201010260 六、發明說明： 【發明所屬之技術領域】 本發明，係有關於對於電漿以及表面處理裝置而以雙 極脈衝狀來進行電力供給之雙極脈衝電源、以及將此雙極 脈衝電源複數台並聯連接所成之電源裝置。 【先前技術】 • 此種雙極脈衝電源，例如係被使用在對於處理基板表 面而形成特定之薄膜的濺鍍裝置中，並週知有：具備供給 . 直流電力之整流電路、和被連接於此整流電路之正負的輸 - 出端處，並由4個的切換元件所成之MOSFET橋接電路者。 , 而後，經由控制手段來使各切換元件適宜動作，並對於身 爲輸出端（電極）之一對的標靶，而以特定之頻率來交互 地切換極性並施加任意之脈衝電壓，而將各標靶分別交互 地切換爲陽極電極、陰極電極，而在陽極電極以及陰極電 Φ 極之間使輝光放電產生，並形成電漿氛圍，而對各標靶作 濺鍍。藉由此，積蓄在標靶表面之電荷，係在施加有相反 之相位電壓時被抵消，而具有能夠得到安定之放電的優點 (例如，專利文獻1 )。 在此種輝光放電中，係週知有會由於某種之原因而產 生電弧放電（異常放電）的事態。若是產生有電弧放電， 則由於電漿（負載）之阻抗係急遽地變小，因此，會引起 急遽之電壓降低，伴隨於此，電流係增加。於此，特別是 當標靶爲鋁等之金屬製的情況時，若是在標靶間局部性地 -5- 201010260 產生有高電弧電流値之電弧放電，則係會產生標靶之被溶 融且放出的部分附著在處理基板表面上之所謂的粒子或是 噴濺（數〜數百//in之塊），而無法進行良好的成膜 〇 因此，在上述之雙極脈衝電源中，係設置有檢測出從 橋接電路而來之輸出電流的檢測電路、和對電弧放電產生 時之電流上升作抑制的電感，當藉由此檢測電路所檢測出 之輸出電流超過了定常輸出電流値時，則對動作中之切換 元件作切換，並將對於該電極之輸出暫時遮斷。而後，若 是過電流鎭靜下來而其之値成爲接近於定常輸出電流値， 則係再度開始對於該電極之輸出。於此情況，若是輸出電 流超過一定之範圍並變化，則係作爲異常放電之前段現象 (微電弧）而掌握，並藉由進行其之消弧處理，而亦可對 於電流變化量爲多之電弧放電的發生作抑制。 [專利文獻1]日本專利第3639605號公報 【發明內容】 [發明所欲解決之課題] 然而，一般而言，由於從直流電力供給源而來之輸出 係具備有定電壓特性’因此’相較於電感成分’電容成分 (capacitance)係成爲更加具有支配性。因此，在電弧放 電發生時，由於電漿負載側之阻抗係變小（依存於情況， 亦會變小至數歐姆以下），因此，輸出與電漿（負載）係 被耦合，並從電容成分而急遽地被放出至輸出側。其結果 -6- 201010260 ，就算是設置電感値爲小之電感’亦無法對電流上升有效 率地作抑制，而有著在短時間（數之間）而流動過電 流（亦即是，電弧放電發生時之每單位時間的電流上升率 爲高）的問題。 當每單位時間之電流上升率爲高的情況時，就算是在 掌握到電流變化量爲較小之狀態並進行微電弧處理時’亦 會有在從根據較電壓變化更遲而發生之電流變化來檢測出 φ 電弧放電起直到將對於電漿之電力供給遮斷爲止的時間內 而流動有大的電弧電流之情況，而使所放出之電弧能量變 大（流動有定常電流値之2倍左右的電流），並使噴濺或 - 是粒子成爲容易發生，特別是，當連續地發生有電弧放電 _ 時，係無法將噴濺或是粒子的發生實質上的作抑制。 因此，本發明之目的，係在於提供一種：能夠對電弧 放電發生時之電流上升作有效的限制，而能夠對於噴濺（ splash )或是粒子之發生作抑制的雙極脈衝電源及將此雙 Ο 極脈衝電源複數台並聯連接所成之電源裝置。 [專利文獻1]日本專利第3 639605號公報（例如，參考 申請專利範圍第1項、段落號碼〇〇 16之記載） [用以解決課題之手段] 爲了解決上述課題，申請專利範圍第1項中所記載之 雙極脈衝電源，係具備有由被連接於從直流電力供給源而 來之正負之直流輸出間的切換元件所構成的橋接電路、和 對橋接電路之各切換元件之ON· OFF的切換作控制之控制 201010260 手段，並對於與電漿接觸之一對的電極而以特定之頻率來 進行雙極脈衝狀之電力供給，其特徵爲：在從前述直流電 力供給源而對橋接電路之正負之直流輸出中的至少其中一 方處，設置有具備ImH之以上之値的電感。 若藉由本發明，則在從直流電力供給源而對於橋接電 路供給直流電力的狀態下，若是經由控制手段，而使構成 橋接電路之切換元件中的對於其中一方之電極作輸出之2 個的切換元件成爲ON，則係對於其中一方之電極進行電 力供給（輸出）。接著，若是在將正對於其中一方之電極 進行輸出的切換元件設爲OFF的同時，將對於另外一方之 電極進行輸出之2個的切換元件設爲ON，則係對於另外一 方之電極進行輸出。藉由反覆進行此控制，而對於與電漿 接觸之一對的電極來以特定之頻率而進行雙極脈衝狀之電 力供給。 而，當由於某些之原因而發生有電弧放電的情況時， 由於電漿之阻抗係急遽地變小，因此會引起急遽之電壓降 低，並伴隨於此而使電流增加。此時，由於係在從直流電 力供給源而對橋接電路之正負之直流輸出中的至少其中一 方處，設置有具備ImH之以上之値的電感28，因此，從直 流電源供給源而來之輸出，係成爲定電流特性，其結果， 電弧放電發生時之每單位時間的電流上升率係被限制。 於此，電弧放電發生時之每單位時間的輸出電流上升 率（Δί)，若是將電感之電感値設爲L，並將對於電極之 輸出電壓設爲V，將電流變化時間設爲At，則係可藉由 201010260 △ i= Δίχν/Ι：而算出。於此情況，在一般之使用於量產 中的電鍍裝置中，由於對於標靶之輸出電力係至少爲5kW 以上，因此，爲了將每單位時間（例如，1 〇〇 // s )之電弧 放電發生時之輸出電流上升率，抑制爲較定常電流値之 20 0%爲更小、更理想係抑制在150%以下，電感係需要具 備有ImH以上之電感値。 然而，當如同上述一般而設置有具備ImH以上的値之 _ 電感的情況時，在對橋接電路之各切換元件作切換時，係 會有產生較通常之放電電壓更高之電壓的情況。亦即是， 由於在電漿中產生有電感成分，因此，在各電極處之極性 - 反轉時，會產生過電壓。由於若是如此這般地產生有過電 壓，則會有誘發電弧放電之虞，因此，若是爲更進而具備 有：輸出箝制電路，其係具備被與從前述直流電力供給源 而來之正負的直流輸出作了並聯連接之電容、和與前述電 感成爲並聯且相互被作了串聯連接之二極體以及電阻，則 • 在極性反轉時，其起初之對於電極的輸出係成爲定電壓特 性，而輸出電流係成爲逐漸地增加，之後（若是輸出電流 達到了特定値），對於電極之輸出係成爲定電流特性。其 結果，在各電極之極性反轉時而產生過電壓一事係被防止 ，而起因於過電流之弧狀放電的發生係被抑制。 另外，前述電極，係爲配置在實施濺鍍法之處理室內 的一對之標靶。 又，若採用下述構成，則爲理想：具備有：輸出短路 用之切換元件，係被設置在從前述直流電力供給源而至橋 201010260 接電路之正負之直流輸出間；和檢測手段，係檢測出前述 一對之電極間的輸出電流；和異常放電檢測手段，係若是 此輸出電流之絕對値超過對於電極之定常輸出電流値，則 將其作爲異常放電發生之前段現象而掌握，若是經由此異 常放電檢測手段而掌握到有異常放電發生之前段現象，貝!1 經由前述輸出短路用之切換元件，來將對於電極之輸出遮 斷，而進行異常放電之消弧處理。 若藉由此，則在進行雙極脈衝狀之電力供給時，不僅 Q 是能夠使切換損失僅在1個的輸出短路用之切換元件處產 生，且相較於對於輸出中之2個的切換元件作控制並進行 異常放電之消弧處理的情況，能夠以良好回應性來進行該 _ 控制，並且，在此處理中，於橋接電路之各切換元件處， 亦幾乎不會產生切換損失，而能夠將其之耐久性提昇。 進而，爲了解決上述課題，申請專利範圍第5項所記 載之電源裝置，係爲將如申請專利範圍第4項所記載之雙 極脈衝電源作了複數台並聯連接的電源裝置，其特徵爲， @ 具備有：統籌控制手段，係當對於被配置在同一之處理室 內的複數對之電極進行雙極脈衝狀之電力供給時，對於各 雙極脈衝電源之輸出短路用的切換元件之ON· OFF的切換 作控制。 若藉由本發明，則由於只要經由統籌控制手段來將各 雙極脈衝電源之各輸出短路用的切換元件作同步即可，因 此，能夠具有充分餘裕地來使橋接電路之切換元件動作， 就算是在各雙極脈衝電源之切換元件或控制電路中存在有 -10- 201010260 切換速度或是控制速度的個體差異，其之同步運轉亦係爲 容易。 [發明之效果] 如同以上所說明一般，在本發明之雙極脈衝電源以及 電源裝置中，電弧放電發生時之電流上升係被有效的限制 ，其結果，能夠對於噴濺（splash )或是粒子之發生有效 # 的作抑制，而可得到成爲能夠進行良好之薄膜形成的效果 【實施方式】 . 參考圖1，E，係爲本發明之雙極脈衝電源，雙極脈衝 電源E，例如係與濺鍍裝置內之處理基板相對向地而被配 置，並用以對於身爲與電漿P相接觸之電極的一對之標靶 ΤΙ、T2而以特定之頻率來進行雙極脈衝狀之電力供給而被 • 使用。雙極脈衝電源E，係由使直流電力之供給成爲可能 的直流電力供給部1、和對於對各標靶ΤΙ、T2之輸出（電 力供給）作控制的震盪部2所構成。於此情況，輸出電壓 之波形，係爲略方形波或是略正弦波。 直流電力供給部1，係具備有··對其之動作進行控制 之第1CPU電路11、和被輸入有商用之交流電力（3相 AC200V或是400V)的輸入部12、和將所輸入之交流電力 作整流並變換爲直流電力之由6個的二極體13 a所成的整流 電路13，並經由正負之直流電力線14a、14b來將直流電力 -11 - 201010260 輸出至震盪部2處。又，在直流電力供給部1中’係被設置 有··被設置在直流電力線14a、14b之間之切換電晶體（切 換元件）15、和被可自由通訊地連接於第1CPU電路11處 ，並對切換電晶體15之ON· OFF作控制的輸出震盪用之驅 動電路16。在直流電力線14a、14b之間’係被設置有檢測 出其之電流、電壓的檢測電路1 7a，藉由檢測電路1 7a所檢 測出之電流、電壓，係成爲經由AD變換電路17b而被輸入 至第1CPU電路1 1處。 另一方面，在震盪部2處，係設置有：可自由通訊地 被連接於第1CPU電路11處之第2CPU電路21、和被連接在 正負之直流電力線14a、14b之間的由4個的第1乃至第4切 換電晶體SW1乃至SW4所成之橋接電路22、和可自由通訊 地被連接於第2CPU電路21，並對各切換電晶體SW1乃至 SW4之ON. OFF之切換進行控制的輸出震盪用之驅動電路 23 ° 而後，若是經由輸出震盪用之驅動電路23，而例如以 使第1以及第4切換電晶體SW1、SW4和第2以及第3切換電 晶體SW2、SW3之ON· OFF的時機反轉的方式，而對各切 換電晶體SW1乃至SW4之切換進行控制，則能夠經由從橋 接電路22而來之輸出電力線24a、2 4b，來對一對之標靶T1 、T2進行雙極脈衝狀之電力供給。在輸出線24a、24b處， 係被連接有檢測出對於一對之標靶ΤΙ、T2的输出電流以及 輸出電壓之檢測電路25，藉由此檢測電路25所檢測出之輸 出電流以及輸出電壓，係成爲經由AD變換電路26而被輸 201010260 入至第2CPU電路21處。 於此，在上述構成之雙極脈衝電源E中，若是於從直 流電力供給部1而輸出有直流電力的狀態下來對各切換電 晶體SW1乃至SW4作切換，則由於該些之切換損耗係成爲 相當大，因此，係有必要以使各切換電晶體SW1乃至SW4 之耐久性提昇的方式來作構成。在本實施形態中，係構成 爲：在從直流電力供給部1而來之正負的直流輸出線14a、 φ 14b之間，設置經由輸出震盪用之驅動電路23而使ON· OFF之切換被作了控制的輸出短路用之切換電晶體S W0， 並在輸出短路用之切換電晶體SW0的短路狀態（對於標靶 - ΤΙ、T2之輸出被遮斷的狀態）下，來進行橋接電路22之各 _ 切換電晶體SW1乃至SW4之切換。 亦即是，如圖2中所示一般，當對於一對之標靶T1、 T2而進行雙極脈衝狀之電力供給的情況時，在切換電晶體 SW0之短路狀態（ON)下，例如將第1以及第4切換電晶體 • SW1、SW4設爲ON，而後，將切換電晶體SW0之短路解除 (OFF)，而對其中一方之標靶T1作輸出（在標靶T1處係 被施加有負的電位）。接著，再度使切換電晶體SW0短路 ，並在將第1以及第4切換電晶體SW1、SW4設爲OFF的同 時，將第2以及第3切換電晶體SW2、SW3設爲ON，而後， 將切換電晶體SW0設爲OFF，而對另外一方之標靶T2作輸 出（在標靶T2處係被施加有負的電位）。 而後，藉由反覆進行將各切換電晶體SW1乃至SW4之 ON、OFF的時機作反轉的上述控制，而在一對之標靶T1、 -13- 201010260 T2之間來以特定之頻率而進行雙極脈衝狀之電力供給。此 時，在將Ar等之濺鍍氣體導入至被保持爲特定壓力之裝置 內的狀態下，以特定之頻率來交互地切換極性並被投入有 電力之一對的標靶ΤΙ、T2，係分別交互地被切換爲陽極電 極、陰極電極，而在陽極電極以及陰極電極之間使輝光放 電產生，並形成電漿氛圍，而能夠對各標靶ΤΙ、T2作濺鍍 〇 藉由此，在進行對於標靶ΤΙ、T2之輸出時所產生的切 @ 換損失，係僅會在切換電晶體SW0處發生，在各切換電晶 體SW1乃至SW4處，係幾乎不會發生切換損失。其結果， 不需使用高功能之切換元件，便可達成高耐久性，並且， _ 係成爲不需要像是在4個的切換元件處均產生有切換損失 的情況時一般之充分的放熱機構，而能夠謀求低成本化。 在上述一般之輝光放電中，係週知有會由於某種之原 因而產生電弧放電的事態。若是產生有電弧放電，則由於 電漿之阻抗係急遽地變小，因此，會引起急遽之電壓降低 Θ ，伴隨於此，電流係增加。