Power silicon carbide-based semiconductor device with improved short circuit capability and method for making such device - Patents.com
Patent Information
- Application Number
- JP2023512345
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2020-08-27
- Filing Date
- 2021-08-16
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2041-08-16
Smart Images

Figure 0007584630000001 
Figure 0007584630000002 
Figure 0007584630000003
Abstract
Description
[Technical field]
[0001] This application claims priority to U.S. Patent Application No. 17 / 004,531, filed August 27, 2020, the entire contents of which are incorporated herein by reference.
[0002] The present invention relates to semiconductor devices, and more particularly to power semiconductor devices. [Background technology]
[0003] A metal oxide semiconductor field effect transistor ("MOSFET") is a well-known type of transistor that can be used as a switching device. A MOSFET includes source and drain regions separated by a channel region, and a gate electrode separated from the channel region by a thin gate insulating layer, such as a silicon oxide layer. A MOSFET can be turned on or off by applying a gate bias voltage to the gate electrode. When a MOSFET is turned on by application of a gate bias voltage equal to or greater than the device's "threshold" voltage, current conducts through the MOSFET's channel region between the source and drain regions. When the bias voltage is removed from the gate electrode (or reduced below the threshold voltage level), current no longer conducts through the channel region. As an example, an n-type MOSFET has n-type source and drain regions and a p-type channel. An n-type MOSFET can be designed to turn on when a gate bias voltage sufficient to create a conductive n-type inversion layer in the p-type channel region that electrically connects the n-type source and drain regions is applied to the gate electrode, thereby allowing majority carrier conduction between the source and drain regions. A MOSFET can also be designed as a "normally on" device, which is turned off by applying a gate bias voltage to the gate electrode.
[0004] Because the gate electrode of a MOSFET is insulated from the channel region by a gate insulating layer, a minimal gate current is required to keep the MOSFET in the on state or to switch it between the on and off states. The gate current during switching is kept small because the gate electrode forms a capacitor with the channel region. Thus, only a small charge and discharge current is required during switching, allowing for a simpler gate drive circuit.
[0005] There is an increasing demand for high power MOSFETs and other gate electrode controlled semiconductor devices that can carry large currents (e.g., hundreds of amperes) in the "on" state and block large voltages (e.g., thousands of volts) in the reverse blocking state. To support high current densities and block such high voltages, power MOSFETs typically have a vertical structure with a source and drain on either side of a thick semiconductor layer structure. For very high power applications, power MOSFETs are typically formed in wide bandgap semiconductor material systems. As used herein, the term "wide bandgap semiconductor" encompasses any semiconductor with a bandgap of at least 1.4 eV, such as, for example, silicon carbide (SiC). Silicon carbide has several advantageous properties including, for example, high electric field breakdown strength, high thermal conductivity, high electron mobility, high melting point, and high saturated electron drift velocity. Compared to devices formed in other semiconductor materials, such as, for example, silicon, electronic devices formed in silicon carbide can have the ability to operate at higher temperatures, higher power densities, higher speeds, higher power levels, and / or higher radiation densities. Summary of the Invention [Means for solving the problem]
[0006] According to some embodiments of the present invention, there is provided a power semiconductor device having a semiconductor layer structure including a silicon carbide drift region having a first conductivity type, a first well located on top of the silicon carbide drift region and doped with a dopant having a second conductivity type different from the first conductivity type, a second well located on top of the silicon carbide drift region and spaced apart from the first well and doped with a dopant having the second conductivity type, and a JFET region in the silicon carbide drift region between the first well and the second well, wherein the doping concentration of the JFET region exceeds the doping concentration of the silicon carbide drift region, and the maximum width of the upper half of the JFET region is at least 30% greater than the minimum width of the lower half of the JFET region.
[0007] In some embodiments, the first well can include a first main well and a first side well between the first main well and the JFET region, and the second well can include a second main well and a second side well between the second main well. The first side well can include a first channel region, and the second side well can include a second channel region.
[0008] In some embodiments, the depth of the first lateral well from the top surface of the semiconductor layer structure is at least 1.0 micron.
[0009] In some embodiments, the power semiconductor device may further include a first source / drain contact on a lower surface of the silicon carbide drift region, a source / drain region having a first conductivity type in an upper portion of the first main well and extending to an upper surface of the first well, a second source / drain contact on an upper surface of the source / drain region, a gate insulating layer on the JFET region and on the first well, and a gate electrode on the gate insulating layer.
[0010] In some embodiments, a depth of the JFET region from a top surface of the semiconductor layer structure may be greater than or equal to a depth of the first lateral well from a top surface of the semiconductor layer structure.
[0011] In some embodiments, the peak doping concentration of the first lateral well is about 5×10 17 / cm 3 and occurs at a depth of 0.8 to 1.2 microns below the top surface of the semiconductor layer structure.
[0012] In some embodiments, the depth of the first lateral well from the top surface of the semiconductor layer structure may be less than 1.6 microns.
[0013] In some embodiments, the depth of the JFET region from the top surface of the semiconductor layer structure may be 1.0 to 1.2 times the depth of the first lateral well from the top surface of the semiconductor layer structure.
[0014] In some embodiments, the peak doping concentration in the top 0.2 microns of the first lateral well may be at least an order of magnitude lower than the peak doping concentration of the first lateral well.
[0015] In some embodiments, the maximum width of the top of the JFET region may be at least 50% greater than the minimum width of the bottom of the JFET region.
[0016] In some embodiments, a peak doping concentration of the first lateral well may occur at a first depth from a top surface of the semiconductor layer structure, and a doping concentration of the JFET region at the first depth from a top surface of the semiconductor layer structure may be at least an order of magnitude lower than the peak doping concentration of the first lateral well.
[0017] In some embodiments, the first lateral well may have a depth of 1.0 to 1.5 microns from the top surface of the semiconductor layer structure, and the JFET region may have a depth of 1.0 to 1.7 microns from the top surface of the semiconductor layer structure, and the JFET region may have a depth at least as deep as the depth of the first lateral well.
[0018] In some embodiments, the distance between the first side well and the second side well may be less than 1.5 microns.
[0019] In some embodiments, the JFET region may include an implanted region implanted with a dopant having a first conductivity type.
[0020] In some embodiments, the maximum width of the top half of the JFET region may be 40% to 80% greater than the minimum width of the bottom half of the JFET region.
[0021] In some embodiments, the width of the bottom half of the JFET region may be less than 50% of the depth of the JFET region.
[0022] In some embodiments, the JFET region may extend below the first well.
[0023] According to a further embodiment of the present invention, there is provided a power semiconductor device having a semiconductor layer structure including a silicon carbide drift region having a first conductivity type, a first well located on top of the silicon carbide drift region and doped with a dopant having a second conductivity type different from the first conductivity type, a second well located on top of the silicon carbide drift region and spaced apart from the first well and doped with a dopant having the second conductivity type, and a JFET region in the silicon carbide drift region between the first well and the second well, wherein a depth of the JFET region from a top surface of the semiconductor layer structure exceeds a depth of the first well from a top surface of the semiconductor layer structure and a maximum width of an upper half of the JFET region is greater than a minimum width of a lower half of the JFET region.
[0024] In some embodiments, the doping concentration of the JFET region may exceed the doping concentration of the silicon carbide drift region, and the depth of the first well from a top surface of the semiconductor layer structure may be at least 1.0 micron, and the peak doping concentration of the first well is at least an order of magnitude greater than the doping concentration of the first well at a depth of 0.2 microns from a top surface of the semiconductor layer structure.
[0025] In some embodiments, the peak doping concentration of the first well is 5×10 17 / cm 3and occurs at a depth of 0.8 to 1.2 microns below the top surface of the semiconductor layer structure.
[0026] In some embodiments, the depth of the first well from the top surface of the semiconductor layer structure may be between 1.0 and 1.6 microns.
[0027] In some embodiments, the depth of the JFET region from the top surface of the semiconductor layer structure may be less than 1.2 times the depth of the first well from the top surface of the semiconductor layer structure.
[0028] In some embodiments, the maximum width of the top half of the JFET region may be at least 50% greater than the minimum width of the bottom half of the JFET region.
