TWI442567B - Charged balanced devices with shielded gate trench - Google Patents

Charged balanced devices with shielded gate trench Download PDF

Info

Publication number
TWI442567B
TWI442567B TW99101558A TW99101558A TWI442567B TW I442567 B TWI442567 B TW I442567B TW 99101558 A TW99101558 A TW 99101558A TW 99101558 A TW99101558 A TW 99101558A TW I442567 B TWI442567 B TW I442567B
Authority
TW
Taiwan
Prior art keywords
gate
trench
region
epitaxial layer
substrate
Prior art date
Application number
TW99101558A
Other languages
Chinese (zh)
Other versions
TW201029182A (en
Inventor
Francois Hebert
Original Assignee
Alpha & Omega Semiconductor
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from US12/321,435 external-priority patent/US7893488B2/en
Application filed by Alpha & Omega Semiconductor filed Critical Alpha & Omega Semiconductor
Publication of TW201029182A publication Critical patent/TW201029182A/en
Application granted granted Critical
Publication of TWI442567B publication Critical patent/TWI442567B/en

Links

Landscapes

  • Insulated Gate Type Field-Effect Transistor (AREA)
  • Metal-Oxide And Bipolar Metal-Oxide Semiconductor Integrated Circuits (AREA)
  • Electrodes Of Semiconductors (AREA)
  • Junction Field-Effect Transistors (AREA)

Description

帶有遮罩柵極溝道的電荷平衡裝置Charge balancing device with a masked gate channel

本發明涉及一種垂直半導體功率裝置,特別涉及一種具有單一薄外延層,依靠先進製造來實現的,可用於製備各種尺寸的帶有超級結結構和遮罩了柵極溝道的電荷平衡的垂直功率裝置,通過簡單、靈活的製作工藝,適用於不同的擊穿電壓。

The present invention relates to a vertical semiconductor power device, and more particularly to a single thin epitaxial layer, which is realized by advanced manufacturing, and can be used for preparing various sizes of vertical power with a super junction structure and a charge balance of a gate channel. The device is suitable for different breakdown voltages through a simple and flexible manufacturing process.

傳統的製造技術和裝置結構,雖然在減小的串聯電阻的同時,能進一步提高擊穿電壓,但仍然面臨許多技術難題。由於傳統高功率裝置的結構特點,通常需要多個費時、複雜和昂貴的製作過程,因此高壓半導體功率裝置的實際應用和實用性都是有限的。正如下文將要討論的那樣,高壓功率裝置的製作工藝都很複雜,而且產量和收益都很低。另外,半導體功率裝置通常並不是用原始半導體晶片製作,而是用帶有外延層的預處理晶片製作而成。這無疑增加了半導體功率裝置的製作成本。而且其功能和性能特徵也取決於,形成外延層時所用的工藝參數。因此,對於依賴於原始預處理晶片的功率裝置,這種預處理的晶片的使用,進一步局限了這些功率裝置的可製造性以及生產的靈活性。
相對於傳統工藝而言,超級結技術具有在不增加漏-源電阻Rdson的同時,獲得更高的擊穿電壓等優點。對於標準的功率電晶體單元,擊穿電壓很大程度上依賴於低摻雜的漂流層。因此,漂流層越厚,所能承受的額定電壓越高,但漏-源電阻Rdson卻大幅增加。在傳統功率裝置中,漏-源電阻Rdson與擊穿電壓BV近似複合以下函數關係:
Rdson BV2.5
相比之下,帶有超級結結構的裝置漂流區中達到了電荷平衡。漏-源電阻Rdson與擊穿電壓BV複合一個更加便於應用的函數關係式,即:
Rdson BV
因此在高壓裝置應用中,需要通過設計和生產帶有超級結結構的半導體功率裝置,以便降低漏-源電阻Rdson,同時獲得高擊穿電壓,提升裝置性能。漂流區中溝道附近的區域,帶有相反的導電類型。只要溝道附近的區域同樣摻雜相反的導電類型,漂流區的相對摻雜濃度就會比較高。在關閉狀態時,這兩個區域中的電荷相互抵消,漂流區呈耗盡狀態,可以承受高電壓,這被稱為超級結效應。在開啟狀態時,由於漂流區的摻雜濃度較高,所以其漏-源電阻Rdson比較低。
然而在製造功率裝置方面,傳統的超級結技術仍然會遇到許多技術上的難題與局限性。更確切地說,一些傳統結構中都要求帶有多外延層和/或掩埋層。根據以前的製作工藝,許多裝置結構都需要多次進行背部刻蝕和化學機械拋光(CMP)工藝。此外,這些製作工藝處理裝置的過程,有時並不符合標準的鑄造工藝。例如,有些標準的高產量半導體鑄造廠都具有氧化物化學機械拋光(CMP),但有些超級結技術中需要用到的矽化學機械拋光(CMP)卻沒有。因此,這些裝置的結構特點和製作工藝決定了,它們並不適用於從低壓到高壓的裝置應用。換言之,某些工藝成本過高,並且/或者工藝太過冗長複雜,並不適用於高額定電壓的裝置應用。下文還將繼續討論,這些具有不同結構特點、通過各種工藝製造的傳統裝置,都帶有阻礙這些裝置在市場需求中實際應用的困難和局限。
由於標準的VDMOS並不具備電荷平衡的功能特點,因此適用於高壓的半導體功率裝置的傳統類型包括帶有如第1A圖所示的標準結構的裝置。根據I-V(電流-電壓)性能測試,以及對這種類型裝置的模擬分析進一步證實:正是出於這個原因,擊穿電壓才沒有超過一維品質因數,即詹森極限。為了滿足高擊穿電壓的要求,帶有這種結構的裝置通常漏極漂流區的摻雜濃度較低,致使其導通電阻相對較高。為了降低導通電阻,這種裝置的晶片尺寸通常都很大。鑒於以上所述的缺點:晶片成本過高(每個晶圓上的晶片數量太少)以及不適用於標準封裝中的較大的晶片,因此儘管這種裝置的製作工藝簡單,而且生產成本不高,然而對於標準封裝中高電流、低阻抗的應用要求,它們卻並不滿足。
半導體功率裝置的第二種類型是帶有二維電荷平衡的結構,這種裝置對於給定的阻抗,可獲得高於詹森極限的擊穿電壓,或對於給定的擊穿電壓,可獲得低於詹森極限的電阻率(導通電阻Rdson×裝置面積)。這種類型的裝置結構通常稱為超級結技術裝置。在超級結結構中,基於在氧化物旁路的裝置中的PN結和靜電場起電板技術,在一個垂直裝置的漂流漏極區中,平行於電流方向上的電荷平衡,可以使裝置獲得更高的擊穿電壓。
第1B圖為一個帶有超級結裝置的橫截面視圖,通過增大漂流區中的漏極摻雜濃度,在保持擊穿電壓不變的情況下,降低裝置的電阻率(Rsp=阻抗×有源區)。通過在漏極中形成P-型(對於n-溝道裝置)垂直立柱,導致高壓下漏極在水準方向完全耗盡,在N+襯底處從漏極高壓中夾斷並遮罩溝道,從而達到電荷平衡。歐洲專利0053854(1982)和美國專利4,754,310中都已經提到了這種技術,具體在該專利的第13圖以及美國專利5,216,275中。之前的這些公開說明書中,垂直超級結都是作為N和P型摻雜物的垂直立柱。在垂直DMOS裝置中,如附圖所示,通過摻雜一個帶有側壁的結構,形成其中一個摻雜立柱,獲得垂直電荷平衡。除了摻雜立柱,美國專利4134123和美國專利6037632還提出使用摻雜浮島來增加擊穿電壓或降低電阻。超級結的這種裝置結構仍然通過耗盡P-區,遮罩柵極/溝道不受漏極影響。但由於電荷存儲和轉換等問題,這種浮島結構仍然受到很多技術難題的局限。
對於上述的超級結型裝置,由於其製作方法工序繁多、有些工序進度緩慢而且產量很低,因此要製備這種裝置通常相當複雜、昂貴,而且需要很長的加工時間。確切地說,這些工序包含多個外延層和掩埋層。部分結構甚至要求溝道深度要穿過整個漂流區,並且大多數工藝都需要進行背部刻蝕或化學機械拋光。總之,這些傳統結構和製作方法製作緩慢而且成本昂貴,並不經濟實用,不適於廣泛應用。
本專利申請為由本專利的發明人申請的美國專利12/005,878的部份接續申請案,其中提出了一種在深溝道中生長的帶有電荷平衡外延立柱的超級結裝置。溝道金屬氧化物半導體場效應管(MOSFET)形成在深溝道以及深溝道周圍區域上方的頂部外延層中。但是這種裝置的溝道柵極所處的電場較高,容易因電壓擊穿而受損。
因此,除了要改進這種超級結裝置的結構和製作工藝,還需要在擊穿時遮罩有源單元的敏感柵極。第1C-1圖至第1C-3圖表示美國專利6,635,906所述的在外延層的大部分層中,帶有P-浮島1的裝置。但是這些浮島不能自對準到柵極或溝道上,而且在電壓擊穿時,並不能有效地保護敏感溝道柵極。Takaya等人在2005年舉行的第17屆功率半導體裝置&積體電路國際論壇上發表的《浮島與厚底部氧化物溝道柵極金屬氧化物半導體場效應管(FITMOS)》一文中提出了一種結構,如第1D圖所示,這種結構表示,為了使漏極和在溝道柵極底部的P-區達到電荷平衡而植入的浮動P-區,可以於將柵極從P-區中分離出來。但是由於這些位於溝道柵極下方的P-植入區,與帶有厚底部氧化物的柵極溝道接觸,因此可能會減少開路時通過的電流量。
因此,在功率半導體裝置設計和製造工藝中,為了解決上述困難與局限,有必要找到一種新的功率裝置結構和製造方法。

Conventional manufacturing techniques and device structures, while reducing the breakdown voltage while reducing the series resistance, still face many technical challenges. Due to the structural characteristics of conventional high-power devices, multiple time-consuming, complicated, and expensive fabrication processes are often required, so the practical application and practicality of high-voltage semiconductor power devices are limited. As will be discussed below, high voltage power devices are complicated to manufacture and have low yields and benefits. In addition, semiconductor power devices are typically fabricated without the use of an original semiconductor wafer, but with a pre-processed wafer with an epitaxial layer. This undoubtedly increases the manufacturing cost of the semiconductor power device. Moreover, its functional and performance characteristics also depend on the process parameters used to form the epitaxial layer. Thus, for power devices that rely on the original pre-processed wafer, the use of such pre-processed wafers further limits the manufacturability and flexibility of production of these power devices.
Compared with the conventional process, the super junction technology has the advantages of obtaining a higher breakdown voltage while increasing the drain-source resistance Rdson. For standard power transistor cells, the breakdown voltage is highly dependent on the low doped drift layer. Therefore, the thicker the drift layer, the higher the rated voltage that can withstand, but the drain-source resistance Rdson increases significantly. In the conventional power device, the drain-source resistance Rdson and the breakdown voltage BV are approximately combined with the following functional relationship:
Rdson BV2.5
In contrast, charge balancing is achieved in the drift zone of a device with a superjunction structure. The drain-source resistance Rdson is combined with the breakdown voltage BV to be a more convenient functional relationship, namely:
Rdson BV
Therefore, in high voltage device applications, it is necessary to design and produce a semiconductor power device with a super junction structure in order to reduce the drain-source resistance Rdson while achieving a high breakdown voltage and improving device performance. The area near the channel in the drift zone has the opposite conductivity type. As long as the region near the channel is also doped with the opposite conductivity type, the relative doping concentration of the drift region will be higher. In the off state, the charges in these two regions cancel each other out, and the drift region is depleted and can withstand high voltage, which is called super junction effect. In the on state, since the doping concentration of the drift region is high, the drain-source resistance Rdson is relatively low.
However, in the manufacture of power devices, the traditional super junction technology still encounters many technical problems and limitations. Rather, multiple epitaxial layers and/or buried layers are required in some conventional structures. According to previous manufacturing processes, many device structures require multiple back etching and chemical mechanical polishing (CMP) processes. Moreover, these processes for fabricating process devices sometimes do not conform to standard casting processes. For example, some standard high-volume semiconductor foundries have oxide chemical mechanical polishing (CMP), but some of the super-junction techniques require hydrazine chemical mechanical polishing (CMP). Therefore, the structural characteristics and manufacturing process of these devices are determined, and they are not suitable for device applications from low pressure to high pressure. In other words, some processes are too costly and/or the process is too long and complicated to be suitable for high voltage rated device applications. As will be discussed further below, these conventional devices, which have different structural features and are manufactured by various processes, have difficulties and limitations that hinder the practical application of these devices in market demand.
Since the standard VDMOS does not have the functional characteristics of charge balancing, the conventional type of semiconductor power device suitable for high voltage includes a device having a standard structure as shown in Fig. 1A. According to the IV (current-voltage) performance test, and the simulation analysis of this type of device further confirms that it is for this reason that the breakdown voltage does not exceed the one-dimensional quality factor, the Jensen limit. In order to meet the high breakdown voltage requirements, devices with such a structure typically have a lower doping concentration in the drain drift region, resulting in a relatively higher on-resistance. In order to reduce the on-resistance, the wafer size of such devices is usually large. In view of the above disadvantages: the cost of the wafer is too high (the number of wafers on each wafer is too small) and is not suitable for the larger wafers in the standard package, although the manufacturing process of the device is simple and the production cost is not High, however, they are not satisfactory for high current, low impedance applications in standard packages.
The second type of semiconductor power device is a two-dimensional charge-balanced structure that achieves a breakdown voltage above the Jensen limit for a given impedance, or for a given breakdown voltage. Resistivity below the Jensen limit (on-resistance Rdson x device area). This type of device structure is commonly referred to as a superjunction technology device. In a superjunction structure, based on the PN junction and electrostatic field electrification technique in an oxide bypass device, in a drifting drain region of a vertical device, parallel to the charge balance in the direction of the current, the device can be obtained Higher breakdown voltage.
Figure 1B is a cross-sectional view of a device with a super junction. By increasing the drain doping concentration in the drift region, the resistivity of the device is reduced while maintaining the breakdown voltage (Rsp = impedance × Source area). By forming a P-type (for an n-channel device) vertical pillar in the drain, the drain is completely depleted in the horizontal direction at high voltage, and the channel is pinched and masked from the drain high voltage at the N+ substrate. Thereby achieving charge balance. This technique has been mentioned in the European Patent No. 0 053 854 (1982) and in the U.S. Patent No. 4,754,310, the entire disclosure of which is incorporated herein by reference. In the previous publications, the vertical superjunctions are vertical columns as N and P type dopants. In a vertical DMOS device, as shown in the drawing, one of the doped pillars is formed by doping a structure with sidewalls to obtain a vertical charge balance. In addition to doping columns, U.S. Patent No. 4,134,123 and U.S. Patent No. 6,037,632 also teach the use of doped floating islands to increase the breakdown voltage or reduce the resistance. This device structure of the super junction still passes through the depletion of the P-region, shielding the gate/channel from the drain. However, due to problems such as charge storage and conversion, this floating island structure is still limited by many technical problems.
For the super junction device described above, the preparation of such a device is generally complicated, expensive, and requires a long processing time due to the numerous processes, slow progress, and low yield. Specifically, these processes include a plurality of epitaxial layers and buried layers. Part of the structure even requires that the channel depth pass through the entire drift zone, and most processes require back etching or chemical mechanical polishing. In summary, these traditional structures and fabrication methods are slow and expensive to manufacture, are not economical, and are not suitable for a wide range of applications.
This patent application is a continuation-in-part of U.S. Patent Application Serial No. 12/005,878, the entire disclosure of which is incorporated herein by reference. A trench metal oxide semiconductor field effect transistor (MOSFET) is formed in the top epitaxial layer above the deep trench and the region around the deep trench. However, the electric field of the channel gate of such a device is high and is easily damaged by voltage breakdown.
Therefore, in addition to improving the structure and fabrication process of such a superjunction device, it is also necessary to mask the sensitive gate of the active cell during breakdown. Figures 1C-1 through 1C-3 show a device with a P-floating island 1 in most of the layers of the epitaxial layer described in U.S. Patent No. 6,635,906. However, these floating islands cannot self-align to the gate or the channel, and do not effectively protect the sensitive channel gate during voltage breakdown. Takaya et al. presented a "Floating Island and Thick Bottom Oxide Channel Gate Metal Oxide Semiconductor Field Effect Transistor (FITMOS)" at the 17th International Forum on Power Semiconductor Devices & Integrated Circuits, held in 2005. The structure, as shown in FIG. 1D, shows that the floating P-region implanted in order to achieve charge balance at the drain and the P-region at the bottom of the trench gate can be used to remove the gate from the P-region. Separated in the middle. However, since these P-emitter regions under the gate of the trench are in contact with the gate channel with a thick bottom oxide, the amount of current passing through the open circuit may be reduced.
Therefore, in the design and manufacturing process of power semiconductor devices, in order to solve the above difficulties and limitations, it is necessary to find a new power device structure and manufacturing method.

