the proposed IMAL-Net can achieve an AUC score of 93.8% ± 1.1% and a recall score of 93.8% ± 2.8% for identification of invasive lung adenocarcinoma. CONCLUSIONS: It can be concluded that fusing semantic features and radiomic features can achieve obvious improvements in the invasiveness classification task. Moreover, by learning more fine-grained semantic features and highlighting the most important radiomics features, the proposed attention and FSM mechanisms not only can further improve the performance but also can be used for both visual explanations and objective analysis of the classification results.

High-Mobility Hydrogenated Fluorine-Doped Indium Oxide Film for Passivating Contacts c-Si Solar Cells.

Broadband transparent conductive oxide layers with high electron mobility (µe) are essential to further enhance crystalline silicon (c-Si) solar cell performances. Although metallic cation-doped In2O3 thin films with high µe (>60 cm2 V-1 s-1) have been extensively investigated, the research regarding anion doping is still under development. In particular, fluorine-doped indium oxide (IFO) shows promising optoelectrical properties; however, they have not been tested on c-Si solar cells with passivating contacts. Here, we investigate the properties of hydrogenated IFO (IFO:H) films processed at low substrate temperature and power density by varying the water vapor pressure during deposition. The optimized IFO:H shows a remarkably high µe of 87 cm2 V-1 s-1, a carrier density of 1.2 × 1020 cm-3, and resistivity of 6.2 × 10-4 Ω cm. Then, we analyzed the compositional, structural, and optoelectrical properties of the optimal IFO:H film. The high quality of the layer was confirmed by the low Urbach energy of 197 meV, compared to 444 meV obtained on the reference indium tin oxide. We implemented IFO:H into different front/back-contacted solar cells with passivating contacts processed at high and low temperatures, obtaining a significant short-circuit current gain of 1.53 mA cm-2. The best solar cell shows a conversion efficiency of 21.1%.

Improvement of the SiOx passivation layer for high-efficiency Si/PEDOT:PSS heterojunction solar cells.

Interfacial properties currently hinder the performance of Si/organic heterojunction solar cells for an alternative to high-efficiency and low-cost photovoltaics. Here, we present a simple and repeatable wet oxidation method for developing the surface passivation layer, SiOx, on the Si surface for the fabrication of high-efficiency Si/poly(3,4-ethylene-dioxythiophene):polystyrenesulfonate (PEDOT:PSS) heterojunction solar cells. The uniform and dense SiOx thin layer introduced by the oxidizing aqueous solution of H2O2 or HNO3 provided the better surface passivation and stronger wettability of the Si surface, compared to those in the native oxide case. These two types of progress helped create a lower defect density at the Si/PEDOT:PSS interface and thus a high-quality p-n junction with a lower interface recombination velocity. As a result, the HNO3-oxidized device displayed better performance with a power conversion efficiency (PCE) of 11%, representing a 28.96% enhancement from the PCE of 8.53% in the native oxide case. The effects on the performance of the Si/PEDOT:PSS hybrid solar cells of the wet oxidation treatment procedure, including the differences in surface roughness and wettability of the Si substrate, the quality and thickness of the SiOx, etc., were explored extensively. Such a simple and controllable oxidizing treatment could be an effective way to promote the interfacial properties that are an important cornerstone for more efficient Si/organic hybrid solar cells.