201133572 六、發明說明： 【發明所屬之技術領域】 本發明是關於一種於非單晶半導體膜上照射（重疊 (overlap)照射）多次線束形狀的脈衝雷射並使其移動: 進行結晶化的結晶半導體膜的製造方法。 【先前技術】 通常於電視（Television，TV)或個人電腦（pers〇nalBACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pulse laser that is irradiated (overlapped) on a non-single crystal semiconductor film by a plurality of strand shapes and moved: crystallization A method of producing a crystalline semiconductor film. [Prior Art] Usually on TV (Television, TV) or personal computer (pers〇nal
Computer，PC )顯示器中所使用的薄膜電晶體 ✓ 、” 卜 0 日 C amorphous )矽（以下稱為a-矽）構成，利用某些方法使 矽結晶化（以下稱為p-矽）而加以利用，藉此可使作為 臈電晶體（Thin Film Transistor，TFT )的性能明顯地提高’。 目前，作為低溫下的Si結晶化製程，準分子雷射退火^ 已得到實用化，赚地於面向行動電話等的小二 器的用途中，而且正在進行於大畫面顯示器等中的實用化。、 該雷射退火法為如下方法：藉由對非單晶半導體膜昭 脈?能量的準分子雷射，而使吸收光能量的半’導 體成為熔融解熔雜態，其彳纽急秘卻並 : 了處理寬廣之區域，而相對地於_方«^ 面純-_射整形躲束形狀的_雷射 ^ :設置有單晶半導體膜的設置台移動來進行脈衝雷= --於掃描方二二：專:：= 4 201133572 ^οου/ριι 此’可對大尺寸的半導體膜進行雷射退火處理。再者，於 專利文獻1中，雷射的依次操作所伴隨的結晶性的不均句 性（不均）成為產生元件間不均的原因，因此使脈衝雷射 的掃描方向上的通道區域的尺寸S、與脈衝雷射的掃描間 距P大致為S = nP (n為除〇以外的整數）’形成為結晶性 Si膜的結晶分佈在脈衝雷射光的掃描方向上週期性地變化 的圖案’使各薄膜電晶體的通道區域中的結晶性Si膜的結 晶性分佈的圖案的週期性變化相等。 並且，於先前的使用線束的雷射退火處理中，將脈衝 雷射的掃描方向的光束寬度固定在〇35 mm〜0.4 mm 左 右，並將每個脈衝對基板的發送量設定為光束寬度的3% 〜8%左右，為了確保多個薄膜電晶體的性能的均勻性而 認為必須儘可能地增加雷射的照射次數。 例如於液晶顯示器（Liquid Crystal Display，LCD )用 半導體膜中，將重疊率設定為92%〜95% (照射次數為i2 次〜20次，掃描間距為32以瓜〜如，於有機發光 二極體（Organic Light-Emitting Diode，OLED )半導體膜 中，將重疊率设疋為93.8°/。〜97% (照射次數為π次〜33 次’掃描間距為25 //m〜12 #π〇。 [先前技術文獻] [專利文獻] [專利文獻1]日本專利特開平1〇 163495號公報 然而，本發明者等人進行研究的結果發現，掃描間距 值越小’則雷射的照射次數越增加，但實際上如在規定條 201133572 件下照射次數為8次左右般，若照射次數到達某次數以 上，則結晶粒徑不會增大而飽和。即，即便使照射次數增 加至必需次數以上，亦無法有效地利用雷射輸出，並會導 致結晶化處理時間的增加。 另外’若使光束寬度增大到必需寬度以上，則由於雷 射脈衝能量固定’因此為了獲得規定的能量密度，必須縮 短線束長度’在處理大尺寸的半導體膜時’處理效率降低。 【發明内容】 _ 本案發明是以上述情況為背景而完成的，其目的在於 提供一種能夠適當地確定雷射脈衝的照射次數及脈衝寬度 而高效率地進行雷射退火處理的結晶半導體膜的製 法。 即，本發明的結晶半導體膜的製造方法於非單晶半導 體膜上相對地掃描並照射線束形狀的脈衝雷射而進行結晶 化；其特徵在於：上述脈衝雷射於掃描方向的光束剖面: 狀中具有強度均勻的平坦部（光束寬度a)，將藉由該脈衝 雷射照射而結晶化的半導體膜所形成的電晶體的上述掃描 方向的通道區域寬度設為b ; w 上述脈衝雷射具有照射脈衝能量密度E，該照射脈衝 月匕量密度E低於藉由該脈衝雷射照射而於上述非單晶半導 體獏上產生微結晶化的照射脈衝能量密度； 曰將藉由上述照射脈衝能量密度E的脈衝雷射照射而結 曰曰粒徑成長達到飽和時的照射次數設為n0，使脈衝雷射的 照射次數η為（n〇—1)以上； 6 201133572 jjou/pii 動量射的上述掃财向上的每舰衝的移 中具射如上舰，料财㈣絲剖面形狀 聂大处-:的平坦部（光束寬度a)。該平坦部相對於 最大月巧強度，可利用9G%以上之區域表示。 ^ 〇 π否產生微結晶化可藉自電子顯微賴片等來判 若，照射脈衝能量密度的值A於產生微結晶化的 度達到Γ/t會變得極小，而使作為半導體的電子移動 而2粒卜枰由I照射脈衝能量密度E的脈衝雷射照射 射態是指各粒徑-致且即便增加照 成县若雷射照射次數未達（η0—υ，則結晶粒役的 ’不同粒徑的結晶混合存在’而產生電子移動 又 '纟。根據相同的理由，較理想岐η0以上。 更理想的是2·η〇^顯降低。而且，根據相同的理由， 若將藉由上述脈衝雷射照射而結晶化的半導體膜上 201133572 脈衝雷射⑽向的通道輯寬度料b，則使 藉此，各通即母個脈衝的移動量C為b/2以下。 上，可減射脈衝接縫為2條或3條以 於^，且為b ^能不均。另一方面，若移動量c大 條，若移動奮士則通道區域中的上述接縫為1條或2 或1條，通’則通道區域中的上述接縫為0條 二通道£域中的電晶體的性能不均增大。 衝雷射次數η及每個脈衝的移動量C，脈 以下。若過於增大光束寬度，則於使能量 小^因此处二Li ΐ衝雷射的長抽方向上的光束長度變 降低。此以一次掃描進行處理的面積變小，處理效率 並且’較理想的是脈衝雷射掃描方向的通道區域寬度 可維r常·ΙΓ右縮小電晶體的區域寬度、即電晶體，則 而^在電晶體中流通的時間，提高訊號處理速度’ 而可獲侍性能優異的薄膜半導體。 =本發明的處理對象的半導體並不限定於規定材 w Si作為較佳的材料。並且，脈衝雷射可舉出準 分子雷射作為較佳的雷射。 [發明效果] 如以上所說明’根據本發明的結晶半導體膜的製造方 法’其是於非單晶半導體膜上相對地掃描並照射線束形狀 的脈衝雷射而進行結晶化的結晶半導麵的製造方法；其 8 201133572 i pij. 中 膜所形成的電晶趙的上述掃描方向的通道 ^述脈衝雷射具有照射脈衝能量密度e，該 此量密度E低於藉由紐衝雷龜射而於上述曰 體膜上產生微結晶化的照射脈衝能量密度； 曰曰 將藉由上述照射脈衝能量密度E的脈衝雷射昭 長ST7時的照射次數設為n〇，使脈衝雷射： …、射久數η為（n0 — 1 )以上· 動署H脈衝雷射的上述掃描方向上的每個脈衝的移 c ‘、、、以下，因此能夠藉由適當的脈衝雷射昭射夂 ，及每個脈衝的移動量而高效率地進行雷射退火處理'。另 衝雷射ίΐΐ寬度設為適當的值，可獲得充分的 、’’長又而且/、有忐夠進行高效率的處理的效果。 為讓本發明之上述特徵和優點能更明顯易懂 舉實施例，並配合所附圖式作詳細說明如下。 ， 【實施方式】 以下，對本發明的一個實施形態進行說明。 圖1表示㈣置於軸台i上的基板騎包含線束狀 ^子雷射的脈衝雷射3的狀態。基板上形成有非晶^ 等非早晶半導舰2。脈衝#射3具有線束長度l及光束 寬度a ’藉由使移動台i以規㈣距移動，而使脈衝雷射3 201133572 /plf 2進1掃插並崎定間距及照射錢照射於鱗晶半導_ 的平坦部，將該平坦部的寬度麵為光束寬度a。強度 — ft脈衝雷射3在照射至非單晶半導體膜2上時， 該非單晶半導體膜2不會微結晶化的照射脈衝能量 表7^照射脈衝能量密度無雷祕衝照射所引 =。粒#大小的關係圖。在照射脈衝能量密度低的區 ，中，隨著照射脈衝能量密度增大而結晶粒徑變大。例如 右·、、、射脈衝此量密度大於此中途的照射脈衝能量密度 E卜則結晶粒徑急遽變大。另—方面，若照射脈衝能量密 度增大至某種程度，則即便照射脈衝能量密度增大至其以 上，結晶粒徑亦幾乎不增大，若超過某照射脈衝能量密度 E2,則結晶粒徑急遽變小而產生微結晶化。因此上述照射 脈衝能量密度E可利用ESE2表示。 〇將照射脈衝能量密度設定為上述E之值，並照射至非 單晶半導體臈2時，即便將照射次數設定為某次數以上， 結晶粒徑成長亦會達到飽和。結晶粒徑成長的飽和根據掃 •fe 式電子顯微鏡（Scanning Electron Microscopy ’ SEM)照 片來判定。 圖4是表示在將照射脈衝能量密度£設定為上述照射 脈衝能量密度E1或照射脈衝能量密度E2的情況下，結晶 201133572 ί==::=Τ-照射脈衝能量密 晶粒徑變大，但若達到某照射次數， 照射次數以上時不再進行而達到飽和: 為本發明的照射次數η0。 〇.、、、射-人數表不 相對於上述騎次數nG，設定為㈤ 有效且高效率^^。。藉此’能夠使非單晶半導體膜2 膜中於射照射而結晶化的結晶化半導體 別膜半導體。在薄膜半導體中，分 訂有規疋的通道區域寬度b，該間隔較佳為設定為 Μ卜0 n、、、a43；1。）圖5 (c)。各薄膜半導體10具有源極 η㈣紙♦、及位於源極、汲極間的通道部13，該通道部 上if二日f的掃描方向寬度成為通道區域寬度b。若對 昭i r ::體膜2以掃描間距(每個脈衝的移動量)c 丄射3並使其移動，則對應於每個脈衝的移動而 於結晶化半導體膜上出現光束接縫3a。 (&)表不使每個脈衝的移動量c大於上述通道區 b _缺接縫％的產生狀況。於此例中光束接 的性部13或出現1條，使薄膜半導體1〇 5 (b)表示使每個脈衝的移動量c大於上述通道區 201133572 域寬度b的1/2時的光束接縫3a的產生狀況。於此例中， 光束接縫3a在通道部13中出現1條或2條，雖然薄膜半 導體10的性能不均減少，但並未充分減少。 圖5 (c)是本發明中規定的圖，表示使每個脈衝的移 動量c為上述通道區域寬度的1/2以下時的光束接縫3a的 產生狀況。於此例中，光束接縫3a在通道部13中出現2 條或3條’而有效地減少薄膜半導體1〇的性能不均。 於上述每個脈衝的移動量c中，在將照射次數設定為 人的清況下，光束寬度a是以a=n.c表示。藉由上述設 疋可將母個脈衝的移動量c設定為較小，另外，照射次數 亦為可良好地進行結晶化的次數，而不會增多到必需次數 以上。其結果可使光束寬度減小至例如5〇〇 以下，其 、、、=»果可增大光束長度而高效率地處理大尺寸的非單晶半導 體膜。 [實例1] 以下’對本發明的一個實例進行說明。 將50 mn厚的非晶si作為非單晶半導體膜，按以下條 件改變照射次數並照射脈衝雷射。 準分子雷射：LSX315C/波長308 nm、頻率3〇〇Hz 光束尺寸：光束長度500 mmx光束寬度〇 13 mm 光束寬度為最大能量強度的90%以上的平坦部 知*描間距.32.5 //m〜6.5 //m 照射脈衝能量密度：320 mJ/cm2 通道區域寬度：40 12 20113357¾ 於上述脈衝雷射中，照射脈衝能量密度達到產生微結 晶的照射脈衝能量密度以下，確認自照射次數4次至照射 次數8次為止結晶粒徑逐漸成長，但照射次數8次以後， 結晶粒控成長達到飽和。 藉由SEM照片對以規定照射次數照射脈衝雷射的部 位進行觀察，將該照片示於圖6。如圖6所示，在照射次 數8次時’良好地結晶化’即便將照射次數增加至12、16、 20次時，亦幾乎未發現結晶粒徑增加。 圖7是表示對應於照射次數的結晶粒徑的變化圖，直 至達到照射次數8次為止，結晶粒徑隨著照射次數的增加 而增大。在照射次數8次以後，未發現結晶粒徑增大。 雖…:本發明已以實施例揭露如上，然其並非用以限定 本發明，任何所屬技術領域中具有通常知識者， ，之精神和範圍内，當可作些許之更動與潤 ,明之保護範圍當視後附之申請專利範圍所界定者為準。 【圖式簡單說明】 體膜圖個實施形態中_㈣半導 的圖圖2陳是表雜衝雷射的掃描方⑽光束剖面形狀 衝雷：密度與-度二 201133572 圖5 (a)〜圖5 (c)同樣是表示每個脈衝的移動量 與通道區域寬度的關係中光束接縫的產生狀況的圖。 圖6是表示本發明的一個實施例中的結晶化半導體的 圖式代用照片。 圖7同樣是表示粒徑變化相對於照射次數的關係的圖 表。 【主要元件符號說明】 I :移動台 2:非單晶半導體膜 3:脈衝雷射 3a :光束接縫 10 :薄膜半導體 II :源極 12 :汲極 13 :通道部 a:光束寬度 b:通道區域寬度 c:每個脈衝的移動量 L :線束長度 14Computer, PC) The thin film transistor used in the display, "C amorphous" (hereinafter referred to as a-矽), which is crystallized by some methods (hereinafter referred to as p-矽). By utilizing this, the performance as a Thin Film Transistor (TFT) can be significantly improved. At present, as a Si crystallization process at a low temperature, excimer laser annealing ^ has been put into practical use, earning a face In the use of a small device such as a mobile phone, it is being put into practical use in a large-screen display or the like. The laser annealing method is a method in which a quasi-molecular ray is excited by a non-single-crystal semiconductor film. Shooting, so that the half-conductor that absorbs the light energy becomes a melting and demelting impurity, and its 彳 急 急 : : : : : : : : : : : : : : : : : : 处理 处理 处理 处理 处理 处理 处理 处理 处理 处理 处理 处理 处理 处理 处理 处理 处理 处理_Laser^: Set the