Silicon carbide MOSFET integrated with SCR and manufacturing method thereof
By integrating an SCR structure into a silicon carbide MOSFET and optimizing the electric field using a PNPN four-layer structure and a termination region, the transient overvoltage problem at the gate-source end under high voltage and high power scenarios was solved, achieving high reliability and stability of the device.
Patent Information
- Application Number
- CN202511682011.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-17
- Publication Date
- 2026-01-20
- Estimated Expiration
- 2045-11-17
AI Technical Summary
Existing silicon carbide MOSFETs are susceptible to transient overvoltage and surge stress at the gate and source terminals in high-voltage and high-power scenarios, leading to dielectric breakdown and reduced reliability. Existing protection schemes have inconsistent responses and are prone to false triggering, affecting device lifespan.
The silicon carbide MOSFET with integrated SCR structure achieves fast discharge by integrating a PNPN four-layer structure between the gate and source, utilizing the resistance path between the body region and the epitaxial layer, and optimizing the electric field distribution by arranging dielectric isolation trenches and field limiting rings in the termination region, matching the forward and reverse trigger voltage thresholds, and ensuring stability and reliability under high-frequency switching.
Without increasing area and losses, the surge protection and reliability of the device are improved, the probability of false triggering is reduced, the trigger threshold is kept stable, and the long-term reliability and application adaptability of the whole device are enhanced.
Smart Images

Figure CN121368159A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of power semiconductor devices, and in particular to a silicon carbide MOSFET integrated with an SCR and a manufacturing method thereof. BACKGROUND
[0002] Silicon carbide (SiC) power MOSFETs have advantages of high breakdown field, high thermal conductivity, and high switching speed, and have been widely used in high-voltage and high-power scenarios such as vehicle electric drive, charging piles, photovoltaic inverters, and industrial power supplies. In such applications, the gate terminal of the device is subjected to complex conditions such as high voltage rate of change, common mode oscillation, abnormal driving, hot plug, inductive back sweep, and load short circuit, and is prone to transient overvoltage and surge stress at the gate-source terminal. Due to the limited thickness of the SiC-SiO2 gate dielectric and the limited electric field margin, if there is no effective and rapid discharge path, the gate oxide stress, dielectric trap accumulation, and threshold drift are prone to occur, and in severe cases, dielectric breakdown is triggered, affecting the reliability and service life of the device.
[0003] Existing protection ideas mainly include: connecting transient suppression devices such as TVS and RC absorption networks outside the chip, or introducing a gate-source clamping unit inside the chip. The external scheme is limited by the parasitic parameters of the leads and wiring, and there is often a deviation between the response window and the actual surge waveform of the device, and the consistency is greatly affected by system assembly; the internal scheme is faster in response, but generally faces a series of problems such as difficulty in accurately matching the trigger voltage with the driving window, imbalance in forward and reverse trigger characteristics, easy mis-triggering under high voltage rate of change (dv / dt), and shutdown failure or even parasitic latch caused by too low holding current. At the same time, if the protection structure invades the active channel region or occupies too much layout boundary, it may also have an adverse effect on the on-resistance and switching loss of the main device. SUMMARY
[0004] In order to overcome the defects of the prior art, the technical problem to be solved by the present application is to provide a silicon carbide MOSFET integrated with an SCR and a manufacturing method thereof, which adopts the following technical solutions: In one aspect, the present application provides a silicon carbide MOSFET integrated with an SCR, comprising: an N-type silicon carbide substrate and an N-type epitaxial layer thereon; a P-type body region, a heavily doped N+ source region, and a heavily doped P+ region formed in the N-type epitaxial layer; a gate dielectric layer and a gate electrode covering the P-type body region; a source electrode electrically connected to the N+ source region; a drain electrode electrically connected to the N-type epitaxial layer or the substrate; a PNPN four-layer structure composed of the P+ region, the N-type epitaxial layer, the P-type body region, and the N+ source region is integrated between the gate electrode and the source electrode. wherein the P+ region is in ohmic contact with the source metal; and the operation of the PNPN four-layer structure is equivalently characterized by a cross-coupled first and second triodes; wherein the conduction path of the first triode comprises the N+ source region, the P-type body region and the N-type epitaxial layer, and the conduction path of the second triode comprises the P+ region, the N-type epitaxial layer and the P-type body region; when overvoltage surge occurs at the gate or source, the PNPN four-layer structure is triggered to form a low-impedance discharge channel between the gate and the source.