[0029] In some embodiments, a peak doping concentration of the first well may occur at a first depth from a top surface of the semiconductor layer structure, and a doping concentration of the JFET region at the first depth from a top surface of the semiconductor layer structure may be at least an order of magnitude lower than the peak doping concentration of the first well.
[0030] In some embodiments, the maximum width of the top half of the JFET region may be 40% to 80% greater than the minimum width of the bottom half of the JFET region.
[0031] In some embodiments, the JFET region may extend below the first well.
[0032] According to yet another embodiment of the present invention, there is provided a power semiconductor device having a semiconductor layer structure including a silicon carbide drift region having a first conductivity type, a first well located on top of the silicon carbide drift region and doped with a dopant having a second conductivity type different from the first conductivity type, a second well located on top of the silicon carbide drift region and spaced apart from the first well and doped with a dopant having the second conductivity type, and a JFET region in the silicon carbide drift region between the first well and the second well. The doping concentration of the JFET region exceeds the doping concentration of the silicon carbide drift region. The peak doping concentration of the first well occurs at a first depth from a top surface of the semiconductor layer structure, and the doping concentration of the JFET region at the first depth is at least an order of magnitude lower than the peak doping concentration of the first well. The depth of the JFET region from a top surface of the semiconductor layer structure exceeds the depth of the first well from a top surface of the semiconductor layer structure.
[0033] In some embodiments, the first peak doping concentration is 5×10 17 / cm 3 and may occur at a depth of 0.7 to 1.2 microns below the top surface of the semiconductor layer structure.
[0034] In some embodiments, the depth of the JFET region from the top surface of the semiconductor layer structure may be less than 1.2 times the depth of the first well from the top surface of the semiconductor layer structure.
[0035] In some embodiments, the peak doping concentration in the top 0.2 microns of the first well may be at least an order of magnitude lower than the peak doping concentration of the first well.
[0036] In some embodiments, the maximum width of the top half of the JFET region may be at least 50% greater than the minimum width of the bottom half of the JFET region.
[0037] In some embodiments, the JFET region may include an implanted region implanted with dopants of the first conductivity type.
[0038] In some embodiments, the JFET region may extend below the first well.
[0039] According to further embodiments of the present invention, methods of forming a power semiconductor device are provided. According to these methods, a silicon carbide drift region having a first conductivity type is provided. An upper portion of the silicon carbide drift region is implanted with a dopant of a second conductivity type to form a first and a second well. An upper portion of the silicon carbide drift region between the first and second wells is implanted with a dopant of the first conductivity type via ion implantation to form a JFET region between the first and second wells, the silicon carbide drift region and the first and second wells being part of a semiconductor layer structure. A minimum width of the upper half of the JFET region is at least 30% greater than a minimum width of a lower half of the JFET region.
[0040] In some embodiments, the first well can include a first main well and a first side well between the first main well and the JFET region, the first side well including the first channel region, the second well can include a second main well and a second side well between the second main well and the JFET region, the second side well including the second channel region, and a depth of the first side well from a top surface of the semiconductor layer structure can be at least 1.0 micron.
[0041] In some embodiments, the peak doping concentration of the first lateral well is about 5×10 17 / cm 3 and may occur at a depth of 0.8 to 1.2 microns below the top surface of the semiconductor layer structure.
[0042] In some embodiments, the depth of the JFET region from the top surface of the semiconductor layer structure may be 1.0 to 1.2 times the depth of the first lateral well from the top surface of the semiconductor layer structure.
[0043] In some embodiments, the peak doping concentration in the top 0.2 microns of the first lateral well may be at least an order of magnitude lower than the peak doping concentration of the first lateral well.
[0044] In some embodiments, the JFET region may be formed prior to the formation of the first and second wells.
[0045] In some embodiments, the JFET region may extend under the first and second wells. [Brief description of the drawings]
[0046] [Figure 1A] FIG. 1 is a schematic cross-sectional view of a unit cell of a conventional power MOSFET. [Figure 1B] FIG. 1B is a graph showing the doping profile of the lateral p-well of the unit cell of FIG. 1A as a function of depth. [Figure 2A] FIG. 1 is a schematic cross-sectional view of a unit cell of another conventional power MOSFET. [Figure 2B] FIG. 2B is a graph showing the doping profile of the lateral p-well of the unit cell of FIG. 2A as a function of depth. [Figure 3A] FIG. 1 is a schematic cross-sectional view of a unit cell of yet another conventional power MOSFET. [Figure 3B] FIG. 3B is a graph showing the doping profile of the lateral p-well of the unit cell of FIG. 3A as a function of depth. [Figure 4] 1 is a schematic plan view of a semiconductor wafer including a plurality of power MOSFETs according to an embodiment of the present invention. [Figure 5A] 5 is a schematic plan view of one of the power MOSFETs included in the semiconductor wafer of FIG. 4. [Figure 5B] FIG. 5B is a schematic plan view of the power MOSFET of FIG. 5A with the source meterization removed. [Figure 6A] FIG. 5C is a schematic plan view of a portion of a unit cell of the power MOSFET of FIGS. 5A and 5B. [Figure 6B] FIG. 6B is a schematic cross-sectional view taken along line 6B-6B of FIG. 6A. [Figure 6C] 7 is a graph illustrating the doping profile of the channel region of the unit cell of FIGS. 6A-6B as a function of depth. [Figure 6D] FIG. 7 is a graph showing the doping profile of the JFET region of the unit cell of FIGS. 6A-6B as a function of depth. [Figure 6E] FIG. 6C is a schematic cross-sectional view of a modified version of the unit cell of FIG. 6B. [Figure 7] FIG. 1 is a circuit diagram showing conditions used in a test of the short-circuit withstand capability of a power MOSFET. [Figure 8] 4 is a graph showing simulated drain current and lattice temperature as a function of time for both a conventional MOSFET and a MOSFET according to an embodiment of the present invention. [Figure 9A] 1A to 1C are schematic cross-sectional views showing a method for manufacturing a power MOSFET according to an embodiment of the present invention. [Figure 9B] 1A to 1C are schematic cross-sectional views showing a method for manufacturing a power MOSFET according to an embodiment of the present invention. [Figure 9C] 1A to 1C are schematic cross-sectional views showing a method for manufacturing a power MOSFET according to an embodiment of the present invention. [Figure 9D] 1A to 1C are schematic cross-sectional views showing a method for manufacturing a power MOSFET according to an embodiment of the present invention. [Figure 9E] 1A to 1C are schematic cross-sectional views showing a method for manufacturing a power MOSFET according to an embodiment of the present invention. [Figure 9F] 1A to 1C are schematic cross-sectional views showing a method for manufacturing a power MOSFET according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0047] Power silicon carbide MOSFETs are used today in applications requiring high voltage blocking, such as blocking several thousand volts. An example is a 10A / cm2 MOSFET that blocks voltages of at least 10KV. 2Silicon carbide MOSFETs are commercially available that are rated for current densities of 1000 MHz or more. To form such devices, multiple "unit cell" MOSFET transistors are typically formed that are electrically connected in parallel. For high power applications, a large number (e.g., hundreds or thousands) of these unit cells are typically provided on a single semiconductor substrate, the top side of which is formed with a gate electrode pattern that acts as the gate electrode for all of the unit cells. The opposite (bottom) side of the semiconductor substrate serves as a common drain for all of the unit cells of the device. Multiple source contacts are formed on source regions of the semiconductor layer structure exposed within the openings in the gate electrode pattern. These source contacts are also electrically connected to each other and serve as a common source. The resulting device has three terminals, a common source terminal, a common drain terminal, and a common gate electrode that serve as terminals for all of the individual unit cell transistors. It will be understood that the above description is for n-type MOSFETs, and that for p-type MOSFETs the locations of the drain and source are reversed.
[0048] There is a constant demand to improve the performance of power MOSFETs and other gate-controlled power semiconductor devices. Some important performance parameters of power MOSFETs are the mobility of carriers in the channel region, the threshold voltage, the on-state resistance, the high frequency switching speed, the device reliability, and the "short circuit capability" of the device.
[0049] Figure 1A is a schematic cross-sectional view of a conventional silicon carbide power MOSFET unit cell transistor 100. (Figure 1A also shows portions of two adjacent unit cells.) Figure 1B is a graph showing the doping profile of the lateral p-well of unit cell transistor 100 as a function of depth from the top surface of the device's semiconductor layer structure.
