201001728 六、發明說明： 【發明所屬_^技術名頁威】 本發明係有關光伏打裝置及其製造方法。更特別來 說，本發明有關薄膜多晶光伏打裝置及方法、薄膜光伏打 5 裝置中之透明傳導層，且有關此等層之製造方法、薄膜奈 米晶光伏打裝置之製造方法、化學氣相沉積反應器及使用 其來製造薄膜光伏打裝置之方法。 【^tr 冬好;3 長久以來不斷致力追求更有效率且具成本競爭力的方 10 式將光、特別是日光轉換成電力。然而，對於大部份用途， 利用光伏打胞元產生的電能成本仍高於相競爭的方法所產 生之成本。 %付的无伏打胞元 π工文叫7¾ 1示馮兵;f目到· 15 20 車乂低的效率’另一主要問題係為此等裝置的製造成本 【考务明内容】 為此，本發明之一目的係提供以相對較高的效率將光 轉換成電能之相對較低成本的光伏打裝置。 少其光伏打446逐年作4Μ，其有部分係使用稀 本發明：士的材料，且其價格已隨時間而顯著增高。因此， 含且低^目的係提供—利用對於制财取得之最為豐 I低成本的部分㈣之光伏 來材料短缺及價格增高所苦。 ⑽U不又未 下不括錢域天、或清晨紐晚料利條件 易取侍直接、明亮的日光。此等條件下’許多先前光 3 201001728 伏打裝置的功率輸出位準係遠遠落在所想要位準之下。 因此，本發明另一目的係提供即便在此等不利條件下 仍保有相對較高輸出之光伏打裝置。 部分先前光伏打裝置之另一問題係為其需要相對較大 5 量的材料’及/或所須使用的材料形式造價昂貴。 為此，本發明另一目的係提供一使用很少量關鍵材料 且利用具有相對較低造價之此等材料的形式之裝置。 已經不利地影響先前裝置的製造成本之另一問題係為 裝置之相對較低的製造速度。 10 因此，本發明另一目的係提供一可以相對較高速度製 造之光伏打裝置，及一用於達成該目的之製程。 雖然部分先前光伏打裝置使用體塊材料、諸如體塊狀 的矽，已取而代之發展出使用薄膜之其他裝置。一種此裝 置係為一使用非晶矽的薄膜光伏打胞元。雖然非晶矽是相 15 對較低成本的材料，其具有特定缺陷。其他薄膜裝置亦有 缺陷。本發明之一目的係提供一可盡量消除這些缺陷之光 伏打裝置。 譬如，非晶薄膜裝置嚴重受到缺乏穩定性所苦。特定 言之，此等裝置的能量轉換效率在其首次使用於強光中之 20 後不久即顯著地減低。效率持續減低，直到經過一段時間 之後，效率穩定在遠比其首次製造時所具有者更低之一位 準為止。 本發明之一目的係提供一可避免或盡量降低此劣化之 裝置。 4 201001728 ^光伏打裝置之另一問題係為其頻譜響 差=，回應於較長波長的曰光隔區之電流輸出量係遠 減^有較短波長的日光隔區，結果p、有實質地小於可取 仔太陽能的完整頻譜範圍被用來產生電力。 為此，本發明另一 波長之光具有顯著經改 較大的電力輸出。 5 目的係提供-對於太陽頻譜的較長 良的響應之光伏打裝置，因此提供201001728 VI. Description of the Invention: [Inventions _^Technical Name Page] The present invention relates to a photovoltaic device and a method of manufacturing the same. More particularly, the present invention relates to a thin film polycrystalline photovoltaic device and method, a transparent conductive layer in a thin film photovoltaic device, and a method for manufacturing the same, a method for manufacturing a thin film nanocrystalline photovoltaic device, and a chemical gas A phase deposition reactor and method of using the same to fabricate a thin film photovoltaic device. [^tr Winter is good; 3 For a long time, we have been striving to pursue more efficient and cost-competitive ways to convert light, especially daylight, into electricity. However, for most applications, the cost of electricity generated using photovoltaic cells is still higher than the cost of competing methods. %付的伏伏打元元工文叫73⁄4 1示冯兵; f目到· 15 20 The low efficiency of the car's other major problem is the manufacturing cost of such devices [test details] One object of the present invention is to provide a relatively low cost photovoltaic device that converts light into electrical energy with relatively high efficiency. Less of its photovoltaics hit 446 year after year, some of which use the rare invention: the material of the scholar, and its price has increased significantly over time. Therefore, the purpose of providing low-quality products is to make use of the photovoltaics that are the most cost-effective for the production of materials (4). (10) U does not include the money field, or the early morning and late night conditions. It is easy to take direct, bright daylight. Under these conditions, the power output level of many previous light 3 201001728 voltaic devices is far below the desired level. Accordingly, it is another object of the present invention to provide a photovoltaic device that retains a relatively high output even under such adverse conditions. Another problem with some prior photovoltaic devices is that they require a relatively large amount of material' and/or the form of material to be used is expensive. To this end, another object of the present invention is to provide a device that uses a small amount of critical material and utilizes the form of such materials having relatively low cost. Another problem that has adversely affected the manufacturing cost of previous devices is the relatively low manufacturing speed of the device. 10 Accordingly, it is another object of the present invention to provide a photovoltaic device that can be fabricated at relatively high speeds, and a process for accomplishing the same. While some prior photovoltaic devices have used bulk material, such as bulk rafts, other devices that use films have been developed. One such device is a thin film photovoltaic cell using amorphous germanium. Although amorphous germanium is a relatively low cost material, it has specific drawbacks. Other thin film devices are also defective. It is an object of the present invention to provide a voltaic device that minimizes these drawbacks. For example, amorphous thin film devices are severely suffering from lack of stability. In particular, the energy conversion efficiencies of such devices are significantly reduced shortly after their first use in glare 20 . Efficiency continues to decrease until after a period of time, efficiency is stable far below one of the first time it was manufactured. It is an object of the present invention to provide a device that avoids or minimizes this degradation. 4 201001728 ^ Another problem with photovoltaic devices is their spectral response =, the current output in response to longer wavelengths of the luminescence compartment is far less than the solar region with shorter wavelengths, the result is p, there is substance The complete spectral range of less than the available solar energy is used to generate electricity. To this end, another wavelength of light of the present invention has a significantly modified power output. 5 The purpose is to provide a long-lasting response to the solar spectrum, and therefore provide
10 15 先前薄膜裝置所存在的另一問題係在於透明傳導 (TCO)層。此層常譬如用來作為該裝置所用之前電極。立應 為透明藉以讓光以最小衰減進入胞元。 “ 透明傳導層常藉由被施加至玻璃支擇構件的表面之娘 摻雜氧化錫所組成。經摻雜氧化錫具有特定缺點，包括由 於其錫含量所導致之相對較高的成本，並已在許多案例中 被經掺雜氧化鋅所取代。氧化錫或氧化鋅的表面通常被钱 刻藉以粗化塗覆物的表面並進行“光困陷”。 基於不同理由，先前TCO層的效能並不如其應有水 準。光的透射、尤其在較長波長中已經受限。比起接收直 接曰光的情形’當如陰天令所接收的光呈現擴散時，此問 題將更加嚴重。 20 》此，本發明另—目㈣提供-經改良的透明傳導淨 覆物，其透射來自擴散光及直接光兩者之在較長波長㈣ 收之更大百分比的光。一項目的亦在於提供比起氧化錫塗 層物更便宜且更穩定之塗覆物。 為了企圖改良非晶石夕袭置，已提議使用多晶石夕作為非 201001728 晶矽的取代物。 一種此物質係為微晶矽，以矽中的微晶體為大多數。 另一提議已在於產生微晶矽裝置。然而，咸信尚未發展出 任一類型商業上成功的裝置。 5 過去製造微晶或奈米晶材料之製程的一項問題係在於 過份緩慢及昂貴。並且，先前尚未得知一併入有奈米晶發 之適當複合結構。 因此，本發明另一目的係提供一相對較高速度之用於 製造奈米晶裝置的方法。 10 化學氣相沉積（CVD)製程中生產奈米晶材料及其他結 晶材料之另一問題係在於：當許多基材被同時放置在一反 應器中時，難以相同的濃度將一製程氣體輸送至基材的所 有部份。 因此，本發明另一目的係提供一其中使氣體平均地輸送 15 橫越可供形成薄膜之基材的整體表面之CVD反應器及方法。 根據本發明，藉由下列的數種不同新特性來達成上述 目的。 首先，提供一新奈米晶光伏打裝置，其中主動層具有 被嵌入具有另一形式之矽的一基質中之一有效量的奈米晶 20 矽。通常，基質係由非晶矽或者非晶矽與微晶矽的一組合 所組成。 一實施例中，奈米晶薄膜層係為裝置中的唯一主動 層。另一實施例中，奈米晶層係與一層的非晶矽呈縱列狀 使用以形成一具有改良特徵的複合光伏打裝置。 201001728 “上述的單一或縱列結構皆可配合使用—此處稱為 層（透明傳導性光困陷氧化物)，，之經改良的TCO層。 此層由兩組件所組成；一第一相對較厚層的-金屬性氧化 广諸如，，、屯粹形式的氧化鋅，及一位於其上之較佳遠為較 碌雜金輕氧化物層，諸如雜有小百分比的諸如 .呂：金屬之魏鋅。薄的經摻雜氧化物作為—電傳導層， 而純巩化辞作為經改良的光困陷層且其比起經摻雜氧化鋅 而言對於錢㈣缺祕。 乳化辞 10 15 TCLO層亦可用來分離上述縱列裝置中的奈米晶及非 晶層。 ^奈米晶層可由數種不同方法形成。然而，—較佳方法 係為可在單—反應器中同時處理一顯著數量的玻璃面板之 化學氣相沉積(CVD)方法。 可有利地藉由相對於矽烷含量大幅地增加氫含量來顯 著地更改用來作為製程氣體之矽烷及氫混合物。此外，利 用一電漿來增強製程。藉由射頻(RF)電能發展出電漿，且 破塗覆的每單位面積面板之能量濃度係大幅地增加高於過 去所用的最大位準。其他製裎變數亦實質地増大。 較仏地，在反應器中利用一將氣體分配至複數個中处 導管之歧管使製程氣體平均且快速地輸送至多重面板的表 面’各中空導管在側邊具有多重之平均分隔的細微孔藉以 在分佈於面板表面上方的許多區位將氣體以小噴注輪迸至 面板表面。 該方法的結果係為用於形成奈米晶材料之製程中的顯 20 201001728 著速度增加、及製造成本的對應降低。 本發明的不同特性之結果係為產生一經改良的高效 能、低成本光伏打模組，其有希望可顯著地降低將光轉換 成電力之成本。 5 藉由下文描述及圖式來敘述或得知本發明之上述目的 及優點。 圖式簡單說明 第1圖為根據本發明所構成之一光伏打面板或模組的 正視立體圖； 10 第2圖為第1圖所示的面板之後視立體圖； 第3圖為根據本發明所構成之一薄膜光伏打胞元的一 實施例之放大部份示意橫剖視圖； 第4圖為根據本發明所構成之一光伏打胞元的另一實 施例之類似第3圖的圖式； 15 第5及6圖為顯示本發明之光伏打裝置的特性之圖形； 第7A及7B圖為顯示主要由微晶矽（第7A圖）及奈米晶矽 (第7B圖）所組成之一光伏打胞元的主動薄膜層的一部分之 晶結構的橫剖視放大不意圖， 第8圖為切過根據本發明所製造之一層奈米晶材料的 20 一橫剖面之顯微照片； 第9圖為所顯示的奈米晶材料但具有較低放大率之類 似於第8圖的顯微照片； 第10圖為根據本發明所構成之一化學氣相沉積反應器 的立體部份示意圖； 201001728 第11圖為第10圖所示之結構的—部分之部分為示意性 的放大端視圖； 第12圖為亦可用來進行本發明的製程之一先前技術反 應器構造之類似第11圖的圖式； 5 帛13圖為可用來作為薄膜光伏打裝置所用的-建造區 塊之一經傳V性塗覆的玻璃面板之放大示意橫剖視圖，包 括此文他處所描述者；及 第14圖為顯不第13圖所示裝置的不同層之特徵特性的 比較之圖形。 10 【實施方式】 光伏打模組 第1圖為根據本發明所構成的一光伏打面板或模組之 正視立體圖。面板30為標準尺寸，近似26吋寬度及55吋高 度（0.666公尺乘以1.4〇公尺），其可用來安裝在屋頂上、作為 15 建築物側邊之鋪面、定位在野外作為太陽能場中的主動元 件、及類似用途。 該等面板特別適合用於將日光轉換成電能，但亦可以 較小或較大尺寸用來將來自室内照明及其他來源的光轉換 成電能。 ' 20 各面板30具有一前玻璃板32、及一與其固接之後破螭 板34(請見第2圖）。各玻璃板係具有一邊界區域％，其可配 合在一框架中用來安裝面板。後玻璃板34係用來保護前破 璃板32的金屬化後表面，並對於面板添加結構性強度。 面板係具有一對電引線38及40以用來將面板32連接至 9 201001728 其他面板、或連接至諸如DC至AC轉換器、儲存電池、或其 他利用裝置等設備。 面板30包括一形成於前玻璃板32的内表面上之薄膜光 伏打結構。該薄膜結構係藉由熟知的“刻劃，，製程被分離成 50個平行互連的條帶42。 各條帶上的光伏打胞元係序列式連接至下個條帶上的 月匕元使传全部50個胞元橫越面板30寬度被序列式連接至彼 此藉以將各條帶中所產生的電壓添加至下一個所產生的電 壓。若譬如各條帶在完全、直接日光中產生近似6〇〇毫伏特 、面板產生的總電壓為近似3 0伏特。然而，電壓輸出將 依據連接在一起之胞元數量、及所使用的胞元類型而定。 經序列式連接的條帶之邊緣側的一接觸部係藉由一紹 各條T44電性連接至一接合箱48 ,而另一鋁條帶46將胞元 的相對邊緣連接至同一接合箱。該等接觸部係在接合箱中 缚·接至電引線38及40。 處於完全、直接曰光時來自面板30的電流輸出將依據 所使用胞元的類型及面板的總表面積而定。 面板30—般將在直接日光中產生大於〗〇〇瓦特的電力。 、面板30為黑色。然而，面板可利用該技藝熟知的技術 致賦予黑色以外的顏色。面板30可被製成使其不透射光通 &面板，或可被製成透射其所接收之光的一部分藉以容許 邠分的光進入一有可能在其中作為外側窗口或裴飾面板之 建築物内部中。 米晶裝詈 10 201001728 薄膜光伏打胞元50的 第3圖為根據本發明所構成之 大幅放大切除示意橫剖視圖。 第3圖所示的胞元5〇係 ^ 坡埤基材52所組成，玻璃基 材526又a十成在則碩56所示方向接 门接收光並予以透射至玻璃基 材上表面上的溥膜層，而在 ^ ^ ^ , ”中破轉換成電能。 應/主思玻璃板52的底表面诵 ^ ^ ^ + 囱通㊉為平坦，因為其來自於 破璃可為普通的鹼石灰 一使用汙式玻埚製程之製造設施。 兩者皆容易取得且具有中 玻璃、但較佳為低鐵含量破璃， 等成本。 10 15Another problem with previous thin film devices is the transparent conduction (TCO) layer. This layer is often used as the front electrode for the device. It should be transparent so that light enters the cell with minimal attenuation. "The transparent conductive layer is often composed of mother-doped tin oxide applied to the surface of the glass-retaining member. The doped tin oxide has certain disadvantages, including relatively high cost due to its tin content, and has In many cases it is replaced by doped zinc oxide. The surface of tin oxide or zinc oxide is often used to roughen the surface of the coating and "light trap". For different reasons, the effectiveness of the previous TCO layer It is not as good as it should be. The transmission of light, especially at longer wavelengths, has been limited. This is a more serious problem when the light received is diffused as in the case of direct sunlight. 