多晶硅太阳电池表面织构工艺优化

Optimization of Polysilicon Solar Cell Texture-making Technology

在太阳电池制备工艺中,多晶硅表面常需要进行绒面处理,在各种不同的工艺中,各向同性腐蚀法由于成本低且易于操作而适宜于大规模生产,已成为当前研究的重点。但不同配比的试剂在不同作用时间及工艺下会得到不同的效果,如何科学地评价这些差异,进而得到最佳的工艺参数用于指导生产实践,具有重要的意义。由于各参数间复杂的非线性耦合关系,使得参数的精确确定成为难题。在充分理解遗传算法(Genetic algorithm,GA)作用机理的基础上,以人工神经网络(Artificial neural network,ANN)为辅助手段,以腐蚀速度和光反射率为评价标准,从理论上提出了最佳工艺参数的确定方法,大量多晶硅表面制绒试验已经证明,此方法具有快捷、准确的优点。

Isotropic acidic etching method is primary one in polysilicon texture-making methods for its low cost and convenient operation and has become a hot spot of present research. Different reagent ratio and different reaction time lead to different results and coupling relations between variables is complex, so finding out suitable processing parameters is important and meaningful. The etching speed and lights reflex ration is used as judge criterion, helped by genetic algorithm（GA） and artificial neural network（ANN）, the besl treating parameters can be deduced out in theo ry, tests have manifest its effectiveness and accuracy.