An IGBT device
By designing the electric field transition layer and drift region in IGBT devices, controlling the average gradient of the energy band occupied by the minority carriers of the electric field transition layer, the problem of tailing current in traditional IGBT devices is solved, and lower shutdown loss and higher performance advantages are achieved.
Patent Information
- Application Number
- CN202411907011.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-23
- Publication Date
- 2025-04-04
AI Technical Summary
Traditional IGBT devices have tailing current problems, which limits their performance and application promotion.
By designing the electric field transition layer and drift region in the IGBT device, the average gradient of the energy band occupied by the minority carriers of the electric field transition layer can be controlled, and the rapid removal of minority carriers can be achieved and the duration of the tailing current is reduced.
It effectively reduces the shutdown loss of IGBT devices, improves the performance of the device, and is suitable for mass production.
Smart Images

Figure CN119767700A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of semiconductor technology, and in particular, to an IGBT device. Background Art
[0002] Super IGBT has gone through three stages of research. Due to its obvious tail current, SuperIGBT cannot be mass-produced industrially.
[0003] 1. Introduction:
[0004] Super IGBT technology research was first proposed in 2004 and has gone through three stages of research. The research results show that Super IGBT has performance advantages and development potential, but due to the problem of device turn-off tail current, Super IGBT only remains at the research level and has not been achieved in industrial mass production. In particular, the second and third stages of research only have simulation results, and no substantial process verification and product development have been carried out.
[0005] 1.1 The first stage: Planar Super IGBT (P-pillar is connected to P-body)
[0006] The first phase of Super IGBT research took the planar Super IGBT (P-pillar connected to P-body) as the object, and carried out quantitative analysis from the simulation and experimental levels. The research results not only verified the feasibility of Super IGBT technology, but also showed the turn-off loss advantage of Super IGBT. Compared with the traditional planar FS IGBT, the planar Super IGBT has a better compromise performance. However, there are still three problems in the research at this stage: 1. When the SJBT is forward-conducted, there is a unipolar conduction mode similar to SJMOSFET, and the on-state voltage drop of the SJBT is not effectively reduced, which limits the performance of the SJBT; 2. The Super IGBT in this study adopts a planar structure, and the device Pitch is too large, which limits the performance of the Super IGBT; 3. The optimization of the SuperIGBT transition zone has not been studied, and the device turn-off loss has not been effectively optimized.
[0007] 1.1.1 Simulation study:
[0008] In 2004, ABB engineer Friedhelm D. Bauer first proposed the concept of Super IGBT and published a related paper [1]. In the paper, Super IGBT was named Super junction bipolar transistor (SJBT). The SJBT structure is based on a planar cell design (such as Figure 1The simulation results show that the 1200VSJBT has smaller switching losses, and the device performance FoM (Vcesat*Eoff) is about 5.60V*mJ (at 600V / 96A), but the result is worse than the 3.15V*mJ of Infineon's latest H7 series 1200V IGBT products with the same specifications.
[0009] In 2006, Fairchild Semiconductor (Korea) engineer Kwang-Hoon Oh continued his research on Super IGBT[2] and introduced field stop layer (FSL) based on previous work. Figure 2 As shown. The simulation results show that the 1200VSuper IGBT can achieve a good compromise between on-state characteristics and off-state characteristics, with an on-state voltage drop of 1.6V and a turn-off loss of 20μJ / A. Through conversion, it can be seen that the Super IGBT performance figure of merit FoM is optimized to about 4.03V*mJ, which is worse than the figure of merit of 3.15V*mJ of Infineon's latest H7 series 1200V IGBT products of the same specification.
[0010] In 2007, Professor Florin Udrea's team at Cambridge University conducted an optimization simulation of the SJBT[5], focusing on the quantitative study of the N-pillar and P-pillar concentrations in the superjunction region. Figure 3 As shown. Compared with the traditional 1200V FSIGBT, the Eoff of the Super IGBT of the same specification is optimized. Through conversion, it can be seen that the performance figure of merit FoM is about 4.48V*mJ, which is worse than the figure of merit of 3.15V*mJ of Infineon's latest H7 series 1200V IGBT products of the same specification.
[0011] In 2010, Florin Udrea’s team used finite element simulation software to simulate the switching characteristics of Super IGBT [3]. The research results revealed that the device is close to a unipolar conduction mode during the switching process. The above conduction mechanism can accelerate the device’s turn-off process and effectively reduce the device’s turn-off loss.
[0012] 1.1.2 Experimental verification:
[0013] In 2013, Lorenzo M. Selgi of STMicroelectronics experimentally verified the performance of Super IGBT[8]. The device structure is as follows: Figure 4-A and Figure 4-BAs shown. In this work, the Super IGBT is realized based on the process route of "multiple epitaxy + ion implantation", in which the cell structure adopts a planar gate structure. Limited by the actual manufacturing process capabilities, this work only studies the performance of the Super IGBT with a specification of 650V. Experimental results show that under 450V / 15A switching conditions, the 650V Super IGBTFoM value is improved by 41%, which is only 3.47V*mJ. According to theoretical calculations, under 75A current conditions, the FoM value rapidly degenerates to 17V*mJ. The above FoM is far greater than the acceptable level in practical applications, which seriously limits the product development progress of Super IGBT.
[0014] In 2016, Kwang-Hoon Oh of TRinno Technolog of South Korea conducted experimental verification of 650V planar Super IGBT
[10] . The device structure is as follows: Figure 5-A and Figure 5-B As shown. The experimental results show that when the collector current density is 250A / cm2, the IGBT on-state voltage drop Vcesat is 1.4V, and the IGBT tail current time can be reduced to 35ns. This work verifies the technical potential of planar Super IGBT from an experimental level, and the FoM of 650V Super IGBT is optimized to 2.35V*mJ, which is the same as the figure of merit of Infineon's latest product 2.44V*mJ. However, due to the increase in process complexity, the relevant research is not worth mass production.
[0015] 1.2 The second stage: Trench type Super IGBT (P-pillar is connected to P-body)
[0016] The research object of this stage was adjusted from planar Super IGBT (P-pillar connected to P-body) to trench Super IGBT (P-pillar connected to P-body). The above device structure optimization solved the performance limitation problem caused by the excessive pitch of planar Super IGBT, and further improved the comprehensive performance of Super IGBT. However, there are still three problems in the research of this stage: 1. When SJBT is forward-conducted, there is a unipolar conduction mode similar to SJ MOSFET, and the on-state voltage drop of SJBT is not effectively reduced, which limits the performance of SJBT; 2. Trench Super IGBT (P-pillar connected to P-body) has only simulation research, and process development and verification have not been carried out; 3. The optimization of Super IGBT transition zone has not been studied, and the device turn-off loss has not been effectively optimized.
[0017] 1.2.1 Simulation study:
[0018] In 2015, ABB engineer Friedhelm D. Bauer conducted research on trench-type Super IGBTs. Figure 6 As shown in the figure. The research results show that compared with Super Junction MOSFET, trench Super IGBT has lower chip cost and shows obvious competitive advantages[9]. In this work, the FoM of 650V Super IGBT is optimized to 4.28V*mJ (under 350V / 96A switching conditions), which is worse than the figure of merit of 2.44V*mJ of Infineon's latest 650V IGBT product with the same specification.
[0019] In 2018, Professor Gong Min’s team at Sichuan University conducted research on “carrier storage effect enhancement” Super IGBT
[12] . The device structure is as follows: Figure 7 As shown. This work is based on TCAD simulation. The results show that after the Schottky junction is introduced into the emitter, the P-pillar Fermi level can be increased, thereby achieving the effect of enhancing the carrier storage effect. In the new structure, the FoM of the 1200VSuper IGBT performance is optimized to 4.03V*mJ (under 600V / 96A switching conditions), which is worse than the FoM of the same specification 1200V IGBT product of Infineon's latest H7 series of Infineon.
[0020] In 2019, Zhigang Wang from Southwest Jiaotong University focused on analyzing the performance of the Super IGBT turn-off process
[13] . The device structure is as follows: Figure 8 As shown. Based on TCAD simulation, the influence of N-pillar and P-pillar doping concentration on turn-off loss is quantitatively analyzed, and a quantitative model is established to predict the turn-off loss and on-state voltage drop of the device. In this work, the FoM of 1200VSuper IGBT performance is 6.34V*mJ (under 600V / 96A switching conditions), which is worse than the FoM of 3.15V*mJ of Infineon's latest 1200V IGBT product of the same specification.