本發明一方面是為了提出一種新改良過的裝置結構和製作方法,通過簡單、便捷的製作工序,在漂流區中形成摻雜立柱,實現電荷平衡。無需背部刻蝕或化學機械拋光,精簡了加工步驟,只需形成一個單一的薄外延層,外延層同時生長在深溝道中和深溝道上方,以及深溝道周圍區域的頂面上,形成超級結結構。在溝道中的外延層部分形成外延立柱。在深溝道上方以及深溝道周圍區域表面上方的外延層部分,形成薄的頂部外延層,溝道金屬氧化物半導體場效應管單元就形成在這個頂部外延層中。這兩部分外延層可以作為單一外延層同時生長。電晶體單元的溝道柵極進一步被遮罩,一旦發生電壓擊穿,摻雜的遮罩區通過溝道柵極植入到柵極下方的漂流區,形成了自校準摻雜遮罩區,從而遮罩敏感柵極,解決了上述困難和局限。摻雜的遮罩區降低了溝道柵極處的峰值電場;還減慢了碰撞電離速度,增加了擊穿電壓。最終的結構提升了電參數的可靠性和穩定性。摻雜的遮罩區形成在溝道柵極下方的聚積區下方,並不接觸溝道柵極。在柵極溝道下方有一個額外的摻雜層,其導電類型和聚積區的導電類型相同,此摻雜層可以確保摻雜的遮罩區沒有接觸到柵極溝道,從而使裝置開啟時通過的電流更多。
本發明的另一方面在於,本發明所述的超級結結構和形狀可用來靈活調整所需的擊穿電壓的範圍。其製作工藝簡便,可使用標準的處理模組和設備通過標準工藝,方便地製備。由於此結構的電晶體部分,例如溝道柵極雙擴散金屬氧化物半導體(DMOS),是自校準的,因此還可進一步簡化製作工藝。上述技術難題與局限就會迎刃而解。
確切地說,本發明的一個方面在於提出了一種新改良過的裝置結構和製作方法,以便在深溝道中形成一個外延層,並且此外延層帶有一層薄的頂部外延層部分,覆蓋在裝置頂面上。此外延層的一部分也作為金屬氧化物半導體場效應管(在n-溝道金屬氧化物半導體場效應管的情況下為p-型)的本體區。此外,在這個頂部薄外延層中形成的金屬氧化物半導體場效應管單元,為溝道金屬氧化物半導體場效應管。溝道柵極通過帶有任選的溝道側壁和溝道底部摻雜植入區的頂部薄外延層打開,以消除可能受溝道柵極的深度以及外延層的摻雜濃度影響的溝道性能的敏感性。在用柵極多晶矽層填充柵極溝道之前,通過柵極溝道,將多個摻雜遮罩區植入到柵極下方的漂流區中。摻雜遮罩區的導電類型與金屬氧化物半導體場效應管的本體區的導電類型相同,並且摻雜遮罩區還擔負柵極遮罩摻雜區的作用,與柵極溝道自校準。摻雜遮罩區可以是浮島,或者是被連接(偏置)到深溝道中的外延層,從而也就被連接到了本體區。特別的是浮島的情況並不太理想,原因是浮動捕獲電荷,並使裝置漂流;被捕獲的電荷需要花時間擴散出去,這就減慢了電轉換。電晶體單元的性能可以通過簡單、便捷的製作工藝來控制和調整。本發明所述的超級結結構可以通過進一步的改進,應用到更廣泛的領域。
本發明的另一方面在於,提出了一種新改良過的裝置結構和製作方法,以便在一個薄的頂層上形成電晶體單元,其中薄頂層作為外延層覆蓋在深溝道上方,以及深溝道周圍和深溝道上方的頂面上。穿過深溝道側壁的離子注入(用和填充深溝道外延層的導電類型相反的離子)可以調節深溝道周圍漂流區的摻雜濃度,以便調整和控制包括電荷平衡、漏-源電阻Rdson以及擊穿電壓在內的裝置性能參數。因此,離子注入提供了一種電荷控制的方法,可以進一步調整和調諧半導體功率裝置的性能,以便用於不同類型的應用。
本發明的另一方面在於,提出了一種新改良過的裝置結構和製作方法,以便在薄的頂部P-外延層上,形成帶有淺溝道柵極的功率電晶體單元,其中薄的頂部P-外延層位於垂直溝道上方的頂面周圍區域上,覆蓋在垂直溝道上方。通過溝道底部摻雜植入以及側壁摻雜植入,就可以靈活地調整裝置溝道的性能。側壁摻雜植入以及溝道底部摻雜植入,用於補償P-外延,並保護適當的積聚以及溝道區域。在用多晶矽柵極層填充柵極溝道之前,要通過柵極溝道的底面進行離子注入。使用垂直注入,形成柵極遮罩摻雜區,以便在電壓擊穿時,遮罩敏感的溝道柵極。
本發明的另一方面在於,提出了一種新改良過的裝置結構和製作方法,以便在一個薄的頂層中,形成帶有較深的溝道柵極的功率電晶體單元,其中薄頂層作為一個外延層,位於外延立柱上方的頂面周圍區域上,覆蓋在外延立柱上方。溝道柵極穿過頂部薄外延層,並延伸至襯底區,這樣一來,就不再需要用於連接聚積區的溝道底部摻雜植入了。穿過柵極溝道底面注入的柵極遮罩摻雜區形成校準的摻雜區,仍然可以遮罩溝道柵極,用於在電壓擊穿時遮罩敏感的溝道柵極。溝道底部摻雜注入仍然可用於確保柵極遮罩摻雜區不接觸柵極溝道。
本發明的一個較佳實施例簡要說明了一種半導體功率裝置,包括一個帶有多個深溝道的半導體襯底。用一個外延層填充深溝道;此外延層還包括一個同時生長的頂部外延層,覆蓋在深溝道頂面上方和半導體襯底上方的區域上。外延層的導電類型與半導體襯底相反。在頂部外延層中,形成多個溝道金屬氧化物半導體場效應管單元,頂部外延層作為本體區,半導體襯底作為漏極區,通過深溝道中的外延層與半導體襯底中的區域之間的電荷平衡,獲得超級結效應。每個溝道金屬氧化物半導體場效應管單元還包括設置在下方的一個溝道柵極和一個柵極遮罩摻雜區,與每一個溝道柵極自校準,並且每一個溝道金屬氧化物半導體場效應管單元都會在電壓擊穿時,遮罩溝道柵極。在一個典型實施例中,溝道金屬氧化物半導體場效應管單元的每個溝道柵極,都通過頂部外延層開口,並用一種柵極介質材料和一種柵極導電材料填充。在另一個典型實施例中,溝道金屬氧化物半導體場效應管單元的每個溝道柵極,都穿過頂部外延層,進入半導體襯底的頂部,半導體襯底中具有一個柵極溝道,其深度大於或等於頂部外延層的厚度,並且用一種柵極介質材料和一種柵極導電材料填充。在另一個典型實施例中,溝道柵極還包括位於溝道柵極側壁周圍的柵極側壁摻雜區,以及柵極溝道下方的柵極-底部摻雜區,其中柵極側壁摻雜區和柵極-底部摻雜區的導電類型與半導體襯底中的導電類型一致。在另一個典型實施例中,半導體襯底還包括深溝道周圍的區域,其摻雜濃度梯度橫向分佈,從周圍區域開始摻雜濃度逐漸降低,在深溝道的側壁附近,濃度迅速降低。在另一個典型實施例中,每個金屬氧化物半導體場效應管電晶體單元,在溝道柵極的側壁以及溝道柵極下方的柵極-底部摻雜區周圍,還帶有柵極側壁摻雜區,其中柵極側壁摻雜區和柵極-底部摻雜區的導電類型都與半導體襯底中的導電類型相同。在另一個典型實施例中,深溝道在半導體襯底的底面附近,漏極接觸摻雜區圍繞在深溝道的底部附近,用於連接漏極電極。在另一個典型實施例中,半導體功率裝置還包括一個底部金屬層,構成一個漏極電極,接觸漏極接頭摻雜區。在另一個典型實施例中,溝道金屬氧化物場效應管單元的溝道柵極和深溝道,都用外延層填充,並且進一步將外延層加工成帶有柵極遮罩摻雜區的條紋,作為浮動摻雜區設置在溝道柵極的條紋下方。在另一個典型實施例中,溝道金屬氧化物場效應管單元的溝道柵極還加工成帶有錯位凸出部的條紋,朝著用外延層填充的深溝道,交替延伸至溝道柵極的對邊上,以便在延伸的溝道柵極下面,通過設置在深溝道中的外延層,將柵極遮罩摻雜區電連接至電晶體單元的本體區。
本發明還提出了一種在半導體襯底上製備半導體功率裝置的方法。該方法包括以下步驟:a)製備半導體襯底; b)在半導體襯底上開通多個深溝道,並生長一個外延層填充深溝道,用頂部外延層覆蓋半導體襯底的頂面,其中外延深溝道中的外延層部分和頂部外延層是同時生長的單層,其中外延層的導電類型與半導體襯底的導電類型相同;c)在頂部外延層中形成多個溝道金屬氧化物半導體場效應管單元,通過開通多個溝道柵極,在溝道柵極下方植入多個柵極遮罩摻雜區,以便在電壓擊穿半導體功率裝置時遮罩電晶體單元的溝道柵極,頂部外延層起本體區的作用,半導體襯底起漏極區的作用,通過深溝道中的外延層部分和半導體襯底中側向深溝道的襯底部分之間的電荷平衡,獲得超級結效應。在一個典型實施例中,本方法還包括通過帶有第一導電類型摻雜物的深溝道側壁植入,在深溝道之間的半導體襯底區中,形成水準濃度梯度,並通過調整深溝道側壁植入,改變半導體功率裝置的性能。在另一個典型實施例中,本方法還包括將一種導電類型與半導體襯底相同的摻雜物,植入到柵極溝道的側壁和底部。在另一個典型實施例中,製備半導體襯底的工序包括製備單層半導體襯底,其中開通多個深溝道的工序包括在單層半導體襯底中開通多個深溝道。在另一個典型實施例中,製備半導體襯底的工序包括製備底部襯底,以及在底部襯底上生長頂部襯底層,頂部襯底層的導電類型與底部襯底的導電類型相同。在另一個典型實施例中,本方法還包括在深溝道的底部進行重摻雜,以便在生長外延層之前,形成漏極接觸區;研磨襯底背部,露出漏極接觸區。在另一個典型實施例中,本方法還包括在形成多個溝道金屬氧化物半導體場效應管單元之前,對外延層的頂面進行部分化學機械拋光,以使其平滑。

