201208071 六、發明說明： 【發明所屬之技術領域】 本發明係關於一處理晶圓之方法。本發明亦關於一組合體， 其包含一晶圓及一支撐結構體，其中該支撐結構體包含至少一形 狀記憶聚合物薄膜。 【先前技術】 使半導體晶圓變薄可能是令人期待的’譬如因為較薄之晶圓 可以降低熱阻抗，並使可靠度增加。使晶圓變薄可以減少來自於 熱效應之應力，因而使可靠度增加。厚的晶圓亦會對封裝之尺寸 產生不良之影響。於有些情況中，厚的晶圓會對效能產生不良之 影響。因此，基於各種理由，使半導體晶圓變薄是令人期待的。 通常，使晶圓變薄之做法係將晶圓表面接觸一堅硬且平坦’ 含有液體泥衆之旋轉水平轉盤。該泥漿液可含有具化學蚀刻劑， 如氨水、氟化物或其等之組合物等之研磨媒劑。晶圓然後保持與 該媒劑接觸，直到一定量之基質被移除後而達到所要之厚度為止。 已知黏附性讀料在進行晶圓薄化之製程中可被使用作為拖 住晶圓，例如經由凸塊技術處理過之晶圓，之支撐結構體。 關於此方面’美國專利案US 7,226,8i2B2揭露一處理晶圓之 方法’於其帽-作為犧牲之聚合物以噴賤之方式覆被在晶圓之 凸塊面上以形成-薄層。，然後利用—黏附層將該晶圓之凸塊面透 過該作為犧牲之聚合物貼附在-晶圓之支撐結構體上，以支持晶 圓於背面之加工處理1晶圓經過薄化之後，將晶圓暴露於高溫 之下以便_作為之聚合物㈣溫之方式分解成氣體。該作 為犧牲之聚合物之分解使__對晶圓凸細之歸性減少， 結果當支撐性之基質與晶圓分開時，雜層即與該支撐性之基質 -同和該晶®之凸塊面分1此核需要關剪力、楔人力及/或 201208071 剝離力以將該支撐結構和變薄之晶圓表面分開。此外，該方法無 法重複使用該支撐結構體。 美國專利案US2006/0046433 A1教導一處理晶圓之方法，其 包括之步驟有：在晶圓上製造多個突出體，並將所述之突出體與 一固定物相嚙合，並將該晶圓經歷研磨處理以使該晶圓變薄。該 固定物具有多個產生干擾之突出體，其等適合於由待薄化處理晶 圓未經薄化之表面向外延伸之表面特性。於有些實施例中，可使 用形狀記憶材料中之開口，其等於加熱時使該等凸塊更牢固地嚙 合在待薄化處理之晶圓上。 因為晶圓被製作地愈來愈薄，故晶圓損壞之嚴重性可能會急 遽增加。當晶圓之厚度低於ιοομιη時，晶圓便容易出現破裂、表 面被磨光及表面不規則性。經由凸塊技術處理過之晶圓可能會有 另外關於凸塊之間之區域上出現表面呈現凹痕之問題。尤其是對 於非常薄之晶圓，該等晶圓容易於從薄化用固定物上移除時受到 損壞。 儘管科技之現況，但是仍令人期待提供處理晶圓更佳之方 法’尤其是提供使晶圓變薄更佳之方法，於其中將晶圓從支撐結 構體上移除所需要之機械力是制，而且無紐續之清潔步驟將 支撐結構體之殘留物從晶圓表面上移除。 【發明内容】 本發明提供-處理晶1]之方法，諸如使晶圓變薄之方法，於 其中二晶圓’於其表面上具有多個突出體，被—支撐結構體所托 住以降低該晶圓於機械性之處理過程，如研磨中所產生 支撐結猶包含至少-具有―内㈣及-核面之形狀記憶 ί中該形狀記憶聚合物之外表面形成該支撐結構體 (外部表面。於_加工之後，可依—無應力難中應用很低之 4 201208071 剝離力將所述之支撐結構體大致完全地從晶圓表面上移除。該支 撐結構體因此可重複被使用，此特性使廢物及整體成本減少。 由其最廣泛之意義而言’本發明提供一處理晶圓之方法，其 所包含之步驟有： a) 提供一晶圓及一支撐結構體， 其中該晶圓於其表面上具有多個突出體， 且於其中該支撐結構體包含至少一具有一内表面及一外表面 之形狀記憶聚合物薄膜，其中該形狀記憶聚合物之外表面形成 該支撐結構體之外部表面’且所述之外部表面所呈現之平坦度 指數(FI)低於0.02，且所述之形狀記憶聚合物薄膜所具有之平 均薄膜厚度為10至1200μηι ; b) 於0.1至100兆帕(MPa)之壓力下，較佳者在丨至丨⑻MPa之壓力 下，及於一高於至少一形狀記憶聚合物之玻璃轉化溫度(Tg)之 溫度下使該等突出體接觸該支撐結構體之外部表面以形成一 組合體； c) 於0.1至100 MPa之壓力下，較佳者在1至1〇〇 MPa之壓力下，將 該形成之組合體冷卻至高於2〇°c並低於至少一形狀記憶聚合 物之玻璃轉化溫度(Tg)之溫度。 本發明之另一面向係為一組合體，其包含一晶圓及一支撐結 構體’其中該組合體係依據本發明之方法所製得。 本發明進一步之面向係為一組合體，其包含一晶圓及一支撐 結構體，其中該晶圓於其表面上具有多個突出體，且該支撐結構 體包含至少一具有一内表面及一外表面之形狀記憶聚合物薄膜， 其中所述之外表面被貼在該晶圓之突出體上，其條件為該形狀記 憶聚合物薄膜之外表面形成該支撐結構體之外部表面，且所述之 外部表面在貼附於晶圓之前所呈現之平坦度指數低於〇〇2,且所述 201208071 之形狀記憶聚合物薄膜所具有之平均薄膜厚度為10至1200 μηι。 如上文中所述，本發明之一面向係為一處理晶圓之方法，其 所包含之步驟有： a)提供一晶圓及一支撐結構體， 其中該晶圓於其表面上具有多個突出體， 且於其中該支撐結構體包含至少—具有一内表面及一外表面 之形狀記憶聚合物薄膜，其中該形狀記憶聚合物之外表面形成 該支撐結構體之外部表面，且所述之外部表面所呈現之平坦度 指數(FI)低於〇.〇2，且所述之形狀記憶聚合物薄膜所具有之平 均薄膜厚度為1〇至1200 μηι ; ' b)於0.1至100 MPa之壓力下，較佳者在丨至丨⑽河以之壓力下及 於一高於至少一形狀記憶聚合物之玻璃轉化溫度(Tg)之溫度下 使該等突出體接觸該支撐結構體之外部表面以形成一組合體； 幻於〇.1至1〇〇]\^&之壓力下，較佳者在1至1〇〇]^1>&之壓力下，將 該形成之組合體冷卻至高於2(Tc並低於至少一形狀記憶聚合 物之玻璃轉化溫度(Tg)之溫度。 & 本發明中所❹之術詞「平坦度指數」，被定義為支撐結構 體外部表面之滅度與雜記憶聚合物薄膜平均軸厚度之比 率。 平坦度指數可依圖i至圖3所示之方法測定之。本發 〜度」，係卿平均值而言，其被定線g 及波合Vn距離Dn之算術平均數。 ^發日种所使狀躺「波峰Pn」，係指㈣表面任何凸出 外=發明中所使用之術詞「波谷、」’則係指介於支撑 、，。構體外部表面凸出處間之任何凹口處。 本發明中所使狀躺「距軌」，储_波味及相鄰波 201208071 以和形狀記憶薄膜3之縱向延伸2呈大致 svn之高度差值，其等係 直角之方向所測得。 換句話說，距離Dn被定義為兩條大致上與縱向延伸2平行延伸 之直線間之距離，其係以和縱向延伸2呈大致直角之方向所測得， 其中直線相對S波辛!^，而另一直線則相對於相鄰之波谷V。 本發明中所使用之術詞「縱向延伸面」，意指—大致上與"支 掉結構體之表面平行延伸之平面，其中支撐結_表面 並夫被納入者虚。 粗链度可例如沿著-料支撐結構麟部表面上通常為〇〇8 cm之直線加以測量。較佳者為，本發明情使狀躺「粗縫度」， 係為在支#結獅外部表面上15個不同位置所測得15個不同測量 值之算術平均數。_度刊用—施“非_式雷射表面粗度 儀’其裝備有一AF2000型之自動對焦感測$，以〇 5 _/s之測量 速度測定之。所測得之數據可利用SQlar map 3丨1()之影像 分析軟體(高斯滤波器0.8 mm)分析之，其中使用到微粗链過滤，其 臨界值為2.5 μιη。 形狀記憶聚合物薄膜之平均薄膜厚度丁可依圖3所示之方法測 定之。平均薄膜厚度Τ係多個薄膜厚度值凡之算術平均值，其中每 一薄膜厚度值凡係以和介於兩相反對立點間之形狀記憶薄膜3之 縱向延伸2呈大致直肖之方向觸得，其巾—點位於形狀記憶聚合 物薄膜之内表面上，且其對應點位於該形狀記憶聚合物薄膜對立 之外表面上。 本發明中所使用之術詞「多個」，係指沿著該支撐結構體^縱 切面10個不同位置上至少1〇個不同之測量點。 平均薄膜厚度Τ之測定可例如利用該支撐結構體之一切面（尺 寸為1 on X 1 cm)，並將其分成5個相等之帶狀區域(每個區域為2 201208071 mm)進行纟JL每*該形狀記憶聚合物薄膜之平均薄膜厚度τ係沿 著該縱切面所収。沿著縱切面之平均薄膜厚度何利用掃描式電 子顯微鏡(SEM)測定之，較佳為使用一掃描式電子顯微鏡腿 JSM-6060 SEM ° 支撐結構體之外部表面可被視為—平坦或為—非微結構之表 面。術詞「微結構」’係指包含微結構特徵之表面，其中術詞「微 結構特徵」係指-表面之特徵，其等至少具有—維尺寸，較佳為 所有三維尺寸(例如高度、長度、寬度或直徑)小於i咖。 於一實施例中，本發明方法之步驟b)包含於4至5〇Mpa之壓力 下使支撐結制之㈣表面及晶圓於高於形狀記憶聚合物玻璃轉 化溫度(Tg)之溫度下接觸30秒至100分鐘，如丨至5分鐘。 較佳者為，使支撐結構體之外部表面與突出體接觸於9〇。匸至 180。(：之溫度下，更佳者為1〇〇cC至16〇<χ：<溫度下，且最佳者為 110 C至140°C之溫度下，及/或於4至3〇 MPa之塵力下，更佳者為5 至10 MPa之壓力下，且最佳者為5.5至8 MPa之壓力下。 使支撐結構體之外部表面與突出體接觸於1〇〇。€至16〇。(；：之溫 度下，及/或於5.5至8 MPa之壓力下是有利的，因為該等工序條件 造成晶圓與支撐結構體間之附著性更佳。 於另一實施例中，本發明方法之步驟c)包含於4至5〇MPa之壓 力下使該形成之組合體冷卻至高於2(rc之溫度且低於至少一形狀 記憶聚合物之玻璃轉化溫度(Tg)之溫度。 較佳者為，於4至30 MPa之壓力下，更佳者為5至1〇 MPa之壓 力下，且最佳者為5.5至8 MPa之壓力下使該形成之組合體冷卻至 高於20°C之溫度下，較佳者為高於3〇〇c^溫度下，且更佳者為高 於40°C之溫度下並低於至少一形狀記憶聚合物之玻璃轉化溫度(％) 之溫度。 g 201208071 於部份實施例中，本發明方法包含另外一步驟d)，其係將晶圓 進行研磨以使所述之晶圓變薄，以達到低於100μιη之厚度較佳。 為使支撐結構體從晶圓表面移除/釋出，本發明方法可包含另 外一步驟，其係將支撐結構體暴露於一高於該形狀記憶聚合物之 玻璃轉化溫度(Tg)之溫度下一段足以將所述支撐結構體從晶圓移 除之時間。 關於此方面，有利之作法係使用支撐結構體，其具有顯示平 坦度指數(FI)低於〇·〇2之外部表面，因為所述表面之平坦性質可使 支撐結構體以非常有效之方式被移除。支撐結構體，其具有顯示 平坦度指數(FI)低於0.02之外部表面，可依一無應力過程中應用很 低之剥離力大致完全地從晶圓表面上移除。 較佳者為’支撐結紐之移除係經由將其絲於高於該形狀 記憶聚合物之玻璃轉化溫度(Tg)至少1〇〇c之溫度下，較佳者為至少 20°C之溫度下，及尤佳者為至少3〇c>c之溫度下。 當使用適當之溫度時，本發明之支撐結可於短時間内， 較佳者為少於15分鐘内，更佳者為少於1〇分鐘内，201208071 VI. Description of the Invention: [Technical Field of the Invention] The present invention relates to a method of processing a wafer. The invention also relates to an assembly comprising a wafer and a support structure, wherein the support structure comprises at least one shape memory polymer film. [Prior Art] Thinning a semiconductor wafer may be desirable, for example, because a thinner wafer can reduce thermal impedance and increase reliability. Thinning the wafer reduces stress from thermal effects and thus increases reliability. Thick wafers can also adversely affect the size of the package. In some cases, thick wafers can have a negative impact on performance. Therefore, it is desirable to thin a semiconductor wafer for various reasons. Typically, thinning the wafer is to bring the wafer surface into contact with a hard, flat, rotating horizontal turntable containing liquid mud. The slurry may contain a grinding agent having a chemical etchant such as ammonia, fluoride or the like. The wafer then remains in contact with the vehicle until a certain amount of substrate has been removed to the desired thickness. Adhesive read materials are known to be used as a drag wafer in a wafer thinning process, such as a wafer processed via bump technology, to support the structure. In this regard, U.S. Patent No. 7,226,8i2B2 discloses a method of processing a wafer in which a cap-sacrificial polymer is overlaid on a bump surface of a wafer to form a thin layer. Then, the bump surface of the wafer is attached to the support structure of the wafer by using the adhesion layer to support the processing of the wafer on the back side. After the wafer is thinned, The wafer is exposed to a high temperature to be decomposed into a gas as a polymer (4). The decomposition of the sacrificial polymer reduces the dependence of the __ on the wafer bump. As a result, when the supporting substrate is separated from the wafer, the hybrid layer is the same as the supporting substrate - the bump of the crystal Face 1 This core requires shear force, wedge force and/or 201208071 peel force to separate the support structure from the thinned wafer surface. Moreover, this method cannot reuse the support structure. US Patent Application No. 2006/0046433 A1 teaches a method of processing a wafer, comprising the steps of: fabricating a plurality of protrusions on a wafer, and engaging the protrusions with a fixture, and the wafer A grinding process is performed to thin the wafer. The fixture has a plurality of interference-producing protrusions, which are suitable for surface characteristics extending outward from the surface to be thinned by the wafer to be thinned. In some embodiments, openings in the shape memory material can be used which are equivalent to heating the bumps