故而，在本實施形態中，係在 第2CPU電路21處，可自由通訊地設置被輸入有藉由檢測 電路25所檢測出之輸出電流以及輸出電壓的電弧檢測控制 電路27 (參考圖1)，若是輸出電流超過一定之範圍而變 化，則作爲電弧放電之前段現象（微電弧）來掌握，並藉 由進行其之消弧處理，而對電弧電流爲大之電弧放電的發 生作抑制。 於此，若是如同先前技術之雙極脈衝電源一般，從直 -14- 201010260 流電力供給部而來之輸出係具備有定電壓特性，則相較於 電感成分，電容成分（capacitance)係成爲更加具有支配 性。如圖3中所示一般，若是電容成分（capacitance )係 成爲支配性，則在電弧放電發生時，由於電漿負載側之阻 抗係變小，因此，輸出與電漿負載係被耦合，並從電容成 分而急遽地被放出至輸出側。因此，在從檢測手段所致之 電弧放電的檢出起直到對於電極之輸出被遮斷爲止的時間 ❹ 內，係流動有大的電弧電流。其結果，若是無法藉由1次 的處理來將電弧放電消弧，則在每1次的微電弧處理之進 行中，電弧電流値係變高（被放出之電弧能量係變大）， -並成爲容易產生噴濺或是粒子。 在本實施形態中，係設爲在負的直流輸出線14b處， 設置了具備有ImH以上、較理想爲2mH以上之値的電感28 (參考圖1)。而，微電弧處理時之電流上升率，係設爲 被限制在較定常電流値之200%爲更小、更理想爲限制在 • 150%以下》亦即是，電弧放電發生時之輸出電流上升率（ △i)，若是將電感28之電感値設爲L，並將對於標靶T1、 T2之輸出電壓設爲V，將電流變化時間設爲At，則係可藉 由Δί = Δίχν/ί而算出。於此情況，假設對於一對之標 靶ΤΙ、Τ2的輸出電壓係爲500V，輸出電流係爲100Α，微 電弧處理（輸出遮斷）時間設爲200 # s，則爲了將從檢測 出過電流起直到將輸出遮斷爲止的電流上升率設爲15 0%， △ i係成爲50Α。在此種情況下，係只要將具有2mH之電感 的電感28連接於負的直流輸出線14b處即可。 -15- 201010260 另外，在本實施形態中，係設爲在負的直流输出線 14b處設置ImH以上、更理想爲2mH以上之値的電感28，但 是，係並不被限定於此，而亦可設置在正的直流輸出線 14a處，或是分別設置在正負之兩直流輸出線14a、14b處 〇 而後，如圖4中所示一般，當藉由檢測電路25所檢測 出之輸出電流la超過定常輸出電流値Ic時，係經由電弧檢 測控制電路27而作爲電弧放電發生之前段現象來掌握，並 經由第2CPU電路21以及電弧檢測控制電路27，來經由輸 出震盪用之驅動電路23來使輸出短路用之切換電晶體SW0 短路（ON)。此時，藉由在直流輸出線14b設置有電感28 一事，從直流電源供給部1而來之輸出係成爲定電流特性 ，而電弧放電之發生時的電流上升率係被限制。 當輸出短路用之切換電晶體SW0被短路（ON)時，橋 接電路22之各切換電晶體SW1乃至SW4，係被保持在對於 任一方之標靶ΤΙ、T2的輸出狀態下，但是，藉由使切換電 晶體SW0短路，對於標靶ΤΙ、T2之輸出係被遮斷（微電弧 處理）° 接著，在特定時間經過後（數；as〜數百#s)，解除 輸出短路用之切換電晶體SW0的短路（OFF )，並因應於 各切換電晶體SW1乃至SW4之動作狀態，來再度開始對於 任一方之標靶ΤΙ、T2的輸出。此時，經由電弧檢測控制電 路27而判斷輸出電流Va是否超過了定常輸出電流値Vc，若 是尙未超過定常輸出電流値Vc，則經由輸出震盪用驅動電 -16- 201010260 路23，來使輸出短路用之切換電晶體SW0再度短路。 當就算是反覆進行複數次之此種一連串的微電弧處理 ，輸出電流la亦仍維持在超過定常輸出電流値Ic的狀態、 或是輸出電流la超過預先所設定之特定値，則係判斷發生 有會誘發噴濺或粒子之產生的電弧放電，並經由從第 1CPU電路11而來之控制，來將切換電晶體15設爲ON，並 停止從直流電力供給部1而來之輸出（強電弧（hard arc) φ 處理）。在此處理之間，亦同樣的，藉由電弧電流値係被 保持爲較定常電流値之2 0 0%爲更小一事（參考圖5 )、和 能夠相較於對輸出中之2個的切換電晶體SW1乃至SW4作切 -換並進行電弧放電之消弧處理的情況而以更好的回應性來 進行其之輸出遮斷的控制一事，此兩者間之相輔相成，而 使被放出之電弧能量變小，而能夠對於噴濺或粒子之產生 有效地作抑制。於此處理之間，在橋接電路22之各切換電 晶體SW1乃至SW4處，由於係幾乎不會產生切換損失，因 〇 此，能夠將其之耐久性更進一步地提昇。 然而，在設置了如同上述一般而具有ImH以上之値的 電感28的情況時，如圖6(a)中所示一般，當對橋接電路 22之各切換元件SW1乃至SW4而以特定之頻率（例如， 5kHz)來作切換時，係產生較通常之放電電壓Vc爲更高的 電壓Va。亦即是，由於在電漿P中產生有電感成分，因此 ，在各標靶ΤΙ、T2處之極性反轉時會產生過電壓。若是產 生此種過電壓，則會有引發電弧放電之虞。 因此，在本實施形態中，係設爲設置了輸出箝制電路 -17- 201010260 29，其係將由被連接於從直流電力供給部1而來之正負的 直流輸出線14a、14b處之電容器C、和爲與電感28並聯且 相互被作了串聯連接之二極體D以及電阻R，作了連接所成 者。藉由此，如圖6(a)以及圖6(b)中所示一般（另外 ，在圖6中，係僅展示有在其中一方之標靶T1處的輸出電 壓以及輸出電流的變化），各標靶ΤΙ、T2之極性反轉時， 於起初，藉由將電源側作爲陰極而連接了的二極體D，電 感28係被短路，對於各標靶ΤΙ、T2之輸出係成爲定電壓特 性，輸出電流Ac係逐漸的增加（參考圖6(b))。而後， 若是輸出電流Ac達到了對應於設定電力之特定値，則上述 輸出係成爲定電流特性。其結果，在各標靶ΤΙ、T2處之極 性反轉時而產生過電壓一事係被防止，而起因於過電流之 電弧放電的發生係被抑制。於此情況，作爲電容器C，係 使用有5〜20# F者，作爲電阻R，係使用有數Ω〜10Ω之 範圍者。 接著，參考圖7以及圖8，針對將本發明之雙極脈衝電 源E作複數台並聯連接所成的電源裝置作說明。ES，係爲 本發明之電源裝置，此電源裝置ES，例如係被使用在具備 有下述之構成的磁控管濺鍍裝置（以下，稱爲「濺鍍裝置 」）3中。 濺鍍裝置3，係具有經由旋轉式幫浦、渦輪分子幫浦 等之真空排氣手段（未圖示）而能保持特定之真空度（例 如，10_5Pa)的真空處理室31，而構成濺鍍室（處理室） 32。在真空處理室31之上部，例如係被設置有將在FPD製 201010260 造時所被使用之大面積的處理基板S電位性地保持爲浮動 狀態之基板支持器33。在真空處理室31中，係又被設置有 將製程氣體導入至濺鏟室32內之氣體導入管（未圖示）’ 並能夠將由Ar等之希有氣體所成之濺鍍氣體、或是在藉由 反應性濺鍍而形成特定之薄膜的情況時而因應於欲形成在 處理基板S之表面的薄膜之組成而被適宜選擇之〇2、N2或 H20等的反應性氣體，導入至處理室32中。 0 在濺鍍室32中，係與處理基板S相對向的，而以等間 隔來並排設置有複數枚（於本實施形態中，係爲8枚）之 標！G41a乃至41h。各標|G41a乃至41h，係由Al、Ti、Mo、 _ 銦以及錫之氧化物（ITO)、或是銦以及錫之合金等的因 應於欲形成在處理基板S之表面處的薄膜之組成而藉由週 知之方法所製作者，並係被形成爲例如略直方體（俯視時 爲長方形）等的相同形狀。 各標靶41 a乃至41h，係在濺鍍中，藉由銦或是錫等之 • 焊接材料，而被接合於用以將標靶41 a乃至41 h作冷卻的背 板上。各標靶41a乃至41h，係以使未使用時之濺鍍面位置 於與處理基板S平行之同一平面上的方式，而經由絕緣構 件來設置在真空處理室31中。又，在標靶41 a乃至41 h之後 方（與濺鍍面相背向之側），係被配置有具備週知之構造 的磁石組裝體（未圖示），藉由捕捉在各標靶41 a乃至41h 之前方（濺鍍面）側所電離的電子及經由濺鏟所產生之二 次電子，而能提高在各檩靶41 a乃至41 h的前方之電子密度 ，並提高電漿密度，而能夠提高濺鍍速率。 -19- 201010260 各標靶41 a乃至41h，係以相鄰之2枚來構成一對之標 靶（41a與41b、41c與41d、41e與41f、41g與41h )，並對 於各個一對之標靶41 a乃至41h，而分配設置有上述實施形 態之雙極脈衝電源E1乃至E4，從雙極脈衝電源E1乃至E4 而來之輸出線24a、24b，係被連接於各一對的標靶41a、 4115(41(：以及41(1、416以及41卜418以及4111)。藉由此 ，可經由雙極脈衝電源E1乃至E4，來對於各一對之標靶 41 a乃至41h而交互地改變極性並進行雙極脈衝狀之電力供 給。 在本實施形態中，係爲了安定地在標靶41 a乃至41 h之 前方產生電漿，而以使相互鄰接之標靶41 a乃至41 h的極性 相互反轉的方式，來使各雙極脈衝電源E1乃至E4同步並供 給電力（參考圖5)。爲了進行此同步運轉，係設置有由 各雙極脈衝電源E1乃至E4之被可自由通訊地與第2CPU電 路21作了連接之CPU所成的統籌控制手段5。 而後，在各雙極脈衝電源E1乃至E4之輸出短路用的切 換電晶體SW0之短路狀態下，於各雙極脈衝電源E1乃至E4 的每一中，將第1以及第4切換電晶體SW1、SW4，和第2以 及第3切換電晶體SW2、SW3，其兩者間之ON· PFF的時機 作反轉，同時，以使對於相互鄰接之標靶41 a乃至41 h的極 性作反轉的方式，來使各切換電晶體SW1乃至SW4動作， 之後，藉由從統籌控制手段5而來之輸出，切換電晶體 SW0之短路係被解除，並對—對的標靶中之各一方的41a 、41c、41e、41g進行輸出。 -20- 201010260 接著，藉由從統籌控制手段5而來之輸出，而將各雙 極脈衝電源E1乃至E4之輸出短路用的切換電晶體SW0短路 ，並對各切換電晶體SW1乃至SW4作切換，而後，藉由從 統籌控制手段而來之輸出，來解除切換電晶體SW0之短路 ，並對另外一方之各標靶41b、41d、41f、41h進行輸出。 而後，藉由反覆進行上述控制，在各標靶41 a乃至41 h處， 係以特定之頻率而被進行雙極脈衝狀之電力供給，並被作 〇 同步運轉。 於此同步運轉時，由於只要經由統籌控制手段5來將 對於各雙極脈衝電源E1乃至E4之輸出短路用的切換元件 - SW0之ON · OFF的切換時機作同步即可，因此，能夠具有 . 充分餘裕地來使各雙極脈衝電源E1乃至E4之切換元件SW1 乃至SW4動作，就算是在各雙極脈衝電源之切換元件或控 制電路中存在有個體差異，其之同步運轉亦係爲容易。 又，各雙極脈衝電源E1乃至E4，係構成爲：於濺鍍中 β ，當在任一個的雙極脈衝電源中，藉由檢測電路25所檢測 出之輸出電流la爲超過定常輸出電流値Ic時，則藉由該雙 極脈衝電源之電弧檢測控制電路23所致之输出短路用的切 換電晶體SW0之切換，來進行上述之微電弧處理。 當在任1個的雙極脈衝電源處而進行微電弧處理時， 若是被連接有從此雙極脈衝電源而來之輸出纜線24a、24b 的一對之標靶，和與此一對之標靶相鄰接之被連接有從其 他之雙極脈衝電源而來的輸出纜線24a、24b之其他的標靶 ，其兩者間之電位係爲相互一致，則係能夠容易地將電弧 •21 - 201010260 放電作消弧。 在本實施形態中，當在任一個的雙極脈衝電源El乃至 E4中而開始了微電弧處理時，係將此經由統籌控制手段5 來輸出至對相鄰接之標靶進行輸出之雙極脈衝電源的第 2CPU電路21處。於此情況，係設爲：經由該第2CPU電路 21，藉由輸出震盪用之驅動電路23，輸出短路用之切換電 晶體SW0係暫時被短路，因應於各切換電晶體SW1乃至 SW4之動作狀態，以使電位成爲相互一致的方式，各切換 @ 電晶體SW1乃至SW4之動作的時機係被作變更，而，輸出 短路用之切換電晶體SW0之短路係被解除，並被輸出至標 耙處。 _ 另外，在本實施形態中，雖係針對爲了將各雙極脈衝 電源E1乃至E4作同步運轉而設置有統籌控制手段者來作了 說明，但是，亦可設爲將任一個的第2CPU電路21作爲統 籌控制手段來構成（主電源），並經由此統籌控制手段之 輸出，來對於其他之雙極脈衝電源E2乃至E4 (副電源）之 @ 動作作控制。 【圖式簡單說明】 [圖1]對本發明之雙極脈衝電源的構成作槪略展示之圖 [圖2]對本發明之雙極脈衝電源的輸出控制作說明之圖 〇 [圖3]對先前技術之雙極脈衝電源處的微電弧處理時之 電流變化作說明之圖。 -22- 201010260 [圖4]對本發明之雙極脈衝電源處的微電弧處理作說明 之圖。 [圖5]對本發明之雙極脈衝電源處的微電弧處理時之電 流變化作說明之圖。 [圖6] (a)以及（b)，係爲對對於其中一方之電極的 輸出電壓以及輸出電流之波形作說明之圖。 [圖7]對使用有本發明之電源裝置的濺鍍裝置作槪略性 φ 說明的圖。 [圖8]對本發明之電源裝置的輸出控制作說明之圖。 .