One aspect of the present invention is to provide a newly improved device structure and fabrication method for forming a doped column in a drift region to achieve charge balance by a simple and convenient fabrication process. No need for back etching or chemical mechanical polishing, the processing steps are simplified, only a single thin epitaxial layer is formed, and the epitaxial layer is simultaneously grown in the deep trench and over the deep trench, and on the top surface of the region around the deep trench to form a super junction structure. . An epitaxial pillar is formed in the epitaxial layer portion of the channel. A thin top epitaxial layer is formed over the deep trench and over the epitaxial layer portion above the surface of the deep trench region, and a trench metal oxide semiconductor field effect transistor cell is formed in this top epitaxial layer. The two epitaxial layers can be grown simultaneously as a single epitaxial layer. The trench gate of the transistor unit is further masked, and once a voltage breakdown occurs, the doped mask region is implanted through the trench gate into the drift region under the gate to form a self-calibrated doped mask region. Thus masking the sensitive gate solves the above difficulties and limitations. The doped mask region reduces the peak electric field at the gate of the trench; it also slows down the impact ionization rate and increases the breakdown voltage. The final structure improves the reliability and stability of the electrical parameters. The doped mask region is formed below the accumulation region below the trench gate and does not contact the trench gate. There is an additional doped layer under the gate channel, the conductivity type and the conductivity type of the accumulation region are the same, the doped layer can ensure that the doped mask region does not contact the gate channel, so that when the device is turned on The current passed is more.
Another aspect of the present invention is that the super junction structure and shape of the present invention can be used to flexibly adjust the range of breakdown voltages required. The manufacturing process is simple and can be conveniently prepared through standard processes using standard processing modules and equipment. Since the transistor portion of this structure, such as a trench gate double-diffused metal oxide semiconductor (DMOS), is self-calibrating, the fabrication process can be further simplified. The above technical problems and limitations will be solved.
Specifically, one aspect of the present invention is to provide a new and improved device structure and fabrication method for forming an epitaxial layer in a deep trench, and the epitaxial layer has a thin top epitaxial layer portion overlying the device. On the surface. Further, a part of the extension layer also serves as a body region of a metal oxide semiconductor field effect transistor (p-type in the case of an n-channel metal oxide semiconductor field effect transistor). Further, the metal oxide semiconductor field effect transistor unit formed in this top thin epitaxial layer is a channel metal oxide semiconductor field effect transistor. The trench gate is opened through a thin epitaxial layer with an optional trench sidewall and a trench bottom doped implant region to eliminate channels that may be affected by the depth of the trench gate and the doping concentration of the epitaxial layer Sensitivity to performance. A plurality of doped mask regions are implanted into the drift region below the gate through the gate trench prior to filling the gate trench with the gate polysilicon layer. The conductivity type of the doped mask region is the same as that of the body region of the MOSFET, and the doped mask region also functions as a gate mask doping region and self-aligns with the gate channel. The doped mask region can be a floating island or an epitaxial layer that is connected (biased) into the deep trench and thus connected to the body region. In particular, the situation of the floating island is not ideal because the floating traps the charge and causes the device to drift; the trapped charge takes time to diffuse out, which slows down the electrical conversion. The performance of the transistor unit can be controlled and adjusted by a simple and convenient manufacturing process. The super junction structure of the present invention can be applied to a wider range of fields by further improvement.
Another aspect of the present invention is to provide a new and improved apparatus structure and method for forming a transistor unit on a thin top layer, wherein a thin top layer is overlying the deep trench as an epitaxial layer, and around the deep trench and On the top surface above the deep channel. Ion implantation through the deep trench sidewalls (using ions of opposite conductivity types to fill the deep trench epitaxial layer) can adjust the doping concentration of the drift region around the deep trench to adjust and control including charge balance, drain-source resistance Rdson, and blow Device performance parameters including voltage. Thus, ion implantation provides a method of charge control that can further tune and tune the performance of semiconductor power devices for use in different types of applications.
Another aspect of the present invention is to provide a new and improved device structure and fabrication method for forming a power transistor unit with a shallow trench gate on a thin top P- epitaxial layer, wherein the thin top The P- epitaxial layer is located on the area around the top surface above the vertical channel and overlies the vertical channel. By channel doping implantation and sidewall doping implantation, the performance of the device channel can be flexibly adjusted. Sidewall doping implants and channel bottom doping implants are used to compensate for P- epitaxy and to protect proper accumulation and channel regions. Ion implantation is performed through the bottom surface of the gate trench before filling the gate trench with the polysilicon gate layer. A vertical mask is used to form a gate mask doped region to mask the sensitive trench gate during voltage breakdown.
Another aspect of the present invention is to provide a new and improved apparatus structure and method for forming a power transistor unit with a deep trench gate in a thin top layer, wherein the thin top layer serves as a The epitaxial layer is located on the area around the top surface above the epitaxial column and covers the top of the epitaxial column. The trench gate passes through the top thin epitaxial layer and extends to the substrate region, thus eliminating the need for doping of the bottom of the trench for the connection of the accumulation region. The gate mask doped regions implanted through the bottom surface of the gate trench form a calibrated doped region that still masks the trench gate for masking the sensitive trench gate during voltage breakdown. The trench bottom doping implant can still be used to ensure that the gate mask doped regions do not contact the gate trench.
A preferred embodiment of the invention briefly illustrates a semiconductor power device including a semiconductor substrate with a plurality of deep trenches. The deep trench is filled with an epitaxial layer; the epitaxial layer further includes a simultaneously grown top epitaxial layer overlying the top surface of the deep trench and over the semiconductor substrate. The conductivity type of the epitaxial layer is opposite to that of the semiconductor substrate. In the top epitaxial layer, a plurality of trench metal oxide semiconductor field effect transistor cells are formed, the top epitaxial layer is used as a body region, and the semiconductor substrate is used as a drain region, between the epitaxial layer in the deep trench and the region in the semiconductor substrate The charge balance is obtained to obtain a super junction effect. Each trench MOSFET unit further includes a trench gate disposed below and a gate mask doped region, self-aligned with each trench gate, and each channel metal oxide The semiconductor field effect transistor unit masks the trench gate during voltage breakdown. In a typical embodiment, each of the trench gates of the trench MOSFET is opened through the top epitaxial layer and filled with a gate dielectric material and a gate conductive material. In another exemplary embodiment, each trench gate of the trench MOSFET passes through the top epitaxial layer into the top of the semiconductor substrate with a gate trench in the semiconductor substrate The depth is greater than or equal to the thickness of the top epitaxial layer and is filled with a gate dielectric material and a gate conductive material. In another exemplary embodiment, the trench gate further includes a gate sidewall doped region around the trench gate sidewall and a gate-bottom doped region under the gate trench, wherein the gate sidewall doping The conductivity type of the region and the gate-bottom doping region is identical to the conductivity type in the semiconductor substrate. In another exemplary embodiment, the semiconductor substrate further includes a region around the deep trench whose doping concentration gradient is laterally distributed, and the doping concentration gradually decreases from the surrounding region, and the concentration rapidly decreases near the sidewall of the deep trench. In another exemplary embodiment, each MOSFET transistor cell has a gate sidewall around the sidewall of the trench gate and the gate-bottom doping region below the trench gate. The doped region, wherein the gate sidewall doped region and the gate-bottom doped region have the same conductivity type as in the semiconductor substrate. In another exemplary embodiment, the deep trench is near the bottom surface of the semiconductor substrate and the drain contact doped region surrounds the bottom of the deep trench for connecting the drain electrode. In another exemplary embodiment, the semiconductor power device further includes a bottom metal layer that forms a drain electrode that contacts the drain junction doped region. In another exemplary embodiment, the channel gate and deep trench of the trench metal oxide field effect transistor cell are filled with an epitaxial layer and the epitaxial layer is further processed into stripes with a gate mask doped region. As a floating doped region, it is disposed under the stripe of the trench gate. In another exemplary embodiment, the trench gate of the trench metal oxide field effect transistor cell is also processed into stripes with misaligned projections that extend alternately to the trench gate toward deep trenches filled with epitaxial layers. The opposite side of the pole is such that under the extended trench gate, the gate mask doped region is electrically connected to the body region of the transistor unit by an epitaxial layer disposed in the deep trench.
The present invention also provides a method of fabricating a semiconductor power device on a semiconductor substrate. The method comprises the steps of: a) preparing a semiconductor substrate; b) opening a plurality of deep trenches on the semiconductor substrate, and growing an epitaxial layer to fill the deep trenches, and covering the top surface of the semiconductor substrate with the top epitaxial layer, wherein the epitaxial deep trenches The epitaxial layer portion and the top epitaxial layer in the track are single layers grown simultaneously, wherein the epitaxial layer has the same conductivity type as the semiconductor substrate; c) a plurality of channel metal oxide semiconductor field effect transistors are formed in the top epitaxial layer a cell, by opening a plurality of trench gates, implanting a plurality of gate mask doping regions under the trench gates to mask a trench gate of the transistor unit when the voltage breakdowns the semiconductor power device, top The epitaxial layer functions as a body region, the semiconductor substrate functions as a drain region, and a superjunction effect is obtained by charge balance between the epitaxial layer portion in the deep trench and the substrate portion of the lateral deep trench in the semiconductor substrate. In a typical embodiment, the method further includes forming a level concentration gradient in the semiconductor substrate region between the deep trenches by deep channel sidewall implants with dopants of the first conductivity type, and adjusting the deep channel Sidewall implantation changes the performance of semiconductor power devices. In another exemplary embodiment, the method further includes implanting a dopant of the same conductivity type as the semiconductor substrate to the sidewalls and bottom of the gate trench. In another exemplary embodiment, the process of preparing a semiconductor substrate includes preparing a single layer semiconductor substrate, wherein the step of opening a plurality of deep trenches includes opening a plurality of deep trenches in the single layer semiconductor substrate. In another exemplary embodiment, the process of preparing a semiconductor substrate includes preparing a bottom substrate, and growing a top substrate layer on the underlying substrate, the top substrate layer having the same conductivity type as the underlying substrate. In another exemplary embodiment, the method further includes heavily doping at the bottom of the deep trench to form a drain contact region prior to growing the epitaxial layer; grinding the back of the substrate to expose the drain contact region. In another exemplary embodiment, the method further includes partially chemical mechanical polishing the top surface of the epitaxial layer to smooth it prior to forming the plurality of trench metal oxide semiconductor field effect transistor cells.