more firmly to the wafer to be thinned. As wafers are made thinner and thinner, the severity of wafer damage can increase dramatically. When the thickness of the wafer is lower than ιοομιη, the wafer is prone to cracking, surface polishing, and surface irregularities. Wafers that have been processed via bumping techniques may have additional problems with the appearance of dents on the areas between the bumps. Especially for very thin wafers, the wafers are susceptible to damage when removed from the thinning fixture. Despite the state of the art, it is still expected to provide a better way to handle wafers', especially to provide a better way to thin the wafer, in which the mechanical force required to remove the wafer from the support structure is Moreover, the cleaning step of the support structure removes the residue of the support structure from the wafer surface. SUMMARY OF THE INVENTION The present invention provides a method of treating a crystal 1], such as a method of thinning a wafer, wherein two wafers have a plurality of protrusions on a surface thereof, which are supported by a support structure to reduce The support process of the wafer during mechanical processing, such as grinding, still contains at least - a shape memory having an inner (four) and a nuclear surface, wherein the outer surface of the shape memory polymer forms the support structure (external surface After the processing, the support structure can be removed from the wafer surface substantially completely according to the low-stress 4 201208071 peel force. The support structure can be repeatedly used. The characteristics reduce waste and overall cost. In its broadest sense, the present invention provides a method of processing a wafer comprising the steps of: a) providing a wafer and a support structure, wherein the wafer Having a plurality of protrusions on a surface thereof, and wherein the support structure comprises at least one shape memory polymer film having an inner surface and an outer surface, wherein the shape memory polymer has a surface shape The outer surface of the support structure' and the outer surface exhibits a flatness index (FI) of less than 0.02, and the shape memory polymer film has an average film thickness of 10 to 1200 μm; b) At a pressure of 0.1 to 100 MPa, preferably at a pressure of 丨 to 8 (MPa) MPa, and at a temperature above the glass transition temperature (Tg) of at least one shape memory polymer Contacting the outer surface of the support structure to form a combination; c) cooling the formed assembly to a temperature higher than 2〇 under a pressure of 0.1 to 100 MPa, preferably at a pressure of 1 to 1 MPa. °c is lower than the temperature of the glass transition temperature (Tg) of at least one shape memory polymer. Another aspect of the present invention is an assembly comprising a wafer and a support structure wherein the combination is made in accordance with the method of the present invention. The invention further relates to an assembly comprising a wafer and a support structure, wherein the wafer has a plurality of protrusions on a surface thereof, and the support structure comprises at least one having an inner surface and a a shape memory polymer film of the outer surface, wherein the outer surface is attached to the protrusion of the wafer, provided that the outer surface of the shape memory polymer film forms an outer surface of the support structure, and The outer surface exhibits a flatness index lower than 〇〇2 before being attached to the wafer, and the 201208071 shape memory polymer film has an average film thickness of 10 to 1200 μη. As described above, one aspect of the present invention is directed to a method of processing a wafer, the steps comprising the steps of: a) providing a wafer and a support structure, wherein the wafer has a plurality of protrusions on a surface thereof And a shape memory polymer film having at least an inner surface and an outer surface, wherein the outer surface of the shape memory polymer forms an outer surface of the support structure, and the outer portion The flatness index (FI) exhibited by the surface is lower than 〇.〇2, and the shape memory polymer film has an average film thickness of 1 〇 to 1200 μηι; ' b) at a pressure of 0.1 to 100 MPa. Preferably, the protrusions are brought into contact with the outer surface of the support structure at a temperature above the pressure of the glass transition temperature (Tg) of the at least one shape memory polymer at a pressure of the helium (10) river to form Under the pressure of a combination of 1 to 1 〇〇]\^&, preferably, the formed assembly is cooled to a higher pressure under the pressure of 1 to 1 〇〇]^1>&; 2 (Tc is lower than the glass transition of at least one shape memory polymer Temperature (Tg) Temperature & The term "flatness index" as used in the present invention is defined as the ratio of the degree of extinction of the outer surface of the support structure to the average axial thickness of the hetero memory polymer film. According to the method shown in Fig. i to Fig. 3, the present invention is the average of the alignment g and the Vn distance Dn. "Crest Pn" means any surface of the (4) surface. The word "valley," as used in the invention, refers to any notch between the support and the convex surface of the outer surface of the structure. The lying lie "track", the storage _ wave odor and adjacent wave 201208071 and the longitudinal extension 2 of the shape memory film 3 are substantially svn height difference, which is measured in the direction of the right angle. In other words, The distance Dn is defined as the distance between two straight lines extending substantially parallel to the longitudinal extension 2, which is measured in a direction substantially perpendicular to the longitudinal extension 2, wherein the straight line is opposite to Sboxin! Relative to the adjacent trough V. The term used in the present invention " "Expansion surface" means a plane extending substantially parallel to the surface of the structure, wherein the support knot is superficial. The thick chain can be, for example, along the surface of the support structure. It is usually measured on a straight line of 〇〇 8 cm. Preferably, the present invention has a "rough seam" which is 15 different measurements measured at 15 different positions on the outer surface of the branch lion. The arithmetic mean of the value. _ _ _ _ "non- _ type laser surface roughness meter" equipped with an AF2000 type of autofocus sensing $, measured at a measuring speed of 〇 5 _ / s. The data can be analyzed using the SQlar map 3丨1() image analysis software (Gaussian filter 0.8 mm), which uses a micro-thick chain filter with a threshold of 2.5 μηη. The average film thickness of the shape memory polymer film can be measured by the method shown in Fig. 3. The average film thickness is the arithmetic mean of the plurality of film thickness values, wherein each film thickness value is in a substantially straight direction with the longitudinal extension 2 of the shape memory film 3 between the opposite opposite points. The towel-point is located on the inner surface of the shape memory polymer film, and its corresponding point is on the opposite surface of the shape memory polymer film. The term "plurality" as used in the present