[0005] Further improved, the P-type body region in the lateral area under the N+ source region and close to the gate side forms a first base series resistance; Further improved, the P-type body region in the lateral area under the N+ source region and close to the gate side forms a first base series resistance; Further improved, the P-type body region in the lateral area under the N+ source region and close to the gate side forms a first base series resistance;
[0006] Further improved, the lateral spacing between the P+ region and the N+ source region and the doping concentration of the P+ region are configured such that the forward trigger voltage threshold of the PNPN four-layer structure is not lower than +18V and avalanche trigger occurs at +20V, and the absolute value of the reverse trigger voltage threshold is not lower than 5V and avalanche trigger occurs at 7V, to match the forward +18V and reverse 5V asymmetric drive window of the electrically controlled main drive.
[0007] Further improved, the PNPN four-layer structure is arranged at the outer edge of the cell region and / or the termination region, and is electrically isolated from the active channel region by a dielectric isolation trench or a junction termination structure.
[0008] Further improved, the termination region comprises at least one of a field limiting ring, a junction termination extension, or a field plate structure electrically connected to the gate electrode.
[0009] Further improved, the doping concentration and depth of the P-type body region, the spacing between the P+ region and the N+ source region, and the contact area between the P+ region and the source metal are cooperatively configured to make the holding current of the PNPN four-layer structure higher than the parasitic latch trigger current of the MOSFET at the highest operating junction temperature.
[0010] Further improved, a low-impedance ohmic contact opening linearly extending in the source direction is provided between the P+ region in ohmic contact with the source metal and the source metal, for accelerating carrier extraction after the surge subsides; and a field plate electrically connected to the gate electrode is provided on the sidewall of the dielectric isolation trench.
[0011] Another aspect of the present application provides a method for fabricating a silicon carbide MOSFET integrated with an SCR, for fabricating the silicon carbide MOSFET integrated with an SCR as claimed in any one of the preceding aspects, comprising the following steps: S1: forming a P-type body region on an epitaxial layer of N-type silicon carbide by photolithography and ion implantation; S2: depositing and patterning a gate stack, depositing polysilicon on the gate side and anisotropically etching to form a side wall; using the above-mentioned side wall as a self-aligned mask, performing N-type ion implantation to form a heavily doped N+ source region; S3: performing P-type ion implantation on the body contact side to form a heavily doped P+ region; S4: performing high-temperature activation annealing to activate the implanted impurities in the above-mentioned P-type body region, N+ source region and P+ region; S5: performing a sacrificial oxidation process, growing a gate dielectric layer after removing the sacrificial oxide layer, and performing interface passivation in a nitrogen-containing oxide atmosphere; S6: depositing an interlayer dielectric and etching to form contact holes, so that the source contact hole simultaneously contacts the above-mentioned N+ source region and the adjacent above-mentioned P+ region, and a gate lead hole is formed; S7: depositing a conductive metal and patterning to form a front metal interconnection to form a source electrode and a gate electrode; forming a field plate electrically connected to the above-mentioned gate electrode in the termination region; and making the above-mentioned P+ region ohmic contact with the above-mentioned source metal; S8: thinning the back surface of the substrate and forming a drain electrode.
[0012] As a further improvement, at least one of the following process conditions is met: Condition one: in steps S1 and S3, the implanted element is Al, multiple ion implantations are used, and the total dose is in the range of 5x1011 cm-2 to 1x1013 cm-2, with a single implantation energy in the range of 50 keV to 650 keV; 13 cm -2 15 cm -2 Condition two: in step S2, the implanted element is N, multiple ion implantations are used, and the total dose is in the range of 5x1011 cm-2 to 5x1013 cm-2, with a single implantation energy in the range of 50 keV to 200 keV; 13 cm -2 14 cm -2 Condition three: the polysilicon side wall formed in step S2 has a thickness of 0.3 μm to 0.7 μm; Condition four: in step S5, the thickness of the sacrificial oxide layer is 15-25 nm, the thickness of the gate dielectric layer is 30-50 nm, and the interface passivation is performed in a nitrogen-containing oxide atmosphere at 1200-1350°C after the growth of the gate dielectric layer; Condition five: in step S3, the forward triggering voltage threshold and the reverse triggering voltage threshold of the PNPN four-layer structure are respectively set by adjusting the energy and dose of the P-type ion implantation and controlling the photolithographic alignment pitch between the P+ region and the N+ source region.
[0013] Compared with the prior art, the application has the following beneficial effects: Firstly, the application integrates the PNPN four-layer structure composed of the P+ region, the N-type epitaxial layer, the P-type body region and the N+ source region between the gate and the source, and uses the natural resistance path of the body region and the epitaxial layer to bear the base series resistance, so as to realize the fast controllable triggering and low resistance discharge when the surge comes. The structure reuses the existing doping and layers of the power MOSFET, is arranged at the outer edge of the cell and / or the terminal region and is isolated from the active channel region, so that the protection unit and the main power unit do not interfere with each other in layout and carrier transport, thereby improving the survival margin of the gate oxide under complex working conditions and the overall surge resistance of the device without significantly increasing the area and loss.