stage with a single crystal semiconductor film to move to pulse Ray = - on the scanning side 22:Special::= 4 201133572 ^οου/ριι This can be used to laser the large size semiconductor film Annealing treatment. Furthermore, in Patent Document 1 In the middle, the unevenness of the crystallinity (unevenness) accompanying the sequential operation of the laser causes the unevenness between the elements, so that the size S of the channel region in the scanning direction of the pulse laser and the pulse laser are made. The scanning pitch P is approximately S = nP (n is an integer other than )) 'formed as a pattern in which the crystal distribution of the crystalline Si film periodically changes in the scanning direction of the pulsed laser light' to make the channels of the respective thin film transistors The periodic variation of the pattern of the crystallinity distribution of the crystalline Si film in the region is equal. Further, in the laser annealing treatment using the wire harness, the beam width in the scanning direction of the pulse laser is fixed at 〇35 mm to 0.4. The distance between each pulse and the substrate is set to about 3% to 8% of the beam width. In order to ensure the uniformity of the performance of the plurality of thin film transistors, it is considered that the number of times of laser irradiation must be increased as much as possible. For example, in a semiconductor film for a liquid crystal display (LCD), the overlap ratio is set to 92% to 95% (the number of times of irradiation is i2 times to 20 times, and the scanning pitch is 32 to melons, for example, In the Organic Light-Emitting Diode (OLED) semiconductor film, the overlap ratio is set to 93.8°/~97% (the number of times of irradiation is π times to 33 times) and the scanning pitch is 25 //m~12 # 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 As the number of times increases, the number of times of irradiation is about 8 times in the predetermined section 201133572. When the number of times of irradiation reaches a certain number of times or more, the crystal grain size does not increase and is saturated. That is, even if the number of times of irradiation is increased to a required number or more, the laser output cannot be effectively utilized, and the crystallization processing time is increased. Further, if the beam width is increased to a necessary width or more, since the laser pulse energy is fixed ', in order to obtain a predetermined energy density, it is necessary to shorten the wire harness length' when processing a large-sized semiconductor film, and the processing efficiency is lowered. SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and an object of the invention is to provide a method for producing a crystalline semiconductor film which can efficiently perform laser annealing treatment by appropriately determining the number of times of irradiation of a laser pulse and the pulse width. . In other words, the method for producing a crystalline semiconductor film of the present invention is performed by relatively scanning a non-single-crystal semiconductor film and irradiating a pulsed laser beam of a