invention refers to at least one different measurement point along 10 different positions of the longitudinal section of the support structure. The average film thickness Τ can be measured, for example, by using all the faces of the support structure (size 1 on X 1 cm) and dividing it into five equal strip-shaped regions (each area is 2 201208071 mm) for 纟JL per * The average film thickness τ of the shape memory polymer film is taken along the longitudinal section. The average film thickness along the longitudinal section is determined by scanning electron microscopy (SEM), preferably using a scanning electron microscope leg JSM-6060 SEM ° The outer surface of the support structure can be regarded as - flat or - Non-microstructured surface. The term "microstructure" refers to a surface containing microstructure features, wherein the term "microstructural feature" refers to a feature of a surface that has at least a dimensional dimension, preferably all three dimensional dimensions (eg, height, length). , width or diameter) is less than i coffee. In one embodiment, step b) of the method of the present invention comprises contacting the supported (4) surface and the wafer at a temperature above the glass transition temperature (Tg) of the shape memory polymer under a pressure of 4 to 5 MPa. 30 seconds to 100 minutes, such as 丨 to 5 minutes. Preferably, the outer surface of the support structure is brought into contact with the protrusions at 9 turns.匸 to 180. (At the temperature, preferably 1 〇〇 cC to 16 〇 < χ: < temperature, and the best is 110 C to 140 ° C, and / or 4 to 3 MPa Under the dust force, the pressure is preferably 5 to 10 MPa, and the optimum is 5.5 to 8 MPa. The outer surface of the support structure is in contact with the protrusion at 1 〇〇. (;: at a temperature, and/or at a pressure of 5.5 to 8 MPa, because the process conditions result in better adhesion between the wafer and the support structure. In another embodiment, Step c) of the inventive method comprises cooling the formed assembly to a temperature above 2 (rc temperature and below the glass transition temperature (Tg) of at least one shape memory polymer at a pressure of 4 to 5 MPa. Preferably, the formed assembly is cooled to a temperature higher than 20 ° C under a pressure of 4 to 30 MPa, more preferably 5 to 1 MPa, and most preferably at a pressure of 5.5 to 8 MPa. The temperature is preferably higher than 3 〇〇 c ^ temperature, and more preferably higher than 40 ° C and lower than the glass transition temperature of at least one shape memory polymer (% The temperature of the process. g 201208071 In some embodiments, the method of the present invention comprises an additional step d) of grinding the wafer to thin the wafer to a thickness of less than 100 μm. In order to remove/release the support structure from the wafer surface, the method of the present invention may comprise the additional step of exposing the support structure to a temperature above the glass transition temperature (Tg) of the shape memory polymer. a period of time sufficient to remove the support structure from the wafer. In this regard, it is advantageous to use a support structure having an outer surface exhibiting a flatness index (FI) lower than 〇·〇2, because The flat nature of the surface allows the support structure to be removed in a very efficient manner. The support structure has an outer surface exhibiting a flatness index (FI) of less than 0.02, which can be applied in a low stress-free process. The peel force is substantially completely removed from the wafer surface. Preferably, the 'supporting knot is removed by filamenting it to a glass transition temperature (Tg) higher than the shape memory polymer by at least 1 〇〇c. Temperature, preferably It is at a temperature of at least 20 ° C, and particularly preferably at a temperature of at least 3 〇 c > c. When a suitable temperature is used, the support of the present invention can be used in a short period of time, preferably less than 15 Within minutes, the better is less than 1 minute,
於5分軸，大致完全地«隊面上歸。内及尤佳者I 本發明中所使用之術詞「大致完全地」，較佳係指，少於5 wt.-%，較佳者為少於丨wt._% ’更佳者為少於〇 5 ^ ，及尤佳者 為少於0.1 wt.·%，以支撐結構體之總重量為依據，於支撐結構體 從晶圓表面上移除後殘留在晶圓表面上。 本發明情使用之術詞「_力」，較佳係㈣祕附在一 起（表面(支撐結構體之外部表面和晶面)時所需要之9〇。剝離 斤k之9G剥離力可於23 C下依據美國材料試驗協會(ASTM) 2-〇4試驗方法’利用-TXT咖拉伸測試儀(由英國Surrey， StableMicr〇Systems所取得），其使用5公斤之荷重元和25麗/論 201208071 之衝頭速度，所測得。 如本發明中所使用之「剝離力」，如9〇。剥離力，被視為是低 的，如於23°C下彼此分離支撐結構體和晶騎需要之麟力，如 90。剥離力，低於〇.〇1 N/mm，較佳者為低於〇 〇〇5 N/mm，更佳者 為低於0.002 N/mm，及特別佳者為低於〇 〇〇1 N/mm時。 令人期待者為，本發明支撐結構體外部表面之平坦度指數(FI) 要低於_，*佳者為低於0.005，及尤佳者為低於〇 〇〇1，因為所 述之表面非常平坦之性質可於無應力過程中應用—非常低之剥離 力，如低於0.005N/mm之90°剝離力，使支撐結構體大致完全地從 晶圓表面上移除。 如上文中所述，本發明支撐結構體之外部表面係由一形狀記 憶聚合物薄膜所形成β 如本發明中所使用之「形狀記憶聚合物」，係指對刺激有反 應之聚合物材料。當應用一外部刺激時，該等材料能夠改變其等 之形狀^由溫度變化所啟動形狀之改變可被視為溫度謗發型之形 狀記憶效應。 本發明中該至少一形狀記憶聚合物較佳者為具有介於4〇及 200°C之間，較佳者為介於50&150〇(：之間，及更佳者為介於1〇〇 及120。(：之間之玻璃轉化溫度（Tg)，其係以差示掃描熱量法 (Differential Scanning Calorimetry)所測定。該形狀記憶聚合物薄膜 之玻璃轉化溫度較佳者可利用一TA Instruments Q2000 MDSC於下 列條件下測定之：取少量大約10 mg之形狀記憶聚合物，並將其密 封於一鋁製樣品盤中，接著將其放入20eC之儀器中，並將其冷卻 至-90°C。經過於此溫度下平衡2〇分鐘之後，將樣本以每分鐘1〇〇c 之加熱速率加熱至2〇〇χ：。 於本發明之某些實施例中，形狀記憶聚合物係為一反應產 201208071 物，其係由固化一可固化之組成物所形成，該組成物包含有 i) 至少一可交聯之組成份，其於固化時形成一彈性體；及 ii) 一聚合物粉末，其以一使形狀維持之含量分佈於所述之可交聯 之組成份内，其於固化之彈性體中仍舊是各自分離，且具有一 低於該固化彈性體分解溫度之熔點溫度。 本發明中該至少一可交聯之組成份較佳者係由可固化之碎氧 樹脂組成物所選出。有許多不同類型之可固化矽氧樹脂組成物可 供使用。例如，可供使用者有熱固化型之矽氧樹脂組成物、濕氣 固化型之矽氧樹脂組成物及光固化型之矽氧樹脂組成物等。多形 態固化之矽氧樹脂組成物’例如光及濕氣雙重固化之組成物，抑 或熱及濕氣之雙重固化矽氧樹脂組成物亦是有用的。 令人期待的是，可固化之矽氧樹脂組成物係熱固化組成物。 尤其此類熱固化矽組成物包括可反應之聚有機妙氧燒 (polyorganosiloxanes)，其含有可反應之官能基，如乙埽基(vinyl) 或丙婦基(allyl)，或（曱基)丙晞酸醋基((meth)acrylate)。除該可反應 之聚有機矽氧烷之外，此類熱固化之組成物還包括矽氫化合物 (silicon hydride)之交聯劑及有機金屬之矽氫化催化劑。 可被使用於熱固化矽氧樹脂組成物中適用之可反應聚有機珍 氧烷之範例包括該等化合物，其等符合下列通式(I)之結構：On the 5th axis, roughly completely «team face. The term "substantially complete" as used in the present invention, preferably means less than 5 wt.-%, preferably less than 丨wt._% 'more preferably less于〇5 ^ , and especially less than 0.1 wt.·%, based on the total weight of the support structure, remains on the surface of the wafer after the support structure is removed from the wafer surface. The term "_force" used in the present invention is preferably (4) secretly attached to the surface (the outer surface and the crystal surface of the support structure). 9G peeling force can be peeled off at 23 C under the American Society for Testing and Materials (ASTM) 2-〇4 test method 'Using -TXT coffee tensile tester (obtained by Surrey, StableMicr〇Systems, UK), which uses 5 kg of load weight and 25 Li / on 201208071 The speed of the punch is measured. The "peeling force" as used in the present invention, such as 9 〇. The peeling force is considered to be low, as required to separate the support structure and the crystal ride at 23 ° C. The force of Lin, such as 90. Peeling force, lower than 〇.