Semiconductor device and manufacturing method for same

By strategically positioning the boundary between low and high carbon vacancy density regions in the drift region of IGBTs, the semiconductor device optimizes both on-resistance and turn-off loss, addressing the trade-off challenge in silicon carbide-based IGBTs.

WO2024241706A1PCT designated stage expired Publication Date: 2024-11-28HITACHI LTD
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
PCT/JP2024/013139
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2023-05-19
Filing Date
2024-03-29
Publication Date
2024-11-28

AI Technical Summary

Technical Problem

Existing semiconductor devices with insulated gate bipolar transistors (IGBTs) using silicon carbide face a trade-off between reducing on-resistance and turn-off loss, as extending minority carrier lifetime in the drift region increases on-resistance but decreases turn-off loss, and vice versa.

Method used

A semiconductor device with a drift region having a low carbon vacancy density region and a high carbon vacancy density region, where the boundary between the two is positioned deeper than the end of the depletion layer, allowing for controlled carbon vacancy distribution to optimize both on-resistance and turn-off loss.

Benefits of technology

This configuration effectively reduces both on-resistance and turn-off loss, improving the trade-off relationship between them by extending hole lifetime in the low carbon vacancy region and maintaining short hole lifetime in the high carbon vacancy region.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure JP2024013139_28112024_PF_FP_ABST
    Figure JP2024013139_28112024_PF_FP_ABST
Patent Text Reader

Abstract

The present invention provides a technology that enables improvement of a trade-off relationship of the reduction of on-resistance and the reduction of turn-off loss. On the assumption that a low-carbon vacancy density region 20 having a low carbon vacancy density is formed on the upper surface side of a drift region 4, while a high-carbon vacancy density region 30 having a high carbon vacancy density is formed on the lower surface side of the drift region 4, the position of an isodensity line 100, which is included in a boundary region 40 between the low-carbon vacancy density region 20 and the high-carbon vacancy density region 30 and where the carbon vacancy concentration corresponds to half the density of the carbon vacancy density of the high-carbon vacancy density region 30, is located deeper than an end portion 200 of a depletion layer extending to the drift region 4, when an IGBT is turned off.
Need to check novelty before this filing date? Find Prior Art

Description

Semiconductor device and manufacturing method thereof

[0001] The present invention relates to a semiconductor device and a manufacturing method thereof, and more particularly to a technology that is effective when applied to a semiconductor device having an insulated gate bipolar transistor and a manufacturing method thereof.

[0002] To improve the energy efficiency of power electronics devices, low-loss power semiconductor elements using wide-gap semiconductor materials such as silicon carbide (SiC), gallium nitride (GaN), and diamond are being researched. Silicon carbide and gallium nitride have breakdown field strengths approximately 10 times higher than silicon (Si), so for a power semiconductor element with the same breakdown voltage, the film thickness of the drift region can be reduced to one-tenth of that of silicon. By thinning the drift region in this way, the resistance value in the drift region can be significantly reduced, thereby lowering the on-resistance of the entire element.

[0003] Wide-gap semiconductor materials are applicable to unipolar elements such as Schottky barrier diodes (SBDs) and power metal oxide field effect transistors (MOSFETs), as well as bipolar elements such as pn junction diodes and insulated gate bipolar transistors (IGBTs). In particular, bipolar elements using silicon carbide are expected to achieve low conduction loss in ultra-high voltage applications exceeding 6.5 kV.

[0004] In silicon carbide, the minority carrier lifetime in the material after the epitaxial crystal layer has finished growing (as-grown state, as-grown means as-grown) is significantly shorter than that of silicon. Therefore, in order to improve the conductivity modulation effect in bipolar devices and reduce resistance during on-state operation, a process to increase the minority carrier lifetime in the drift region is carried out during device fabrication. On the other hand, if the minority carrier lifetime in the entire drift region is long, many residual carriers are discharged during the turn-off operation, when the device switches from the on state to the off state, resulting in large turn-off losses. Therefore, a "local lifetime control structure" in which only the minority carrier lifetime on the collector side of the drift region is locally short can reduce on-state voltage and turn-off losses.

[0005] Patent Document 1 describes a device in which the edge of a depletion layer formed in a drift region when an operating voltage is applied is located in a region where the minority carrier lifetime is short.

[0006] Furthermore, Patent Document 2 describes a method for forming a region for extending the minority carrier lifetime inside a silicon carbide epitaxial layer, in which the depth to which the carrier lifetime extension region reaches from the implantation surface of the silicon carbide epitaxial layer is arbitrarily changed by changing the annealing temperature or annealing time of annealing for diffusing and activating carbon atoms ion-implanted into the silicon carbide epitaxial layer, thereby making the thickness of the region thicker than other regions of the epitaxial layer.

[0007] Japanese Patent No. 3622405 Japanese Patent Application Laid-Open No. 2021-19157

[0008] However, in the above-mentioned Patent Document 1, there exists a region in the depletion layer where the minority carrier lifetime is short. Because there are no residual carriers in the depletion layer, even if the minority carrier lifetime in the depletion layer is short, it does not affect the suppression of turn-on loss. However, on the other hand, it causes an increase in on-resistance, which worsens the trade-off relationship between on-resistance and turn-off loss.

[0009] Furthermore, the above-mentioned Patent Document 2 does not disclose any description of the relationship between the carrier lifetime extension region and the depletion layer, and therefore, similar to the above-mentioned Patent Document 1, there is a possibility that this may cause a deterioration in the trade-off relationship between on-resistance and turn-off loss.

[0010] In view of the above problems, an object of the present invention is to provide a technique capable of improving the trade-off between reducing on-resistance and reducing turn-off loss.