20 》 Thus, the present invention provides a modified transparent conductive net coating that transmits a greater percentage of light from both the diffused light and the direct light at longer wavelengths (four). One purpose is to provide a ratio A tin oxide coating is a cheaper and more stable coating. In an attempt to improve the amorphous stone, it has been proposed to use polycrystalline as a substitute for non-201001728 wafer. One of the materials is microcrystalline The majority of the microcrystals in the crucible are already in place. Another proposal has been to create a microcrystalline germanium device. However, Xianxin has not developed any type of commercially successful device. 5 In the past, processes for manufacturing microcrystalline or nanocrystalline materials. One problem is that it is excessively slow and expensive. Also, a suitable composite structure incorporating nanocrystals has not been previously known. Therefore, another object of the present invention is to provide a relatively high speed for the manufacture of nanometers. Method of crystal device. 10 Another problem in the production of nanocrystalline materials and other crystalline materials in the chemical vapor deposition (CVD) process is that when many substrates are placed in a reactor at the same time, it is difficult to achieve the same concentration. A process gas is delivered to all portions of the substrate. Accordingly, it is another object of the present invention to provide a CVD reactor and method in which a gas is uniformly transported across an entire surface of a substrate from which a film can be formed. The invention achieves the above object by several different new features as follows. First, a new nanocrystalline photovoltaic device is provided, wherein the active layer has been embedded with another form. An effective amount of nanocrystals 20 一 in a matrix of ruthenium. Generally, the matrix is composed of amorphous ruthenium or a combination of amorphous ruthenium and microcrystalline ruthenium. In one embodiment, the nanocrystalline thin film layer is The only active layer in the device. In another embodiment, the nanocrystalline layer is layered with a layer of amorphous germanium to form a composite photovoltaic device having improved features. 201001728 "The single or columnar structure described above All can be used together - referred to herein as a layer (transparent conductive light trapped oxide), an improved TCO layer. The layer consists of two components; a first relatively thick layer of metal-oxidized, such as, a pure form of zinc oxide, and a layer of light oxide oxide thereon , such as a small percentage of such as: Lu: metal Wei zinc. The thin doped oxide acts as an electrically conductive layer, and purely as a modified trapped layer of light, and it is less secretive than money (d). Emulsification Word 10 15 The TCLO layer can also be used to separate the nanocrystalline and amorphous layers in the above column device. The nanocrystalline layer can be formed by several different methods. However, the preferred method is a chemical vapor deposition (CVD) process which can simultaneously process a significant number of glass panels in a single reactor. It is advantageously possible to significantly modify the mixture of decane and hydrogen used as a process gas by substantially increasing the hydrogen content relative to the decane content. In addition, a plasma is used to enhance the process. The plasma is developed by radio frequency (RF) electrical energy, and the energy concentration per unit area of the broken coating is greatly increased above the maximum level used in the past. Other system variables are also substantial. Increasingly, a manifold that distributes gas to a plurality of central conduits is used in the reactor to deliver process gas to the surface of the multiple panels on average and quickly. 'The hollow conduits have multiple subdivisions of fine separation on the sides. The holes are used to circulate the gas to the panel surface with a small injection jet at a number of locations distributed over the surface of the panel. The result of this method is the increase in speed and the corresponding reduction in manufacturing cost in the process for forming nanocrystalline materials. The result of the different characteristics of the present invention is to produce an improved high performance, low cost photovoltaic module that is promising to significantly reduce the cost of converting light into electricity. The above objects and advantages of the present invention will be described or become apparent from the following description and drawings. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a front perspective view of a photovoltaic panel or module constructed in accordance with the present invention; 10 is a rear perspective view of the panel shown in FIG. 1; and FIG. 3 is a perspective view of the panel according to the present invention. An enlarged view of an embodiment of a thin film photovoltaic cell is schematically cross-sectional view; FIG. 4 is a view similar to FIG. 3 of another embodiment of a photovoltaic cell constructed in accordance with the present invention; 5 and 6 are graphs showing the characteristics of the photovoltaic device of the present invention; FIGS. 7A and 7B are diagrams showing photovoltaics mainly composed of microcrystalline germanium (Fig. 7A) and nanocrystalline germanium (Fig. 7B). The cross-sectional view of the crystal structure of a portion of the active thin film layer of the cell is not intended to be enlarged, and FIG. 8 is a photomicrograph of a cross-section of a layer of nanocrystalline material produced according to the present invention; The nanocrystalline material is shown but has a lower magnification similar to the photomicrograph of Fig. 8; Fig. 10 is a schematic perspective view of a chemical vapor deposition reactor constructed according to the present invention; 201001728 The picture shows the part of the structure shown in Figure 10. Is a schematic enlarged end view; Fig. 12 is a view similar to Fig. 11 of one of the prior art reactor configurations which can also be used to carry out the process of the present invention; 5 帛13 is a view which can be used as a thin film photovoltaic device - an enlarged schematic cross-sectional view of a V-coated glass panel of one of the building blocks, including those described elsewhere herein; and Figure 14 is a comparison of the characteristic characteristics of the different layers of the device shown in Figure 13 Graphics. [Embodiment] Photovoltaic Module FIG. 1 is a front perspective view of a photovoltaic panel or module constructed in accordance with the present invention. Panel 30 is a standard size, approximately 26 inches wide and 55 feet high (0.666 meters by 1.4 inches), which can be used to mount on the roof, as a pavement on the side of 15 buildings, and positioned in the field as a solar field. Active components, and similar uses. These panels are particularly well suited for converting daylight into electrical energy, but can also be used to convert light from indoor lighting and other sources into electrical energy in smaller or larger sizes. '20 Each panel 30 has a front glass panel 32 and a breakout panel 34 after it is secured (see Figure 2). Each of the glass sheets has a boundary area % which can be fitted in a frame for mounting the panel. The rear glass sheet 34 is used to protect the metallized back surface of the front glass panel 32 and to add structural strength to the panel. The panel has a pair of electrical leads 38 and 40 for connecting panel 32 to 9 201001728 other panels, or to devices such as DC to AC converters, storage batteries, or other utilization devices. The panel 30 includes a thin film photovoltaic structure formed on the inner surface of the front glass plate 32. The film structure is separated into 50 parallel interconnected strips 42 by well-known "scouring. The photovoltaic cell lines on each strip are serially connected to the lunar element on the next strip. Passing all 50 cells across the width of the panel 30 are serially connected to each other to add the voltage generated in each strip to the voltage generated by the next. If, for example, each strip produces an approximation in complete, direct sunlight At 6 〇〇 millivolts, the total voltage produced by the panel is approximately 30 volts. However, the voltage output will depend on the number of cells connected together and the type of cell used. One contact on the side is electrically connected to a junction box 48 by a strip T44, and the other aluminum strip 46 connects the opposite edges of the cells to the same junction box. The contacts are in the junction box. Bonding to electrical leads 38 and 40. The current output from panel 30 in full, direct calendering will depend on the type of cell used and the total surface area of the panel. Panel 30 will generally produce greater than direct sunlight. 