〇1 N/mm, preferably less than 〇〇〇5 N/mm, more preferably less than 0.002 N/mm, and especially good Below 〇〇〇1 N/mm. It is expected that the flatness index (FI) of the outer surface of the support structure of the present invention is lower than _, * is preferably less than 0.005, and particularly preferably is low. In 〇〇〇1, because the surface is very flat, the properties can be applied in the stress-free process—very low peel force, such as 90° peeling below 0.005N/mm. The force causes the support structure to be substantially completely removed from the surface of the wafer. As described above, the outer surface of the support structure of the present invention is formed of a shape memory polymer film as "shape" as used in the present invention. "Memory polymer" means a polymer material that responds to stimuli. When an external stimulus is applied, the material can change its shape. The change in shape initiated by temperature change can be regarded as the shape of the temperature 谤 hairstyle. The memory effect. The at least one shape memory polymer in the present invention preferably has a relationship between 4 〇 and 200 ° C, preferably between 50 & 150 〇 (:, and more preferably Between 1 and 120. (: glass transition temperature (Tg), which is determined by Differential Scanning Calorimetry. The glass transition temperature of the shape memory polymer film is preferably used. A TA Instruments Q2000 MDSC was measured under the following conditions: a small amount of approximately 10 mg of shape memory polymer was taken and sealed in an aluminum sample pan, which was then placed in a 20 eC instrument and Cool to -90 ° C. After 2 minutes of equilibration at this temperature, the sample is heated to 2 Torr at a heating rate of 1 ° C per minute: In some embodiments of the invention, shape memory The polymer is a reaction product of 201208071, which is formed by curing a curable composition comprising i) at least one crosslinkable component which forms an elastomer upon curing; and ii a polymer powder which is distributed in the crosslinkable component in a form-maintaining content which is still separated in the cured elastomer and has a decomposition temperature lower than the cured elastomer The melting point temperature. Preferably, the at least one crosslinkable component of the present invention is selected from the curable oxy-compound composition. There are many different types of curable epoxy resin compositions that can be used. For example, the user may have a thermosetting epoxy resin composition, a moisture-curable epoxy resin composition, and a photocurable epoxy resin composition. Multi-form cured epoxy resin compositions such as a dual curing composition of light and moisture, or a dual curing epoxy resin composition of heat and moisture are also useful. It is expected that the curable epoxy resin composition is a thermally cured composition. In particular, such heat-curing ruthenium compositions include reactive polyorganosiloxanes containing reactive functional groups such as vinyl or allyl or (fluorenyl) propyl. (meth)acrylate. In addition to the reactive polyorganosiloxane, such thermally cured compositions also include a crosslinker of a silicon hydride and a rhodium hydrogenation catalyst of an organometallic. Examples of reactive polyorganosiloxanes which may be employed in the thermally curable epoxy resin composition include such compounds which conform to the structure of the following general formula (I):
R5 R1 I I. (SiO)n—Si-R R1 R3—SiO- I R2 R5 R2 通式⑴ 其中R1、R2、R3及R5可為相同或相異’且為經取代或未經取代 之含有1至20個碳原子之碳氫化合物(hydrocarb〇n)殘基或碳氫化 201208071 合物氧基(hydrocarbonoxy)殘基，該等R殘基中至少有兩個殘基， 且被期待地超過兩個殘基為可反應之官能基，如烯烴基(olefmic)， 其包括乙缔基(vinyl)、（甲基）丙婦酸醋基((meth)acrylate)、馬來酸 基(maleate)及肉桂酸基(cinnamate)。例如，當一或數個前述之R殘 基(R1、R2、R3及R5)並非所需要之可反應官能基之其中一者時， 則該等殘基可選自奴基類(alkyl) ’如甲基(methyl)、丙基(propyl)、 丁基(butyl)及戊基(pentyl);烯基類(alkenyl)，如乙埽基(vinyl)及丙 烯基(allyl);環烷基類(CyCl〇aikyl)，如環己基(CyCi〇heXyl)及環庚基 (cycloheptyl);芳香基類（aryi)，如苯基(phenyl);芳香燒基類 (arylalkyl) ’如貝塔苯乙基（betaphenylethyl);燒芳香基類 (alkylaryl),及碳氫化合物氧基類(hydrocarbonoxy)，如燒氧基 (alkoxy)、芳香氧基(aryi〇xy)、烷芳香氧基(alkaryl〇xy)、芳香烷氧 基(aryalkoxy) ’且所期待者為甲氧基(meth〇Xy)、乙氧基(eth〇Xy)或 經基(hydroxy)等。同樣適用者還有前述殘基之任意一者，其有數 個或全部之氫原子被例如齒素，如氟或氯，所取代。前述R殘基 之其中一或數個亦可為氫原子，其所提供需要之可反應官能基如 前文所示，而且氫原子之存在不會阻礙聚有機矽氧烷於本發明中 表現之能力。期待地，於上文通式中之R3為： O R4R5 R1 I I. (SiO)n-Si-R R1 R3—SiO- I R2 R5 R2 Formula (1) wherein R1, R2, R3 and R5 may be the same or different and are substituted or unsubstituted. Hydrocarbon(R) residues of 1 to 20 carbon atoms or hydrocarbons 201208071 Hydroxyoxy residues having at least two residues in the R residues, and are expected to exceed two Residues are reactive functional groups, such as olefmic, which include vinyl, (meth) acrylate, maleate, and Cinnamate. For example, when one or more of the aforementioned R residues (R1, R2, R3 and R5) are not one of the desired reactive functional groups, then the residues may be selected from the group 'alkyl' Such as methyl, propyl, butyl and pentyl; alkenyl, such as vinyl and allyl; cycloalkyl (CyCl〇aikyl), such as CyCi〇heXyl and cycloheptyl; aromatic (aryi), such as phenyl (phenyl); aromatic alkyl (arylalkyl) such as beta phenethyl ( Betaphenylethyl); alkylaryl, and hydrocarbonoxy, such as alkoxy, aryi〇xy, alkaryl〇xy, aromatic Alkoxy (aryalkoxy) ' is expected to be methoxy (meth〇Xy), ethoxy (eth〇Xy) or hydroxy. Also suitable for use is any one of the aforementioned residues, wherein several or all of the hydrogen atoms are replaced by, for example, dentate such as fluorine or chlorine. One or more of the aforementioned R residues may also be a hydrogen atom, which provides the desired reactive functional group as shown above, and the presence of a hydrogen atom does not hinder the ability of the polyorganosiloxane to behave in the present invention. . Desirably, R3 in the above formula is: O R4
— II I— II I
R —0—C—C=CH 其中R為，經取代或未經取代，含有1至20個碳原子之碳氫化 合物殘基’且期待地為—燒基，如丙基；且r4 或甲基 (CH3)。 為達到特殊之分予量、黏稠度及其他之化學或物理性質，該 可反應之聚有财氧燒中重複單元之數目可加以改變。通常，n 12 201208071 為整數，例如當η從1至1200，且被期待地從10至1000時，於 25°C下之黏稠度係從大約25 cps至大約2,500,000 cps。 本發明之可反應聚有機矽氧烷可包括，作為其骨幹結構之部 分者，有一或數個雙價之經取代或未經取代，含有1至20個碳原 子之脂肪烴(aliphatic)、環脂肪烴(cycloaliphatic)或芳香碳氫化合物 (aromatic hydrocarbon)之殘基，其等可被一含有雜原子之鍵結所中 斷。該雜原子可包括氮(N)、氧(0)或硫(S)等原子。於該等適用之 雙價碳氫化合物殘基當中包括有亞烷基（alkylenes)、聚烯烴 (polyolefins) '聚酸（polyethers)、聚酯（polyesters)、聚氨醋 (polyurethanes)及其等之組合物。 令人期待者為，可交聯之組成份包括一化合物(可反應之聚有 機矽氧烷），其具有下列之通式(II): MA R5 (CH30)2.c—SiO—(SiO)n——Si-(OCH3)2-cR —0—C—C=CH wherein R is a substituted or unsubstituted hydrocarbon residue having 1 to 20 carbon atoms′ and is desirably a —alkyl group such as propyl; and r 4 or A Base (CH3). In order to achieve a particular amount, viscosity, and other chemical or physical properties, the number of repeating units in the reactive polyoxygenate can be varied. Typically, n 12 201208071 is an integer, such as when η is from 1 to 1200, and desirably from 10 to 1000, the viscosity at 25 ° C is from about 25 cps to about 2,500,000 cps. The reactive polyorganosiloxane of the present invention may comprise, as part of its backbone structure, one or several divalent substituted or unsubstituted aliphatic hydrocarbons having from 1 to 20 carbon atoms. Residues of cycloaliphatic or aromatic hydrocarbons, etc., may be interrupted by a bond containing a hetero atom. The hetero atom may include an atom such as nitrogen (N), oxygen (0) or sulfur (S). Among these suitable divalent hydrocarbon residues are alkylenes, polyolefins, polyethers, polyesters, polyurethanes, and the like. combination. It is expected that the crosslinkable component comprises a compound (reactive polyorganosiloxane) having the following general formula (II): MA R5 (CH30)2.c-SiO-(SiO) n——Si-(OCH3)2-c