[0014] Secondly, the application cooperatively sets the lateral pitch and doping of the P+ region and the N+ source region to respectively dominate the forward and reverse triggering mechanisms, so that the triggering voltage thresholds in the forward and reverse directions can be matched with the asymmetric driving window, the device remains non-conductive during normal driving and small amplitude oscillation, and quickly enters conduction during the out-of-bound transient state. Meanwhile, the protection unit is arranged in the terminal region and cooperates with at least one of the dielectric isolation trench, the junction termination extension, the field limiting ring and the field plate electrically connected to the gate, so as to optimize the terminal electric field and effectively suppress the displacement current coupling caused by the voltage rate of change, thereby ensuring the stability of the triggering threshold, significantly reducing the probability of false triggering, and balancing the triggering certainty and the switching quality of the main device in the high-frequency switching scenario.
[0015] Thirdly, in terms of layout and contact design, the application provides an efficient extraction channel for the injected carriers during triggering by linearly extending the low-resistance ohmic contact opening in the source direction, and designs the maintenance current of the SCR in a safe interval higher than the parasitic latch triggering current of the device in terms of structure and doping, so that the device can be quickly and reliably turned off after the surge subsides. In addition, the gate interface stability and the repeatability of manufacturing are improved by the steps of sacrificial oxidation and interface passivation in the process flow. Thus, the protection unit balances the triggering speed, the controllability of turn-off and the robustness of temperature conditions, and significantly improves the long-term reliability and application adaptability of the whole machine. BRIEF DESCRIPTION OF DRAWINGS
[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.
[0017] Figure 1 This is a schematic diagram of the SCR equivalent circuit of the present invention; Figure 2 This is a schematic diagram of the cross-sectional structure of the device after step S2 is performed according to the method of the present invention; Figure 3 This is a schematic diagram of the cross-sectional structure of the device after step S3 of the present invention is performed; Figure 4 This is a cross-sectional view of the MOSFET structure of the present invention; Figure 5 This is a flowchart of the method of the present invention.
[0018] Figure label: 101 - N-type silicon carbide substrate; 102 - N-type epitaxial layer; 103 - P-type body region; 104 - Heavily doped N+ source region; 105 - Heavily doped P+ region; 106 - N-type polysilicon gate; 107 - Gate dielectric layer; 108 - Source electrode; 109 - Drain electrode. Detailed Implementation
[0019] To facilitate understanding by those skilled in the art, the structure of the present invention will now be described in further detail with reference to the accompanying drawings: This invention aims to provide a silicon carbide MOSFET with integrated SCR and its fabrication method. The technical solution of this invention will be described in detail below with reference to the accompanying drawings. It should be noted that the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without inventive effort are within the scope of protection of this invention.
[0020] like Figures 2-4 As shown, a silicon carbide MOSFET with integrated SCR of the present invention includes an N-type silicon carbide substrate 101 and an N-type epitaxial layer 102 thereon. A P-type body region 103, a heavily doped N+ source region 104, and a heavily doped P+ region 105 are formed in the N-type epitaxial layer 102. The gate dielectric layer 107 covering the P-type body region 103 and the gate electrode together constitute the control part of the device. The source electrode 108 electrically connected to the N+ source region and the drain electrode 109 electrically connected to the N-type epitaxial layer 102 or the substrate together constitute the current transmission terminal of the device.
[0021] In particular, between the gate electrode and the source electrode 108, there is integrated a PNPN four-layer structure composed of a P+ region, an N-type epitaxial layer 102, a P-type body region 103, and an N+ source region. The P+ region forms an ohmic contact with the source metal. The working mechanism of the PNPN four-layer structure is equivalent to that of a first transistor and a second transistor in cross-coupling. The emitter of the first transistor is the N+ source region, the base is the P-type body region 103, and the collector is the N-type epitaxial layer 102; the emitter of the second transistor is the P+ region, the base is the N-type epitaxial layer 102, and the collector is the P-type body region 103. When an overvoltage surge occurs at the gate or the source, the PNPN four-layer structure is triggered, and a low-resistance discharge channel is formed between the gate and the source. In the equivalent representation, the N-type region on the gate side can be regarded as the emitter of the first transistor, including the N-type polysilicon gate 106 and the adjacent N+ source region, corresponding to the layout.
[0022] Specifically, the formation of the above-mentioned PNPN four-layer structure is the result of the optimization of device design and process. The P-type body region 103 not only serves as the channel body region of the MOSFET, but also as the P-type base layer in the PNPN structure and the base region of the first transistor. The N+ source region not only serves as the source contact region of the MOSFET, but also as the emitter region of the first transistor; the P+ region forms an ohmic contact with the source metal, constituting the emitter of the second transistor; at the circuit equivalent level, the gate polysilicon, the P-type body region 103, and the N+ source region can be regarded as a heterojunction structure participating in the triggering process. The function of the P+ region in the above-mentioned embodiment is different from that of the conventional body contact. It forms the emitter region of the second transistor through ohmic contact with the source metal. The N-type epitaxial layer 102 simultaneously serves as the drift region of the MOSFET and the collector region of the two transistors. Through this multiplexing design, the protection structure does not need to occupy a large chip area, and deep integration with the main power device is achieved.