〗 〇〇 Watt's electricity. The panel 30 is black. However, the panel can impart a color other than black using techniques well known in the art. The panel 30 can be made such that it does not transmit light and panel panels, or can be made to transmit light it receives. Part of it allows the light to enter into the interior of a building where it is possible to act as an outside window or enamel veneer. Fig. 3 201001728 The third picture of thin film photovoltaic cell 50 is constructed in accordance with the present invention. The enlarged cross-sectional view is enlarged and enlarged. The cell 5 shown in Fig. 3 is composed of a sloping substrate 52, and the glass substrate 526 is again received in the direction indicated by the glory 56 and transmitted to the light. The ruthenium layer on the surface of the glass substrate is broken into electric energy in ^^^,". The bottom surface of the glass plate 52 should be flat. Because it is from the glass, it can be a common soda lime. A manufacturing facility using a dirty glass bottle process. Both are easy to obtain and have medium glass, but preferably low iron content, and the like. 10 15
I 玻璃的上表面59較佳如藓由 戈精由蝕刻呈‘‘紋路狀，，，藉以盡 量加大留置於胞元内之光量並“產生電能。 玻璃的底或前表面53且右_ & ” ’一虱化矽（Si〇2)的一塗覆物 54。玻璃的前表面53通常為平坦 l而1定望復物盡罝減少反 射且盡量加大經由前表面進人_中之日光吸收。 第3圖所示沉積在破璃上类；l 喝上表面上之不同層的材料係由 波狀線所料。這錢的波敎麵純代表表面被粗 化，但表面中的凹部較佳將極小以使其無法以肉眼看見且 在圖中不可見。 玻璃基材52的上紋路狀表面上之第_層观為另_| 2〇 的二氧化矽，其如熟知般作為一擴散障壁。 根據本心明的特性，不採用平常的TCO(透明傳導氧 化物），已提供有-複合層，其由一具有一祕刻上表面之 第-層m ‘本徵”氧化鋅(iZn〇)及—位於氧化鋅頂上的 實質較薄層62所組成。層62係由捧雜有從1 5%至5%的一已 11 201001728 知金屬性換雜材料使層62具電傳導性之氧化辞所組成。氧化 鋅可如熟知般摻雜有銘、鎵、銦或删。然而，較佳係為紹。 由本徵氧化鋅與傳導性氧化鋅組成之複合層在此處較 佳將稱為“TCLQ”層。其極為有利處係在於其提供形成光 5伏打胞元的前接觸部所需要之電傳導層，但顯著地改良光 的透射’特別是在較長波長中尤然且顯著地比諸如氧化錫 (Sb〇2)等物質更穩定，其當受到CVD反應器製造中所使用 的氫大氣4係 < 損失其部分的透射性(如細^篇㈣。 TCLO層不只改良直接光的透射，且亦改良擴散光的透射。 使用此TCLO複合層極為有利。一使用此TCL〇複合層 之月已凡係相較於先前1(：〇塗覆物而言造成輸出電流密度的 顯著增加。 &形成於層62頂上之層64係為胞元的主動層，且其利用 本發明的另—新穎特性，亦即，層64含有-有效量的奈求 15 晶。 將層64顯示為遠比其他層更厚，只單純為了容易顯示 其内部結構。 層64的下部分係為p_型奈米晶石夕，上部分為卜型奈米晶 夕而中間段為“i”或本徵性（未摻雜）奈米晶矽。 2〇 ·第7Ag|示意地顯示層64若由微晶破製成以高度放大率 觀看寺之情形。長且細薄的微晶94係欲入-主要由非晶石夕 組成之基質96中。 第⑺圖顯示相同的層64若主要由奈米晶矽製成之情 形。結晶結構係由具有類似微晶寬度的寬度但高度顯著較 12 201001728 短之“顆粒，，製成。 矽“奈米晶”在此處係被界定為半奈米（〇 5 nm)至近似 十奈米（1〇腿)之間寬度、及從十至—百奈米（1〇咖至⑽ nm)長度的個別晶體。 5 ‘‘微晶”係被界定為約為相同寬度或更寬、最多達到一 百奈米（100 nm)及從一百至四百奈米（1〇〇 nm至細nm)長 度之矽的晶體。 非晶矽當然為非晶性。 在用於製作層64的非晶材料之下述製程中，其中形成 10有奈米晶之基質係以非晶性為大多數，但如上述，有可能 額外出現部分微晶。 —有效董的奈米晶材料係為主動層中至少3〇%重量 的材料。一大多數”的奈米晶材料係超過50%。雖無上限， 在π於約70%的數量，進一步增加係可能提供 遞減的回 15 報。因此，較佳範圍係為30%至70%奈米晶材料。 對於裝置50的主動層64提供一有效量、或一大多數的 奈米晶材料係極為有利。奈米晶材料的出現咸信可顯著地 改良太陽能電池胞元的量子效率（QE)。此外’可以比先前 溥膜矽裝置實質更快的速率來沉積奈米晶層以形成商業上 20 可接受之裝置。 尚且，相較於非晶矽胞元，如此處所述般製成的奈米晶 矽胞元係提供效率上的顯著改良。如熟妒，#晶矽裝置的效 率係隨時間而變成劣化。本發明實質地消除此嚴重缺陷。 位於主動層糾頂上之下一層66係為另〆層具傳導性之 13 201001728 經摻雜氧化鋅。其用來作為一擴散障壁及一抗反射塗覆物。 最終的塗覆物係為銘68的一塗覆物，其作為用於光伏 打胞元之背電極。 第5及6圖顯示相較於非晶矽裝置而言之奈米晶及微晶 5 矽裝置的經改良特徵之圖形。 第5圖中的虛線曲線8G係、為相對於所接收光的波長所缚 製顯示典型非晶胞元的量子效率_之繪圖。繪圖係涵蓋: 四百至-千-百奈米_麵至11()()·)之電磁頻譜區。 顯然奈米晶響應曲線82在曲線8〇與82之間的區域料中 10提供遠為較大的量子效率及光成為電能的轉換。譬如在 nm曲線82的QE呵於〇·2’而對於曲線8〇的邱則接近 為零。這代表晶性裝置在電流及電壓輸出上優於非晶裝置 之顯著增加。 第6圖在9G顯不藉由奈米晶材料之電輸出的百分比劣 而曲線88顯示一非晶裝置隨時間經過的劣化，兩者皆 曝路在均等於直接日光5 5倍強烈度(丨_邮之很強烈的人 工光底下。 ^在日光中之實際使用中，非晶裝置通常在30至80天中 2〇釔疋下來，然後在其剩餘壽命期間不會進一步劣化。 &同的時間期間中’本發明的奈米晶裝置顯示並無劣 化或顯示極小劣化。 &咸彳D非晶裝置與本發明者之間的轉換效率差異將持續 又tt年的期間’且事實上將歷時裝置有效壽命的時程。 J微曰曰石夕展現出奈米晶材料的部分相同優點’其製 14 201001728 造所需要的時間彼顯著地大於使用本專利申請案的新穎製 程來製作奈米晶材料之時間，其將詳述於下文。 並且’咸信奈米晶裝置將供應每胞元之一較高電壓輸 出（550至600毫伏特，相較於先前的4〇〇至5〇〇毫伏特）及一較 5高電流輸出（譬如，高於20且最多達到26毫安培每平方公分 的胞元表面積’相較於先前的小於2〇毫安培）。 第8及9圖係為一奈米晶矽層在橫剖面中由穿透式電子 顯微鏡攝取的顯微照片。 第9圖顯不晶體對準於箭頭方向，其實質地垂直於位居 10 下方及右方之基材。 比起第9圖的照片更加放大之第8圖係顯示處於近似原 子位準之材料的相同表面之一部分。 第8圖的每個黑點代表一原子。一條規則線的原子代表 —晶格構。兩圖中皆可看出具有奈米晶顆粒。 15 看見數奈米至數十奈米大小的晶域（crystalline domains)。其被非晶性本質之不規則顆粒邊界所圍繞。顆粒 邊界及懸鍵(dangling bonds)係以原子氫(未圖示）被鈍化。可 藉由紅外線光譜術來評估氫含量。 第9圖顯示生長在彼此上方之奈米晶矽顆粒。玻璃基材 2〇 位於顆粒底部。此結構係増加薄膜内側之光的吸收及内部 光散射。此特性係增高胞元的短路電流密度。 因為奈米晶顆粒的對準本質及顆粒邊界及重組中心 (懸鍵、陷阱）的有效鈍化，係以經過奈米晶顆粒之最小散射 發生少數電荷載體(正及負光載流子）的運送。 15 201001728 可藉由諸如拉曼光譜術(Rarnan Spectroscopy)、X射線 散射分析、高解析度穿透式電子顯微分析及/或紅外線光譜 術等已知技術來偵測奈米晶矽在一層中之出現及數量。 縱列裝晋 第4圖顯示一類似於第3圖所示的裝置50之縱列光伏打 裝置70 ’差異在於一非晶矽層首先形成於基材上，然後— 奈米晶層形成於非晶層的頂上。 10 15 20 確切來說，非晶矽層72形成於傳導氧化鋅層62上。然 後，另一TCLO複合層形成於非晶矽層的頂上。TCL〇複合 物係由頂上具有—經摻雜ZnCKcZn。)層76之iZn〇層74所組 成。複合層74、76作為與第—TCLO層相同的用途。奈米晶 矽層64建造於TCLO結構74、76頂上。 縱列裝置70利用非晶材料及奈米晶材料的不同帶隙來 ^供比諸如第3圖所示裝置5〇等奈米晶裝置更高之輸 壓及功率位準。 电 曰曰曰可產生每胞元_毫伏特的電壓，而非 ;曰層^產生_毫伏特。胞元的總電壓係為個別電壓的她 電流輪出通常將略為較低，咖^ ㈣毫安培_平方公分，但功率輸 置同 縱列裝置可有利地並未 樣大的轉換效率劣化。 製造製裎 X到與一單獨為非晶矽裝 上文弟3及4圖所示的|, 凌置可以至少兩不同方式製造 16 201001728 一方式利用化學氣相沉積(CVD)，另一方式藉由濺鍍。 化學氣相沉穑 第10圖為一化學氣相沉積(CVD)反應器100的立體圖， 其可很有利地用來製造本發明的光伏打裝置。 5 反應器100包括絕緣側壁102及104，一頂壁106及一底 壁108。亦提供一後壁但因為已切除反應器的往後延伸部分 以提高裝置在圖中的可見度而未顯示於第1〇圖。 提供一鉸接式門11()。鉸接式門110當反應器處於使用 中時被緊緊地關閉，並具有一氣密性密封件以當處於操作 10 中時幫助反應器室中維持一很低壓力。 一習知真空泵112設有其進入反應器底部中之入口開口 以生成及維持反應器中的很低壓力。反應器係在操作期間由 熟知的部件加熱至相對較高的内部溫度，最多達到45(rc。 第10至12圖顯示用於輸送氣體至將在反應器中被塗覆 15 的玻璃基材之兩不同結構性配置。第10及11圖示意性顯示 藉由本發明的一特性所改良之一結構，而第12圖示意地顯 示一習知反應器的氣體輸送系統。 首先參照第12圖所示的習知結構，將被塗覆的玻璃板 係成對配置於反應器内側。第12圖為顯示諸如對132、134 2〇 等五對之橫剖端視圖。玻璃板係配置為以其前表面接觸一 很平的鋁板136、138、140、142或144、及一下支樓角产構 件145、及一上通路構件143。構件145及143較佳由一在高 溫將維持其形狀之諸如氟聚合物如杜邦（Dup〇m)“鐵弗龍 (Teflon)”等非傳導性塑料材料製成。將被塗覆的破璃板之 17 201001728 外側表面係曝露於反應器中的氣體，而前表面則齊平抵住 金屬板136等’未接觸於氣體^正是這些外側表面將被塗覆。 各鋁板支撐結構136、143、145亦提供一電傳導性框架 以將RF信號施加至玻璃板來形成一電漿，將如下文所述。 5 再參照第12圖，製程氣體被輸送至一歧管140，在其底 壁中具有許多很小的孔。製程氣體在一輕微正壓力下被進 給至歧”40中以將氣體往下輸送經過小孔來形成許多小 氣體喷注141。氣體往下流向真空泵112且在途中與玻璃板 的經曝露表面產生接觸。玻璃板很緊密地分佈，通常彼此 10 呈6至32mm的一距離。 射頻(RF)電信號的-供源146、147或148等係連接至每 隔—個用於固持住玻璃板對之板136。中間對如第12圖所示 被接地。這些RF供源在破螭板之間的開放空間中建立一電 磁场。如熟知’ RF場生成—電漿於含氣製程氣體中藉以輔 15助氣體的組份沉積在經曝露的氣體板表面上以形成薄膜塗 覆物。 k去已用來產生溥膜非晶石夕塗覆物之典变製程中， 與虱及氬混合之矽烷氣體係經由歧管14〇及氣體喷注141被 進給至反應器室内。 2〇 其上形成有主動矽層之玻璃板較佳係已在前表面上塗 覆有一一氧化矽抗反射塗覆物，且在將被塗覆的相對表面 上塗覆有- T C 0塗覆物。這些塗覆物通常由減鍵設備形成。 藉由RF產^ n產生諸如13.56百萬赫料適當頻率或 該頻率的倍數之-RF信號以在氬，院-氮氣體混合物中發 18 201001728 展出一電漿。這造成矽烷變成解離為矽及氫，所產生的材 料具有公式Si+2H2。 在氣體中首先使用約3%的三曱硼以產生矽的p層，其 後形成未經摻雜或本徵層的矽’然後以約3%重量比率將膦 5 (PH3)混入於起始氣體以形成η-型矽層。 所產生的產物較佳被送到一濺鑛線以鋪設透明傳導性 氧化物(TCO)及金屬電極層來完成裝置。 如同熟知，可在製程中的不同點進行中間刻劃步驟藉 以开/成個別胞元42(第1圖）並如上述使胞元序列式互連。這 10 些步驟已為人熟知而此處不予詳述。 雖然可使用第12圖所示的結構來製造本發明的光伏打 結構’較佳的反應器顯示於第⑺及丨丨圖。 如第ίο及11圖所示，下降的中空氣體分配導管118、 122、126、130等係散佈於諸如由金屬板115、U3、⑴及 15 1〇9及氟聚合物通路121、125、127、129等及角鋼1〇5所支 撐之板119、120等相鄰對的玻璃板之間。這些導管係被定 形為類似玻璃板且具有近似相同的面積。導管各者在兩側 壁中具有複數個彳艮小的孔142(第1〇圖），但只在一側壁中具 有孔之導管118除外，該壁係面對玻璃板12〇的左表面。類 20似地，位於反應器的相對端之一分配導管（未圖示）係只在面 朝最近的玻璃板之表面中具有孔。 如第11圖的117、123、124所示，譬如，導管壁中的小 孔各產生一微小噴注的氣體，其被導往將被塗覆的玻璃板 表面之一者。利用此方式，氣體被實質平均地分配橫越玻 19 201001728 璃板各者的整體表面藉以確保新鮮氣體以近似相等濃度抵 達各玻璃板的實質全部區域。因為若無導管118、122等則當 氣體從反應器頂部流至底部時氣體濃度可能實質地改變之 事實，咸信這可容許使用遠比原本可能者更高之氣體流率。 5 較佳地，氣體自一歧管114輸送經過一入口硬管116， 其可經由譬如在入口周圍設有適當真空緊密性密封件之後 壁進入殼體。若入口 116位於反應器100的後壁中，其將背 離第11圖的紙面。然而，為方便圖示，其被顯示成自左方 進入。 10 利用此構造，藉由使耦合件116切斷連接並使金屬支撐 軌道107中的單元滑過反應器的前開口以令整體歧管114及 所有分配導管118、122等可以一單元自反應器殼體被移 除，所以整體單元可被移除及重新放置以供清理。 較佳地，歧管114及其他組件皆由一可抵抗反應器中所 15 使用氣體及電漿的腐蝕之金屬製成，並可在反應器所使用 溫度下維持其強度及形狀。不銹鋼係為此用途的一種理想 材料，但也可使用銘。 如第11圖所示，金屬框架及分配導管之電性連接係不 同於第12圖所示者。第11圖配置中，一 RF產生器13卜133、 20 138或137係連接至玻璃框架各者且分配導管被接地。可能 將連接予以反轉，其中使分配導管板連接至產生器且玻璃 板接地。其目的在於確保一分配導管與玻璃板之間沒有障 壁阻止電漿形成於兩者之間以輔助沉積製程。 基於下述目的，提供一信號供應單元139以將一負3伏 20 201001728 特DC偏壓及一低頻率調變信號供應至各RF供源。 製程參數 不論疋使用第11或12圖的氣體分配系統，用於製造奈 米晶裝置之製程參數係仰賴對於先前慣例產生大幅背離。 5 確切來說，製程的兩參數相當具有顯著意義。第一， 製程氣體中之氫量對於矽烷量的比值應為從近似十至約五 百。這大幅地不同於用來製造非晶或微晶矽之先前慣例， 其中氫對於矽烷的比值咸信小於十。 第二，雖然典型習知技術慣例咸信使用約為半千瓦每 1〇平料尺待塗覆㈣之RFi》率輸出，根據本發明，功率位 準將增南至從近似—至十千瓦每平方公尺的範圍。 雖然上述參數咸信最具顯著意義，其他參數亦應被顯 著地往上調整。 製程氣體輸入流率應被增加最多達到五十標準聲每分 15鐘。這大幅列於咸信過去已採狀遠為較低的流率。 並且，溫度範圍應從200。〇至400。〇的平常範圍增加至 小於200°C至45(TC的範圍。 最後，工作壓力應位於衫達近似十托耳(㈣的最大 值之範圍。 2〇 製程開始前真空泵將反應ϋ排空至1G.6毫托 絲壓力。這係在製程開始之前自反應器清除所有過多: 载i體。 當輸送製程氣體時，反應器中的工作壓力 10托耳降錢.3毫托耳的範圍中。 持在從 21 201001728 根據另一特性，藉由自供源139將一負三伏特DC(-3V DC)偏壓信號供應至各產生器的輸入來修改從RF供應器輸 送至反應器之RF信號。 此外，各RF供源應為脈衝式以約一百循環每秒的速率 5 將其關斷及接通。 提供對於各供源之負三伏特DC偏壓藉以當離子轟擊 被塗覆的玻璃表面時將離子減慢藉以防止過度轟擊對於塗 覆物之損害。 RF供源各者以約為一百循環每秒的低速率之脈動係提 10 供功率輸送中的間隙，咸信其有助於奈米晶的發展且亦有 助於防止被形成的晶層的固定離子轟擊所致之損害。 製程氣體混合物中氫對於矽烷的比值之大幅增加咸信 係極為有利。咸信氫的大幅增加將加速其中沉積有薄膜之 表面上的長晶。並且，因為氫係以充填鬆弛鍵(loose bonds) 15 來盡量減少重組之已知用途使用於混合物中，其以此等龐 大數量使用咸信係具有確保潔淨、新鮮的氫可供充沛取用 以盡量加大有益的氫活性並盡量減少石夕層中的重組之進一 步有益目的。 由於上述之緣故，若氫位準維持在上文提供的範圍 20 中、且若功率位準類似地維持在一高位準，咸信可確保一 大多數的奈米晶矽將被形成於裝置的主動層中。 不只所產生的裝置具有上文對於奈米晶矽所描述的全 部細微特徵，而且製造製程咸信遠比可用於光伏打胞元的 生產處理之先前製程更快速。事實上，雖然咸信實際所達 22 201001728 成用以形成非晶及微晶矽以產生可接受的商業產品之最高 速率係為約一埃每秒，咸信本製程將以最多達到三十五埃 每秒的速率產生奈米晶材料。 下列疋一顯不典型參數組的三項範例之表格，其可在 使用本發9月製造奈米晶光伏打裝置中被有效利用。 參數 範例1 範例2 範例3 (a)氫對於矽烷比值 30 80 250 以扎每平万公尺為單位之RF功率 800 1000 3000 (C)標準升每分鐘之流率 50 125 380 ⑷溫度範圍 190 200 250 (e)壓力範圍 0_5m托耳 1托耳 10托耳 ⑴脈衝率 1-100 Hz 1 KHz 1〇〇 κηΓ 濺鐘方法 濺鑛製私中，玻璃基材具有前表面上之二氧化石夕抗反射 材料的塗覆物、及相對表面上之如上述的一tcl〇塗覆物。 板被放置在一金屬（鋁或銀）支撐板上之第一濺鍍室 中 ^^供源係附接至板而室被排空且基材被加熱。 基材隨後被移至一第二室在其中生長奈米晶矽層。所 使用的乾材為純石夕。室中的賤鑛氣體係為氯、氮，及依順 序具有氡之聚甲硼或二硼烷。 或者，一經掺雜;5夕乾材可配合使用氬加上氫的一混合 15 物作為濺鍍氣體。 第三及第四室含有一純矽靶材，其具有氫及氬作為濺 鍍氣體以形成本徵層。 第五室含有一矽靶材’其具有混合於氫加上氬濺鍍氣 體之膦氣，或-經磷摻雜的石夕把材，其具有氮加上氮。 23 201001728 使用第六室將板退火及熱處理最多達到250°C。 使用第七室使板冷卻下來，且使用第八室來卸載該板。 隨後’較佳在進一步的濺鍍室中，添加TCO及金屬接 觸層。 5 濺鍍室中亦應使用如上文對於CVD製程所指定之相對 較高位準的氫氣及相對較高RF功率位準及偏壓及控制電壓 及氣體流率。 透明傳導性糸困陷（TCLO)某柑 第13圖為一TCLO塗覆基材150的橫剖視切除圖，其形 10 成類似上文第2及3圖所示者之薄膜光伏打裝置的基礎。 