參見第2圖所示金屬氧化物半導體場效應管裝置100的橫截面視圖,提出了本發明在結構和生產製造方面的新思路。金屬氧化物半導體場效應管裝置100的詳細說明將在下文第3圖仲介紹。金屬氧化物半導體場效應管裝置100位於襯底105上,襯底105中含有一個N+摻雜底部區域120,起漏極接觸區的作用,通過用外延層填充的深溝道130(如第3圖所示,經背部研磨)摻雜。襯底105中還含有一個頂部部分125,深溝道130就形成在頂部部分125中。例如對於一個n-溝道金屬氧化物半導體場效應管,襯底105為n-型,在深溝道中的外延層為p-型。金屬氧化物半導體場效應管電晶體單元位於單一薄外延層上,填充外延立柱溝道130,並覆蓋在P-外延立柱周圍的頂面上,將P-外延填充物填充在立柱溝道中。頂面上方的薄的P-外延層部分也作為本體區,圍繞在用柵極多晶矽填充的溝道柵極145周圍。P-本體區150還圍繞著位於溝道柵極145周圍的源極區155。溝道柵極145用柵極氧化物層140 襯墊,用多晶矽填充,並被帶有接觸開口的絕緣層160覆蓋,以便通過源極接觸金屬連接溝道柵極145之間的源極-本體區域。溝道柵極145被柵極-遮罩摻雜區144遮罩,柵極-遮罩摻雜區144是在用柵極多晶矽填充溝道之前,通過柵極溝道植入的。因此,柵極-遮罩摻雜區144與溝道柵極145自校準。柵極-遮罩區144的導電類型與填充在外延立柱溝道130中的外延層的導電類型相同。
如第2圖所示的裝置帶有單一薄外延層,以便形成溝道柵極,其中溝道柵極的溝道中用柵極多晶矽填充,並通過它形成開口。這種新結構實現了超級結的性能要求,例如不超過“詹森極限”,擊穿電壓不隨生長在起始襯底上的外延層的厚度變化而變化等。絕對擊穿電壓的因素是,溝道在半導體襯底中的深度,以及襯底區之間的外延立柱溝道中的電荷平衡。外延矽生長的厚度僅僅是在矽襯底中刻蝕的深溝道寬度的函數。傳統裝置必須將外延層生長為漂流區,此漂流區的厚度與所需的擊穿電壓成比例,因此傳統裝置並不具備上述柔性。
圖中所示的結構尺寸靈活可變,並且通過簡便的製造方法就可以生產出這種裝置。例如,要製作一個在詹森極限以下、低電阻率、擊穿電壓寬範圍可變(比如200V至900V)的裝置,可以通過生長幾微米的單一外延矽層,刻蝕深度與所需擊穿電壓成比例的單一溝道刻蝕(>200V大約10-15微米,>600V大約40-50微米,>900V大約70-90微米)。此外,裝置位於外延層130頂部上的電晶體部分的結構,是根據溝道柵極雙擴散金屬氧化物半導體裝置而來的,其中裝置結構自校準,製作方法方便、簡單。本裝置的敏感溝道柵極145部分距離溝道130上方的接縫較遠,這也提高了裝置的可靠性,並且省去了不必要的化學機械拋光過程。
參見第3圖,金屬氧化物半導體場效應管裝置100的橫截面視圖,依靠新穎設計的思路以及第2圖所示的基本結構,根據第13A圖至第13N圖所述的工藝製作而成。金屬氧化物半導體場效應管裝置100位於N型襯底上,包括一個N+摻雜底部區120作為漏極接觸區,在底部漏極電極110上方,與其直接接觸。通過含有外延層130的深溝道摻雜漏極接觸區120。用一個P-外延層填充每個深溝道,並覆蓋在溝道周圍和溝道上方的頂面上。金屬氧化物半導體場效應管電晶體單元位於單一薄P-外延層上,單一薄P-外延層填充在外延立柱溝道130中,並覆蓋在P-外延立柱周圍的頂面上。頂面上方的薄P-外延層由溝道柵極145周圍的P-本體區150構成,帶有柵極多晶矽的溝道柵極145填充在溝道中,溝道通過頂部外延層130開口。P-本體區還包圍著溝道柵極145周圍的源極區155。用柵極氧化物層160填充溝道柵極145,並用帶有接頭開口的絕緣層160覆蓋溝道柵極145,以使金屬阻擋層165上方的源極接觸金屬170接觸溝道柵極145之間的源極-本體區。p-型柵極遮罩摻雜區144進一步遮罩溝道柵極145,並在柵極多晶矽填充柵極溝道之前,穿過柵極溝道植入到N-襯底區125中。在金屬氧化物半導體場效應管裝置發生電壓擊穿時,柵極遮罩摻雜區144保護敏感的柵極145。P-外延立柱130周圍的N襯底區125可以用N-摻雜物通過深溝道130的側壁植入,以便獲得水準摻雜濃度梯度,並控制N-立柱電荷。
通過使填充在溝道中P-外延層的電荷在水準方向上平衡,來獲得超級結效應或電荷平衡,即沿垂直於垂直金屬氧化物半導體場效應管結構的n-型漂流區125中的漏極電流流向,獲得電荷平衡,當金屬氧化物半導體場效應管處於截止狀態時,漏極電流耗盡。換言之,填充在溝道中的P-外延層的電量,與N襯底附近的N-漂流區的電量基本相等,在製作公差範圍內。N-漂流區中電量的控制和調節可以通過摻雜N-襯底,或摻雜N-襯底與植入在深溝道側壁中的任何其他N-摻雜離子。對於理想狀況,目標電量是每平方釐米P=N=1E12個原子。在製作過程中,通過植入濃度、植入退火、襯底摻雜濃度、外延摻雜濃度、溝道深度、寬度和形狀、及其它處理工序的參數等對電量控制地越靈活,裝置結構越優化,便於調諧獲得給定擊穿電壓下的較低電阻率。
金屬氧化物半導體場效應管電晶體單元還包括沿柵極側壁的N型摻雜植入區135-S,以及柵極溝道底部下面的N型摻雜植入區135-B。圍繞在柵極145周圍的側壁和底部摻雜植入區,可以用於消除金屬氧化物半導體場效應管裝置溝道,對於溝道深度和P-外延摻雜濃度的敏感性。這種新型結構的實施例是考慮到,要在P-外延層裏形成高性能的金屬氧化物半導體場效應管結構的基礎上提出來的。外延層同最小的或沒有背部刻蝕的P-外延層一同生長。一個金屬氧化物半導體場效應管要工作,必須使源極的導電類型與漏極一致,與本體相反,並有一個聚積區將溝道連接到漏極上。實現了溝道柵極垂直金屬氧化物半導體場效應管結構後,源極位於頂部,溝道沿柵極溝道的側壁,形成在源極下方本體區中。聚積區必須形成在本體區和漏極之間。對於本發明所述的新型的高壓裝置,當生長在N襯底的頂部水準表面上的P-外延很厚時,很難形成高性能的垂直溝道柵極金屬氧化物半導體場效應管。如果P-外延層很厚,柵極溝道為了穿過N-漂流漏極區,就必須很深。深溝道與厚的P本體區相結合,會使溝道變長、溝道電阻增高,最終導致垂直雙擴散金屬氧化物半導體結構的性能降低。因此,在本發明的實施例中,遇到P-外延層的情況時,要在柵極溝道側壁和底部植入額外的摻雜物,使柵極溝道的厚度比一般0.8至1.5微米範圍內的典型的柵極溝道厚度,厚1至3微米。這些額外的摻雜植入物是為了補償柵極溝道附近的聚積區和漏極區中的P-外延區,以便獲得高性能的、短溝道的垂直溝道雙擴散金屬氧化物半導體裝置。因此,在加工金屬氧化物半導體場效應管裝置之前,在柵極溝道中植入額外的傾斜和非傾斜植入物,會使高性能的溝道柵極金屬氧化物半導體場效應管裝置,不再依賴於這些區域中的P-外延層厚度和摻雜濃度。在柵極溝道底部的n-型摻雜植入物135-B也可以用來保護柵極遮罩區144不與柵極溝道145接觸。
應注意的是,第3圖中的實施例表示一個穿過P-外延層的柵極溝道,以及額外的N-型植入物135-S、135-B,可以用於優化金屬氧化物半導體場效應管的性能,而無需完全補償P-摻雜區,即在柵極溝道側壁上的P-外延層。植入物最好是磷和砷或銻。能量應在50KeV至200KeV範圍內。與底部植入物之間的傾斜角應為零度,與側壁植入物之間的傾斜角為+/-5至15度。植入劑量應在1E11至1E13範圍內。額外的P-型本體植入物可用于形成本體區150,並使溝道區保持在沿溝道柵極145側壁的方向上。
第4圖是一個橫截面視圖,表示一種類似於第3圖所示的金屬氧化物半導體場效應管裝置的一個可選實施例,不同之處在於N-襯底區125’的側壁沒有植入N摻雜物,以便通過製作過程實現電荷控制功能。由於假設初始N-襯底的摻雜濃度足夠大,以便與深溝道中生長的P型外延層達到電荷平衡,因此本實施例並不需要將額外的N-摻雜區,引入到深溝道的側壁中。當摻雜濃度的實際值可以達到所需的電荷平衡,即達到N電荷的絕對值=P電荷=1E12個粒子/cm2時,初始N-襯底的摻雜濃度就足夠了。當在所需的公差限制範圍內,襯底濃度可以實現電荷平衡時(例如,當出現N-襯底的摻雜濃度充足的情況的重複性大於+/-10%時),就不一定必須靠摻雜植入物來實現電荷控制。
第5圖是一個橫截面視圖,表示一種類似於第3圖所示的金屬氧化物半導體場效應管裝置的一個可選實施例,不同之處在於金屬氧化物半導體場效應管裝置並不包含側壁,以及第3圖所示的溝道底部摻雜植入區135-B和135-S。當溝道柵極145的深度較大,並在外延層130下方延伸進襯底區125時,就不再需要使用溝道側壁和溝道底部摻雜植入區,來消除溝道對溝道柵極深度的敏感性。
第6圖是一個橫截面視圖,表示一種類似於第3圖所示的金屬氧化物半導體場效應管裝置的一個可選實施例,不同之處在於金屬氧化物半導體場效應管裝置的溝道柵極的深度較淺,小於外延層的深度。金屬氧化物半導體場效應管裝置包括一個柵極溝道側壁和柵極溝道底部摻雜植入區135-S和135-B,分別用於補償P-外延層130,並確保裝置具有適當的聚積區和溝道區。本實施例是基於以下結構,金屬氧化物半導體場效應管裝置具有厚P-外延層或淺柵極溝道,或兼而有之。柵極溝道並沒有到達N漏極區。為了確保電晶體正常、高效的工作,柵極溝道中較低的部分必須作為N摻雜區135-B進行摻雜,以便將沿柵極溝道的側壁,在本體區中形成的有源溝道,與漏極相連接。
傳統晶片都具有重摻雜的襯底,以及輕摻雜的頂層。然而由一個普通晶片製成的如第2圖至第6圖所示的裝置,一開始卻並沒有外延層。這雖然可以節省一大筆晶片成本,但卻多出了通過深溝道和背部研磨晶片,進行底部摻雜的額外工序。另外,第7圖至第8圖所示的裝置使用一個帶有重摻雜N+底部襯底121的傳統晶片,以及生長在N+底部襯底121上方的次重摻雜N-型頂部襯底層126。在一個傳統晶片中,這種N-型頂部襯底層126通常被認為是一個外延層,在本專利中,為了避免產生混淆,將其稱為頂部襯底層。第7圖是一個橫截面視圖,表示一種類似於第3圖所示的金屬氧化物半導體場效應管裝置的一個可選實施例,不同之處在於用外延層填充的深溝道130現在位於頂部襯底層126中,並延伸到重摻雜的底部襯底區121。不再需要,通過一個獨立的摻雜植入過程形成如第3圖所示的獨立漏極接觸區120。相反,在本實施例中,一個重摻雜N+底部襯底區121用作漏極接頭,還有一個N-型頂部襯底層126生長在N+底部襯底區121的頂部。與傳統晶片相比,為了節省成本,頂部襯底區的厚度一般較小。本實施例並不一定要求進行背部研磨。金屬漏極電極110可以形成在重摻雜底部襯底區121下方。
在深溝道底部的漏極接觸摻雜植入過程可省略,因此非常顯著地簡化了製作過程。
第8圖是一個橫截面視圖,表示一種類似於第7圖所示的金屬氧化物半導體場效應管裝置的一個可選實施例,不同之處在用P-外延層填充的深溝道130的厚度小於N+底部襯底121。
第9圖表示本發明半導體功率裝置的條形結構的俯視圖。隨外延層130一同生長的外延深溝道形成一個條形結構。外延深溝道130的輪廓用點劃線表示。電晶體單元所包含的溝道柵極145也形成一個線性條形結構,溝道柵極145由源極區155周圍的柵極氧化物層140填充,並被本體區150包圍。自校準的柵極遮罩摻雜區(圖中沒有明確指出)也作為浮動條紋,形成在溝道柵極145下方。
第10圖表示另一種不同的電晶體單元結構的可選實施例。柵極遮罩P摻雜區144應通過將溝道柵極145作為十字溝道柵極,延伸至如第10圖所示的電晶體單元的某部分中P-立柱130區,連接在本體區150下方的P-摻雜外延立柱130上,而不是在溝道柵極145下方將柵極遮罩摻雜區144加工成浮動區域。第11圖為一種類似的實施例,不同之處在於延伸的溝道柵極145帶有錯位凸出部145-TB,以降低漏-源導通電阻Rdson,改善裝置的可製造性(在填充十字形柵極溝道時,可能會出現空洞問題)。第12圖表示與第11圖相同的結構,解釋說明柵極遮罩摻雜區凸出部144-TB如何在柵極溝道145下方進行自校準,以及如何通過擴散接觸P-摻雜立柱150。在主柵極條紋145下方,和垂直於主柵極條紋的柵極凸出部145-TB下方,植入柵極遮罩摻雜區144。通過插入柵極遮罩摻雜區凸出部144-TB和p-外延立柱130之間的接觸區,P-遮罩摻雜區144電接觸到p-本體150區。錯位結構降低了對溝道寬度的影響。只要電流流經溝道柵極凸出部145-TB的另一側,錯位凸出部還可以獲得更好的分佈電流。
參見第13A圖至第13N圖的一系列側面橫截面視圖,用來說明如第3圖所示的電荷平衡的半導體功率裝置的製作步驟。第13A圖表示初始矽襯底包括一個阻抗約為10ohm/cm N襯底205。襯底205最初並沒有外延層。設置或熱生長厚度約為0.1至1.5微米的一層硬掩膜氧化層212。然後用臨界尺寸在1至5微米範圍內的溝道掩膜(圖中沒有表示出),進行氧化物刻蝕,開通多個溝道刻蝕窗,然後除去光致抗蝕劑。使用矽刻蝕,對於工作電壓約為650伏的裝置,要開通深度約為40至50 微米的深溝道214。根據刻蝕器的類型和刻蝕化學反應,光致抗蝕劑掩膜也可以用於形成刻蝕圖案並開通溝道,而無需使用如圖所示的硬掩膜氧化層212。溝道開口可以在1至5微米範圍內,但大多數裝置應用中都採用3微米比較合適(溝道開口由之前提到的溝道掩膜決定)。然後進行晶片清洗。在第13B圖中,通過氧化物設置或熱生長工藝,形成一個正形投影的氧化層215。如果在底部表面上的氧化層較厚,那麼就採用可選的反應離子刻蝕的各向異性刻蝕,從溝道底部表面上清除氧化物。如果沒有採用可選的反應離子刻蝕工藝,那麼氧化層215的厚度就在0.015至0.1微米之間,如果採用了可選的反應離子刻蝕工藝,那麼氧化層215的厚度就在0.0151至0.4微米之間。為了在深溝道214下方直接形成漏極接觸區220,要進行漏極接觸植入,就是在沿相對於溝道側壁零傾斜角的方向植入N+離子,即垂直植入,植入劑量大於1E15。用磷或砷等N-型離子,植入漏極接觸區220。氧化層215沿側壁方向,保護側壁不受高劑量的漏極接觸植入物的影響。
在第13C圖中,用磷等N-型離子植入溝道側壁,以便設置N區中的摻雜濃度。根據溝道深度,傾斜著旋轉植入,植入劑量為5E11至2E13、傾斜角為5至15度,以便在溝道中形成N區225。在第13D圖中,在很低的氧氣和/或氮氣環境下,1050至1200攝氏度高溫退火30至60分鐘,可以使N+漏極接觸區220擴散,側壁植入N-區225水準擴散。N-區225形成水準N-型濃度梯度,濃度在深溝道側壁附近最大。為了獲得電荷平衡(超級結效應),連同(將要生長的)P-外延層230,可以通過側壁植入,調節襯底205中深溝道旁邊的區域的N-型濃度。也可選擇對於側壁植入,最初用所需的N-型濃度形成襯底205,以獲得超級結效應。在第13E圖中,刻蝕除去氧化層212和215,並生長一個P-外延層230,其中P摻雜濃度為1E15至1E16甚至更高。P-外延層230的厚度足夠填充溝道214。溝道214寬約3微米,在N-區225頂部上方的外延層230的厚度約為1.5至2.0微米。在第13F圖中,厚度約為0.5至1.5微米的氧化層作為硬掩膜層228設置,利用柵極溝道掩膜(圖中沒有表示出),刻蝕硬掩膜氧化層228,然後除去光致抗蝕劑。柵極溝道的寬度一般在0.4至1.5微米的範圍內。利用矽刻蝕的方法通過P-外延層230,刻蝕溝道柵極開口232,溝道深度約為1至2.5微米,可能會穿過P-外延層230,進入設置在溝道212中的外延立柱230之間的N-摻雜區225。晶片清洗,隨後還可進行圓孔刻蝕,以便使柵極溝道結構更加平滑,然後清洗下一個晶片。
在第13G圖中,除去氧化硬掩膜228,然後設置一個薄螢幕層234,覆蓋柵極溝道232的側壁以及底面。深P-型植入硼離子(B11),能量在200至600KeV之間,劑量在1E12至1E13之間,零傾斜角植入,以便在N-摻雜立柱225中的柵極溝道232下方,形成柵極遮罩P-摻雜區244。在第13H圖中,可以選擇N-型柵極溝道側壁植入,傾斜角(植入角)在+/-5至7度之間,用於補償P-外延層230,如果柵極溝道232太淺的話,就用零傾斜角的n-型柵極溝道底部植入,補償P-外延層230,或者確保柵極遮罩P-摻雜區244沒有接觸柵極溝道232。植入物進入柵極溝道側壁和底面,分別形成側壁和底面摻雜區235-S和235-B,消除金屬氧化物半導體場效應管裝置的溝道對於溝道柵極深度以及P-外延層230的摻雜濃度/厚度的敏感性。在第13I圖中,除去螢幕氧化層234,生長一個厚度在0.01至0.1微米之間的柵極氧化層240,具體厚度取決於裝置的額定電壓。在柵極溝道232中設置柵極多晶矽層245。柵極多晶矽層245最好是用原位N+摻雜的方法;如果沒有使用原位摻雜,那麼就通過離子植入或擴散摻雜多晶矽層245。從溝道柵極245周圍的頂面開始,對柵極多晶矽層245進行背部刻蝕。
在第13J圖中,可以使用本體掩膜(圖中沒有表示出),本體植入劑量在3E12至1E14之間的硼,然後在1000至1500攝氏度下進行本體驅動,在溝道柵極245周圍的外延層230中,形成P-本體區250。本體植入可以和本體區之間形成良好的接觸,還可以確保金屬氧化物半導體場效應管溝道區始終位於柵極側壁植入235-S上方。第13K圖表示進行源極摻雜植入。源極植入掩膜(圖中沒有表示出)可以用於保護此位置形成P-本體接觸。用砷離子等源極摻雜離子在能量約為70KeV、劑量約為4E15、零度傾斜角時進行源極植入,然後在800至950攝氏度下,進行源極退火操作,以便擴散源極區255。在第13L圖中,通過低溫氧化物設置(LTO)形成的介質層260和含有硼酸的矽玻璃(BPSG)層260形成在頂面上,然後進行含有硼酸的矽玻璃流水作業。使用接觸掩膜(圖中沒有表示出),進行氧化刻蝕,通過含有硼酸的矽玻璃層260刻蝕出接觸開口。P+本體接觸植入是可選的,然後在本體接觸植入後回流。在第13M圖中,設置勢壘金屬,覆蓋在帶有勢壘金屬層265的頂面,然後設置厚金屬,形成源極金屬層270。金屬掩膜(圖中沒有表示出)用於刻蝕源極金屬260和柵極金屬(圖中沒有表示出)並形成圖案。設置介質層使裝置表面鈍化,鈍化層的圖案用於形成結合區開口(圖中沒有表示出),整個過程就完成了,並且完成了最終的合鑄。為了簡便,這些標準的製作過程就不在此詳述了。在第13N圖中,通過背部研磨,從襯底底面,除去襯底205的低摻雜部分,然後形成背部金屬層210,以便當摻雜濃度較高時,接觸漏極區220。可以通過在晶片背面直接設置TiNiAg層形成背部金屬層210。背部研磨過程的厚度控制可達幾微米甚至是1微米,能夠進行可靠的背部接觸,形成漏極電極層210,以便接觸N+漏極接觸區220。
儘管本發明已經提出了現有的較佳實施例,但這些公開內容並不應成為局限。本領域的技術人員,閱讀上述說明之後,必定可以掌握其他各種變化和修正。例如,儘管上述實施例使用的是n-溝道裝置,但是通過改變半導體區域的導電類型,就可以將本發明應用於p-溝道裝置。因此,所附的申請專利範圍書涵蓋的全部變化和修正都屬於本發明的保護範圍和真實意圖。
儘管本發明的內容已經通過上述優選實施例作了詳細介紹,但應當認識到上述的描述不應被認為是對本發明的限制。在本領域技術人員閱讀了上述內容後,對於本發明的多種修改和替代都將是顯而易見的。因此,本發明的保護範圍應由所附的申請專利範圍來限定。

