Thus far, no reports have been made on the correlation between photovoltaics and electroluminescence in light-emitting diodes. With machine learning assistance, we here illustrate the relationship between photovoltaics and electroluminescence of quantum dot light-emitting diodes (QLEDs) by analyzing the measurements of over 200 devices, including J–V–L, photovoltaics, and time-resolved electroluminescence (TREL) test. By applying a decision tree analysis of 17 extracted features of photovoltaics test and TREL curves, we clarify the key features of open-circuit voltage (V_oc) and short-circuit current (I_sc) under varied illuminated light intensities that correlate with maximum external quantum efficiency (EQE_max) of QLED devices. These photovoltaic features are discussed from the perspective of carrier injection and recombination. In addition, the exciton formation rate (r) derived from TREL curves also affects the EQE_max. The machine learning assisted methodology is also able to predict EQE_max of the QLED with a coefficient of determination of 0.78 with an artificial neural network model.

中文翻译：

---

关联光伏和量子点发光二极管电致发光的机器学习

迄今为止，还没有关于光伏发电和发光二极管电致发光之间的相关性的报道。在机器学习的帮助下，我们通过分析 200 多种器件（包括 J-V-L、光伏和时间分辨电致发光 (TREL)）的测量结果来说明光伏和量子点发光二极管 (QLED) 电致发光之间的关系测试。通过对光伏测试和 TREL 曲线的 17 个提取特征进行决策树分析，我们阐明了与最大外部光强度相关的不同照明光强度下的开路电压 ( V _oc ) 和短路电流 ( I _sc ) 的关键特征QLED 器件的量子效率 ( EQE _max )。这些光伏特性是从载流子注入和复合的角度讨论的。此外，由TREL曲线得出的激子形成速率( r )也会影响EQE _max。机器学习辅助方法还能够通过人工神经网络模型预测QLED 的EQE_最大值，确定系数为 0.78。