[0050] As shown in FIG. 1A, the unit cell transistor 100 includes an n-type silicon carbide semiconductor substrate 110. -) silicon carbide drift regions 120 are provided on a substrate 110. The tops of the n-type silicon carbide drift regions 120 may be doped p-type by ion implantation to form p-wells 130. Each p-well 130 includes a main p-well 132 and a pair of side p-wells 134. The main p-wells 132 are 5×10 18 / cm 3 ~5×10 19 / cm 3 The dopant concentration is generally higher the deeper the p-well 132 extends into the silicon carbide drift region 120. The side p-wells 134, which serve as the channel region 136 of the unit cell 100, are more lightly doped with a p-type dopant than the main p-well 132. The doping concentration profile of each side p-well 134 is shown in FIG. 1B (described below). Each p-well 130 is formed by ion implantation using two ion implantation steps.
[0051] Heavily doped (n + A heavily doped (n) n-type silicon carbide source region 140 is formed on top of the p-well 130. The n-type source region 140 may be formed by ion implantation. + ) The n-type silicon carbide region 140 serves as a source region for the unit cell transistor 100. The drift region 120 and the substrate 110 together serve as a common drain region for the unit cell transistor 100. The n-type silicon carbide substrate 110, the n-type silicon carbide drift region 120, the p-well 130, and the n-type source region 140 formed therein may together comprise a semiconductor layer structure 150 for the unit cell transistor 100.
[0052] An n-type epitaxial pattern 160 is formed on the semiconductor layer structure 150. The n-type epitaxial pattern 160 has a low doping concentration (e.g., 1×10 16 / cm 3 ~1×10 17 / cm 3). The n-type epitaxial pattern 160 may be, for example, 0.5 to 1.5 microns thick. A silicon oxide gate insulating layer 170 is formed on the top surface of the n-type epitaxial pattern 160. A gate electrode 172 is formed on the gate insulating layer 170 opposite the n-type epitaxial pattern 160. A source contact 180 is formed in an opening in the n-type epitaxial pattern 160 to contact the heavily doped n-type source region 140. A drain contact 190 is formed on the bottom surface of the substrate 110.
[0053] As mentioned above, the channel region 136 is provided in the lateral p-well 134. The channel region 136 electrically connects the n-type source region 140 to the drift region 120 when a sufficient bias voltage is applied to the gate electrode 172. When such a bias voltage is applied to the gate electrode 172, current can flow from the n-type source region 140, through the channel region 136, and into the portion of the drift region 120 under the gate electrode 172, referred to as the "JFET region" 122 of the device.
[0054] 1B is a graph illustrating the doping profile of the lateral p-well 134 of the unit cell transistor 100 as a function of depth from the top surface of the semiconductor layer structure 150. The p-well 130 is formed by implanting aluminum ions into the silicon carbide drift region 120. As shown in FIG. 1B, the lateral p-well 134 has a doping profile of approximately 1×10 18 / cm 3 and then at a depth of about 0.5 microns from the n-type epitaxial pattern 160, the n-type epitaxial pattern 160 has a doping concentration of approximately 5×10 19 / cm 3 The side p-well 134 may have a graded doping profile where the doping concentration decreases rapidly as depth increases. The side p-well 134 may be implanted using a relatively low implant energy (e.g., 300-350 keV) to provide a relatively high doping concentration at the surface of the side p-well 134.
[0055] As mentioned above, two important performance parameters of a power MOSFET are the mobility of carriers in the channel region of the unit cell transistor and the threshold voltage of the device. The doping profile of the channel region 136 can have a significant effect on both of these performance parameters. To have high channel mobility, a low doping concentration directly under the gate insulating layer 170 may be desirable. Thus, the channel mobility can be improved by providing a relatively lightly doped n-type epitaxial pattern 160 directly under the gate insulating layer 170. Such a design also results in a lower threshold voltage. Furthermore, because the damage of ion implantation is not directly under the gate insulating layer 170, damage caused by such ion implantation is further removed from the gate insulating layer 170 and therefore may have less impact on device performance.
[0056] MOSFETs formed from unit cell transistor 100 can exhibit good performance. However, MOSFETs can be expensive to manufacture because of the need to grow a first set of semiconductor layers, remove the structure from a growth reactor to form p-well 130 and n-type source region 140 via ion implantation, and then return the structure to the growth reactor to form n-type epitaxial pattern 160.
[0057] FIG. 2A is a schematic cross-sectional view of another conventional silicon carbide power MOSFET unit cell 200 (and portions of two additional unit cells). As shown in FIG. 2A, the unit cell transistor 200 includes an n-type silicon carbide semiconductor substrate 210 and a lightly doped n-type (n - The tops of the n-type silicon carbide drift regions 220 are doped p-type by ion implantation to form p-wells 230, each of which includes a main p-well 232 and a pair of side p-wells 234. +2.) An n-type silicon carbide source region 240 is formed on top of the p-well 230 by ion implantation. The substrate 210, drift region 220, p-well 230, and n-source region 240 together may comprise a semiconductor layer structure 250 of the unit cell transistor 200. An insulating pattern 270 is gated on the semiconductor layer structure 250. A gate electrode 272 is formed on the gate insulating layer 270. A source contact 280 is formed on the source region 240, and a drain contact 290 is formed on the underside of the substrate 210. A channel region 236 is provided in the side p-well 234 that electrically connects the source region 240 to the drift region 220 when a sufficient bias voltage is applied to the gate electrode 272.
[0058] 1A, except that the n-type epitaxial pattern 160 of unit cell transistor 100 is omitted in unit cell transistor 200. Additionally, the dimensions and doping concentrations of various regions of unit cell transistor 200 differ from the dimensions and doping concentrations of corresponding regions of unit cell transistor 100.
[0059] 2B is a graph illustrating the doping profile of the lateral p-well 234 of the unit cell transistor 200 as a function of depth from the top surface of the semiconductor layer structure 250. As shown in FIG. 2B, the dopant concentration in the channel region 236 just below the surface of the semiconductor layer structure 250 is approximately 1×10 17 / cm 3 The dopant concentration is about 1×10 at a depth of about 0.5 microns below the surface of the semiconductor layer structure 250. 19 / cm 3 The doping concentration then decreases to a peak doping concentration of 1 x 10 at a depth of approximately 0.7 microns. 17 / cm 3 It drops to less than
[0060] A power MOSFET formed with the unit cell transistor 200 may exhibit good performance, but the performance may not be as good as that of the power MOSFET formed with the unit cell transistor 100 of FIG. 1A. The p-well 230 may be formed via a higher energy ion implantation process (e.g., the implantation energy may be about 450 keV), which may result in increased damage to the semiconductor layer structure 250, which may result in reduced performance of the power MOSFET formed with the unit cell transistor 200. In addition, the damage to the semiconductor lattice may be directly below the gate insulating layer 270, where the damage may have the greatest impact on device performance. In addition, the doping profile of FIG. 2B may be inferior to the doping profile of FIG. 1B, which may result in a corresponding reduction in performance. However, the power MOSFET formed with the unit cell transistor 200 may be significantly less expensive to manufacture compared to the power MOSFET formed with the unit cell transistor 100.
[0061] Another potential problem with a power MOSFET formed with unit cell transistor 200 is that it may exhibit reduced "short circuit capability" compared to a power MOSFET formed with unit cell transistor 100. The "short circuit capability" of a power MOSFET (or other gate-controlled power semiconductor device) refers to the amount of time the device can operate when exposed to a short circuit condition before it is damaged or destroyed. Under short circuit conditions, the drain current increases dramatically. The large drain current causes a spike in the internal temperature of the device due to the large amount of power that is dissipated in the semiconductor layer structure as the large current passes through the device. How quickly the device heats up under short circuit conditions is a function of the thermal dissipation characteristics of the device's packaging and the operating conditions. For example, if a power MOSFET formed with unit cell 200 conducts 500 amps at a voltage of 1200 volts, the power is 1200V x 500A = 60 kilowatts. A power MOSFET with a typical package may have a thermal impedance of, for example, 0.01°C / W. Thus, for such a MOSFET, operation at 60 kilowatts will heat the device to approximately 600°C (60 kilowatts * 0.01°C / W = 600°C). Typically, a MOSFET can only withstand such temperatures without failure for a very short time, e.g., 1 microsecond. In contrast, the same MOSFET might be able to operate at 200°C for 10 hours without failure.