Referring to the cross-sectional view of the metal oxide semiconductor field effect transistor device 100 shown in Fig. 2, a new idea of the structure and manufacturing of the present invention is proposed. A detailed description of the metal oxide semiconductor field effect transistor device 100 will be described later in FIG. The metal oxide semiconductor field effect transistor device 100 is disposed on a substrate 105 having an N+ doped bottom region 120 therein functioning as a drain contact region through a deep trench 130 filled with an epitaxial layer (as shown in FIG. 3) Shown by back grinding). The substrate 105 also contains a top portion 125 in which the deep trench 130 is formed. For example, for an n-channel MOSFET, the substrate 105 is n-type and the epitaxial layer in the deep trench is p-type. The metal oxide semiconductor field effect transistor transistor unit is located on a single thin epitaxial layer, fills the epitaxial pillar channel 130, and covers the top surface around the P-epitaxial pillar, and fills the P- epitaxial filler in the pillar channel. A thin portion of the P- epitaxial layer above the top surface also serves as a body region surrounding the trench gate 145 filled with gate polysilicon. The P-body region 150 also surrounds the source region 155 located around the trench gate 145. The trench gate 145 is padded with a gate oxide layer 140, filled with a polysilicon, and covered by an insulating layer 160 with a contact opening to connect the source-body between the trench gates 145 through the source contact metal region. The trench gate 145 is masked by a gate-mask doping region 144 that is implanted through the gate trench before filling the trench with a gate polysilicon. Thus, the gate-mask doping region 144 and the trench gate 145 are self-aligned. The conductivity type of the gate-mask region 144 is the same as the conductivity type of the epitaxial layer filled in the epitaxial pillar channel 130.
The device as shown in Figure 2 has a single thin epitaxial layer to form a trench gate in which the channel of the trench gate is filled with gate polysilicon and through which openings are formed. This new structure achieves the performance requirements of the super junction, such as not exceeding the "Jensen limit", the breakdown voltage does not vary with the thickness of the epitaxial layer grown on the starting substrate, and the like. The absolute breakdown voltage is due to the depth of the channel in the semiconductor substrate and the charge balance in the epitaxial column channel between the substrate regions. The thickness of the epitaxial germanium growth is only a function of the deep trench width etched in the germanium substrate. Conventional devices must grow the epitaxial layer as a drift region, the thickness of which is proportional to the required breakdown voltage, so conventional devices do not have the flexibility described above.
The structure shown in the figure is flexible in size and can be produced by a simple manufacturing method. For example, to create a device that is below Jensen's limit, low resistivity, and wide range of breakdown voltages (eg, 200V to 900V), can be etched to a desired depth by a single epitaxial layer of a few microns. A single channel etch with a proportional voltage (>200V approximately 10-15 microns, >600V approximately 40-50 microns, >900V approximately 70-90 microns). In addition, the structure of the transistor portion on the top of the epitaxial layer 130 is based on the channel gate double-diffused metal oxide semiconductor device, wherein the device structure is self-aligned, and the fabrication method is convenient and simple. The sensitive channel gate 145 portion of the device is remote from the seam above the channel 130, which also increases the reliability of the device and eliminates unnecessary chemical mechanical polishing processes.
Referring to Fig. 3, a cross-sectional view of the MOSFET device 100 is fabricated according to the process described in Figs. 13A through 13N, depending on the novel design concept and the basic structure shown in Fig. 2. The MOSFET device 100 is located on an N-type substrate and includes an N+ doped bottom region 120 as a drain contact region over which the top drain electrode 110 is in direct contact. The drain contact region 120 is doped through a deep channel containing epitaxial layer 130. Each deep trench is filled with a P- epitaxial layer and covers the top surface around the trench and above the trench. The metal oxide semiconductor field effect transistor transistor unit is on a single thin P- epitaxial layer, and a single thin P- epitaxial layer is filled in the epitaxial pillar channel 130 and overlying the top surface around the P-extension pillar. A thin P- epitaxial layer above the top surface is formed by a P-body region 150 around the trench gate 145, a trench gate 145 with a gate polysilicon is filled in the trench, and the trench is opened through the top epitaxial layer 130. The P-body region also surrounds the source region 155 around the trench gate 145. The trench gate 145 is filled with a gate oxide layer 160, and the trench gate 145 is covered with an insulating layer 160 having a tab opening such that the source contact metal 170 over the metal barrier layer 165 contacts the trench gate 145. Source-body area. The p-type gate mask doping region 144 further masks the trench gate 145 and is implanted into the N-substrate region 125 through the gate trench before the gate polysilicon fills the gate trench. The gate mask doping region 144 protects the sensitive gate 145 when a voltage breakdown occurs in the MOSFET device. The N substrate region 125 around the P-epitaxial pillar 130 can be implanted through the sidewall of the deep trench 130 with an N-dopant to obtain a level doping concentration gradient and control the N-column charge.
A superjunction effect or charge balance is obtained by balancing the charge of the P- epitaxial layer filled in the channel in the horizontal direction, that is, a leak in the n-type drift region 125 perpendicular to the vertical metal oxide semiconductor field effect transistor structure. The polar current flows to obtain charge balance, and when the metal oxide semiconductor field effect transistor is in an off state, the drain current is depleted. In other words, the amount of charge of the P- epitaxial layer filled in the channel is substantially equal to the amount of charge of the N-drift region near the N substrate, within the manufacturing tolerances. The control and regulation of the charge in the N-drift zone can be by doping the N-substrate, or doping the N-substrate with any other N-doped ions implanted in the deep channel sidewalls. For ideal conditions, the target charge is P = N = 1E12 atoms per square centimeter. In the manufacturing process, the power control is more flexible by the implantation concentration, implantation annealing, substrate doping concentration, epitaxial doping concentration, channel depth, width and shape, and other processing parameters, and the device structure is more flexible. Optimized for easy tuning to achieve lower resistivity at a given breakdown voltage.
The metal oxide semiconductor field effect transistor transistor unit further includes an N-type doped implant region 135-S along the gate sidewall and an N-type doped implant region 135-B under the bottom of the gate trench. The sidewall and bottom doped implant regions surrounding the gate 145 can be used to eliminate metal oxide semiconductor FET device channels, sensitivity to channel depth and P- epitaxial doping concentration. Embodiments of this novel structure are contemplated to be based on the formation of a high performance metal oxide semiconductor field effect transistor structure in a P- epitaxial layer. The epitaxial layer grows with the smallest or no back-etched P- epitaxial layer. To operate a metal-oxide-semiconductor field effect transistor, the source conductivity type must be consistent with the drain, opposite the body, and an accumulation region connects the channel to the drain. After the trench gate vertical metal oxide semiconductor field effect transistor structure is realized, the source is at the top, and the channel is along the sidewall of the gate channel, and is formed in the body region below the source. The accumulation region must be formed between the body region and the drain. For the novel high voltage device of the present invention, when the P-epitaxial growth on the top level surface of the N substrate is thick, it is difficult to form a high performance vertical channel gate metal oxide semiconductor field effect transistor. If the P- epitaxial layer is thick, the gate trench must be deep in order to pass through the N-drift drain region. The combination of the deep channel and the thick P body region causes the channel to become longer and the channel resistance to increase, eventually resulting in a decrease in the performance of the vertical double-diffused metal oxide semiconductor structure. Therefore, in the embodiment of the present invention, in the case of the P- epitaxial layer, additional dopants are implanted in the sidewalls and the bottom of the gate trench so that the thickness of the gate trench is generally 0.8 to 1.5 μm. Typical gate channel thickness in the range is 1 to 3 microns thick. These additional doped implants are used to compensate for the P-epitaxial regions in the accumulation and drain regions near the gate channel in order to obtain a high performance, short channel vertical channel double diffused metal oxide semiconductor device. . Therefore, prior to processing the MOSFET device, implanting additional tilted and non-tilted implants in the gate trench will result in a high performance trench gate MOSFET device, It is again dependent on the P- epitaxial layer thickness and doping concentration in these regions. The n-type dopant implant 135-B at the bottom of the gate trench can also be used to protect the gate mask region 144 from contact with the gate channel 145.
It should be noted that the embodiment in Figure 3 shows a gate channel through the P- epitaxial layer, as well as additional N-type implants 135-S, 135-B, which can be used to optimize metal oxides. The performance of the semiconductor FET without completely compensating for the P-doped region, ie the P- epitaxial layer on the sidewall of the gate trench. The implant is preferably phosphorus and arsenic or antimony. The energy should be in the range of 50 KeV to 200 KeV. The angle of inclination with the bottom implant should be zero degrees and the angle of inclination with the sidewall implants is +/- 5 to 15 degrees. The implant dose should be in the range of 1E11 to 1E13. An additional P-type body implant can be used to form the body region 150 and maintain the channel region in a direction along the sidewalls of the trench gate 145.
Figure 4 is a cross-sectional view showing an alternative embodiment of a metal oxide semiconductor field effect transistor device similar to that shown in Figure 3, except that the sidewalls of the N-substrate region 125' are not implanted. N dopants to achieve charge control functions through the fabrication process. Since it is assumed that the doping concentration of the initial N-substrate is sufficiently large to achieve charge balance with the P-type epitaxial layer grown in the deep channel, this embodiment does not require introduction of an additional N-doped region into the sidewall of the deep trench. in. When the actual value of the doping concentration can reach the desired charge balance, that is, the absolute value of the N charge = P charge = 1E12 particles / cm2, the doping concentration of the initial N-substrate is sufficient. When the substrate concentration can achieve charge balance within the required tolerance limits (for example, when the repeatability of the N-substrate with sufficient doping concentration is greater than +/- 10%), it is not necessary to Charge control is achieved by doping the implant.
Figure 5 is a cross-sectional view showing an alternative embodiment of a MOSFET device similar to that shown in Figure 3, except that the MOSFET device does not include sidewalls. And the channel bottom doping implant regions 135-B and 135-S shown in FIG. When the depth of the trench gate 145 is large and extends below the epitaxial layer 130 into the substrate region 125, it is no longer necessary to use the trench sidewall and the trench bottom doped implant region to eliminate the channel-to-channel. Sensitivity of gate depth.
Figure 6 is a cross-sectional view showing an alternative embodiment of a metal oxide semiconductor field effect transistor device similar to that shown in Figure 3, except that the trench of the metal oxide semiconductor field effect transistor device The depth of the pole is shallower than the depth of the epitaxial layer. The metal oxide semiconductor field effect transistor device includes a gate trench sidewall and a gate trench bottom doped implant region 135-S and 135-B for respectively compensating the P- epitaxial layer 130 and ensuring that the device has appropriate Accumulation zone and channel zone. This embodiment is based on the structure that the metal oxide semiconductor field effect transistor device has a thick P- epitaxial layer or a shallow gate channel, or both. The gate channel does not reach the N drain region. In order to ensure proper and efficient operation of the transistor, the lower portion of the gate trench must be doped as an N-doped region 135-B to form an active trench in the body region along the sidewall of the gate trench. The channel is connected to the drain.
Conventional wafers have heavily doped substrates, as well as lightly doped top layers. However, the device shown in Figures 2 through 6 made of a conventional wafer does not have an epitaxial layer at first. Although this can save a lot of wafer cost, it has an extra process of bottom doping through deep trench and back grinding of the wafer. In addition, the apparatus shown in Figures 7 through 8 uses a conventional wafer with a heavily doped N+ underlying substrate 121, and a sub-heavy doped N-type top substrate layer 126 grown over the N+ underlying substrate 121. . In a conventional wafer, such an N-type top substrate layer 126 is generally considered to be an epitaxial layer, and in this patent, to avoid confusion, it is referred to as a top substrate layer. Figure 7 is a cross-sectional view showing an alternative embodiment of a MOSFET device similar to that shown in Figure 3, except that the deep trench 130 filled with the epitaxial layer is now located at the top lining The bottom layer 126 extends into the heavily doped underlying substrate region 121. It is no longer necessary to form the individual drain contact regions 120 as shown in FIG. 3 by a separate doping implantation process. In contrast, in the present embodiment, one heavily doped N+ underlying substrate region 121 serves as a drain junction, and an N-type top substrate layer 126 is grown on top of the N+ underlying substrate region 121. In order to save cost, the thickness of the top substrate region is generally small compared to conventional wafers. This embodiment does not necessarily require back grinding. Metal drain electrode 110 may be formed under heavily doped underlying substrate region 121.
The drain contact doping implantation process at the bottom of the deep trench can be omitted, thus greatly simplifying the fabrication process.
Figure 8 is a cross-sectional view showing an alternative embodiment of a metal oxide semiconductor field effect transistor device similar to that shown in Figure 7, except that the thickness of the deep trench 130 filled with the P- epitaxial layer Less than the N+ underlying substrate 121.
Fig. 9 is a plan view showing the strip structure of the semiconductor power device of the present invention. The epitaxial deep channel grown with the epitaxial layer 130 forms a strip structure. The outline of the epitaxial deep trench 130 is indicated by a chain line. The channel gate 145 included in the transistor unit also forms a linear strip structure, and the trench gate 145 is filled with the gate oxide layer 140 around the source region 155 and surrounded by the body region 150. A self-calibrated gate mask doped region (not explicitly shown) is also formed as a floating strip under the trench gate 145.
Figure 10 shows an alternative embodiment of another different transistor unit structure. The gate mask P-doped region 144 should be extended to the P-column 130 region of a portion of the transistor unit as shown in FIG. 10 by using the trench gate 145 as a cross-channel gate, connected to the body region. The gate mask doping region 144 is processed into a floating region on the P-doped epitaxial pillar 130 below 150 instead of below the trench gate 145. Figure 11 is a similar embodiment except that the extended trench gate 145 is provided with misalignment projections 145-TB to reduce the drain-source on-resistance Rdson, improving the manufacturability of the device (in filling ten When the glyph gate channel is used, a hole problem may occur). Fig. 12 shows the same structure as Fig. 11, explaining how the gate mask doped region protrusions 144-TB are self-aligned under the gate trenches 145, and how the P-doped pillars 150 are contacted by diffusion. . A gate mask doping region 144 is implanted under the main gate strips 145 and below the gate bumps 145-TB perpendicular to the main gate strips. The P-mask doped region 144 is electrically contacted to the p-body 150 region by interposing a contact region between the gate mask doped region protrusion 144-TB and the p-extension pillar 130. The misalignment structure reduces the effect on the channel width. As long as current flows through the other side of the channel gate projections 145-TB, the misalignment projections can also obtain a better distributed current.
Referring to a series of side cross-sectional views of Figures 13A through 13N, the fabrication steps of the charge balancing semiconductor power device as shown in Figure 3 are illustrated. Figure 13A shows that the initial germanium substrate comprises a substrate 205 having an impedance of about 10 ohm/cm. Substrate 205 does not initially have an epitaxial layer. A hard mask oxide layer 212 having a thickness of about 0.1 to 1.5 microns is disposed or thermally grown. An oxide etch is then performed using a trench mask (not shown) having a critical dimension in the range of 1 to 5 microns, opening a plurality of channel etch windows, and then removing the photoresist. With germanium etching, a deep trench 214 having a depth of about 40 to 50 microns is turned on for a device operating at about 650 volts. Depending on the type of etcher and the etch chemistry, a photoresist mask can also be used to form the etch pattern and turn on the trench without the use of a hard mask oxide layer 212 as shown. The channel opening can be in the range of 1 to 5 microns, but 3 micron is preferred in most device applications (the channel opening is determined by the previously mentioned channel mask). Wafer cleaning is then performed. In Fig. 13B, a positively projected oxide layer 215 is formed by an oxide setting or a thermal growth process. If the oxide layer on the bottom surface is thicker, then an anisotropic etch of the optional reactive ion etch is used to remove oxide from the bottom surface of the trench. If an optional reactive ion etching process is not employed, the thickness of the oxide layer 215 is between 0.015 and 0.1 microns. If an alternative reactive ion etching process is employed, the thickness of the oxide layer 215 is between 0.0151 and 0.4. Between microns. In order to form the drain contact region 220 directly under the deep trench 214, the drain contact implant is performed, that is, the N+ ion is implanted in a direction perpendicular to the tilt angle of the trench sidewall, that is, the vertical implant is implanted at a dose greater than 1E15. . The drain contact region 220 is implanted with an N-type ion such as phosphorus or arsenic. The oxide layer 215 is oriented along the sidewalls to protect the sidewalls from high doses of drain contact implants.
In Fig. 13C, N-type ions such as phosphorus are implanted into the channel sidewalls to set the doping concentration in the N region. Depending on the depth of the channel, the tilt implant is tilted with an implant dose of 5E11 to 2E13 and a tilt angle of 5 to 15 degrees to form an N region 225 in the channel. In Figure 13D, annealing at a high temperature of 1050 to 1200 degrees Celsius for 30 to 60 minutes in a very low oxygen and/or nitrogen atmosphere allows the N+ drain contact region 220 to diffuse and the sidewall implant N-zone 225 level diffusion. The N-zone 225 forms a level N-type concentration gradient with a concentration that is greatest near the sidewalls of the deep trench. To achieve charge balance (super junction effect), along with the P- epitaxial layer 230 (to be grown), the N-type concentration of the region beside the deep channel in the substrate 205 can be adjusted by sidewall implantation. Alternatively, for sidewall implantation, substrate 205 is initially formed with the desired N-type concentration to achieve a superjunction effect. In Fig. 13E, the oxide layers 212 and 215 are etched away, and a P- epitaxial layer 230 is grown in which the P doping concentration is 1E15 to 1E16 or even higher. The thickness of the P- epitaxial layer 230 is sufficient to fill the trench 214. Channel 214 is about 3 microns wide and epitaxial layer 230 over the top of N-region 225 has a thickness of about 1.5 to 2.0 microns. In Fig. 13F, an oxide layer having a thickness of about 0.5 to 1.5 μm is provided as a hard mask layer 228, and a hard mask oxide layer 228 is etched by a gate trench mask (not shown) and then removed. Photoresist. The width of the gate channel is typically in the range of 0.4 to 1.5 microns. The trench gate opening 232 is etched through the P- epitaxial layer 230 by a germanium etch, with a channel depth of about 1 to 2.5 microns, possibly passing through the P- epitaxial layer 230 into the channel 212. An N-doped region 225 between the epitaxial pillars 230. Wafer cleaning followed by a circular hole etch to smooth the gate trench structure and then clean the next wafer.
In Fig. 13G, the oxidized hard mask 228 is removed, and then a thin screen layer 234 is provided covering the sidewalls and bottom surface of the gate trench 232. Deep P-type implanted with boron ion (B11), energy between 200 and 600 KeV, dose between 1E12 and 1E13, implanted at zero tilt angle for under gate channel 232 in N-doped column 225 A gate mask P-doped region 244 is formed. In Figure 13H, an N-type gate channel sidewall implant can be selected with a tilt angle (implant angle) between +/- 5 and 7 degrees to compensate for the P- epitaxial layer 230 if the gate trench If the track 232 is too shallow, the bottom of the n-type gate trench with a zero tilt angle is implanted to compensate for the P- epitaxial layer 230, or to ensure that the gate mask P-doped region 244 does not contact the gate trench 232. The implant enters the sidewalls and the bottom surface of the gate trench to form sidewall and bottom doped regions 235-S and 235-B, respectively, eliminating the channel of the MOSFET device for the channel gate depth and P-epitaxial The sensitivity of the doping concentration/thickness of layer 230. In Fig. 13I, the gate oxide layer 234 is removed, and a gate oxide layer 240 having a thickness of between 0.01 and 0.1 micrometers is grown, the thickness of which depends on the rated voltage of the device. A gate polysilicon layer 245 is disposed in the gate trench 232. The gate polysilicon layer 245 is preferably doped by in-situ N+; if in-situ doping is not used, the polysilicon layer 245 is doped by ion implantation or diffusion. The gate polysilicon layer 245 is back etched from the top surface around the trench gate 245.
In Fig. 13J, a bulk mask (not shown) can be used, the body is implanted with boron at a dose between 3E12 and 1E14, and then bulk driven at 1000 to 1500 degrees Celsius, around the trench gate 245. In the epitaxial layer 230, a P-body region 250 is formed. The body implant can form a good contact with the body region and also ensure that the metal oxide semiconductor field effect transistor channel region is always above the gate sidewall implant 235-S. Figure 13K shows the source doping implant. A source implant mask (not shown) can be used to protect this location from forming a P-body contact. The source is implanted with a source-doped ion such as arsenic ion at a potential of about 70 KeV, a dose of about 4E15, and a zero-degree tilt angle, and then a source annealing operation is performed at 800 to 950 degrees Celsius to diffuse the source region 255. . In Fig. 13L, a dielectric layer 260 formed by a low temperature oxide arrangement (LTO) and a beryllium glass (BPSG) layer 260 containing boric acid are formed on the top surface, and then a bismuth glass flow operation containing boric acid is performed. An etch mask is performed using a contact mask (not shown) to etch the contact opening through a bismuth glass layer 260 containing boric acid. P+ body contact implantation is optional and then reflows after body contact implantation. In Fig. 13M, a barrier metal is provided covering the top surface of the barrier metal layer 265, and then a thick metal is formed to form the source metal layer 270. A metal mask (not shown) is used to etch the source metal 260 and the gate metal (not shown) and form a pattern. The dielectric layer is provided to passivate the surface of the device, and the pattern of the passivation layer is used to form the opening of the bonding region (not shown), the entire process is completed, and the final casting is completed. For the sake of simplicity, the production process of these standards is not detailed here. In Fig. 13N, the low doped portion of the substrate 205 is removed from the bottom surface of the substrate by back grinding, and then the back metal layer 210 is formed to contact the drain region 220 when the doping concentration is high. The back metal layer 210 can be formed by directly providing a TiNiAg layer on the back surface of the wafer. The thickness of the back grinding process can be controlled to a few microns or even 1 micron, enabling reliable back contact to form the drain electrode layer 210 to contact the N+ drain contact region 220.
Although the present invention has been presented in its preferred embodiments, these disclosures are not intended to be limiting. Those skilled in the art will be able to grasp various other changes and modifications after reading the above description. For example, although the above embodiment uses an n-channel device, the present invention can be applied to a p-channel device by changing the conductivity type of the semiconductor region. All changes and modifications that come within the scope of the appended claims are therefore intended to be
Although the present invention has been described in detail by the preferred embodiments thereof, it should be understood that the foregoing description should not be construed as limiting. Various modifications and alterations of the present invention will be apparent to those skilled in the art. Therefore, the scope of the invention should be limited by the scope of the appended claims.