[0021] 1.2.2 Experimental verification:
[0022] There are many research reports on the trench type Super IGBT (P-pillar is connected to P-body) structure, but the relevant structure only has simulation results and no complete experimental verification data support.
[0023] 1.3 The third stage: Trench type Super IGBT (P-pillar and P-body are not connected)
[0024] Based on the second phase of Super IGBT research, this phase of trench Super IGBT research adopted a solution where P-pillar and P-body were not connected. The simulation results showed that the problem of SJ MOSFET-like unipolar conduction mode during the forward conduction of Super IGBT was successfully solved, effectively reducing the on-state voltage drop of Super IGBT. However, there are still two problems in this phase of research: 1. The research on trench Super IGBT (P-pillar and P-body are not connected) only has simulation analysis, and no process development verification has been carried out; 2. The optimization of Super IGBT transition zone has not been studied, and the device turn-off loss has not been effectively optimized.
[0025] 1.3.1 Simulation study:
[0026] In 2015, Professor Florin Udrea's team changed the IGBT structure from planar to trench based on the original research. Fig. 9 As shown in the figure, the performance of the trench Super IGBT was further verified by simulation [4]. The research results show that the Super IGBT has smaller leakage and turn-off losses under high temperature conditions of 200°C. In addition, the structure in which the P-Pillar and Pbody of the Super IGBT super junction area are not connected is more conducive to obtaining a better compromise characteristic, that is, reducing the turn-off loss without sacrificing the forward on-state performance. In this work, under high temperature conditions of 200°C, the performance figure of merit FoM of the 1200V Super IGBT is 11.2V*mJ, which is about 29% lower than that of the FS IGBT (FoM is 15.68V*m) (under 620V / 96A switching conditions). This result is worse than the figure of merit of Infineon's latest 1200V IGBT product with the same specification.
[0027] 1.3.2 Experimental verification:
[0028] Although many studies have been done in the field of Super IGBT, the research on trench Super IGBT is mostly at the simulation level. More importantly, the research on the structure where the P-pillar and P-body are not connected has only limited simulation, and no complete research, experimental verification and optimization have been carried out.
[0029] After three phases of research, the technical feasibility of Super IGBT was fully discussed, and the device performance was optimized for multiple iterations. However, the performance of the above devices is still insufficient in actual use. Among them, the performance figure of merit FoM of 650V SuperIGBT was optimized to 2.35V*mJ (400V / 96A); the performance figure of merit FoM of 1200V Super IGBT was reduced to 4.03V*mJ (600V / 96A).
[0030] Super IGBT devices have gone through three stages of research. However, Super IGBT devices still have obvious tail current, which limits the performance and application promotion of Super IGBT devices.
[0031] Therefore, the tail current of the traditional IGBT device is relatively large, which limits the performance of the IGBT device and is a technical problem that technical personnel in this field urgently need to solve.
[0032] The above information disclosed in the background section is only for enhancing understanding of the background of the present application and therefore it may contain information that does not form the prior art known to a person of ordinary skill in the art. Summary of the invention
[0033] The present application provides an IGBT device to solve the technical problem of large tail current in traditional IGBT devices.
[0034] The present application provides an IGBT device, comprising:
[0035] A substrate of a first doping type, serving as a collector region;
[0036] An electric field stop layer of a second doping type is formed on the substrate;
[0037] An electric field transition layer of a second doping type is formed on the electric field stop layer;
[0038] A drift region of a second doping type is formed on the electric field transition layer;
[0039] A column region of a first doping type is formed in the drift region, wherein a lower end surface of the column region is spaced from a lower surface of the drift region; and an upper end surface of the column region is spaced from an upper surface of the drift region;
[0040] The thickness W of the electric field transition layer and the doping concentration N of the substrate as the collector region are Collector The following relationship is satisfied:
[0041]
[0042] The present application also provides an IGBT device, comprising:
[0043] An IGBT device, comprising:
[0044] A substrate of a first doping type, serving as a collector region;
[0045] An electric field stop layer of a second doping type is formed on the substrate;
[0046] An electric field transition layer of a second doping type is formed on the electric field stop layer;
[0047] A drift region of a second doping type is formed on the electric field transition layer;
[0048] A column region of a first doping type is formed in the drift region, a lower end surface of the column region and a lower surface of the drift region are spaced apart, and an upper end surface of the column region and an upper surface of the drift region are spaced apart;
[0049] The average gradient K of the energy band occupied by the minority carriers in the electric field transition layer 3 It is precisely controlled to control the removal rate of minority carriers from the electric field transition layer;
[0050] The minority carriers in the electric field transition layer are the minority carriers remaining in the electric field transition layer after the carriers in the collector region are injected into the electric field transition layer during the shutdown process of the IGBT device.
[0051] Due to the adoption of the above technical solution, this application has the following technical effects:
[0052] The IGBT device of the present application not only has a faster removal rate of minority carriers in the electric field transition layer, which shortens the duration of the tail current and reduces the turn-off loss, but also has low difficulty in actual process implementation and is suitable for mass production. BRIEF DESCRIPTION OF THE DRAWINGS
[0053] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:
[0054] Figure 1 A schematic diagram of the first Super IGBT device in the first stage of the background technology;
[0055] Figure 2 A schematic diagram of a second Super IGBT device in the first stage of the background technology;
[0056] Figure 3 A schematic diagram of the third Super IGBT device in the first stage of the background technology;
[0057] Figure 4-A A schematic diagram of the fourth Super IGBT device in the first stage of the background technology;
[0058] Figure 4-B for Figure 4-A Simulation and partial enlargement of the image;
[0059] Figure 5-AA schematic diagram of the fifth Super IGBT device in the first stage of the background technology;
[0060] Figure 5-B for Figure 5-A Simulation diagram of
[0061] Figure 6 A schematic diagram of the first Super IGBT device in the second stage of the background technology;
[0062] Figure 7 A schematic diagram of two positions of a second Super IGBT device in the second stage of the background technology;
[0063] Figure 8 A schematic diagram of a third Super IGBT device in the second stage of the background technology;
[0064] Fig. 9 A schematic diagram of a fourth Super IGBT device in the second stage of the background technology;
[0065] Fig.10 A schematic diagram of an IGBT device which is the basis for the development of the IGBT device of the present application;
[0066] Fig.11 for Fig.10 Schematic diagram of the current during the turn-off process of the IGBT device;
[0067] Figure 12-A for Fig.10 3D schematic diagram of the conduction band Ec of the IGBT device during the current tailing phase;
[0068] Figure 12-B for Fig.10 3D schematic diagram of the electron quasi-Fermi level Efn in the current tailing phase of the IGBT device;
[0069] Figure 12-C for Fig.10 3D schematic diagram of the hole quasi-Fermi level Efp in the current tailing phase of the IGBT device;
[0070] Figure 12-D for Fig.10 3D schematic diagram of the valence band Ev of the IGBT device during the current tailing phase;
[0071] Figure 13-A for Fig.10 Two-dimensional energy band diagram of the IGBT device at the Y=0 position;
[0072] Figure 13-B for Figure 13-A The middle rectangular dotted box is a partial schematic diagram of the vertical axis enlarged;
[0073] Figure 14-A A schematic diagram of an IGBT device according to an embodiment of the present application;
[0074] Figure 14-B The IGBT device and Fig.10 Valence band Ev diagram of IGBT device;
[0075] Fig.15 for Fig.10 Typical waveform diagram of the turn-off process of the IGBT device;
[0076] Figure 16-A for Fig.10 The electronic current diagram corresponding to the 9 key moments of the IGBT device's turn-off process;
[0077] Figure 16-B for Fig.10 The hole current diagram corresponding to the 9 key moments of the IGBT device's turn-off process;
[0078] Fig.17 For the IGBT device of this application based on TCAD software and Fig.10 The process simulation comparison diagram of the IGBT device shown;
[0079] Fig.18 The IGBT device of this application at the experimental level and Fig.10 The IGBT device performance comparison chart is shown.
[0080] Reference numerals:
[0081] In the prior art:
[0082] P-type substrate 10-1, N-type drift region 10-2, P-type column region 10-3, back metal 10-4;
[0083] This application:
[0084] Substrate 1, electric field cutoff layer 2, electric field transition layer 3, drift region 4, column region 5, collector metal 6,
[0085] Body region 71 , trench gate 72 , emitter 73 , and ohmic contact region 74 . DETAILED DESCRIPTION
[0086] In order to better understand this application, it is necessary to fully understand the research process and theoretical derivation of this application. The applicant has placed the research process and theoretical derivation of this application after the embodiments. Please check if necessary.