為了圖式簡單起見’第13圖中已省略該等裝置中所使 用之二氧化石夕層。 TCLO基材係包括一玻璃基材構件152，其具有一層純 氧化鋅（iZnO)154及位於頂上之一層傳導性氧化鋅 15 15b(cZnO: A1)。 較佳如第2及3圖所示被添加有二氧化矽層之基材15〇 係形成一可以極理想的建造區塊被製造及供應至光伏打胞 元製造商之獨特產品。其亦可以胞元製程的一部份由此等 製造商製造。 20 玻璃基材152較佳係為低鐵玻璃，但可為鹼石灰玻璃， 或對於包括可見光的電磁輻射呈透明之另一材料，並可承 受最多達到近似450°c的溫度且對於氫電漿具有低的反應 性（reactivity)。 為了增強困陷光的能力，玻璃的上表面16〇應諸如藉由 24 201001728 蝕刻被粗化。所產生的表面不規則部之尺寸理想上將與被 透射的光之波長王現相同數量級、或接近可以合理成本取 得的尺寸。 形成於基材152上之不同層較佳係在序列中以數個濺 5鍍室所組成之一直列式濺鍍系統中由一系列的濺鍍步驟形 成，類似於一組裝線。 任何給定濺鍍室中所使用的濺鍍製程可為_DC濺鍍 f 製程、一RF濺鍍製程、一磁控濺鍍製程或任何其他適當的 濺鍍製程。 10 毯氧化鋅之層 再度參照第13圖，已知稱為本徵氧化辞或i Zn〇之一層 I54的貫質純、未摻雜氧化鋅(Zn〇)係沉積在玻璃基材152的 、、’文路狀上表面160上。所使用的濺鍍製程較佳係為一具有一 鋪覆的DC偏壓之rf濺鍍製程。 15 範例 ϋ 基材152放置在一RF濺鍍室中。基材152譬如為近似26 吋寬乘以55吋長、及近似3至1〇 mm厚的一玻璃板。室為近 似32吋寬乘以60吋長乘以1〇吋高。一形成為長方形區塊之 99.999%或更高純度的Zn〇靶材係在室内被附接至一電 極至壁係接地，而電極則耗合至一rf產生器。 室隨後被實質地排空。一較佳為氬的惰性氣體係被導 入以使室中的壓力成為最多達到在近似〇〇1至〇.03毫巴 (milhbar)。整體室被加熱至近似2〇(rc。 射頻波隨後藉由RF產生器產生並施加至祀材電極的背 25 201001728 表面。射頻波具有近似300瓦每五吋_平方的功率密度、及 較佳13.56 MHz的頻率或其整數倍數，但可選自一寬廣的頻 率頻譜。 這些條件下，氬氣係被轉換成一電漿狀態。近似負3伏 5 特的一DC_偏壓被施加至電極。這強力地吸引及加速帶正電 的氬原子核。氬原子核被加速朝向電極，並衝擊211〇靶材 的表面。此衝擊係敲鬆被沉積在玻璃板152上之個別zn〇分 子以形成一薄膜。 藉由此製程，iZnO層154以近似1至5埃每秒生長。所想 10 要的厚度為500nm至750nm(50〇〇埃至7500埃）。依需要，層 154可比此範圍更厚或更薄. 當其沉積在表面上，ZnO分子自然地形成六角形晶 體’垂直於基材152的上表面160。此規則結構稱為一“纖維 紋路’，。一般而言，一纖維紋路係存在於當一晶學軸線沿著 15 一晶性集合體的某較佳方向被對準之時。 表面姓釗 為了具有改良的光困陷特徵，極想要蝕刻iZn〇層的頂 表面162。為了蝕刻表面162，可使用諸如藉由將表面沾浸 於一池中之蝕刻等已知的製程。蝕刻係產生一具有2至3〇〇 20 微米範圍尺寸的斷面之規則紋路狀表面。 或者，層162可在其於濺鍍製程中生長之時受到蝕刻。 這具有數項優點。-主要優點係在於：其將藉由消除餘刻 專用的一特定步驟來達成與一酸池相同之紋路狀結果但需 要較少時間。不只消除了酸蝕刻本身的時間，且包括與 26 201001728 自濺鍍室移除裝置150、進行酸蝕刻、將裝置放置在相同戈 一不同濺鍍室中以供下個步驟用、及重新生成—高户真* 相關之所有時間及能量。 mi 另一優點係為：藉由將iZn〇層154曝露於一蝕刻劑如3 5 先前技術酸池中所作之情形，係留下一固有的髒污表面 來自蝕刻劑的殘留化學物將被留下，而需要一額外清理製 程。藉由消除後生長蝕刻步驟及其蝕刻劑，將具有較小化 學殘留物需要移除。 為了同時地生長及蝕刻iZn◦層，在濺鍍步驟中作出— 10 小修改。並非只將一諸如氬等惰性氣體插入濺鍍室中，亦 週期性注射一少量的水蒸氣。 水蒸氣初始係呈氣態以H2〇的分子存在。然而，當曝露 於濺鍍室内的條件時，其解離成H+及OH-的離子。當〇H-離 子衝擊於生長層154上時係與Zn〇起反應以形成水、氧及純 15 辞，其被釋放至室大氣中且由真空泵所通泄。 因為OH離子與氧化辞反應，而非只作為一供另一製程 所用之觸媒，水應予歸還。僅藉由在製程全程中不時地將 更多水蒸氣添加至濺鍍室中來予以達成。利用此方式，Zn〇 層在其生長時被連續地姓刻，結果一經钱刻表面305比起一 ° 酸池所產生者具有遠為更少的殘留物。 化鋅之層 摻有再度參照第13圖’表面162頂上’藉由濺鑛來沉積輕度 1〜雜質之-薄層15_ZnQ。如上述，所使用的雜質較 •、銪，但可為鎵、銦、或硼。此層154稱為傳導性氧化鋅 27 201001728 或c-ZnO。層154較佳係為近似25〇·厚且具 從二分之一至三分之一厚度。 層的 習知技藝的技術通常係提供粗略以2至5%重 雜質之心〇。本發财，比率為近似15%。藉此只輕= 雜氧化鋅，c-ZnO比拓2 5 s〇/a-j/· 又7 起至5/。！摻雜的氧化鋅更容易 長波長的光，如第14圖所—m 勿透射車乂 圖所不，同時保留2至5%經摻雜氧 鋅之所想要的傳導性特徵。 10 15 20 為了控制摻雜的百分比’可採用多種方法的-者。 =Μ圭將為一以所想要的確切比值含有辞及鋁之物 。，此方式，如該技藝所習知，被沉積的層將具有鱼 靶材相同比值之成份。 /、 料材可為純辞製成，如同用來雜i-Zn〇層之無 材3有摻雜劑的氣體係在濺錢製程期間被注射至濺於 '室内。藉由已知機構’摻雜劑將被併人所形成的新層中。又 k具有&夠使單—室中的後續操作改變摻雜比值 =改㈣材之優點。尚且，亦可能在層156生長全程的不 同時間改變摻雜劑的百分比。 第14圖藉由虛線曲線164顯示2至抓經摻雜氧化辞的 ^射特徵。如其中顯示，對於具有最多達到近似麵肺 、/’ =之光，2至5〇/〇經摻雜氧化鋅係透射近似Μ。〆。的光。然 而二當波長增加超過8〇〇nm ’ 2至5%經摻雜的氧化鋅係使光 顯著地衰減。 相反地，曲線168顯示輕度摻雜（1.5%)的氧化辞之光透 、對衣小於近似8〇〇nm的波長，透射曲線係與對於2 28 201001728 n雜孔化鋅之透射曲線實質地相同。然而 雜的氧化鋅係繼續透射處於繼至謂nm區中的波長= 百分比的光。口右 大 负對於大於近似110〇1101波長的光，輕 雜的氧化鋅係_衰減光_¥部分。 度检 5 1-¾ 導層 。根據本發明的另一特性，傳導層156的段165製成較薄 或在J區中被完全移除，而形成延伸通過或部份地通過層 156之孔或凹坑。這些孔咸信可容許透射更多的光而不顯著 地降低層156的電流攜載產能。較佳地，孔165將粗略地覆 10蓋層156的表面積之耻寫。淨結果咸信係域於一給定 所接收光量之光伏打裝置的電流輸出之一增加。 可利用蝕刻、諸如藉由沾浸至一酸池中來製作孔165。 另一製程係在濺鍍製程期間以來自解離水蒸氣的ΟΗ- 離子姓刻之時同時地生長層156,如同上文對於同時生長及 蝕刻層154所描述。The upper surface 59 of the I glass is preferably etched in a ''grain shape' by gemin, thereby maximizing the amount of light remaining in the cell and "generating electrical energy. The bottom or front surface of the glass 53 and right _ & ; "A coating 54 of a bismuth (Si〇2). The front surface 53 of the glass is generally flat and l is expected to reduce the reflection as much as possible and to maximize the absorption of sunlight through the front surface. Figure 3 is deposited on the broken glass; l The material that is used to drink the different layers on the surface is determined by the wavy line. The wavefront of this money is purely representative of the surface being roughened, but the recesses in the surface are preferably so small that they are not visible to the naked eye and are not visible in the figure. The first layer on the upper grain-like surface of the glass substrate 52 is another cerium oxide, which is known as a diffusion barrier as is well known. According to the characteristics of the present invention, instead of the usual TCO (transparent conductive oxide), a composite layer having a first layer m' intrinsic zinc oxide (iZn〇) having a secret upper surface has been provided. And consisting of a substantially thin layer 62 on top of the zinc oxide layer 62. The layer 62 is composed of a 5% to 5% of a 11 201001728 metal-replaced material to make the layer 62 electrically conductive. The zinc oxide may be doped with inscriptions, gallium, indium or by subtraction as well known. However, it is preferred that the composite layer composed of intrinsic zinc oxide and conductive zinc oxide is referred to herein as "TCLQ". "Layer. It is extremely advantageous in that it provides the electrically conductive layer required to form the front contact of the light 5 volt cell, but significantly improves the transmission of light', especially at longer wavelengths, particularly and significantly Substances such as tin oxide (Sb〇2) are more stable, and they are subjected to the hydrogen atmosphere 4 system used in the manufacture of CVD reactors. Loss of partial transmittance (eg, detail). The TCLO layer not only improves direct light. Transmission, and also improve the transmission of diffused light. It is extremely advantageous to use this TCLO composite layer. The use of this TCL 〇 composite layer has resulted in a significant increase in output current density compared to the previous 1 (: 〇 coating. & layer 64 formed on top of layer 62 is the active layer of the cell, And it utilizes another novel feature of the present invention, i.e., layer 64 contains an effective amount of 15 crystals. Layer 64 is shown to be much thicker than the other layers, simply for ease of display of its internal structure. The lower part is p_type nanocrystalline spar, the upper part is the type of nano-crystal and the middle part is "i" or intrinsic (undoped) nanocrystalline. 2〇·7Ag| If the ground display layer 64 is made of microcrystals, the temple is viewed at a high magnification. The long and thin microcrystals 94 are intended to be in the matrix 96 mainly composed of amorphous austenite. The figure (7) shows the same. If the layer 64 is mainly made of nanocrystalline germanium, the crystalline structure is made of "particles, which are shorter in width than the width of the microcrystals but are significantly shorter than 12 201001728. The "nano crystal" is here. Defined as a width between half nanometer (〇5 nm) to approximately ten nanometers (1 leg), and from ten to - hundred nanometers 1 〇 to (10) nm) individual crystals. 5 ''Microcrystalline' is defined as approximately the same width or wider, up to 100 nanometers (100 nm) and from one hundred to four hundred nanometers ( A crystal having a length of from 1 nm to a fine nm. Amorphous germanium is of course amorphous. In the following process for forming an amorphous material of layer 64, a matrix having 10 nanocrystals is formed therein. Amorphous is the majority, but as mentioned above, it is possible to additionally have some crystallites. - Effective nanocrystalline material is at least 3% by weight of the active layer. A majority of the nanocrystalline material system More than 50%. Although there is no upper limit, in the amount of π of about 70%, further increase may provide a declining return of 15 reports. Therefore, the preferred range is from 30% to 70% of nanocrystalline material. It is highly advantageous to provide an active amount, or a majority of the nanocrystalline material, to the active layer 64 of the device 50. The appearance of nanocrystalline materials can significantly improve the quantum efficiency (QE) of solar cell cells. In addition, the nanocrystalline layer can be deposited at a substantially faster rate than the prior art diaphragm unit to form a commercially acceptable device. Moreover, nanocrystalline stem cell lines made as described herein provide a significant improvement in efficiency compared to amorphous stem cells. As is well known, the efficiency of the #晶晶装置 becomes degraded over time. The present invention substantially eliminates this serious drawback. The layer 66 on the lower layer of the active layer is electrically conductive. 