[0011] In one embodiment, a semiconductor device includes an insulated gate bipolar transistor (IGBT), the insulated gate bipolar transistor having a drift region of a first conductivity type, the drift region including silicon carbide as a constituent material. The drift region includes a first region having a first carbon vacancy density, a second region having a second carbon vacancy density higher than the first density, and a boundary region sandwiched between the first and second regions, the boundary region having a higher carbon vacancy density than the first region and a lower carbon vacancy density than the second region. In this case, a position of an isopycnal included in the boundary region, where the isopycnal corresponds to a carbon vacancy concentration half the second density, is deeper than an end of a depletion layer extending into the drift region when the IGBT is turned off.

[0012] In one embodiment, a method for manufacturing a semiconductor device includes an insulated gate bipolar transistor in which an end of a depletion layer extending when the transistor is turned off remains within a drift region, and the method includes the steps of: (a) forming a drift region of a first conductivity type containing silicon carbide as a constituent material; (b) introducing carbon atoms into an upper surface of the drift region; and (c) diffusing the carbon atoms into the drift region.

[0013] After the step (c), the drift region has a first region having a carbon vacancy density of a first density, a second region having a carbon vacancy density of a second density higher than the first density, and a boundary region sandwiched between the first region and the second region and having a carbon vacancy density higher than that of the first region and lower than that of the second region. At this time, the position of the isopycnal included in the boundary region, which corresponds to a carbon vacancy concentration that is half the second density, is deeper than the end of the depletion layer that extends when the semiconductor device is turned off.

[0014] According to one embodiment, it is possible to achieve an improvement that achieves both a reduction in on-voltage and a reduction in turn-off loss, which are in a trade-off relationship with each other.

[0015] FIG. 1 is a diagram showing an active cell of an IGBT in an embodiment; FIG. 2 is a diagram showing a manufacturing process of a semiconductor device in an embodiment; FIG. 3 is a diagram showing a manufacturing process of a semiconductor device following FIG. 4; FIG. 5 is a diagram showing a manufacturing process of a semiconductor device following FIG. 6; FIG. 6 is a diagram showing a manufacturing method variation; FIG. 7 is a graph showing simulation results of changes in current and voltage over time at turn-off in each case where the lifetime of minority carriers in a high carbon vacancy density region, which is a "local lifetime control structure", is changed; FIG. 8 is a diagram showing structural variation 1; FIG. 9 is a diagram showing structural variation 2; FIG. 10 is a diagram showing structural variation 3;

[0016] In all the drawings for explaining the embodiments, the same components are generally designated by the same reference numerals, and repeated explanations thereof will be omitted. In addition, hatching may be used even in plan views to make the drawings easier to understand.

[0017] In this embodiment, a semiconductor device including an insulated gate bipolar transistor having an n-type channel structure in which the drift region and emitter region are made of n-type semiconductor regions and the body region is made of p-type semiconductor regions will be described as an example. However, it goes without saying that the technical idea of ​​this embodiment can also be applied to an insulated gate bipolar transistor having a p-type channel structure in which the conductivity type is reversed.

[0018] <Study by the Inventor> In an insulated gate bipolar transistor (hereinafter sometimes referred to as an IGBT), when a gate voltage equal to or greater than the threshold voltage is applied to the gate electrode to turn on the IGBT, holes, which are minority carriers, are injected from the collector region to the drift region. As a result, electrons, which are majority carriers, are injected from the emitter region into the drift region, as if attracted by the injected holes. As a result, a conductivity modulation phenomenon occurs in the drift region, reducing the on-resistance of the IGBT.

[0019] On the other hand, in an IGBT, when a gate voltage smaller than the threshold voltage is applied to the gate electrode to turn the IGBT from on to off, the IGBT turns off after all of the electrons and holes injected into the drift region are swept out of the drift region. In other words, when an IGBT is turned off, all of the electrons and holes injected into the drift region must be swept out of the drift region, and the tail current caused by this sweeping flows before the IGBT turns off.

[0020] In an IGBT that operates in this manner, the longer the lifetime of holes, which are minority carriers, the lower the on-resistance. This is because if the lifetime of holes injected into the drift region is long, the number of holes that attract electrons can be increased, and as a result, the number of electrons attracted to the drift region increases, thereby reducing the on-resistance. Therefore, in order to reduce the on-resistance of an IGBT, it is desirable to extend the lifetime of holes.

[0021] On the other hand, when an IGBT is turned off, turn-off loss occurs. The cause of this turn-off loss is the tail current. Therefore, in order to reduce turn-off loss, it is necessary to reduce the tail current. This tail current is a current that occurs when all of the electrons and holes injected into the drift region are swept out of the drift region, and the shorter the hole lifetime, the more the tail current can be reduced. This is because if the hole lifetime is short, more holes will disappear before being swept out of the drift region, and as a result, fewer holes will be swept out of the drift region, which means that the tail current will be reduced.

[0022] From the above, in an IGBT, the longer the lifetime of holes, the more the on-resistance can be reduced, while the shorter the lifetime of holes, the more the turn-off loss can be reduced. In other words, in an IGBT, the reduction in on-resistance and the reduction in turn-off loss are in a trade-off relationship. Therefore, the present inventors are studying improvements to achieve both the reduction in on-resistance and the reduction in turn-off loss, which are in a trade-off relationship.

[0023] For example, the drift region can be formed by using an epitaxial growth method, but it is known that in IGBTs using silicon carbide as a constituent material, the lifetime of holes in the "as-grown" drift region is significantly shorter than in IGBTs using silicon as a constituent material. Therefore, if an "as-grown" drift region is used as is, the short lifetime of holes reduces turn-off loss, but also increases on-resistance. For this reason, it is not desirable to use an "as-grown" drift region as is in IGBTs using silicon carbide as a constituent material.