[0087] Embodiment 1
[0088] like Figure 14-AAs shown, the IGBT device of the embodiment of the present application includes:
[0089] A substrate 1 of a first doping type, serving as a collector region;
[0090] An electric field stop layer 2 of a second doping type, formed on the substrate 1;
[0091] An electric field transition layer 3 of a second doping type is formed on the electric field stop layer 2;
[0092] A drift region 4 of a second doping type, formed on the electric field transition layer 3;
[0093] A column region 5 of a first doping type is formed in the drift region 4, wherein a lower end surface of the column region 5 is spaced from a lower surface of the drift region 4, and an upper end surface of the column region 5 is spaced from an upper surface of the drift region 4;
[0094] Wherein: the interface between the electric field transition layer 3 and the drift region 4 is the lower surface of the depletion region formed in the turn-off stage of the IGBT device;
[0095] The average gradient K of the energy band occupied by the minority carriers in the electric field transition layer 3 It is precisely controlled to control the removal speed of the minority carriers in the electric field transition layer; the minority carriers in the electric field transition layer are the minority carriers remaining in the electric field transition layer after the carriers in the collector region are injected into the electric field transition layer during the shutdown process of the IGBT device.
[0096] The IGBT device of the embodiment of the present application has the following characteristics:
[0097] Firstly, the interface position between the electric field transition layer and the drift region is specially designed by the inventors. The interface between the electric field transition layer and the drift region is the lower surface of the depletion region formed in the turn-off phase of the IGBT device.
[0098] During the IGBT device shutdown process, as the voltage rises, the depletion region will expand, causing the electric field formed in the depletion region to widen. The depletion region is located in the drift region and does not enter the electric field transition layer, so that during the IGBT device shutdown phase, the edge of the formed depletion region is the interface between the drift region and the electric field transition layer, and the minority carriers in the drift region are depleted.
[0099] Since the interface between the electric field transition layer and the drift region is the lower surface of the depletion region formed during the IGBT device turn-off phase, the minority carriers in the electric field transition layer will not be affected by the expansion of the depletion region, and the drift region will not affect the minority carriers in the electric field transition layer.
[0100] Secondly, the thickness of the electric field transition layer 3 and the doping concentration of the substrate 1 as the collector region are also specially designed by the inventor. The design requirements for the thickness of the electric field transition layer 3 and the doping concentration of the substrate 1 as the collector region are that the average gradient K of the energy band occupied by the minority carriers of the electric field transition layer 3 It is precisely controlled, that is, the morphology of the energy band occupied by the minority carriers in the electric field transition layer is controlled. Among them, the minority carriers in the electric field transition layer refer to the minority carriers that remain in the electric field transition layer after the carriers in the collector region are injected into the electric field transition layer during the IGBT device shutdown process. Because the minority carriers in the electric field transition layer itself account for a small proportion, they can be ignored.
[0101] In the semiconductor energy band theory system, the depletion rate of minority carriers has a corresponding relationship with the morphology of the energy band occupied by the minority carriers. 3 The precise control of the energy band occupied by the minority carriers in the electric field transition layer can achieve precise control of the morphology. This can control the removal rate of the minority carriers in the electric field transition layer, and further control the duration of the tail current, making the duration of the tail current shorter, thereby reducing the turn-off loss.
[0102] In implementation, the average gradient K of the energy band occupied by the minority carriers in the electric field transition layer is 3 The value range is greater than or equal to 50eV / cm and less than or equal to 500eV / cm.
[0103] After simulation, the average gradient K of the energy band occupied by the minority carriers in the electric field transition layer is 3 The value range can achieve a tail current duration of less than 100ns.
[0104] In implementation, the average gradient K of the energy band occupied by the minority carriers in the electric field transition layer is 3 , the thickness of the electric field transition layer, and the doping concentration of the substrate 1 as the collector region satisfy a preset relationship, so that the thickness of the electric field transition layer and the doping concentration of the substrate 1 as the collector region are affected by the average gradient K of the energy band occupied by the minority carriers of the electric field transition layer. 3 limitation.
[0105] The inventors of this application define the average gradient K of the energy band occupied by the minority carriers in the electric field transition layer in the semiconductor energy band theory system as 3 A relationship is established between the thickness of the electric field transition layer and the doping concentration of the substrate 1 as the collector region as a preset relationship. The average gradient K of the energy band occupied by the minority carriers in the electric field transition layer in the semiconductor energy band theory is 3This relatively abstract parameter establishes a connection with the specific and intuitive thickness of the electric field transition layer and the doping concentration of the substrate 1 as the collector area, providing conditions for the process implementation.
[0106] The following describes the N-channel IGBT device and the P-channel IGBT device respectively:
[0107] In the first implementation mode, when the IGBT device is an N-channel IGBT device, the electric field transition layer is N-type, the minority carriers of the N-type electric field transition layer are valence band holes, and the energy band occupied by the minority carriers of the N-type electric field transition layer is the valence band; correspondingly:
[0108] The average gradient K of the energy band occupied by the minority carriers in the N-type electric field transition layer 3 is the average gradient ΔEv / dx of the valence band of the electric field transition layer; the average gradient ΔEv / dx of the valence band of the electric field transition layer satisfies the following relationship:
[0109]
[0110] The above relationship is transformed into:
[0111] Wherein, ΔEv is the valence band change of the electric field transition layer, the x-axis is the coordinate axis of the depth direction of the N-type electric field transition layer and the positive direction of the x-axis is the direction in which the depth of the electric field transition layer increases, dx indicates that the integral variable is x, q is the basic charge or elementary charge, K is the Boltzmann constant, and L 空穴 is the hole diffusion length, W N is the thickness of the N-type electric field transition layer, N Collector is the doping concentration of the substrate 1 as the collector region, ε si is the dielectric constant of the Si substrate.
[0112] The inventors of the present application have determined the average gradient K of the energy band occupied by the minority carriers in the electric field transition layer. 3 The technical concept and theoretical derivation that are precisely controlled are applied to N-channel IGBT devices. The average gradient K of the energy band occupied by minority carriers in the N-type electric field transition layer is 3 is the average gradient ΔEv / dx of the valence band of the electric field transition layer.
[0113] In practical application, this is achieved through the following steps:
[0114] First, the average gradient K of the energy band occupied by minority carriers in the electric field transition layer 3 The value range is greater than or equal to 50 eV / cm and less than or equal to 500 eV / cm. First, determine the specific value of the average gradient ΔEv / dx of the valence band of an electric field transition layer;
[0115] After that, under the constraints of the above preset relationship, there are two ways:
[0116] The first method: combined with the thickness W of the electric field transition layer N In the case of Collector Finally, by the thickness W of the electric field transition layer N Based on the doping concentration of the substrate 1 as the collector area, process biasing and simulation are performed to select the IGBT device with the best effect.
[0117] The second method: combining the doping concentration N of the substrate 1 as the collector region Collector The best matching thickness W of the electric field transition layer is obtained. N Finally, by the thickness W of the electric field transition layer N Based on the doping concentration of the substrate 1 as the collector area, process biasing and simulation are performed to select the IGBT device with the best effect.
[0118] Specifically, the value range of the average gradient ΔEv / dx of the valence band of the electric field transition layer is greater than or equal to 50 eV / cm and less than or equal to 500 eV / cm.
[0119] Through simulation, the value range of the average gradient ΔEv / dx of the valence band of the electric field transition layer can achieve a tail current duration of less than 100 ns.
[0120] The average gradient K of the energy band occupied by the minority carriers in the electric field transition layer 3 (The first implementation corresponds to the average gradient ΔEv / dx of the valence band of the electric field transition layer) In the range of values, the larger the value, the shorter the duration of the tail current. Because, for the manufacturing process of different IGBT devices, not only the tail current but also the comprehensive performance of the IGBT device needs to be considered. Therefore, a more convenient method is to find the K value suitable for the IGBT device through simulation. 3 The value of, and the corresponding doping concentration and thickness of the electric field transition layer.
[0121] In summary, the device structure of the IGBT device of the present application is simple; and the present application reduces the average gradient K of the energy band occupied by the minority carriers of the electric field transition layer. 3 (The first implementation corresponds to the precise control of the average gradient ΔEv / dx of the valence band of the electric field transition layer) to achieve the control of the removal rate of excess carriers of the electric field transition layer, and thus to achieve the control of the duration of the tail current. In actual operation, the average gradient K of the energy band occupied by the minority carriers of the electric field transition layer 3The precise control is achieved by controlling the thickness of the electric field transition layer and the doping concentration of the substrate 1 as the collector region. It is easy to implement at the actual process level and is suitable for mass production.
