201216446 六、發明說明： 【發明所屬之技術領域】 [0001] 本發明係關於一種功率模組，特別關於一種應用於電源 變換器之功率模組。 【先前技術·】 [0002] 高效率和高功率密度一直是業界對電源變換器的要求。 高效率意味著減少能耗，利於節能減排保護環境，並減 少使用成本。高功率密度則意味著體積小、重量輕，減 少運輸成本和空間需求，從而減少建設成本。因此，電 源領域對高效率、高功率密度的追求將永不停息。 電源變換器由於用途不同，其種類較多。由轉換電能類 型來分，其可分為：非隔離型AC/DC電源變換器，例如， 由一個用於功率因數校正（下稱PFC電路）的AC/DC轉換 電路組成；非隔離型DC/DC電源變換器；隔離型DC/DC變 換器；隔離型AC/DC電源變換器，例如，由一個PFC電路 加一個或者多個DC/DC變換器而成；DC/AC、AC/AC等等 。由於需要轉換的電能性質和轉換的級數不同，各種變 換器的容易達成的功率密度和效率也不盡相同。以隔離 型AC/DC電源變換器為例，目前業界普遍的功率密度為 10W/inch3，效率為90%左右。非隔離型AC/DC電源變換 器、隔離型DC/DC變換器和DC/AC的效率和功率密度則會 更高些。 電源變換器的高效率意味著低能耗。如效率90%時，其轉 換能耗約為整個電源變換器總輸入能量的10%。而效率 91%的電源變換器，其轉換能耗則降低為總輸入能量的9% 。也就是說，效率每提升一個點，其能耗就較90%效率的 099134587 表單編號 A0101 第 4 頁/共 44 頁 0992060394-0 201216446 電源變換器降低ίο%，極為可觀。事實上，電源變換器效 率提升的努力常常以0. 5%甚至0. 1%的量級進行。 電源變換器的能耗主要由通態損耗和開關損耗特別是有 源器件的開關損耗組成。開關損耗受工作頻率的影響較 大。電源變換器，特別是開關電源變換器，為降低音頻 噪音，其工作頻率通常在20kHz以上。其實際工作頻率的 選擇受無源器件特別是磁元件的影響較大。若磁元件體 積小，為了可靠工作，通常需要高頻率來降低其工作磁 通密度從而帶來高開關損耗；或者減小磁性元件中線組 〇 的線徑並增加匝數，從而增加通態損耗，均帶來高損耗 。反之，若磁元件體積大，則可以在保證可靠工作的前 提下降低工作頻率從而降低開關損耗；也可以增加磁性 元件中線組的線徑或者減小匝數，從而降低通態損耗， 以降低總損耗，得到高效率。 因此，不難理解，提升電源内部的空間利用率，是得到 高功率密度或者高效率的關鍵因素之一。空間利用率越 高，留給對電源變換效率很重要的無源器件特別是磁性 ® 元件的空間就越大，就更容易使用到大體積的無源元件 ，從而提升電源效率。也可以通過使用大體積的無源器 件來增加電源總功率，從而提升電源變換器的功率密度 。所以，高的電源空間利用率，更易於在特定功率密度 下達成高效率或者在特定效率下達成高功率密度，也有 機會高功率密度和高效率兼顧。 半導體器件是決定電源變換器效率的重要因素之一。但 使用半導體器件，往往不可避免的需要使用對電變換效 率無益的額外材料，如：保護半導體的封裝材料、幫助 099134587 表單編號A0101 第5頁/共44頁 0992060394-0 201216446 散熱的散熱器、固定半導體器件的夾具等等。這些材料 在電源變換器内部的比例越大，電源的内部空間利用率 就越差。而目前優秀的產品，已經很好地利用了電源内 π空間。也正因為此，功率半導體器件佔用的空間體積 ，占電源總體積的比重也越來越大，也越來越被重視。 目前業界有很多先進技術被提出，如優化散熱器，簡化 女裴等等來減少散熱器及其安裝造成的空間佔用。例如 通過新的絕緣墊片技術，捨棄螺絲、夾具等，來減小體 積’以改善電源設計。 為進一步提升電源性能，需要潘續提高空間利用率。半 導體器件本身的封裝空間利用率成為瓶頸。而集成功率 模組（Integrated Power Module, ΙΡΜ)，將多個半 導體器件集成在一個器件封裝裏，為提升封裝内的空間 利用率提供了可能。集成模組因為應用的不同，集成内 容也不盡相同：有將單個功率半導體器件與其控制器或 者驅動集成在一起的；有僅將多個功率半辱體器件集成 在一起的；有將多個半導體器件與其相應控制器或者驅 動集成在一起的》集成内蓉不同，導致考量點和難易程 度不盡相同。為有所區別，下文中提及的功率模組中， 至少包含2個功率器件，強調多個功率器件的集成。 功率模組通常集成功率器件在某些場合更會集成一些控 制、驅動元器件。常用功率器件有M〇SFET，IGBT， POWER Diode等，而控制，驅動元件常包含一些三極管 ，1C，被動元件等。由於將多個器件變成一個器件功 率模組具備使用方便、平均無故障時間長等等優勢在 很多場合被應用。由於功率模組將多個功率器件集中在 099134587 表單編號A0101 S R百/从石 0992060394-0 201216446 -起’熱量多且多點分佈，其熱管理因此變得很關鍵。 眾多現有技術，报多是在散熱能力上做優化。 現有技術一，如圖1所示，為一典型的功率模組30内部裁 面圖。該已有技術將元器件32、34和弓i線框架（1咖 frame) 35組裝。以部分功率器件的晶片Μ、％為例， 其正面電極可以通過引線鍵合（wireb〇nding)，麵片 釺焊（copper strap bonding)等方式和引線框架實 現電氣連接；其背面可以通過釺焊，銀膠，燒結，環氣 膠等方式實現和lead-frame之間的電和/或機械連接。 〇 元器件和引線框架麵裝舞以後，將需要保護的區域使用 封料（molding compound) 36包覆，以便起到機械， 防塵，防潮，絕緣保護之功用。此結構具有價格低的優 勢。 該現有技術，其散熱面由封料絕緣，該散熱面同時擔負 機械保護之角色，所以，厚度也會較大，通常大於〇· 5瓜田 。通常而言’ molding compound的熱導率在lW/m. K左 右。從晶片表面向case的傳導熱阻的計算式為， 〇 導熱系數χ面積/ w •幻x 乂(瓣/) 對於10mm乘以10mm見方的面積，厚度在0.5mm，封料的 熱導率假設為lW/m. K時，熱阻就高達5K/W。由此可見’ 通常而言’此類封裝的散熱性能較差，即，以功率半導 體器件為例，從晶片的junction至case的熱阻（Rjc) 較大《而且，由於封料較低的導熱係數，其橫向熱擴散 099134587 表單編號 AG1G1 % 7 1/^ 44 1 〇992〇6〇394 〇 201216446 的能力也較低，因此往往會出現熱集中點（熱點），危 害器件的可靠性以及使用壽命。 所以，現有技術一的散熱能力較差，不適合散熱要求高 的場合。為了優化功率模組的性能，有更多已有技術被 提出。 現有技術二，如圖2所示’在現有技術一的基礎上，在 molding的一側，增加一散熱單元31 ’由於此散熱單元 的熱導率較高，例如銅的熱導率高於3〇〇w/m. κ，因此， 這樣可以使得模組的均溫性能有所增加，在一定程度上 緩解熱點問題’以此增加：模:組的熱管理能力。但由於該 散熱單元通常也被要求電絕緣，其與導線框架35間往往 被填充封料36。由於molding工藝的限制，該封料層的 厚度一般大於0· 2mm，通常需要在〇 3mm以上，按照現有 技術一中的計算方法，其l〇mmXl〇mm面積對應的熱阻在 3K/W左右。即，此結構的整體散熱性能雖有較大改善， 但依然較差。 . η .. :. ：.. . .. 現有技術二，如圖3所示：在覆铜^陶瓷基板（Direct Bonded Copper，DBC) 31a上系歲電路圖形，此DBC板 作為元器件的安裝載板，將元器件32、34與胍板組裝， 針對部分半導體晶片需要使用引線鍵合（wire b〇nding )工藝兀成半導體晶片32、34正面電極和dBC基板/引線 框架35的電信號連接。此結構的本f是在現有技術二的 基礎上’採用熱導係數較高的陶竞介質層，來替代封料 層。由於f用的二氧化二㈣兗的導熱係數在24W/mK左 右，這相對於封料的1W/m· κ有很大的改進。對於1〇隨乘 以lOnrn見方的面積的DBC板（陶究厚度〇38賴，兩側銅 0992060394-0 099134587 表單編號A0101 第8頁/共44頁 201216446 厚均為0· 3mm)其熱阻為0. 17K/W，相對現有技術一中所 舉例的5K/W有較大提升，減少90%以上》 但由於所有元器件32、34均需安裝至DBC板，因此所需的 DBC載板的面積比較大，而DBC的價格比較昂責，因此成 本相對較高。更由於DBC的生產工藝為高溫燒結，為高能 耗產品，大面積DBC板的使用也不符合當前綠色環保的科 技進步潮流。而且三氧化二鋁，的導熱係數（〜24W/m. κ 左右），雖然較封料（通常低於lW/m.K)，有很大改善 ，但是，和金屬相比（例如銅300W/m. K左右）依然相 〇 去甚遠，導致橫向熱擴散能力不夠好，所以其熱均勻性 往往不佳。因此，散熱性能奶然有進一步提升的空間。 現有技術四，如圖4所示，此庭構在現有技術三的基礎上 有所改進，在DBC板組裝元器件31a的另外一侧再組裝一 個散熱單元（散熱器）31b，這可以使得激組的均溫性能 有所提升。但是，由於大面積DBC板的應用，由此可能存 在的由DBC，散熱單元31b，封料36之間由丨於熱膨脹係數 (coefficient of thermal expansion, CTE)失配 〇 ( mi smatch )而引起的魅曲變形（warpage )會比較大 ，並可能導致可靠性的降低。如果DBC尺寸過大，且DBC 和散熱單元31b採用通用的焊料（Solder)的方式進行 時，更有可能發生焊料層氣泡過多等缺陷。而且，成本 高的問題，依然沒有解決。 現有技術五，如圖5所示，在現有技術四上集成了控制器 件或者驅動器件。由於控制器件和驅動器件自身功耗往 往不高，而又對溫度較敏感，所以通常被設計成與溫度 較高的材料熱絕緣。該現有技術就是將控制器件38或者 099134587 表單編號 A0101 第 9 頁/共 44 頁 0992060394-0 201216446 驅動器件部分作為一個單元（通過PCB板集成或者IC)， 通過熱絕緣體（PCB、Molding料或者專用填充料等等熱 絕緣體，通常導熱係數小於lW/m.K)與散熱單元31a連 接。絕緣體的生成方式可以是粘結、填充或者在表面上 鍍膜等等。這樣一來，控制器件或者驅動器件等自身功 耗小且對熱敏感的器件，就可以在封裝體中獨善其身， 而減小被功率器件的高溫影響，使得其可以集成到功率 模組中，並被可靠使用。 如前所述，目前的功率模組由於考慮通用性的影響，通 常將外殼設計成絕緣的，以簡化散熱器的安裝和選擇。 即便如現有技術四一樣，外殼是電良導體（如銅），其 往往也被設計成電絕緣。因此，模組内的金屬材料，例 如銅，往往僅僅被用作導電（lead frame，DBC銅層） 或者散熱（銅熱沉）之單一功用，少有將銅層在導電同 時，兼作和環境直接換熱的案例。因此，材料的潛能並 沒有被完整挖掘，從而降低了空間利用率。 而且，為了簡化用戶安裝散熱器，功率模組往往允許用 螺絲或者夾具將其固定到散熱器上。所以功率模組通常 被設計成允許承受較大的機械應力。為了可靠使用，功 率模組通常設計成較厚的封料以允許承受較大應力，這 樣就增加了厚度，也增加了材料成本，並大大降低空間 利用率。而且，功率模組也通常要求自己有較高的表面 平整度，以減小散熱器安裝時的應力，從而導致更大的 設計成本和模具成本。 以上可知，目前，功率模組已有技術依然有著各類問題 :如散熱性能不佳，材料浪費，可靠性設計困難，電性 099134587 表單編號A0101 第10頁/共44頁 0992060394-0 201216446 能未能充分發揮，過於強調設計的通用性而導致過度設 計（over design)，經濟性能不高等這樣或那樣的問 題。特別是其空間利用率不足，限制了其在高功率密度 或者高效率場合的應用推廣。 因此，現有技術中的功率模組方案，其性能尚不能很好 滿足高功率密度或者高效率電源的需求。 針對每一種半導體封裝，其初期投入很高。比如Molding 的模具成本，產線架設成本等等。所以，要得到價格合 理的半導體封裝，往往需要很大的產品量來支撐，以消 〇 化初期投入，並降低生產成本。所以，目前的功率模組 ，往往用在一些應用標準化的場合。如圖6所示的IGBT三 相橋模組。它被廣泛應用在逆變器、變頻器等等場合， 因為這些場合的電路很標準化，需求很一致，量也就很 大。所以，半導體廠家可以自行給出標準化封裝，以供 客戶選擇使用。 在電源變換器場合，也有功率模組被成功使用，如圖7所 示之雙相整流橋。由於絕大部分AC/DC電源變換器，需要 〇 w 輸入整流橋，所以，功率模組的需要量很大。且整流電 路很標準化，半導體廠家可以給出標準化封裝，以供客 戶選擇使用。 但是，電源變換器其他部分的功率半導體器件，雖然也 有眾多廠家嘗試給出功率模組，但很少被推廣使用。除 了如前所述，現有技術的性能有不足之外，另一個很重 要的原因是電源變換器的電路結構複雜，很難標準化。 若只針對一種電路設計給出功率模組，其量較少，成本 代價較高，也就限制了應用。 099134587 表單編號A0101 第11頁/共44頁 0992060394-0 201216446 因此，為進一步提升電源變換器的功率密度或者變換效 率，需要空間利用率高的、成本合理的功率模組解決方 案。目前的已有技術尚不能很好滿足。 【發明内容】 [0003] 有鑑於上述課題，本發明提出了一種適合電源變換器的 功率模組，用以提升功率密度或效率的解決方案，並給 出了支持該解決方案的功率模組實施方案。該方案適合 於功率密度大於15w/inch3、或者最高效率高於91%的電 源變換器，尤其適合功率密度大於20w/inch3，或者最高 效率高於93%的電源變換器場合。 為達上述目的，依據本發明之一種功率模組包含一第一 散熱單元、一第一功率器件、一導熱絕緣材料層、一第 二功率器件、一引線框架以及一封料。第一散熱單元具 有一第一區及一第二區。第一功率器件設置於第一區。 導熱絕緣材料層設置於第二區並具有一絕緣層。第二功 率器件藉由導熱絕緣材料層設置於散熱單元。引線框架 與第一功率器件及第二功率器件之至少一電性連接。封 料係包覆第一功率器件、導熱絕緣材料層、第二功率器 件及引線框架之一部分。第一散熱單元與第一功率器件 及第二功率器件之至少一電性連接。 承上所述，由於本發明之功率模組集成了複數功率器件 ，故可大幅提升功率密度或效率。另外，由於本發明之 第一功率器件非藉由導熱絕緣材料層設置於散熱單元， 而該導熱絕緣材料層通常可由導熱基板實現，故可降低 導熱基板之成本。此外，通過本發明所揭露的，用以提 升電源變換器功率密度或者效率的封裝方法和結構，可 099134587 表單編號 A0101 第 12 頁/共 44 頁 0992060394-0 201216446 以獲得與現有技術相比，更佳的熱性能，電性能，經濟 性能，EMC性能與更高的可靠性。其内部空間利用率很高 ，使用方便，非常有利於提高變換器功率密度或者效率 。而本發明給出的具體功率模組具體實施，也非常可行 有效。本發明非常適合用以提升電源變換器的整體性能 和性價比。 【實施方式】 [0004] Ο ο 099134587 以下將參照相關圖式，說明依本發明較佳實施例之一種 功率模組，其中相同的元件將以相同的參照符號加以說 明。 請參照圖8所示，本發明較佳實施例之一種功率模組10可 例如應用於電源變換器（power converter)或是其他 需要功率變換的裝置上。其中，電源變換器可為交流/直 流（AC/DC)或直流/交流（DC/AC)變換器或隔離型DC/ DC變換器。若應用於電源變換器上，功率模組10則可應 用於電源變換器之功率因數校正部分（power factor correction，PFC)、DC/DC—次側部分（以下稱 D2D_Pri)或DC/DC二次側部分（以下稱D2D_Sec)。 功率模組10係為一封裝體，包含一第一散熱單元（heat sink) 11、一第一功率器件（p.ower chip) 12、一導 熱絕緣材料層13、一第二功率器件14、一引線框架（ lead frame) 15以及一封料（molding material) 16 。第一散熱單元11設置於封裝體之一底侧，並具有一第 一區111及一第二區112。第一功率器件12設置於第一區 111，導熱絕緣材料層13設置於第二區112。第二功率器 件14設置於導熱絕緣材料層13並與引線框架1 5電性連接 表單編號A0101 第13頁/共44頁 0992060394-0 201216446 。第一散熱單元11係與第一功率器件12及第二功率器件 14之至少一電性連接。封料i 6係包覆第一功率器件丨2、 導熱絕緣材料層13、第二功率器件14及引線框架15之至 少一部分，並構成為封裝體的主要外觀。 第一散熱單元11可以是一獨立部件或與引線框架15 一體 成型，並可為電和熱的良導體，例如銅。於此，散熱單 兀11係作為第一功率器件12的載板。第一散熱單元^可 兀全没置於封料16内、或部分位於封料16外、或完全位 於封料16外。 第功率器件例如為M=0:SF，ET的、.器件，.對於一個mosfET的 器件而言，其通常有兩個相對平行的面：上表面和下表 面。上表面上往往會設置兩個電極，s〇urce*gate，而 下表面電極為drain，下表面利用一鍵接材料層17可直接 與散熱單元11組裝，鍵接材料層17可包含釺焊的焊料、 導電銀膠、或燒結金屬材料等。 第一散熱單元11自身的傳導熱阻通常也非常低，因此， 可以獲得非常低的器件結點至第散熱單元丨丨外殼的熱 阻（Rjc)，且，由於第一散爇單元u的熱容較大，因此 ，功率器件的抗熱衝擊的性能也很優良。總而言之即 直接組裝至第一散熱單元丨丨的第一功率器件12的熱性能 非常優良。且由於第一散熱單元丨！的存在，功率模組1〇 的熱會較均勻，更有利於熱管理。當然，此處僅以功率 器件為例進行描述。 由於本實施例之封裝類型為電源内部使用，為達成更高 空間利用率和提升功率模組1G性能，該模組表面無需與 099134587 内部電路全部電絕緣。以降低絕緣成本和絕緣造成的空 表單編號A0101 第14頁/共44頁 0992060394-0 201216446 間浪費，散熱能力衰減等不良。所以在一些具體場合， 可以直接利用第一散熱單元11作為導電通道，由於第一 散熱單元11通常為銅、鋁等電的優良導體，且厚度相對 較厚，其導電性能極佳。因此，可以獲得更佳的電氣性 能，減小發熱量，從而進一步改善封裝體的熱性能。更 進一步，第一散熱單元11可以直接作為引腳（Pin)使用 ，或者與至少一個引腳相連，即，引腳可以是和第一散 熱單元11為一體成型的，或者引腳和第一散熱單元11通 過導線接合（wire bonding)、焊接、釺焊、導電膠枯 〇 接等方式實現良好的電連接，以更充分利用該表面之電 良導體。這樣大大減小了器件到第一散熱單元11的熱阻 ，也使第一散熱單元11這個電良導體同時被發掘熱和電 的能力。從而提升空間利用率，以利於提升電源變換器 功率密度或變換效率。 由於D2D_Pri、D2D_Sec等場合中，全橋電路極為常用。 所以，本實施例之功率模組10可被用在全橋電路中。圖9 為全橋電路示意圖，要滿足該應用，功率模組10至少要 ❹ 能夠排布下8個功能引腳，即Vin、GND、VA、VB、G1、 G2、G3、G4 ° 為進一步提升功率模組10性能，充分發掘潛力，功率模 組10當具備雙面散熱能力。在本實施例中，功率模組之 兩個最大的主表面，一前表面（封料16) A1和一後表面 (散熱單元11及封料16) A2，均能用來散熱。這樣就可 以大大增加有效散熱能力，更容易在低損耗場合下自行 散熱而無需額外散熱器，大大提升電源的内部空間利用 率。為了實現更好的散熱特性，封料的厚度越薄越好。 099134587 表單編號A0101 第15頁/共44頁 0992060394-0 201216446 為減少使用時的機械應力，以使模組更容易設計得薄， 該功率模組也可以不必預設螺絲安裝孔。以進一步提升 空間利用率。若需安裝額外散熱器，可選擇無螺絲之解 決方案，如直接粘結等。 這樣一來，本實施例之功率模組10將大大提升該類型封 裝的量，也很符合目前和未來電源變換器的需求，並能 提升電源變換器的空間利用率，從而提升電源的功率密 度或者效率。 另外，請續參照圖8所示，第二功率器件14係藉由一導熱 絕緣材料層13設置於第一散熱單元11上，而非直接置於 散熱單元11。導熱絕緣材料層13可具有一絕緣層132，比 如用陶瓷片絕緣。導熱絕緣材料層13比如為金屬基板或 金屬化陶兗基板，例如覆銅陶竟基板（direct bonded copper, DBC)、金屬化陶瓷片上組裝厚銅電路層、覆 銘陶究基板（direct bonded aluminum, DBA)、銘 基板、銅基板，或其他形式的高導熱基板。於此導熱絕 緣材料層13以DBC基板為例，導熱絕緣材料層13可包含一 導熱層131、一絕緣層132及一線路層133，其中導熱層 131及線路層133可為銅，絕緣層132可為陶瓷。 以常用的DBC板為例，相對於現有技術，由於本發明僅有 一部分元器件（第二功率器件14)，安裝於導熱絕緣材 料層13上，因為搭載在其上的元器件數量減少，DBC板面 積也可以相應減小，如此可以降低封裝的材料成本，提 高封裝的經濟性能。