[0062] To protect the MOSFET from such failures, control circuitry may be provided that senses when a short-circuit condition is occurring and, in response, reduces the gate voltage (e.g., to 0 volts). A short-circuit condition is not a normal operating condition and typically occurs because a larger system containing the MOSFET is not operating as intended. However, the short-circuit capability of a MOSFET is important because, if a short-circuit condition does occur, the control system must be able to shut off the gate voltage quickly to prevent failure of the device. The shorter the duration of the short-circuit capability, the faster the control circuitry must be able to act.
[0063] FIG. 3A is a schematic cross-sectional view of a unit cell 300 (and portions of two additional unit cells) of another conventional silicon carbide power MOSFET designed for improved short-circuit performance.
[0064] As shown in FIG. 3A, the unit cell transistor 300 includes an n-type silicon carbide semiconductor substrate 310 and a lightly doped n-type (n - The upper portions of the n-type silicon carbide drift regions 320 are doped p-type by ion implantation to form p-wells 330, each including a main p-well 332 and a pair of side p-wells 334. The upper portions of the p-wells 330 are doped p-type by ion implantation. + 3. A MOSFET 310 is formed on the drift region 320 ...
[0065] The p-well 330 is formed via channeling ion implantation, which allows for the use of relatively low implantation energies while forming a deep p-well 330. By using channeling ion implantation, it is possible to make the top of the semiconductor layer structure 350 only lightly doped p-type, or even leave it n-type. The side well 334 and the main well 332 can have the same or different doping profiles. If different doping profiles are provided, the main well 332 may be more heavily doped than the side well 334. At least the side well 334 is formed using low energy channeling ion implantation techniques. The p-well 330 can have a depth D of 1.5 microns to 6.0 microns or more. The width W of the JFET region 322 is larger than normal due to the increased depth of the p-well 330.
[0066] 3B is a graph showing the (p-type dopant) doping profile of the side well 334 of the unit cell transistor 300 as a function of depth D from the top surface of the semiconductor layer structure 350. As shown in FIG. 3B, the p-type dopant concentration just below the surface of the side p-well 334 is about 1×10 15 / cm 3 The p-type dopant concentration increases rapidly over the first 0.9 microns, reaching approximately 8×10 16 / cm 3 The peak doping concentration is then about 3 x 10 at a depth of about 2.8 microns. 16 / cm 3 The doping concentration then decreases very slowly to a doping concentration of 0.01 μm. The doping concentration then decreases rapidly to below detection levels at a depth of 3.5 microns.
[0067] Although the device of FIG. 3A can exhibit improved shot circuit capability, the device is formed using channeling ion implantation, which requires specialized equipment and increases fabrication time. Furthermore, the maximum doping concentration achievable using channeling ion implantation is approximately 1×10 17 / cm 3 , which may be insufficient to deplete the JFET region 322 of the device under short circuit operating conditions. The MOSFET of FIG. 3A also has a wider JFET region 322, which reduces the number of unit cells that can be formed in a given area of the wafer.
[0068] According to an embodiment of the present invention, power MOSFETs and other gate-controlled devices are provided that can exhibit improved short circuit robustness while still providing very high levels of performance in terms of blocking voltage, switching time, and on-state resistance. This improved performance can be achieved by using high energy ion implantation to form a relatively deep p-well having a highly doped bottom portion in the semiconductor layer structure of the device. Due to the high doping level at the bottom of the p-well, the bottom of the p-well widens laterally, reducing the width of at least a portion of the bottom of the JFET region. The deeper, highly doped p-well and the reduced width of the bottom of the JFET region increase the size and width of the depletion region formed in the JFET region, resulting in a better pinch-off effect and lowering the saturation drain current. The lower saturation current reduces the power loss in the device when operating under short circuit conditions, and therefore reduces the internal temperature rise of the device. Thus, the short circuit robustness of the device can be improved compared to conventional devices.
[0069] The narrowing of the bottom of the JFET region due to the higher doping level at the bottom of the p-well can act to increase the specific on-resistance of the device, which is undesirable. To counteract the increase in on-resistance, the doping concentration of the JFET region can be increased from conventional levels and the depth of the JFET region can be increased. For example, the JFET region can be at least as deep as the deep p-well. This deeper, more highly doped JFET region can mitigate the increase in the specific on-resistance of the device. The deeper, more highly doped JFET region can also reduce the value of the specific on-resistance at high temperatures, which helps to reduce the rate at which the internal temperature of the device increases under short circuit conditions (i.e., the temperature coefficient of the device is reduced). This can further improve the short circuit withstand capability of the device.
[0070] Additionally, narrowing the bottom of the JFET region helps shield the gate electrode from the drain, resulting in a lower gate-drain intrinsic capacitance level for the device. The ratio of gate-source to gate-drain intrinsic capacitance is one of the most important factors for reliable high frequency switching performance, and therefore reducing the gate-drain intrinsic capacitance can improve device performance. Additionally, a deeper, more heavily doped p-well increases the size and width of the depletion region that forms in the JFET region of the device, reducing the electric field values of both the JFET region and the gate insulating layer during operation. Since one known failure mechanism for power MOSFETs and other gate controlled switching devices is the breakdown of the gate insulating layer due to prolonged exposure to high electric fields, improved shielding can improve the lifetime, and therefore the reliability, of devices according to embodiments of the present invention.
[0071] According to some embodiments of the present invention, a power semiconductor device is provided that includes a semiconductor layer structure having a silicon carbide drift region having a first conductivity type (e.g., n-type) and first and second wells located on top of the drift region, the first and second wells being doped with a dopant having a second conductivity type (e.g., p-type). A JFET region is defined in the silicon carbide drift region between the first well and the second well. The first well includes a first main well and a first side well between the first main well and the JFET region, and the second well includes a second main well and a second side well between the second main well and the JFET region. The first and second side wells include respective first and second channel regions. A doping concentration of the JFET region exceeds a doping concentration of the silicon carbide drift region, and a maximum width of an upper portion of the JFET region is greater than a minimum width of a lower portion of the JFET region. In exemplary embodiments, the maximum width of the top of the JFET region may be at least 30%, at least 40%, at least 50%, at least 60%, at least 80%, or at least 100% greater than the minimum width of the bottom of the JFET region. In such embodiments, the top of the JFET region may be the top half of the JFET region and the bottom of the JFET region may be the bottom half of the JFET region. For example, the maximum width of the top half of the JFET region may be between 40% and 80% greater than the minimum width of the bottom half of the JFET region. In each of the above embodiments, the maximum width of the top of the JFET region may be less than three times (i.e., 300%) the minimum width of the bottom of the JFET region.
[0072] In some embodiments, the depth of the JFET region from the top surface of the semiconductor layer structure may be equal to or greater than the depth of the first side well from the top surface of the semiconductor layer structure. The depth of the first side well from the top surface of the semiconductor layer structure may be 1.0 to 1.6 microns in some embodiments. The depth of the JFET region from the top surface of the semiconductor layer structure may be 1.0 to 1.2 times the depth of the first side well. For example, the depth of the first side well may be 1 to 1.5 microns from the top surface of the semiconductor layer structure and the depth of the JFET region may be 1 to 1.7 microns, at least as deep as the depth of the first side well. The JFET region may extend below the side well in some embodiments.
[0073] The peak doping concentration of the first lateral well is 5×10 17 / cm 3 or even 1×10 18 / cm 3 and in some embodiments may occur at a depth of 0.8 to 1.2 microns below the top surface of the semiconductor layer structure. Further, the peak doping concentration in the top 0.2 microns of the first lateral well may be at least an order of magnitude lower than the peak doping concentration of the first lateral well.
[0074] The width of the bottom of the JFET region may be less than the depth of the JFET region. In exemplary embodiments, the width of the bottom of the JFET region may be less than 40%, less than 50%, less than 60%, less than 70%, or less than 80% of the depth of the JFET region. The width of the top of the JFET region may be greater or less than the depth of the JFET region. In exemplary embodiments, the width of the top of the JFET region may be between 70% and 130% of the depth of the JFET region, between 80% and 120% of the depth of the JFET region, or between 90% and 110% of the depth of the JFET region.