13 201001728 Doped zinc oxide. It is used as a diffusion barrier and an anti-reflective coating. The final coating was a coating of Ming 68 which served as the back electrode for the photovoltaic cells. Figures 5 and 6 show graphs of improved features of nanocrystalline and microcrystalline 5 germanium devices compared to amorphous germanium devices. The dashed curve 8G in Fig. 5 is a plot showing the quantum efficiency of a typical amorphous cell with respect to the wavelength of the received light. The drawing system covers: the electromagnetic spectrum area of four hundred to one thousand to one hundred meters to the surface of 11 () (). It is apparent that the nanocrystal response curve 82 provides a much larger quantum efficiency and conversion of light into electrical energy in the region between the curves 8A and 82. For example, QE in the nm curve 82 is lower than 〇·2' and Qiu in the curve 8〇 is close to zero. This represents a significant increase in the current and voltage output of the crystalline device over amorphous devices. Figure 6 shows that the 9G is not inferior to the percentage of the electrical output of the nanocrystalline material. Curve 88 shows the deterioration of an amorphous device over time. Both exposures are equal to 5 5 times the intensity of direct sunlight (丨_ The post is very intense under artificial light. ^In practical use in daylight, amorphous devices usually fall 2 times in 30 to 80 days and then do not deteriorate further during the remaining life. During the period, the nanocrystalline device of the present invention showed no deterioration or showed minimal deterioration. The difference in conversion efficiency between the salty D-amorphous device and the inventor will continue for a period of tt years and will actually last for a while. The time course of the effective life of the device. J micro 曰曰石夕 exhibits some of the same advantages of the nanocrystalline material. The time required for the manufacture of 14 201001728 is significantly greater than the use of the novel process of this patent application to make nanocrystals. The time of the material, which will be detailed below, and 'the salty nanocrystal device will supply a higher voltage output per cell (550 to 600 millivolts compared to the previous 4 to 5 millimeters Volt) and a higher than 5 high The flow output (for example, a cell surface area greater than 20 and up to 26 mA per square centimeter is less than 2 mA amps compared to the previous one). Figures 8 and 9 are a nanocrystalline layer in a cross section. A photomicrograph taken by a transmission electron microscope. Figure 9 shows that the crystal is aligned in the direction of the arrow, which is substantially perpendicular to the substrate below and to the right of 10. It is more magnified than the photo in Figure 9. Figure 8 shows a portion of the same surface of the material at approximately the atomic level. Each black dot in Figure 8 represents an atom. The atom of a regular line represents the crystal lattice. Rice crystal particles. 15 A few nanometers to tens of nanometers of crystalline domains are seen. They are surrounded by amorphous nature of irregular grain boundaries. Particle boundaries and dangling bonds are atomic hydrogen. (not shown) is passivated. The hydrogen content can be evaluated by infrared spectroscopy. Figure 9 shows the nanocrystalline granules grown on top of each other. The glass substrate 2 〇 is located at the bottom of the granule. This structure is added to the inside of the film. Absorption of light and inside Light scattering. This characteristic is to increase the short-circuit current density of the cell. Because the alignment nature of the nanocrystalline particles and the effective passivation of the particle boundary and the recombination center (dwelling bonds, traps) occur through the minimum scattering of the nanocrystalline particles. Transport of a small number of charge carriers (positive and negative photocarriers) 15 201001728 by means of Raman spectroscopy, X-ray scattering analysis, high-resolution transmission electron microscopy and/or infrared spectroscopy Known techniques such as techniques to detect the presence and number of nanocrystals in a layer. The fourth column of the vertical array shows a similarity to the tandem photovoltaic device 70 of the device 50 shown in Fig. 3. The amorphous germanium layer is first formed on the substrate, and then a nanocrystalline layer is formed on top of the amorphous layer. 10 15 20 Specifically, the amorphous germanium layer 72 is formed on the conductive zinc oxide layer 62. Then, another TCLO composite layer is formed on top of the amorphous germanium layer. The TCL ruthenium complex has a doped ZnCKcZn on top. The iZn germanium layer 74 of layer 76 is comprised. The composite layers 74, 76 serve the same purpose as the first TCLO layer. A nanocrystalline layer 64 is formed on top of the TCLO structures 74,76. The tandem device 70 utilizes different band gaps of the amorphous material and the nanocrystalline material to provide higher input voltage and power levels than nanocrystalline devices such as the device 5 of Figure 3. The enthalpy produces a voltage of _ millivolts per cell, instead of 曰 millivolts. The total voltage of the cell is the individual voltage. Her current turnout will usually be slightly lower, but the power output and the tandem device can advantageously have no large conversion efficiency degradation. Manufacture of 裎X to a separate amorphous 矽 上文 上文 上文 上文 上文 上文 3 3 上文 上文 , , , , , , , , , 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 By sputtering. Chemical Vapor Deposition Fig. 10 is a perspective view of a chemical vapor deposition (CVD) reactor 100 that can be advantageously used to fabricate the photovoltaic device of the present invention. 5 Reactor 100 includes insulated sidewalls 102 and 104, a top wall 106 and a bottom wall 108. A rear wall is also provided but is not shown in Figure 1 because the rearward extension of the reactor has been removed to increase the visibility of the device in the drawing. An articulated door 11 () is provided. The articulated door 110 is tightly closed when the reactor is in use and has a hermetic seal to assist in maintaining a very low pressure in the reactor chamber when in operation 10. A conventional vacuum pump 112 is provided with an inlet opening into the bottom of the reactor to create and maintain a very low pressure in the reactor. The reactor is heated during operation to a relatively high internal temperature by well-known components up to 45 (rc. Figures 10 through 12 show the glass substrate used to deliver gas to the coating 15 to be coated in the reactor. Two different structural configurations. Figures 10 and 11 schematically show one structure improved by a feature of the present invention, and Figure 12 schematically shows a conventional gas delivery system of the reactor. A conventional structure is shown in which the coated glass sheets are arranged in pairs on the inside of the reactor. Fig. 12 is a cross-sectional end view showing five pairs such as pairs 132, 134 2 , etc. The glass sheet is configured to be The front surface contacts a very flat aluminum plate 136, 138, 140, 142 or 144, and a lower gusset member 145, and an upper access member 143. The members 145 and 143 are preferably maintained by a shape that will maintain its shape at elevated temperatures. The fluoropolymer is made of a non-conductive plastic material such as Dup〇m "Teflon". The outer surface of the coated glass panel 17 201001728 is exposed to the gas in the reactor, and The front surface is flush against the metal plate 136, etc. It is the gas that exactly these outer surfaces will be coated. Each of the aluminum plate support structures 136, 143, 145 also provides an electrically conductive frame to apply an RF signal to the glass sheet to form a plasma, as will be described below. Referring again to Fig. 12, the process gas is delivered to a manifold 140 having a plurality of small holes in its bottom wall. The process gas is fed to the manifold 40 at a slight positive pressure to transport the gas downward. A plurality of small gas jets 141 are formed through the apertures. The gases flow down the vacuum pump 112 and come into contact with the exposed surface of the glass sheets on the way. The glass sheets are closely spaced, typically at a distance of 6 to 32 mm from each other. The (RF) electrical signal source 146, 147 or 148 is connected to every other plate 136 for holding the glass plate pair. The intermediate pair is grounded as shown in Fig. 12. These RF sources are broken. An electromagnetic field is established in the open space between the rafts. As is well known, 'RF field generation—the plasma is deposited in the gas-containing process gas by means of a component of the auxiliary gas on the surface of the exposed gas plate to form a thin film coating. K. has been used to produce yttrium film amorphous In the process of the coating process, the decane gas system mixed with helium and argon is fed into the reactor chamber via the manifold 14 and the gas injection 141. 