導熱高分子之加工性及機械性質研究

行政院國家科學委員會專題研究計畫 成果報告

導熱高分子之加工性及機械性質研究

研究成果報告(精簡版)

計 畫 類 別 ： 個別型

計 畫 編 號 ： NSC 95-2221-E-151-060-

執 行 期 間 ： 95 年 08 月 01 日至 96 年 10 月 31 日

執 行 單 位 ： 國立高雄應用科技大學模具工程系

計 畫 主 持 人 ： 吳政憲

計畫參與人員： 碩士班研究生-兼任助理：吳士杰、洪智仁、鄭竹宏

報 告 附 件 ： 出席國際會議研究心得報告及發表論文

處 理 方 式 ： 本計畫可公開查詢

中 華 民 國 97 年 01 月 29 日

附件一

行政院國家科學委員會補助專題研究計畫

▉ 成 果 報 告

□期中進度報告

導熱高分子之加工性及機械性質研究

計畫類別：▉ 個別型計畫 □ 整合型計畫

計畫編號：

NSC 95-2221-E-151-060-

執行期間：

95 年 8 月 1 日至 96 年 10 月 31 日

計畫主持人：

吳政憲

計畫參與人員：吳士杰、洪智仁、鄭竹宏

成果報告類型(依經費核定清單規定繳交)：▉精簡報告 □完整報告

本成果報告包括以下應繳交之附件：

□赴國外出差或研習心得報告一份

□赴大陸地區出差或研習心得報告一份

▉出席國際學術會議心得報告及發表之論文各一份

□國際合作研究計畫國外研究報告書一份

處理方式：除產學合作研究計畫、提升產業技術及人才培育研究計畫、

列管計畫及下列情形者外，得立即公開查詢

□涉及專利或其他智慧財產權，□一年□二年後可公開查詢

執行單位：

國立高雄應用科技大學

中華民國九十七年一月二十一日

2

行政院國家科學委員會專題研究計畫成果報告

導熱高分子之加工性及機械性質研究

Processability and Mechanical Properties of

Thermally Conductive Polymers

計畫編號：NSC 95－2221－E－151－060－

執行期限：95 年 8 月 1 日至 96 年 10 月 31 日

主持人：吳政憲

計畫參與人員：吳士杰、洪智仁、鄭竹宏

中文摘要

絕大多數的高分子材料都屬於絕熱性材料，要提高高分子材料的導熱性，可在高分子

材料中添加導熱性能良好的填充材料。高分子可改善加工條件來增加其熱傳導係數，但常

會降低其物理性質，因此導熱高分子的製造上需要兼顧導熱性質、物理性能和成本之考量。

本計畫包含導熱高分子製作與分析、模具設計與製作、微結構的成形性量測、製程參

數的影響、射出成形縫合線分析等五部分。進行結果整理歸納後，將來可應用到許多需要

高導熱性高分子產品的地方。實驗材料採用聚丙烯（PP）、加入 20%與 30%的填充材，進

行材料流動性之比較，並利用田口實驗進行規劃，分析各參數對成形性的影響。另外，我

們也進行熱壓成形實驗，探討熱壓參數和模仁粗糙度對熱壓複製性的影響。

關鍵字: 導熱高分子、機械性質、射出成形、熱壓成形、縫合線

Abstract

Most polymers are non-conductive. In order to improve their conductivities, some fillers are

added. However, it might reduce the mechanical strength. To increase the application, mechanical

behaviors of thermally conductive polymers have to be improved. This can be conducted with the

optimization of process parameters.

This project includes preparation of thermally conductive polymers, mold design and

fabrication, measurement of microstructure replication performance, study of parametric effects,

the weldline analysis.

Keywords: thermally conductive polymer, mechanical property, injection molding, hot