I I5 I (CH3)c r5 (CH3)c 通式(II) 其中MA係為甲基丙烯醯基氧基丙基(methacryloxypropyl)，n係從 1至1200，且c為0或1 ; R5為一經取代或未經取代，含有1至 20個碳原子之碳氫化合物殘基，或含有1至20個碳原子之碳氫化 合物氧基殘基。 可反應之聚有機矽氧烷應該存在之含量為大約50至大約95 重量百分比，且被期待之含量為大約60至80重量百分比，其係 以可固化組成物之總含量為依據。 矽氫交聯劑化合物亦可係攙混使用，且特別有用於熱固化之 組成物中。此等材料可由相當廣泛之化合物中選出，儘管該交聯 13 201208071 劑被期待要符合下列之通式(in): R7- R10 Ri〇' I I |1— ° — (Si — 0)x—(SiO)y— R10 R10 Rio R1 通式(in) 其中R7、R8及R9中至少有兩個殘基為氫原子；此外，R7、R8及 R9可為相同或相異，且可為經取代或未經取代，含有丨至20個碳 原子之碳氫化合物殘基，該碳氫化合物殘基包括該些於上文中針 對通式(I)所定義之化合物，因此矽氫基(SiH)可位於終端，侧端或 兩者皆是；R1Q亦可為一經取代或未經取代，含有1至20個碳原 子之碳氫化合物殘基’包括該些於上文中針對通式⑴所定義之化 合物’且被期待為一烷基，如甲基4係為一從1〇至1〇〇〇之整數； 且y為一從1至20之整數。令人期待者為’不是氫原子之r基係 為甲基。矽氫交聯劑應該存在足夠之含量以達到所希望之交聯 量，且被期待之含量為該可固化組成物之大約1至大約1〇重量百 分比。 適用之有機金屬矽氫化(hydrosilation)催化劑可選自任意貴金 屬或含有貴金屬之催化劑，其可有效啟動熱矽氫化固化反應。特 別適用者係為柏(platinum)及姥(rhodium)催化劑，其等可有效催化 與矽鍵結之氫原子及與矽鍵結之烯烴基間之加成反應。 其他於本發明中適用之催化劑種類包括有機之姥(rhodium)及 銘(platinum)之醇化物。釘(ruthenium)、把(palladium)、鐵(osmium) 及銥(iridium)等之複合物亦可被考慮》有機金屬矽氫化催化劑可使 用任意有效之含量以完成熱固化。令人期待者為，該催化劑存在 之含量為0.025至大約1.0重量百分比，其係以可固化組成物之總 含量為依據。該催化劑之組合亦可被納入考慮。 201208071 令人期待者為，本發明中該至少一可交聯之組成份存在 量為20至80重量百分比，且更令人期待者為，其存在之^ 40至60重量百分比，其係以可固化組成物之總含量為依據。 固化(硫化)反應可被定義為任何之處置，其使彈性體之黏祠户 增加’並使張力強度和模數，及應變破壞(strain_t〇_failure)増加二 此過程可以被描料聚合物分子間之交聯反應，其亦包括鍵之延 長與交聯。液體含乙埽基之酯類或醚類皆被包括於適用之 脂聚合物當中。 樹 本發明之聚合物粒子於固化之彈性體中仍舊是各自分離的， 並具有一低於該固化彈性體分解溫度之熔化溫度。 本發明中所使用於表示聚合物粒子之「熔化溫度」該術語， 較佳者係指聚合物粒子經歷從固體至液體狀態改變時之溫度。熔 化溫度可以差示掃描熱量法(DSC)測定之，其中該溶化溫度被定義 為一 DSC曲線之曲折點。 本發明中所使用於表示固化彈性體之「分解溫度」該術語， 係指彈性體經歷重量減損超過10 wt.%，較佳者為超過2〇 wt.%時 之溫度。分解溫度可以熱重量分析法(TGA)測定之。 於適用之聚合物粒子中，如聚合物粉末，係為聚烯烴 (polyolefins)或埽烴共聚物（copolyoleflns)，如聚乙婦 (polyethylene)、聚丙烯（polypropylene)、聚乙埽共聚·丙締 (polyethylene-eo-propylene)、聚 丁二烯（polybutadiene)(72% 為順 式，28%為反式）、聚己内酯(卩〇1>^^瓜1&(^〇1^)、同排聚（1-丁烯) (isotactic poly(l-butene))、對排聚丙埽（Syndi〇tactic polypropylene)、聚(1-癸婦)(poly(l-decene))、聚（乙烯共聚-1·丁晞) (poly(ethylene-co-l-butene))、聚（乙稀-共聚-乙酸乙烯酯） (poly(ethylene-co-vinylacetate))、聚丁稀己二酸(polybutylene adipic 15 201208071 acid)、聚（α-甲基苯乙烯-共聚_甲基苯乙埽）(p〇iy(a methyl 1，4-聚丁二烯(trans l，4-p〇lybutadiene)或反式 i，4-聚異戊二烯(trans l，4-P〇lyiS〇prene)等。本發明聚合物粒子之粒徑廣泛變化從％奈 米達到大約100微米。令人期待者為，聚合物粒子所具有之粒徑 範圍從5至大約10微米。 粒徑可以雷射繞射法利用一 Mastersizer 2〇〇〇(由Malvern instruments Ltd所生產，並依據米氏理論(Mie)所計算)測定之。 本發明中所使用之術詞「粒徑」，係指d5〇粒子直徑。D5〇 代表粒子直徑，其定義為50%之粒子大於該直徑，而另外5〇%之 粒子則小於該直徑。 令人期待者為，聚合物粒子係以一使形狀維持之含量分佈於 該可交聯之組成份内，較佳含量為i至8〇 wt %，更佳含量為2〇 至60 wt.%，及更佳含量為3〇至5〇 wt.%，其係以可固化組成物之 總含量為依據。 本發明之形狀記憶聚合物可依任意方法製備之。美國專利申 請案No. 2004/0266940 A1論述一用於製備本發明形狀記憶聚合物 特別佳之方法，其被參考納入本發明之文獻中。 於本發明進一步之實施例中，該形狀記憶聚合物係為一形狀 記憶環氧樹脂(epoxy)聚合物。較佳之形狀記憶環氧樹脂聚合物係 選自於由固化一可固化組成物時所形成之反應產物，其包含芳香 烴二環氧化物(aromatic diepoxide)(剛性環氧樹脂）、脂肪烴二環氧 化物(aliphatic diepoxide)(彈性環氧樹脂）及二胺類(diamine)固化 劑。 適用之芳香烴二環氧化物包括雙酚A環氧樹脂單體之二縮水 甘/由謎(diglycidyl ether) ’ 其可向 Hexion Speciality Chemicals 以商 201208071 品名EPON 826購買取得；適用之脂肪烴二環氧化物包括新戊二醇 一縮水甘油醚(neopentyl glyC〇i diglycidyl ether)(NGDE)，其可向 TCI America購買取得；適用之二胺類固化劑包括聚（丙二醇)雙(2_ 氨丙基)½) (poly(propylene glyc〇1)bis(2-amin〇pr〇pyI)ether)，其可向I I5 I (CH3)c r5 (CH3)c Formula (II) wherein MA is methacryloxypropyl, n is from 1 to 1200, and c is 0 or 1; R5 is A substituted or unsubstituted hydrocarbon residue having 1 to 20 carbon atoms or a hydrocarbon oxy residue having 1 to 20 carbon atoms. The reactive polyorganosiloxane should be present in an amount from about 50 to about 95 weight percent, and is expected to be present in an amount from about 60 to 80 weight percent based on the total amount of the curable composition. The hydrazine crosslinking agent compound can also be used in combination, and is particularly useful in compositions for thermal curing. These materials can be selected from a wide variety of compounds, although the cross-linking 13 201208071 agent is expected to conform to the following general formula (in): R7- R10 Ri〇' II |1— ° — (Si — 0)x—( SiO)y— R10 R10 Rio R1 Formula (in) wherein at least two of R7, R8 and R9 are a hydrogen atom; in addition, R7, R8 and R9 may be the same or different and may be substituted or Unsubstituted, a hydrocarbon residue having from 丨 to 20 carbon atoms, the hydrocarbon residue including the compound defined above for the formula (I), such that the hydrazine group (SiH) can be located Terminal, side or both; R1Q may also be a substituted or unsubstituted hydrocarbon residue containing from 1 to 20 carbon atoms 'including those compounds defined above for formula (1)' And is expected to be an alkyl group, such as methyl 4 is an integer from 1 〇 to 1 ;; and y is an integer from 1 to 20. It is expected that the r group which is not a hydrogen atom is a methyl group. The rhodium crosslinking agent should be present in an amount sufficient to achieve the desired amount of crosslinking and is expected to be present in an amount from about 1 to about 1% by weight of the curable composition. Suitable organometallic hydrosilation catalysts can be selected from any noble metal or noble metal containing catalyst which is effective to initiate the hot hydrogenation curing reaction. Particularly suitable for use are platinum and rhodium catalysts which are effective for catalyzing the addition reaction between a hydrogen atom bonded to a hydrazine and an olefin group bonded to a hydrazine. Other types of catalysts suitable for use in the present invention include organic rhodium and platinum alkoxides. Composites of ruthenium, palladium, osmium, and iridium may also be considered. The organometallic ruthenium hydrogenation catalyst may be used in any effective amount to effect thermal curing. It is expected that the catalyst will be present in an amount from 0.025 to about 1.0 weight percent based on the total amount of the curable composition. Combinations of the catalysts can also be considered. 201208071 It is expected that the at least one crosslinkable component is present in the present invention in an amount of from 20 to 80 weight percent, and more desirably, from 40 to 60 weight percent, which is The total content of the cured composition is based on. The curing (vulcanization) reaction can be defined as any treatment that increases the bond of the elastomer and increases the tensile strength and modulus, and the strain damage (strain_t〇_failure). The process can be traced to the polymer. Intermolecular cross-linking reactions, which also include elongation and cross-linking of bonds. Esters or ethers containing ethyl hydrazino groups are included in the suitable lipid polymer. The polymer particles of the present invention are still separately separated in the cured elastomer and have a melting temperature below the decomposition temperature of the cured elastomer. The term "melting temperature" as used in the present invention is used to mean the temperature at which the polymer particles undergo a change from a solid to a liquid state. The melting temperature can be determined by differential scanning calorimetry (DSC), which is defined as the tortuosity point of a DSC curve. The term "decomposition temperature" as used in the present invention to mean a cured elastomer means a temperature at which the elastomer undergoes a weight loss of more than 10 wt.%, preferably more than 2 〇 wt.