[0122] In the prior art, in order to reduce the size of the tail current of the IGBT device, the method adopted is to reduce the number of minority carriers, that is, to control the number of minority carriers generated in the electric field transition layer.
[0123] The technical concept of the present application is different. The technical concept of the present application is to remove the minority carriers in the electric field transition layer from the electric field transition layer at a faster moving speed and move them into the electric field cutoff layer. In the present application, the thickness of the N-type electric field transition layer and the doping concentration of the substrate as the collector region are limited by the average gradient ΔEv / dx of the valence band of the electric field transition layer, that is, the removal speed of the minority carriers in the electric field transition layer is controlled, so that the removal speed is faster.
[0124] As a first preferred embodiment, the average gradient ΔEv / dx of the valence band of the electric field transition layer is 50.00 eV / cm; correspondingly:
[0125] After simulation, the average gradient ΔEv / dx of the valence band of the electric field transition layer can achieve a tail current duration of less than 87ns and a turn-off loss of 1.05mJ.
[0126] As a second preferred embodiment, the average gradient ΔEv / dx of the valence band of the electric field transition layer is 162.00 eV / cm; correspondingly:
[0127] It should be noted that when the average gradient ΔEv / dx of the valence band of the electric field transition layer is 162.00 eV / cm, there may be multiple sets of values for the thickness of the electric field transition layer 3 and the doping concentration of the substrate 1 as the collector region.
[0128] After simulation, the value of the average gradient ΔEv / dx of the valence band of the electric field transition layer can achieve a tail current duration of less than 64ns and a turn-off loss of 1.01mJ.
[0129] As a third preferred embodiment, the average gradient ΔEv / dx of the valence band of the electric field transition layer is 346.00 eV / cm; correspondingly:
[0130] It should be noted that when the average gradient ΔEv / dx of the valence band of the electric field transition layer is 346.00 eV / cm, there may be multiple sets of values for the thickness of the electric field transition layer 3 and the doping concentration of the substrate 1 as the collector region.
[0131] After simulation, the value of the average gradient ΔEv / dx of the valence band of the above-mentioned electric field transition layer can achieve a tail current duration of less than 46ns and a turn-off loss of 0.96mJ. The value of the average gradient ΔEv / dx of the valence band of the electric field transition layer is the same, and the tail current duration is not much different. However, the impact of other performances of IGBT devices may be different. Therefore, IGBT devices need to consider a variety of factors to determine the value of the average gradient ΔEv / dx of the valence band of the electric field transition layer, the thickness of the electric field transition layer, and the doping concentration of the substrate 1 as the collector region.
[0132] As a fourth preferred embodiment, the average gradient ΔEv / dx of the valence band of the electric field transition layer is 500.00 eV / cm; correspondingly:
[0133] After simulation, the average gradient ΔEv / dx of the valence band of the electric field transition layer can achieve a tail current duration of less than 39ns and a turn-off loss of 0.9mJ.
[0134] In the second implementation mode, when the IGBT device is a P-channel IGBT device, the electric field transition layer is of P type, the minority carriers of the P-type electric field transition layer are conduction band electrons, and the energy band occupied by the minority carriers of the P-type electric field transition layer is the conduction band; correspondingly:
[0135] The average gradient K of the energy band occupied by the minority carriers in the P-type electric field transition layer 3 The average gradient ΔEc / dx of the conduction band of the P-type electric field transition layer; the average gradient ΔEc / dx of the conduction band of the electric field transition layer satisfies the following relationship:
[0136]
[0137] Wherein, ΔEc is the conduction band change of the electric field transition layer, the x-axis is the coordinate axis of the depth direction of the P-type electric field transition layer and the positive direction of the x-axis is the direction of increasing depth from the electric field transition layer, dx indicates that the integral variable is x, q is the basic charge or elementary charge, K is the Boltzmann constant, and L 电子 is the electron diffusion length, W P is the thickness of the P-type electric field transition layer, N Collector is the doping concentration of the substrate 1 as the collector region, ε si is the dielectric constant of the Si substrate.
[0138] Specifically, the value range of the average gradient ΔEc / dx of the conduction band of the electric field transition layer is greater than or equal to 50 eV / cm and less than or equal to 500 eV / cm.
[0139] In implementation, the upper surface of the depletion region formed in the turn-off stage of the IGBT device is the plane where the lower end surface of the column region 5 in the drift region 4 is located.
[0140] In this way, the best balance between device withstand voltage and loss can be achieved.
[0141] In practice, the average gradient K of the energy band occupied by the minority carriers of the electric field cutoff layer 2 is 2 , the average gradient K of the energy band occupied by the minority carriers in the electric field transition layer 3 The following relations are satisfied:
[0142] K 3 is a positive value, K 2 is a negative value; and K 2 The absolute value range is greater than or equal to 2×K 3 Less than or equal to 10×K 3 ;
[0143] The minority carriers of the electric field stop layer 2 are the minority carriers remaining in the electric field stop layer after the carriers in the collector region are injected into the electric field stop layer during the shutdown process of the IGBT device.
[0144] K 2 The absolute value is much higher than K3, which indicates that the IGBT device has better pressure-bearing capacity.
[0145] In an implementation, the doping concentration of the drift region is less than the doping concentration of the electric field transition layer and less than the doping concentration of the electric field stop layer.
[0146] The doping concentration of the drift region is set according to the pressure bearing requirements of the drift region of the IGBT device, and the doping concentration of the electric field cutoff layer is set according to the requirements of the electric field cutoff of the IGBT device. In this application, the small effect of the doping concentration of the electric field transition layer on the tail current of the IGBT device is ignored. Therefore, the doping concentration of the electric field transition layer is greater than the doping concentration of the drift region and less than the doping concentration of the electric field cutoff layer, so as to achieve a relatively smooth transition.
[0147] In implementation, the doping of the electric field stop layer 2 may be uniform doping or non-uniform doping;
[0148] The thickness of the electric field cutoff layer 2 is in the range of 1 micron or more and 3 microns or less, and the doping concentration of the electric field cutoff layer 2 is in the range of 1×10 15 / cm 3 Less than or equal to 5×10 17 / cm 3 ;
[0149] The doping of the drift region 4 may be uniform doping or non-uniform doping;
[0150] The doping concentration of the drift region 4 is in the range of greater than or equal to 1×10 13 / cm 3Less than or equal to 1×10 15 / cm 3 .
[0151] The inventor of the present application has conducted a process verification simulation based on TCAD software on the IGBT device of the present application:
[0152] The IGBT device of this application based on TCAD software and Fig.10 The process simulation results of the IGBT device shown in Fig.17 As shown: the red dotted line ① represents the curve of the voltage VCE between the collector and the emitter of the traditional structure, and the red solid line ③ represents the curve of the collector current IC of the traditional structure; the green dotted line ② represents the curve of the voltage VCE between the collector and the emitter of the IGBT device of the present application, and the green solid line ④ represents the curve of the collector current IC of the IGBT device of the present application.
[0153] It can be seen from the simulation results that the tail current of the IGBT device of the present application is significantly improved, and the current fall time is reduced from 100ns to 46ns; correspondingly, the turn-off loss is reduced by 0.12mJ (1.08mJ reduced to 0.96mJ), and the turn-off performance of the IGBT device of the present application is significantly improved.
[0154] The inventor of the present application also conducted experimental verification of the device performance of the IGBT device of the present application:
[0155] like Fig.18 As shown, the red dotted line ① represents the curve of the voltage VCE between the collector and the emitter of the traditional structure, and the red solid line ③ represents the curve of the collector current IC of the traditional structure; the green dotted line ② represents the curve of the voltage VCE between the collector and the emitter of the IGBT device of the present application, and the green solid line ④ represents the curve of the collector current IC of the IGBT device of the present application.
[0156] Experimental data show that the current tail of the 650V IGBT device of the present application is shortened by 49ns (92ns reduced to 43ns), the corresponding turn-off loss is reduced by 0.21mJ (0.91mJ reduced to 0.70mJ), and the performance improvement reaches 23%.
[0157] The forward on-state voltage drop of the 650V IGBT device of the present application is 1.65V.
[0158] Under the conditions of 400V / 75A, the performance figure of merit FoM of the 650V IGBT device of the present application is 1.16V*mJ.
[0159] Correspondingly, under the condition of 96A high current, the FoM of the 650V IGBT device of the present application is only 1.48V*mJ.
[0160] Compared with the latest research results of other teams (FoM is 2.35V*mJ), the overall performance of the device has been improved by more than 37%, and is better than the performance figure of 2.44V*mJ of Infineon's latest H7 series products of the same specification.
