VARTM製程之灌注監控及IC載板之翹曲模擬

Infusion Monitoring in VARTM Process and Warpage Simulation of IC Substrate

徐宏斌(Hung-Pin Hsu)

指導教授 ： 姚遠

國立清華大學/工學院/化學工程學系/碩士(2021年)

摘要

複合材料由於質輕及強度高，抗腐蝕性佳，近年來廣泛地受到各行各業的青睞。但由於高材料成本與產品價值，減少預產品中的缺陷是業界的首要任務。但是近年來複合材料的檢測技術只能檢測那些已經完成交聯反應的預產品，無法提前就檢測出缺陷，這樣無形中增加許多材料成本。本文提供一個能在真空輔助樹脂轉注成型製程中實時檢測缺陷的方法，當樹脂灌入模腔中，往往會同時伴隨著交聯反應的進行，此時樹脂會放熱，這樣熱影像儀就被動地收到溫度的變化圖，也就是說若有缺陷存在於產品中，其表面就會因熱傳速率不均一而出現溫度差，同時在熱影像上就會呈現缺陷的位置與大小。為了確認缺陷是存在肉眼不可見之產品內部，拍攝時會同時採用熱影像儀與攝影機兩種設備，之後再用一系列的影像處(Image processing)理方法來解決像是沒對齊、噪點(Noise)等問題，接著將缺陷面積與時間作圖，若未來需要，可利用將統計過程控制(Statistical Process Control)來實現自動化。最後為了驗證實驗正確性，使用Moldex 3D軟體照著實驗參數做一次模擬，得到相同的結果。
在本文的第二部分中，基於現今電子裝置的製作重心在於更輕薄，一種無核芯(coreless)的基版技術被研發及發展、稱為嵌入式基板(Embedded Trace Substrate, ETS)，並廣泛應用於各式電子產品。然而，此種設計因為不同材料間的相異的物理性質會有嚴重的缺陷，例如在回焊過程(Reflow soldering)中，會有一段加溫降溫過程，此時銅和其他材質就會因為不同的熱膨脹係數而造成翹曲，進而影響後續封裝製程(package processing)。近年來，有限元素分析法(Finite Element Analysis, FEA)是一種受歡迎且有效的方法幫助研究人員預測板彎翹曲和機械性質的研究。生產者可以應用有限元素方法去模擬基板的改良設計，以及能提供一個特定的翹曲數值滿足後續開發所需。儘管如此，對於高精度的模擬需耗費龐大的計算成本，因此，如果必須多次執行模擬研究，則模擬研究將成為漫長而艱鉅的任務。在此篇論文中，透過ANSYS軟體提出了三種等效方法來降低有限元素法分析的時間與難度，並設計出三種不同的案例來預測板彎翹曲。有別於傳統透過不斷try and error或是實驗設計法(DOE)，需要透過大量實驗來得到較佳的翹曲結果，本文提出的方法結果顯示與實際實驗相符且精準。若將來能將此方法運用於工業上，使得在封裝製程的預處理步驟中，能直接藉由參數的調整進行板彎翹曲的優化，減少不必要時間及材料成本，並加速產品上市時間。

關鍵字

複合材料 ； 真空輔助樹脂轉注成形 ； 熱影像 ； 嵌入式基板 ； 回焊製程 ； 有限元素模型 ； 翹曲模擬

並列摘要

Fabric composite materials are widely used in various applications, because of the characteristics of lightweight and strong strength. Due to high raw material cost and final product value, reducing defects in products is critically important. However, it is difficult to conduct real-time monitoring during the manufacturing process before completely curing. This study provides a method for online defect detection in vacuum assist resin transfer molding (VARTM), a popular process for manufacturing large composite parts in many industries. When thermosetting polymer resins, such as epoxy, are used as matrix to bind the reinforcement material, the filling process of VARTM is often accompanied with curing.
The temperature changes occurred in VARTM can be recorded using an infrared camera. When a subsurface void occurred, the surface temperature captured by the infrared camera often behaves different from the surroundings. Therefore, the thermograms recorded by the infrared camera can disclose voids inside the product during filling. In addition, a visual camera is also used to record the filling process. The thermal and visual images captured at the same time point are aligned by image registration[1]. After binarization, these two images are compared by subtracting one from the other. In doing this, the defective region is highlighted. By monitoring the changes in the area of highlighted region, the void can be detected with a statistical control chart. The proposed method was applied to a simulated VARTM process and a real experiment. The results of both illustrates the effectiveness of this method.
In the second part of this thesis, the IC substrate warpage problem during the reflow process problem is studied. As electronic devices become lighter and smaller, a coreless substrate technology, called Embedded Trace Substrate (ETS) was developed to meet the market requirement. However, this design often causes severe warpage due to the large difference in materials parameters of the build-up substrates. Recently, finite element analysis (FEA) is a popular and effective method used for substrate warpage predictions and mechanical studies. Nevertheless, the computational resources needed for high-fidelity simulation are extremely expensive and time-consuming. Hence the simulation study becomes a long and tedious task if it has to be performed many times, e.g., sensitivity analysis and warpage optimization.
In this paper, three different equivalent strategies are used for decreasing computational costs in terms of time and difficulty of FEA, and implement the aforementioned strategies on three different cases for simulating substrate warpages. Compared with the real geometry mapping simulation, these cases significantly decrease the demands of computational resources. Besides, we gain accurate warpage prediction results as validated by real substrate experiments. In conclusion, the results show that the methodology for substrate simulation in this paper is practical, effective, and computationally feasible.

並列關鍵字

composite materials ； real time monitoring ； VARTM ； filling simulation ； ETS ； reflow process ； warpage simulation

參考文獻

[1] C. Nash, P. Karve, D. Adams, S. Mahadevan, and G. Thorne, "Real-time cure monitoring of fiber-reinforced polymer composites using infrared thermography and recursive Bayesian filtering," Composites Part B: Engineering, vol. 198, p. 108241, 2020.

[2] L. Xueshu and C. Fei, "A review of void formation and its effects on the mechanical performance of carbon fiber reinforced plastic," Engineering Transactions, vol. 64, no. 1, pp. 33–51, 2016.

[3] H. Kaiser and V. Karbhari, "Non-destructive testing techniques for FRP rehabilitated concrete. I: a critical review," International Journal of Materials and Product Technology, vol. 21, no. 5, pp. 349-384, 2004.

[4] R. Prakash, "Non-destructive testing of composites," Composites, vol. 11, no. 4, pp. 217-224, 1980.

[5] R. Shrestha, S. Sfarra, S. Ridolfi, G. Gargiulo, and W. Kim, "A numerical–thermal–thermographic NDT evaluation of an ancient marquetry integrated with X-ray and XRF surveys," Journal of Thermal Analysis and Calorimetry, pp. 1-15, 2021.