| 大綱 |
摘要
本研究旨在應用特殊電腦數值控制氣囊拋光機，建立光學玻璃表面品質拋光策略，研究中透由氧化鈰聚胺脂拋光墊(LP66)於石英光學玻璃表面進行拋光研究，期可達到穩定控制材料移除率目的，同時找出提升光學玻璃表面品質的拋光策略。研究發現，高精度光學玻璃於拋光過程中，特定的刀具路徑常會伴隨中頻空間頻率誤差，使鏡片表面產生定頻紋理或不規則波紋，弱化了成像品質。為此，本論文為了降低中頻空間頻率誤差，首先將拋光各種參數組進行最佳化，其拋光參數包含工具下壓量、主軸轉速、路徑間距與表面進給速率等。接著基於拋光基本公式-普雷斯頓方程式與赫茲接觸理論，將拋光主軸與光學玻璃表面接觸的拋光行為視為彈性固體的接觸，建立刀具影響函數理論模型。
本研究進一步嘗試探討與預測出穩定材料移除深度的理想拋光參數組合。研究中，經由所建立模型，將數值分析模擬結果，透過實際拋光進行數值比較，並據以修正其數學模型。研究結果發現，氣囊拋光刀具下壓量，僅會對於材料移除量有所影響，對於表面紋理的品質並不會造成太大的改變。研究結果發現，主軸轉速與進給率的配置為主要的影響因子，使其在拋光過程中穩定進行拋光作業，實現高精度的光學表面紋理品質。
本研究發現改變路徑間距參數，可有效改善定頻紋理與抑制不規則中頻空頻空間頻率誤差，其光學玻璃表面紋理可實現均方根粗度值達1.6
nm。透由所建立的刀具影響函數理論模型之模擬結果與實驗結果比較，再次證實路徑間距的設定確實對於光學玻璃表面紋理品質造成影響，其同位置所受到的重疊數值，運算結果為4.82。接著將拋光路徑間距所產生的疊加效應影響納入理論模型，再加入形貌誤差修正實驗驗證，其模擬與實驗的最大深度差值為123.6
nm，兩者結果相似約85%，亦即證明理論模型可用於模擬拋光製程的最佳參數條件，研究成果將可實現穩定控制材料移除量之目的，此項技術將可應用於光學玻璃元件拋光製程技術。
關鍵字：電腦數值控制、氣囊拋光、中頻空間頻率、石英、光學玻璃
Abstract
This study presents an optical polishing quality
strategy using automated computer numerical control
(CNC) bonnet polishing to produce a high-quality level
surface at a stable material removal rate (MRR). Owing
to the specified toolpaths integral to CNC bonnet
polishing, the resulting polished surfaces generally
have mid-spatial-frequency (MSF) errors, which can occur
as periodic surface ripples or irregular waviness. These
surface textures could degrade image quality in many
devices or cause defocusing and energy loss in
high-energy laser systems. In this study, the MSF
surface textures of fused silica are investigated by
developing a CNC bonnet polishing technique using a
cerium-oxide-filled polyurethane pad (LP66) that employs
a cellular polyurethane material designed to handle the
high flatness and surface finishing requirements of
optical glass materials. Various combinations of tool
offset, head speed, track spacing and surface feed rate
settings are studied to determine the optimal polishing
parameters for minimizing MSF errors. To further explore
and predict the ideal polishing parameters for a stable
material removal rate, a tool influence function (TIF)
mathematical model is built based on the Preston
equation of material removal rate for polishing, with
the head spindle contact and polishing behavior of
optical glass surfaces treated as contact between
elastic solids. Then, polishing is simulated using the
mathematical model, the material removal characteristics
of the simulated and experimental results are compared
and the mathematical model is modified accordingly. The
experimental results demonstrate that the head speed and
surface feed rate significantly affect the surface
texture during bonnet polishing. Although the tool
offset does not cause surface texturing, it does affect
the material removal rate. A series of optimization
experiments is conducted, ultimately leading to the
effective removal of irregular surface ripples and a
reduction in MSF errors. By optimizing the polishing
parameters, a high surface quality with extremely
accurate optical performance is achieved, along with a
root-mean-square error of 1.6 nm. The effect of TIF
superposition is considered in the polishing strategy
because the track spacing distance experiment results
showed that the superposition of the polishing path was
responsible for differences between the simulated and
experimental results. First, the superposition value is
calculated as 4.82 for the same point by superposition
of the modeled TIF. Then, the TIF model is adjusted to
include the impact of superposition, and the output of
the TIF model is recalculated to further refine the
polishing strategy. Finally, the material removal
function is modified to attain an MRR difference value
of 0.12
mm.
When the results are compared, the simulated material
removal depth accounts for up to 85% of the experimental
material removal depth. These experimental and simulated
results indicate that the TIF model can provide
preliminary predictions of the effects of various
combinations of polishing parameters on the material
removal rate and demonstrate the potential applications
of LP66 in CNC bonnet polishing for optical glass
component processing technologies.
Keywords:
computer numerical control, bonnet polishing,
mid-spatial frequency, fused silica, optical glass
|