[0023] In a preferred embodiment of the present application, the P-type body region 103 forms a first base series resistance in the lateral section near the gate side below the N+ source region. The N-type epitaxial lateral section near the source side forms a second base series resistance. The first and second base series resistances are used to generate a voltage drop thereon during an overvoltage surge transient to trigger the conduction of the opposite transistor.
[0024] As preferred, the resistance of the first base series resistance is mainly determined by the doping concentration, depth and lateral distance of the P-body region 103 in the path. In device design, the doping profile of this region is precisely controlled by TCAD simulation tools to ensure that the small leakage current flowing through this region under normal gate voltage, such as +18V or -5V, is not enough to generate a voltage on the resistance that is sufficient to turn on the opposite transistor. However, when a surge voltage comes, such as +20V or -7V, the sharply increased avalanche current or displacement current will flow through this resistance, and a voltage drop exceeding the PN junction opening voltage will be quickly established on it, thereby reliably triggering the SCR action.
[0025] Further, the lateral distance between the P+ region and the N+ source region and the doping concentration of the P+ region are cooperatively configured so that the forward triggering voltage threshold of the PNPN four-layer structure is not lower than +18V and avalanche triggering occurs at +20V, and the absolute value of the reverse triggering voltage threshold is not lower than 5V and avalanche triggering occurs at 7V, to match the asymmetric driving window of the electrically controlled main drive, with a forward driving window of +18V and a reverse driving window of -5V.
[0026] Specifically, the implementation of this asymmetric triggering characteristic is the key to the application of the application to meet specific application requirements. The forward triggering voltage (gate positive source negative) is mainly determined by the avalanche breakdown voltage of the emitter junction of the second transistor. By increasing the doping concentration of the P+ region, the avalanche breakdown voltage of the junction can be reduced, thereby setting the forward triggering point at +20V, which is slightly higher than the normal on-state driving voltage +18V. The reverse triggering voltage (gate negative source positive) is mainly determined by the avalanche breakdown voltage of the emitter junction of the first transistor, which usually has a high breakdown voltage. By precisely controlling the lateral distance between the P+ region and the N+ source region, the resistance path and electric field distribution of the P-body region 103 can be adjusted, thereby controlling the absolute value of the reverse triggering voltage to be -7V, which is slightly higher than the off-state driving voltage -5V. This fine doping and spacing cooperative design ensures that the protection structure is completely invisible during normal operation and responds immediately when the overvoltage occurs.
[0027] In an embodiment of the application, the PNPN four-layer structure is arranged at the outer edge of the cell region and / or the termination region and is electrically isolated by a dielectric isolation trench or a junction termination structure.
[0028] As preferred, the PNPN protection structure is arranged at the periphery or termination region of the cell region, which can effectively avoid interference with the carrier transport of the active channel region, thereby maintaining the excellent on-state and switching characteristics of the power MOSFET itself. The dielectric isolation trench, such as a deep trench formed by reactive ion etching and filled with SiO2, can provide excellent electrical isolation effect. The junction termination isolation uses the depletion region of the reverse biased PN junction to achieve isolation, which has the advantage of simple process. This layout strategy realizes the perfect coexistence of protection function and mainstream capability.
[0029] Preferably, the termination region can include at least one of field limiting rings, junction termination extensions, etc. As a preferred option, a field plate structure electrically connected to the gate electrode can also be included to further optimize the electric field.
[0030] Specifically, the design of the termination structure is critical to the device's voltage withstand and the stability of the protection structure. The field limiting rings or the junction termination extensions lower the peak electric field at the termination surface by laterally broadening the depletion region, preventing premature breakdown. The field plate electrically connected to the gate electrode further optimizes the electric field distribution at the termination region through its underlying electric field modulation effect. More importantly, the presence of the field plate can suppress the displacement current induced at the termination region during high-speed switching due to high dv / dt, which can mistakenly trigger the SCR structure. Therefore, the termination structure not only serves the voltage withstand, but also directly participates in the reliability and robustness construction of the protection function.
[0031] In a preferred option of the present application, the doping concentration and depth of the P-body region 103, the spacing between the P+ region and the N+ source region, and the contact area between the P+ region and the source metal are cooperatively configured to make the holding current of the PNPN four-layer structure higher than the parasitic latch-up trigger current of the MOSFET at the highest operating junction temperature.