linear beam shape to perform crystallization; wherein the pulse beam is laser beam in the scanning direction: a flat portion having a uniform intensity (beam width a), and a width of a channel region in the scanning direction of the transistor formed by the semiconductor film crystallized by the pulsed laser irradiation is set to b; w Irradiating the pulse energy density E, the irradiation pulse monthly density E is lower than the irradiation pulse energy density generated by the pulsed laser irradiation on the non-single-crystal semiconductor germanium; 曰 will be irradiated by the above-mentioned irradiation pulse energy When the pulsed laser irradiation of the density E is irradiated, the number of times of irradiation when the crucible particle size is saturated is set to n0, and the number of irradiations η of the pulsed laser is (n〇-1) or more; 6 201133572 jjou/pii Each ship rushing in the sweeping direction has the same ship as the ship, and the grain (4) wire section shape Nie Da-- the flat part (beam width a). The flat portion can be expressed by an area of 9 G% or more with respect to the maximum monthly strength. ^ 〇 π No microcrystallization can be judged by electron microscopy, etc., and the value A of the irradiation pulse energy density becomes extremely small when the degree of microcrystallization becomes Γ/t, and the electrons as semiconductors become extremely small. The pulsed laser illuminating oscillating state in which the pulse energy density E is irradiated by I is two particle diameters, and even if the number of laser irradiations in the illuminating county is not reached (η0-υ, the crystal granules are 'The crystals of different particle sizes are mixed and there is electron movement and '纟. For the same reason, it is more desirable than 岐η0. More preferably, 2·η〇^ is reduced. Moreover, for the same reason, if it will be borrowed The channel width width b of the 201133572 pulsed laser (10) on the semiconductor film crystallized by the pulsed laser irradiation is such that the movement amount C of each of the mother pulses is b/2 or less. The reduction pulse joint is 2 or 3 for ^, and b ^ can be uneven. On the other hand, if the movement amount c is large, if the movement is slow, the seam in the passage area is 1 or 2 or 1, the above joint in the channel area is 0, 2 channels, electricity in the £ domain The performance of the body is not uniform. The number of shots η and the amount of movement of each pulse C, below the pulse. If the beam width is too large, the energy is made small, so the long direction of the Li rushing laser is taken. The length of the beam on the upper side is reduced. The area processed by one scan becomes smaller, and the processing efficiency and 'preferably the width of the channel region in the pulsed laser scanning direction can be narrowed to the right. In the case of the transistor, the time during which the transistor is circulated, the signal processing speed is increased, and the thin film semiconductor having excellent performance can be obtained. The semiconductor to be processed according to the present invention is not limited to the predetermined