2 The glass plate on which the active layer is formed is compared. The preferred surface has been coated with a cerium oxide antireflective coating on the front surface and coated with a -TC 0 coating on the opposite surface to be coated. These coatings are typically formed by a key reduction device. An RF signal, such as an appropriate frequency of 13.56 megahertz or a multiple of this frequency, is generated by the RF generator to exhibit a plasma in the argon, gas-nitrogen mixture at 18 201001728. This causes the decane to become dissociated. Hydrogen, the material produced has the formula Si+2H2. First, about 3% of tri-boron-boron is used in the gas to produce a p-layer of germanium, after which an undoped or intrinsic layer of germanium is formed, and then phosphine 5 (PH3) is mixed in at about 3% by weight. The gas forms an η-type ruthenium layer. The resulting product is preferably sent to a splash line to lay a transparent conductive oxide (TCO) and metal electrode layer to complete the device. As is well known, intermediate scribing steps can be performed at different points in the process by opening/forming individual cells 42 (Fig. 1) and sequentially interconnecting the cells as described above. These 10 steps are well known and will not be detailed here. Although the structure shown in Fig. 12 can be used to fabricate the photovoltaic structure of the present invention, a preferred reactor is shown in (7) and in the drawings. As shown in Figs. 11 and 11, the descending hollow gas distribution conduits 118, 122, 126, 130, etc. are interspersed, for example, by metal plates 115, U3, (1) and 15 1〇9 and fluoropolymer passages 121, 125, 127. 129, etc. and the angled steel 1 〇 5 supported by the plates 119, 120 and other adjacent glass plates. These conduits are shaped like glass sheets and have approximately the same area. Each of the conduits has a plurality of small bores 142 in the side walls (Fig. 1), except for the conduit 118 having a bore in one of the side walls that faces the left surface of the glass sheet 12〇. Similarly, one of the dispensing conduits (not shown) at the opposite end of the reactor has holes only in the surface of the glass sheet facing the nearest glass. As shown in 117, 123, 124 of Fig. 11, for example, the small holes in the wall of the conduit each produce a tiny injection of gas which is directed to one of the surfaces of the glass sheet to be coated. In this manner, the gas is substantially evenly distributed across the entire surface of each of the glass sheets to ensure that fresh gas reaches substantially all of the substantial area of each glass sheet at approximately equal concentrations. This is because, if there are no conduits 118, 122, etc., the gas concentration may change substantially as the gas flows from the top of the reactor to the bottom, which allows for the use of gas flow rates that are much higher than would otherwise be possible. Preferably, the gas is conveyed from a manifold 114 through an inlet tube 116 which can enter the housing via, for example, a suitable vacuum tight seal around the inlet. If the inlet 116 is located in the rear wall of the reactor 100, it will face away from the paper of Figure 11. However, for convenience of illustration, it is shown entering from the left. With this configuration, the integral manifold 114 and all of the distribution conduits 118, 122, etc. can be self-reactored by disconnecting the coupling member 116 and sliding the unit in the metal support rail 107 through the front opening of the reactor. The housing is removed so the unit can be removed and repositioned for cleaning. Preferably, the manifold 114 and other components are made of a metal that is resistant to corrosion by gases and plasmas used in the reactor and maintains its strength and shape at the temperatures used in the reactor. Stainless steel is an ideal material for this purpose, but it can also be used. As shown in Fig. 11, the electrical connection between the metal frame and the distribution duct is different from that shown in Fig. 12. In the configuration of Fig. 11, an RF generator 13 133, 20 138 or 137 is attached to each of the glass frames and the distribution conduit is grounded. It is possible to reverse the connection where the distribution conduit plate is connected to the generator and the glass plate is grounded. The purpose is to ensure that there is no barrier between a dispensing conduit and the glass sheet to prevent plasma formation between the two to aid in the deposition process. A signal supply unit 139 is provided to supply a negative 3 volt 20 201001728 special DC bias and a low frequency modulation signal to each RF supply based on the following purpose. Process Parameters Regardless of the gas distribution system of Figure 11 or 12, the process parameters used to fabricate the nanocrystal device rely on a significant deviation from previous practices. 5 Specifically, the two parameters of the process are quite significant. First, the ratio of the amount of hydrogen in the process gas to the amount of decane should be from about ten to about five hundred. This is largely different from previous conventions used to make amorphous or microcrystalline germanium, where the ratio of hydrogen to decane is less than ten. Second, although the typical conventional technical practice uses a RFi rate output of about half a kilowatt per 1 inch of flat material to be coated (four), according to the present invention, the power level will increase from approximately to ten kilowatts per square. The range of meters. Although the above parameters are most significant, other parameters should be significantly adjusted upwards. The process gas input flow rate should be increased by up to fifty standard sounds per minute for 15 minutes. This is largely due to the fact that Xianxin has taken a much lower flow rate in the past. Also, the temperature range should be from 200. 〇 to 400. The normal range of 〇 is increased to less than 200 ° C to 45 (the range of TC. Finally, the working pressure should be in the range of approximately ten Torr (4). The vacuum pump will vent the reaction 1 to 1G before the start of the process. .6 millitors pressure. This removes all excess from the reactor before the start of the process: loading the body. When the process gas is delivered, the working pressure in the reactor is 10 Torr and the range is 3 mTorr. According to another feature, from 21 201001728, the RF signal delivered from the RF supply to the reactor is modified by supplying a negative three volt DC (-3V DC) bias signal from the supply 139 to the inputs of the respective generators. In addition, each RF supply should be pulsed to turn it off and on at a rate of about one hundred cycles per second. A negative three volt DC bias is provided for each supply source to thereby bombard the coated glass surface. The ions are slowed down to prevent excessive bombardment damage to the coating. Each of the RF supply sources provides a gap in the power supply at a low rate of about one hundred cycles per second. The development of nanocrystals also helps prevent the formation of Damage caused by fixed ion bombardment of the layer. The large increase in the ratio of hydrogen to decane in the process gas mixture is extremely advantageous. The large increase in salt hydrogen will accelerate the growth of crystals on the surface where the film is deposited. Hydrogen is used in mixtures to fill loose bonds 15 to minimize the known use of recombination. The use of such a large amount of salty letters ensures that clean, fresh hydrogen is available for maximum use. Useful hydrogen activity and minimizing the further beneficial purpose of recombination in the layer. For the above reasons, if the hydrogen level is maintained in the range 20 provided above, and if the power level is similarly maintained at a high level, The letter ensures that a majority of the nanocrystals will be formed in the active layer of the device. Not only does the resulting device have all of the fine features described above for the nanocrystal, but the manufacturing process is more than available. The previous process for the production of photovoltaic cells is faster. In fact, although the actual letter 22 201001728 is used to form amorphous and microcrystalline The highest rate of acceptable commercial products is about one angstrom per second, and the salty letter process will produce nanocrystalline materials at rates up to thirty-five angstroms per second. The following three items of the atypical parameter set A sample table that can be effectively utilized in the fabrication of nanocrystalline photovoltaic devices using this September. Parameter Example 1 Example 2 Example 3 (a) Hydrogen to decane ratio 30 80 250 in units of tens of meters per square meter RF power 800 1000 3000 (C) Standard liter per minute flow rate 50 125 380 (4) Temperature range 190 200 250 (e) Pressure range 0_5m Torr 1 Torr 10 Torr (1) Pulse rate 1-100 Hz 1 KHz 1〇 〇κηΓ Splashing method In the case of splashing, the glass substrate has a coating of a SiO2 anti-reflective material on the front surface, and a tcl 〇 coating as described above on the opposite surface. The plate is placed in a first sputtering chamber on a metal (aluminum or silver) support plate. The source is attached to the plate and the chamber is evacuated and the substrate is heated. The substrate is then moved to a second chamber where the nanocrystalline layer is grown. The dry material used is pure stone. The bismuth ore gas system in the chamber is chlorine, nitrogen, and polyboron or diborane in the order of ruthenium. Alternatively, once doped, the 5th dry material can be used as a sputtering gas in combination with a mixture of argon and hydrogen. The third and fourth chambers contain a pure ruthenium target having hydrogen and argon as a sputtering gas to form an intrinsic layer. The fifth chamber contains a target of 'phosphorus gas mixed with hydrogen plus argon-sputtered gas, or - a phosphorus-doped stone material having nitrogen plus nitrogen. 