[0024] In this regard, carbon vacancies exist in drift regions that use silicon carbide as a constituent material. These carbon vacancies function as trap levels that capture holes. Therefore, in drift regions with a high carbon vacancy density, the number of trap levels that capture holes increases, resulting in a shorter hole lifetime. In other words, in drift regions with a low carbon vacancy density, the number of trap levels that capture holes decreases, resulting in a longer hole lifetime.

[0025] Considering the above, it is conceivable to introduce carbon into the drift region in the "as-grown state" and fill the carbon vacancies with the introduced carbon. In this case, it is possible to reduce the carbon vacancies that function as trap levels that capture holes, thereby lengthening the lifetime of the holes. That is, the present inventors are considering extending the lifetime of holes by introducing carbon into the drift region. This is because extending the lifetime of holes can reduce the on-resistance of the IGBT.

[0026] However, doping carbon throughout the entire drift region increases the lifetime of holes throughout the drift region, resulting in increased turn-off loss. Therefore, to achieve a balance between the trade-off between reduced on-resistance and reduced turn-off loss, it is important to determine the depth at which carbon is doped into the drift region.

[0027] That is, on the premise that a low carbon vacancy density region in which the carbon vacancy density is reduced by introducing carbon and a high carbon vacancy density region in which the carbon vacancy density is increased by maintaining the “as-grown state” without introducing carbon are provided in the drift region, it is important to determine the position (depth) of the boundary region between the low carbon vacancy density region and the high carbon vacancy density region.

[0028] In this regard, the present inventors have discovered, as a novel finding, that in specifying the depth of the boundary region described above, in order to achieve improvements toward achieving both a reduction in on-resistance and a reduction in turn-off loss, which are in a trade-off relationship with each other, it is important to specify the relationship with the end of the depletion layer that extends into the drift region when the IGBT is turned off.

[0029] For example, consider a case where the depth of the boundary region described above is shallower than the edge of the depletion layer extending into the drift region when the IGBT is turned off. In this case, the low carbon vacancy density region becomes narrower, which is disadvantageous from the perspective of reducing on-resistance. On the other hand, the electric field within the depletion layer sweeps out residual carriers and they do not exist inside the depletion layer. Therefore, even if there is a high carbon vacancy density region in the depletion layer where the hole lifetime is short, the presence of this high carbon vacancy density region within the depletion layer does not affect the improvement of turn-off loss.

[0030] In other words, from the perspective of achieving improvements to reduce on-resistance and turn-off loss, which are in a trade-off relationship with each other, a configuration in which the depth of the boundary region is located at a position shallower than the edge of the depletion layer cannot be said to be appropriate.

[0031] Therefore, in this embodiment, in order to achieve both a reduction in on-resistance and a reduction in turn-off loss, which are in a trade-off relationship with each other, a device is devised to appropriately define the depth of the boundary region between the low carbon vacancy density region and the high carbon vacancy density region in relation to the edge of the depletion layer. The technical concept of this device in this embodiment will be described below.

[0032] <Basic Concept of the Embodiment> The basic concept of the present embodiment is based on the premise that, in an IGBT using silicon carbide as a constituent material, a low carbon vacancy density region in which the carbon vacancy density is reduced by introducing carbon and a high carbon vacancy density region in which the carbon vacancy density is increased by maintaining the "as-grown state" without introducing carbon are provided in the drift region. The basic concept is that the position of the boundary region between the low carbon vacancy density region and the high carbon vacancy density region is located deeper than the end of the depletion layer extending into the drift region when the IGBT is turned off.

[0033] According to this basic concept, the thickness of the low carbon vacancy density region, where holes have a long lifetime, can be increased, thereby reducing the on-resistance of the IGBT. On the other hand, according to the basic concept, the high carbon vacancy density region, where holes have a short lifetime, is provided below the depletion layer. Therefore, by adjusting the thickness of the high carbon vacancy density region, the turn-off loss due to the tail current can be reduced. That is, according to the basic concept, by positioning the boundary region deeper than the edge of the depletion layer, it is possible to achieve improvements toward both reduced on-resistance and reduced turn-off loss, which are in a trade-off relationship with each other. Below, we will explain embodiments that embody this basic concept.

[0034] <Embodiment> <<Configuration of Semiconductor Device>> FIG. 1 is a cross-sectional view that schematically shows an active cell of an “n-channel SiC-IGBT” (hereinafter, sometimes abbreviated as IGBT) included in a semiconductor device in an embodiment.

[0035] 1 , the IGBT has an n-type drift region 4 containing n-type impurities (donors) such as nitrogen and phosphorus, and the drift region 4 contains silicon carbide as a main constituent material. This drift region 4 has, for example, a buffer region 3a and a low-concentration drift region 3b, with the buffer region 3a formed below the low-concentration drift region 3b. That is, the drift region 4 has the low-concentration drift region 3b and a buffer region 3a located below the low-concentration drift region 3b, in contact with the low-concentration drift region 3b, and having a higher impurity concentration than the low-concentration drift region 3b. The buffer region 3a is not necessarily required, but is provided to improve the breakdown voltage of the IGBT and suppress conduction loss (reduce the on-resistance of the IGBT).

[0036] The lower layer of the buffer region 3a is p + A p-type collector region 2 is provided. + A collector electrode 1 is provided below the collector region 2. + The collector region 2 is made of a p-type semiconductor region containing p-type impurities such as aluminum or boron.

[0037] Next, a body region 5, which is a p-type semiconductor region containing aluminum, boron, etc., is formed inside the low-concentration drift region 3b, and an n-type semiconductor region containing nitrogen, phosphorus, etc. is formed inside the body region 5. + A type emitter region 6 is formed in the body region 5. A body contact region 7 having a higher impurity concentration than the body region 5 is formed in the contact opening of the body region 5.