[0161] Obviously, the above experimental results fully verify the advantages of the IGBT device of the present application. The present application can effectively improve the turn-off performance of the device, thereby improving the high-frequency performance of the device, so that the device can meet the needs of high-frequency applications.
[0162] The IGBT device of the present application also includes:
[0163] like Figure 14-A As shown, the collector metal 6 is formed on the back side of the substrate 1;
[0164] A body region 71 of a first doping type is formed on the drift region 4;
[0165] Trench gate 72;
[0166] An emitter 73 of a second doping type;
[0167] The ohmic contact region 74 of the first doping type is of the first doping type.
[0168] Specifically, the first doping type is P type, and the second doping type is N type; or, the first doping type is N type, and the second doping type is P type.
[0169] Embodiment 2
[0170] like Figure 14-A As shown, the IGBT device of the present application includes:
[0171] A substrate 1 of a first doping type, serving as a collector region;
[0172] An electric field stop layer 2 of a second doping type, formed on the substrate 1;
[0173] An electric field transition layer 3 of a second doping type is formed on the electric field stop layer 2;
[0174] A drift region 4 of a second doping type, formed on the electric field transition layer 3;
[0175] A column region 5 of a first doping type is formed in the drift region 4, wherein a lower end surface of the column region 5 is spaced from a lower surface of the drift region 4; and an upper end surface of the column region 5 is spaced from an upper surface of the drift region 4;
[0176] Wherein, the interface between the electric field transition layer and the drift region is the lower surface of the depletion region formed in the turn-off phase of the IGBT device;
[0177] The thickness W of the electric field transition layer and the doping concentration N of the substrate 1 as the collector region Collector The following relationship is satisfied:
[0178]
[0179] The IGBT device of the embodiment of the present application has the following characteristics:
[0180] Firstly, the interface position between the electric field transition layer and the drift region is specially designed by the inventors. The interface between the electric field transition layer and the drift region is the lower surface of the depletion region formed in the turn-off phase of the IGBT device.
[0181] Secondly, the thickness of the electric field transition layer 3 and the doping concentration of the substrate 1 as the collector region are also specially designed by the inventors. The average gradient K of the energy band occupied by the minority carriers in the electric field transition layer is 3 Under the technical concept of precise control, combined with theoretical derivation and simulation, the thickness W of the electric field transition layer and the doping concentration N of the substrate 1 as the collector region are obtained. Collector Requirements:
[0182] The removal speed of minority carriers in the electric field transition layer that meets the above requirements is faster, so that the duration of the tail current is shorter and the turn-off loss is smaller.
[0183] The device structure of the IGBT device of the present application is simple, and the average gradient K of the energy band occupied by the minority carriers in the electric field transition layer is 3 The removal rate of the minority carriers in the electric field transition layer is precisely controlled. In actual operation, the thickness of the electric field transition layer and the doping concentration of the substrate 1 as the collector region are controlled by the average gradient K of the energy band occupied by the minority carriers in the electric field transition layer. 3 The actual process is simple to implement and is suitable for mass production.
[0184] In summary, the IGBT device of the embodiment of the present application not only has a faster removal rate of minority carriers in the electric field transition layer, which shortens the duration of the tail current and reduces the turn-off loss; it is also easy to implement at the actual process level and is suitable for mass production.
[0185] As the first preferred method,
[0186] After simulation, it can be achieved that the tail current duration is less than 87ns and the turn-off loss is 1.05mJ.
[0187] As a second preferred method,
[0188] After simulation, the value of the average gradient ΔEv / dx of the valence band of the electric field transition layer can achieve a tail current duration of less than 64ns and a turn-off loss of 1.01mJ.
[0189] As a third preferred method,
[0190] After simulation, it can be achieved that the tail current duration is less than 46ns and the turn-off loss is 0.96mJ.
[0191] As a fourth preferred method,
[0192] After simulation, it can be achieved that the tail current duration is less than 39ns and the turn-off loss is 0.9mJ.
[0193] From the above four optimal methods, we can conclude that:
[0194] After simulation, it can be achieved that the tail current duration is between 39ns and 87ns, and the turn-off loss is between 0.9mJ and 1.05mJ.
[0195] Based on the above four optimal methods, we can also conclude that:
[0196] The tail current duration can be between 46ns and 64ns, and the turn-off loss can be between 0.96mJ and 1.01mJ.
[0197] In an implementation, the doping concentration of the drift region is less than the doping concentration of the electric field transition layer and less than the doping concentration of the electric field stop layer.
[0198] The doping concentration of the drift region is set according to the pressure bearing requirement of the drift region of the IGBT device, and the doping concentration of the electric field cutoff layer is set according to the electric field cutoff requirement of the IGBT device.
[0199] In implementation, the doping of the electric field stop layer 2 may be uniform doping or non-uniform doping;
[0200] The thickness of the electric field cutoff layer 2 is in the range of 1 micron or more and 2 microns or less, and the doping concentration of the electric field cutoff layer 2 is in the range of 1×10 15 / cm 3 Less than or equal to 5×10 17 / cm 3 ;
[0201] The doping of the drift region 4 may be uniform doping or non-uniform doping;
[0202] The doping concentration of the drift region 4 is in the range of greater than or equal to 1×10 13 / cm 3 Less than or equal to 1×10 15 / cm 3 .
[0203] In implementation, the upper surface of the depletion region formed in the turn-off stage of the IGBT device is the plane where the lower end surface of the column region 5 in the drift region 4 is located.
[0204] The inventors of this application define the average gradient K of the energy band occupied by the minority carriers in the electric field transition layer in the semiconductor energy band theory system as 3 A relationship is established between the thickness of the electric field transition layer and the doping concentration of the substrate 1 as the collector region as a preset relationship. The average gradient K of the energy band occupied by the minority carriers in the electric field transition layer in the semiconductor energy band theory is 3 This relatively abstract parameter establishes a connection with the specific and intuitive thickness of the electric field transition layer and the doping concentration of the substrate 1 as the collector area, providing conditions for the process implementation.
[0205] The following describes the N-channel IGBT device and the P-channel IGBT device respectively:
[0206] In the first implementation mode, when the IGBT device is an N-channel IGBT device, the electric field transition layer is N-type, the minority carriers of the N-type electric field transition layer are valence band holes, and the energy band occupied by the minority carriers of the N-type electric field transition layer is the valence band; correspondingly:
[0207] The average gradient K of the energy band occupied by the minority carriers in the N-type electric field transition layer 3 is the average gradient ΔEv / dx of the valence band of the electric field transition layer; the thickness of the N-type electric field transition layer, the doping concentration of the substrate 1 as the collector region, and the average gradient ΔEv / dx of the valence band of the electric field transition layer satisfy the following relationship:
[0208]
[0209] Among them, W N is the thickness of the N-type electric field transition layer, N Collector is the doping concentration of the substrate 1 as the collector region, q is the basic charge or elementary charge, K is the Boltzmann constant, and L 空穴 is the hole diffusion length, ΔEv is the valence band change of the electric field transition layer, the x-axis is the coordinate axis of the depth direction of the N-type electric field transition layer and the positive direction of the x-axis is the direction in which the depth of the electric field transition layer increases, dx indicates that the integral variable is x, ε si is the dielectric constant of the Si substrate, and q is the basic charge or elementary charge.
[0210] The inventors of the present application have determined the average gradient K of the energy band occupied by the minority carriers in the electric field transition layer. 3The technical concept and theoretical derivation that are precisely controlled are applied to N-channel IGBT devices. The average gradient K of the energy band occupied by minority carriers in the N-type electric field transition layer is 3 is the average gradient ΔEv / dx of the valence band of the electric field transition layer.
[0211] In the second implementation mode, when the IGBT device is a P-channel IGBT device, the electric field transition layer is of P type, the minority carriers of the P-type electric field transition layer are conduction band electrons, and the energy band occupied by the minority carriers of the P-type electric field transition layer is the conduction band; correspondingly:
[0212] The average gradient K of the energy band occupied by the minority carriers in the P-type electric field transition layer 3 is the average gradient ΔEc / dx of the conduction band of the P-type electric field transition layer; the thickness of the P-type electric field transition layer, the doping concentration of the substrate 1 as the collector region, and the average gradient ΔEc / dx of the conduction band of the electric field transition layer satisfy the following relationship:
[0213]
[0214] Among them, W P is the thickness of the P-type electric field transition layer, N Collector is the doping concentration of the substrate 1 as the collector region, q is the basic charge or elementary charge, K is the Boltzmann constant, and L 电子 is the electron diffusion length, ΔEc is the conduction band change of the electric field transition layer, the x-axis is the coordinate axis of the depth direction of the P-type electric field transition layer and the positive direction of the x-axis is the direction of increasing depth from the electric field transition layer, dx indicates that the integral variable is x, ε si is the dielectric constant of the Si substrate, and q is the basic charge or elementary charge.