embossing, weld line

一、前言

有一些公司開始應用導熱高分子來做商業生產，例如 Cool Polymers 公司開始用導熱高

分子製造一些散熱器，可以用來加工出一些鋁或其他金屬，甚至陶瓷材料，均因技術或經

濟因素無法製造的電子元件。

高分子若要廣泛使用，必需先提高其加工性和機械性質，但大部分的研究都在討論此

類高分子的導熱和導電性質，鮮少有人去探討此類材料的加工性和機械性質。因此本研究

不只要探討此類材料的導熱性，更重要的是要探討其加工性和機械性質。

影響複材性質的因素有很多，包含基材和填充物的物理性質、填充物的含量、尺寸、

形狀和方向、基材和填充物的附著狀態[1]，當然這會和複材的製作條件有關。當碳被使用

為填充物時，型態從碳黑到石墨纖維，本身的拉伸強度會有很大的變化，當然也會影響到

複材的物理性質。研究也發現填充物的顆粒大小、形狀和排向，均會影響複材的機械強度

[2]。複材在加工中，如押出或射出成形時，對於深寬比大於 1 的填充物，會影響填充物的

排向，產生複材的異向性強度。成型過程剪切力也會造成填充物的斷裂，進而改變複材的

機械性質[3]。另外，成型條件也會影響基材和填充物之間的黏著性。有研究發現[4]，模溫

增加會提高複材的機械性質，原因就在於黏著性的增強。

目前用來大量生產塑膠產品的方法有三種：

z

射出成型

z

射出壓縮成型

z

熱壓成型

熱壓成型是一種製程簡單，且又可以大量複製塑膠元件的一種加工方法，以下就針對

熱壓成型製程的步驟做一簡單的說明[5]。

(一)加熱階段－加熱模具使塑膠粒或塑膠薄膜升溫達到玻璃轉化溫度以上約

10~20℃左右。

(二)保壓階段－在塑料與模仁上施加壓力使塑料成型，爲避免成型品發生過度的收縮或翹

曲，須將模板保持在一定的壓力下。

(三)冷卻階段－利用模溫機將冷卻水送入模具並且使冷卻水循環，加速成型品的冷卻。

二、實驗設備與步驟

2.1 實驗設備

微射出成形實驗採用

FANUC 50 噸全電式射出機，模溫機為暐吉水循環機 WMB-10，

最高溫度可達

120℃，烘料機使用晏邦 HD-12。

2.2 模具

2.2.1射出成形強度測試用模具

本研究所採用的模具有五個試片的模穴如圖

1，而圖 2 記錄了五個試片的尺寸。

2.2.2熱壓成形模具

對於熱壓成形模具的設計方面，考慮到成本、時間、加工等因素，所以把熱壓成形可能會

遇到的需求，考慮到同一副射出成形與射出壓縮成形的模具上，將三種製程模具整合在同

一模具上，以便進行熱壓成形實驗與其他實驗，如圖

3 所示，在此不移除套筒與導柱，因

為可藉由套筒與導柱維持公母模平整。

本文選用材料為低碳鋼，俗稱鐵，先把材料铣成長寬為20mm×20mm長條狀，再使用線

切割機加工，此線切割機使用直徑0.2mm的銅線，然後在模仁上面切7條溝槽，溝寬0.34mm，

溝深3mm，溝槽間距2.2mm，模仁厚8mm，如圖4。割完實際尺寸為長寬為20mm×20mm厚

6mm正方形模仁，上面有7條溝，溝寬0.25mm，溝深3mm，溝槽間距2.28mm，原本溝寬設

計為0.34mm是因為避免直徑0.2mm的銅線放電割完會變寬，所以留有裕度。

2.3 檢測設備

成型品幾何形狀檢測，是以光學顯微鏡(OM)和薄膜厚度量測儀量測，分別檢測成品之

外觀與成型後之尺寸。

4

2.4 射出成形強度測試

2.4.1 田口品質實驗

經由田口方法的實驗規劃，可以減少實驗中不確定參數的調整，避免直接影響到時間

和成本的浪費，且應用田口品質實驗方法可以知道哪些製程參數的影響較大，並且找出最

佳的微射出製程參數組合。其步驟如下：

(1) 選擇可控因子與水準

首先，先選擇可能會影響拉伸試片抗拉強度的控制因子(Control factor)，本實驗選擇的

控制因子為模具溫度、融膠溫度、射出速度、保壓壓力、保壓時間和冷卻時間(如表 2、3

所示)。再配合水準(Level)的分配，由於考量到塑膠材料的特性與機台的參數設定範圍，並

且要在所有試片都能夠完整充填之下設定其水準的範圍值。而實驗目標是希望有縫合線的

試片，能夠在抗拉強度方面越大越好，不要因為縫合線造成的缺陷和添加填充材，導致結

構強度變弱，所以本研究選用望大型(Larger the better)的目標函數。

(2) 選用適合的直交表

因為實驗所選擇的控制因子有六個，而水準數有三個，所以本研究選用L

(2

×3

)的直

交表。第一行是2水準行，也就是行內只能配置兩個水準。而實驗所選定的六個可控因子都

是三水準數，為避免資訊的不足影響到實驗結果，所以第一行將不配置任何因子，剩下的

七行3水準中，將第八行設為空行。而在建構直交表之前，要先觀察和計算自由度(Degree of

Freedom)，然後決定最適合的直交表，做為實驗所用。各因子的自由度是因子的水準數減1，

總自由度則是全部的實驗數減1，誤差的自由度為總自由度減每個因子的自由度之和。

(3)最佳化參數組合

最佳參數組合也就是各因子中，每個水準之平均η的最大值，亦可以利用因子回應圖，

清楚得知各因子中最大的η值。

2.4.2拉伸試驗

拉伸試驗機的設定為使用

5mm/min 的定速率，以一定的速度進行拉伸，直到試片斷裂

為止，且每一個試片皆為：夾距

15mm，標距 10mm。為了使實驗更具準確性，將在每一種

條件下，做五次拉伸試驗並將差異最大的值捨去後，取剩下三組數據求出其平均值。由於

實驗的試片尺寸小且薄，因此在做夾持時要非常的注意，須避免試片斷掉、拉到、夾持時

未夾緊或試片夾歪等因素，都有可能會影響到拉伸試驗的準確性。

2.5 熱壓成形性測試

2.5.1實際模仁粗糙度

根據田口法找出了粗糙度最佳與最差的兩組參數，並用這兩組參數按設計去線切割模

仁，所割出的模仁如圖5、圖6，再使用表面粗度儀HOMMEL TESTER T500量測，最差參

數(B3 C1 D1 E1 F3)所割出的模仁量測所得粗糙度為2.2μm、2.0μm、2.3μm、2.2μm、2.1μm，

平均:2.16μm；最佳參數(B1 C3 D2 E3 F2)所割出的模仁量測所得粗糙度為1.4μm、1.5μm、

1.4μm、1.5μm、1.5μm，平均:1.46μm。

2.5.2熱壓成形高度

根據之前田口法所得兩組參數(最佳和最差)，用來線切割兩塊模仁去熱壓，熱壓參數採

用，如表

3，使用第 1 組參數成型高度比第 2 組來的差，第 3 組壓力太大，導致快穿透塑

料，所以也不佳，所以採用第

2 組參數熱壓(材料溫度 140℃、熱壓力量 15KN、熱壓時間

180S、脫模溫度 80℃)，熱壓結果，B3 C1 D1 E1 F3 (粗糙度最大的模仁)熱壓成型高

度:1.934mm，如圖 7，B1 C3 D2 E3 F2 (粗糙度最小的模仁)熱壓成型高度:2.228mm，如圖 8。

三、結果與討論

3.1 射出成形強度測試

3.1.1 田口實驗結果

根據相關文獻的研究顯示有縫合線的抗拉強度會較無縫合線的抗拉強度弱，因此，本

實驗只針對無縫合線的試片做田口實驗，再用無縫合線的最佳製程參數來成型有縫合線的

試片。經由田口實驗分析之後，可分別得到不含填充材、含填充材

20%和 30%的微射出成

型最佳製程參數。其最佳製程參數結果如下：

(1)無縫合線的 No.1 拉伸試片

不含填充材且無縫合線的最佳製程參數為A

B

C

D

E

F

，含填充材

20%無縫合線的最

佳製程參數為A

B

C

D

E

F

，含填充材

30%無縫合線的最佳製程參數為A

B

C

D

E

F

。

(2)無縫合線的 No.3 拉伸試片

不含填充材且無縫合線的最佳製程參數為A

B

C

D

E

F

，含填充材

20%無縫合線的最

佳製程參數為A

B

C

D

E

F

，含填充材

30%無縫合線的最佳製程參數為A

B

C

D

E

F

。

3.2 無縫合線拉伸試驗分析

依據無縫合線

No.1 和 No.3 試片的最佳製參數設定，分別將 3 種不同含填充材的試片

進行拉伸試驗。分析結果如下：

(1)無縫合線的 No.1 試片

與實驗之前所預估的情況相同，亦即含填充材

30%的抗拉強度大於含填充材 20%的抗

拉強度，而含填充材

20%的抗拉強度又遠大於不含填充材的 PP 塑料。

(2)無縫合線的 No.3 試片

在

No.3 試片方面，PP 塑料、含填充材 20%和含填充材 30%的抗拉強度則是沒有像 No.1

試片有那麼明顯的差異，但含填充材

30%的抗拉強度還是略大於含填充材 20%的抗拉強

度，而含填充材

20%的抗拉強度又略大於 PP 塑料。由光學顯微鏡(OM) 的觀察中發現，會

造成此一原因有可能是因為

No.3 試片的拉伸處截面積小，因此大量的填充材集中在拉伸

處，導致用來做為基材的

PP 塑料不足，使得強度未能明顯提昇。

3.3 有縫合線拉伸試驗分析

有縫合線的拉伸試片成型條件是依照無縫合線的最佳製程參數來進行，圖

15 和 16 為

No.1 與 No.3 試片在拉伸實驗後所得之抗拉強度，實驗結果分析如下：

(1)有縫合線的 No.1 拉伸試片

PP 塑料在不含填充材時，抗拉強度略大於含填充材 20%之試片，而含填充材 20%之試