[0038] Next, n + A gate insulating film 8 is formed so as to cover the n-type emitter region 6, the body region 5 and the low-concentration drift region 3b, and a gate electrode 9 is provided so as to cover the gate insulating film 8. + An emitter electrode 10 is provided so as to cover the emitter region 6 and the body contact region 7 , and an interlayer insulating film 11 is formed to insulate the gate electrode 9 from the emitter electrode 10 .

[0039] In an IGBT configured as described above, the drift region 4 contains silicon carbide as a main constituent material, and as a result, carbon vacancies are inevitably present. In the drift region 4 in the embodied embodiment, the above-described carbon vacancy density has an intentional distribution.

[0040] In other words, the drift region 4 in the embodied mode has a low carbon vacancy density region 20 having a first carbon vacancy density, a high carbon vacancy density region 30 having a second carbon vacancy density higher than the first density, and a boundary region 40 sandwiched between the low carbon vacancy density region 20 and the high carbon vacancy density region 30, and having a carbon vacancy density higher than that of the low carbon vacancy density region 20 and a carbon vacancy density lower than that of the high carbon vacancy density region 30. For example, the first density is 1×10 13 / cm 3 and the second density is less than 1×10 13 / cm 3 That's all.

[0041] In this embodiment, the low carbon vacancy density region 20 is provided in the low-concentration drift region 3b. Meanwhile, the high carbon vacancy density region 30 is provided in a part of the low-concentration drift region 3b and in the buffer region 3a. The position of the isopycnal 100, which is included in the boundary region 40 and corresponds to a carbon vacancy concentration that is half the second density, is located deeper than the end 200 of the depletion layer that extends into the drift region 4 when the IGBT is turned off. In this embodiment, the isopycnal 100 exists in the low-concentration drift region 3b, and the end 200 of the depletion layer also exists in the low-concentration drift region 3b.

[0042] For example, the graph on the right side of FIG. 1 shows the depth-wise distribution (dotted line) of the carbon vacancy density contained in the drift region 4 and the depth-wise distribution (dashed dotted line) of the electric field generated in the drift region 4 when the IGBT is turned off with a power supply voltage applied between the collector electrode 1 and the emitter electrode 10. From the graph on the right side of FIG. 1 , it can be seen that a low carbon vacancy density region 20, which has a low carbon vacancy density, extends from the upper surface side of the drift region 4. A high carbon vacancy density region 30, which has a high carbon vacancy density, is formed below the low carbon vacancy density region 20, with a boundary region 40 sandwiched therebetween. In this case, within the boundary region 40, an isopycnal 100, where the carbon vacancy density is half that of the high carbon vacancy density region 30, is located deeper than an end 200 of the depletion layer, where the electric field generated in the drift region 4 becomes "0".

[0043] The carbon vacancies function as trap levels that capture minority carriers (holes). Therefore, the lifetime of minority carriers in the low carbon vacancy density region 20, where the carbon vacancy density is low, is longer than the lifetime of minority carriers in the high carbon vacancy density region 30, where the carbon vacancy density is high. As described above, the semiconductor device in this embodiment has an IGBT in which the drift region 4 is configured to include regions with different minority carrier lifetimes.

[0044] <<Method of Manufacturing a Semiconductor Device>> Next, a method of manufacturing a semiconductor device in an embodiment will be described. First, as shown in FIG. 2, an n-type bulk substrate 12 made of silicon carbide is prepared. Then, a p-type bulk substrate 12 is epitaxially grown on the upper surface of the bulk substrate 12. + The collector region 2, the buffer region 3a, and the low concentration drift region 3b are grown in this order.

[0045] Here, p + The impurity concentration of the collector region 2 is, for example, 1×10 18 / cm 3 The impurity concentration of the buffer region 3a is higher than the impurity concentration of the low-concentration drift region 3b. The impurity concentration of the low-concentration drift region 3b is, for example, 5×10 15 / cm 3 The temperature in the epitaxial growth method is, for example, 1700° C. or higher, and p + The carbon vacancy density in each of the collector region 2, the buffer region 3a, and the low-concentration drift region 3b is 1×10 13 / cm 3 The epitaxial growth is performed so that the thickness of the low-concentration drift region 3 b is approximately 50 μm to 100 μm when the breakdown voltage of the IGBT is 6.5 kV, for example. Furthermore, when the breakdown voltage of the IGBT is 20 kV, for example, the thickness of the low-concentration drift region 3 b is approximately 180 μm to 250 μm.

[0046] 3, carbon atoms are ion-implanted from the upper surface of the low-concentration drift region 3b. The implantation dose of carbon atoms is, for example, 1×10 15 / cm 2 This is all. Thereafter, as shown in FIG. 4 , the carbon atoms are diffused into the low-concentration drift region 3 b by heat treatment. This allows carbon atoms to fill the carbon vacancies present in the region from the top surface of the low-concentration drift region 3 b to a predetermined depth, thereby reducing the carbon vacancy density in that region. The heat treatment temperature (annealing temperature) is, for example, 1300° C. or higher, and the heat treatment time is, for example, 60 minutes or longer.

[0047] 4, in the embodied embodiment, there are formed a low carbon vacancy density region 20 having a first carbon vacancy density, a high carbon vacancy density region 30 having a second carbon vacancy density higher than the first density, and a boundary region 40 sandwiched between the low carbon vacancy density region 20 and the high carbon vacancy density region 30 and having a higher carbon vacancy density than the low carbon vacancy density region 20 and a lower carbon vacancy density than the high carbon vacancy density region 30. Note that in FIG. 4, for example, an isopycnal 100 included in the boundary region 40, which corresponds to a carbon vacancy concentration that is half the second density, is illustrated.