【主要元件符號說明】 1 '·直流電力供給部 2 :震盪部 22 :橋接電路 24a ' 24b ··輸出纜線 • 25 :輸出電流、電壓檢測電路 27 :電弧檢測控制電路 E :雙極脈衝電源 SW0乃至SW4 :切換元件 ΤΙ、T2 :電極（標靶） -23-201010260 VI. Description of the Invention: [Technical Field] The present invention relates to a bipolar pulse power supply for supplying power to a plasma and a surface treatment apparatus in a bipolar pulse shape, and a plurality of bipolar pulse power sources The power supply unit is connected in parallel. [Prior Art] • Such a bipolar pulse power source is used, for example, in a sputtering apparatus for forming a specific thin film on a surface of a substrate, and is known to have a rectifying circuit for supplying DC power and to be connected thereto. The MOSFET bridge circuit of the positive and negative output terminals of the rectifier circuit and formed by four switching elements. Then, the switching elements are appropriately operated by the control means, and for the target which is one pair of the output terminals (electrodes), the polarity is alternately switched at a specific frequency and an arbitrary pulse voltage is applied, and each The targets are alternately switched to the anode electrode and the cathode electrode, respectively, and a glow discharge is generated between the anode electrode and the cathode electric Φ pole, and a plasma atmosphere is formed, and each target is sputtered. As a result, the electric charge accumulated on the surface of the target is canceled when the opposite phase voltage is applied, and there is an advantage that a stable discharge can be obtained (for example, Patent Document 1). In such glow discharge, it is known that an arc discharge (abnormal discharge) occurs for some reason. If an arc discharge occurs, the impedance of the plasma (load) is rapidly reduced, so that the voltage of the sudden voltage is lowered, and accordingly, the current system is increased. Here, in particular, when the target is made of a metal such as aluminum, if a high arc current 値 arc discharge occurs locally between the targets -5 - 201010260, the target is melted and The released portion adheres to so-called particles or splatters (blocks of several to several hundred//in) on the surface of the substrate, and it is impossible to perform good film formation. Therefore, in the above-described bipolar pulse power source, the system is provided. a detection circuit for detecting an output current from the bridge circuit and an inductance for suppressing a rise in current when the arc discharge is generated, and when the output current detected by the detection circuit exceeds the constant output current 値, Switching between the switching elements in the action and temporarily blocking the output of the electrode. Then, if the overcurrent is quiet and the enthalpy becomes close to the constant output current 値, the output to the electrode is started again. In this case, if the output current exceeds a certain range and changes, it is grasped as a phenomenon before the abnormal discharge (micro-arc), and by performing the arc-extinguishing process thereof, it is also possible to perform an arc with a large amount of current change. The occurrence of discharge is suppressed. [Patent Document 1] Japanese Patent No. 3369960 [Invention] [Problems to be Solved by the Invention] However, in general, an output system derived from a DC power supply source has a constant voltage characteristic 'so' The capacitance component 'capacitance' is more dominant. Therefore, when the arc discharge occurs, the impedance on the load side of the plasma becomes small (depending on the case, it is also reduced to a few ohms or less), so the output is coupled to the plasma (load), and the capacitance component is removed. And is eagerly released to the output side. As a result, -6-201010260, even if the inductor is set to a small inductor, it cannot effectively suppress the current rise, but has a current flowing in a short time (between numbers) (that is, the arc discharge occurs). The problem of the current rise rate per unit time is high. When the current rise rate per unit time is high, even when the state of the current change is small and the micro-arc treatment is performed, there is also a change in current from a later change according to the voltage change. It is detected that a large arc current flows during the time from the φ arc discharge until the power supply to the plasma is interrupted, and the released arc energy is increased (the flow has a constant current of about 2 times) The current is) and the splash or the particles are likely to occur. In particular, when an arc discharge occurs continuously, the occurrence of sputtering or particle formation cannot be substantially suppressed. Accordingly, it is an object of the present invention to provide a bipolar pulse power source capable of suppressing the rise of current during arc discharge and capable of suppressing splash or particle generation and the double Ο Pole pulse power supply multiple units connected in parallel to form a power supply unit. [Patent Document 1] Japanese Patent No. 3 639605 (for example, refer to the description of Patent Application No. 1 and Paragraph No. 16) [Means for Solving the Problem] In order to solve the above problem, the first application of the patent scope is The bipolar pulse power supply described in the present invention includes a bridge circuit including a switching element connected between positive and negative DC outputs from a DC power supply source, and ON/OFF of each switching element of the bridge circuit. The switching is controlled by the control method 201010260, and the bipolar pulse-like power supply is performed at a specific frequency for the electrode in contact with the plasma, characterized in that the bridge circuit is connected from the DC power supply source. At least one of the positive and negative DC outputs is provided with an inductance of more than or equal to ImH. According to the present invention, when DC power is supplied to the bridge circuit from the DC power supply source, two of the switching elements constituting the bridge circuit are outputted by the control means. When the element is turned