23 201001728 Use the sixth chamber to anneal and heat the plate up to 250 °C. The seventh chamber was used to cool the plate and the eighth chamber was used to unload the plate. Subsequently, it is preferred to add a TCO and a metal contact layer in a further sputtering chamber. 5 The relatively high level of hydrogen and relatively high RF power levels and bias and control voltages and gas flow rates as specified above for the CVD process should also be used in the sputtering chamber. TCLO, a TCLO, Figure 13 is a cross-sectional view of a TCLO coated substrate 150, which is shaped like a thin film photovoltaic device similar to those shown in Figures 2 and 3 above. basis. For the sake of simplicity of the drawings, the layer of dioxide used in these devices has been omitted from Fig. 13. The TCLO substrate comprises a glass substrate member 152 having a layer of pure zinc oxide (iZnO) 154 and a layer of conductive zinc oxide 15 15b (cZnO: A1) on top. Preferably, the substrate 15 to which the ruthenium dioxide layer is added as shown in Figures 2 and 3 forms a unique product in which the highly desirable building blocks are manufactured and supplied to photovoltaic cell manufacturers. It can also be manufactured by a manufacturer such as a part of the cell process. 20 Glass substrate 152 is preferably low iron glass, but may be soda lime glass, or another material that is transparent to electromagnetic radiation including visible light, and can withstand temperatures up to approximately 450 ° C and for hydrogen plasma Has low reactivity. In order to enhance the ability to trap light, the upper surface 16 of the glass should be roughened by etching, for example, by 24 201001728. The resulting surface irregularities will desirably be of the same order of magnitude as the wavelength of the transmitted light, or a size that is reasonably achievable. The different layers formed on substrate 152 are preferably formed by a series of sputtering steps in a sequential sputtering system consisting of several sputtering chambers in a sequence, similar to an assembly line. The sputtering process used in any given sputtering chamber can be a _DC sputtering f process, an RF sputtering process, a magnetron sputtering process, or any other suitable sputtering process. 10 The layer of zinc oxide of the blanket is again referred to FIG. 13, and a pure, undoped zinc oxide (Zn) which is known as an intrinsic oxide or a layer I54 of i Zn〇 is deposited on the glass substrate 152. , 'The path is on the upper surface 160. The sputtering process used is preferably an rf sputtering process with a blanket DC bias. 15 Example 基材 Substrate 152 is placed in an RF sputtering chamber. Substrate 152 is, for example, a glass sheet approximately 26 inches wide by 55 inches long and approximately 3 to 1 mm thick. The chamber is approximately 32 inches wide by 60 inches long by 1 inch high. A Zn 〇 target of 99.999% or higher purity formed into a rectangular block is attached indoors to an electrode to the wall ground, and the electrode is consumed to an rf generator. The chamber is then substantially emptied. An inert gas system, preferably argon, is introduced to bring the pressure in the chamber up to approximately 〇〇1 to 03.03 mbarhbar. The bulk chamber is heated to approximately 2 〇 (rc. The RF wave is then generated by the RF generator and applied to the back surface of the coffin electrode 25 201001728. The RF wave has a power density of approximately 300 watts per 吋 平方 square, and preferably The frequency of 13.56 MHz or an integral multiple thereof, but may be selected from a broad frequency spectrum. Under these conditions, the argon gas is converted into a plasma state. A DC_bias bias of approximately 5 volts is applied to the electrode. This strongly attracts and accelerates the positively charged argon nucleus. The argon nucleus is accelerated toward the electrode and impacts the surface of the 211 〇 target. This impact knocks out individual zn 〇 molecules deposited on the glass plate 152 to form a thin film. By this process, the iZnO layer 154 is grown at approximately 1 to 5 angstroms per second. The desired thickness is from 500 nm to 750 nm (50 angstroms to 7500 angstroms). If desired, the layer 154 may be thicker than this range or Thinner. When deposited on the surface, the ZnO molecules naturally form a hexagonal crystal 'perpendicular to the upper surface 160 of the substrate 152. This regular structure is referred to as a "fibre pattern". In general, a fiber line exists. When a crystal axis When a preferred direction of the 15-crystal aggregate is aligned, the surface surname 钊 is intended to have a modified light trapping feature, and it is highly desirable to etch the top surface 162 of the iZn layer. To etch the surface 162, for example, The etching process produces a regular grain-like surface having a cross-section having a size in the range of 2 to 3 〇〇 20 μm by a known process such as etching the surface into a bath. Alternatively, the layer 162 can be splashed in it. It is etched during growth in the plating process. This has several advantages. The main advantage is that it will achieve the same grain-like results as an acid bath by eliminating a specific step dedicated to the remainder but requiring less time. Not only does the time of the acid etch itself be eliminated, but it also includes the 26 201001728 self-sputter chamber removal device 150, acid etching, placement of the device in the same different sputtering chamber for the next step, and regeneration - All the time and energy associated with Takahashi*. Another advantage of mi is that by exposing the iZn〇 layer 154 to an etchant such as the 3 5 prior acid bath, an inherently dirty surface is left behind. From Eclipse The residual chemicals of the agent will be left behind, and an additional cleaning process is required. By eliminating the post-growth etching step and its etchant, there will be less chemical residues to remove. In order to simultaneously grow and etch the iZn layer, In the sputtering step, a small modification is made. Not only an inert gas such as argon is inserted into the sputtering chamber, but also a small amount of water vapor is periodically injected. The water vapor is initially present in a gaseous state as a molecule of H2〇. When exposed to conditions in the sputtering chamber, it dissociates into ions of H+ and OH-. When 〇H- ions impinge on the growth layer 154, they react with Zn to form water, oxygen and pure 15 words. It is released into the atmosphere of the chamber and is vented by a vacuum pump. Since the OH ion reacts with the oxidizing word, not just as a catalyst for another process, the water should be returned. This is achieved only by adding more water vapor to the sputtering chamber from time to time throughout the process. In this manner, the Zn layer is continuously engraved as it grows, with the result that the surface 305 has a much less residue than the one produced by the acid pool. The zinc layer is doped with a thin layer of 15_ZnQ deposited by splashing to re-deposit the light 1~ impurity layer by referring to Fig. 13 'top surface 162'. As described above, the impurities used are more than ?, but may be gallium, indium, or boron. This layer 154 is referred to as conductive zinc oxide 27 201001728 or c-ZnO. Layer 154 is preferably about 25 Å thick and has a thickness from one-half to one-third. The skill of the layer of conventional techniques is generally to provide a rough core of 2 to 5% impurities. This is a fortune, the ratio is approximately 15%. Therefore, only light = zinc oxide, c-ZnO ratio of 2 5 s 〇 / a-j / · 7 up to 5 /. ! The doped zinc oxide is more prone to long-wavelength light, as shown in Figure 14, which does not transmit the desired conductivity characteristics of 2 to 5% of the doped oxy-zinc. 10 15 20 In order to control the percentage of doping, a variety of methods can be used. =Μ圭 will contain the words and the aluminum in the exact ratio you want. In this manner, as is well known in the art, the deposited layer will have the same ratio of components of the fish target. /, the material can be made purely, as used in the miscellaneous i-Zn layer, the dopant system gas is injected into the room during the splash process. The dopants will be formed in a new layer by the known mechanism. Again k has & enough to make the subsequent operation in the single-chamber change the doping ratio = the advantage of the (four) material. Still, it is also possible to vary the percentage of dopants during the entire growth of layer 156. Figure 14 shows the characteristics of the 2 to the doping oxidation word by the dashed curve 164. As shown therein, for light having up to approximately the approximate lung, /' = light, the 2 to 5 〇/〇-doped zinc oxide transmission approximates Μ. Hey. Light. However, when the wavelength is increased