%. The decomposition temperature can be determined by thermogravimetric analysis (TGA). Among the suitable polymer particles, such as polymer powder, are polyolefins or copolyoleflns, such as polyethylene, polypropylene, polyethylene copolymer, and polypropylene. (polyethylene-eo-propylene), polybutadiene (72% cis, 28% trans), polycaprolactone (卩〇1>^^瓜1&(^〇1^), Isotactic poly(l-butene), Syndi〇tactic polypropylene, poly(l-decene), poly(ethylene copolymerization) -1·丁晞) (poly(ethylene-co-l-butene)), poly(ethylene-co-vinylacetate), polybutylene adipic 15 201208071 acid), poly(α-methylstyrene-co-methyl acetophenone) (p〇iy(a methyl 1,4-polybutadiene (trans l,4-p〇lybutadiene) or trans i,4-polyisoprene (trans l,4-P〇lyiS〇prene), etc. The particle size of the polymer particles of the invention varies widely from about 100 microns to about 100 microns. It is expected that the polymer Particle size range 5 to about 10 μm. The particle size can be determined by laser diffraction using a Mastersizer 2 (manufactured by Malvern Instruments Ltd and calculated according to Mie's theory (Mie)). The word "particle size" refers to the particle diameter of d5〇. D5〇 represents the particle diameter, which is defined as 50% of the particles are larger than the diameter, and the other 5〇% of the particles are smaller than the diameter. The particle system is distributed in the crosslinkable component in a form which maintains the shape, preferably in an amount of from i to 8 wt%, more preferably from 2 to 60 wt.%, and more preferably in a range of from 3 to 60 wt.%. Up to 5 〇 wt.%, based on the total content of the curable composition. The shape memory polymer of the present invention can be prepared by any method. US Patent Application No. 2004/0266940 A1 discusses a preparation for this A particularly preferred method of inventing a shape memory polymer is incorporated herein by reference. In a further embodiment of the invention, the shape memory polymer is a shape memory epoxy polymer. Memory epoxy resin system A reaction product formed by curing a curable composition, comprising an aromatic diepoxide (rigid epoxy resin), an aliphatic diepoxide (elastic epoxy) Resin) and diamine curing agent. Suitable aromatic hydrocarbon diepoxides include bisphenol A epoxy resin diglycidyl ether' which can be purchased from Hexion Speciality Chemicals under the trade name 201208071 under the name EPON 826; suitable aliphatic hydrocarbon second ring Oxides include neopentyl gly C〇i diglycidyl ether (NGDE), which is commercially available from TCI America; suitable diamine curing agents include poly(propylene glycol) bis(2_aminopropyl) 1⁄2) (poly(propylene glyc〇1)bis(2-amin〇pr〇pyI)ether)
Huntsman以商品名jeffamine d_23〇購買取得。 美國專利申請案No. 2008/0262188 A1論述一用於製備本發明 形狀圮憶環氧樹脂聚合物特別佳之方法，其被參考納入本發明之 文獻中。 於本發明之一實施例中，位於晶圓表面上，較佳為位於晶圓 正面上之突出體係為焊接凸塊。如本發明中所使用之術詞「洋接 凸塊」包括大的焊接形成物，其被用於將平面結構結合在一起， 並將該等結構連接導電》令人期待者為，該等焊接凸塊所具有於 空間之任何方向上最大長度要介於2〇卜111至5〇〇 μπΐ2範圍内。較佳 者為，該等焊接凸塊於空間任何方向上最大長度在針對於第一級 半導體封裝之應用上要介於25 μιη至1〇〇 μπι之範圍内，且在針對於 某些第一級和所有第二級半導體封裝之應用上要介於15〇卜111至 400 μιη之範圍内，其中第一級之應用關於半導體坪接凸塊直接貼 附在封裝基板之表面上，且第二級之應用關於整個封裝組合體貼 附在印刷電路板之表面上。焊接凸塊之長度係以掃描式電子顯微 鏡(SEM)測定之，較佳為使用一掃描式電子顯微鏡JE〇LJSM_6〇6〇 SEM。 支撐結構體之形狀記憶聚合物係一層薄膜，其所且有之平均 薄膜厚度為10至1200 μιη。含有形狀記憶聚合物薄膜之支撐結構體 於背面研磨時被用來支撐晶圓。形狀記憶聚合物薄膜於進行背面 研磨過程時保護位於晶圓表面上之突出體，如凸塊，其中該形狀 記憶聚合物薄膜之厚度通常至少比該突出體之高度，如凸塊(空間 17 201208071 任何方向上最大長度)大兩倍，以雜位於該晶圓正面上之突出體 及其他突出結構體受到有效之保護。 目前已發現，平均薄膜厚度之·_41μιη，較佳 為is至5〇0 μιη，及特別佳為3〇至遞阿時，晶圓表面之突出處於 研磨過程巾可有效地被保護，且應用_少於Q篇N/mm之低9〇。剥 離力’支撐結構體可^無應力過程中自晶圓表面大致完全地移除。 應該要注意的是本發明之支撐結構體可或不可包含至少一承 載薄膜。 如果當支撐結構體包含至少一承載薄膜時，則所述之承載薄 膜則係貼附在該支撐結構體形狀記憶聚合物之内表面上。 有利者為，使用本發明方法中包含至少一承載薄膜之支撐結 構體，因為該承載薄膜減少支撐結構體自晶圓表面移除時所需要 之剝離力。此外’承載薄膜給予支撐結構體結構上之完整性及/或 剛性，及/或確保支撐結構體之再利用性。 本發明之至少一承載薄膜係具有紫外線透光性及/或可包含至 夕一聚合物’其選自聚乙稀（P〇lyethyleneS)、聚丙缔 (polypropylenes)、聚碳酸酉旨(p〇iyearb〇nates)、聚 g旨類(polyesters)、 聚對苯二甲酸乙二酯（p〇lyethyleneterephthalates)、聚氣乙浠 (polyvinylchlorides)、乙婦(ethylene)和乙酸乙晞酿(vinyl acetate)之 共聚合物’及/或其等之組合物。承載薄膜可包含或由一、二或超 過二層不同之層體所組成，其中每一層體可包含或可由至少一前 述聚合物所組成。 令人期待者’該承載薄膜之平均厚度範圍為由50至200 μιη， 較佳為由70至175 μηι，及更佳者為由90至140 μιη。 承載薄膜之平均薄膜厚度被定義為多個薄膜厚度值之算術平 均數’其中每一薄膜厚度值係以和介於兩相反對立點間之承載薄 201208071 膜 之縱向延伸面呈大致直角之方向所 薄膜之内表面上，且其對應馳糾& 糾㈣承載 W田1 ί 承裁薄膜之相對外表面上。 Γ Γ 多個」’例如沿著該支撐結構體-縱切 面10個不同位置上至少10個不同之測量點。 、一錄«膜之平均薄祕度之·可例如該支撐結構體 ^切面(尺寸為1⑽X lem)，並將其分成5個相等之帶狀區域(每 :區域為2mm)進行之，且每次該承_膜之平均雜厚度係沿著 該縱切面賴定。承麟膜沿著縱切私料祕厚度可利用掃 描式電子顯㈣(随），較佳域用—掃描式電子顯微細0L JSM-6060 SEM測定。 本發明進-步之面向係為-組合體，其包含—晶圓及一支撐 結構體，其中該組合體係依據本發明之方法所製得。 本發明之另-面向係為一組合體，其包含一晶圓及一支撐結 構體’其中該晶圓於其表面上具有多個突出體，且該支撐結構體 包含至少一具有一内表面及一外表面之形狀記憶聚合物薄膜其 中所述之外表面被貼在該晶®之突㈣上，其條件為該形狀記憶 聚合物薄膜之外表面形成該支撐結構體之外部表面，且所述之外 部表面在貼附於晶圓之前所呈現之平坦度指數低於〇〇2，且所述之 形狀記憶聚合物薄膜所具有之平均薄膜厚度為1〇至12〇〇 μπι。 【實施方式】 圖1顯示如本發明方法中所使用之一支撐結構體之剖面圖。所 述支撐結構體包含一形狀記憶聚合物薄膜3，其所具有之平均薄膜 厚度介於10 μιη至1200 μιη之範圍内。圖2描述顯示於圖1之支撐結 構體1部份之放大剖面圖。形狀記憶聚合物薄膜3所具有之平坦度 指數低於0.02，其中平坦度指數被定義為支撐結構體丨外部表面之 粗糙度與形狀記憶聚合物薄膜3平均薄膜厚度之比率。 201208071 圖2描述顯示於圖丨之支撐結構體丨部份之放大剖面圖，並示出 形狀圮憶聚合物薄膜3之粗糙度。外表面理所當然呈現不平坦，並 顯示出特殊之粗糙度。形狀記憶聚合物薄膜3表面之粗糙度反映在 形狀記憶聚合物薄膜3之切面上，其等在厚度上或多或少具有些微 之差異。此差異性之厚度反映在於圖3中所顯示之支撐結構體1之 4面圖上’其所呈現之形式有位於形狀記憶聚合物薄膜3/支撐結構 體1之外表面/外部表面上之波峰Pl、p2、p3及波谷％、V2、Vs、乂4。 本發明中所使用之術詞「粗糙度」，係指粗糙平均值而言，其被 疋義為波辛Pn及波谷Vn距離Dn之算術平均數。本發明中所使用之 術詞:波轨」，係指外部表面任何凸域，而本發明中所使用 讀柯「波谷Vn」，縣齡於支獅構體外部表面凸出處間之 任何，處。本發明中所使用之術詞「距離A」，係指—波峰^ 及相鄰波谷Vn之高度之差值，其等係以和形狀記憶薄膜3之縱向延 伸2呈大致直角之方向所測得。 粗縫度可利$ -Solarius非接觸式之雷射表面粗度冑，其裝備 有AF2〇〇〇型之自動斜焦感測器，以〇 5 _/s之測量速度測定 之。所測得之數據可_SGlarmapuniversal3 U(^影像分析軟體 (高斯滤波器G.8 mm)分析之，其中使用到微粗糖過滤，其臨界 2.5 um 〇 ^ 、形狀記憶聚合物薄膜之平均薄膜厚度τ可依圖3所示之 疋（。平均薄膜厚度τ係多個薄膜厚度值τη之算術平均值，其中矣 -薄膜厚度值τη係以和介於兩相反對立關之形狀記憶聚合物 膜3《縱向延伸2呈大致直角之方向所測得，其中—點位於形狀 憶聚合物薄膜之岐面上，且雜應齡於該雜 薄 膜對立之外表面上。 丨〜取口物溥 上又中所使用之術詞「多個」，可例如指於沿著該支撐結構 201208071 體1之一縱切面10個不同位置上至少10個不同之測量點。 平均薄膜厚度T之測定可例如利用該支撐結構體之一切面（尺 寸為1 cm X lcm)，並將其分成5個相等之帶狀區域(每個區域為2 mm)進行之’且每次該形狀記憶聚合物薄膜之平均薄膜厚度丁係沿 著該縱切面所測定。沿著縱切面之平均薄膜厚度了可利用掃描式電 子顯微鏡（SEM)測定，較佳為使用一掃描式電子顯微鏡jEOL JSM-6060 SEM。 實施例 實施例中使用到下列材料： 經凸塊技術處理過之晶圓 晶圓類型：VPA 18 Sn3.oAgo.5Cu氮化物鈍化； 經凸塊技術處理過之晶圓小方塊4 X 4 cm，具有96 μιη直徑長度之 凸塊； 凸塊間距：225 μιη ; 名義經分割芯片尺寸：7 X 7 mm ; 切縫寬度：750 μιη。 形狀記憶聚合物薄膜(SMP薄膜） 該形狀記憶聚合物薄膜係以紫外光固化法所製得，係為一可 紫外光固化之組成物，其包含有 61 wt.-%之聚二甲基矽氧烷型之聚合物， 28wt.-°/〇之丙烯酸異冰片酯， 10 wt.-%之氣相式二氧化矽， 1 wt.-%之光起始劑。 使用面積為4x4 cm之形狀記憶聚合物薄膜，其中該形狀記憶 聚合物薄膜所呈現之玻璃轉化溫度(Tg)大約為110。(：，其係由差示 掃描熱量法(DSC)所測得。 21 201208071 承載薄膜 帅且面積為4 X 4 cm，被作 一聚酯薄膜，平均之膜厚度為125 為承載薄膜使用。 於實施例中使關不暇撐結構體。該等支撐結構體於所有 情況中皆包含前文中所述之形狀記憶聚合物(SMp)薄膜，其且有_ 内表面及-外表面，其中卿狀記憶聚合物薄膜之絲面形成該 支撐結構體之外部表面。每-形狀記憶聚合物薄膜之粗糙度、平 均薄膜厚度及平坦度指數皆示於表4。於有些實施例中，該支撐 結構體亦包含-承載薄膜’其中該承載薄膜被貼附在該形狀記憶 聚合物薄膜之内表面上。 粗糙度 粗糙度係沿著一位於支撐結構體外部表面上〇 〇8 ^⑺之直線所 測定。於表1中所示之粗糙度係為在支撐結構體外部表面上15個不 同位置所測得15個不同測量值之算術平均數。粗糙度係利用一 Solarius非接觸式之雷射表面粗度儀，其裝備有一AF2〇〇〇型之自動 對焦感測器，以0.5 mm/s之測量速度所測定。所測得之數據係利用 solar map universal 3.1.10之影像分析軟體(高斯濾波器〇 8⑺叫分 析之’其中使用到微粗糙過濾，其臨界值為2.5 μιη。 形狀記憶聚合物薄膜之平均薄膜厚度Τ 形狀記憶聚合物薄膜之平均薄膜厚度τ係沿著於圖3中所示之 縱剖面所測定。於表1中所示之平均薄膜厚度Τ係沿著該形狀記憶 聚合物薄膜之縱剖面10個不同位置上所測得到10個薄膜厚度值Τη 之算術平均數。沿著該縱剖面之平均薄膜厚度Τ係以掃描式電子顯 微鏡方法(SEM)，使用一掃描式電子顯微鏡JEOL JSM-6060 SEM， 所測定。 平坦度指數(FI) 22 201208071 平坦度和數被疋義為支撐結構體外部表面之粗链度與形狀記 憶聚合物薄膜平均薄膜厚度之比率。 形成一支撐結構體與晶圓之組合體 一支撐結構體與晶圓之組合體係以使晶圓之經由凸塊技術處 理過之表面於6 MPa之壓力下及l3〇t之溫度下接觸支撐結構體之 外部表面15分鐘所形成。然後，該組合體接著於6Mpa之壓力下被 冷卻不到2分鐘而達至22。(：之溫度。 支撐結構體之移除 將支撐結構體與晶圓之組合體暴露於125。(：之溫度下2分鐘以 活化形狀記憶聚合物薄膜。 然後將支撐結構體與晶圓之組合體冷卻至23°c並應用剝離力 將支撐結構體從晶圓表面上移除。於表丨中所述之剝離力係剝開貼 附在一起之表面（支撐結構體之外部表面和經凸塊技術處理過之晶 圓表面）時所需要之9〇。剝離力。所述之90。剝離力係於23°C下依據 美國材料試驗協會(ASTM) D6862-04試驗方法，利用一TXT plus 拉伸測試儀(由英國Surrey ’ Stable Micro Systems所取得)所測得， 其使用5 kg之荷重元和25 mm/分鐘之衝頭速度。 【圖式簡單說明】 本發明較佳之實施例以下列之圖式說明之。 圖1顯示本發明支撐結構體之剖面圖。 圖2 顯示於圖1所示之支撐結構體部份之放大剖面圖。 圖3顯示於圖1所示之支撐結構體部份之另一放大剖面圖。 【主要元件符號說明】 1 支撐結構體 2 縱向延伸 23 201208071 3 形狀記憶聚合物薄膜 Pi ' P2 ' P3 波峰 Vi ' V2 ' v3 ' v4 波谷 Di ' D2 距離 Ti ' T2 ' T3 厚度值 24 201208071 一袼 1? 給 不可 90°剝離力 [N/mm] 〇-釋出順暢 0-釋出順暢 0-釋出順暢 0-釋出順暢 0.052 承載薄膜 杯r 杯r 沒有 沒有 SMP平坦度指 數FI 0.003 0.003 0.011 0.015 0.09 SMP平均薄膜 厚度[μπι] 500 500 350 280 250 SMP粗糙度 im]_ 1.49 1.49 〇 rn rn 寸· 實施例1 實施例2 實施例3 實施例4 比較實施例 。金樂 Tvg^s 吨每今钿丧罐叫* $-#『fv砸«_铋冢(¥嫁駟«^)觀罅珑miv SO农vad}鵪骒甸刼+v?