on, power is supplied (output) to one of the electrodes. Then, when the switching elements that output the electrodes of one of the electrodes are turned OFF, and the switching elements that output the other electrode are turned ON, the electrodes of the other side are output. By repeating this control, a bipolar pulsed power supply is performed at a specific frequency for the electrode in contact with the plasma. On the other hand, when an arc discharge occurs for some reason, since the impedance of the plasma is rapidly reduced, the voltage of the sudden voltage is lowered, and the current is increased accordingly. In this case, since at least one of the positive and negative DC outputs to the bridge circuit from the DC power supply source is provided with the inductance 28 having a radius of 1 or more, the output from the DC power supply source is provided. The constant current characteristics are obtained, and as a result, the rate of increase in current per unit time at the time of occurrence of arc discharge is limited. Here, the output current increase rate (Δί) per unit time at the time of arc discharge occurs, if the inductance 値 of the inductance is L, and the output voltage of the electrode is V, and the current change time is At, It can be calculated by 201010260 Δ i= Δίχν/Ι:. In this case, in the plating apparatus generally used in mass production, since the output power to the target is at least 5 kW or more, in order to discharge the arc per unit time (for example, 1 〇〇//s) The rate of increase of the output current at the time of occurrence is suppressed to be less than 20% of the constant current 値, and more preferably 150% or less, and the inductance system needs to have an inductance I of ImH or more. However, when the inductance of 値 or more is provided as described above, when switching the switching elements of the bridge circuit, a voltage higher than the normal discharge voltage may be generated. That is, since an inductance component is generated in the plasma, an overvoltage is generated when the polarity of each electrode is reversed. If an overvoltage is generated in this way, the arc discharge is induced. Therefore, if the output clamp circuit is further provided, it is provided with positive and negative direct currents from the DC power supply source. The output is a capacitor connected in parallel, and a diode and a resistor connected in parallel with the above-mentioned inductor and connected in series. When the polarity is reversed, the output of the electrode is initially a constant voltage characteristic. The output current is gradually increased, and then (if the output current reaches a certain level), the output of the electrode becomes a constant current characteristic. As a result, an overvoltage is generated when the polarity of each electrode is reversed, and the occurrence of an arc discharge due to an overcurrent is suppressed. Further, the electrode is a pair of targets disposed in a processing chamber where a sputtering method is performed. Further, it is preferable to provide a switching element for outputting a short circuit between the positive and negative DC outputs connected to the circuit from the DC power supply source to the bridge 201010260, and the detecting means. The output current between the pair of electrodes is detected; and the abnormal discharge detecting means is such that if the absolute value of the output current exceeds the constant output current 对于 for the electrode, it is grasped as a phenomenon before the abnormal discharge occurs, if This abnormal discharge detecting means grasps the phenomenon before the occurrence of the abnormal discharge, and the arc extinguishing process of the abnormal discharge is performed by blocking the output of the electrode via the switching element for the output short circuit. By this, when bipolar pulse-shaped power supply is performed, not only Q is able to cause switching loss to be generated in only one switching element for output short-circuit, but also switching to two of the outputs. In the case where the component is controlled and the arc extinguishing process of the abnormal discharge is performed, the _ control can be performed with good responsiveness, and in this process, switching loss is hardly generated at each switching element of the bridge circuit, and Can improve its durability. Further, in order to solve the above problems, the power supply device according to the fifth aspect of the invention is a power supply device in which a plurality of bipolar pulse power sources as described in the fourth aspect of the patent application are connected in parallel, and is characterized in that @ 有 有: Coordinated control means, when bipolar pulse-shaped power is supplied to the electrodes of a plurality of pairs disposed in the same processing chamber, ON/OFF of the switching element for short-circuiting the output of each bipolar pulse power supply Switching is controlled. According to the present invention, it is only necessary to synchronize the switching elements for short-circuiting the respective outputs of the respective bipolar pulse power supplies via the integrated control means, so that the switching elements of the bridge circuit can be operated with sufficient margin, even if In the switching element or control circuit of each bipolar pulse power supply, there is an individual difference of the switching speed of -10-201010260 or the control speed, and the synchronous operation is also easy. [Effects of the Invention] As described above, in the bipolar pulse power source and the power supply device of the present invention, the current rise at the time of arc discharge is effectively limited, and as a result, it is possible to spatter or particles. The effect of the effective # can be suppressed, and the effect of forming a good film can be obtained. [Embodiment] Referring to FIG. 1, E, the bipolar pulse power supply of the present invention, the bipolar pulse power supply E, for example, The processing substrate in the sputtering apparatus is disposed opposite to the ground, and is used for bipolar pulse-like power supply at a specific frequency for a pair of targets ΤΙ, T2 that are electrodes that are in contact with the plasma P. And being used. The bipolar pulse power supply E is composed of a DC power supply unit 1 that enables supply of DC power, and an oscillation unit 2 that controls the output (power supply) of each of the targets ΤΙ and T2. In this case, the waveform of the output voltage is a slightly square wave or a slightly sine wave. The DC power supply unit 1 includes a first CPU circuit 11 that controls the operation of the DC power supply unit 1 and an input unit 12 to which commercial AC power (3-phase AC 200V or 400 V) is input, and an input exchange. The electric power is rectified and converted into a rectifying circuit 13 composed of six diodes 13 a of DC power, and the DC power -11 - 201010260 is output to the oscillating portion 2 via the positive and negative DC power lines 14a and 14b. Further, in the DC power supply unit 1, a switching transistor (switching element) 15 provided between the DC power lines 14a and 14b is provided, and is connected to the first CPU circuit 11 in a freely communicable manner. The drive circuit 16 for output oscillation for controlling the ON/OFF of the switching transistor 15 is also performed. The detection circuit 17a that detects the current and voltage detected between the DC power lines 14a and 14b is input via the AD conversion circuit 17b by the current and voltage detected by the detection circuit 17a. To the first CPU circuit 1 1 . On the other hand, the oscillation unit 2 is provided with four CPU circuits 21 that are freely communicably connected to the first CPU circuit 11, and four that are connected between the positive and negative DC power lines 14a and 14b. The bridge circuit 22 formed by the first and fourth switching transistors SW1 and SW4, and the output of the second CPU circuit 21 that is freely communicable, and that controls the switching of the ON/OFF of each of the switching transistors SW1 and SW4 The drive circuit 23 for oscillation is turned ON and OFF, for example, by the drive circuit 23 for output oscillation, for example, the first and fourth switching transistors SW1 and SW4 and the second and third switching transistors SW2 and SW3. When the switching of each switching transistor SW1 or SW4 is controlled by the timing reversal, the pair of targets T1 and T2 can be doubled via the output power lines 24a and 24b from the bridge circuit 22. Extremely pulsed power supply. At the output lines 24a, 24b, a detection circuit 25 that detects an output current and an output voltage for a pair of target ΤΙ, T2 is connected, whereby the output current and the output voltage detected by the detection circuit 25 are detected. It is transmitted to the second CPU circuit 21 via the AD conversion circuit 26 through the 201010260. In the above-described bipolar pulse power supply E, when the DC power is output from the DC power supply unit 1 and the switching transistors SW1 to SW4 are switched, the switching loss is reduced. Since it is quite large, it is necessary to configure the switching transistors SW1 to SW4 to have improved durability. In the present embodiment, the ON/OFF switching is performed between the positive and negative DC output lines 14a and φ 14b from the DC power supply unit 1 via the drive circuit 23 for output oscillation. The control output short-circuit switching transistor S W0 is used, and in the short-circuit state of the switching transistor SW0 for output short-circuit (the state in which the output of the target - ΤΙ, T2 is blocked), the bridge circuit 22 is performed. Each _ switches the switching of the transistors SW1 to SW4. That is, as shown in FIG. 2, when bipolar pulse-like power supply is performed for a pair of targets T1 and T2, in the short-circuit state (ON) of the switching transistor SW0, for example, The first and fourth switching transistors • SW1 and SW4 are turned on, and then the short circuit of the switching transistor SW0 is turned off (OFF), and one of the targets T1 is output (the target T1 is applied with the target T1). Negative potential). Then, the switching transistor SW0 is again short-circuited, and the second and third switching transistors SW2 and SW3 are turned ON while the first and fourth switching transistors SW1 and SW4 are turned off, and then switched. The transistor SW0 is set to OFF, and the other target T2 is output (a negative potential is applied to the target T2). Then, by repeating the above-described control of inverting the timing of turning ON and OFF of each of the switching transistors SW1 and SW4, a predetermined frequency is performed between the pair of targets T1, -13-201010260T2. Bipolar pulsed power supply. At this time, in a state where a sputtering gas such as Ar is introduced into a device held at a specific pressure, the polarity is alternately switched at a specific frequency, and the target ΤΙ and T2 of one pair of electric power are input. They are alternately switched to an anode electrode and a cathode electrode, and a glow discharge is generated between the anode electrode and the cathode electrode to form a plasma atmosphere, and each target ΤΙ and T2 can be sputtered, thereby The switching loss generated when the output of the target ΤΙ and T2 is performed is only generated at the switching transistor SW0, and switching loss is hardly generated at each switching transistor SW1 or SW4. As a result, it