[0032] Specifically, preventing latch-up is a prerequisite for the successful application of the SCR structure in power devices. The holding current refers to the minimum current that can maintain the on-state of the SCR once it is turned on. If the holding current is too low, the SCR may not be able to turn off after a surge, leading to device failure. By increasing the doping concentration and depth of the P-body region 103, the current gain of the NPN triode can be reduced. By optimizing the spacing between the P+ region and the N+ source region, the carrier recombination can be increased. By increasing the contact area between the P+ region and the source metal, a high-efficiency hole extraction path can be provided. The synergistic effect of these measures significantly improves the holding current. In terms of design goals, it must be ensured that this holding current is higher than the trigger current of any parasitic latch-up that may be caused by the structure of the MOSFET itself throughout the entire operating temperature range, for example, -55°C to 175°C, thereby fundamentally eliminating the risk of latch-up.
[0033] In an optimized embodiment of the present application, in order to further improve the recovery speed and working stability of the device after surge events, the layout and terminal of the PNPN structure are designed in a targeted manner. On the one hand, between the P+ region and the source metal, the ohmic contact opening is intentionally designed to extend linearly in a direction parallel to the source metal strip, rather than the traditional square or circular point contact. This linearly extended low-resistance ohmic contact opening significantly increases the extraction area of holes, providing a fast discharge channel with extremely low resistance for the large number of holes accumulated in the P+ region after the surge subsides. This can effectively speed up the turn-off process of the SCR, significantly shorten the holding time, and thus reduce the risk of device failure caused by SCR holding on.
[0034] On the other hand, a field plate structure is formed on the sidewall of the dielectric isolation trench by depositing polysilicon or metal electrically connected to the gate electrode. The field plate can effectively modulate the surface electric field distribution near the isolation region and suppress the displacement current caused by parasitic capacitance coupling under high dv / dt. This design can avoid the mis-triggering of the integrated SCR protection structure due to the dv / dt effect during the normal high-speed switching process of the device, thereby ensuring the reliability and robustness of the main power circuit.
[0035] In a more specific embodiment, a low-resistance ohmic contact opening extending linearly in the source direction is provided between the P+ region and the source metal to accelerate carrier extraction after the surge subsides, and a field plate electrically connected to the gate electrode is provided on the sidewall of the dielectric isolation trench.
[0036] Specifically, the design of the low-resistance ohmic contact opening aims to provide a fast discharge channel for the large number of holes injected into the P+ region during SCR conduction. When the surge voltage disappears and the anode current tries to drop below the holding current, this low-resistance path can quickly extract the remaining carriers, destroy the holding condition of the SCR, and thus quickly and reliably turn off the SCR, shorten the holding time, and improve the dynamic reliability of the device. At the same time, the field plate provided on the sidewall of the dielectric isolation trench and electrically connected to the gate electrode precisely modulates the surface electric field near the isolation region. In high dv / dt events, the field plate can effectively suppress the displacement current caused by parasitic capacitance coupling, preventing it from reaching the trigger threshold of the SCR, thereby avoiding mis-triggering during normal switching and further enhancing the stability of the circuit.
[0037] As Figure 5 shown, the present application also provides a method for manufacturing a silicon carbide MOSFET integrated with an SCR. The following will take the preparation process of an N-type SiC MOSFET suitable for 1200V voltage level as an example to elaborate the method in detail. The method comprises the following steps: S1: Form P-type body region 103 by lithography and ion implantation on the epitaxial layer of N-type silicon carbide.
[0038] Specifically, first, an N-type 4H-SiC epitaxial wafer with appropriate thickness and doping concentration is prepared. An N-type polysilicon layer is deposited by thermal oxidation and LPCVD as a hard mask. Lithography is performed using a Pwell lithography plate to define the area of P-type body region 103, and then dry etching is performed to remove the polysilicon in the exposed area. Next, a layer of SiO2 about 50 nm thick is deposited by PECVD in the etched window as an ion implantation mask. The implant element is selected as aluminum, and four times of implantation is adopted to obtain the ideal doping profile. The first implantation dose is 2x10 13 cm -2 , and the energy is 610 keV; the dose and energy of the subsequent three implantations are gradually reduced to ensure the formation of approximately uniform P-type doping distribution in the longitudinal direction, and the total dose is controlled at about 8x10 13 cm -2 . After implantation, the SiO2 mask is removed using HF solution.
[0039] S2: Deposit and pattern the gate stack, deposit polysilicon on the gate side and anisotropically etch to form side walls; perform N-type ion implantation with the side walls as a self-aligned mask to form heavily doped N+ source region 104. As shown in Figure 2 .