[0215] In practice, the average gradient K of the energy band occupied by the minority carriers of the electric field cutoff layer 2 is 2 , the average gradient K of the energy band occupied by the minority carriers in the electric field transition layer 3 The following relations are satisfied:
[0216] K 3 is a positive value, K 2 is a negative value; and K 2 The absolute value range is greater than or equal to 2×K 3 Less than or equal to 10×K 3 .
[0217] The minority carriers of the electric field stop layer 2 are the minority carriers remaining in the electric field stop layer after the carriers in the collector region are injected into the electric field stop layer during the shutdown process of the IGBT device.
[0218] K 2The absolute value is much higher than K3, which indicates that the IGBT device has better pressure-bearing capacity.
[0219] The research process and theoretical derivation of this application are as follows:
[0220] As can be seen from the background technology, Super IGBT devices have gone through three stages of research. However, Super IGBT devices still have obvious tail current, which limits the performance and application promotion of Super IGBT devices.
[0221] In order to solve the above tail current problem, the inventor of this application starts from the device physics and focuses on analyzing the physical mechanism of the tail current of the SuperIGBT device, thereby proposing a targeted solution to fully tap the technical potential of the Super IGBT device and complete technical verification and process development from the actual process level.
[0222] The inventors of this application analyze the semiconductor band theory as follows:
[0223] In this study, trench-type Super IGBT devices are taken as the research object. Compared with traditional FS IGBT devices, SuperIGBT devices introduce a P-pillar (i.e., P-type pillar region) structure in the drift region to achieve super junction structure design.
[0224] In order to effectively suppress the problem of forward unipolar conduction, a structure in which the P-pillar and P-body are not connected is adopted, such as Fig.10 As shown. P-type substrate 10-1, N-type drift region 10-2, P-type pillar region 10-3, P-type body region 10-4, back metal 10-5. P-pillar is Fig.10 The P-type column region 10-3; the P-body is the P-type body region 10-4.
[0225] IGBT devices are mainly used in inductive load circuits. During the shutdown process of Super IGBT devices, the inductive load topology makes the state switching of IGBT devices include two processes: the voltage rising stage (process "I") and the current falling stage (process "II"), such as Fig.11 For high-voltage and high-power IGBT devices, the current tailing during the current drop phase is the key to the device performance and directly affects the magnitude of the device turn-off loss. This study will also focus on the turn-off current tailing phase of Super IGBT devices.
[0226] The Super IGBT device includes a P-channel Super IGBT device and an N-channel Super IGBT device. Fig.10It is an N-channel Super IGBT device. The following description will take the N-channel Super IGBT device as the research object.
[0227] against Fig.10 The two-dimensional structure of the Super IGBT device is shown in Figure 2. The 3D energy band characteristics of the current tailing stage (process “II”) are shown in Figure 2. Figure 12-A , Figure 12-B , Figure 12-C , Figure 12-D As shown. Figure 12-A , Figure 12-B , Figure 12-C , Figure 12-D In the energy band diagram, the X coordinate represents the lateral size of the Super IGBT device (i.e. Fig.10 The Y coordinate represents the longitudinal dimension of the SuperIGBT device (i.e. Fig.10 The Z coordinate represents the electronic energy level corresponding to the energy band diagram of the Super IGBT device. Obviously, as the depth from the bottom of the device increases, the energy levels of the conduction band (Ec), electron quasi-Fermi level (Efn), hole quasi-Fermi level (Efp) and valence band (Ev) inside the device increase rapidly, and a high carrier barrier appears inside the device, which prevents minority carriers in the bottom area of the device from freely entering the space electric field area.
[0228] In order to analyze the internal energy band characteristics of Super IGBT in more detail, the two-dimensional energy band diagram at the Y=0 position is extracted, as shown in Figure 13-A As shown in the figure, Band Diagram is an energy band diagram. According to the energy band characteristics at different locations inside the device, the device can be divided into four regions: P-pillar region (region A), N-type depletion region (region B), N-type transition region (region C), and N-type field cutoff region (region D). Due to the strong electric field in the P-pillar region, the energy band diagram shows an obvious gradient distribution. In order to focus on analyzing the energy band characteristics of the transition region, the energy band diagram of the corresponding region is partially enlarged, as shown in the figure below. Figure 13-B shown.
[0229] like Figure 13-B As shown, it is obvious that with the increase of the depth from the back of the device, the position of the electron quasi-Fermi level (Efn) remains basically unchanged, the conduction band (Ec) shows an upward trend, and the corresponding electron concentration decreases rapidly. Differently, with the increase of the depth from the back of the device, both the valence band (Ev) and the hole quasi-Fermi level (Efp) show an upward trend, and the change of hole concentration needs to be considered comprehensively. In particular, for the area in the dotted rectangular box, there is a "potential barrier-like extreme zone" at the position shown by point A of the valence band (Ev), and the hole enrichment effect will appear in this area. Under the action of the spatial electric field, the holes in the above-mentioned area are gradually extracted, thereby generating a tail current.
[0230] For the turn-off characteristics of IGBT devices, its tail current is mainly related to the excess minority carriers (holes) remaining in the above transition region. To achieve a smaller tail current, on the one hand, it is necessary to reduce the total amount of excess minority carriers remaining in the transition region, and on the other hand, it is necessary to accelerate the removal speed of excess minority carriers.
[0231] In traditional IGBT devices (such as Fig.10 In the IGBT device shown in the figure, the electric field in the transition zone is hardly changed by the increase of external voltage and remains almost zero. The energy band in the transition zone remains basically flat, and the drift speed of the carriers, conduction band electrons and valence band holes, is basically close to zero, and the excess carriers become "difficult to drift carriers".
[0232] So far, the reason why traditional IGBT devices have a large tail current during the shutdown process has been analyzed.
[0233] Fig.10 The physical characteristics of the IGBT device (trench-type Super IGBT device) during the turn-off process are analyzed as follows:
[0234] The typical waveform of a trench type Super IGBT device during the turn-off process is as follows: Fig.15 As shown, the red curve ① represents the Super IGBT collector current, and the green curve ② represents the Super IGBT collector voltage.
[0235] In order to observe the device turn-off process in more detail, the turn-off process is divided into 8 time periods, corresponding to the time from t0 to t8. Obviously, there is a significant current tail in the time period from t4 to t8. If the duration of the above tail current can be reduced, the turn-off loss of the Super IGBT can be effectively reduced.
[0236] against Fig.10 The IGBT device turn-off process shown in Figure 1 shows the current distribution inside the device at 9 critical moments. Figure 16-A and Figure 16-B As shown, the current in the time period t4-t8 is mainly hole current, and the source of the hole current is the excess holes below the P-Pillar.
[0237] When the internal electric field of the device extends to the P-pillar, the electric field will no longer expand. At this time, the hole depletion below the P-pillar can only be achieved through the electron-hole recombination mechanism.
[0238] In order to effectively reduce the above-mentioned current tail, the design of the transition zone below the P-Pillar is particularly critical to the performance improvement of the Super IGBT.
[0239] The structural design ideas of the IGBT device of this application are as follows:
[0240] Therefore, when actually designing IGBT devices in this application, accelerating the depletion of "difficult-to-drift carriers" can achieve the goal of optimizing the tail current of IGBT devices. For N-type (i.e., N-channel) IGBT devices, "difficult-to-drift carriers" mainly refer to the holes remaining in the transition zone after the holes in the collector region are injected into the transition zone during the shutdown process of the IGBT device, that is, the holes are the majority carriers for the collector region, and the holes are the minority carriers for the transition zone. "Difficult-to-drift carriers" also include holes in the transition zone itself, which account for a very small proportion and can be ignored.
[0241] In addition, for P-type (i.e. P-channel) IGBT devices, "difficult-to-drift carriers" mainly refer to the electrons remaining in the transition region after the electrons in the collector region are injected into the transition region during the IGBT device shutdown process, that is, the electrons are majority carriers for the collector region and minority carriers for the transition region. "Difficult-to-drift carriers" also include electrons in the transition region itself, which account for a very small proportion and can be ignored.
[0242] That is, in the design concept of the present application, the minority carriers in the transition region are the minority carriers remaining in the transition region after the carriers in the collector region are injected into the transition region during the shutdown process of the IGBT device.