is possible to achieve high durability without using a high-function switching element, and it is not necessary to have a sufficient heat release mechanism as in the case where switching loss occurs in all four switching elements. It is possible to reduce costs. In the above general glow discharge, it is known that an arc discharge occurs due to a certain origin. If an arc discharge occurs, the impedance of the plasma is rapidly reduced, so that the voltage of the sudden voltage is lowered, and accordingly, the current system increases. Therefore, in the second embodiment, the second CPU circuit 21 is provided with an arc detection control circuit 27 (refer to FIG. 1) to which the output current and the output voltage detected by the detection circuit 25 are input, (see FIG. 1). If the output current changes beyond a certain range, it is grasped as a phenomenon before the arc discharge (micro-arc), and by performing the arc-extinguishing process thereof, the occurrence of arcing with a large arc current is suppressed. Here, as in the case of the prior art bipolar pulse power supply, the output system from the direct-14-201010260 power supply unit has a constant voltage characteristic, and the capacitance component becomes more compact than the inductance component. Dominant. As shown in FIG. 3, in general, if the capacitance component is dominant, when the arc discharge occurs, since the impedance on the plasma load side becomes small, the output is coupled to the plasma load system, and The capacitor component is rushed out to the output side. Therefore, a large arc current flows during the time from the detection of the arc discharge by the detecting means until the output of the electrode is blocked. As a result, if the arc discharge cannot be extinguished by one treatment, the arc current becomes high during the progress of the micro-arc treatment (the arc energy to be released becomes large), and It becomes easy to produce splashes or particles. In the present embodiment, an inductance 28 having a thickness of 1 mH or more, more preferably 2 mH or more is provided at the negative DC output line 14b (refer to FIG. 1). However, the current rise rate during the micro-arc treatment is set to be limited to 200% of the constant current 値, and more preferably to be less than 150%, that is, the output current rises when the arc discharge occurs. Rate ( Δi), if the inductance 値 of the inductor 28 is set to L, and the output voltage for the targets T1 and T2 is set to V, and the current change time is set to At, Δί = Δίχν/ί And calculate. In this case, it is assumed that the output voltage of the pair of targets ΤΙ and Τ2 is 500 V, the output current is 100 Α, and the micro-arc processing (output occlusion) time is set to 200 # s, in order to detect an overcurrent. The current increase rate until the output is blocked is set to 15%, and Δi is 50Α. In this case, it is only necessary to connect the inductor 28 having an inductance of 2 mH to the negative DC output line 14b. -15- 201010260 In the present embodiment, the inductance 28 of ImH or more, more preferably 2 mH or more is provided in the negative DC output line 14b. However, the present invention is not limited thereto. It can be disposed at the positive DC output line 14a, or at the positive and negative DC output lines 14a, 14b, respectively, and then, as shown in FIG. 4, when the output current is detected by the detection circuit 25 When the constant output current 値Ic is exceeded, the arc detection control circuit 27 grasps the phenomenon before the arc discharge occurs, and the second CPU circuit 21 and the arc detection control circuit 27 pass the output drive circuit 23 for output oscillation. The switching transistor SW0 for output short-circuit is short-circuited (ON). At this time, by providing the inductance 28 to the DC output line 14b, the output from the DC power supply unit 1 has a constant current characteristic, and the current increase rate at the time of occurrence of the arc discharge is limited. When the switching transistor SW0 for output short-circuit is short-circuited (ON), each switching transistor SW1 or SW4 of the bridge circuit 22 is held in the output state of the target ΤΙ, T2 for either side, but by The switching transistor SW0 is short-circuited, and the output of the target ΤΙ and T2 is blocked (micro-arc treatment). Then, after a certain time elapses (number; as ~ hundreds of #s), the output short-circuit switching power is released. The short circuit (OFF) of the crystal SW0 is used to restart the output of the target ΤΙ and T2 for any one of the switching transistors SW1 and SW4. At this time, it is determined whether or not the output current Va exceeds the constant output current 値Vc via the arc detection control circuit 27, and if the constant output current 値Vc is not exceeded, the output is output via the output oscillation drive--16-201010260 path 23. The switching transistor SW0 for short circuit is short-circuited again. Even if the series of micro-arc treatments are repeated repeatedly, the output current la is maintained at a state exceeding the constant output current 値Ic, or the output current la exceeds a predetermined threshold, which is determined to occur. The arc discharge which causes the occurrence of splashing or particles is controlled by the first CPU circuit 11, and the switching transistor 15 is turned ON, and the output from the DC power supply unit 1 is stopped (strong arc ( Hard arc) φ processing). Also between the processes, the arc current is maintained to be less than 200% of the constant current 値 (refer to FIG. 5), and can be compared with the two of the outputs. Switching the transistors SW1 and SW4 to cut-and-change and performing the arc-extinguishing process of the arc discharge, and