[0075] Next, with reference to FIGS. 4 to 9F, illustrative embodiments of power semiconductor devices according to embodiments of the present invention will be described.
[0076] 4 is a schematic plan view of a wafer 400 including a plurality of power MOSFETs 410 according to an embodiment of the invention. The power MOSFETs 410 may be formed in rows and columns and spaced apart from one another so that the wafer 400 can be later singulated (e.g., diced) to separate the individual power MOSFETs 410 for packaging and testing. The wafer 400 may, in some embodiments, include, for example, a 4H silicon carbide substrate with one or more silicon carbide layers formed thereon (e.g., by epitaxial growth). Other semiconductor layers (e.g., polysilicon layers), insulating layers, and / or metal layers may be formed on the silicon carbide semiconductor layer structure to form the power MOSFETs 410.
[0077] Figure 5A is a schematic plan view of one of the power MOSFETs 410 included in wafer 400 of Figure 4. Figure 5B is a schematic plan view of power MOSFET 410 of Figure 5A with source meterization and gate bond pads removed.
[0078] As shown in Figure 5A, a gate bond pad 420 and one or more source bond pads 430-1, 430-2 may be formed on the top surface of the semiconductor layer structure of MOSFET 410. A drain bond pad 440 (represented by a dashed box in Figure 5A) may be provided on the bottom side of MOSFET 410. Each bond pad 420, 430, 440 may be formed of a metal such as aluminum to which bond wires may be readily attached by conventional techniques such as thermocompression or soldering.
[0079] As described in more detail below, a source contact is provided to contact a source region of the semiconductor layer structure of the MOSFET 410. The source contact may be the bottom of a source metal pattern 432 that extends across most of the top surface of the MOSFET 410. The source metal pattern 432 is illustrated by a dashed box in FIG. 5A because a significant portion of the source metal pattern 432 is covered by the protective layer 450. The source bond pads 430-1, 430-2 are the portions of the source metal pattern 432 that are exposed through openings 452 in the protective layer 450. FIG. 5A also shows bond wires 460 that can be used to connect the gate bond pad 420 and the source bond pads 430-1, 430-2 to external circuitry or the like.
[0080] As shown in FIG. 5B, a gate electrode pattern 422 may be provided that includes a gate pad 424, a number of gate fingers 426, and one or more gate buses 428 electrically connecting the gate fingers 426 to the gate pad 424. The gate pad 424 may be directly below and electrically connected to a gate bond pad 420, and the gate fingers 426 may extend horizontally across the entire device. An insulating layer (not shown) may cover the gate fingers 426 and the gate bus 428. A source metal pattern 432 may be formed on the insulating layer above the gate fingers 426. A source contact of the source metal pattern 432 extends downward between the gate fingers 426 through an opening in the insulating layer (not shown) and contacts a corresponding source region of the semiconductor layer structure. The MOSFET 410 includes a number of unit cell transistors 500 arranged in parallel. To provide context, the position of one unit cell 500 is shown in FIG. 5B.
[0081] Figure 6A is a schematic plan view of a unit cell transistor 500 of the power MOSFET 410 of Figures 5A and 5B. Figure 6B is a schematic cross-sectional view taken along line 6B-6B of Figure 6A. It will be understood that Figure 6B shows one complete unit cell 500 and portions of two additional unit cells on either side of it to provide context.
[0082] 6A-6B, the unit cell transistor 500 may be formed on an n-type silicon carbide semiconductor substrate 510, such as a single crystal 4H silicon carbide semiconductor substrate that is heavily doped with n-type impurities. The doping concentration of the substrate 510 may be, for example, 1×10 18 atoms / cm 3 ~1×10 21 atoms / cm 3 Although other doping concentrations may be used. As used herein, the "doping concentration" of a semiconductor material refers to the number of dopant atoms present within one cubic centimeter of the semiconductor material that cause the semiconductor material to have a particular conductivity type (i.e., either n-type or p-type) as measured using standard measurement techniques such as Secondary Ion Mass Spectrometry (SIMS). For n-type semiconductor materials, references to the doping concentration refer to the concentration of n-type dopants, and for p-type semiconductor materials, references to the doping concentration refer to the concentration of p-type dopants. Substrate 510 may be of any suitable thickness (e.g., 100-500 microns thick). Substrate 510 may be partially or completely removed in some embodiments.
[0083] On the substrate 510, a lightly doped n-type (n - ) silicon carbide drift region 520 is provided. The n-type silicon carbide drift region 520 may be formed, for example, by epitaxial growth on the silicon carbide substrate 510. The n-type silicon carbide drift region 520 may have a density of, for example, 1×10 14 ~5×10 16 Dopant / cm 3The doping concentration may vary depending on the voltage blocking rating of the device, with devices with higher voltage blocking ratings typically having a lower doping concentration in the drift region 520. For example, a MOSFET with a voltage blocking rating of 10 KV or greater may have a doping concentration of 1×10 14 ~5×10 14 Dopant / cm 3 while MOSFETs with voltage blocking ratings of 500V or greater may have drift region doping concentrations of 1×10 16 ~5×10 16 Dopant / cm 3 The n-type silicon carbide drift region 520 may have a drift region doping concentration of, for example, about 1.5 to 4.0 times higher than the doping concentration of the remainder of the drift region 520. The n-type silicon carbide drift region 520 may be a thick region having a vertical height above the substrate 510 of, for example, 3 to 100 microns. Although not shown in FIG. 6B, in some embodiments, the top of the n-type silicon carbide drift region 520 may be more highly doped than the remainder of the drift region 520 to provide a current spreading layer at the top of the drift region 520. The doping concentration of this current spreading layer may be, for example, about 1.5 to 4.0 times higher than the doping concentration of the remainder of the drift region 520.
[0084] A p-well 530 is formed on top of the n-type drift region 520 (or the current spreading layer, if provided). A JFET region 522 is defined on top of the drift region 520 between adjacent p-wells 530. The p-wells 530 may include a main well 532 and side wells 534 on either side of the main well 532. The main well 532 and the side wells 534 may be identically doped or differently doped. A channel region 536 of the unit cell 500 is within the side p-well 534. The channel region 536 electrically connects an n-type source region 540 (described below) to the JFET region 522 when a sufficient bias voltage is applied to the gate electrode 572.
[0085] The p-well 530 is formed by ion implantation. As known to those skilled in the art, ions, such as n-type or p-type dopants, can be implanted into a semiconductor layer or region by ionizing a desired ion species and accelerating the ions with a predetermined kinetic energy as an ion beam toward the surface of the semiconductor layer in an ion implantation target chamber. Based on the predetermined kinetic energy, the desired ion species can penetrate the semiconductor layer. The ions are implanted at different depths into the semiconductor layer such that the predetermined kinetic energy provides an implantation "profile" in which the ion concentration varies as a function of depth. The dopant can be, for example, Al. + Or N + ions, but any suitable dopant ions may be used. In some embodiments, the implantation may be performed at different temperatures, such as, for example, at a temperature of 75° C. or higher.
[0086] 6C is a graph showing an exemplary doping profile of the side well 534. As shown in FIG. 6C, the side well 534 has a doping profile of about 1×10 near its surface. 17 / cm 3 6C, the doping concentration is about 5×10 at a depth of about 1 micron below the surface of the p-well 530. 18 / cm 3 The doping concentration drops off sharply at depths below about 1 micron, falling below the background n-type doping concentration level at the top of the drift region 520 at a depth of about 1.4 microns.
[0087] The doping profile shown in FIG. 6C can be achieved by using higher energy ion implants. For example, ion implants may be performed with implant energies between 500 keV and 2 MeV. Using these higher implant energies, the lower portion of side well 534 can be doped more heavily than its upper portion to achieve an implant depth of 1.0 micron or more, and thus the depth of side p-well 534 can be between 0.8 and 2.0 microns in illustrative embodiments, between 1.0 and 1.7 microns in other illustrative embodiments, and between 1.1 and 1.5 microns in still other illustrative embodiments. The peak doping concentration of side well 534 can be between 8×10 in some embodiments. 17 ~1×10 19 Dopant / cm 3 In another exemplary embodiment, the 18 ~8×10 18 Dopant / cm 3 The peak doping concentration may occur at a depth of 0.7 to 1.5 microns below the top surface of semiconductor layer structure 550 in some embodiments, and at a depth of 0.8 to 1.2 microns in other embodiments.