且，由於DBC面積的減小，由於DBC 和散熱單元11，封料1 6之間熱脹系數（c 〇 e f f i c i e n t of thermal expansion, CTE)不一致而導致的撓曲 099134587 表單編號A0101 第16頁/共44頁 0992060394-0 201216446 (warpage)現象也會有所緩解。 ^疋因為由於不同材料 CTE之間的適配而引起的撓曲通常隨— L γ 者尺寸的增加而加劇 。如此，可以降低封裝體内的應力，從而進步提言封 裝體的可靠性。所以，由於部分器 ，门 .^ 干（第—功率器件12 )已經直接與散熱單元11相連，相 +於現有技術，本發 明之功率杈組需要絕緣的材料明顯減少，不盡降低了成 本’更提升了熱管理能力，還更有利 又’刊於減少各材料CTE不 匹配造成的可靠性設計難度。201216446 VI. Description of the Invention: [Technical Field of the Invention] [0001] The present invention relates to a power module, and more particularly to a power module applied to a power converter. [Prior Art·] [0002] High efficiency and high power density have always been the requirements of the power converters in the industry. High efficiency means reducing energy consumption, saving energy and reducing emissions, and reducing operating costs. High power density means small size and light weight, reducing transportation costs and space requirements, thus reducing construction costs. Therefore, the pursuit of high efficiency and high power density in the power field will never stop. Power converters have many types due to their different uses. Divided by the type of converted electrical energy, it can be divided into: non-isolated AC/DC power converter, for example, consisting of an AC/DC converter circuit for power factor correction (hereinafter referred to as PFC circuit); non-isolated DC/ DC power converter; isolated DC/DC converter; isolated AC/DC power converter, for example, one PFC circuit plus one or more DC/DC converters; DC/AC, AC/AC, etc. . Due to the nature of the electrical energy that needs to be converted and the number of stages of conversion, the easily achieved power density and efficiency of the various converters are also different. Taking an isolated AC/DC power converter as an example, the current power density is generally 10 W/inch 3 and the efficiency is about 90%. The efficiency and power density of non-isolated AC/DC power converters, isolated DC/DC converters, and DC/AC are higher. The high efficiency of the power converter means low energy consumption. At 90% efficiency, the conversion energy consumption is approximately 10% of the total input energy of the entire power converter. For a power converter with 91% efficiency, the conversion energy consumption is reduced to 9% of the total input energy. That is to say, for every point of efficiency improvement, the energy consumption is 90% more efficient. 099134587 Form No. A0101 Page 4 of 44 0992060394-0 201216446 The power converter is reduced by ίο%, which is extremely impressive. In fact, the efficiency of the power converter is often performed on the order of 0.5% or 0.1%. The power consumption of a power converter is mainly composed of on-state losses and switching losses, especially the switching losses of active devices. Switching losses are greatly affected by the operating frequency. Power converters, especially switching power converters, typically operate at frequencies above 20 kHz to reduce audio noise. The choice of its actual operating frequency is greatly influenced by passive components, especially magnetic components. If the magnetic component is small in size, in order to work reliably, high frequency is usually required to reduce the working magnetic flux density to cause high switching loss; or reduce the wire diameter of the wire group in the magnetic component and increase the number of turns, thereby increasing the on-state loss. Both bring high losses. Conversely, if the magnetic component is bulky, the operating frequency can be reduced to reduce the switching loss while ensuring reliable operation; the wire diameter of the wire group in the magnetic component can be increased or the number of turns can be reduced, thereby reducing the on-state loss and reducing the on-state loss. Total loss, get high efficiency. Therefore, it is not difficult to understand that increasing the space utilization inside the power supply is one of the key factors for obtaining high power density or high efficiency. The higher the space utilization, the more space is left for passive components, especially magnetic ® components, which are important for power conversion efficiency, making it easier to use large-volume passive components to improve power efficiency. It is also possible to increase the power density of the power converter by using a large volume of passive components to increase the total power of the power supply. Therefore, high power space utilization makes it easier to achieve high efficiency at a specific power density or achieve high power density at a specific efficiency, as well as the opportunity for high power density and high efficiency. Semiconductor devices are one of the important factors determining the efficiency of power converters. However, when using semiconductor devices, it is inevitable to use additional materials that are not beneficial to the efficiency of electrical conversion, such as: packaging materials for protecting semiconductors, help 099134587 Form No. A0101 Page 5 / Total 44 Page 0992060394-0 201216446 Heat sink, fixed Fixtures for semiconductor devices and the like. The greater the proportion of these materials inside the power converter, the worse the internal space utilization of the power supply. At present, excellent products have made good use of the π space in the power supply. For this reason, the volume of space occupied by power semiconductor devices, which accounts for a larger proportion of the total volume of power supplies, is also becoming more and more important. At present, many advanced technologies have been proposed in the industry, such as optimizing the radiator, simplifying the niece, etc. to reduce the space occupied by the radiator and its installation. For example, by using new insulating gasket technology, discard screws, clamps, etc. to reduce the volume' to improve the power supply design. In order to further improve the power performance, it is necessary to continue to improve space utilization. The package space utilization of the semiconductor device itself becomes a bottleneck. The integrated power module (ΙΡΜ) integrates multiple semiconductor devices in a single device package, making it possible to increase the space utilization within the package. Integrated modules vary in application content due to different applications: there is a single power semiconductor device integrated with its controller or driver; there are only multiple power and abusive devices integrated; there will be multiple The integration of semiconductor devices with their respective controllers or drivers is different, resulting in different considerations and difficulty levels. To make a difference, the power modules mentioned below contain at least 2 power devices, emphasizing the integration of multiple power devices. Power modules typically integrate power devices and in some cases integrate control and drive components. Commonly used power devices include M〇SFET, IGBT, POWER Diode, etc., while control, drive components often include some triodes, 1C, passive components, and so on. The ability to turn multiple devices into a single device power module is easy to use, has a long mean time between failures, and so on. Since the power module concentrates multiple power devices on 099134587 Form No. A0101 S R hundred / from Shi 0992060394-0 201216446 - the heat is more and more distributed, so its thermal management becomes critical. Many existing technologies, the report is mostly optimized in the heat dissipation capacity. The prior art 1, as shown in FIG. 1, is a typical internal view of the power module 30. This prior art assembles the components 32, 34 and the bow frame 35. Taking the wafer Μ and % of the partial power device as an example, the front electrode can be electrically connected to the lead frame by wire bonding, copper strap bonding, etc.; the back surface can be soldered by soldering. , silver glue, sintering, epoxy glue, etc. to achieve electrical and / or mechanical connection between the lead-frame.