100‧‧‧金屬氧化物半導體場效應管裝置
105‧‧‧襯底
110‧‧‧金屬漏極電極
120‧‧‧N+摻雜底部區域
121‧‧‧N+底部襯底
125‧‧‧頂部部分、N襯底區、n-型漂流區
126‧‧‧N-型頂部襯底層
130‧‧‧深溝道、外延層
135-S‧‧‧N型摻雜植入區
135-B‧‧‧N型摻雜植入區
140‧‧‧柵極氧化物層
144‧‧‧柵極-遮罩摻雜區
145‧‧‧溝道柵極
145-TB、144-TB‧‧‧凸出部
150‧‧‧本體區
155‧‧‧源極區
160‧‧‧絕緣層
165‧‧‧金屬阻擋層
170‧‧‧源極接屬金屬
205‧‧‧襯底
210‧‧‧背部金屬層、漏極電極層
212‧‧‧硬掩膜氧化層
214‧‧‧深溝道
215‧‧‧氧化層
220‧‧‧漏極區、N+漏極接觸區
225‧‧‧N-摻雜區、N-摻雜立柱
228‧‧‧硬掩膜層
230‧‧‧P-外延層
232‧‧‧柵極溝道
234‧‧‧薄螢幕層
235-S‧‧‧側壁
235-B‧‧‧底面摻雜區
240‧‧‧柵極氧化層
244‧‧‧P-摻雜區
245‧‧‧柵極多晶矽層
250‧‧‧P-本體區
255‧‧‧源極區
260‧‧‧介質層、矽玻璃層
270‧‧‧源極金屬層
100‧‧‧Metal Oxide Semiconductor Field Effect Device
105‧‧‧Substrate
110‧‧‧Metal drain electrode
120‧‧‧N+ doped bottom region
121‧‧‧N+ bottom substrate
125‧‧‧Top part, N-substrate area, n-type drifting area
126‧‧‧N-type top substrate layer
130‧‧‧deep channel, epitaxial layer
135-S‧‧‧N type doping implant area
135-B‧‧‧N type doping implanted area
140‧‧‧Gate oxide layer
144‧‧‧Gate-mask doped area
145‧‧‧channel gate
145-TB, 144-TB‧‧‧ bulging
150‧‧‧ body area
155‧‧‧ source area
160‧‧‧Insulation
165‧‧‧Metal barrier
170‧‧‧Source is attached to metal
205‧‧‧Substrate
210‧‧‧Back metal layer, drain electrode layer
212‧‧‧ Hard mask oxide
214‧‧‧deep channel
215‧‧‧Oxide layer
220‧‧‧Drain region, N+ drain contact region
225‧‧‧N-doped region, N-doped column
228‧‧‧hard mask layer
230‧‧‧P-epitaxial layer
232‧‧‧Gate channel
234‧‧‧ Thin screen layer
235-S‧‧‧ side wall
235-B‧‧‧Bottom doped area
240‧‧‧ gate oxide
244‧‧‧P-doped zone
245‧‧‧Gate polysilicon layer
250‧‧‧P-body area
255‧‧‧ source area
260‧‧‧Medium layer, glass layer
270‧‧‧ source metal layer