performing the control of the output interruption thereof with better responsiveness, the two complement each other, and the released The arc energy is reduced, and it is possible to effectively suppress the generation of splashes or particles. Between this processing, at each of the switching transistors SW1 and SW4 of the bridge circuit 22, since switching loss is hardly generated, the durability thereof can be further improved. However, in the case where the inductor 28 having the I or more than the above is provided, as shown in FIG. 6(a), when the switching elements SW1 to SW4 of the bridge circuit 22 are at a specific frequency ( For example, when 5 kHz is used for switching, a voltage Va higher than the normal discharge voltage Vc is generated. That is, since the inductance component is generated in the plasma P, an overvoltage is generated when the polarity of each of the targets ΤΙ and T2 is reversed. If such an overvoltage is generated, there is a risk of causing an arc discharge. Therefore, in the present embodiment, the output clamp circuit -17-201010260 29 is provided, which is connected to the capacitor C connected to the positive and negative DC output lines 14a and 14b from the DC power supply unit 1. And a diode D and a resistor R which are connected in parallel with the inductor 28 and connected in series to each other are connected. Thereby, as shown in FIG. 6(a) and FIG. 6(b), in general (in FIG. 6, only the output voltage and the output current change at the target T1 on one of the sides are shown), When the polarity of each of the target ΤΙ and T2 is reversed, the inductor 28 is short-circuited by the diode D connected to the power supply side as a cathode, and the output of each target ΤΙ and T2 becomes a constant voltage. Characteristic, the output current Ac is gradually increased (refer to Fig. 6(b)). Then, if the output current Ac reaches a specific enthalpy corresponding to the set power, the output is a constant current characteristic. As a result, an overvoltage is generated when the polarity of each target ΤΙ and T2 is reversed, and the occurrence of arc discharge due to overcurrent is suppressed. In this case, as the capacitor C, 5 to 20 #F is used, and as the resistor R, a range of several Ω to 10 Ω is used. Next, a power supply device in which the bipolar pulse power source E of the present invention is connected in parallel in a plurality of stages will be described with reference to Figs. 7 and 8. The ES is the power supply device of the present invention, and the power supply device ES is used, for example, in a magnetron sputtering device (hereinafter referred to as "sputtering device") 3 having the following configuration. The sputtering apparatus 3 has a vacuum processing chamber 31 capable of maintaining a specific degree of vacuum (for example, 10_5 Pa) via a vacuum exhausting means (not shown) such as a rotary pump or a turbo molecular pump, and is configured to perform sputtering. Room (processing room) 32. In the upper portion of the vacuum processing chamber 31, for example, a substrate holder 33 for holding a large-area processing substrate S used for the production of the FPD 201010260 in a floating state is provided. In the vacuum processing chamber 31, a gas introduction pipe (not shown) that introduces the process gas into the splash shovel 32 is provided, and a sputtering gas formed of a gas such as Ar can be provided, or When a specific thin film is formed by reactive sputtering, a reactive gas such as 〇 2, N 2 or H 20 which is appropriately selected in accordance with the composition of the thin film to be formed on the surface of the processing substrate S is introduced into the processing chamber. 32. In the sputtering chamber 32, the processing substrate S is opposed to each other, and a plurality of pieces (eight in the present embodiment) are arranged side by side in equal intervals! G41a or even 41h. Each of the marks |G41a or 41h is composed of an alloy of Al, Ti, Mo, _Indium and tin oxide (ITO), or an alloy of indium and tin, which is formed in accordance with the film to be formed on the surface of the substrate S to be processed. The person who is produced by a known method is formed into the same shape such as a substantially rectangular parallelepiped (rectangular in plan view). Each of the targets 41a to 41h is bonded to a backing plate for cooling the target 41a or 41h by sputtering or the like. Each of the targets 41a to 41h is disposed in the vacuum processing chamber 31 via an insulating member so that the sputtering surface when not in use is positioned on the same plane as the processing substrate S. Further, after the target 41a is 41 h or later (on the side opposite to the sputtering surface), a magnet assembly (not shown) having a well-known structure is disposed, and is captured by each target 41a. Even the electrons ionized on the side (sputtered surface) before 41h and the secondary electrons generated by the spatter can increase the electron density in front of each target 41a or 41h and increase the plasma density. Can increase the sputtering rate. -19- 201010260 Each target 41 a or 41 h is a pair of adjacent targets (41a and 41b, 41c and 41d, 41e and 41f, 41g and 41h), and for each pair The target 41a is even 41h, and the bipolar pulse power supplies E1 and E4 of the above embodiment are allocated, and the output lines 24a and 24b from the bipolar pulse power source E1 to E4 are connected to each pair of targets. 