片抗拉強度又大於含填充材

30%之拉伸試片。會造成此一現象的原因極有可能是因為縫合

線的成形是由兩邊的融膠流動波前相遇結合而成，且在縫合線的邊緣會因殘餘的空氣或雜

質的存在而產生一

V 形缺口(V-notch)進而造成缺陷，因此其本身的抗拉強度就已經不如沒

有縫合線時的抗拉強度，再加上添加填充材後，兩邊的融膠波前在結合後內部填充材排列

雜亂無章而無法有效的鍵結在一起，或是因為融膠波前會產生噴泉效應(Fountain flow)，使

得波前位置的填充材配向呈現互相平行結合，導致含填充材

20%和 30%的強度反而下降。

若比較之後可以發現不含填充材的試片在沒有縫合線和有縫合線的抗拉強度約差

2MPa 左

右，而含填充材

20%和 30%的抗拉強度則與無縫合線的抗拉強度相距甚遠。

(2)有縫合線的 No.3 拉伸試片

在

No.3 有縫合線的試片之抗拉強度大小則為含填充材 20%大於含填充材 30%，而含填

充材

30%的抗拉強度又大於不含填充材的 PP 塑料。由光學顯微鏡的觀察下發現，No.3 的

試片在縫合線處看到兩邊波前融膠相接處的填充材排列較其他處為雜亂。而有縫合線且不

含填充材的

PP 抗拉強度比無縫合線的抗拉強度低很多，可能是因為 No.3 的截面積小

(0.3mm×0.3mm)，因此縫合線和 V 形缺口對抗拉強度的影響就比有縫合線 No.1 的試片更大。

6

四、結論與自評

本研究利用純PP和含填充材 20%及 30%三種材料，進行微結構成型實驗，再藉由田口

L

直交表進行實驗，並探討射出對成型性的影響，經過ANOVA分析後可得到以下的結論

與重要成果：

(1)無縫合線的 No.1 試片由田口實驗可得：

a、不含填充材的 PP 塑料成型重要因子為：保壓壓力＞保壓時間＞冷卻時間＞射出速度

＞模具溫度＞融膠溫度。

b、含填充材 20%的成型重要因子為：融膠溫度＞模具溫度＞射出速度＞保壓壓力＞保壓

時間＞冷卻時間。

c、含填充材 30%的成型重要因子為：模具溫度＞保壓時間＞冷卻時間＞射出速度＞保壓

壓力＞融膠溫度。

(2)無縫合線的 No.3 試片經由田口實驗結果可得知：

a、純 PP 塑料成型重要因子為：融膠溫度＞保壓壓力＞模具溫度＞冷卻時間＞射出速度

＞保壓時間。

b、含填充材 20%的成型重要因子為：模具溫度＞保壓壓力＞融膠溫度＞射出速度＞保壓

時間＞冷卻時間。

c、含填充材 30%的成型重要因子為：模具溫度＞射出速度＞保壓壓力＞冷卻時間＞保壓

時間＞融膠溫度。

(3)對 No.1 和 No.3 無縫合線的試片而言，添加 20%和 30%的填充材的確可以有效的增加試

片之抗拉強度，尤其是對

No.1 無縫合線試片的效果最為顯著。

在熱壓實驗部分，得到的結論包括：

(1)表面粗糙的模仁熱壓成型高度較低，表面較光滑的模仁熱壓成型高度較高。

(2)熱壓的參數控制也極為重要，溫度不可超過塑料玻璃轉化溫度太多，熱壓力量的大小也

是個重點。

(3)熱壓出的微結構形狀整體而言，不管是側看或橫看都是由兩邊往中成型越高，像圓弧型，

看其單一溝槽的成行也是由兩邊往中間成型越高，像圓弧型。

參與本計畫的研究生經由此研究過程，培養出規劃實驗、分析實驗結果與解決問題之

能力，對於目前正在撰寫畢業論文上有相當大之助益。

五、致謝

本研究承蒙國科會計畫補助，計畫編號：NSC 95-2221-E-151-060。

六、參考文獻

[1] T.S. Chow, J. Mat. Sci., 15, pp. 1873 (1980).

[2] Mahesh Gupta and Wang, K.K., “Fiber orientation and mechanical properties of

short-fiber-reinforced injection-molded composites: Simulated and experimental results＂,

Polym. Comp., 14(5), pp. 367 (1993).

[3] S.-Y. Fua, B. Lauke, E. Mäder, C.-Y. Yue, X. Hu “Tensile properties of short-glass-fiber-

and short-carbon-fiber -reinforced polypropylene composites”, Composites: Part A, 31, pp.

1117–1125 (2000).

[4] J.L. Kardos, F.S.Cheng, and T.L. Tolbert, “Tailoring the interface in graphite- reinforced

polycarbonate”, Polym. Eng. Sci., 13(6), pp. 455 (1973).

[5] K.M.B.Jansen, “Heat Transfer in Injection Moulding Systems with Insulation Layers and

Heating Elements”, Int. J. Heat Mass Transfer, Vol.38, No.2, pp.309-316 (1995).

七、圖表彙整

圖

1. 模仁

(mm) No.1 No.2 No.3 No.4 No.5

寬度

5 0.3 0.3 2 10

厚度

2 2 0.3 0.3 0.3

圖

2. 試片尺寸圖

圖

3. 熱壓模具結構圖

圖5. 最差參數所割模仁

圖6. 最佳參數所割模仁

圖7. 粗糙度最大的模仁之熱壓成品

圖8. 粗糙度最小的模仁之熱壓成品

圖

9. 有縫合線 No.1 試片抗拉強度圖

圖

10. 有縫合線 No.3 試片抗拉強度圖

8

附件二

可供推廣之研發成果資料表

□ 可申請專利 ▉ 可技術移轉

日期：97年1月21日

國科會補助計畫

計畫名稱：導熱高分子之加工性及機械性質研究

計畫主持人：吳政憲

計畫編號：NSC 95-2221-E-151-060 學門領域：高分子加工

技術/創作名稱

導熱高分子加工技術

發明人/創作人

吳政憲

中文：絕大多數的高分子材料都屬於絕熱性材料，要提高高分子材

料的導熱性，可在高分子材料中添加導熱性能良好的填充材料。高

分子可改善加工條件來增加其熱傳導係數，但常會降低其物理性

質，因而降低其應用性。

本技術包含導熱高分子製作與分析、模具設計與製作、微結構的成

形性量測、製程參數的影響、射出成形縫合線分析等五部分。應用

此技術，可製造出兼顧導熱性質、物理性能和成本之導熱高分子。

技術說明

英文：Most polymers are non-conductive. In order to improve their

conductivities, some fillers are added. However, it might reduce the

mechanical strength. To increase the application, mechanical behaviors

of thermally conductive polymers have to be improved. This can be

conducted with the optimization of process parameters.

This technology includes preparation of thermally conductive

polymers, mold design and fabrication, measurement of microstructure

replication performance, study of parametric effects, the weldline

analysis.

可利用之產業及

可開發之產品

可利用之產業:電子、電機、光電與生醫產業

可開發之產品:微機電和電腦周邊的相關產品

技術特點

導熱高分子加工技術可以降低導熱產品的製造成本，並改善成品機

械性質。

推廣及運用的價值

應用在光電及製造產業，可增進產品的品質，提高其附加價值。

※ 1.每項研發成果請填寫一式二份，一份隨成果報告送繳本會，一份送 貴單位

研發成果推廣單位（如技術移轉中心）

。

※ 2.本項研發成果若尚未申請專利，請勿揭露可申請專利之主要內容。

※

3.本表若不敷使用，請自行影印使用。

附件三

行政院國家科學委員會補助國內專家學者出席國際學術會議報告

96 年 10 月 14 日

報告人姓名

吳政憲

服務機構

及職稱

國立高雄應用科技大學模具工程系

副教授

時間

會議

地點

96 年 10 月 7 日至 11 日

韓國 大田

本會核定

補助文號

NSC 95-2221-E-151-060

會議

名稱

(中文) 第十屆 AMPT 國際研討會

(英文) The 10th International Conference on ADVANCES IN MATERIALS

AND PROCESSING TECHNOLOGIES (AMPT 2007)