[0243] The inventor of this application has studied the semiconductor theory system. In the semiconductor theory system, the hole depletion rate and the morphology of the valence band have a strict corresponding relationship. Figure 13-B In the energy band diagram, different valence band positions represent different minority carrier (valence band hole) energies. For valence band holes, the higher the valence band position, the lower their energy; that is, Figure 13-B In the valence band curve Ev, the higher the position, the lower the energy. Holes near the valence band will move to low-energy positions, and the greater the spatial gradient of the valence band, the more obvious the above movement trend will be. Therefore, in order to accelerate the minority carrier depletion process in the transition zone, a feasible method is to increase the gradient distribution of the valence band in the transition zone, so that the "difficult-to-drift holes" in the transition zone can be removed by self-acceleration.
[0244] Therefore, it can be seen that in order to improve the tail current problem during the turn-off process of IGBT devices, the control of the energy band gradient, especially the valence band gradient, in the area below the P-Pillar (i.e., the transition zone) is the core. Related work has not been reported in previous studies, or quantitative research and analysis has been conducted.
[0245] Theoretically, in order to accelerate the depletion process of "difficult-to-drift holes" in the N-type transition region, the average gradient ΔEv / dx of the valence band in the transition region needs to be as large as possible to ensure that the IGBT device does not generate obvious tail current, thereby achieving turn-off energy loss optimization.
[0246] For N-type (i.e. N-channel) IGBT devices, "difficult-to-drift carriers" mainly refer to the holes that remain in the transition region after the holes in the collector region are injected into the transition region, that is, the holes are majority carriers for the collector region, and the holes are minority carriers for the transition region. Correspondingly, as the average gradient ΔEv / dx of the valence band in the transition region increases, the removal rate of "difficult-to-drift holes" in the transition region is also greater, making the duration of the tail current generated by the excess holes shorter.
[0247] Similarly, for P-type (i.e. P-channel) IGBT devices, "difficult-to-drift carriers" mainly refer to the electrons remaining in the transition region after the collector region electrons are injected into the transition region, that is, the electrons are majority carriers for the collector region, and the electrons are minority carriers for the transition region. Correspondingly, as the average gradient ΔEc / dx of the conduction band in the transition region increases, the removal speed of the "difficult-to-drift electrons" in the transition region is also faster, making the duration of the tail current generated by the excess electrons shorter.
[0248] According to the energy band theory of semiconductors:
[0249] For an N-channel IGBT device, the average gradient ΔEv / dx of the valence band in the transition region is defined as
[0250]
[0251] The average gradient ΔEv / dx of the valence band in the transition region is an abstract parameter. The inventors realized that it is necessary to make this abstract parameter concrete and intuitive. Therefore, after a simple deduction,
[0252] Where ΔEv is the valence band change in the transition zone, the x-axis is the coordinate axis in the thickness direction of the transition zone and the positive direction of the x-axis is the direction in which the depth of the transition zone increases, dx indicates that the variable of the integration is x, W N is the thickness of the N-type transition zone, q is the basic charge or elementary charge, and ψ is the electric potential.
[0253] Part 1: Based on Poisson's equation, the average electric field of the semiconductor can be derived. The derivation process is as follows:
[0254] Poisson's equation in semiconductor theory:
[0255]
[0256] Where ρ is the charge density, ε si is the dielectric constant of the Si substrate of the IGBT device, q is the basic charge or elementary charge, p is the hole concentration, n is the electron concentration, N D is the N-type doping concentration, N A is the P-type doping concentration and ψ is the electric potential.
[0257] According to Poisson's equation, it can be deduced that:
[0258]
[0259] In the above formula, the carriers in the transition region are mainly holes, so p is smaller than n and N. D 、N A It is several orders of magnitude larger, so it can be approximately considered that p-n+N D -N A If p is equal to p, the above formula is simplified to:
[0260] Corresponding to Fig.10 The transition zone structure of the IGBT device, the thickness of the transition zone is W N , the electric potential formula corresponds to:
[0261]
[0262] Due to the average electric field is the average value of the electric potential, that is:
[0263] Part 2: Corresponding to Fig.10 The transition region structure of the IGBT device, the thickness of the transition region is W N , the total amount of “difficult drift carriers” in the transition region Q s The expression is:
[0264]
[0265] It can be concluded that:
[0266]
[0267] Among them, N Collector is the doping concentration of the substrate 1 as the collector region, ρ is the charge density, k is the Boltzmann constant, and L 空穴 is the hole diffusion length.
[0268] The total amount of “difficult drift carriers” in the transition region is Q s These are the holes that remain in the transition region after the holes in the collector region are injected into the transition region.
[0269] Part 3: The average gradient ΔEv / dx of the valence band in the transition region is equal to the absolute value of the average electric field. Considering the uniformity of the physical units, the relationship is as follows:
[0270]
[0271] Afterwards, Bring in It turns out that:
[0272] Let’s deduce:
[0273]
[0274]
[0275] Right now
[0276] After deformation, we can get:
[0277] The inventor of the present application has further thought of setting the transition region and the drift region separately based on a lot of creative work, and the thickness of the transition region and the doping concentration of the collector region meet a preset relationship. That is, the drift region sets the doping concentration and thickness according to the pressure bearing capacity requirements; the transition region is no longer a part of the drift region, but a separate layer structure - the electric field transition layer, and the thickness of the electric field transition layer and the doping concentration of the collector region meet a certain relationship.
[0278] At this point, the core relationship of the present application is formed. When the IGBT device is an N-channel IGBT device, the electric field transition layer is N-type, the minority carriers of the N-type electric field transition layer are valence band holes, and the energy band occupied by the minority carriers of the N-type electric field transition layer is the valence band; correspondingly:
[0279] The average gradient K of the energy band occupied by the minority carriers in the N-type electric field transition layer 3 is the average gradient ΔEv / dx of the valence band of the N-type electric field transition layer; the average gradient ΔEv / dx of the valence band of the electric field transition layer satisfies the following relationship:
[0280]
[0281] Wherein, Ev is the valence band, the x-axis is the coordinate axis of the depth direction of the N-type electric field transition layer and the positive direction of the x-axis is the direction in which the depth of the electric field transition layer increases, dx indicates that the integral variable is x, W N is the thickness of the N-type electric field transition layer, N D is the doping concentration of the N-type electric field transition layer, ε si is the dielectric constant of the Si substrate, and q is the basic charge or elementary charge.
[0282] For a P-channel IGBT device, the corresponding:
[0283] The average gradient K of the energy band occupied by the minority carriers in the P-type electric field transition layer 3 The average gradient ΔEc / dx of the conduction band of the P-type electric field transition layer; the average gradient ΔEc / dx of the conduction band of the electric field transition layer satisfies the following relationship:
[0284]
[0285] Among them, L 电子 is the electron diffusion length, W P is the thickness of the P-type electric field transition layer.
[0286]
[0287] The range of the average gradient Ev / dx of the valence band of the electric field transition layer of the IGBT device of the present application is greater than or equal to 50 eV / cm and less than or equal to 500 eV / cm. Obviously, in formula (2), after selecting a value for the average gradient ΔEv / dx of the valence band of the field transition layer, only W N and N Collector are variables, and the constraints between them can be obtained directly.
[0288] In practical applications, we can first determine a specific value of a target (ΔEv / dx), and then determine the thickness W of the electric field transition layer under the constraints of the above formula: N and the collector region doping concentration N Collector The value of one of them can get the best matching value of the other parameter. Finally, by setting the thickness W of the transition zone N and the collector region doping concentration N Collector Based on the value of , process deviation and simulation are performed to select the IGBT device with the best effect.
[0289] like Figure 14-B As shown, compared with the traditional SJ IGBT structure, the average gradient ΔEv / dx of the valence band of the electric field transition layer of the IGBT device of the present application is increased by one order of magnitude (from the traditional 36.95eV / cm to 346.00eV / cm). The present application can significantly accelerate the removal speed of the "difficult-to-drift holes" in the transition region, thereby achieving a shorter duration of the tail current generated by the excess carriers.
[0290] Conclusion: From the background technology, we can know that Super IGBT devices have gone through three stages of research. However, Super IGBT devices still have obvious tail current, which limits the performance and application promotion of Super IGBT devices.
[0291] In order to solve the above-mentioned tail current problem, the inventor of this application starts from the device physics and focuses on analyzing the physical mechanism of the tail current of the SuperIGBT device, so as to propose a targeted solution. In order to fully tap the technical potential of the Super IGBT device, and complete the technical verification and process development from the actual process level. Starting from the physics of semiconductor devices, this application conducts performance analysis through physical analysis, TCAD simulation, experimental verification and other means, and finds that the tail current of the Super IGBT device mainly comes from the storage effect of excess minority carriers in the transition zone below the device P-pillar (i.e., the P-type column area). Among them, the transition zone is the part of the epitaxial layer located below the P-type column area and above the substrate.