以后 After the component and lead frame are worn, the area to be protected is covered with a molding compound 36 to provide mechanical, dustproof, moisture-proof, and insulation protection. This structure has the advantage of low price. In the prior art, the heat dissipating surface is insulated by the sealing material, and the heat dissipating surface plays the role of mechanical protection at the same time, so the thickness is also large, and is usually larger than 〇·5 瓜田. Generally, the thermal conductivity of the ' molding compound is around lW/m. K. The calculation of the conduction thermal resistance from the surface of the wafer to the case is: 〇thermal conductivity χ area / w • illusion x 乂 (valve /) For 10 mm by 10 mm square area, thickness is 0.5 mm, thermal conductivity of the sealing material is assumed When it is lW/m. K, the thermal resistance is as high as 5K/W. It can be seen that 'generally' such packages have poor heat dissipation performance, that is, in the case of power semiconductor devices, the thermal resistance (Rjc) from the junction of the wafer to the case is large. Moreover, due to the lower thermal conductivity of the sealing material. , its lateral thermal diffusion 099134587 Form No. AG1G1 % 7 1 / ^ 44 1 〇 992 〇 6 〇 〇 16 201216446 is also less capable, so there will often be heat concentration points (hot spots), jeopardizing the reliability and service life of the device. Therefore, the heat dissipation capability of the prior art 1 is poor, and it is not suitable for occasions with high heat dissipation requirements. In order to optimize the performance of the power module, more existing technologies have been proposed. In the prior art 2, as shown in FIG. 2, on the basis of the prior art, on the side of the molding, a heat dissipating unit 31' is added. Since the heat conductivity of the heat dissipating unit is high, for example, the thermal conductivity of copper is higher than 3 〇〇w/m. κ, therefore, this can increase the average temperature performance of the module, and alleviate the hotspot problem to a certain extent'. This increases: the thermal management capability of the model: group. However, since the heat dissipating unit is also generally required to be electrically insulated, it is often filled with the sealing material 36 between the heat dissipating unit and the lead frame 35. Due to the limitation of the molding process, the thickness of the sealing layer is generally greater than 0. 2 mm, and generally needs to be more than 3 mm. According to the calculation method of the prior art, the thermal resistance corresponding to the area of l〇mmXl〇mm is about 3K/W. . That is, although the overall heat dissipation performance of this structure is greatly improved, it is still poor. η .. :. :.. . . . . . 2, the prior art 2, as shown in Figure 3: on the copper-clad ceramic substrate (Direct Bonded Copper, DBC) 31a is the old circuit figure, this DBC board as a component installation The carrier board assembles the components 32, 34 and the dummy board, and the electrical wiring connection of the semiconductor wafer 32, 34 front electrode and the dBC substrate/lead frame 35 is required for a part of the semiconductor wafer by a wire bonding process. . The f of this structure is based on the prior art 2, using a ceramic layer with a higher thermal conductivity to replace the sealing layer. Since the thermal conductivity of di(tetra)phosphorus oxide for f is around 24 W/mK, this is a significant improvement over the 1 W/m·κ of the sealant. For the DBC board with 1 〇 见 On On On On On On 陶 陶 陶 陶 陶 陶 陶 陶 陶 陶 陶 陶 陶 陶 陶 陶 陶 陶 陶 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 0. 17K/W, compared with the 5K/W exemplified in the prior art, there is a large increase, which is more than 90%. However, since all components 32 and 34 need to be mounted to the DBC board, the required DBC carrier board is required. The area is relatively large, and the price of DBC is relatively high, so the cost is relatively high. Moreover, because DBC's production process is high-temperature sintering, it is a high-energy-consuming product, and the use of large-area DBC boards does not conform to the current trend of green technology advancement. Moreover, the thermal conductivity of aluminum oxide, which is about ~24 W/m. κ, is greatly improved compared with the sealing material (usually lower than lW/mK), but compared with metal (for example, copper 300 W/m. K is still far away, resulting in a lack of lateral thermal diffusion capacity, so its thermal uniformity is often poor. Therefore, the heat dissipation performance has room for further improvement. Prior art 4, as shown in FIG. 4, the court structure is improved on the basis of the prior art 3, and a heat dissipating unit (heat sink) 31b is assembled on the other side of the DBC board assembling component 31a, which can make The group's average temperature performance has improved. However, due to the application of the large-area DBC board, there may be a relationship between the DBC, the heat dissipating unit 31b, and the sealing material 36 due to the coefficient of thermal expansion (CTE) mismatch mi( mi smatch ). The warpage will be large and may result in reduced reliability. If the DBC size is too large, and the DBC and the heat sink unit 31b are used in a general solder (Solder) manner, defects such as excessive bubbles in the solder layer are more likely to occur. Moreover, the problem of high cost remains unresolved. Prior art 5, as shown in Fig. 5, a controller device or a driving device is integrated in the prior art 4. Since the control device and the driver device are inherently less power consuming and temperature sensitive, they are typically designed to be thermally insulated from higher temperature materials. The prior art is to use the control device 38 or the 099134587 form number A0101 page 9/44 page 0992060394-0 201216446 drive device portion as a unit (through PCB board integration or IC), through a thermal insulator (PCB, Molding material or dedicated padding) A thermal insulator such as a material, usually having a thermal conductivity of less than 1 W/mK, is connected to the heat dissipating unit 31a. The insulator can be formed by bonding, filling or plating on the surface, and the like. In this way, a control device or a driving device, such as a device that consumes less power and is sensitive to heat, can be isolated in the package, and can be reduced in the high temperature effect of the power device, so that it can be integrated into the power module, and Used reliably. As mentioned earlier, current power modules are typically designed to be insulated to account for versatility, simplifying the installation and selection of the heat sink. Even though the outer casing is a good conductor (e.g., copper) as in the prior art four, it is often designed to be electrically insulated. Therefore, the metal material in the module, such as copper, is often only used as a single function of a conductive (lead frame, DBC copper layer) or heat dissipation (copper heat sink), and the copper layer is electrically conductive, and the environment is directly Case of heat transfer. As a result, the potential of the material is not fully excavated, reducing space utilization. Moreover, to simplify the user's installation of the heat sink, the power module often allows it to be attached to the heat sink with screws or clamps. Therefore, power modules are usually designed to withstand large mechanical stresses. For reliable use, power modules are typically designed to be thicker to allow for greater stress, which increases thickness, increases material costs, and greatly reduces space utilization. Moreover, power modules often require higher surface flatness to reduce stress during heat sink installation, resulting in greater design and mold costs. As can be seen from the above, at present, the power module has various problems in the prior art: such as poor heat dissipation performance, material waste, and reliability design difficulty, electrical 099134587 Form No. A0101 Page 10 / Total 44 Page 0992060394-0 201216446 Can fully play, over-emphasizing the versatility of design, resulting in over design, economic performance is not high or the like. In particular, its insufficient space utilization limits its application in high power density or high efficiency applications. Therefore, the performance of the power module solution in the prior art cannot meet the requirements of high power density or high efficiency power supply. For each type of semiconductor package, its initial investment is high. For example, Molding's mold cost, production line erection cost, and so on. Therefore, in order to obtain a reasonably priced semiconductor package, a large amount of product is often required to support the initial investment and reduce the production cost. Therefore, current power modules are often used in some applications where standardization is required. The IGBT three-phase bridge module shown in Figure 6. It is widely used in inverters, inverters, etc., because the circuits in these cases are standardized, the requirements are very consistent, and the amount is large. Therefore, semiconductor manufacturers can provide their own standardized packaging for customers to choose. In the case of power converters, power modules have also been successfully used, as shown in Figure 7 for a two-phase rectifier bridge. Since most AC/DC power converters require a 整流 w input rectifier bridge, the power module is in great demand. And the rectifier circuit is standardized, and semiconductor manufacturers can provide a standardized package for customers to choose. However, power semiconductor devices in other parts of the power converter, although many manufacturers try to give power modules, are rarely promoted. In addition to the performance of the prior art as described above, another important reason is that the power converter has a complicated circuit structure and is difficult to standardize. If a power module is given for only one circuit design, the amount is small and the cost is high, which limits the application. 