[0040] Specifically, after removing the polysilicon hard mask, an N-type polysilicon layer with a thickness of 0.5 μm is directly deposited by LPCVD. Subsequently, the polysilicon layer is etched by anisotropic dry etching to form polysilicon side walls with a width of about 0.5 μm on the side walls of the gate pattern. The thickness of this side wall precisely controls the alignment of the subsequent N+ source region and channel region. Then, about 50 nm of SiO2 is deposited again by PECVD as an implantation mask. Nitrogen ion implantation is performed with the polysilicon side walls as a self-aligned mask to form heavily doped N+ source region 104. Again, four times of implantation is adopted, the first implantation dose is 1.1x10 14 cm -2 , the energy is 175 keV, and the dose and energy of the subsequent implantations are gradually reduced, and the total dose is about 3x10 14 cm -2 . After implantation, the SiO2 mask is removed.
[0041] S3: Perform P-type ion implantation on the body contact side to form heavily doped P+ region 105. As shown in Figure 3 .
[0042] Specifically, a SiO2 implantation masking layer is deposited again via PECVD. A photolithography pattern is used to define the regions where heavily doped P+ contacts and SCR structure P+ emitters need to be formed. High-dose aluminum ion implantation is then performed, using four implantations, with the first implantation dose being 8 × 10⁻⁶. 14 cm -2 The energy was 610 keV, with subsequent doses and energy decreasing, and the total dose was approximately 2 × 10⁻⁶. 15 cm -2 The heavily doped P+ region 105 formed in this step is used partly for conventional bulk ohmic contacts and partly for the SCR structure, and is connected to the gate through subsequent metallization.
[0043] S4: Perform high-temperature activation annealing to activate the injected impurities in the P-type body region 103, the N+ source region, and the P+ region.
[0044] Specifically, high-temperature annealing is required to activate all impurities introduced by ion implantation and repair lattice damage. A standard carbon film protection process is used to form a carbon film on the SiC wafer surface to prevent SiC surface decomposition at high temperatures. Subsequently, high-temperature annealing is performed at 1650°C for 30 minutes under an argon atmosphere. After annealing, the carbon film on the surface is removed by oxygen plasma ashing.
[0045] S5: Perform sacrificial oxidation process, remove the sacrificial oxide layer, grow gate dielectric layer 107, and perform interface passivation in a nitrogen oxide atmosphere.
[0046] Specifically, a sacrificial oxide layer approximately 20 nm thick is first grown on the SiC surface via thermal oxidation. This layer is used to absorb potentially damaged areas of the surface. It is then completely removed by etching with an HF solution. Next, a gate oxide layer approximately 40 nm thick is grown in a humid oxygen atmosphere at 1000°C to 1100°C. Subsequently, the device undergoes interface passivation treatment at a high temperature of 1300°C in a nitric oxide atmosphere for approximately 1 hour. This process significantly reduces the SiC-SiO2 interface state density, improving channel mobility and gate oxide reliability.
[0047] S6: Deposit interlayer medium and etch to form contact holes, so that the source contact holes simultaneously contact the N+ source region and the adjacent P+ region, and open gate lead holes.
[0048] Specifically, LTO and BPSG are deposited sequentially using LPCVD to form an interlayer dielectric layer. Planarization is then performed. Source contact holes are formed using an ohmic contact photolithography plate and dry etching; these contact holes are designed to simultaneously expose the N+ source region and the adjacent P+ contact region. Simultaneously, the gate polysilicon lead contact region is etched using a gate lead hole photolithography plate.
[0049] S7: Deposit conductive metal and pattern it to form a front metal interconnect to form the source electrode 108 and the gate electrode; simultaneously form a field plate electrically connected to the gate electrode in the terminal region; make the P+ region ohmic in contact with the source metal.
[0050] Specifically, the process begins with cleaning and resist removal, followed by the sequential deposition of Ti, Ni, and Ag metal stacks using a magnetron sputtering system, with a total thickness of approximately 4 μm. The front-side metal is patterned using photolithography and wet / dry etching to form the pressure points for the source electrode 108 and the gate electrode. During patterning, a metal field plate connected to the gate electrode is retained above the termination region. It is particularly important to emphasize that in the SCR region, the P+ region forms an ohmic contact with the source metal. The PNPN four-layer structure achieves its protective function through the inherent coupling of the P+ region, the N-type epitaxial layer 102, the P-type body region 103, and the N+ source region. Subsequently, rapid thermal annealing is performed in an inert gas atmosphere at approximately 1000°C to form excellent ohmic contacts.
[0051] S8: Thin the back side of the substrate and form the drain electrode 109.
[0052] Specifically, the wafer is flipped, and the back side of the N+ substrate is mechanically ground and chemically mechanically polished to reduce its thickness to approximately 170 μm. Then, a Ni-Ag metal stack with a thickness of approximately 1 μm is sequentially deposited on the back side by electron beam evaporation or sputtering. Finally, laser annealing is performed to form a low-resistance back drain ohmic contact.
