直接以合作企業台灣晶技公司新石英振盪器所需的性能與尺寸為載具，本計畫擬建立石英晶片之精密分析與微加工技術，提供合作企業開發微小化頻率元件所需之核心技術。首先，基於薄石英板之基本振動理論，本計畫將建立石英晶片振盪特性模型，應用有限元素模擬軟體進行石英振盪器特性的模擬分析；在石英晶片的生產製程上，採用電化學放電方法針對石英晶片進行微鑽孔精密加工，以克服石英材料難加工之特性；並使用超快雷射進行石英基材的改質、鑽孔、切割、減薄及金屬沉積等加工製程，透過超快雷射的高瞬間功率特性，進行石英基材的高精密加工；接著使用蝕刻技術，針對雷射加工後之石英晶體進行二次加工，移除因雷射加工熱能影響材料特性的區域，以提升製程精度。為了補償模擬分析準確度，使用機器學習等人工智慧演算法，建立石英晶片之非線性參數、等效電氣參數以及Data-driven Model等模型，可作為石英性能分析使用；最後針對加工機台建立加工品質預測模型，監控製程中之加工品質，並透過基因演算法建立之製程參數優化模型，提升加工品質。本計畫預期能夠協助台灣晶技公司建立新一代微小頻率元件製造技術，提升本土國際級產業的核心競爭力。 ;Based on the performance and dimension specs of a next quartz oscillator of partner corporation, TXC Corp., this proposal aims at developing precision analyzing and micro-machining technologies for miniaturizing quartz oscillators, which are deemed as core technologies for continuing miniaturizing frequency devices. First, according to the fundamental vibration theory for a thin quartz plate, this study will develop a mathematical model for the plate’s vibration characteristics and apply a commercially available finite element software to perform simulation analysis on the plate’s oscillation features. In the regard of machining the hard-to-machine thin quartz plates, two potential approaches, ECDM (electrochemical discharge machining) and ultrafast laser machining followed by wet chemical etching, are proposed for micro-drilling, narrow-slit cutting and thickness thinning processing. The ultrafast laser pulses are also employed to irradiate quartz plate for property modification, as a pre-processing step for the follow-up chemical etching processing. This secondary treatment is able to etch out the heat-affected zone from laser machining and enhance both dimension accuracy and quartz quality. In order to improve accuracy of simulation results, machine learning algorithm is utilized to establish prediction models such as the nonlinear parameter model, equivalent electrical parameter model and data-driven model. The obtained parameters also useful to simulation analysis. Finally, we propose to develop a quality prediction model for supervising and controlling machining qualities during the processing course. In addition, based on the genetic algorithm, a process parameter optimization model will also be established for the improvement of machining qualities. The results of this work are believed to be beneficial to TXC Corp., for enhancing TXC’s long-term core competitiveness in fabricating next-generations miniature frequency devices.