[0048] Next, as shown in FIG. 5, the body region 5 and n + The emitter region 6 and the body contact region 7 are formed by, for example, an ion implantation method using a photoresist as a patterning mask. The impurity concentration of the body region 5 is, for example, 2×10 16 / cm 3 From 2 x 10 18 / cm 3 About that. + The impurity concentration of each of the emitter region 6 and the body contact region 7 is, for example, 1×10 19 / cm 3 The conductive impurities are implanted at a high concentration so as to achieve the above-mentioned condition. After a protective film made of, for example, carbon is formed on the surface, heat treatment is performed at, for example, 1700° C. to 1900° C. to activate the conductive impurities introduced by ion implantation.

[0049] Next, for example, the protective film is removed by oxygen plasma treatment or the like, and then the gate insulating film 8 is formed on the low-concentration drift region 3b by, for example, wet oxidation, dry oxidation, or deposition of a silicon oxide film by CVD (Chemical Vapor Deposition).

[0050] Next, the gate electrode 9 is formed by forming a gate insulating film 8 and then depositing a polysilicon film directly thereon by the CVD method, or by depositing an amorphous silicon film by the CVD method and then transforming the amorphous silicon film into a polysilicon film by heat treatment. Then, an interlayer insulating film 11 made of a silicon oxide film is formed by the CVD method, and then a contact portion is opened in the interlayer insulating film 11 by using, for example, a dry etching method. As a result, an n-type SiO 2 film is formed at the bottom of the contact portion. + The type emitter region 6 and the body contact region 7 are exposed from the interlayer insulating film 11 .

[0051] Thereafter, the emitter electrode 10 made of a metal film such as aluminum, titanium, or nickel is formed by, for example, sputtering or metal vapor deposition. A part of the emitter electrode 10 is buried in the contact portion. + The gate electrode 1 is electrically connected to the emitter region 6 and the body contact region 7 .

[0052] Before forming the emitter electrode 10, the n-type silicon nitride film exposed at the bottom of the contact portion is + A silicide layer may be formed to cover the upper surfaces of the emitter region 6 and the body contact region 7. The silicide layer can be formed by using, for example, salicide technology.

[0053] Next, as shown in FIG. 6, the bulk substrate 12 is removed by a grinding process. + The lower surface of the collector region 2 is exposed. + A semiconductor substrate consisting of a collector region 2, a buffer region 3a, and a low-concentration drift region 3b can be obtained.

[0054] Here, an example has been described in which the bulk substrate 12 is removed after forming device structures such as the gate electrode 9 on the semiconductor substrate. However, if the drift region 4 (epitaxial growth layer) on the bulk substrate 12 has sufficient strength, the bulk substrate 12 may be removed after forming the drift region 4 as shown in FIG. 2, and then the device structures such as the body region 5 and the gate electrode 9 may be formed.

[0055] Next, as shown in FIG. + A collector electrode 1 is formed on the lower surface of the p-type collector region 2. The collector electrode 1 can be formed from a metal film such as aluminum, titanium, nickel, or gold by using, for example, a sputtering method or a metal vapor deposition method. + A silicide layer may be formed to cover the lower surface of the collector region 2. The silicide layer can be formed by, for example, laser annealing. In this manner, the semiconductor device according to the embodied embodiment can be manufactured.

[0056] The manufacturing method of the semiconductor device in the embodied mode can be summarized as follows. That is, the manufacturing method of the semiconductor device in the embodied mode is a manufacturing method of a semiconductor device including an IGBT in which the end of the depletion layer that extends when turned off remains inside the drift region 4. In this case, the manufacturing method of the semiconductor device includes the steps of: (a) forming an n-type drift region 4 containing silicon carbide as a constituent material; (b) introducing carbon atoms into the upper surface of the drift region 4; and (c) diffusing the carbon atoms into the drift region 4.

[0057] In a completed state of the semiconductor device after the step (c), the drift region 4 has a low carbon vacancy density region having a first carbon vacancy density, a high carbon vacancy density region having a second carbon vacancy density higher than the first density, and a boundary region sandwiched between the low carbon vacancy density region and the high carbon vacancy density region and having a higher carbon vacancy density than the low carbon vacancy density region and a lower carbon vacancy density than the high carbon vacancy density region. The position of the isopycnal included in the boundary region, where the isopycnal corresponds to a carbon vacancy concentration half the second density, is deeper than the end of the depletion layer that extends when the semiconductor device is turned off.

[0058] <<Modification of Manufacturing Method>> Next, a modification of the manufacturing method will be described. In this modification, as shown in FIG. +After the collector region 2, buffer region 3a, and low-concentration drift region 3b are grown in sequence, a silicon oxide film 50 is formed by thermal oxidation, as shown in Fig. 8. In this case, the thermal oxidation process decomposes silicon carbide on the upper surface of the low-concentration drift region 3b. As a result, the silicon produced by the decomposition forms the silicon oxide film 50 as silicon dioxide, while the carbon produced by the decomposition exists as atoms on the upper surface of the low-concentration drift region 3b, as in Fig. 3, which shows the embodiment.

[0059] At the same time, the carbon atoms are diffused by thermal oxidation, filling the carbon vacancies present in the region from the top surface of the low-concentration drift region 3 b to a predetermined depth with the carbon atoms. As a result, the carbon vacancy density in the region decreases. The temperature of the thermal oxidation is, for example, 1300° C. or higher, and the time of the thermal oxidation is, for example, 60 minutes or longer.

[0060] Thereafter, the silicon oxide film 50 is removed using, for example, hydrofluoric acid, and then a device structure (element structure) is formed in the same manner as in the above-described embodiment.

[0061] In the method for manufacturing a semiconductor device according to this modification, the generation of carbon atoms and the filling of carbon vacancies by the diffusion of the carbon atoms can be performed in the same thermal oxidation step. Furthermore, the step of removing the silicon oxide film 50 is also performed as a substrate surface cleaning step in the embodied mode, so that the semiconductor device can be manufactured with a reduced number of steps according to this modification.