099134587 Form No. A0101 Page 11 of 44 0992060394-0 201216446 Therefore, in order to further improve the power density or conversion efficiency of the power converter, a space-efficient and cost-effective power module solution is required. The current state of the art is not well met. SUMMARY OF THE INVENTION [0003] In view of the above problems, the present invention proposes a power module suitable for a power converter, a solution for improving power density or efficiency, and provides a power module implementation supporting the solution. Program. This solution is suitable for power converters with a power density greater than 15w/inch3 or a maximum efficiency higher than 91%, especially for power converters with a power density greater than 20w/inch3 or a maximum efficiency greater than 93%. To achieve the above object, a power module according to the present invention comprises a first heat dissipating unit, a first power device, a layer of thermally conductive insulating material, a second power device, a lead frame, and a material. The first heat dissipation unit has a first zone and a second zone. The first power device is disposed in the first region. The layer of thermally conductive insulating material is disposed in the second region and has an insulating layer. The second power device is disposed on the heat dissipation unit by a layer of thermally conductive insulating material. The lead frame is electrically connected to at least one of the first power device and the second power device. The seal coats the first power device, the layer of thermally conductive insulating material, the second power device, and a portion of the lead frame. The first heat dissipation unit is electrically connected to at least one of the first power device and the second power device. As described above, since the power module of the present invention integrates a plurality of power devices, the power density or efficiency can be greatly improved. In addition, since the first power device of the present invention is not disposed on the heat dissipating unit by the layer of the thermally conductive insulating material, and the layer of the thermally conductive insulating material can be realized by the thermally conductive substrate, the cost of the thermally conductive substrate can be reduced. In addition, the packaging method and structure for improving the power density or efficiency of the power converter disclosed by the present invention can be obtained by the method of 099134587 Form No. A0101, Page 12/44, 0992060394-0 201216446, to obtain more compared with the prior art. Good thermal performance, electrical performance, economic performance, EMC performance and higher reliability. Its internal space utilization is very high and easy to use, which is very beneficial to improve the power density or efficiency of the converter. The specific implementation of the specific power module given by the present invention is also very feasible and effective. The invention is well suited for improving the overall performance and cost performance of a power converter. [Embodiment] A power module in accordance with a preferred embodiment of the present invention will be described with reference to the accompanying drawings, wherein like elements will be described with the same reference numerals. Referring to Figure 8, a power module 10 in accordance with a preferred embodiment of the present invention can be applied, for example, to a power converter or other device that requires power conversion. Among them, the power converter can be an AC/DC or DC/AC converter or an isolated DC/DC converter. If applied to a power converter, the power module 10 can be applied to a power factor correction (PFC), a DC/DC-sub-side portion (hereinafter referred to as D2D_Pri) or a DC/DC quadratic power converter. Side part (hereinafter referred to as D2D_Sec). The power module 10 is a package, and includes a first heat sink 11 , a first power device (p. ower chip) 12 , a thermal conductive material layer 13 , a second power device 14 , and a first heat sink 11 . A lead frame 15 and a molding material 16 are provided. The first heat dissipating unit 11 is disposed on one of the bottom sides of the package and has a first area 111 and a second area 112. The first power device 12 is disposed in the first region 111, and the thermally conductive insulating material layer 13 is disposed in the second region 112. The second power device 14 is disposed on the thermally conductive insulating material layer 13 and electrically connected to the lead frame 115. Form No. A0101 Page 13 of 44 0992060394-0 201216446. The first heat dissipation unit 11 is electrically connected to at least one of the first power device 12 and the second power device 14. The sealing material i 6 is coated with at least a portion of the first power device 2, the thermally conductive insulating material layer 13, the second power device 14, and the lead frame 15, and is configured as a main appearance of the package. The first heat dissipating unit 11 may be a separate component or integrally formed with the lead frame 15, and may be a good conductor of electricity and heat, such as copper. Here, the heat sink unit 11 serves as a carrier of the first power device 12. The first heat dissipating unit can be disposed not within the encapsulant 16, or partially outside of the encapsulant 16, or entirely outside the encapsulant 16. The first power device is, for example, M = 0: SF, ET, .. device. For a mosfET device, it usually has two relatively parallel faces: the upper surface and the lower surface. Two electrodes are often disposed on the upper surface, s〇urce*gate, and the lower surface electrode is drain, and the lower surface is directly assembled with the heat dissipation unit 11 by using a bonding material layer 17, and the bonding material layer 17 may include soldered Solder, conductive silver paste, or sintered metal material. The conduction heat resistance of the first heat dissipation unit 11 itself is also generally very low, and therefore, the thermal resistance (Rjc) of the very low device node to the heat dissipation unit casing can be obtained, and, due to the heat of the first heat dissipation unit u The capacitance is large, and therefore, the thermal shock resistance of the power device is also excellent. In summary, the thermal performance of the first power device 12 directly assembled to the first heat sink unit is very excellent. And because of the first heat sink unit! The existence of the power module 1 〇 will be more uniform, which is more conducive to thermal management. Of course, only the power device is taken as an example here. Since the package type of the embodiment is used internally by the power supply, in order to achieve higher space utilization and improve the performance of the power module 1G, the surface of the module does not need to be completely electrically insulated from the internal circuit of 099134587. To reduce the insulation cost and the insulation caused by the empty form number A0101 page 14 / a total of 44 pages 0992060394-0 201216446 waste, heat dissipation and other attenuation. Therefore, in some specific occasions, the first heat dissipating unit 11 can be directly used as a conductive path. Since the first heat dissipating unit 11 is generally an excellent conductor of copper, aluminum, etc., and has a relatively thick thickness, the conductive performance is excellent. Therefore, it is possible to obtain better electrical performance and reduce heat generation, thereby further improving the thermal performance of the package. Further, the first heat dissipation unit 11 can be directly used as a pin (Pin) or connected to at least one pin, that is, the pin can be integrally formed with the first heat dissipation unit 11, or the pin and the first heat dissipation. The unit 11 achieves a good electrical connection by means of wire bonding, soldering, soldering, conductive bonding, etc., to make better use of the electrical conductor of the surface. This greatly reduces the thermal resistance of the device to the first heat dissipating unit 11, and also enables the first heat dissipating unit 11 to be exposed to heat and electricity at the same time. Thereby improving the space utilization, in order to improve the power converter power density or conversion efficiency. Due to D2D_Pri, D2D_Sec and other occasions, full-bridge circuits are extremely common. Therefore, the power module 10 of the present embodiment can be used in a full bridge circuit. Figure 9 is a schematic diagram of a full-bridge circuit. To meet this application, the power module 10 must be capable of arranging at least eight function pins, namely Vin, GND, VA, VB, G1, G2, G3, and G4 ° for further power boosting. The performance of the module 10 fully exploits the potential, and the power module 10 has a double-sided heat dissipation capability. In this embodiment, the two largest major surfaces of the power module, a front surface (sealing material 16) A1 and a rear surface (heat dissipating unit 11 and sealing material 16) A2, can be used for heat dissipation. This greatly increases the effective heat dissipation capability, making it easier to dissipate itself in low-loss situations without the need for additional heat sinks, greatly increasing the internal space utilization of the power supply. In order to achieve better heat dissipation characteristics, the thinner the thickness of the sealing material, the better. 