第1A圖至第1B圖表示通過傳統方法製作的傳統垂直功率裝置結構的橫截面視圖;
第1C-1圖至第1C-3圖表示在沒有與柵極和柵極溝道校準的大塊外延層中形成的浮島的橫截面視圖;
第1D圖表示在連接溝道的柵極溝道下方的摻雜區的橫截面視圖;
第2圖至第8圖為對應本發明的不同實施例,帶有超級結結構的高壓功率裝置的橫截面視圖;
第9圖至第12圖表示用於排列溝道遮罩摻雜區的各種不同佈局結構的俯視圖;
第13A圖至第13N圖表示本發明用於製作高壓功率裝置的加工工序的橫截面視圖,這種高壓功率裝置類似於第3圖所示,帶有超級結結構以及自校準的溝道遮罩摻雜區。

1A to 1B are cross-sectional views showing the structure of a conventional vertical power device fabricated by a conventional method;
1C-1 through 1C-3 illustrate cross-sectional views of floating islands formed in a bulk epitaxial layer not aligned with gate and gate trenches;
1D is a cross-sectional view showing a doped region under the gate channel connecting the trenches;
2 to 8 are cross-sectional views of a high voltage power device with a super junction structure in accordance with various embodiments of the present invention;
9 through 12 are plan views showing various layout structures for arranging channel mask doped regions;
13A through 13N are cross-sectional views showing the processing steps of the present invention for fabricating a high voltage power device similar to that shown in Fig. 3, with a super junction structure and a self-aligning channel mask. Doped area.

100‧‧‧金屬氧化物半導體場效應管裝置 100‧‧‧Metal Oxide Semiconductor Field Effect Device

110‧‧‧金屬漏極電極 110‧‧‧Metal drain electrode

120‧‧‧N+摻雜底部區域 120‧‧‧N+ doped bottom region

125‧‧‧頂部部分、N襯底區、n-型漂流區 125‧‧‧Top part, N-substrate area, n-type drifting area

130‧‧‧深溝道、外延層 130‧‧‧deep channel, epitaxial layer

135-S‧‧‧N型摻雜植入區 135-S‧‧‧N type doping implant area

135-B‧‧‧N型摻雜植入區 135-B‧‧‧N type doping implanted area

140‧‧‧柵極氧化物層 140‧‧‧Gate oxide layer

144‧‧‧柵極-遮罩摻雜區 144‧‧‧Gate-mask doped area

145‧‧‧溝道柵極 145‧‧‧channel gate

150‧‧‧本體區 150‧‧‧ body area

155‧‧‧源極區 155‧‧‧ source area

160‧‧‧絕緣層 160‧‧‧Insulation

165‧‧‧金屬阻擋層 165‧‧‧Metal barrier

170‧‧‧源極接屬金屬 170‧‧‧Source is attached to metal

Claims (29)