[0062] 1 , a low carbon vacancy density region 20 having a low carbon vacancy density is formed on the upper surface side of the drift region 4, and a high carbon vacancy density region 30 having a high carbon vacancy density is formed on the lower surface side of the drift region 4. The characteristic of the embodiment is that, for example, as shown in FIG. 1 , an isopycnal 100 included in a boundary region 40 between the low carbon vacancy density region 20 and the high carbon vacancy density region 30, the isopycnal 100 having a carbon vacancy concentration that is half the carbon vacancy density of the high carbon vacancy density region 30, is positioned deeper than an end 200 of a depletion layer extending into the drift region 4 when the IGBT is turned off.

[0063] Here, since carbon vacancies act as a factor that shortens the lifetime of minority carriers in silicon carbide, the lifetime of minority carriers in low carbon vacancy density region 20, where the carbon vacancy density is low, is longer than that in high carbon vacancy density region 30, where the carbon vacancy density is high.

[0064] Therefore, the presence of the low carbon vacancy density region 20 can reduce the on-resistance during operation of the IGBT. On the other hand, when the IGBT is turned off, there are no residual carriers in the depletion layer extending to the drift region 4, and so the low carbon vacancy density region 20 does not affect the turn-off loss as long as it is within the depletion layer. As a result, according to the feature of the embodied aspect, it is possible to improve the trade-off relationship between the on-resistance (on-voltage) during operation of the IGBT and the turn-off loss when the IGBT is turned off.

[0065] For example, the depth Wd (cm) of the depletion layer formed in the IGBT when a voltage is applied can be expressed by the following formula, because the difference in impurity concentration between the low-concentration drift region 3b and the body region 5 is large and the built-in voltage generated in the pn junction of silicon carbide is much smaller than the power supply voltage: Wd = {(2 εs Vcc) / (q Nd)} 0.5

[0066] However, the power supply voltage is Vcc (V), the impurity concentration of the low-concentration drift region 3b is Nd (1 / cm 3 ), the amount of charge of an electron is q (C), and the dielectric constant of silicon carbide is ε (F / cm).

[0067] Therefore, when the thickness of the low carbon vacancy density region 20 is W (cm), and the characteristic point (Wd<W) in the embodiment is combined with the above-mentioned relational expression, the power supply voltage Vcc that realizes the characteristic point satisfies the following condition: Vcc<{(q·Nd) / (2·εs)}·W 2

[0068] Furthermore, the high carbon vacancy density region 30 functions as a region where the lifetime of minority carriers is short, which is called a “local lifetime control structure.” That is, in the embodied embodiment, the high carbon vacancy density region 30 constitutes a “local lifetime control structure.”

[0069] FIG. 9 is a graph showing the simulation results of the time variations in current and voltage at turn-off when the lifetime of minority carriers in the high carbon vacancy density region 30, which is the "local lifetime control structure", is changed.

[0070] When the lifetime of minority carriers in the high carbon vacancy density region 30 is set to be the same as that in the low carbon vacancy density region 20, that is, when there is no “local lifetime control structure”, the collector current decreases with time, but the slope of the decrease becomes gentle midway, and a tail current is generated.

[0071] In contrast, when the lifetime of the minority carriers in the high carbon vacancy density region 30 is reduced, in the case of 5 μs shown by the dashed line, the waveform is almost the same as that in the case where the “local lifetime control structure” is not provided, but in the case of 2 μs shown by the dashed line and 0.5 μs shown by the dashed line, it can be seen that the tail current is reduced.

[0072] That is, by setting the lifetime of minority carriers in the high carbon vacancy density region 30 to about 2 μs or less, it is possible to obtain the effect of reducing the turn-off loss by the "local lifetime control structure". To this end, the carbon vacancy density in the high carbon vacancy density region 30 must be set to 1×10 13 / cm 3 In this regard, since the carbon vacancy density of the high carbon vacancy density region 30 is the same as the carbon vacancy density of the epitaxially grown layer in the as-grown state, the carbon vacancy density during epitaxial growth is 1×10 13 / cm 3 By growing the crystal in the above manner, a "local lifetime controlled structure" can be realized in the embodied embodiment.

[0073] In the embodiment, silicon carbide is used as the semiconductor material. In this case, carbon vacancies, which act as a factor in reducing the lifetime of minority carriers, form a "deep level" 0.63 eV from the conduction band of silicon carbide. This level is deeper than the level formed by crystal defects, which act as a factor in reducing the lifetime of minority carriers in silicon. Therefore, according to the embodiment, the "localized lifetime control structure" can be maintained even at higher temperatures, reducing turn-off loss. IGBTs are power switching elements that are used while passing large currents. Therefore, the IGBTs reach high temperatures during operation. Therefore, by using silicon carbide as the semiconductor material, the effect of the "localized lifetime control structure" can be obtained even at high temperatures, thereby resulting in an IGBT with a wider operating range.

[0074] As described above, according to the features of the embodiment, it is possible to improve the trade-off between reduced on-resistance and reduced turn-off loss. In particular, by using silicon carbide as the semiconductor material, the trade-off can be improved even at high temperatures.

[0075] <<Structural Modification 1>> Next, Modification 1 will be described. Fig. 10 is a cross-sectional view showing an active cell of an IGBT according to Modification 1. The IGBT according to Modification 1 differs from the IGBT in the above-described embodiment in that, as shown in Fig. 10, in Modification 1, the position of an isopycnal 100, which indicates a carbon vacancy concentration that is half the carbon vacancy density of the high carbon vacancy density region 30, is located inside the buffer region 3a, and an end 200 of the depletion layer is also located inside the buffer region 3a.

[0076] Such a device structure is realized by increasing the heat treatment temperature and heat treatment time in the heat treatment (annealing) after implanting carbon atoms shown in FIG. 4 in the manufacturing process of the IGBT in the embodied embodiment, compared to the embodied embodiment, thereby diffusing the carbon atoms deeper.

