Optical Simulation of External Quantum Efficiency Spectra
Optical Simulation of External Quantum Efficiency Spectra
复制标题
外部量子效率谱的光学模拟
DOI：
10.1007/978-3-319-95138-6_3
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
R. Collins
中科院分区：
文献类型：
--
作者：
P. Koirala;Abdel;P. Aryal;P. Pradhan;Z. Huang;N. Podraza;S. Marsillac;R. Collins
介绍了非原位椭圆偏振光谱(SE)技术在薄膜太阳电池S介电性能和结构参数测量中的应用，利用最小二乘回归分析方法建立了S太阳电池的完整光学模型，实现了外部量子效率的模拟和测量的对比。通过这种比较，可以详细了解薄膜光伏技术中光学和电子收集和损耗的来源，并因此了解潜在的性能限制。给出了氢化非晶硅(a-Si：H)、碲化镉(CdTe)和铜铟镓二硒化物(CuIn1-xGaxSe2；CIGS)三种商业化薄膜技术的实例。在a-Si：H太阳电池的研究中，基于SE模型的EQE模拟与测量的EQE的比较表明，p/i/n界面附近的光生载流子产生的电损耗是由氧化层厚度大于空穴收集长度引起的。这里还通过EQE测量和模拟的比较，展示了当p/i界面处加入改善了电学质量的原晶Si：Hi层时，p/i界面附近增强的载流子收集。对于层状结构的CdS/CdTe异质结太阳电池，通过玻璃进行SE，并在综合光学性质和多层SE分析的基础上产生了EQE谱的模拟。在这种情况下，在模拟和测量的EQE之间观察到的偏差可以帮助改进单元的光学模型。应用这些方法，当具有带隙以上能量的光子没有被电池的有源层吸收时发生的光学损失可以与当这些有源层中光生的电子和空穴没有被收集时发生的电子损失区分开来。用SE研究了同时含有标准厚度和薄吸收层S的CIGS/CDS异质结太阳电池。数据分析对CIGS来说更具挑战性，因为需要提取吸收层Ga分布以获得准确的光学模型。对于有标准厚度吸收层的电池，模拟和测量的EQE之间有很好的一致性，后者是在假设100%从有源层收集的情况下进行的。然而，对于具有薄吸收层的电池，模拟和测量的EQE之间的差异可以归因于钼背接触附近的电子-空穴复合造成的损耗。当在EQE模拟中引入载波收集的概率分布时，该模拟与测量结果更接近。除了本研究中提出的SE的单点能力外，还存在提供短路电流密度预测的多层厚度和元件层特性的高分辨率映射能力。这种测绘能力是由于多通道椭偏仪的高速[<；每次测量(ψ，Δ)光谱1次S]而实现的。
Applications of ex situ spectroscopic ellipsometry (SE) are presented for determination of the parameters that describe the dielectric function and structure of thin film solar cell s. Complete optical models of solar cell s developed using least squares regression analysis of the SE data enable external quantum efficiency (EQE) simulations for comparison with measurements. Through this comparison, it becomes possible to understand in detail the origins of optical and electronic collection and losses in thin film photovoltaics technologies and, as a result, the underlying performance limitations. Examples of this approach are presented for the three commercialized thin film technologies of hydrogenated amorphous silicon (a-Si:H) , cadmium telluride (CdTe ), and copper indium-gallium diselenide (CuIn1–xGaxSe2; CIGS ). In the studies ofa-Si:H solar cells, a comparison between the EQE simulation based on the SE model and the measured EQE suggests electrical losses from photo-generated carriers near thep/iandi/ninterface s, the latter caused by ani-layer thickness greater than the hole collection length. Also demonstrated here through comparisons of EQE measurements and simulation is enhanced carrier collection near thep/iinterface when a protocrystalline Si:Hi-layer of improved electrical quality is incorporated at the interface. For a CdS/CdTe heterojunction solar cell in the superstrate configuration, SE is performed through the glass, and simulations of the EQE spectra have been generated on the basis of comprehensive optical property and multilayer analysis by SE. In this case, observed deviations between simulated and measured EQE can assist in refining the optical model of the cell. Applying these methods, the optical loss es that occur when photons with above-bandgap energies are not absorbed within the cell’s active layers can be distinguished from electronic losses that occur when electrons and holes photo-generated within these active layers are not collected. CIGS/CdS heterojunction solar cells incorporating both standard thickness and thin absorber s are also studied using SE. Data analysis is more challenging for CIGS because of the need to extract absorber layer Ga profiles for accurate optical models. For cells with standard thickness absorbers, excellent agreement is found between the simulated and measured EQE, the latter under the assumption of 100% collection from the active layers. For cells with thin absorbers, however, the difference observed between the simulated and measured EQE can be assigned to losses via electron-hole recombination near the Mo back contact. When a probability profile for carrier collection is introduced into the EQE simulation, closer agreement between this simulation and the measurement is observed. In addition to a single spot capability of SE as presented in this study, a capability also exists for high resolution mapping of multilayer thicknesses and component layer characteristics that provide short-circuit current density predictions. The mapping capability is made possible due to the high speeds [<1 s per measurement of (ψ, Δ) spectra] of multichannel ellipsometer s.