material w Si as a preferable material. Further, the pulsed laser may be an excimer laser as a preferred laser. [Effect of the Invention] As described above, the method for producing a crystalline semiconductor film according to the present invention is a relatively non-single-crystal semiconductor film. A method for producing a crystal semi-conducting surface which is scanned and irradiated with a pulsed laser beam in the shape of a wire harness; 8 201133572 i pij. a channel of the above-mentioned scanning direction of the electro-crystal Zhao formed by the film ^ The pulsed laser has an irradiation pulse energy density e, which is lower than an energy density of an irradiation pulse which is microcrystallized on the above-mentioned body film by a Newton tortoise; the energy of the irradiation pulse is The number of times of irradiation of the pulse laser of the density E is set to n〇, so that the pulse laser: ..., the number of shots η is (n0 - 1) or more, and each of the scanning directions of the H-pulse laser is emitted. Since the pulses are shifted by c ', , and below, the laser annealing treatment can be efficiently performed by an appropriate pulsed laser illuminating 夂 and the amount of movement of each pulse. Appropriate values provide sufficient, long-lasting, and/or effective, high-efficiency treatments. The above-described features and advantages of the present invention will be more apparent and understood. The formula is described in detail below. [Embodiment] Hereinafter, an embodiment of the present invention will be described. Fig. 1 shows a state in which (4) a substrate placed on the stage i rides a pulsed laser 3 including a beam-like sub-laser. A non-earthing semi-conducting ship 2 such as an amorphous metal is formed on the substrate. Pulse #射3 has a harness length l and a beam width a' by moving the mobile station i at a (four) distance, so that the pulse laser 3 201133572 /plf 2 into 1 sweep and the spacing and illumination money are irradiated to the squam The flat portion of the semiconducting_, the width surface of the flat portion is the beam width a. Intensity - When the ft pulsed laser 3 is irradiated onto the non-single-crystal semiconductor film 2, the non-single-crystal semiconductor film 2 is not microcrystallized by the irradiation pulse energy. Diagram of the size of the grain #. In the region where the irradiation pulse energy density is low, the crystal grain size becomes larger as the irradiation pulse energy density increases. For example, the right-,, and-shot pulses have a density greater than the energy density of the irradiation pulse in the middle of the pulse, and the crystal grain size becomes sharper and larger. On the other hand, if the energy density of the irradiation pulse is increased to some extent, the crystal grain size hardly increases even if the energy density of the irradiation pulse is increased above it, and if it exceeds the energy density E2 of the irradiation pulse, the crystal grain size is The rush is small and microcrystallized. Therefore, the above-mentioned irradiation pulse energy density E can be expressed by ESE2. When the irradiation pulse energy density is set to the value of the above E and is irradiated to the non-single crystal semiconductor 臈2, even if the number of times of irradiation is set to a certain number or more, the crystal grain size growth is saturated. The saturation of the crystal grain size growth was judged by a scanning electron microscope (Scanning Electron Microscopy' SEM). 