[0077] In this first modification, as shown in FIG. 10, the drift region 4 has a low-concentration drift region 3b and a buffer region 3a that is located below the low-concentration drift region 3b, contacts the low-concentration drift region 3b, and has a higher impurity concentration than the low-concentration drift region 3b.

[0078] In this case, in this modified example 1, the low carbon vacancy density region 20 is provided in a part of the buffer region 3a and the low concentration drift region 3b, the high carbon vacancy density region 30 is provided in the buffer region 3a, the isopycnal 100 exists in the buffer region 3a, and the end 200 of the depletion layer also exists in the buffer region 3a.

[0079] In an IGBT configured as described above, when the IGBT is turned off, the gradient of the depth-wise distribution of the electric field generated in drift region 4 between collector electrode 1 and emitter electrode 10 becomes steeper in buffer region 3a with a high impurity concentration, thereby suppressing the expansion of the depletion layer even when the power supply voltage is increased. Therefore, even when the power supply voltage is increased, the position of isodensity line 100 can be made greater than the depth of edge 200 of the depletion layer. Therefore, according to this first modification, the trade-off between reduced on-resistance and reduced turn-off loss can be improved in a power supply voltage range greater than the power supply voltage range in the embodiment.

[0080] <<Structural Modification 2>> Next, Modification 2 will be described. FIG. 11 is a cross-sectional view showing an active cell of an IGBT according to Modification 2. The IGBT according to Modification 2 differs from the IGBT of the above-described embodiment in that a high-concentration drift region 3c is provided between the low-concentration drift region 3b and the buffer region 3a. The high-concentration drift region 3c has an impurity concentration higher than that of the low-concentration drift region 3b and a lower impurity concentration than that of the buffer region 3a. Another difference is that, as shown in FIG. 11 , in Modification 2, an isopycnal 100 corresponding to a carbon vacancy concentration half that of the high carbon vacancy density region 30 is located within the high-concentration drift region 3c, and an end 200 of the depletion layer is also located within the high-concentration drift region 3c.

[0081] Such a device structure can be realized by increasing the supply of impurity doping source gas for a predetermined initial period when forming the low-concentration drift region 3b using the epitaxial growth method shown in Figure 2 during the manufacturing process of the IGBT in the embodied embodiment. Alternatively, the high-concentration drift region 3c may be grown after forming the buffer region 3a, and then the low-concentration drift region 3b may be formed. Furthermore, if the drift region 4 on the bulk substrate 12 has sufficient strength, it is possible to form a p + After forming the collector region 2, the buffer region 3a, and the low-concentration drift region 3b in this order, the bulk substrate 12 is removed. + Alternatively, high-concentration drift region 3 c may be formed by ion-implanting n-type impurities (donors) such as nitrogen or phosphorus from the lower surface side (back surface side) of collector region 2 .

[0082] In this second modification, as shown in FIG. 11 , the drift region 4 includes a low-concentration drift region 3 b, a high-concentration drift region 3 c located below the low-concentration drift region 3 b and in contact with the low-concentration drift region 3 b, and having a higher impurity concentration than the low-concentration drift region 3 b, and a buffer region 3 a located below the high-concentration drift region 3 c and in contact with the high-concentration drift region 3 c, and having a higher impurity concentration than the high-concentration drift region 3 c.

[0083] In this case, in this modified example 2, a low carbon vacancy density region 20 is provided in part of the high-concentration drift region 3c and in the low-concentration drift region 3b, a high carbon vacancy density region 30 is provided in part of the high-concentration drift region 3c and in the buffer region 3a, the isopycnal 100 exists in the high-concentration drift region 3c, and the end 200 of the depletion layer also exists in the high-concentration drift region 3c.

[0084] In an IGBT configured as described above, when the IGBT is turned off, the gradient of the electric field in the depth direction generated in drift region 4 between collector electrode 1 and emitter electrode 10 becomes steeper in high-concentration drift region 3c having a high impurity concentration, thereby suppressing the expansion of the depletion layer even when the power supply voltage is increased. Therefore, even when the power supply voltage is increased, the position of isodensity line 100 can be made greater than the depth of edge 200 of the depletion layer. Therefore, according to this second modification, the trade-off between reduced on-resistance and reduced turn-off loss can be improved over a power supply voltage range greater than that of the embodiment.

[0085] Furthermore, because the impurity concentration of the high-concentration drift region 3c is lower than that of the buffer region 3a, the gradient of the electric field in the depletion layer is gentler than in the IGBT of the above-described modification 1. Therefore, when the IGBT is turned off, the gradient of the change over time in the voltage applied between the collector electrode 1 and the emitter electrode 10 after the depletion layer reaches the high-concentration drift region 3c can be made gentler than in the IGBT of modification 1. As a result, modification 2 has the advantage of being able to suppress the generation of noise and the like due to the change over time in the voltage.

[0086] <<Structural Modification 3>> Next, Modification 3 will be described. Fig. 12 is a cross-sectional view showing an active cell in Modification 3. The IGBT in Modification 3 differs from the IGBT in the above-described embodiment in that Modification 3 employs a trench structure for the gate electrode 9. Specifically, in Modification 3, as shown in Fig. 12, a trench is formed in the gate electrode 9 from the surface to the n-type. + A trench 15 is formed penetrating the n-type emitter region 6 and the body region 5. + A gate insulating film 8 is formed so as to cover the emitter region 6 and the trench 15 , and a gate electrode 9 is provided so as to cover the gate insulating film 8 .

[0087] Such a device structure is realized by the IGBT manufacturing process in the embodiment, in which the body region 5, n +This can be achieved by forming the emitter region 6 and the body contact region 7, activating the conductive impurities introduced by ion implantation, and then forming the trench 15 by, for example, dry etching.