[0053] This invention is in Figure 1 Based on the fundamental principle of the MOSFET device M1, two coupled transistors T1 and T2 are connected between the gate and source. When a surge occurs at the gate, the avalanche current generated by the emitter of T1 flows through resistor R1. When the voltage across R1 reaches the PN junction turn-on voltage, the lower transistor T2 turns on, and the collector of T2 continuously provides current to the base of T1, forming positive feedback and thus dissipating the surge from the gate to the source. Similarly, when a surge current occurs at the source, the current generated by the collector of T2 creates a voltage across R2 that reaches the PN junction turn-on voltage, causing T1 to turn on, forming positive feedback and dissipating the current.
[0054] In a preferred embodiment of the method of the present invention, the above process satisfies at least one of the following process conditions to achieve better performance and controllability: Condition 1: In steps S1 and S3, the implanted element is Al, and multiple ion implantations are performed, with a total dose range of 5 × 10⁻⁶. 13 cm -2 Up to 1×10 15 cm -2The single implantation energy ranges from 50 keV to 650 keV. This wide range of parameters provides clear guidance for those skilled in the art to flexibly adjust the parameters according to different withstand voltage levels and trigger voltage requirements. The core of this approach lies in forming an optimized doping profile through multiple implantations.
[0055] Condition 2: In step S2, the implanted element is N, and multiple ion implantations are performed, with a total dose range of 5 × 10⁻⁶. 13 cm -2 Up to 5×10 14 cm -2 The single injection energy range is from 50keV to 200keV. This parameter range ensures that the N+ source region can form a good ohmic contact and form a PN junction with the P-type body region 103 with controllable junction depth, which is the basis for precisely controlling the channel length and SCR triggering characteristics.
[0056] Condition 3: The thickness of the polysilicon sidewall formed in step S2 is 0.3 μm to 0.7 μm. This dimension directly determines the distance between the junction and the gate edge of the self-aligned N+ source region, and is key to controlling the effective channel length, thereby balancing the device's on-resistance and threshold voltage.
[0057] Condition 4: In step S5, the thickness of the sacrificial oxide layer is 15 nm to 25 nm, the thickness of the gate dielectric layer 107 is 30 nm to 50 nm, and interface passivation is performed in a nitrogen-containing oxide atmosphere at 1200 °C to 1350 °C after the gate dielectric is grown. This set of parameters is the process guarantee for obtaining a gate oxide layer with high reliability and low interface state density, which is crucial for the long-term stable operation and switching efficiency of the device.
[0058] Condition 5: In step S3, by adjusting the energy and dose of P-type ion implantation and controlling the photolithographic alignment distance between the P+ and N+ source regions, the forward and reverse trigger voltage thresholds of the PNPN four-layer structure are set respectively. This is a direct process for achieving asymmetric and adjustable trigger voltages. By adjusting the P+ implantation conditions, the forward trigger voltage can be mainly controlled; by controlling the distance between P+ and N+, the reverse trigger voltage can be mainly controlled, thus enabling customized designs for different driving voltage windows.
[0059] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.
Claims
1. A silicon carbide MOSFET integrated with an SCR, comprising: an N-type silicon carbide substrate and an N-type epitaxial layer thereon; a P-type body region, a heavily doped N+ source region and a heavily doped P+ region formed in the N-type epitaxial layer; a gate dielectric layer and a gate electrode covering the P-type body region; a source electrode electrically connected to the N+ source region; a drain electrode electrically connected to the N-type epitaxial layer or substrate; characterized in that between the gate electrode and the source electrode, a PNPN four-layer structure is integrated, which is composed of the P+ region, the N-type epitaxial layer, the P-type body region and the N+ source region; wherein the P+ region is in ohmic contact with the source metal; and the working mechanism of the PNPN four-layer structure is equivalent to that of a first transistor and a second transistor cross-coupled; wherein the conduction path of the first transistor comprises the N+ source region, the P-type body region and the N-type epitaxial layer, and the conduction path of the second transistor comprises the P+ region, the N-type epitaxial layer and the P-type body region; when an overvoltage surge occurs at the gate or source, the PNPN four-layer structure is triggered to form a low-resistance discharge channel between the gate and the source.
2. The silicon carbide MOSFET integrated with an SCR according to claim 1, characterized by: the P-type body region in the lateral area near the gate side below the N+ source region constitutes a first base series resistance; the N-type epitaxial layer in the lateral area below the P-type body region and adjacent to the P+ region constitutes a second base series resistance; the first and second base series resistances are used to generate a voltage drop thereon under an overvoltage surge transient to trigger the conduction of the opposite transistor.