發表

論文

題目

第一篇(中文) 高分子微射出成形和微熱壓成形之參數研究

(英文) Parametric Study of Injection Molding and Hot Embossing In

Polymer Microfabrication

第二篇(中文) 射出/壓縮液態複材成形之研究

(英文) The Study of Injection/Compression Liquid Composite Molding

一、參加會議經過

AMPT國際研討會起源自1993年的Ireland所舉辦，前三屆(1993,1995,1995)都由Dublin City

University在Dublin所舉辦，第四屆(1997)在Portugal舉辦，第五屆(1998)在Malaysia舉辦，第六

屆(2001)在Spain舉辦，第七屆(2003)又回到Dublin舉辦，第八屆(2004)則在Poland舉辦，去年

(2006)在美國舉辦；今年為第十屆，由韓國Korea Advanced Institute of Science and Technology

(KAIST)的National Research Laboratory for Computer Aided Materials Processing所舉辦。

AMPT研討會的主題包含各種材料(如金屬、陶瓷與複合材料等)的製程技術，其中包含

forging, micro-forming, rolling and hydro-forming, extrusion, bending, sheet metal forming,

drawing, casting, material removal processes, grinding, non-traditional machining, powder

metallurgy, micro-electro-mechanical-system, welding, laser processing, polishing, injection

molding, processing of composites, thin film technology, material property and formability, tool life

and coating, 以及電腦輔助設計與建模(computer aided design and modeling)等等。發表論文來

自世界各地共23個國家，其中包含Australia, Brazil, China, Canada, Croatia, France, Germany,

Hong Kong, India, Iran, Ireland, Italy, Japan, Malaysia, Nigeria, Poland, Portugal, Saudi Arabia,

Spain, Thailand, Taiwan, UK, 以及 USA。 共有370篇摘要被接受，其論文經過會議委員會peer

review後, 共計有209篇論文被接受到研討會進行發表。大會並邀請 POSCO的Dr. Ohjoon

Kwon 與Hyundai Motors 的Dr. Moon-Sik Kwon 做為the plenary speakers，並進行了兩場關於

“Steel-what we have to know about additionally” 與 “Hyundai-Kia Motor, Towards Future Global

Leadership”的座談會。

會議於10月7日開始報到，總共約數百位來自世界各地的學者及研究人員參加，於前兩天

的會程中共有28個Section，約200多篇口頭報告（oral）論文，會議的安排相當地充實，共有

Forging, Microforming, Meterial removal process , Casting, FEA, Powder metallurgy, Rolling and

Hydroforming, Sheet metal forming, Welding, MEMS, Drawing, Laser processing, Polishing,

Injection molding, Extrusion, Tool life and coating, Non-traditional machining, Composites,

Material property and formability, Bending and CAD, Grinding 以及Thin film 等多個會場同時

進行論文之口頭發表。第三天大會舉辦到POSCO與現代汽車進行工廠的參觀，下午並到舉辦

學校KAIST進行實驗室的參觀。最後一天共有8個Section的口頭報告（oral）論文，包含Rolling

and Hydroforming, Sheet metal forming, Tool life and coating, Non-traditional machining, Material

property and formability, Welding, Material removal process 以及Composites等多個會場同時進

行論文之口頭發表。

論 文 報 告 於 10 月 7 日 揭 開 序 幕 ， 第 一 天 論 文 發 表 的 領 域 包 括 Forging, Microforming,

Meterial removal process , Casting, FEA, Powder metallurgy, Rolling and Hydroforming, Sheet

metal forming, Welding, MEMS，本人和所指導的研究生在MEMS section有發表一篇論文，

“Parametric Study of Injection Molding and Hot Embossing In Polymer Microfabrication”，此

Section由來自台灣的鄭江河教授與韓國的C.D.Cho教授共同主持，共有3篇來自韓國與5篇來自

台灣的論文進行發表。第二天10月8日論文發表的領域包括Drawing, Laser processing, Polishing,

Injection molding, Extrusion, Tool life and coating, Non-traditional machining, Composites,

Material property and formability, Bending and CAD, Grinding 以及Thin film等section，本人和所

指導的研究生在Injection molding Section有發表一篇論文，此Section剛好由本人所主持。第三

天筆者參加大會所舉辦的現代汽車進行工廠的參觀，下午並到韓國KAIST進行實驗室的參

觀。第四天10月11日，口頭論文內容包含以下主題，分別為Rolling and Hydroforming, Sheet

metal forming, Tool life and coating, Non-traditional machining, Material property and formability,

Welding, Material removal process 以及Composites等section。

二、與會心得

筆者認為這次的研討會內容十分充實，而且與會者來自於世界各地，在參與全程會議後，

筆 者 認 為 這 次 會 議 的 最 大 貢 獻 在 於 有 很 多 最 新 的 研 究 成 果 發 表 於 下 列 領 域 ： Forging,

Microforming, Material removal process , Casting, FEA, Powder metallurgy, Rolling and

Hydroforming, Sheet metal forming, Welding, MEMS, Drawing, Laser processing, Polishing,

Injection molding, Extrusion, Tool life and coating, Non-traditional machining, Composites,

Material property and formability, Bending and CAD, Grinding 、 Thin film 、 Rolling and

Hydroforming, Sheet metal forming, Tool life and coating, Non-traditional machining, Material

property and formability, Welding, Material removal process 以及 Composites。於四天的會議中，

筆者有相當多的機會和來自於世界各地的研究學者，交換意見與討論，特別是針對目前筆者

從事的研究，得到很多好的建議和肯定。

三、建議

這是一次相當有規劃，內容充實的學術研討會，且每年都會持續舉辦，若能多鼓勵台灣

的學者參與，對於加強台灣在材料和加工技術領域的國際能見度必有相當的助益。若台灣爭

取將來的主辦權，將會激勵整個台灣學術界對此領域的研究，亦提升國家的研究形象。

四、攜回資料名稱及內容

此次本人攜回一冊發表論文摘要全集，內容包括所有論文的摘要，也攜回一片光碟，內

容包括所有論文的全文。

Please cite this article in press as: Wu, C.-H., Pan, Y.-R., The study of injection/compression liquid composite molding, J. Mater. Process. Tech. (2008), doi:10.1016/j.jmatprotec.2007.11.241

ARTICLE IN PRESS

PROTEC-11585; No. of Pages 6

j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y x x x ( 2 0 0 8 ) xxx–xxx

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / j m a t p r o t e c

The study of injection/compression liquid

composite molding

Cheng-Hsien Wu

_{, Yu-Rey Pan}

a_{Department of Mold and Die Engineering, National Kaohsiung University of Applied Sciences, Taiwan}

b_{Department of Mechanical and Automation Engineering, Da-Yeh University, Taiwan}

a r t i c l e

i n f o

Keywords:

Liquid composite molding Finite element method Permeability

Compression

a b s t r a c t

Liquid composite molding (LCM) is one of the most important processes to produce reinforced plastics. To reduce the injection pressure and improve the part quality, injec-tion/compression liquid composite molding (I/CLCM) was applied in this study. The research was conducted through modeling, numerical simulation, and experimental analysis. Control-volume finite-element method (CVFEM) has been widely used in LCM simulation. This research applied this numerical approach to simulate the I/CLCM processes. The sim-ulation results were verified through experimental analysis. An analytical solution was derived to calculate the required injection pressure and filling time. A three-point bend-ing test was also conducted to compare the mechanical properties of LCM parts and I/CLCM parts.

© 2007 Elsevier B.V. All rights reserved.

1.

Introduction

There are various manufacturing processes to produce rein-forced plastics. Liquid composite molding (LCM) is one of the most important processes (Gokce and Advani, 2004). In producing large surface area parts with low fiber per-meability, long mold filling time is needed, i.e. the cycle time is large. Moreover, the resin might solidify before the filling period ends. To prevent the short shot, increas-ing the injection pressure is a possible choice. However, the equipment cost is increased. Excessive injection pres-sure would also produce the fiber deformation or the fiber wash-out, and it affects the quality of the reinforced plastics.

The main goal of the proposed research is to provide a novel approach, injection/compression liquid composite molding (I/CLCM), which can reduce the injection pressure and improve the part quality. The research will be conducted through mod-eling, numerical simulation and experimental analysis.

∗_{Corresponding author. Tel.: +886 7381 4526/5429; fax: +886 7383 5015.}

E-mail addresses:[email protected](C.-H. Wu),[email protected](Y.-R. Pan).