[0088] It will be appreciated that the p-well 530 may be formed using a multi-step ion implantation process. For example, in a first ion implantation step, a relatively low dose of aluminum ions may be implanted at a relatively low implantation energy level. In a second ion implantation step, a medium dose of aluminum ions may be implanted at a medium implantation energy level. In a third ion implantation step, a high dose of aluminum ions may be implanted at a high implantation energy level. In some embodiments, the high dose of aluminum ions may be at least one order of magnitude higher than the low dose of aluminum ions, or at least two orders of magnitude higher than the low dose of aluminum ions. In some embodiments, the high dose of aluminum ions may be at least one order of magnitude higher than the medium dose of aluminum ions. In some embodiments, the high dose of aluminum ions may be implanted using an implantation energy that is at least two times, or at least three times, the implantation energy used to implant the medium dose of aluminum ions. In some embodiments, the high dose of aluminum ions may be implanted using an implant energy that is at least 4 times, at least 5 times, at least 6 times, at least 7 times, or at least 8 times the implant energy used to implant the low dose of aluminum ions. Ions other than aluminum may also be used.
[0089] As best seen in FIG. 6B, the higher doping concentration in the bottom portion of the lateral well 534 causes the lateral p-well 534 to expand laterally. As a result, the top of the JFET region 522 defined between the p-wells 530 is wider than at least a portion of the bottom of the JFET region 522. Here, the top of the JFET region 522 refers to the top half of the JFET region 522, and the bottom of the JFET region 522 refers to the bottom half of the JFET region 522. For example, as shown in FIG. 6B, the top of the JFET region 522 has a first maximum width W 1 Although the width of the top of the JFET region 522 is shown to be constant in FIG. 6B, in an actual device there will typically be some amount of variation in width, with the width W1 is the maximum width of the top of the JFET region 522. The width of the bottom of the JFET region 522 is variable, with the minimum width W 2 is the maximum width W of the top of the JFET region 522 1 In some embodiments, the maximum width W of the top of the JFET region 522 is smaller than 1 In another embodiment, the maximum width W of the top of the JFET region 522 may be 0.8 to 3.0 microns. 1 In yet another embodiment, the maximum width W of the top of the JFET region 522 may be 1.0 to 2.0 microns. 1 In some embodiments, the minimum width W of the bottom of the JFET region 522 may be 1.0 to 1.5 microns. 2 In another embodiment, the minimum width W of the bottom of the JFET region 522 is 0.4 to 1.5 microns. 2 In some embodiments, the width W 2 is the width W 1 In another embodiment, the width W 2 is the width W 1 In some embodiments, the minimum width of the bottom half of the JFET region 522 may be between 35% and 90% of the average width of the top half of the JFET region 522, or between 45% and 65% of the average width of the top half of the JFET region 522.
[0090] As discussed above, the narrowed area at the bottom of the JFET region 522 may act to increase the specific on-resistance of a device including the unit cell 500. To counteract the increase in on-resistance, a deeper JFET region 522 may be provided in the device, the JFET region 522 being at least as deep as the lateral p-well 534. Additionally or alternatively, the doping concentration of the JFET region 522 may be higher than conventional levels. The deeper, more highly doped JFET region 522 may also reduce the value of the specific on-resistance at high temperatures, which helps to reduce the rate at which the internal temperature of the device increases under short circuit conditions (i.e., the temperature coefficient of the device is reduced). This may further improve the short circuit capability of the device. The deeper, more highly doped JFET region 522 may be formed by ion implantation in some embodiments. FIG. 6B illustrates an embodiment in which the depth of the JFET region 522 is approximately the same as the depth of the p-well 530. Figure 6E shows a modified version 500' of unit cell 500 that includes a JFET region 522' having a depth that is about 20% deeper than the depth of p-well 530. The embodiment of Figure 6E may be formed by forming JFET region 522 by ion implantation before forming p-well 530 by ion implantation.
[0091] 6D is a graph illustrating an example doping profile of the JFET region 522. As shown in FIG. 6D, additional n-type dopants (e.g., nitrogen ions) are added at approximately 6×10 16 / cm 3 In the example of FIG. 6D, a total of six ion implantation steps, each with a different implantation energy, were performed to dope the JFET region 522, resulting in the six peaks shown in FIG. 6D. Implanted nitrogen ions tend to be less deflected in the semiconductor lattice than aluminum ions, and therefore in the exemplary embodiment of FIG. 6D, more implantation steps were used to obtain a fairly constant doping profile throughout the JFET region 522. In this example, the background n-type doping concentration in the drift region 520 is approximately 2×10 16 / cm 3 , and thus JFET region 522 has a doping concentration that is 2-3 times higher than the background doping concentration. The doping concentration drops off sharply at depths below about 1.25 microns and is below the background n-type doping concentration level at a depth of about 1.35 microns. As can be seen by comparing FIG. 6C with FIG. 6D, the depth of JFET region 522 is approximately equal to the depth of side p-well 534. The depth of a region formed by ion implantation, such as p-well 530 (or a portion thereof) or JFET region 522, refers to the depth below the surface of that region (or another reference surface) where the doping concentration drops to a level equal to the background doping concentration. The background doping concentration may be of the same conductivity type or a different conductivity type.
[0092] In some embodiments, the depth of the JFET region 522 may be 0.8 to 1.5 times the depth of the lateral p-well 534. In some embodiments, the depth of the JFET region 522 may be 1.0 to 1.4 times the depth of the lateral p-well 534. In yet other embodiments, the depth of the JFET region 522 may be 1.0 to 1.2 times the depth of the lateral p-well 534. In some embodiments, the depth of the JFET region 522 may be 0.8 to 2.0 microns. In other embodiments, the depth of the JFET region 522 may be 1.0 to 1.7 microns.
[0093] The width of the bottom of the JFET region 522 may be less than the depth of the JFET region 522. In an exemplary embodiment, the width of the bottom of the JFET region 522 may be less than 40%, less than 50%, less than 60%, less than 70%, or less than 80% of the depth of the JFET region 522. The width of the top of the JFET region 522 may be greater or less than the depth of the JFET region 522. In an exemplary embodiment, the width of the top of the JFET region 522 may be 70%-130% of the depth of the JFET region 522, 80%-120% of the depth of the JFET region 522, or 90%-110% of the depth of the JFET region 522.
[0094] The main well 532 and the side wells 534 may be formed in the same process and may have the same p-type doping concentration as a function of depth. However, it will be appreciated that in other embodiments, the main well 532 and the side wells 534 may be formed to have different doping profiles. For example, the p-type dopant concentration at the top of the main well 532 may be 5×10 18 / cm 3 ~5×10 19 / cm 3 Additional shallow ion implants may be performed to increase the
[0095] After the p-well 530 is formed, a heavily doped (n + ) An n-type silicon carbide source region 540 may be formed on top of the p-well 530. The n-type source region 540 may be formed by ion implantation. The substrate 510, drift region 520 (including the JFET region 522 and any current spreading layers), p-well 530, and source region 540 together comprise a semiconductor layer structure 550.
[0096] The ion implantation steps used to form the p-well 530, JFET region 522, and source region 540 may be performed in any suitable order. For example, ion implantation into the JFET region 522 may occur before or after the formation of the p-well 530. Typically, the p-well 530 is a heavily doped (n + ) is formed before the n-type silicon carbide source region 540, but this does not have to be the case.
[0097] After the n-type source region 540 is formed, a gate insulating layer 570 may be formed on the top surface of the semiconductor layer structure 550. The gate insulating layer 570 may be formed directly on the semiconductor layer structure 550 without an intervening n-type epitaxial layer. The gate insulating layer 570 may include, for example, a silicon oxide layer, although other insulating materials may be used. A gate electrode 572 is formed on the gate insulating layer 570. The gate electrode 572 may include, for example, conductive gate fingers that function as gate electrodes for the multiple unit cell transistors 500.