4 is a view showing that when the irradiation pulse energy density is set to the above-described irradiation pulse energy density E1 or irradiation pulse energy density E2, the crystal 201133572 ί==::=Τ-irradiation pulse energy dense crystal grain size becomes large, but If the number of times of irradiation is reached, the number of times of irradiation is no longer performed and the saturation is reached: the number of irradiations η0 of the present invention. 〇.,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, . Thereby, the crystallized semiconductor film semiconductor which can be crystallized by irradiation in the film of the non-single-crystal semiconductor film 2 can be used. In the thin film semiconductor, a prescribed channel region width b is defined, and the interval is preferably set to 0 n, , a43; ) Figure 5 (c). Each of the thin film semiconductors 10 has a source η (tetra) paper ♦ and a channel portion 13 between the source and the drain, and the width of the scanning direction of the if two days f on the channel portion becomes the channel region width b. When the scanning aperture (the amount of movement per pulse) c is shot and moved by the scanning aperture (the amount of movement per pulse) c, the beam seam 3a appears on the crystallized semiconductor film corresponding to the movement of each pulse. The (&) table does not cause the movement amount c of each pulse to be larger than the generation condition of the above-mentioned passage area b__ seam %. In this example, the beam-connecting portion 13 or one strip is present, so that the thin film semiconductor 1〇5 (b) represents the beam seam when the amount of movement c of each pulse is larger than 1/2 of the width b of the region of the channel region 201133572. The production status of 3a. In this example, one or two of the beam seams 3a appear in the channel portion 13, and although the performance unevenness of the film semiconductor 10 is reduced, it is not sufficiently reduced. Fig. 5 (c) is a view defined in the present invention, showing a state in which the beam joint 3a is generated when the movement amount c of each pulse is 1/2 or less of the width of the channel region. In this example, the beam seam 3a appears in the channel portion 13 by two or three strips' to effectively reduce the performance unevenness of the thin film semiconductor. In the above-described movement amount c of each pulse, the beam width a is represented by a = n. c in the case where the number of irradiations is set to be human. According to the above configuration, the amount of movement c of the mother pulse can be set small, and the number of times of irradiation can be a number of times of crystallization which can be favorably performed, without increasing the number