A statistical method to optimize the chemical etching process of Zinc Oxide thin films

Zinc Oxide (ZnO) is an attractive material for micro and nanoscale devices. Its desirable semiconductor, piezoelectric, and optical properties make it useful in applications ranging from microphones to missile warning systems to biometric sensors. This work introduces a demonstration of blending statistics and chemical etching of thin films to identify the dominant factors, and interaction between factors, and develop statistically enhanced models on etch rate and selectivity of ZnO thin films. Over other mineral acids, ammonium chloride (NH4Cl) solutions have commonly been used to wet etch microscale ZnO devices because of their controllable etch rate and near-linear behavior. Etchant concentration and temperature were found to have a significant effect on etch rate. Moreover, this is the first demonstration that has identified multifactor interactions between temperature and concentration and between temperature and agitation. A linear model was developed relating etch rate and its variance against these significant factors and multifactor interactions. An average selectivity of 73:1 was measured with none of the experimental factors having a significant effect on the selectivity. This statistical study captures the significant variance observed by other researchers. Furthermore, it enables statistically enhanced microfabrication processes for other materials.

氧化锌（Zinc Oxide, ZnO）是一种极具应用潜力的微纳尺度器件材料。其优异的半导体性能、压电性能与光学性能，使其可应用于麦克风、导弹预警系统、生物识别传感器等诸多领域。本研究通过结合统计分析与薄膜化学刻蚀技术，对关键影响因子及其交互作用进行了甄别，并针对氧化锌薄膜的刻蚀速率与刻蚀选择性构建了统计优化模型。
相较于其他无机酸，氯化铵（ammonium chloride, NH4Cl）溶液因刻蚀速率可控且具备近线性刻蚀特性，常被用于微尺度氧化锌器件的湿法刻蚀。研究发现，刻蚀液浓度与反应温度对刻蚀速率具有显著影响。此外，本研究首次探明了温度与浓度、温度与搅拌之间的多因子交互作用。基于上述关键影响因子与多因子交互作用，构建了关联刻蚀速率及其方差的线性模型。实验测得平均刻蚀选择性为73:1，且所有实验因子对刻蚀选择性均无显著影响。本统计研究复现了其他研究者所观测到的显著方差特征。此外，该研究可为其他材料的统计优化微纳加工工艺提供参考。