099134587 Form No. A0101 Page 15 of 44 0992060394-0 201216446 In order to reduce the mechanical stress during use, the module can be designed to be thinner, and the power module does not need to be preset. To further improve space utilization. If you need to install additional heat sinks, you can choose a screwless solution, such as direct bonding. In this way, the power module 10 of the embodiment will greatly increase the amount of the package of the type, and is also in line with the current and future power converter requirements, and can improve the space utilization of the power converter, thereby increasing the power density of the power source. Or efficiency. In addition, referring to FIG. 8, the second power device 14 is disposed on the first heat dissipation unit 11 by a layer of thermally conductive insulating material 13 instead of directly disposed on the heat dissipation unit 11. The thermally conductive insulating material layer 13 may have an insulating layer 132, such as insulated with a ceramic sheet. The thermally conductive insulating material layer 13 is, for example, a metal substrate or a metallized ceramic substrate, such as a direct bonded copper (DBC), a metallized ceramic sheet, a thick copper circuit layer, and a direct bonded aluminum. DBA), Ming substrate, copper substrate, or other forms of high thermal conductivity substrate. The thermally conductive insulating material layer 13 is exemplified by the DBC substrate. The thermally conductive insulating material layer 13 can include a heat conducting layer 131, an insulating layer 132, and a wiring layer 133. The heat conducting layer 131 and the wiring layer 133 can be copper and the insulating layer 132. Can be ceramic. Taking a conventional DBC board as an example, compared with the prior art, since only a part of components (second power device 14) of the present invention are mounted on the heat conductive insulating material layer 13, the number of components mounted thereon is reduced, DBC The board area can also be reduced accordingly, which can reduce the material cost of the package and improve the economic performance of the package. Moreover, due to the reduction of the DBC area, the deflection due to the inconsistent coefficient of thermal expansion (CTE) between the DBC and the heat dissipating unit 11 and the sealing material is 099134587 Form No. A0101 Page 16 / Total 44 pages 0992060394-0 201216446 (warpage) phenomenon will also be alleviated. ^疋Because the deflection due to the adaptation between CTEs of different materials is usually exacerbated by the increase in the size of the -L γ. In this way, the stress in the package can be reduced, thereby improving the reliability of the package. Therefore, due to the partial device, the gate (the power device 12) has been directly connected to the heat dissipating unit 11, and in the prior art, the power stack of the present invention requires a significant reduction in the material of the insulation, which reduces the cost. It also enhances the thermal management capability, and is more beneficial and 'published in the reliability design difficulty caused by reducing the CTE mismatch of each material.
在實際應用中，有-些對散熱要求非常苛刻的場合還 可以選用導熱係數更高的(不低於lw/m K，尤以大於 1.2W/m.K乃至大於uwa.K為佳）封料16，如此，可 以增加封料一侧的散熱能力’從而實現更優良的雙面散 熱’進一步提升整個封裝體的散熱能力。 圖10為该封裝類型的另一種擴展應用，可以將散熱單元 表面進行絕緣處理，使第一散熱單元^完全由封料16包 覆，使其任一表面不外露、或是藉由一絶緣體使散熱單 元11與外界隔離’以便使用在希望絕緣的場合。In practical applications, there are some places where the heat dissipation requirements are very demanding, and it is also possible to use a higher thermal conductivity (not less than lw/m K, especially greater than 1.2 W/mK or even greater than uwa.K). In this way, the heat dissipation capability on the side of the sealing material can be increased to achieve better double-sided heat dissipation, which further improves the heat dissipation capability of the entire package. FIG. 10 is another extended application of the package type, in which the surface of the heat dissipating unit can be insulated, so that the first heat dissipating unit is completely covered by the sealing material 16 so that neither surface is exposed or an insulator is used. The heat dissipating unit 11 is isolated from the outside for use in applications where insulation is desired.
為使功率模組之封裝類型可以擴展鉤更多場合，其可以 被設計成雙排Pin。如圖11所示。當内部電路過於複雜， 以至於需要更多引腳，可以在前面提及的特徵上，再加 一排引腳Ρ2。若此類封裝類型被應用在單排引腳pi就足 夠的場合’則圖中之上排引腳Ρ2可以被設計成散熱用途 眾所周知，電源内部，電壓跳變點越多，造成的電磁輻 射往往就越強，從而給電源電磁相容帶來難度。本發明 之散熱單元11 ’由於具備電特性，而其面積又相對較大 表單編號Α0101 第17頁/共44頁 0992060394-0 201216446 ，所以對電磁輻射帶來隱患。但如果優化設計該散熱單 元11的電特性，反而有機會將其設計成電磁輻射的遮罩 層，更有利於電磁相容。例如，可以將散熱單元Π連接 到電壓靜地點，即：相對來講，該電位相對與大地，比 較安靜，少噪音。如圖9中的Vin和GND，相對與其他電壓 點，就是比較平靜的。將散熱單元11設計成Vin或者GND ，更有利於電磁相容的。但實際操作中，為了便於實現 ，需要功率器件與散熱單元11連接的那個面只有一個電 極，在本實施例中為第一功率器件12。比如M0SFET，其 漏極（Drain)與源極（Source)間承受的電壓往往高 於門極（Gate)與源極（Source)間的電壓，所以，其 器件的源極和門極往往共用一面，而漏極往往獨佔一面 。這樣一來，將漏極作為靜地點的功率器件（第一功率 器件12)，直接與散熱單元11連接，既可以更好地進行 電磁相容，又方便制程。 如圖12，可以將背面的散熱單元]1拓寬/長，甚至折彎， 使其部分超過封料16包覆的部分，以擴大表面積。超出 封料16包覆的散熱單元11的兩面均可以實現和環境的熱 交換，因此，可以進一步加強功率模組1 0的散熱性能。 如圖13，在某些場合下，封裝體内部不僅僅需要搭載一 些功率半導體器件，還需要集成一些控制功能。而控制 線路通常比較複雜，因此需要使用佈線密度更高的基板 ，如PCB板或者1C。在此態樣中，可以將搭載控制線路的 控制器件18，例如高密度佈線板或者控制1C也封裝至封 裝體内。 如圖14，控制器件18可以是導熱係數較低，但是佈線密 099134587 表單編號A0101 第18頁/共44頁 0992060394-0 201216446 度較高的高密度基板。以便可以集成更多的控制功能。 控制器件18通常耐溫等級較功率器件的耐溫等級相比較 低，因此，在控制器件18和散熱單元11之間放置一個絕 熱層（熱導率通常低於〇.5W/m.K) IL。如此，可以降低 控制器件18，以及其上所搭載器件的溫度。 如圖15，上面所述散熱單元11，不限於一整塊，其上可 以根據需要做進一步的分割，以形成一些電路圖形，即 散熱單元11也可以具有多個電極。如此可以進一步增加 功率模組設計的靈活性。 Ο 功率模組ίο由於將多個器件集成在一起，相比分立器件 ，其電流流通回路被大大減少，從而降低了回路電感， 即減少了損耗，又降低了電壓噪音。但仍可以繼續被優 化。如圖16，以所提及全橋電路為例，增加集成一高頻 電容器C至功率模組10内部，以進一步減少回路，降低回 路電感量。 通常電源變換器為了安全可靠，會即時監測功率半導體 的溫度狀態，若溫度過高或者升溫過快，則說明電路有 ® 危險，可以提前採取預防動作，如關閉電源等。分立器 件的溫度檢測，只能在其外部增加溫度感測器，所以， 無法及時反映内部溫度狀態，且溫度感測器的安裝也較 複雜。所以，功率模組中，還可以集成溫度感測器，既 提升了溫度監測效果，又簡化了使用。 如圖17所示，此態樣之功率模組更包含一第二散熱單元 (heat sink) 11a，其設置於第二功率器件14與導熱絕 緣材料層13之間。由於功率器件在工作過程中，例如會 經歷超過正常工作電流數倍以上的暫態衝擊，故，藉由 099134587 表單編號A0101 第19頁/共44頁 0992060394-0 201216446 散熱單元11 a，可以在不增加導熱絕緣材料層13 ( dbc ) 面積的情況下，改善搭載至DBC板上需要承受熱衝擊的元 件的抗熱衝擊能力。另外，引線框架15係延伸與導熱絕 緣材料層13之線路層133連接。 如圖18所示，為了進一步改善導熱絕緣材料層13 (以DBC 板為例）上發熱量較大的元件（例如第二功率器件14 ) 的抗熱衝擊的性能，以及進一步改善DBC上線路的承栽電 流的能力’降低電流傳導阻抗，更可以將引線框架15的 面積增加’通過一導電材料鍵合至DBC的線路層上。利用 此結構開發的一款功率模組的實物照片見圖19 (未經封 料包覆）。其中DBC基板通過:釺焊的方式焊接至散熱單元 11上’而引線框架15同樣通過釺焊的方式和|)阢基板的線 路層實現電氣與機械連接。圖19所示的功,率模組1 〇所使 用的DBC基板，其線路層厚度為0. 3mm，而引線框架15的 厚度為0. 5mm，因此，採用此結構的傳導電阻和直接將晶 片鍵合在DBC線路層上相比降低6〇%点 上如此可以有效 降低模組產熱量，從而提高模組的電性能，改善模組的 散熱性能。 如圖20所不’在功率模組1〇内除了使用導熱能力較好的 DBC基板以外’也可以使用類似銅基板13a等導熱能力較 好的基板。一般銅基板的結構為，在一較厚的銅襯底上 ’生成絕緣層和薄銅線路層。而且絕緣層和薄銅線路層 的層數不以一層為限，可以是多層。在某些場合下可以 實現更南的佈線密度。 一般而言’第~功率器件與第二功率器件皆由導線（ wire bonding)來傳輸訊號，由於導線往往是用鋁導線 099134587 表單編號A0101 第20頁/共44頁 0992060394-0 201216446 (A1 wire)來完成，内阻很大。用金導線（Au wi re ) ，則成本太高。雖然最近工藝有銅導線（Cu wi re )出現 ，但仍舊内阻很大。如圖21所示，為進一步降低封裝内 阻造成的損耗，本發明可以用wireless bond工藝，如 銅片取代wire bond來實現電流傳遞，大大降低了封裝 内阻，且成本也不會太高。本態樣係藉由引線框架15延 伸連結於第一功率器件12及第二功率器件14之至少一而 取代導線。 圖22所示為一進一步改善熱傳遞能力的方案。由於本發 〇 明提及之功率模組，往往是有些器件（例如第一功率器 件12)直接與散熱單元11相連，而有些器件（例如第二 功率器件14)與散熱單元11之間則有絕緣元件（例如具 有絕緣層之導熱絕緣材料層13)，從而導致整個模組中 封料16的厚度不均，也就是說，局部封料16與器件的距 離會比較厚，使封料16的溫度不均勻，從而影響了封料 16表面的散熱能力。圖22中，在封料16較厚的地方增加 熱良導體之一第三散熱單元lib，其係設置於第一散熱單 〇 元11的第一區，從均勻化封料16至器件的厚度，而改善 散熱能力。 另外，如圖25所示，第三散熱單元lib係穿出封料16，並 具有一彎折。第三散熱單元lib穿出封料16而可作為引腳 Pin、或是單純散熱、或是部分作為引腳部分用來散熱。 第三散熱單元lib藉由彎折可減少功率模組10直立時的尺 寸。 實際應用中，若需進一步擴大散熱能力，可以通過圖26 的方式達成。即在功率模組10的第三散熱單元lib上再安 099134587 表單編號A0101 第21頁/共44頁 0992060394-0 201216446 裝一第四散熱單元lie。第四散熱單元lie可藉由焊接、 粘結等方式與第三散熱單元lib連結。由於安裝簡單第四 散熱單元lie的形狀和位置可以不受限定。但實際效果上 ，以保留功率模組10自有表面散熱能力為佳。即，如圖 26，在第四散熱單元lie與功率模組10前表面A1之間保 留一空隙，使得風流可以該空隙中流動，從而使功率模 組前表面和第四散熱單元lie下表面（靠近前表面A1之表 面）均能發揮一定散熱功能。為使該空隙中的風流能夠 達到相當的程度，該空隙厚度可大於1mm，尤以大於2mm 為佳。 為了更好地解釋本發明的意義，進一步借助全橋電路來 進行說明，如前所述，圖9為全橋電路的拓撲圖，圖23和 圖24A至24D分別為其功率模組内部結構和三維示意圖。 其中，圖24A為功率模組10的正面示意圖，圖24B為功率 模組10的背面示意圖，圖24C為功率模組10脫去封料16 的正面示意圖，圖24D為功率模組10脫去封料16的背面示 意圖。 雖然上述實施例係以一第一功率器件12及一第二功率器 件14為例作說明，但並非具限制性，且其中第一功率器 件12所代表的意義為其設置於散熱單元11上，而第二功 率器件14所代表的意義為其藉由一導熱絕緣材料層1 3設 置於散熱單元11上。以下係以二個第一功率器件S1及S2 以及二個第二功率器件S3及S4作說明。 如圖9所示，全橋電路包括4個開關器件S1〜S4，這裏以 MOSFET為例。這四個開關器件組成兩組導電橋臂：S1和 S4組成一組，S2和S3組成一組橋臂；橋臂上管開關器件 099134587 表單編號A0101 第22頁/共44頁 0992060394-0 201216446 S1和S2的Drain端共同連接在電壓高電位點yin (在D2D 應用時’電氣端V i η為直流輸入端，是電愿波形為一個穩 定的直流或者帶有很小紋波的直流），橋臂下管開關器 件S3和S4的Source端共同連接在電壓的低電位點GND ; 而單一橋臂上管的Source和下管的Drain相連接，如S1 和S4橋臂連接於VA ’ S2和S3的橋臂連接於VB，其工作的 基本原理是橋臂的上下管互補導通，如S1開通，S4關斷 ;S1關斷’ S4開通’在開關狀態轉換過程存在短暫時間 都關斷的過程。