*金飨τνθ^® 吨每耷^択該^叫畹铋交a -觀璁垅轅Tvi-交ίΝοο^(Ιί}鵪珲甸刼騣¥蠕駟^-夺一蝽Huntsman is purchased under the trade name jeffamine d_23〇. A particularly preferred method for preparing the shape of the epoxy resin of the present invention is discussed in U.S. Patent Application Serial No. 2008/0262188 A1, which is incorporated herein by reference. In one embodiment of the invention, the protruding system on the surface of the wafer, preferably on the front side of the wafer, is a solder bump. The term "foreign bump" as used in the present invention includes a large solder formation that is used to bond planar structures together and to connect the structures to electrical conduction. The maximum length of the bump in any direction of the space is in the range of 2 111 111 111 to 5 〇〇 μπ ΐ 2 . Preferably, the maximum length of the solder bumps in any direction of the space is in the range of 25 μm to 1 μm μm for the application of the first-level semiconductor package, and is directed to some first The application of the stage and all second-level semiconductor packages is in the range of 15 111 111 to 400 μηη, wherein the application of the first stage is directly attached to the surface of the package substrate with respect to the semiconductor splicing bump, and the second The application of the grade is attached to the surface of the printed circuit board with respect to the entire package assembly. The length of the solder bump is measured by a scanning electron microscope (SEM), preferably using a scanning electron microscope JE〇LJSM_6〇6〇 SEM. The shape memory polymer of the support structure is a film having an average film thickness of 10 to 1200 μm. A support structure comprising a shape memory polymer film is used to support the wafer during back grinding. The shape memory polymer film protects protrusions on the surface of the wafer, such as bumps, during the back grinding process, wherein the thickness of the shape memory polymer film is generally at least higher than the height of the protrusions, such as bumps (space 17 201208071 The maximum length in any direction is twice as large as the protrusions and other protruding structures on the front side of the wafer are effectively protected. It has been found that the average film thickness of _41μιη, preferably is to 5〇0 μιη, and particularly preferably from 3〇 to 递, the protrusion of the wafer surface is effectively protected by the polishing process towel, and application _ Less than 9 pieces of N/mm. The stripping force' support structure can be substantially completely removed from the wafer surface during the stress free process. It should be noted that the support structure of the present invention may or may not contain at least one load bearing film. If the support structure comprises at least one carrier film, then the carrier film is attached to the inner surface of the support structure shape memory polymer. It is advantageous to use at least one support structure carrying a film in the method of the present invention because the carrier film reduces the peel force required to remove the support structure from the wafer surface. In addition, the carrier film imparts structural integrity and/or rigidity to the support structure and/or ensures re-use of the support structure. At least one carrier film of the present invention has ultraviolet light transmissivity and/or may comprise a polymer - selected from the group consisting of polyethylene (P〇lyethyleneS), polypropylenes, and polycarbonates (p〇iyearb) 〇nates), polygs, polyethylene terephthalate (polyethylene chlorides), polyethylene (ethylene) and ethylene acetate A combination of a polymer 'and/or its like. The carrier film may comprise or consist of one, two or more than two different layers, wherein each layer may comprise or consist of at least one of the foregoing polymers. It is desirable that the carrier film has an average thickness ranging from 50 to 200 μηη, preferably from 70 to 175 μηι, and more preferably from 90 to 140 μηη. The average film thickness of the carrier film is defined as the arithmetic mean of the thickness values of the plurality of films, wherein each film thickness value is at a substantially right angle to the longitudinally extending surface of the film bearing 201208071 film between the opposite opposite points. The inner surface of the film, and its corresponding chiseling & correcting (four) bearing W field 1 ί on the opposite outer surface of the film. Γ 多个 A plurality of "", for example, at least 10 different measurement points along 10 different positions of the support structure - longitudinal section. For example, the average thinness of the film can be, for example, the support structure ^ face (size 1 (10) X lem), and divided into 5 equal strip regions (each: 2 mm), and each The average impurity thickness of the film is determined along the longitudinal section. The thickness of the lining film along the slitting private material can be measured by scanning electron display (four) (s), and the preferred field is measured by scanning electron microscopy fine 0L JSM-6060 SEM. The present invention is directed to an assembly comprising a wafer and a support structure, wherein the combination is made in accordance with the method of the present invention. Another aspect of the present invention is an assembly comprising a wafer and a support structure, wherein the wafer has a plurality of protrusions on a surface thereof, and the support structure comprises at least one having an inner surface and An outer surface shape memory polymer film wherein the outer surface is attached to the protrusion (4) of the crystal, provided that the outer surface of the shape memory polymer film forms an outer surface of the support structure, and The outer surface exhibits a flatness index lower than 〇〇2 before being attached to the wafer, and the shape memory polymer film has an average film thickness of 1 〇 to 12 〇〇μπι. [Embodiment] Fig. 1 shows a cross-sectional view of a support structure used in the method of the present invention. The support structure comprises a shape memory polymer film 3 having an average film thickness ranging from 10 μm to 1200 μm. Figure 2 depicts an enlarged cross-sectional view of a portion of the support structure 1 shown in Figure 1. The shape memory polymer film 3 has a flatness index of less than 0.02, wherein the flatness index is defined as the ratio of the roughness of the outer surface of the support structure to the average film thickness of the shape memory polymer film 3. 