41a, 4115 (41 (: and 41 (1, 416 and 41 418 and 4111). Thereby, the pair of targets 41 a or 41h can be interactively exchanged via the bipolar pulse power source E1 or E4 The polarity is changed and the power supply in the form of a bipolar pulse is performed. In the present embodiment, in order to stably generate plasma before the target 41a or 41h, the target 41a or 41h adjacent to each other is made. The polarity is reversed to synchronize the bipolar pulse power supplies E1 and E4 to supply power (refer to Fig. 5). In order to perform this synchronous operation, the bipolar pulse power supplies E1 and E4 are freely communicable. The integrated control means 5 formed by the CPU connected to the second CPU circuit 21. Then, In the short-circuit state of the switching transistor SW0 for short-circuiting the output of each of the bipolar pulse power supplies E1 and E4, the first and fourth switching transistors SW1 and SW4 are provided in each of the bipolar pulse power supplies E1 and E4. In the second and third switching transistors SW2 and SW3, the timing of the ON/PFF between the two is reversed, and the polarity of the target 41a or 41h adjacent to each other is reversed. After the switching transistors SW1 and SW4 are operated, the short circuit of the switching transistor SW0 is released by the output from the integrated control means 5, and 41a and 41c of each of the target targets are 41a and 41c. 41e and 41g are output. -20- 201010260 Next, by switching from the integrated control means 5, the switching transistors SW0 for short-circuiting the outputs of the bipolar pulse power supplies E1 to E4 are short-circuited, and each switching is performed. The crystals SW1 and SW4 are switched, and then the short circuit of the switching transistor SW0 is released by the output from the integrated control means, and the other targets 41b, 41d, 41f, and 41h are outputted. Repeatedly performing the above control, in each The target 41 a to 41 h is supplied with bipolar pulsed power at a specific frequency, and is operated synchronously. In this synchronous operation, it is only necessary to control the bipolar via the integrated control means 5. The switching element for short-circuiting the output of the pulse power supply E1 or E4 - SW0 ON / OFF switching timing can be synchronized. Therefore, it is possible to have sufficient margin to make the switching elements SW1 and SW4 of each of the bipolar pulse power supplies E1 and E4 The operation, even if there are individual differences in the switching elements or control circuits of the bipolar pulse power supplies, the synchronous operation is also easy. Further, each of the bipolar pulse power supplies E1 and E4 is configured to be β during sputtering, and in any of the bipolar pulse power supplies, the output current la detected by the detecting circuit 25 exceeds the constant output current 値Ic. At this time, the above-described micro-arc processing is performed by switching the switching transistor SW0 for output short-circuit by the arc detecting control circuit 23 of the bipolar pulse power supply. When performing micro-arc processing at any of the bipolar pulse power sources, if a pair of targets connected to the output cables 24a, 24b from the bipolar pulse power source are connected, and a pair of targets Other targets connected to the output cables 24a, 24b from other bipolar pulse power sources are adjacent to each other, and the potentials of the two are mutually identical, so that the arc can be easily removed. 201010260 Discharge is used for arc suppression. In the present embodiment, when micro-arc processing is started in any of the bipolar pulse power sources El or E4, this is output to the bipolar pulse for outputting the adjacent target via the integrated control means 5. The second CPU circuit 21 of the power supply. In this case, the switching transistor for outputting the short-circuit is temporarily short-circuited by the drive circuit 23 for outputting the oscillation via the second CPU circuit 21, and the operation state of each switching transistor SW1 to SW4 is performed. In order to make the potentials match each other, the timing of the operation of each of the switching transistors ◆1 to SW4 is changed, and the short-circuit of the switching transistor SW0 for outputting the short-circuit is released, and is output to the standard. . In addition, in the present embodiment, the overall control means is provided for synchronously operating the bipolar pulse power supplies E1 to E4. However, any of the second CPU circuits may be used. 21 is configured as an integrated control means (main power supply), and the output of the integrated control means is used to control the @ action of the other bipolar pulse power supplies E2 and E4 (sub power supply). BRIEF DESCRIPTION OF THE DRAWINGS [Fig. 1] A schematic diagram showing the configuration of a bipolar pulse power supply of the present invention [Fig. 2] A diagram illustrating the output control of the bipolar pulse power supply of the present invention [Fig. 3] The current change during micro-arc processing at the bipolar pulsed power supply of the technology is illustrated. -22- 201010260 [Fig. 4] A diagram illustrating the microarc treatment at the bipolar pulse power supply of the present invention. Fig. 5 is a view for explaining changes in current during microarc processing at the bipolar pulse power supply of the present invention. Fig. 6 (a) and (b) are diagrams for explaining waveforms of an output voltage and an output current of one of the electrodes. Fig. 7 is a view showing a schematic φ of a sputtering apparatus using the power supply unit of the present invention. Fig. 8 is a view for explaining output control of the power supply device of the present invention. [Description of main component symbols] 1 '·DC power supply unit 2 : Oscillation unit 22 : Bridge circuit 24a ' 24b · Output cable • 25 : Output current, voltage detection circuit 27 : Arc detection control circuit E : Bipolar pulse Power supply SW0 or even SW4: switching element ΤΙ, T2: electrode (target) -23-