這樣’ D2D的應用場合下，輸入端Vin- GND之間為直流’而橋▼中韧連接點\ΓΑ，VB的電壓則是開 ： " . ......" . . 關次的跳變，幅值為〇與Vin。 目前大功率M0SFET最典型的’極引出方式為，晶片的背 面為“Drain” ，正面分佈兩個電極，“s〇urce”和“In order to make the package type of the power module expandable, it can be designed as a double-row Pin. As shown in Figure 11. When the internal circuitry is too complex to require more pins, a row of pins Ρ2 can be added to the previously mentioned features. If such a package type is applied in a single row of pins pi is sufficient, then the upper row of pins Ρ2 in the figure can be designed for heat dissipation. It is well known that the internal voltage, the more voltage trip points, the electromagnetic radiation is often caused. The stronger, the difficulty in electromagnetic compatibility of the power supply. The heat dissipating unit 11' of the present invention has a relatively large area due to its electrical characteristics. Form No. Α0101 Page 17 of 44 0992060394-0 201216446, so there is a hidden danger to electromagnetic radiation. However, if the electrical characteristics of the heat dissipating unit 11 are optimized, the organic layer will be designed as a mask layer for electromagnetic radiation, which is more advantageous for electromagnetic compatibility. For example, the heat sink unit Π can be connected to a voltage static location, that is, relatively speaking, the potential is relatively quiet and less noisy relative to the earth. As shown in Figure 9, Vin and GND are relatively quiet compared to other voltage points. Designing the heat sink unit 11 to be Vin or GND is more advantageous for electromagnetic compatibility. However, in actual operation, in order to facilitate the implementation, the surface on which the power device is connected to the heat dissipating unit 11 is required to have only one electrode, which is the first power device 12 in this embodiment. For example, the M0SFET has a voltage between the drain and the source that is often higher than the voltage between the gate and the source. Therefore, the source and the gate of the device often share one side. And the drain tends to be exclusive. In this way, the power device (the first power device 12) having the drain as a static place is directly connected to the heat dissipating unit 11, which can better perform electromagnetic compatibility and facilitate the process. As shown in Fig. 12, the heat dissipating unit 1 on the back surface can be widened/long, or even bent, so that it partially exceeds the portion covered by the sealing material 16 to enlarge the surface area. The heat exchange with the environment can be achieved on both sides of the heat dissipating unit 11 covered by the sealing material 16, so that the heat dissipation performance of the power module 10 can be further enhanced. As shown in Fig. 13, in some cases, not only some power semiconductor devices but also some control functions need to be integrated inside the package. Control lines are often complex, so you need to use a substrate with a higher wiring density, such as a PCB or 1C. In this aspect, the control device 18 on which the control line is mounted, such as the high-density wiring board or the control 1C, can also be packaged into the package. As shown in Fig. 14, the control device 18 may have a lower thermal conductivity, but the wiring is dense. 099134587 Form No. A0101 Page 18/44 page 0992060394-0 201216446 Higher high density substrate. So that you can integrate more control functions. The control device 18 typically has a lower temperature rating than the temperature rating of the power device, and therefore, a thermal insulation layer (thermal conductivity typically below 〇.5 W/m.K) IL is placed between the control device 18 and the heat sink unit 11. Thus, the temperature of the control device 18 and the device mounted thereon can be lowered. As shown in Fig. 15, the heat dissipating unit 11 is not limited to a single block, and may be further divided as needed to form some circuit patterns, that is, the heat dissipating unit 11 may have a plurality of electrodes. This can further increase the flexibility of power module design.功率 Power Module ίο Because of the integration of multiple devices, the current circulation loop is greatly reduced compared to discrete devices, which reduces the loop inductance, which reduces losses and reduces voltage noise. But still can continue to be optimized. As shown in Fig. 16, taking the full bridge circuit mentioned as an example, the integration of a high frequency capacitor C into the interior of the power module 10 is added to further reduce the loop and reduce the loop inductance. In general, the power converter monitors the temperature state of the power semiconductor in time for safety and reliability. If the temperature is too high or the temperature rises too fast, the circuit has a ® danger, and precautionary actions such as turning off the power can be taken in advance. The temperature detection of discrete components can only increase the temperature sensor outside of it, so the internal temperature state cannot be reflected in time, and the temperature sensor installation is also complicated. Therefore, in the power module, the temperature sensor can also be integrated, which not only improves the temperature monitoring effect, but also simplifies the use. As shown in FIG. 17, the power module of this aspect further includes a second heat sink 11a disposed between the second power device 14 and the layer of thermally conductive insulating material 13. Since the power device is in operation, for example, it will experience transient shocks that are more than several times higher than the normal operating current, therefore, by 099134587 Form No. A0101 Page 19/44 Page 0992060394-0 201216446 Heat Dissipating Unit 11 a, can be In the case where the area of the thermally conductive insulating material layer 13 (dbc) is increased, the thermal shock resistance of the component to be subjected to thermal shock to the DBC board is improved. Further, the lead frame 15 is extended to be connected to the wiring layer 133 of the thermally conductive insulating material layer 13. As shown in FIG. 18, in order to further improve the thermal shock resistance of the element having a large heat-generating material layer 13 (for example, the DBC board), and further improving the line on the DBC. The ability to carry current 'reduced current conduction impedance, and the area of lead frame 15 can be increased' is bonded to the circuit layer of DBC through a conductive material. A physical photograph of a power module developed using this structure is shown in Figure 19 (unsealed). The DBC substrate is soldered to the heat dissipating unit 11 by soldering, and the lead frame 15 is also electrically and mechanically connected by means of soldering and the circuit layer of the substrate. The thickness of the lead frame 15 is 0. 5mm, and therefore, the conductive resistance of the structure and the direct transfer of the wafer are as shown in FIG. Bonding on the DBC line layer is reduced by 6〇%, which can effectively reduce the heat generation of the module, thereby improving the electrical performance of the module and improving the heat dissipation performance of the module. As shown in Fig. 20, in addition to the use of a DBC substrate having a good thermal conductivity in the power module 1A, a substrate having a relatively good thermal conductivity such as a copper substrate 13a can be used. Generally, the structure of the copper substrate is such that an insulating layer and a thin copper wiring layer are formed on a thick copper substrate. Further, the number of layers of the insulating layer and the thin copper wiring layer is not limited to one layer, and may be a plurality of layers. In some cases, a more souther wiring density can be achieved. Generally, the 'th power device and the second power device are both transmitted by wire bonding, since the wire is often made of aluminum wire 099134587 Form No. A0101 Page 20 / Total 44 Page 0992060394-0 201216446 (A1 wire) To complete, the internal resistance is very large. With gold wire (Au wi re ), the cost is too high. Although the recent process has a copper wire (Cu wi re ), it still has a large internal resistance. As shown in Fig. 21, in order to further reduce the loss caused by the internal resistance of the package, the present invention can use a wireless bond process, such as a copper chip instead of a wire bond to achieve current transfer, which greatly reduces the internal resistance of the package, and the cost is not too high. This aspect replaces the wires by extending the lead frame 15 to at least one of the first power device 12 and the second power device 14. Figure 22 shows a solution for further improving heat transfer capability. Due to the power module mentioned in the present invention, it is often that some devices (for