Control-volume finite-element method (CVFEM) has been widely used in LCM simulation (Wu et al., 1998). This research applied this numerical approach to simulate the I/CLCM pro-cesses. Numerical simulation was verified with experimental results. An analytical solution was derived to calculate the required injection pressure and filling time. A three-point bending test was also conducted.

2.

Experimental

The experimental setup includes a mold, an injection unit, a compression unit, and a pressure/temperature measurement system. The mold was made of S45C steel and consists of three plates. The middle plate was used as a spacer and it can be used to adjust the preform thickness.

The mold was slightly opened to reduce the required pressure for the resin injection. If the mold opening gap is small, the fiber mats still occupy the entire cavity because

0924-0136/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.jmatprotec.2007.11.241

ARTICLE IN PRESS

PROTEC-11585; No. of Pages 6

2

of the compressibility. After the enough amount of resin had been injected into the cavity, the inlet was closed. Resin was injected into the cavity using a transfer pot con-nected to an air compressor. The injection velocity could be adjusted by controlling the pressure of the air compres-sor. To reduce the humidity of the air, a dehumidifier could be connected to the air compressor. The resin was com-pressed and forced to penetrate and wet the fiber preform. After the compression stage, the mold was kept at a con-stant temperature for a period time to cure the sample. A hot press (TIEN FA CN-50, Taiwan) was used as a compres-sion unit. The heat source to cure the composite was also provided by the hot press which has a temperature control system.

Two different types of sensors, thermocouples and pres-sure transducers, were used in the molding experiments. The cavity temperatures were measured by using 0.076 mm in diameter type-J thermocouples (Omega TT-J-36SLE). This lead diameter is small enough not to interfere with the flow pattern. In order to trace the pressure history of the cavity pressure, two pressure transducers (Kistler 6157A) were used. One is located at the inlet and the other is buried at appropriate position to measure the cavity pressure. Four thermocouples were also used in this experiment. The resin temperature can be monitored. The temperature history for different loca-tions can also be used to verify the flow pattern. Before the resin fills the preform, the thermocouple reveals tempera-ture of the preheated fiber mat. When the cool resin contacts the thermocouple, the temperature abruptly decreases. It indicates that the flow front arrives at the location of the thermocouple.

3.

Material characterization

3.1. Fiber permeability

Permeability measurements are needed to characterize the flow resistance of the fiber reinforcement in I/CLCM processes. The woven glass fiber mats (GC#1515, Wah Lee Industrial Co., Taiwan) was used in this study as the fiber reinforcement. The fiber density is 2.54 g/cm3_{. The mass per unit of area is}

305 g/m2_{for each fiber mat.}

The liquid used in the permeability measurement was

DOP oil (diphenyl-octyl-phthalate). Detailed descriptions of the

permeability measurement equipment and techniques have been reported (Wang et al., 1994). The measured in-plane per-meabilities of the woven glass fiber mats are based on the unidirectional flow method. The data were obtained from several sets of repeating experiments. The permeabilities at porosities of 0.57, 0.62 and 0.66 are 2.7× 10−10_{, 3.5× 10}−10_and

5.6× 10−10m2_{, respectively. The permeability versus porosity}

relationship can be modeled as,

K = c 3

(1− )2 (1)

Here, c = 2.2× 10−10_m2_.

3.2. Resin properties

A mixture of LY564 resin and hardener HY2954 (CIBA-GEIGY), with a weight ratio of 100/35, was used in this study. A simple kinetic model (Kamal and Sourour, 1973) is used in this study as,

d˛

dt = (k1+ k2˛m1)(1− ˛)m2 (2)

k1= A1e−E1/RT

k2= A2e−E2/RT

Here, m1, m2, A1and A2are constants. E1and E2are activation

energies associated with rate constants k1and k2, respectively.

R is the gas constant, T is the absolute temperature and˛ is

the resin conversion. The values of m1, m2, A1, A2, E1and E2

are 0.59, 3.1, 0.6, 6× 104_{,−2 × 10}4_{,−5 × 10}4_{, respectively.}

The viscosity change for a reactive thermosetting material can be expressed by a widely used model (Castro and Macosko, 1982),  = AeE/RT





Here, is the viscosity, Eis the activation energy, R is the gas constant, T is the temperature,˛ is the resin conversion and ˛g is the gel conversion. The values of A␮, E␮, A␯and B␯ are

6× 10−12_{Pa s, 6× 10}4_{, 0.4 and 4.2, respectively.}

4.

Process simulation

A computer simulation package considering resin flow, heat transfer, and chemical reaction based on the control-volume finite-element method has been developed. It can simulate a two-dimensional non-isothermal molding process. Detailed model formulation, numerical scheme, and solution proce-dures have been published elsewhere (Calhoun et al., 1996; Young, 1994). A brief description of the mathematical model is given below.

For thin mold cavities, the flow may be simplified to a two-dimensional problem, but heat transfer is still in the three-dimensional form because heat convection in the pla-nar direction and heat conduction in the gapwise direction are both significant. During mold filling and curing, the ther-mal physical properties may vary from location to location and may change as a function of time. Numerically, this can be handled by assigning appropriate values to the numeri-cal elements during numeri-calculation. For simplicity, we assume isotropic and constant thermal physical properties in deriving the working equations.

If the resin is assumed to be incompressible, we can get the continuity equation for LCM process,

∂ ¯u ∂x+

∂ ¯v

Please cite this article in press as: Wu, C.-H., Pan, Y.-R., The study of injection/compression liquid composite molding, J. Mater. Process. Tech. (2008), doi:10.1016/j.jmatprotec.2007.11.241

ARTICLE IN PRESS

PROTEC-11585; No. of Pages 6

j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y x x x ( 2 0 0 8 ) xxx–xxx

3

The fiber preform is considered to be a rigid porous media. Darcy’s law









⎡

⎢

⎣

⎤

⎥

⎦

can be used to represent the momentum balance without considering the inertia effect. The permeability matrix is expressed as









Here, the velocity components ¯u and ¯v are the gapwise average values in the planar directions, Sij’s are the flow coefficients

with average values of viscosity and permeability in the gap-wise direction, and h is the gapgap-wise thickness.

In this study, the simulation is based on the governing equations without considering the dispersion term. A simpli-fied energy balance equation for LCM processes can then be expressed as

cp∂T_∂t + rcprv · ∇T = k ∇2T +  H ˙m (7) where T is the temperature, is the resin viscosity, ˙m is the reaction rate as a function of phase averaged resin conversion and resin temperature, is the fiber porosity and H is the reaction heat per unit volume. The hybrid material properties are defined as cp = cprwr+ cpfwf  = rf fwr+ rwf k = krkf kfwr+ krwf (8)

where is the density, cp is the specific heat, k is the heat

conduction coefficient. The index r is for resin and f for fiber. ∂˛

∂t + v · ∇˛ =  ˙m (9)

˛ is the conversion of the resin.

An I/CLCM process consists of two stages, i.e., injection and compression. If the mold opening is small, the fiber pre-form may still occupy the entire mold cavity because of its compressibility.

1. Injection stage: If the preform has no local deformation during the injection period, the injection is the same as the filling in LCM. Eqs.(1)–(4)can be used as the governing equations.

2. Compression stage: Because the fiber preform is com-pressed, the two-dimensional continuity equation is

∂ ¯u ∂x+ ∂ ¯v ∂y= ˙h h (10)

where h is the thickness of the cavity and ˙h is the com-pression speed. Combining Eq.(10)with the momentum

Fig. 1 – Locations of thermocouples and pressure transducers.

equation—Darcy’s law, the pressure distribution can be expressed as ∂ ∂x









The pressure field, the temperature distribution, and the conversion change are solved using the control-volume finite-element method. A Chebyshev collocation spectral method is used to solve the energy balance in the gapwise direction.