一種半導體功率裝置,其特徵在於,包括:
一個含有多個深溝道的半導體襯底;
一個填充在所述的深溝道中的外延層,此外延層包括一個同時生長的頂部外延層,覆蓋所述深溝道頂面上的區域,以及所述的半導體襯底,其中外延層的導電類型與半導體襯底相反;
多個溝道金屬氧化物半導體場效應管單元,設置在所述的頂部外延層中,頂部外延層作為本體區,半導體襯底作為漏極區,通過深溝道中的外延層和旁邊的半導體襯底中的區域之間的電荷平衡,獲得超級結效應;以及
每個所述的多個溝道金屬氧化物半導體場效應管單元還包括一個溝道柵極和一個設置在其下方並與每個溝道金屬氧化物半導體場效應管單元的溝道柵極基本校準的柵極遮罩摻雜區,以便在電壓擊穿時,遮罩溝道柵極,其中柵極遮罩摻雜區的導電類型與襯底相反。
A semiconductor power device, comprising:
a semiconductor substrate having a plurality of deep trenches;
An epitaxial layer filled in said deep trench, said epitaxial layer comprising a simultaneously grown top epitaxial layer covering a region on said top surface of said deep trench, and said semiconductor substrate, wherein said epitaxial layer has a conductivity type and The semiconductor substrate is reversed;
a plurality of channel metal oxide semiconductor field effect transistor cells disposed in said top epitaxial layer, a top epitaxial layer as a body region, a semiconductor substrate as a drain region, an epitaxial layer in a deep trench, and a side semiconductor substrate a charge balance between the regions in the middle, obtaining a superjunction effect; and each of the plurality of channel metal oxide semiconductor FET cells further includes a trench gate and a subsurface disposed therewith and each trench The channel gate of the MOSFET has a substantially calibrated gate mask doped region to mask the trench gate during voltage breakdown, wherein the conductivity type of the gate mask doped region Contrary to the substrate.
如申請專利範圍第1項所述的半導體功率裝置,其特徵在於,
所述柵極遮罩摻雜區設置在距溝道柵極的底面一定距離的地方,並不接觸所述的溝道柵極。
The semiconductor power device according to claim 1, wherein
The gate mask doping region is disposed at a distance from the bottom surface of the trench gate and does not contact the trench gate.
如申請專利範圍第1項所述的半導體功率裝置,其特徵在於,還包括:
所述設置在每個溝道柵極下方,用導電類型與襯底相同的摻雜物植入的柵極底部摻雜區,其位於柵極遮罩摻雜區上方。
The semiconductor power device of claim 1, further comprising:
The gate bottom doped region is disposed under each trench gate and is implanted with the same dopant as the substrate of the conductivity type, which is over the gate mask doped region.
如申請專利範圍第1項所述的半導體功率裝置,其特徵在於,
所述溝道柵極位於頂部外延層內,深溝道之間。
The semiconductor power device according to claim 1, wherein
The trench gate is located within the top epitaxial layer, between the deep trenches.
如申請專利範圍第1項所述的半導體功率裝置,其特徵在於,
每個所述的溝道金屬氧化物半導體場效應管單元的所述的溝道柵極,都延伸穿入所述的頂部外延層,柵極溝道的深度小於或等於所述的頂部外延層的厚度。
The semiconductor power device according to claim 1, wherein
The channel gate of each of the channel metal oxide semiconductor field effect transistor cells extends through the top epitaxial layer, and the depth of the gate channel is less than or equal to the top epitaxial layer thickness of.
如申請專利範圍第1項所述的半導體功率裝置,其特徵在於,
所述每個溝道柵極都延伸並穿透所述的頂部外延層,進入所述的半導體襯底的頂部。
The semiconductor power device according to claim 1, wherein
Each of the channel gates extends through the top epitaxial layer and into the top of the semiconductor substrate.
如申請專利範圍第1項所述的半導體功率裝置,其特徵在於,
所述的溝道柵極還包括圍繞所述的溝道柵極側壁的柵極側壁摻雜區,以及在所述的溝道柵極下方的柵極底部摻雜區,其中柵極側壁摻雜區和柵極底部摻雜區的導電類型與半導體襯底的導電類型相同。
The semiconductor power device according to claim 1, wherein
The trench gate further includes a gate sidewall doped region surrounding the trench gate sidewall, and a gate bottom doped region under the trench gate, wherein the gate sidewall doping The conductivity type of the region and the bottom doped region of the gate is the same as the conductivity type of the semiconductor substrate.
如申請專利範圍第1項所述的半導體功率裝置,其特徵在於,
所述的半導體襯底還包括圍繞所述的深溝道的區域,其有一水準摻雜濃度梯度,濃度從深溝道側壁緊鄰的區域開始逐漸減小。
The semiconductor power device according to claim 1, wherein
The semiconductor substrate further includes a region surrounding the deep trench having a level doping concentration gradient that gradually decreases from a region immediately adjacent to the deep trench sidewall.
如申請專利範圍第2項所述的半導體功率裝置,其特徵在於,
每個所述的金屬氧化物半導體場效應電晶體單元還包括圍繞所述的溝道柵極側壁的柵極側壁摻雜區,以及在所述的溝道柵極下方的柵極底部摻雜區,其中柵極側壁摻雜區和柵極底部摻雜區的導電類型與半導體襯底的導電類型相同。
The semiconductor power device according to claim 2, characterized in that
Each of the MOS field effect transistor units further includes a gate sidewall doped region surrounding the trench gate sidewall, and a gate bottom doped region under the trench gate The conductivity type of the gate sidewall doping region and the gate bottom doping region is the same as the conductivity type of the semiconductor substrate.
如申請專利範圍第1項所述的半導體功率裝置,其特徵在於,還包括:
圍繞所述的深溝道的底部位於所述的半導體襯底的底面附近的一個漏極接觸摻雜區。
The semiconductor power device of claim 1, further comprising:
A drain adjacent the bottom surface of the semiconductor substrate surrounding the bottom of the deep trench contacts the doped region.
如申請專利範圍第1項所述的半導體功率裝置,其特徵在於,
所述柵極遮罩摻雜區構成了浮島。
The semiconductor power device according to claim 1, wherein
The gate mask doped regions constitute a floating island.
如申請專利範圍第1項所述的半導體功率裝置,其特徵在於,
所述柵極遮罩摻雜區電連接到金屬氧化物半導體場效應管單元的本體區上。
The semiconductor power device according to claim 1, wherein
The gate mask doped region is electrically connected to the body region of the metal oxide semiconductor field effect transistor unit.
如申請專利範圍第1項所述的半導體功率裝置,其特徵在於,
所述的溝道金屬氧化物半導體場效應管單元的所述的溝道柵極,以及用所述的外延層填充的所述的深溝道,組成條紋的形式,所述的柵極遮罩摻雜區設置在所述的溝道柵極的條紋下方,作為浮動摻雜區。
The semiconductor power device according to claim 1, wherein
Said channel gate of said channel metal oxide semiconductor field effect transistor unit, and said deep channel filled with said epitaxial layer, in the form of stripes, said gate mask doping The impurity region is disposed under the stripe of the channel gate as a floating doping region.
如申請專利範圍第1項所述的半導體功率裝置,其特徵在於,
所述的溝道金屬氧化物半導體場效應管單元的所述的溝道柵極,組成帶有凸出部的條紋的形式,所述的凸出部朝著用所述的外延層填充的所述的深溝道方向延伸,以便將凸出部溝道柵極下方的所述的柵極遮罩摻雜區,通過填充在所述的深溝道中的所述的外延層,電連接到所述的電晶體單元的本體區上。
The semiconductor power device according to claim 1, wherein
The trench gate of the channel metal oxide semiconductor field effect transistor unit is in the form of a stripe having a protrusion toward the portion filled with the epitaxial layer The deep channel direction extends to electrically connect the gate mask doping region under the bump channel gate to the epitaxial layer filled in the deep trench On the body region of the transistor unit.
如申請專利範圍第1項所述的半導體功率裝置,其特徵在於,
所述的溝道金屬氧化物半導體場效應管單元的所述的溝道柵極,還以帶錯位凸出部的條紋的形式,所述的錯位凸出部在所述的溝道柵極的對邊上,交替朝著用所述的外延層填充的所述的深溝道延伸,以便將溝道柵極凸出部下方的所述的柵極遮罩摻雜區,通過填充在所述深溝道中的所述外延層,電連接至所述的電晶體單元的本體區。
The semiconductor power device according to claim 1, wherein
The channel gate of the channel metal oxide semiconductor field effect transistor unit is also in the form of a stripe with a dislocation protrusion, the dislocation protrusion being at the channel gate On the opposite side, alternately extending toward the deep trench filled with the epitaxial layer to fill the gate mask doped region under the trench gate protrusion by filling in the deep trench The epitaxial layer in the track is electrically connected to the body region of the transistor unit.
如申請專利範圍第1項所述的半導體功率裝置,其特徵在於,
所述半導體襯底還包括一個重摻雜的底部襯底和一個生長在底部襯底上方的輕摻雜的頂部襯底,其中深溝道主要形成在頂部襯底中。
The semiconductor power device according to claim 1, wherein
The semiconductor substrate further includes a heavily doped underlying substrate and a lightly doped top substrate grown over the underlying substrate, wherein the deep trench is formed primarily in the top substrate.
如申請專利範圍第12項所述的半導體功率裝置,其特徵在於,
所述深溝道延伸至底部襯底。
The semiconductor power device according to claim 12, characterized in that
The deep channel extends to the underlying substrate.
如申請專利範圍第12項所述的半導體功率裝置,其特徵在於,
所述深溝道延伸進所述的襯底的頂部,但並沒有觸及所述襯底的底部。
The semiconductor power device according to claim 12, characterized in that
The deep channel extends into the top of the substrate but does not touch the bottom of the substrate.
一種半導體功率裝置,其特徵在於,包括:
一個包含深溝道的半導體襯底;
一個填充深溝道並覆蓋在半導體襯底頂面的單一外延層;以及多個形成在半導體表面上方的外延層頂部中的溝道金屬氧化物半導體場效應管單元,其中深溝道旁邊的的一部分半導體襯底,擔負著溝道金屬氧化物半導體場效應管單元的漂流層的作用,並且其中所述的溝道金屬氧化物半導體場效應管單元的溝道柵極,形成在深溝道之間的漂流區上方的一部分外延層中,並通過漂流區和深溝道中的外延層部分之間的電荷平衡,使半導體功率裝置獲得超級結效應;以及
一個柵極遮罩摻雜區,設置在每個溝道柵極下方,並與每個溝道柵極基本校準,用於當每個溝道金屬氧化物半導體場效應管單元發生電壓擊穿時,遮罩溝道柵極。
A semiconductor power device, comprising:
a semiconductor substrate comprising a deep trench;
a single epitaxial layer filling the deep trench and covering the top surface of the semiconductor substrate; and a plurality of trench metal oxide semiconductor field effect transistor cells formed in the top of the epitaxial layer over the semiconductor surface, wherein a portion of the semiconductor beside the deep trench a substrate, which functions as a drift layer of the channel metal oxide semiconductor field effect transistor unit, and wherein the trench gate of the channel metal oxide semiconductor field effect transistor unit forms a drift between the deep trenches a portion of the epitaxial layer above the region, and through the charge balance between the drift region and the epitaxial layer portion of the deep trench, the semiconductor power device obtains a super junction effect; and a gate mask doped region is disposed in each channel Below the gate, and substantially aligned with each of the trench gates, the trench gate is masked when a voltage breakdown occurs in each of the trench MOSFETs.
如申請專利範圍第19項所述的半導體功率裝置,其特徵在於,
所述柵極遮罩摻雜區設置在距溝道柵極的底面有一定距離的位置上,並沒有接觸所述的溝道柵極。
The semiconductor power device according to claim 19, characterized in that
The gate mask doping region is disposed at a distance from the bottom surface of the trench gate and does not contact the trench gate.
一種在半導體襯底上形成半導體功率裝置的方法,其特徵在於,包括:
製備一個半導體襯底;
在半導體襯底中開通數個深溝道,生長一個頂部外延層,用它填充所述的深溝道,覆蓋所述半導體襯底的頂面,其中深溝道中的一部分外延層和所述的頂部外延層,都作為單層同時生長,其中外延層的導電類型與半導體襯底的導電類型相反;以及
在所述的頂部外延層中,通過開通多個柵極溝道,並在所述的柵極溝道下方植入多個柵極遮罩摻雜區,形成多個溝道金屬氧化物半導體場效應管單元,以便當所述的半導體功率裝置發生電壓擊穿時,遮罩所述的電晶體單元的溝道柵極,頂部外延層擔負本體區的作用,半導體襯底擔負漏極區的作用,其中通過深溝道中的一部分外延層和深溝道旁邊的一部分半導體襯底之間達到電荷平衡,獲得超級結效應。
A method of forming a semiconductor power device on a semiconductor substrate, comprising:
Preparing a semiconductor substrate;
Opening a plurality of deep trenches in the semiconductor substrate, growing a top epitaxial layer, filling the deep trenches therewith, covering a top surface of the semiconductor substrate, wherein a portion of the epitaxial layer in the deep trench and the top epitaxial layer Both are grown simultaneously as a single layer, wherein the conductivity type of the epitaxial layer is opposite to that of the semiconductor substrate; and in the top epitaxial layer, by opening a plurality of gate channels, and in the gate trench A plurality of gate mask doped regions are implanted under the track to form a plurality of channel metal oxide semiconductor field effect transistor cells to mask the transistor unit when voltage breakdown occurs in the semiconductor power device The trench gate, the top epitaxial layer acts as a body region, and the semiconductor substrate acts as a drain region, wherein a charge balance is achieved between a portion of the epitaxial layer in the deep trench and a portion of the semiconductor substrate next to the deep trench to obtain a super Junction effect.
如申請專利範圍第21項所述的在半導體襯底上形成半導體功率裝置的方法,其特徵在於,還包括:
通過深溝道的側壁,植入帶有第一導電類型的摻雜物,以便在所述的半導體襯底中所述的深溝道之間的區域中形成水準濃度梯度,並通過調整深溝道側壁植入,調節所述的半導體功率裝置的所述的裝置性能。
The method for forming a semiconductor power device on a semiconductor substrate according to claim 21, further comprising:
A dopant having a first conductivity type is implanted through a sidewall of the deep trench to form a level concentration gradient in a region between the deep trenches in the semiconductor substrate, and by adjusting a deep trench sidewall implant Into, adjusting the device performance of the semiconductor power device.
如申請專利範圍第21項所述的在半導體襯底上形成半導體功率裝置的方法,其特徵在於,
所述的在所述的柵極溝道下方,植入多個柵極遮罩摻雜區的步驟,還包括在距所述的柵極溝道底面下方一定距離處,植入所述的多個柵極遮罩摻雜區,其中所述的柵極遮罩摻雜區並沒有接觸所述的溝道柵極。
A method of forming a semiconductor power device on a semiconductor substrate as described in claim 21, characterized in that
The step of implanting a plurality of gate mask doping regions under the gate channel further includes implanting the plurality of locations below a bottom surface of the gate channel A gate mask doped region, wherein the gate mask doped region does not contact the trench gate.
如申請專利範圍第21項所述的在半導體襯底上形成半導體功率裝置的方法,其特徵在於,還包括:
通過柵極溝道的底部,植入和襯底導電類型相同的摻雜區。
The method for forming a semiconductor power device on a semiconductor substrate according to claim 21, further comprising:
A doped region of the same conductivity type as the substrate is implanted through the bottom of the gate trench.
如申請專利範圍第21項所述的在半導體襯底上形成半導體功率裝置的方法,其特徵在於,還包括:
將和襯底導電類型相同的摻雜物植入到柵極溝道的側壁和底部。
The method for forming a semiconductor power device on a semiconductor substrate according to claim 21, further comprising:
A dopant of the same conductivity type as the substrate is implanted into the sidewalls and bottom of the gate trench.
如申請專利範圍第21項所述的在半導體襯底上形成半導體功率裝置的方法,其特徵在於,還包括:
所述的製備一個半導體襯底的步驟包括製備一個單層半導體襯底,並且其中所述的開通多個深溝道的步驟包括在單層半導體襯底中開通多個深溝道。
The method for forming a semiconductor power device on a semiconductor substrate according to claim 21, further comprising:
The step of preparing a semiconductor substrate includes preparing a single-layer semiconductor substrate, and wherein the step of opening a plurality of deep trenches includes opening a plurality of deep trenches in the single-layer semiconductor substrate.
如申請專利範圍第21項所述的在半導體襯底上形成半導體功率裝置的方法,其特徵在於,還包括:
所述的製備一個半導體襯底的步驟還包括製備一個重摻雜的底部襯底,並在底部襯底上方生長一個頂部襯底層,其中頂部襯底層的導電類型與底部襯底相同。
The method for forming a semiconductor power device on a semiconductor substrate according to claim 21, further comprising:
The step of preparing a semiconductor substrate further includes preparing a heavily doped underlying substrate and growing a top substrate layer over the underlying substrate, wherein the top substrate layer has the same conductivity type as the bottom substrate.
如申請專利範圍第26項所述的在半導體襯底上形成半導體功率裝置的方法,其特徵在於,還包括:
對深溝道底部進行重摻雜,是為了在生長所述的外延層之前,形成漏極接觸區;並且
對襯底進行背部研磨,使漏極接觸區裸露出來。
The method for forming a semiconductor power device on a semiconductor substrate according to claim 26, further comprising:
The deep trench bottom is heavily doped to form a drain contact region before the epitaxial layer is grown; and the substrate is back ground to expose the drain contact region.
如申請專利範圍第21項所述的在半導體襯底上形成半導體功率裝置的方法,其特徵在於,還包括:
在形成所述的多個溝道金屬氧化物半導體場效應管單元的步驟之前,對外延層的頂面進行部分化學機械拋光,以使頂面平滑。
 
The method for forming a semiconductor power device on a semiconductor substrate according to claim 21, further comprising:
Prior to the step of forming the plurality of trench metal oxide semiconductor field effect transistor cells, a partial chemical mechanical polishing of the top surface of the epitaxial layer is performed to smooth the top surface.
TW99101558A 2009-01-21 2010-01-20 Charged balanced devices with shielded gate trench TWI442567B (en)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
US12/321,435 US7893488B2 (en) 2008-08-20 2009-01-21 Charged balanced devices with shielded gate trench

Publications (2)

Publication Number Publication Date
TW201029182A TW201029182A (en) 2010-08-01
TWI442567B true TWI442567B (en) 2014-06-21

Family

ID=42523256

Family Applications (1)

Application Number Title Priority Date Filing Date
TW99101558A TWI442567B (en) 2009-01-21 2010-01-20 Charged balanced devices with shielded gate trench

Country Status (2)

Country Link
CN (1) CN101783346B (en)
TW (1) TWI442567B (en)

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104241341A (en) * 2012-07-27 2014-12-24 俞国庆 High-frequency low-power dissipation power MOS field-effect tube device
CN103681315B (en) * 2012-09-18 2016-08-10 中芯国际集成电路制造(上海)有限公司 The forming method of buried regions
JP6063280B2 (en) * 2013-02-05 2017-01-18 ルネサスエレクトロニクス株式会社 Semiconductor device
CN104979376B (en) * 2014-04-02 2018-12-21 中芯国际集成电路制造(上海)有限公司 Insulated gate bipolar transistor and forming method thereof
CN105336777A (en) * 2014-07-11 2016-02-17 北大方正集团有限公司 Super-junction MOS device and manufacturing method thereof
CN109509721B (en) * 2017-09-14 2021-05-25 联华电子股份有限公司 Semiconductor element and manufacturing method thereof
CN107731908B (en) * 2017-10-24 2020-09-25 贵州芯长征科技有限公司 Shielding gate MOSFET structure for improving voltage resistance and preparation method thereof
CN113851524B (en) * 2021-09-17 2022-06-14 深圳真茂佳半导体有限公司 Multi-source MOS tube shared grid charge balance chip structure and manufacturing method thereof
CN117476770B (en) * 2023-11-16 2024-07-19 华羿微电子股份有限公司 Low-grid charge shielding grid MOSFET device and manufacturing method thereof

Also Published As

Publication number Publication date
CN101783346B (en) 2012-04-18
TW201029182A (en) 2010-08-01
CN101783346A (en) 2010-07-21

Similar Documents

Publication Publication Date Title
US7893488B2 (en) Charged balanced devices with shielded gate trench
TWI442567B (en) Charged balanced devices with shielded gate trench
US8860130B2 (en) Charged balanced devices with shielded gate trench
US8742401B2 (en) Field effect transistor with gated and non-gated trenches
US9876072B2 (en) Configurations and methods for manufacturing charged balanced devices
TWI466194B (en) Buried field ring field effect transistor (buf-fet) integrated with cells implanted with hole supply path
TWI524521B (en) Nano mosfet and fabricating method thereof with trench bottom oxide shielded and third dimensional p-body contact
US8502312B2 (en) Configurations and methods for manufacturing charge balanced devices
US9356134B2 (en) Charged balanced devices with shielded gate trench
TWI493718B (en) Top drain ldmos﹑semiconductor power device and method of manufacturing the same
US10186573B2 (en) Lateral power MOSFET with non-horizontal RESURF structure