example, the first power device 12) are directly connected to the heat dissipation unit 11, and some devices (for example, the second power device 14) and the heat dissipation unit 11 are An insulating component (for example, a layer of thermally conductive insulating material 13 having an insulating layer), resulting in uneven thickness of the sealing material 16 in the entire module, that is, the distance between the partial sealing material 16 and the device is relatively thick, so that the sealing material 16 is The temperature is not uniform, which affects the heat dissipation capability of the surface of the sealing material 16. In FIG. 22, a third heat dissipating unit lib is added to the first portion of the first heat dissipating unit 11 from the uniformity of the sealing material 16 to the thickness of the device. And improve heat dissipation. Further, as shown in Fig. 25, the third heat radiating unit lib passes through the sealing material 16 and has a bend. The third heat dissipating unit lib passes through the sealing material 16 and can be used as a pin pin, or simply dissipates heat, or partially serves as a pin portion for heat dissipation. The third heat dissipating unit lib can reduce the size of the power module 10 when it is erected by bending. In practical applications, if you need to further expand the heat dissipation capacity, you can do so by means of Figure 26. That is, the third heat dissipating unit lib of the power module 10 is re-installed. 099134587 Form No. A0101 Page 21/44 page 0992060394-0 201216446 A fourth heat dissipating unit lie is installed. The fourth heat dissipation unit lie can be coupled to the third heat dissipation unit lib by soldering, bonding, or the like. Due to the simple installation, the shape and position of the fourth heat dissipating unit lie can be unrestricted. However, in actual effect, it is preferable to retain the surface heat dissipation capability of the power module 10. That is, as shown in FIG. 26, a gap is left between the fourth heat dissipation unit lie and the front surface A1 of the power module 10, so that the wind flow can flow in the gap, so that the front surface of the power module and the lower surface of the fourth heat dissipation unit lie ( Close to the surface of the front surface A1) can play a certain heat dissipation function. In order to achieve a considerable degree of wind flow in the void, the void thickness may be greater than 1 mm, especially greater than 2 mm. In order to better explain the meaning of the present invention, further description is made by means of a full bridge circuit. As described above, FIG. 9 is a topology diagram of a full bridge circuit, and FIG. 23 and FIGS. 24A to 24D are respectively a power module internal structure and Three-dimensional schematic. 24A is a front view of the power module 10, FIG. 24B is a rear view of the power module 10, FIG. 24C is a front view of the power module 10 with the sealing material 16 removed, and FIG. 24D is a power module 10 A schematic view of the back side of the material 16. Although the foregoing embodiment is described by taking a first power device 12 and a second power device 14 as an example, it is not limited, and wherein the first power device 12 represents the meaning of being disposed on the heat dissipation unit 11 . The second power device 14 is represented by a thermal conductive material layer 13 disposed on the heat dissipation unit 11. The following description will be made with two first power devices S1 and S2 and two second power devices S3 and S4. As shown in FIG. 9, the full bridge circuit includes four switching devices S1 to S4, and a MOSFET is taken as an example here. The four switching devices form two sets of conductive bridge arms: S1 and S4 form a group, S2 and S3 form a group of bridge arms; Bridge arm upper tube switch devices 099134587 Form No. A0101 Page 22 / Total 44 Page 0992060394-0 201216446 S1 Connected to the Drain terminal of S2 at the high voltage point yin (in the D2D application, the 'electrical terminal V i η is the DC input terminal, the power waveform is a stable DC or DC with little ripple), the bridge arm The source terminals of the lower tube switching devices S3 and S4 are connected in common to the low potential point GND of the voltage; the source of the upper arm of the single bridge is connected to the Drain of the lower tube, such as the S1 and S4 bridge arms connected to the VA 'S2 and S3 The bridge arm is connected to VB, and the basic principle of its operation is that the upper and lower tubes of the bridge arm are complementarily turned on, such as S1 is turned on, and S4 is turned off; S1 is turned off, 'S4 is turned on', and the process of switching state transition is short-lived. In this case, in the application of 'D2D, the input terminal Vin- GND is DC' and the bridge is connected to the toughness point ΓΑ, and the voltage of VB is ON: " . . . . The jump, the amplitude is 〇 and Vin. At present, the most typical 'maximum extraction method for high-power MOSFETs is that the back side of the wafer is “Drain”, the front side is distributed with two electrodes, “s〇urce” and “
Gate” ’其中“Gate”的尺寸較小，例如lmm*lmni。晶 片背面的“Drain”通常預先進行可釺焊處理，而正面的 “Source”和“Gate”往往為:銘金屬化電極，可以通過 鋁/金wire bonding的方式實現和週邊電路的連接。由 ) 於開關器件S1和S2的“Drain”連接於共同的直流電位 點Vin，因此，可以將其直接釺焊至散熱單元丨丨上，而Gate" 'where "Gate" is small in size, for example, lmm*lmni. The "Drain" on the back of the wafer is usually pre-welded, while the "Source" and "Gate" on the front are often: metallized electrodes. The connection to the peripheral circuit is realized by aluminum/gold wire bonding. The "Drain" of the switching devices S1 and S2 is connected to the common DC potential point Vin, so that it can be directly soldered to the heat sink unit ,and
Vi η和外界電連接的Pin也可以直接釺焊至散熱單元^上 ’從而利用該導電極佳的散熱單元11導電，降低電損耗 ’減少封裝體的熱量產生。如此，可以獲得最佳的熱， 電性能。而現有的功率模組，如前述習知的做法為，將 所有四顆M0SFET安裝至DBC板上，隨後，所有M0SFET和 弓丨線框架的電連接均靠wire bonding的方式來實現。如 上文所討論的一樣，現有技術的種種缺陷（散熱差，電 099134587 * „ …、 农單編號A010I 第23頁/共44頁 0992060394-0 201216446 性能差，價格高，可靠性差等等）相比之下一目了然。 本發明在此處的應用更具降低EMI的效果，前面對全橋電 路的基本工作原理分析看。散熱單元1 1連接於直流輸入 端Vin，為很好的靜態電位點，而橋臂中間連接點VA，VB 則為電壓跳變點，大片的散熱單元11可以有效阻斷跳變 信號的傳遞。如此，可以有效減小跳變點對週邊電路的 幹擾，減小測試的Ε ΜI。 如前所述，為了具備更好的EMC特性和散熱性能，將全橋 模組中的開關器件S3、S4置於絕緣層（即導熱絕緣材料 層具有之絕緣層）之上，將開關器件SI、S2直接置於散 熱單元11之上；為了方便生產，並減少生產工差造成的 空間浪費，開關器件S3、S4置於相連的絕緣層之上；為 減少回路電感和方便使用，將S2置於S3外側，將S1置於 S4外側。也就是說，對於圖9所示的全橋電路來講，模組 内部器件按S2-S3-S4-S1或者S1-S4-S3-S2的順序排布 ，性能更為優秀。 以下說明本實施例之功率模組的製造流程，於此，導熱 絕緣材料層係以覆銅陶瓷基板為例，另外，此功率模組 除了搭載功率器件（半導體晶片）外還集成了一些被動 元件，如電阻和電容，且在引線框架其中一些引腳上還 搭載了一個溫度測量電阻，以用作模組過溫保護之用。 具體的製作流程如下··先在散熱單元11上組裝導熱絕緣 材料層1 3的位置以及和引線框架15連接的位置塗上錫膏 ，同樣將導熱絕緣材料層13上需要和引線框架1 5組裝的 位置塗上錫膏，隨後將散熱單元11、導熱絕緣材料層13 和引線框架15按照設定的裝配關係置於一治具中（As- 099134587 表單編號A0101 第24頁/共44頁 0992060394-0 201216446 sembly) ·，然後過回流焊爐（Reflow)使其焊接在一起 ，由此這三個部件形成一個整體’在隨後植晶製程中可 以利用引線框架15進行傳輸與定位；清洗（Flux Cleaning)後，進行植晶安裝所需的半導體器件（如 MOS及Diode )，此處需要著重強調的是部分功率器件放 置在散熱單元11上（如第一功率器件12)，另外一部分The Pin connected to Vi η and the outside can also be directly soldered to the heat dissipating unit to conduct electricity through the heat dissipating unit 11 of the conductive electrode, thereby reducing the electric loss and reducing the heat generation of the package. In this way, the best thermal and electrical properties can be obtained. In the conventional power module, as described above, all four MOSFETs are mounted on the DBC board, and then the electrical connections of all the MOSFETs and the bow frame are implemented by wire bonding. As discussed above, the various drawbacks of the prior art (difference in heat dissipation, electricity 099134587 * „, country order number A010I page 23/total 44 page 0992060394-0 201216446 poor performance, high price, poor reliability, etc.) The application of the present invention is more effective in reducing EMI. The basic working principle of the full-bridge circuit is analyzed. The heat-dissipating unit 1 1 is connected to the DC input terminal Vin, which is a good static potential point. The intermediate connection point VA and VB of the bridge arm are voltage jump points, and the large heat dissipation unit 11 can effectively block the transmission of the hopping signal. Thus, the interference of the trip point to the peripheral circuit can be effectively reduced, and the test is reduced. Ε ΜI. As mentioned above, in order to have better EMC characteristics and heat dissipation performance, the switching devices S3 and S4 in the full-bridge module are placed on the insulating layer (ie, the insulating layer of the thermally conductive insulating material layer), The switching devices SI and S2 are directly placed on the heat dissipating unit 11; in order to facilitate production and reduce space waste caused by production work, the switching devices S3 and S4 are placed on the connected insulating layer; For the sense of convenience and convenience, place S2 outside S3 and S1 outside S4. That is, for the full-bridge circuit shown in Figure 9, the internal components of the module are