5.

Unidirectional injection/compression

LCM

To study the unidirectional I/CLCM, the preform was cut slightly shorter than the cavity length. The fiber preform con-sists of 28 layers of fiber mats. As shown inFig. 1, the shorter preform and the cavity wall formed a rectangular gap next to the inlet. The injected resin filled the gap, then, flowed through the preform in the longitudinal direction together. Three ther-mocouples were inserted at the middle layer of the preform and at positions of 1 cm, 3 cm and 5 cm away from the inlet. One pressure transducer is located at the inlet and the other is 7 cm away from the inlet.

Fig. 2 – Flow front locations versus time relationships for numerical and experimental results.

ARTICLE IN PRESS

PROTEC-11585; No. of Pages 6

4

Fig. 3 – (a) Mold design and (b) locations of thermocouples and the pressure transducer.

A numerical approach was also conducted to verify the validity of the simulation program. A molding experiment was conducted and the front advancement was obtained by mon-itoring the temperature drops of thermocouples or pressure rises of pressure transducers.Fig. 2shows that the trends of numerical and the experimental values are similar, however, there is significant deviation between them. The deviation was believed to be induced by non-uniform fiber deformation near the inlet. Because the resin is injected with high pressure, the fiber preform near the inlet is pushed forward. Squeezing preform reduces the permeability and increases the required filling time.

6.

Radial injection/compression LCM

To study the radial I/CLCM processes, the mold cav-ity is designed to be a cuboid with dimensions of 13 cm× 13 cm × 1 cm. The resin is injected from the top cen-tral part of the cavity as shown inFig. 3(a). The fiber preform was cut to be 13 cm× 13 cm. The preform was punched to have a circular hole with a diameter of 0.8 cm. A pressure transducer was installed at the inlet and three thermocou-ples were buried on the preform surface as shown inFig. 3(b). Three thermocouples were installed at positions which were 1 cm, 3 cm and 5 cm away from the inlet. A molding experi-ment was conducted and the front advanceexperi-ment was obtained by monitoring the temperature drops of thermocouples. The inlet pressure and temperatures for three thermocouples were measured and shown inFig. 4.

Fig. 4 – The inlet pressure and temperature versus time.

A numerical approach was conducted to verify the numer-ical simulation. The results are shown inFig. 5. This figure shows that numerical and experimental values are in good agreement.

7.

Analysis of special I/CLCM processes

For some special LCM or I/CLCM processes, analytical solu-tions can be derived.

7.1. Unidirectional I/CLCM

For a unidirectional I/CLCM process with constant inlet pres-sure, the governing equations for injection stage can be described as follows. Continuity equation: d ¯u dx = 0 (12) Momentum equation: ¯ u = −K_d_dP_x (13)

Here, ¯u is the gapwise average velocity, K is the fiber perme-ability and is the resin viscosity. The boundary conditions are

Entrance : x = 0, P = Pinlet

Melt front : x = xf, P = 0

(14)

Please cite this article in press as: Wu, C.-H., Pan, Y.-R., The study of injection/compression liquid composite molding, J. Mater. Process. Tech. (2008), doi:10.1016/j.jmatprotec.2007.11.241

ARTICLE IN PRESS

PROTEC-11585; No. of Pages 6

j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y x x x ( 2 0 0 8 ) xxx–xxx

5

With boundary conditions Eq.(14), Eq.(13)can be solved. The pressure function of time is

P = Pinlet





Integrating Eq.(12)with time, the melt front location can be obtained as

xf=



2K

 Pinlett. (16)

Here, t is the time. During the injection stage, the preform remains the same as the initial permeability. Here, the ini-tial permeability is K0 and the required filling length before

compression is Linj. The total filling time can be described as

ttotal= L2 inj 2K0Pinlet+ hf− h0 ˙h (17) 7.2. Radial I/CLCM

For a unidirectional I/CLCM process with constant inlet pres-sure, the governing equations for injection stage can be described as follows. Continuity equation: ¯ vr r + d ¯vr dr = 0 (18) Momentum equation: ¯ vr= −K  dP dr (19)

Here ¯vrrepresents the resin gapwise averaged velocity in

the radial direction.

The boundary conditions are

Inlet region : r = rin, P = Pinlet

Melt front : r = rf, P = 0

(20)

Substituting the boundary conditions Eq.(20)into Eq.(19), one obtains P = Pinlet





Substituting the above pressure distribution into the momentum equation and setting r = rf, one obtains the front

velocity as

¯

vr,f= −K_r Pinlet fln(rin/rf).

(22)

Here, the initial permeability is K0. Integrating the front

veloc-ity with time gives the front location rf, of a function of time,

t. The total filling time is

ttotal=  2K0Pinlet









7.3. Filling time comparison between LCM and I/CLCM

A lower filling time is the main advantage of I/CLCM, especially in producing a composite with large surface area. Here, the derived analytical solution of filling time, Eqs.(17)and(23)

can be used for both LCM and I/CLCM.

In this study, the cavity is a cuboid with dimensions of 13 cm× 13 cm × 1 cm. Two cases of radial flow were studied. In case A, 28 layers of fiber mats were placed in the cavity and the compression distance was 2 mm. The required filling time is only 14 s for I/CLCM, however, the required filling time is 38 s for LCM. In case B, 36 layers of fiber mats were placed in the cavity and the compression distance was 5 mm. The required filling time is only 9 s for I/CLCM, however, the required filling time is 78 s for LCM.

Most filling time is occupied by the resin injection, which depends mainly on the preform porosity and its permeability. If the mold is partially opened in the injection stage, the higher permeability results in a shorter injection time.

For LCM, both cases have the same thickness of fiber pre-form. Case A has less layers of fibers than case B does. The resin injection will be much easier for case A than for case B. Therefore, case A has a shorter filling time than case B does. For I/CLCM, the initial porosity dominates the filling time. In case A, the preform thickness before compression is 12 mm and the porosity is 0.72. In case B, the preform thickness before compression is 15 mm and the porosity is 0.76. The latter case has a larger permeability, then the filling time is shorter.

8.

Mechanical behaviors of LCM and I/C

LCM parts

To investigate the mechanical strength of the molded parts, three-point bending tests were conducted for both LCM and I/CLCM specimens. Young’s modulus of LCM and I/CLCM parts are found to be 11 GPa and 14 GPa, respectively. The maximum bending stresses for LCM and I/CLCM parts are 280 MPa and 370 MPa, respectively. I/CLCM parts are shown to have better mechanical properties than LCM parts do.

In a LCM process, the resin has a higher pressure near the inlet. The resin penetrates deeply into the preform with increasing pressure. The preform is less impregnated with resin far from the inlet. Therefore, the resin is unevenly impregnated with resin. It reduces the whole mechanical strength of the part. In an I/CLCM process, the compression motion in the thickness direction helps to distribute the resin more uniformly and increase the mechanical properties.

9.

Conclusions

Some conclusions can be drawn from this study:

1. Injection/compression liquid composite molding is a better method to produce composites with large surface area. It can not only reduce the injection pressure but also improve the part quality.

2. Control-volume finite-element method can be used in I/CLCM simulation. Numerical simulation was verified with experimental results.

ARTICLE IN PRESS

PROTEC-11585; No. of Pages 6

6

3. Analytical solutions for injection pressure and filling time can be derived for both unidirectional flow and radial flow. It can be easily applied for parametric study.

r e f e r e n c e s

Calhoun, D.R., Yalvac, S., Wetters, D.G., Wu, C.-H., Wang, T.J., Tsai, J.S., Lee, L.J., 1996. Mold filling analysis in resin transfer molding. Polymer Composites 17 (2), 251–264.