3. The silicon carbide MOSFET integrated with an SCR according to claim 2, wherein: The lateral spacing between the P+ region and the N+ source region and the doping concentration of the P+ region are configured such that the forward triggering voltage threshold of the PNPN four-layer structure is not lower than +18V and avalanche triggering occurs at +20V, the absolute value of the reverse triggering voltage threshold is not lower than 5V and avalanche triggering occurs at 7V, to match the asymmetric driving window of +18V in the forward direction and 5V in the reverse direction of the electrically controlled main drive. 5V.
4. The silicon carbide MOSFET integrated with an SCR according to claim 3, wherein: the PNPN four-layer structure is arranged at the outer edge of the cell region and / or the termination region; and is electrically isolated from the active channel region by a dielectric isolation trench or a junction termination structure.
5. The silicon carbide MOSFET integrated with an SCR according to claim 4, wherein: the termination region includes at least one of a field limiting ring, a junction termination extension or a field plate structure electrically connected to the gate electrode.
6. The silicon carbide MOSFET integrated with an SCR according to claim 5, wherein: The doping concentration and depth of the P-type body region, the spacing between the P+ region and the N+ source region, and the contact area between the P+ region and the source metal are cooperatively configured to make the holding current of the PNPN four-layer structure higher than the parasitic latch trigger current of the MOSFET at the highest working junction temperature.
7. The silicon carbide MOSFET integrated with an SCR according to claim 6, wherein: A low-resistance ohmic contact opening linearly extending in the source direction is provided between the P+ region in ohmic contact with the source metal and the source metal, which is used to accelerate carrier extraction after the surge subsides; and a field plate electrically connected to the gate electrode is provided on the sidewall of the dielectric isolation trench.
8. A method of fabricating a silicon carbide MOSFET integrated SCR for fabricating a silicon carbide MOSFET integrated SCR as claimed in any one of claims 1 to 7, characterized by, comprising the following steps: S1: forming a P-type body region on an epitaxial layer of N-type silicon carbide by photolithography and ion implantation; S2: depositing and patterning a gate stack, depositing polysilicon on the gate side and anisotropically etching to form a sidewall; using the sidewall as a self-aligned mask, performing N-type ion implantation to form a heavily doped N+ source region; S3: performing P-type ion implantation on the body contact side to form a heavily doped P+ region; S4: performing high-temperature activation annealing to activate the implanted impurities in the P-type body region, N+ source region and P+ region; S5: performing a sacrificial oxidation process, growing a gate dielectric layer after removing the sacrificial oxide layer, and performing interface passivation in a nitrogen-containing oxide atmosphere; S6: depositing interlayer dielectric and etching to form contact holes, so that the source contact hole simultaneously contacts the N+ source region and the adjacent P+ region, and a gate lead hole is formed; S7: depositing conductive metal and patterning to form front metal interconnection, so as to form source electrode and gate electrode; field plate is formed in the terminal region and electrically connected with the gate electrode; the P+ region is ohmically contacted with the source metal; S8: thinning the back of the substrate and forming the drain electrode.
9. The method of fabricating an integrated SCR silicon carbide MOSFET of claim 8, wherein, At least one of the following process conditions is met: Condition one: in step S1 and step S3, the implanted element is Al, multiple ion implantation is adopted, the total dose range is 5x1014cm-2 to 1x1015cm-2, and the single implantation energy range is 50keV to 650keV; 13 cm -2 to 1x1015cm 15 cm -2 -2. Condition two: in step S2, the implanted element is N, multiple ion implantation is used, the total dose range is 5 x 1015cm-2 to 5 x 1016cm-2, and the single implantation energy range is 50 keV to 200 keV. 13 cm -2 to 5 x 1016cm-2 14 cm -2 , and the single implantation energy range is 50 keV to 200 keV. Condition three: the polysilicon side wall formed in step S2 has a thickness of 0.3-0.7 μm; Condition four: in step S5, the thickness of the sacrificial oxide layer is 15-25 nm, the thickness of the gate dielectric layer is 30-50 nm, and the interface passivation is performed in a nitrogen-containing oxide atmosphere at 1200-1350 °C after the growth of the gate dielectric; Condition five: in step S3, by adjusting the energy and dose of P-type ion implantation, and controlling the photolithographic alignment pitch of the P+ region and the N+ source region, the forward trigger voltage threshold and the reverse trigger voltage threshold of the PNPN four-layer structure are set respectively.
Citation Information
Patent Citations
Memory device and forming method thereof
CN119545775A
Source / Drain Features For Multi-Gate Device And Method Of Fabricating Thereof
US20230378298A1
Method of fabricating field effect transistors having self-registering electrical connections between gate electrodes and metallic interconnection lines, and fabrication of integrated circuits containing the transistors
US4035198A
Cited By
Power semiconductor device and preparation method thereof
CN122121269A