by more than 8 〇〇 nm ' 2 to 5% of the doped zinc oxide system, the light is significantly attenuated. Conversely, curve 168 shows a lightly doped (1.5%) oxidized light transmission, a pair of coatings having a wavelength less than approximately 8 〇〇 nm, and a transmission curve and a transmission curve for the 2 28 201001728 n porous zinc. the same. However, the heterogeneous zinc oxide system continues to transmit light at wavelengths = percent in the region to the nm. Right and left large negative For light greater than approximately 110 〇 1101 wavelength, the lightly oxidized zinc oxide _ attenuates the light _¥ portion. Check 5 1-3⁄4 guide layer. According to another feature of the invention, the segment 165 of the conductive layer 156 is made thinner or completely removed in the J region to form holes or dimples that extend through or partially through the layer 156. These holes allow for more light to be transmitted without significantly reducing the current carrying capacity of layer 156. Preferably, the aperture 165 will roughly cover the surface area of the cover layer 156. The net result is an increase in the current output of the photovoltaic device for a given amount of light received. The holes 165 can be made by etching, such as by dipping into an acid bath. Another process is to simultaneously grow layer 156 with the sputum from the dissociated water vapor during the sputtering process as described above for the simultaneous growth and etch layer 154.
利用這些蝕刻方法的任一者，所產生的孔165將具有隨 機的區位，且孔的形狀可能為不規則。一般而言，可接受 隨機的區位及形狀。 應控制钱刻製程藉以不餘除大於50%的表面積，因此 2〇 不會在經摻雜Ζη〇處產生與層156其餘部分呈現隔離之島。 或者，可採用半導體製造技術所已知及使用之光微影 術蝕刻。這具有對於孔的區位提供確切控制且確保不會形 成島之優點。 較佳應使用同時性蝕刻，因為其比起一酸池更可被控 29 201001728 制’亚比光微影軸更為便宜且容易。 如上文已作描述，純氧化辞的層較佳係 鋅層156的2至3倍之間厚度。 W生减 t構利用下列事實：純氧化鋅的“霧霾因子(haze factor)很軸麵摻雜的氧化鋅者，但純氧化辞具有顯 更好的透射特徵，如第14圖所示。因此，藉由純氧化辞製 作大部份TCL()複合塗t物被―較高“霧霾 10 15 20 f-撕強之“光困陷，，(將光子留置於主動層15顯 保持在广位準同時改良了裝置的電流密度輸出。 产改良可在任何薄臈光伏打裝置中產生—電流密 又义σ犯的改良量值係高達30%或更高。 意本=t^:r顯示及描述本發明’應注 改，而不脫離由申料：:置的細節’可作出變化及修 %判祀_界定之本發明的範圍。 【圖式簡單說明】 第1圖為根據本發明所構成之一光 正視立體圖； *田败4梹蛆的 第2圖為第1圖所示的面板之後視立體圖； 第3圖為根據本發明所構成之一薄膜光 實施例之放大部份示意橫㈣圖； ⑽、 〃第4圖為根據本發明所構成之-光伏打胞元的另一實 施例之類似第3圖的圖式； ’ 貫 第5及6圖為_干太& 第7娜圖為顯示置的特性之圖形； 由微曰曰矽(第7A圖)及奈米晶矽 30 201001728 (第7B圖）所組成之一光伏打胞元的主動薄膜層的一部分之 晶結構的橫剖視放大示意圖； 第8圖為切過根據本發明所製造之一層奈米晶材料的 一橫剖面之顯微照片； 5 第9圖為所顯示的奈米晶材料但具有較低放大率之類 似於第8圖的顯微照片； 第10圖為根據本發明所構成之一化學氣相沉積反應器 的立體部份示意圖； 第11圖為第10圖所示之結構的一部分之部分為示意性 10 的放大端視圖； 第12圖為亦可用來進行本發明的製程之一先前技術反 應器構造之類似第11圖的圖式； 第13圖為可用來作為薄膜光伏打裝置所用的一建造區 塊之一經傳導性塗覆的玻璃面板之放大示意橫剖視圖，包 15 括此文他處所描述者；及 第14圖為顯示第13圖所示裝置的不同層之特徵特性的 比較之圖形。 【主要元件符號說明】 30···面板 32···前玻璃板 34···後玻璃板 36···邊界區域 38,40…電引線 42…胞元，條帶 31 201001728 44…銘f|條帶 46…铭條帶 48···接合箱 50…薄膜光伏打胞元 52···玻璃板，玻璃基材 53…玻璃的底或前表面 54…塗覆物 56…箭頭 58…第一層 59…玻璃的上表面 62…傳導氧化辞層 64…胞元的主動層 66·.·位於主動層64頂上之下一層 68"·!呂 70…縱列光伏打裝置 72…非晶矽層 74.. .1ZnO 層 76.. .經摻雜 ZnO(cZno)層 80…典型非晶胞元的量子效率(QE)之繪圖 82…奈米晶響應曲線 84 ··· 80與82之間的區域 88…非晶裝置隨時間經過的劣化 90…藉由奈米晶材料之電輸出的百分比劣化 94···微晶 32 201001728 96…主要由非晶矽組成之基質 100".化學氣相沉積(CVD)反應器 102,104…絕緣側壁 105.··角鋼 106···頂壁 107...金屬支撐軌道 108…底壁 109,111,113,115.·.金屬板 110…鉸接式門 112···真空泵 114,140···歧管 116···入口硬管,辆合件 117,123,124···導管壁中的小孔產生微小喷注的氣體 118,122,126,130.··中空氣體分配導管 119…板 120···玻璃板With either of these etching methods, the resulting holes 165 will have random locations and the shape of the holes may be irregular. In general, random locations and shapes are acceptable. The surface engraving process should be controlled so as not to remove more than 50% of the surface area, so 2〇 does not create islands that are isolated from the rest of layer 156 at the doped Ζη〇. Alternatively, photolithographic etching known and used in semiconductor fabrication techniques can be employed. This has the advantage of providing precise control over the location of the holes and ensuring that islands are not formed. Simultaneous etching should preferably be used because it is more controllable than an acid bath. 29 201001728 The 'abido photomicrograph axis is cheaper and easier. As described above, the layer of pure oxidized is preferably between 2 and 3 times the thickness of the zinc layer 156. The W-subtractive t-structure utilizes the fact that the "haze factor" of pure zinc oxide is very axially doped zinc oxide, but the pure oxidation has a better transmission characteristic, as shown in Fig. 14. Therefore, most of the TCL() composite coatings produced by the pure oxidation word are trapped by the "higher smog 10 15 20 f-", leaving the photons in the active layer 15 The wide-level standard also improves the current density output of the device. The production improvement can be generated in any thin-twist photovoltaic device. The current value is improved by up to 30% or higher. The invention is shown and described in its <RTIgt; </ RTI> <RTIgt; </ RTI> <RTIgt; </ RTI> <RTIgt; </ RTI> <RTIgt; An optical front view of the invention; a second view of the field of Fig. 1 is a rear perspective view of the panel shown in Fig. 1; and Fig. 3 is an enlarged view of an embodiment of the thin film light according to the present invention. Illustrated transverse (four) diagram; (10), 〃 Figure 4 is another embodiment of a photovoltaic cell constructed in accordance with the present invention A pattern similar to Fig. 3; 'Figures 5 and 6 are _ dry too & 7th natogram is a graphic showing the characteristics of the set; by micro 曰曰矽 (Fig. 7A) and nanocrystalline 矽 30 201001728 (Fig. 7B) is a cross-sectional enlarged view showing a crystal structure of a part of an active thin film layer of a photovoltaic cell; Fig. 8 is a cross section of a layer of nanocrystalline material fabricated according to the present invention. Photomicrograph; 5 Figure 9 is a photomicrograph showing the nanocrystalline material but having a lower magnification similar to Figure 8; Figure 10 is a chemical vapor deposition reaction according to the present invention. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 11 is an enlarged end elevational view of a portion of a portion of the structure illustrated in FIG. 10; FIG. 12 is a prior art reactor configuration which may also be used to carry out the process of the present invention. Figure 13 is a schematic cross-sectional view of a conductively coated glass panel that can be used as a building block for a thin film photovoltaic device, including the one described elsewhere herein. And Figure 14 shows the device shown in Figure 13 Graphic of comparison of characteristic characteristics of different layers. [Description of main component symbols] 30···Panel 32··· Front glass plate 34··· Rear glass plate 36···Boundary area 38, 40...Electrical lead 42... Yuan, strip 31 201001728 44... Ming f| Strip 46... Ming strip 48··· Binding box 50... Thin film photovoltaic cell 52···Glass plate, glass substrate 53... Glass bottom or front surface 54 ...coating 56...arrow 58...first layer 59...upper surface 62 of glass...transducing oxidized layer 64...active layer of cell 66·.. located on top of active layer 64 below layer 68"·! Lu 70... Column photovoltaic device 72...amorphous germanium layer 74..1 ZnO layer 76.. doped ZnO (cZno) layer 80...typical amorphous cell quantum efficiency (QE) plot 82...nanocrystal response Curve 84 ··· 80 between regions 80 and 82...attenuation of the amorphous device over time 90...by the percentage of the electrical output of the nanocrystalline material is degraded 94····microcrystals 32 201001728 96...mainly by amorphous germanium Composition of the substrate 100 " chemical vapor deposition (CVD) reactor 102, 104... insulated sidewall 105. · angle steel 106 · · top wall 107 ... metal branch Brace track 108... bottom wall 109, 111, 113, 115.. metal plate 110... articulated door 112···vacuum pump 114, 140··· manifold 116···inlet pipe, fittings 117,123,124···holes in the duct wall Tiny injection gas 118, 122, 126, 130. · Hollow gas distribution conduit 119... plate 120···glass plate
121,125,127,129…氟聚合物通路 131,133,137,138--RF 產生器 132,134 …對 136···鋁板支撐結構 138,142,144…很平的鋁板 139…供源，信號供應單元 14卜··氣體噴注 142···很小的孑L 33 201001728 143…上通路構件，鋁板支撐結構 145…下支撐角度構件，鋁板支撐結構 146,147,148…射頻(RF)電信號的供源 150...TCLO塗覆基材 152…玻璃基材構件 154···主動層,生長層，傳導性氧化鋅或c-ZnO層 156···ΖηΟ薄層，傳導性氧化辞層 160···基材152的上表面 162···ίΖηΟ層的頂表面 164···2至5%經摻雜氧化鋅的光透射特徵 165···孔,傳導層156的段 168…輕度摻雜(1.5%)的氧化辞之光透射特徵 305…經姓刻表面 34121,125,127,129...fluoropolymer passages 131, 133, 137, 138--RF generators 132, 134 ... pairs 136 · · aluminum plate support structure 138, 142, 144... flat aluminum plate 139... supply, signal supply unit 14 · gas injection 142 · · · very small孑L 33 201001728 143... upper access member, aluminum plate support structure 145... lower support angle member, aluminum plate support structure 146, 147, 148... radio frequency (RF) electrical signal supply 150... TCLO coated substrate 152... glass substrate member 154 ··· active layer, growth layer, conductive zinc oxide or c-ZnO layer 156···ΖηΟ thin layer, conductive oxidized layer 160··· top surface of substrate 152···· 164··· 2 to 5% light-transmissive characteristics of doped zinc oxide 165··· hole, segment 168 of conductive layer 156...lightly doped (1.5%) oxidized light transmission characteristic 305... Surface 34