[0098] A source contact 580 may be formed on the heavily doped n-type source region 540 and the well region 530. Although not shown to simplify the drawing, the source contact 580 may be part of a continuous source pattern 432 that extends across the entire top surface of the silicon carbide semiconductor layer structure of the MOSFET 410 (see FIG. 5A). The source contact 580 may include, for example, metals such as nickel, titanium, tungsten, and / or aluminum, and / or alloys, and / or thin layer stacks of these and / or similar materials. A drain contact 590 may be formed on the bottom surface of the substrate 510. The drain contact 590 may include, for example, similar materials as the source contact to form an ohmic contact to the silicon carbide substrate.
[0099] The deep, heavily doped p-well 530 and narrowing of at least a portion of the bottom of the JFET region 522 can provide enhanced shielding compared to conventional devices. This enhanced shielding can reduce the electric field values of the JFET region 522 and the gate insulating layer 570 during on-state operation. Because one known failure mechanism for power MOSFETs and other gate-controlled switching devices is the breakdown of the gate insulating layer 570 due to prolonged exposure to high electric fields, the improved shielding can improve the lifetime and therefore the reliability of devices according to embodiments of the invention. In addition, the enhanced shielding helps to shield the gate electrode 572 from the drain contact 590, thereby reducing the intrinsic gate-to-drain capacitance level of the device, which can improve the high frequency switching performance of devices including the unit cell 500.
[0100] By reducing the width of the bottom of the JFET region 522, the size of the depletion region is increased when the device is operated under short circuit conditions, which improves the pinch-off effect and thereby reduces the level of saturated drain current. Thus, the short circuit capability of the device can be improved compared to conventional devices. Increasing the depth and doping concentration of the JFET region 522 may help offset an increase in the specific on-resistance of the device that may result from providing a deeper, more highly doped side p-well 534. The deeper, more highly doped JFET region 522 can also reduce the value of the specific on-resistance at high temperatures, which helps to reduce the rate at which the internal temperature of the device increases under short circuit conditions (i.e., the temperature coefficient of the device is reduced). This can further improve the short circuit capability of the device.
[0101] Unit cell transistor 500 is an n-type device having a source contact 580 on its top surface and a drain contact 590 on its bottom surface, although it will be understood that in a p-type device, these positions would be reversed. Thus, in parts of the following description (including the claims), the source and drain contacts may be referred to generically as "source / drain contacts," which term generally refers to either the source contact or the drain contact.
[0102] In the exemplary unit cell 500 described above, the doping concentration of the JFET region 522 may exceed the doping concentration of the silicon carbide drift region 520. The minimum width W 1 is the minimum width W of the bottom of the JFET region 522 2The depth of the first lateral well 534 may be at least 1.0 micron and the depth of the JFET region 522 may be equal to or greater than the depth of the first lateral well 534 (e.g., the depth of the JFET region 522 may be 1.0 to 1.2 times the depth of the first lateral well 534). For example, the depth of the first lateral well 534 may be 1 to 1.5 microns and the depth of the JFET region 522 may be 1 to 1.7 microns and may be equal to or greater than the depth of the first lateral well 534. 1 is the minimum width W of the bottom of the JFET region 522 in some embodiments. 2 may be at least 50% greater than
[0103] The peak doping concentration of the first lateral well 534 is 5×10 17 / cm 3 and may occur at a depth of 0.8 to 1.2 microns below the top surface of the semiconductor layer structure 550. Additionally, the peak doping concentration in the top 0.2 microns of the first lateral well 534 may be at least an order of magnitude lower than the peak doping concentration of the first lateral well 534. The peak doping concentration of the first lateral well 534 may occur at a first depth from the top surface of the semiconductor layer structure 550, and the doping concentration of the JFET region 522 at the first depth may be at least an order of magnitude lower than the peak doping concentration of the first lateral well 534.
[0104] FIG. 7 is a circuit diagram illustrating a MOSFET in a short circuit condition. As shown in FIG. 7, a power MOSFET may have its source terminal coupled to ground and a high direct current (DC) voltage (VDC) applied to its drain terminal. As an example, the value of VDC may be 800 volts or some other high value. In this condition, when a voltage Vg exceeding the threshold voltage of the device is applied to the gate electrode, the device begins to conduct and the drain current increases rapidly in light of the large voltage difference between the source and drain terminals. As mentioned above, power dissipation in the device structure increases as the drain current increases, rapidly heating the device and typically resulting in permanent device failure if the short circuit condition is not quickly corrected.
[0105] 8 is a graph showing simulated drain current and lattice temperature as a function of time for both a conventional MOSFET (curves 600 and 602) and a MOSFET in accordance with an embodiment of the invention (curves 610 and 612) operating under short circuit conditions. As shown by curve 600, the drain current of the conventional device rises quickly to about 1050 amperes and then decreases as the drain current saturates due to pinch-off effects from depletion in the JFET region. As shown by curve 602, the temperature in the conventional MOSFET rises by 3×10 -6 In contrast, as shown by curve 610, the drain current of a MOSFET according to an embodiment of the present invention only rises to approximately 675 amperes and then decreases as the drain current saturates, and the temperature within the device (curve 612) rises by 3×10 seconds after the short circuit condition occurs. -6 1500K within seconds, indicating that the semiconductor device according to the embodiment of the present invention can exhibit improved short circuit withstand capability.
[0106] 9A to 9F are schematic cross-sectional views showing a method for manufacturing a power MOSFET according to an embodiment of the present invention. As shown in FIG. 9A, a lightly doped n-type (n -) a silicon carbide drift region 520 may be formed. The substrate 510 may be a silicon carbide semiconductor substrate 510, such as a single crystal 4H silicon carbide semiconductor substrate that is heavily doped with n-type impurities. The n-type silicon carbide drift region 520 may be formed by epitaxial growth, e.g., 1×10 14 ~5×10 16 Dopant / cm 3 The top portion (e.g., the top 2-3 microns) of the n-type silicon carbide drift region 520 may be more heavily doped than the remainder of the drift region 520 to form a current spreading layer on top of the drift region 520.
[0107] As shown in FIG. 9B, a p-well 530 is formed on top of the n-type drift region 520 by ion implantation. A mask 521 (e.g., a thick oxide layer) is formed on the top surface of the drift region 520. P-type dopants are then implanted by ion implantation using the mask 521 as an ion implantation mask. A multi-step implant may be used to achieve a desired doping profile. For example, a low energy implant at a low dose (e.g., 1×10 at 50 keV) may be used. 12 ) can be used to form the top of the p-well 530, followed by a low to medium dose medium energy implant (e.g., 2×10 at 100 keV). 12 ) can be used to form the middle (depth) portion of the p-well 530. A high dose, high energy implant (e.g., 1×10 at 500 keV) can be used to form the middle (depth) portion of the p-well 530. 14 ) can be used to form the lower portion of p-well 530. As shown in FIG. 9B, scattering effects cause p-well 530 to extend laterally beyond the opening formed in mask 521. The more heavily doped lower portion of p-well 530 can extend laterally farther than the more lightly doped middle and upper regions of p-well 530. Although not shown in the drawings, mask 521 may be removed and replaced with a second mask having a narrower opening, and an additional low energy implant at a high dose (e.g., 1×10 at 50 keV) can be performed. 14) can be used to significantly increase the doping concentration at the top of the main p-well 532 without affecting the doping concentration of the side p-well 534.
[0108] As shown in FIG. 9C, the ion implantation mask used to form the p-wells 530 is removed and replaced with a mask 523 that exposes the portions of the drift region 520 that are between the p-wells 530. This region is then implanted with n-type dopants using mask 523 as an ion implantation mask. Again, a multi-step ion implantation process (e.g., 3-10 steps) may be used. The ion implantation steps may use different implant energies but relatively similar doses to provide implants that increase the n-type dopant concentration in the portions of the drift region 520 that are between the p-wells 530 to form the JFET region 522. As discussed above, the depth of the JFET region 522 may be greater than or equal to the depth of the p-wells 530.