[0292] In order to reduce the tail current problem, this application innovatively proposes the electric field transition layer design theory. TCAD simulation and test results show that compared with the latest research results of Super IGBT devices in the prior art, the 650V SuperIGBT device of this application has a 23% lower turn-off loss, and the performance figure of merit is improved to 1.48V*mJ. The device performance has reached the international leading level (better than the latest Infineon H7 product figure of merit 2.44V*mJ). The 650V Super IGBT device of this application has completed technical verification and process development at the actual process level, and has been able to reach the level of mass production, providing a foundation for subsequent widespread application.
[0293] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.
Claims
1. An IGBT device, characterized in that: include: A substrate (1) of a first doping type, serving as a collector region; An electric field stop layer (2) of a second doping type, formed on the substrate (1); An electric field transition layer (3) of a second doping type, formed on the electric field stop layer (2); A drift region (4) of a second doping type is formed on the electric field transition layer (3); A column region (5) of a first doping type is formed in the drift region (4), wherein a lower end surface of the column region (5) and a lower surface of the drift region (4) are spaced apart; and an upper end surface of the column region (5) and an upper surface of the drift region (4) are spaced apart; The thickness W of the electric field transition layer and the doping concentration N of the substrate (1) serving as the collector region are Collector The following relationship is satisfied:
2. The IGBT device according to claim 1, characterized in that: or, 3. The IGBT device according to claim 1, characterized in that: or or or 4. The IGBT device according to any one of claims 1 to 3, characterized in that: The doping concentration of the drift region is less than the doping concentration of the electric field transition layer and less than the doping concentration of the electric field stop layer.
5. The IGBT device according to any one of claims 1 to 3, characterized in that: When the IGBT device is an N-channel IGBT device, the electric field transition layer is N-type, the minority carriers of the N-type electric field transition layer are valence band holes, and the energy band occupied by the minority carriers of the N-type electric field transition layer is the valence band; correspondingly: The thickness of the N-type electric field transition layer, the doping concentration of the substrate (1) as the collector region, and the average gradient ΔEv / dx of the valence band of the N-type electric field transition layer satisfy the following relationship: Among them, W N is the thickness of the N-type electric field transition layer, N Collector is the doping concentration of the substrate (1) serving as the collector region, q is the basic charge or elementary charge, K is the Boltzmann constant, and L 空穴 is the hole diffusion length, ΔEv is the valence band change of the electric field transition layer, the x-axis is the coordinate axis of the depth direction of the N-type electric field transition layer and the positive direction of the x-axis is the direction in which the depth of the electric field transition layer increases, dx indicates that the integral variable is x, ε si is the dielectric constant of the Si substrate.
6. The IGBT device according to any one of claims 1 to 3, characterized in that: When the IGBT device is a P-channel IGBT device, the electric field transition layer is of P type, the minority carriers of the P-type electric field transition layer are conduction band electrons, and the energy band occupied by the minority carriers of the P-type electric field transition layer is the conduction band; correspondingly: The thickness of the P-type electric field transition layer, the doping concentration of the substrate (1) as the collector region, and the average gradient ΔEc / dx of the conduction band of the P-type electric field transition layer satisfy the following relationship: Among them, W P is the thickness of the P-type electric field transition layer, N Collector is the doping concentration of the substrate (1) serving as the collector region, q is the basic charge or elementary charge, K is the Boltzmann constant, and L 电子 is the electron diffusion length, ΔEc is the conduction band change of the electric field transition layer, the x-axis is the coordinate axis of the depth direction of the P-type electric field transition layer and the positive direction of the x-axis is the direction of increasing depth from the electric field transition layer, dx indicates that the integral variable is x, ε si is the dielectric constant of the Si substrate.
7. An IGBT device, characterized in that: include: A substrate (1) of a first doping type, serving as a collector region; An electric field stop layer (2) of a second doping type, formed on the substrate (1); An electric field transition layer (3) of a second doping type, formed on the electric field stop layer (2); A drift region (4) of a second doping type is formed on the electric field transition layer (3); A column region (5) of a first doping type is formed in the drift region (4), a lower end surface of the column region (5) and a lower surface of the drift region (4) are spaced apart, and an upper end surface of the column region (5) and an upper surface of the drift region (4) are spaced apart; The average gradient K3 of the energy band occupied by the minority carriers of the electric field transition layer is precisely controlled to control the removal speed of the minority carriers of the electric field transition layer; The minority carriers in the electric field transition layer are the minority carriers remaining in the electric field transition layer after the carriers in the collector region are injected into the electric field transition layer during the shutdown process of the IGBT device.
8. The IGBT device according to claim 7, characterized in that: The value range of the average gradient K3 of the energy band occupied by the minority carriers of the electric field transition layer is greater than or equal to 50 eV / cm and less than or equal to 500 eV / cm.
9. The IGBT device according to claim 8, characterized in that: The thickness of the electric field transition layer and the doping concentration of the substrate (1) as a collector region are limited by the average gradient K3 of the energy band occupied by the minority carriers of the electric field transition layer.
10. The IGBT device according to claim 9, characterized in that: When the IGBT device is an N-channel IGBT device, the electric field transition layer is N-type, the minority carriers of the N-type electric field transition layer are valence band holes, and the energy band occupied by the minority carriers of the N-type electric field transition layer is the valence band; correspondingly: The average gradient K3 of the energy band occupied by the minority carriers of the N-type electric field transition layer is the average gradient ΔEv / dx of the valence band of the N-type electric field transition layer; the average gradient ΔEv / dx of the valence band of the electric field transition layer satisfies the following relationship: Wherein, ΔEv is the valence band change of the electric field transition layer, the x-axis is the coordinate axis of the depth direction of the N-type electric field transition layer and the positive direction of the x-axis is the direction in which the depth of the electric field transition layer increases, dx indicates that the integral variable is x, q is the basic charge or elementary charge, K is the Boltzmann constant, and L 空穴 is the hole diffusion length, W N is the thickness of the N-type electric field transition layer, N Collector is the doping concentration of the substrate (1) serving as the collector region, ε si is the dielectric constant of the Si substrate.
11. The IGBT device according to claim 9, characterized in that: When the IGBT device is a P-channel IGBT device, the electric field transition layer is of P type, the minority carriers of the P-type electric field transition layer are conduction band electrons, and the energy band occupied by the minority carriers of the P-type electric field transition layer is the conduction band; correspondingly: The average gradient K3 of the energy band occupied by the minority carriers of the P-type electric field transition layer is the average gradient ΔEc / dx of the conduction band of the P-type electric field transition layer; the average gradient ΔEc / dx of the conduction band of the P-type electric field transition layer satisfies the following relationship: Wherein, ΔEc is the conduction band change of the electric field transition layer, the x-axis is the coordinate axis of the depth direction of the P-type electric field transition layer and the positive direction of the x-axis is the direction of increasing depth from the electric field transition layer, dx indicates that the integral variable is x, q is the basic charge or elementary charge, K is the Boltzmann constant, and L 电子 is the electron diffusion length, W P is the thickness of the P-type electric field transition layer, N Collector is the doping concentration of the substrate (1) serving as the collector region, ε si is the dielectric constant of the Si substrate.
12. The IGBT device according to any one of claims 9 to 11, wherein the average gradient K2 of the energy band occupied by the minority carriers of the electric field stop layer (2) and the average gradient K3 of the energy band occupied by the minority carriers of the electric field transition layer satisfy the following relationship: K3 is a positive value, K2 is a negative value; and the absolute value of K2 ranges from greater than or equal to 2×K3 to less than or equal to 10×K3; The minority carriers of the electric field stop layer (2) are the minority carriers remaining in the electric field stop layer after the carriers in the collector region are injected into the electric field stop layer during the shutdown process of the IGBT device.
13. The IGBT device according to any one of claims 8 to 11, characterized in that: in: The interface between the electric field transition layer (3) and the drift region (4) is the lower surface of the depletion region formed in the turn-off phase of the IGBT device; The upper surface of the depletion region formed in the turn-off phase of the IGBT device is the plane where the lower end surface of the column region (5) in the drift region (4) is located.
14. The IGBT device according to any one of claims 8 to 11, characterized in that: The doping concentration of the drift region is less than the doping concentration of the electric field transition layer and less than the doping concentration of the electric field stop layer.