Castro, J.M., Macosko, C.W., 1982. Studies of mold filling and curing in the reaction injection molding process. AIChE Journal 28 (2), 250–260.

Gokce, A., Advani, S.G., 2004. Simultaneous gate and vent location optimization in liquid composite molding processes.

Composites Part A: Applied Science and Manufacturing 35 (12), 1419–1432.

Kamal, M.R., Sourour, S., 1973. Kinetics and thermal

characterization of thermoset cure. Polymer Engineering and Science 13 (1), 59–64.

Wang, T.J., Wu, Cheng-Hsien, Lee, L.J., 1994. In-plane permeability measurement and analysis in liquid composite molding. Polymer Composites 15 (4), 278–288.

Wu, C.-H., Chiu, H.T., Lee, L.J., Nakamura, S., 1998. Simulation of reactive liquid composite molding. International Polymer Processing XIII (4), 389–397.

Young, W.B., 1994. Three-dimensional non-isothermal mold filling simulations in resin transfer molding. Polymer Composites 15 (2), 118–127.

Journal of Mechanical Science and Technology 21 (2007) 1338~1343

Journal of

Mechanical

Science and

Technology

Parametric study of injection molding and hot embossing in

polymer microfabrication

Cheng-Hsien Wu

and Hsien-Chang Kuo

1

Department of Mold and Die Engineering, National Kaohsiung University of Applied Sciences; Taiwan

2

Department of Mechanical Engineering, National Central University; Taiwan

(Manuscript Received May 31, 2007; Revised July 31, 2007; Accepted August 31, 2007)

---

Abstract

In recent years, plastics have begun to show great commercial potential especially in manufacturing micro-structured

parts. Injection molding and hot embossing are two major microfabrication methods. Replication accuracy was

investigated for these two methods. Polymethyl methacrylate (PMMA) was used as the polymer substrate. The

mold insert (or master) was fabricated by LIGA-type method. In this study, hot embossing was found to have

better replication accuracy for microstructure than injection molding. Experiments were also conducted to stud

y the effects of process parameters on the replication quality.

Keywords:

Injection molding; Hot embossing; Microfabrication; MEMS

---

1. Introduction

Many polymer-based microfabrication techniques

have been explored for high-volume production.

Par-ticularly in the field of bio- and chemical-MEMS,

products with microstructure are in great demand

[1-2]. In recent years, plastics have begun to show great

commercial potential especially in manufacturing

micro-structured parts. Injection molding is the most

important process to manufacture plastic parts. While

many prototype plastic micro devices are fabricated

using precision engineering methods, such as laser

machining, micro injection molding is currently being

investigated all over the world [3-4]. An important

advantage of injection molding is that with it we can

make complex geometries in one production step in

an automated process. Many micro devices, such as

watches and camera components, automotive crash,

acceleration, distance sensors, read/write heads of

hard discs, CD drives, medical sensors, pumps,

surgical instruments and telecommunications

compo-nents, have been successfully injection molded.

The injection molding process involves the

injection of a melt polymer into a mold where the

melt cools and solidifies to form a plastic part. It is

generally a three phase process including filling,

packing and cooling phases. After the cavity becomes

stable, the product is ejected from the mold.

Hot embossing is another method of replicating

polymer microstructures. A polymer substrate is

heated above its glass transition temperature. A mold

with a master is then pressed against the substrate,

allowing the pattern to be fully transferred onto the

substrate. After a certain time of holding and cooling,

the substrate is cooled below transition temperature.

The pressed substrate is removed from the mold.

To have better replication accuracy, the common

problem of knowing and accurately controlling the

state of the material during hot embossing must be

solved. When the material characteristics are well

known, the embossing conditions can be correctly

*_{Corresponding author. Tel.: +00 00 0000 0000, Fax.: +00 00 0000 0000}

Cheng-Hsien Wu and Hsien-Chang Kuo / Journal of Mechanical Science and Technology 21(2007) 1338~1343 1339

determined [5-8].

This report describes the application of hot

embossing to produce parts with microstructure. An

embossing machine, designed for microfabrication,

was used to emboss PMMA substrate. Injection

molding was also applied for comparison. Both the

injection molded part and the hot embossed part were

observed under microscope to compare the

replica-tion accuracy.

2. Experimental procedures

2.1 Material

The material used for injection molding is a high

heat injection grade of polymethyl methacrylate

(PMMA, CM-205, from Chi Mei Corp., Taiwan). The

melt flow index is 1.8 g/10 min and the bulk density

0.77 g/cm

_{. The barrel recommended injection}

tem-perature is between 210~250 and the recommended

℃

mold temperature about 50~70 . The

℃

material was

pre-dried at 90

°

C

for 4 hours using a

dehumi-difying drier before molding. The substrate used for

hot embossing is a PMMA sheet (Chi Mei Corp.)

with a thickness of 1.8 mm. The glass transition

temperature of the PMMA sheet is 110 .

℃

2.2 Part geometry:

The microstructure design is based on

micro-grooves. The microstructure consists of three types:

rectangular groove row, square groove array and

circular groove array. The microstructure size

in-cludes 0.05 mm, 0.1 mm, 0.2 mm, 0.3 mm, 0.4 mm

and 0.5 mm (Fig. 1).

2.3 Mold Insert Fabrication

Our photolithography process involves photomask

fabrication, wafer cleaning, spin coating, soft baking,

exposure, post-exposure baking, developing and hard

baking.

Fig. 1. Mold used for hot embossing.

A high-resolution transparency was used as the

mask in photolithography. A silicon wafer was used

as the substrate. The wafer surface was cleaned with a

4:1 H

SO

/H

O

mixture for 10 minutes at 120 to

℃

remove organic contaminants. The wafer surface was

rinsed using deionized (DI) water until the water

resistance was larger than 8 Ω. A 50:1 H

O/HF step

for 10 minutes at room temperature was applied to

remove chemical oxides. The wafer surface was

rinsed using deionized (DI) water again. After wafer

cleaning, the substrate was spun, blown dry using

heated nitrogen and then placed on a hot plate (120

℃

for 3 minutes) to drive off any water vapor on the

surface. This step is called dehydration baking.

To improve the adhesion of resists to the silicon

wafer, hexamethyldisilane (HMDS) is often applied.

HMDS was applied to the wafer by spinning at room

temperature. HMDS was dried by placing the wafer

on a hot plate for 2 minutes at 90 . The next step is

℃

spinning the resist on to the wafer; this should be

done immediately after the HMDS application. A

positive resist, AZ9260, was used in this study. The

resist is dispensed onto the wafer while the wafer is

spinning to produce a uniform layer on the wafer. The

spin speed of the spin coater was increased to 500

rpm with an acceleration of 500 rpm/s for 10 seconds.

For another 30 seconds, spin speed and acceleration

of the spin coater were 300rpm and 300 rpm/s,

respectively. The spin speed at this step determines

the final thickness of the resist (about 50 µm in this

study).

The next three steps in the lithography are the

prebake, wafer exposure and postexposure bake (PEB).

PEB was carried out at 90 for 1 minute. After PEB,

℃

the substrate was again gradually cooled down to

room temperature in order to minimize stress and

prevent the resist from cracking. The substrate was

immersed into the beaker containing the developer

(1:3 AZ400K/DI water) for about 5 minutes. After all

the features were developed, the photoresist was

rinsed in fresh DI water and blown dry with nitrogen.

This gave rise to an AZ9260 mold insert with positive

features.