[0109] As shown in Figure 9D, the ion implantation mask used to form the JFET region 522 is removed and replaced with a mask 541 that exposes selected portions of the main p-well 532. These regions are then implanted with n-type dopants using mask 541 as an ion implantation mask. Finally, the device can be completed by forming a gate insulation layer 570, a gate electrode 572, and source and drain contacts 580 and 590, as shown in Figures 9E and 9F.
[0110] While the above discussion has focused on n-channel MOSFETs, it will be appreciated that in accordance with further embodiments of the present invention, the polarity of each semiconductor layer within each device may be reversed to provide a corresponding p-channel MOSFET. Similarly, while the embodiments of the present invention described above are MOSFETs, it will be appreciated that the techniques disclosed herein may also be used to form insulated gate bipolar junction transistors (IGBTs) that include MOSFETs in accordance with embodiments of the present invention, or other gate-controlled power semiconductor devices.
[0111] In the embodiments described above, the JFET region 522 is formed by ion implantation, but it will be appreciated that in other embodiments, the JFET region 522 may be formed by more heavily doping the top of the drift region 520, which may eliminate the need for an extra ion implantation step. However, if the more heavily doped JFET region 522 is formed during growth of the drift region 520, care may need to be taken to ensure that high electric fields at the edge terminations of the device do not degrade the blocking characteristics of the device.
[0112] Embodiments of the invention are described herein with reference to cross-sectional views showing one or two unit cells of a power switching device. It will be understood that actual implementations will typically include a much larger number of unit cells. However, it will be understood that the invention is not limited to such devices, and that the claims appended hereto encompass, for example, MOSFETs and other power switching devices including a single unit cell. Furthermore, while the present disclosure focuses on silicon carbide devices, it will be understood that embodiments of the invention may also be applicable to devices formed using other wide bandgap semiconductors, such as, for example, gallium nitride, zinc selenide, or any other II-VI or III-V wide bandgap compound semiconductors.
[0113] The present invention has been described above with reference to the accompanying drawings, in which examples of the invention are shown. However, the present invention may be embodied in many different forms and should not be construed as being limited to the examples set forth herein, but rather, these examples are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. In the drawings, the sizes and relative sizes of layers and regions may be exaggerated for clarity. When an element or layer is referred to as being "on," "connected to," or "coupled to" another element or layer, it will be understood that an element or layer may be directly on, connected to, or coupled to the other element or layer, or that there may be intervening elements or layers. In contrast, when an element is referred to as being "directly on," "directly connected to," or "directly coupled to" another element or layer, there are no intervening elements or layers. As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items. Like numbers refer to like elements throughout.
[0114] Although the terms first and second are used herein to describe various regions, layers, and / or elements, it will be understood that these regions, layers, and / or elements should not be limited by these terms. These terms are used only to distinguish one region, layer, or element from another region, layer, or element. Thus, without departing from the scope of the present invention, a first region, layer, or element described below can be referred to as a second region, layer, or element, and similarly, a second region, layer, or element can be referred to as a first region, layer, or element.
[0115] Relative terms such as "lower" or "bottom" and "upper" or "top" may be used herein to describe the relationship of one element to another element as depicted in the figures. It will be understood that the relative terms are intended to encompass different orientations of the device in addition to the orientation depicted in the figures. For example, if the device in the figures is flipped over, an element described as being on the "lower" side of the other element would be oriented on the "upper" side of the other element. Thus, the exemplary term "lower" can encompass both an orientation of "lower" and "upper", depending on the particular orientation of the figure. Similarly, if the device in one of the figures is flipped over, an element described as being "lower" or "directly below" the other element would be oriented "above" the other element. Thus, the exemplary term "lower" or "directly below" can encompass both an orientation of upper and lower.
[0116] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises", "comprising", "includes" and / or "including", as used herein, specify the presence of stated features, elements and / or components, but do not exclude the presence or addition of one or more other features, elements, components and / or groups thereof.
[0117] The embodiments of the invention are described herein with reference to cross-sectional views, which are schematic illustrations. As such, variations from the shapes of the illustrations are expected as a result of, for example, manufacturing techniques and / or tolerances. Thus, the embodiments of the invention should not be construed as limited to the particular shapes of regions shown herein, but should include, for example, deviations in shapes due to manufacturing. For example, implanted regions shown as rectangles typically have rounded or curved features and / or implant concentration gradients at the edges of the implanted region, rather than a binary transition from implanted to non-implanted regions. Thus, the regions shown in the figures are schematic in nature, and their shapes are not intended to represent the actual shapes of regions of a device, and are not intended to limit the scope of the invention.
[0118] It will be understood that the embodiments disclosed herein may be combined, such that features depicted and / or described with respect to a first embodiment may also be included in a second embodiment, and vice versa.
[0119] Although the above embodiments have been described with reference to certain figures, it should be understood that some embodiments of the invention may include additional and / or intervening layers, structures, or elements, and / or certain layers, structures, or elements may be omitted. Although several exemplary embodiments of the invention have been described, those skilled in the art will readily appreciate that many modifications are possible in the exemplary embodiments without substantially departing from the novel teachings and advantages of the invention. Accordingly, all such modifications are intended to be included within the scope of the invention as defined in the claims. Accordingly, the above is illustrative of the invention and should not be construed as limited to the specific embodiments disclosed, and it should be understood that modifications to the disclosed embodiments, as well as other embodiments, are intended to be included within the scope of the appended claims. The invention is defined by the following claims, including equivalents of the claims.
Claims
1. a silicon carbide drift region having a first conductivity type; a first well located over the silicon carbide drift region, the first well doped with a dopant having a second conductivity type different from the first conductivity type; a second well located in the upper portion of the silicon carbide drift region, the second well being spaced from the first well and doped with a dopant having the second conductivity type; a JFET region in the silicon carbide drift region between the first well and the second well; Including, a doping concentration of the JFET region exceeds a doping concentration of the silicon carbide drift region; a maximum width of an upper half of the JFET region is at least 30% greater than a minimum width of a lower half of the JFET region; the minimum width of the bottom half of the JFET region is less than 50% of the depth of the JFET region; Semiconductor layer structure, A power semiconductor device comprising:
2. the first well includes a first main well and a first side well between the first main well and the JFET region, the first side well including a first channel region; the second well includes a second main well and a second side well between the second main well and the JFET region, the second side well including a second channel region; 2. The power semiconductor device according to claim 1.
3. 3. The power semiconductor device of claim 2, wherein said depth of said JFET region from a top surface of said semiconductor layer structure is equal to or greater than a depth of said first lateral well from said top surface of said semiconductor layer structure.
4. 4. The power semiconductor device according to claim 2, wherein the peak doping concentration of the first lateral well exceeds 5×10 / cm and occurs at a depth of 0.8 to 1.2 microns below a top surface of the semiconductor layer structure.
5. 5. A power semiconductor device according to claim 1, wherein the maximum width of the top of the JFET region is at least 50% greater than the minimum width of the bottom of the JFET region.
6. 5. The power semiconductor device of claim 2, wherein the depth of the first side well is between 1.0 and 1.5 microns from a top surface of the semiconductor layer structure, the depth of the JFET region is between 1.0 and 1.7 microns from the top surface of the semiconductor layer structure, and the depth of the JFET region is at least as deep as the depth of the first side well.
7. 7. The power semiconductor device of claim 2, wherein the distance between the first and second side wells is less than 1.5 microns.
8. 8. A power semiconductor device according to claim 1, wherein the maximum width of the top half of the JFET region is between 40% and 80% greater than the minimum width of the bottom half of the JFET region.
9. A power semiconductor device as described in claim 2, wherein a peak doping concentration in the top 0.2 microns of the first side well is at least one order of magnitude lower than a peak doping concentration of the first side well.
10. 10. A power semiconductor device as claimed in claim 1, wherein the JFET region extends beneath the first well.
Citation Information
Patent Citations
Vertical jfet limited silicon carbide power metal oxide semiconductor field effect transistor and method of fabricating vertical jfet limited silicon carbide metal oxide semiconductor field effect transistor
JP2006511961A
Semiconductor apparatus
JP2011159797A
Silicon carbide semiconductor device
JP2011165861A
Semiconductor device and method for manufacturing the same
JP2012235002A
Silicon carbide semiconductor device and silicon carbide semiconductor device manufacturing method
JP2017092367A