S2-S3-S4-S1 or S1- The sequential arrangement of S4-S3-S2 is more excellent. The following describes the manufacturing process of the power module of the present embodiment. Here, the thermal conductive insulating material layer is exemplified by a copper-clad ceramic substrate, and the power module is additionally used. In addition to carrying power devices (semiconductor wafers), some passive components, such as resistors and capacitors, are integrated, and a temperature measuring resistor is also mounted on some of the leads of the lead frame for use as a module for overtemperature protection. The manufacturing process is as follows: First, the position of the heat conductive insulating material layer 13 is assembled on the heat radiating unit 11 and the solder paste is applied to the position where the lead frame 15 is connected, and the heat conductive insulating material layer 13 is also required to be assembled with the lead frame 15. The position is coated with solder paste, and then the heat dissipating unit 11, the thermally conductive insulating material layer 13 and the lead frame 15 are placed in a fixture according to the set assembly relationship (As- 099134587 Form No. A0101 Page 24 / Total 44 Page 0992060394-0 201216 446 sembly) ·, then reflow soldering (Reflow) to make them welded together, so that the three parts form a whole 'transport and positioning can be carried out by the lead frame 15 in the subsequent crystallization process; cleaning (Flux Cleaning) After that, the semiconductor devices (such as MOS and Diode) required for the crystal mounting are performed. Here, it is emphasized that part of the power devices are placed on the heat dissipation unit 11 (such as the first power device 12), and another part.
功率晶片放置在導熱絕緣材料層13上（如第二功率器件 14)，植晶時所用的連接介面材料也是錫膏；在使用單 一功能的植晶機時，由於其不具備抓取表面黏著（SMT) 器件的能力，因此，一些電阻、電容等器件還需要進行 SMT的操作，即：點錫膏（Solder Dispense)後，放 置其他元器件（SMT)由於所用的功率器件的晶片尺寸 較大’採用錫膏（solder paste)進行reflow時有焊 接層的氣孔率較高而帶來工藝性、可靠性不佳的疑慮， 此處採用真空回焊（Vacuum Reflow)使元件和散熱單 元11、導熱絕緣材料層13、引線框架15、晶片、SMT器 件焊接在一起清疼乂Flux.O.le.aning)後，進行打線 接合（Wire bond)作業；包封（Molding或者其他金 屬/陶瓷封裝形式）後即完成主要流程。 於一些在植晶製程時無需使用引線框架15進行定位的 應用下’有機會進一步簡化工藝流程。首先將散熱單元 11、導熱絕緣材料層13、引線框架15上需要的位置施加 錫膏；隨後將所需的元件（功率晶片以及被動SMT元件） 刀別放置於需要的位置上，這步驟可以通過泛用較強的 機台（如集成植晶和表面黏著技術功能的機台）上而一 站式實現，也可以在多個機臺上實現，使用的連接介面 099134587 表單編號A0101 第25頁/共44頁 0992060394-0 201216446 材料可以為錫膏；隨後將放置有元件的散熱單元11、導 熱絕緣材料層13、引線框架15按照設定的裝配關係放置 於一治具中，完成assembly ;隨後真空reflow ;後續的 工藝和上述的工藝流程相同。如此，可以減少re f 1 ow的 次數以及相應的清洗等流程，由於r e f 1 ow次數的減少對 於提高模組的可靠性也有一定的好處。 當然也可以使用焊料片，導電膠，低溫燒結納米銀漿， 等材料代替錫膏作為模組組裝所需的電性/機械連接的介 面材料。在某些場合下甚至可以在同一模組組裝過程中 不同的裝配位置使用超過一種連接介面材料。使用這些 連接材料時，組裝工藝以及流程也需要做對應的調整。 晶片正面電極引出時也可以使用除打線外的其他方式， 例如使用粘結/焊接一金屬（銅）片，實現正面電極和外 部線路的互聯。 綜上，通過本發明所揭露的，用以提升電源變換器功率 密度或者效率的封裝方法和結構，可以獲得與現有技術 相比，更佳的熱性能，電性能，經濟性能，EMC性能與更 高的可靠性。其内部空間利用率很高，使用方便，非常 有利於提高變換器功率密度或者效率。而本發明給出的 具體功率模組具體實施，也非常可行有效。本發明非常 適合用以提升電源變換器的整體性能和性價比。 以上所述僅為舉例性，而非為限制性者。任何未脫離本 發明之精神與範疇，而對其進行之等效修改或變更，均 應包含於後附之申請專利範圍中。 【圖式簡單說明】 [0005] 圖1至圖5為習知之功率模組之不同態樣的示意圖； 099134587 表單編號A0101 第26頁/共44頁 0992060394-0 201216446 圖6顯示一種I GBT三相橋模組； 圖7顯示一種雙相整流橋； 圖8為本發明較佳實施例之一種功率模組的示意圖； 圖9及圖1 6顯示本發明較佳實施例之一種功率模組應用之 全橋電路的不同態樣；以及 圖10至圖15以及圖17至圖26為本發明較佳實施例之功率 模組不同態樣的示意圖。 【主要元件符號說明】 [0006] ❹ 〇 099134587 10、 30 :功率模組 11、 11a、lib、11c、31、31b :散熱單元 111 :第一區 112 :第二區 12 :第一功率器件 13 :導熱絕緣材料層 131 :導熱層 132 :絕緣層 133 :線路層 13a :銅基板 14 :第二功率器件 15、 35 :引線框架 16、 36 :封料 17 :鍵接材料層 18、38 :控制器件 31a :覆銅陶瓷基板 32、34 :元器件 A1 :前表面 表單編號A0101 第27頁/共44頁 0992060394-0 201216446 A2 :後表面 C :電容器 D :厚度 IL :絕熱層 P2、P1 ··引腳 S1〜S4 :開關器件 099134587 表單編號A0101 第28頁/共44頁 0992060394-0The power chip is placed on the layer of thermally conductive insulating material 13 (such as the second power device 14), and the connection interface material used for the seeding is also a solder paste; when using a single function crystallizer, since it does not have a gripping surface adhesion ( SMT) The ability of the device, therefore, some resistors, capacitors and other devices also need to perform SMT operation, that is, after the solder paste (Solder Dispense), the other components (SMT) are placed because of the larger size of the power device used. When reflow is performed with a solder paste, there is a problem that the porosity of the solder layer is high, resulting in poor processability and reliability. Here, Vacuum Reflow is used to insulate the component and the heat dissipation unit 11, and heat conduction. After the material layer 13, the lead frame 15, the wafer, and the SMT device are soldered together, Flux.O.le.aning), wire bonding work; after encapsulation (Molding or other metal/ceramic package form) That is to complete the main process. For applications that do not require the use of leadframe 15 for positioning during the lithography process, there is an opportunity to further streamline the process. First, the solder paste is applied to the required positions on the heat dissipating unit 11, the thermally conductive insulating material layer 13, and the lead frame 15; then the required components (power chips and passive SMT components) are placed in the desired positions, and the steps can be passed. It can be implemented in one-stop operation by using a strong machine (such as a machine with integrated crystallization and surface adhesion technology). It can also be implemented on multiple machines. The connection interface used is 099134587 Form No. A0101 Page 25 / A total of 44 pages 0992060394-0 201216446 material can be solder paste; then placed the component of the heat dissipation unit 11, the thermal insulation material layer 13, the lead frame 15 in a set assembly relationship in a fixture, complete the assembly; then vacuum reflow The subsequent process is the same as the above process. In this way, the number of re f 1 ows and the corresponding cleaning process can be reduced, and the reduction in the number of r e f 1 ows also has certain advantages for improving the reliability of the module. It is of course also possible to use solder sheets, conductive pastes, low-temperature sintered nano-silver pastes, and the like instead of solder paste as the interface material for the electrical/mechanical connection required for module assembly. In some cases it is even possible to use more than one connection interface material at different assembly locations during the same module assembly process. When using these connecting materials, the assembly process and the process also need to be adjusted accordingly. Other methods other than wire bonding can also be used when the front electrode of the wafer is taken out, for example, by bonding/welding a metal (copper) piece to interconnect the front electrode and the external line. In summary, the packaging method and structure for improving the power density or efficiency of the power converter disclosed by the present invention can obtain better thermal performance, electrical performance, economic performance, EMC performance and more than the prior art. High reliability. Its internal space utilization is very high and easy to use, which is very beneficial to improve converter power density or efficiency. The specific implementation of the specific power module given by the present invention is also very feasible and effective. The invention is well suited for improving the overall performance and cost performance of a power converter. The above is intended to be illustrative only and not limiting. Any equivalent modifications or alterations to the spirit and scope of the invention are intended to be included in the scope of the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS [0005] FIGS. 1 to 5 are schematic diagrams showing different aspects of a conventional power module; 099134587 Form No. A0101 Page 26 of 44 0992060394-0 201216446 Figure 6 shows an I GBT three-phase Figure 7 shows a dual-phase rectifier bridge; Figure 8 is a schematic diagram of a power module according to a preferred embodiment of the present invention; Figure 9 and Figure 16 show a power module application according to a preferred embodiment of the present invention. Different aspects of the full bridge circuit; and FIGS. 10 to 15 and 17 to 26 are schematic views of different aspects of the power module according to a preferred embodiment of the present invention. [Main component symbol description] [0006] ❹ 〇 099134587 10, 30: power module 11, 11a, lib, 11c, 31, 31b: heat dissipation unit 111: first region 112: second region 12: first power device 13 : thermally conductive insulating material layer 131 : thermally conductive layer 132 : insulating layer 133 : wiring layer 13 a : copper substrate 14 : second power device 15 , 35 : lead frame 16 , 36 : sealing material 17 : bonding material layer 18 , 38 : control Device 31a: Copper-clad ceramic substrate 32, 34: Component A1: Front surface form No. A0101 Page 27/44 page 0992060394-0 201216446 A2: Rear surface C: Capacitor D: Thickness IL: Insulation layer P2, P1 ·· Pins S1~S4: Switching Device 099134587 Form No. A0101 Page 28/Total 44 Page 0992060394-0