of times required. As a result, the beam width can be reduced to, for example, 5 Å or less, and the length of the beam can be increased to efficiently process the large-sized non-single-crystal semiconductor film. [Example 1] Hereinafter, an example of the present invention will be described. A 50 mn thick amorphous Si was used as the non-single crystal semiconductor film, and the number of times of irradiation was changed and the pulsed laser was irradiated as follows. Excimer laser: LSX315C / wavelength 308 nm, frequency 3 Hz Beam size: beam length 500 mmx beam width 〇 13 mm Beam width is more than 90% of the maximum energy intensity of the flat section * * spacing * 32.5 / m ~6.5 //m Irradiation pulse energy density: 320 mJ/cm2 Channel area width: 40 12 201133573⁄4 In the above-mentioned pulsed laser, the energy density of the irradiation pulse is below the energy density of the irradiation pulse which produces microcrystals, and it is confirmed that the number of irradiations is 4 times to The crystal grain size gradually increased until the number of irradiations was 8 times, but after the number of irradiations was 8 times, the crystal grain growth growth was saturated. The portion irradiated with the pulse laser at a predetermined number of irradiations was observed by an SEM photograph, and the photograph is shown in Fig. 6. As shown in Fig. 6, when the number of irradiations was increased to 12, 16 or 20 times, the crystal grain size was hardly increased even when the number of irradiations was 8 times. Fig. 7 is a graph showing changes in crystal grain size corresponding to the number of times of irradiation until the number of times of irradiation reaches 8 times, and the crystal grain size increases as the number of times of irradiation increases. After 8 times of irradiation, no increase in crystal grain size was observed. The present invention has been disclosed in the above embodiments, but it is not intended to limit the present invention. Any one of ordinary skill in the art, within the spirit and scope, may make some changes and protections. This is subject to the definition of the scope of the patent application. [Simple description of the figure] In the embodiment of the body film diagram, the image of the _(four) semi-conductor is shown in Fig. 2. The scanning side of the surface is laser beam (10) The beam profile shape is washed: density and - degree II 201133572 Figure 5 (a) ~ Fig. 5(c) is a view similarly showing the state of generation of the beam joint in the relationship between the amount of movement of each pulse and the width of the channel region. Fig. 6 is a schematic view showing a substitute of a crystallized semiconductor in an embodiment of the present invention. Fig. 7 is also a graph showing the relationship between the change in particle diameter and the number of times of irradiation. [Description of main component symbols] I: Mobile station 2: Non-single crystal semiconductor film 3: Pulse laser 3a: Beam joint 10: Thin film semiconductor II: Source 12: Deuterium 13: Channel portion a: Beam width b: Channel Area width c: movement amount per pulse L: harness length 14