The final step in the photolithography process is the

postbake which is designed to harden the resist and

improve its etch resistance. The temperature is set at

110 for 10 minutes. The postbaked photoresist

℃

mold is not as strong as metal in withstanding the

high pressure and temperature in injection molding.

Isotropic reactive-ion etching (RIE) was applied in

1340 Cheng-Hsien Wu and Hsien-Chang Kuo / Journal of Mechanical Science and Technology 21(2007) 1339~1344

this study. The photoresist mold acted as a pattern to

electroplate the nickel mold. A thin gold layer was

sputtered to create a conductive area for nickel

growth.

Electroplating was conducted at 50 with a pH of

℃

4, and at a low current density of 4 A/dm

_{in order to}

minimize internal stress in the nickel mold. After

electroplating, the photoresist was stripped with an

ultrasonic shaker. The nickel structure was placed in

acetone and then rinsed with DI water.

2.4 Experimental setup

Injection molding operations were conducted with

an injection molding machine (FANUC ROBOSHOT

S-2000i50A). The machine can offer a clamping force

up to 50 tons. The screw diameter is 22 mm and the

maximum injection volume is 29 cm

_{. A hot}

embo-ssing machine (Fig. 2) was designed, assembled and

calibrated. This machine mainly consists of a force

frame which delivers the embossing force via a

ballscrew connected with a servo motor. The force

frame includes upper mold plate (movable) and lower

mold plate (fixed). A mold (Fig. 1), consisting of two

mold halves, was installed between two mold plates.

The movable half was fixed on the upper mold plate

and the fixed half was fixed on the lower mold plate.

Two cartridge heaters were installed on the fixed half

which has a cavity at the center. The movable half has

cooling channels inside to keep it at a specified mold

temperature. A load cell was applied to monitor the

pressing force. The measured force was also used to

control the embossing at a specified pressing force. A

thermocouple was installed directly on the PMMA

substrate to get feedback on embossing temperature

control.

Fig. 2. Schematic illustration of the hot embossing equipment.

2.5 Molding parameters

In this study, the molding parameters can be

divided into two parts, one is for hot embossing and

the other is for injection molding. All process

para-meters were based on material manufacture's

sugge-stions. The parameters were shown in Table 1.

Table 1. Parameters of hot embossing and injection molding. Embossing temp. Embossing force Embossing time Demolding temp. Hot embossing 120 ℃ 15 kN 180 sec 80 ℃ Melt temp. Injection _velocity Mold temp. Packing _pressure Injection

molding

230 ℃ 130 mm/s 70 ℃ 10 MPa

3. Results

3.1 Comparison with hot embossed parts and

injec-tion molded parts

When the substrate is directly placed between the

two mold halves, the embossed part shows

incom-plete microstructure (Fig. 3a). To have a better

embossing quality, a rubber pad was inserted between

the upper mold half (movable one) and the substrate.

The embossed part showed a better replication quality

as shown in Fig. 3b.

(a) (b)

Fig. 3. Parts with (a) incomplete and (b) complete hot embossing.

(a) (b)

Fig. 4. (a) Rectangular protrusion row and (b) square protru-sion arrays of mold insert.

There are a lot of microstructures on the part. To

simplify the analysis, only the rectangular groove of

0.2 mm and square groove array of 0.5 mm were

Cheng-Hsien Wu and Hsien-Chang Kuo / Journal of Mechanical Science and Technology 21(2007) 1338~1343 1341

measured and discussed. Firstly, the mold insert was

observed using a microscope. The microstructures are

shown in Fig. 4. The microstructures all show a taper

angle because of isotropic reactive-ion etching (RIE).

Rectangular groove row and square groove array of

an embossed part are shown in Figs. 5 and 6.

The injection molded part was also observed for

comparison. Optimization methods were applied for

both processes. Rectangular groove row and square

groove array of an injection molded part are shown in

Figs. 7 and 8. As shown in Fig. 7, the microstructure

(a) (b)

Fig. 5. (a) Topview and (b) sideview of rectangular groove rows of embossed part.

(a) (b)

Fig. 6. (a) Topview and (b) stereoview of square groove arrays of embossed part.

(a) (b)

Fig. 7. (a) Topview and (b) sideview of rectangular groove rows of injection molded part.

(a) (b)

Fig. 8. (a) Topview and (b) stereoview of square groove arrays of injection molded part.

does not have a perfect shape due to incomplete

filling. The embossed part has a larger wall width of

square groove array than the injection molded part

has (in comparison of Fig. 6 and Fig. 8). In this study,

hot embossing seems to provide a better replication

than injection molding does.

3.2 Microstructure Measurement

Rectangular groove rows of the hot embossed part

and the injection molded part were measured. To

measure the microstructure profile, a high performance

surface profiler (XP-2, Ambios Technology, Inc.) was

used. Dimensions of a rectangular groove row are

defined as shown in Fig. 9. From Table 2, it is found

that both a hot embossed part and an injection molded

part do not have a perfect replication of insert

microstructure. The dimension deviations between the

insert and the part are around 5~15%. It shows that

these two methods can be applied to replicate the

insert microstructure.

However, the shape of a hot embossed

micro-structure has a better replication than the shape of an

injection molded microstructure does. As shown in

Fig. 10, the embossed microstructure has similar

sharp corners as the insert does. However, the corners

of an injection molded microstructure are smooth.

Fig. 9. Schematic illustration of title of the rectangular

groove rows.

Table 2. Dimensions of rectangular groove row. Locals of

microstructure Insert Embossed part Injection molded part Width of line (µm) 105.1 89.4 93.6 Height (µm) 81.40 76.3 76.0

width

line

1342 Cheng-Hsien Wu and Hsien-Chang Kuo / Journal of Mechanical Science and Technology 21(2007) 1339~1344

(a)

(b)

(c)

Fig. 10. Surface profiles of rectangular groove rows at the (a) insert, (b) embossed part and (c) injection molded part.

This phenomenon can be explained by sudden

freezing while polymer contacts the cold groove wall.

When the polymer flows into the groove, the skin

layer quickly cools down. The frozen layer slows

down and induces a smooth shape of microstructure.

As for hot embossing, the polymer is strongly pushed

into the groove by the embossing force. This action

creates a better replication of microstructure.

3.3 Parameters analysis

Embossing temperature, embossing force,

em-bossing time and demolding temperature are

inves-tigated to study their effects on replication quality of a

hot embossed part. The dimensions of rectangular

groove row on insert were described in Fig. 9. The

basic set of parameters for embossing temperature,

Table 3. Parameters set of single factor experiment. Factors Level 1 Level 2 Level 3 Embossing temp. (℃) 120 140 160 Embossing force (kN) 10 15 20 Embossing time (sec) 60 180 240 Demolding temp. (℃) 70 80 90 (a) (c) (b) (d)

Fig. 11. Results of single factor analysis by changing (a) embossing temperature, (b) embossing force, (c) embossing time and (d) demolding temperature.

embossing force, embossing time and demolding

temperature are 160℃, 10kN, 240 seconds and 70℃,

respectively. In this study, parametric analysis is

carried out by changing one factor at a time keeping

the others constant as shown in Table 3. The depth of

groove row on insert is 81.40 µm. The replication

heights are shown in Fig. 11.

The results show that embossing force has little

influence on the replication quality in this study. It

can be explained that embossing force of 10 kN is

enough for these experiments. Higher embossing

force does not have better replication. An appropriate

embossing temperature gives a best replication.

Optimization is necessary for setting a best

em-bossing temperature. However, the polymer can be

further pushed into the micro-groove by increasing

embossing time. Higher demolding temperature

induces more shrinkage after ejecting embossed part.

Therefore, the replication height decreases with

demolding temperature.