201208071 Fig. 2 is an enlarged cross-sectional view showing a portion of the support structure shown in Fig. 2, and showing the roughness of the shape memory film 3. The outer surface is of course uneven and shows a special roughness. The roughness of the surface of the shape memory polymer film 3 is reflected on the cut surface of the shape memory polymer film 3, which is more or less slightly different in thickness. The thickness of this difference is reflected in the four-sided view of the support structure 1 shown in Fig. 3, which is in the form of a peak located on the outer surface/outer surface of the shape memory polymer film 3 / support structure 1 Pl, p2, p3 and trough %, V2, Vs, 乂4. The term "roughness" as used in the present invention refers to the arithmetic mean of the distance Rn and the valley Vn distance Dn in terms of a rough average value. The term "wave track" as used in the present invention refers to any convex field on the outer surface, and the reading "valley Vn" used in the present invention is any place between the county and the outer surface of the lion body. . The term "distance A" as used in the present invention refers to the difference between the height of the peak ^ and the adjacent trough Vn, which is measured at a substantially right angle to the longitudinal extension 2 of the shape memory film 3. . The rough seam can be used for the $-Solarius non-contact laser surface roughness 胄, which is equipped with an AF2〇〇〇 type automatic incline sensor, measured at a measuring speed of 〇 5 _/s. The measured data can be analyzed by _SGlarmapuniversal3 U (^ image analysis software (Gaussian filter G.8 mm), which uses micro-sweet sugar filtration, its critical 2.5 um 〇^, average film thickness of shape memory polymer film τ The average film thickness τ is an arithmetic mean of a plurality of film thickness values τη, wherein the 矣-thickness thickness value τη is a shape memory polymer film 3 opposite to the opposite polarity. The longitudinal extension 2 is measured in a substantially right-angled direction, wherein the point is located on the surface of the shape-recalling polymer film, and the impurity is on the opposite surface of the hetero-membrane. 丨〜口口物上上中The term "plurality" as used may, for example, refer to at least 10 different measurement points along 10 different longitudinal planes of the support structure 201208071. The average film thickness T can be determined, for example, by using the support structure. The entire surface of the body (size 1 cm X lcm) and divided into 5 equal strips (2 mm per area) and the average film thickness of the shape memory polymer film Along the longitudinal section The average film thickness along the longitudinal section can be determined by scanning electron microscopy (SEM), preferably using a scanning electron microscope jEOL JSM-6060 SEM. The following materials were used in the examples: Technically processed wafer wafer type: VPA 18 Sn3.oAgo.5Cu nitride passivation; Wafer small block 4 X 4 cm processed by bump technology, with bumps of 96 μm diameter length; bump pitch: 225 μιη; nominal divided chip size: 7 X 7 mm; slit width: 750 μηη. Shape memory polymer film (SMP film) The shape memory polymer film is made by UV curing method. A UV-curable composition comprising 61 wt.-% of a polymer of a polydimethyloxane type, 28 wt.-°/〇 of isobornyl acrylate, 10 wt.-% of a gas phase type II Cerium oxide, 1 wt.-% photoinitiator. A shape memory polymer film having an area of 4 x 4 cm, wherein the shape memory polymer film exhibits a glass transition temperature (Tg) of about 110. (:, It is measured by differential scanning calorimetry (DSC). 21 201208071 The load-bearing film is 4 X 4 cm in area and is made into a polyester film with an average film thickness of 125 for the carrier film. In the embodiment, the structure is not supported. The support structure is used in all cases. All include a shape memory polymer (SMp) film as described above, having an inner surface and an outer surface, wherein the filament surface of the clear memory polymer film forms an outer surface of the support structure. The roughness, average film thickness and flatness index of the memory polymer film are shown in Table 4. In some embodiments, the support structure also includes a carrier film wherein the carrier film is attached to the inner surface of the shape memory polymer film. Roughness Roughness is measured along a line of 〇 8 ^(7) on the outer surface of the support structure. The roughness shown in Table 1 is the arithmetic mean of the 15 different measurements measured at 15 different locations on the outer surface of the support structure. The roughness is based on a Solarius non-contact laser surface roughness meter equipped with an AF2〇〇〇 type autofocus sensor, measured at a measurement speed of 0.5 mm/s. The measured data is obtained using the image analysis software of solar map universal 3.1.10 (Gaussian filter 〇 8 (7) is called 'analysis' which uses micro-roughness filtration with a critical value of 2.5 μη. Average film thickness of shape memory polymer film The average film thickness τ of the shape memory polymer film is determined along the longitudinal section shown in Figure 3. The average film thickness Τ shown in Table 1 is along the longitudinal section 10 of the shape memory polymer film. The arithmetic mean of the 10 film thickness values Τη measured at different positions. The average film thickness along the longitudinal profile is measured by scanning electron microscopy (SEM) using a scanning electron microscope JEOL JSM-6060 SEM Flatness Index (FI) 22 201208071 Flatness and number are defined as the ratio of the thick chain of the outer surface of the support structure to the average film thickness of the shape memory polymer film. Forming a support structure and wafer The assembly-supporting structure and the wafer are combined so that the surface of the wafer treated by the bump technology is contacted at a pressure of 6 MPa and at a temperature of 13 〇t The outer surface of the support structure is formed for 15 minutes. Then, the assembly is cooled under a pressure of 6 MPa for less than 2 minutes to reach 22. (: The temperature. The removal of the support structure will support the structure and the crystal The round combination is exposed to 125. (: at a temperature of 2 minutes to activate the shape memory polymer film. The assembly of the support structure and the wafer is then cooled to 23 ° C and the release structure is applied to the support structure from the wafer. The surface is removed. The peeling force described in the surface is required to peel off the attached surface (the outer surface of the support structure and the wafer surface treated by the bump technology). 90. The peel force is measured at 23 ° C according to the American Society for Testing and Materials (ASTM) D6862-04 test method using a TXT plus tensile tester (obtained by Surrey 'Stable Micro Systems, UK) It is a load weight of 5 kg and a punch speed of 25 mm/min. [Brief Description of the Drawings] Preferred embodiments of the invention are illustrated by the following figures. Figure 1 shows a cross-sectional view of a support structure of the present invention. Figure 2 is shown in Figure 1. Fig. 3 shows another enlarged cross-sectional view of the portion of the support structure shown in Fig. 1. [Main element symbol description] 1 Support structure 2 Longitudinal extension 23 201208071 3 Shape Memory polymer film Pi ' P2 ' P3 wave peak Vi ' V2 ' v3 ' v4 wave valley Di ' D2 distance Ti ' T2 ' T3 thickness value 24 201208071 one 袼 1? give not 90 ° peel force [N / mm] 〇 - release Smooth 0-release smooth 0-release smooth 0-release smooth 0.052 carrying film cup r cup r no SMP flatness index FI 0.003 0.003 0.011 0.015 0.09 SMP average film thickness [μπι] 500 500 350 280 250 SMP roughness Im]_ 1.49 1.49 〇rn rn 寸 · Example 1 Example 2 Example 3 Example 4 Comparative Example. Jinle Tvg^s ton every time 钿 罐 罐 * * * * * "fv砸 «_铋冢 (¥嫁驷«^) Guanlan miv SO farm vad} 鹌骒 刼 刼 + v? * 金飨τνθ^ ® tons per 耷 ^ 択 ^ ^ 畹铋 a a - - - - - - - - - - - vi vi vi vi vi vi vi vi vi vi vi vi vi vi vi - - - - -