[0088] In the IGBT of Modification 3, the cell pitch can be reduced by the trench gate structure, which improves the channel density per unit area and reduces the on-resistance of the IGBT. Furthermore, the crystal planes on the side surfaces of trench 15 are used as the channel, and since silicon carbide has high channel mobility on the side surfaces of the trench, a synergistic effect with the reduction in the cell pitch can be achieved, thereby further reducing the on-resistance.

[0089] The invention made by the inventor has been specifically described above based on the embodiments thereof, but it goes without saying that the present invention is not limited to the above-described embodiments and can be modified in various ways without departing from the spirit of the invention.

[0090] 1 Collector electrode 2 p + collector region 3a buffer region 3b low-concentration drift region 3c high-concentration drift region 4 drift region 5 body region 6 n + Type emitter region 7 Body contact region 8 Gate insulating film 9 Gate electrode 10 Emitter electrode 11 Interlayer insulating film 12 Bulk substrate 15 Trench 20 Low carbon vacancy density region 30 High carbon vacancy density region 40 Boundary region 50 Silicon oxide film 100 Contour 200 End

Claims

1. A semiconductor device including an insulated gate bipolar transistor, the insulated gate bipolar transistor having a drift region of a first conductivity type, the drift region including silicon carbide as a constituent material, the drift region having: a first region having a first carbon vacancy density; a second region having a second carbon vacancy density higher than the first density; and a boundary region sandwiched between the first region and the second region, the boundary region having a higher carbon vacancy density than the first region and a lower carbon vacancy density than the second region, wherein a position of an isopycnal included in the boundary region, the isopycnal corresponding to a carbon vacancy concentration that is half the second density, is located deeper than an end of a depletion layer extending into the drift region when the insulated gate bipolar transistor is turned off.

2. In the semiconductor device according to claim 1, the second density is 1×10 13 / cm 3 That is the semiconductor device.

3. A semiconductor device according to claim 1, wherein the lifetime of minority carriers in said first region is longer than the lifetime of minority carriers in said second region.

4. In the semiconductor device according to claim 1, the voltage applied between the collector and the emitter when the insulated gate bipolar transistor is turned off is Vcc (V), the thickness of the first region is W (cm), and the impurity concentration of the first region is Nd (1 / cm 3 ), the amount of charge of an electron is q (C), and the dielectric constant of the first region is εs (F / cm), Vcc<{(q·Nd) / (2·εs)}·W 2 The semiconductor device according to the present invention satisfies the above relationship.

5. A semiconductor device according to claim 1, wherein the drift region comprises: a low-concentration drift region; and a buffer region located below the low-concentration drift region, in contact with the low-concentration drift region, and having a higher impurity concentration than the low-concentration drift region; the first region is provided within the low-concentration drift region; the second region is provided in a part of the low-concentration drift region and in the buffer region; the isopycnic lines exist within the low-concentration drift region; and an end of the depletion layer exists within the low-concentration drift region.

6. A semiconductor device according to claim 1, wherein the drift region comprises: a low concentration drift region; and a buffer region located below the low concentration drift region, in contact with the low concentration drift region, and having a higher impurity concentration than the low concentration drift region; the first region is provided in a part of the buffer region and in the low concentration drift region; the second region is provided within the buffer region; the isopycnic lines exist within the buffer region; and an end of the depletion layer exists within the buffer region.

7. A semiconductor device according to claim 1, wherein the drift region comprises: a low concentration drift region; a high concentration drift region located below the low concentration drift region, in contact with the low concentration drift region and having a higher impurity concentration than the low concentration drift region; and a buffer region located below the high concentration drift region, in contact with the high concentration drift region and having a higher impurity concentration than the high concentration drift region; the first region is provided in a part of the high concentration drift region and in the low concentration drift region; the second region is provided in a part of the high concentration drift region and in the buffer region; the isopycnic lines exist within the high concentration drift region; and an end of the depletion layer exists within the high concentration drift region.

8. The semiconductor device according to claim 1, wherein the insulated gate bipolar transistor is a trench gate type insulated gate bipolar transistor.

9. A method for manufacturing a semiconductor device including an insulated gate bipolar transistor in which an end of a depletion layer that extends when the transistor is turned off remains inside a drift region, the method comprising: (a) a step of forming a drift region of a first conductivity type containing silicon carbide as a constituent material; (b) a step of introducing carbon atoms into an upper surface of the drift region; and (c) a step of diffusing the carbon atoms into the drift region, wherein after the step (c), the drift region has: a first region having a carbon vacancy density of a first density; a second region having a carbon vacancy density of a second density higher than the first density; and a boundary region sandwiched between the first region and the second region, the carbon vacancy density being higher than that of the first region and lower than that of the second region, and wherein a position of the isopycnic line included in the boundary region, the isopycnic line corresponding to a carbon vacancy concentration that is half the second density, is located deeper than the end of the depletion layer that extends when the transistor is turned off.

10. The method of manufacturing a semiconductor device according to claim 9, wherein in the step (b), the carbon atoms are ion-implanted.

11. The method for manufacturing a semiconductor device according to claim 9, wherein the steps (b) and (c) are carried out by a thermal oxidation process.

12. The method for manufacturing a semiconductor device according to claim 9, wherein the drift region formed in the step (a) has a carbon vacancy density of 1×10 13 (1 / cm 3 ) The method for manufacturing a semiconductor device.

Citation Information

Patent Citations

  • Transistor and manufacturing method of the same

    JP2014029952A

  • Silicon carbide semiconductor device and manufacturing method of silicon carbide semiconductor device

    JP2021019157A

  • Insulated gate bipolar transistor and method for designing same

    WO2011049054A1

  • Semiconductor device and method of manufacturing semiconductor device

    WO2015129430A1