Semiconductor device and method for manufacturing semiconductor device

Abstract

Provided is a semiconductor device comprising a semiconductor substrate having an upper surface and a lower surface and having a drift region of a first conductivity type. The semiconductor substrate has a buffer region of the first conductivity type provided between the drift region and the lower surface and having a doping concentration higher than the doping concentration of the drift region. The buffer region has a decrease portion in which the doping concentration monotonously decreases from the lower surface toward the upper surface. The decrease portion contains hydrogen. In the decease portion, the chemical concentration of the hydrogen also monotonously decreases. In the decrease portion, a logarithmic gradient of the distribution of the doping concentration from the lower surface toward the upper surface is smaller than 0 and not smaller than -250 ( / cm), and the length of the decrease portion is 15 μm or greater and is not greater than the thickness of the semiconductor substrate.

Description

Semiconductor device and method for manufacturing a semiconductor device

[0001] This invention relates to a semiconductor device and a method for manufacturing a semiconductor device.

[0002] Patent Document 1 discloses a technique for adjusting the donor concentration of a semiconductor substrate by implanting hydrogen ions into the semiconductor substrate from the implantation surface so as to penetrate the semiconductor substrate. Patent Document 2 discloses a semiconductor device having multiple hydrogen concentration peaks in the buffer region. [Prior Art Documents] [Patent Documents] [Patent Document 1] International Publication No. 2021 / 070584 [Patent Document 2] International Publication No. 2022 / 224883 General disclosure

[0003] A first embodiment of the present invention provides a semiconductor device comprising a semiconductor substrate having an upper surface and a lower surface, and having a drift region of a first conductivity type. In the above semiconductor device, the semiconductor substrate may have a buffer region of a first conductivity type provided between the drift region and the lower surface, wherein the doping concentration is higher than that of the drift region. In any of the above semiconductor devices, the buffer region may have a reduction portion in which the doping concentration decreases monotonically from the lower surface to the upper surface. In any of the above semiconductor devices, the reduction portion may contain hydrogen, and the chemical concentration of hydrogen in the reduction portion may also decrease monotonically. In the reduction portion of any of the above semiconductor devices, the logarithmic gradient of the distribution of the doping concentration from the lower surface to the upper surface may be less than 0 and -250 ( / cm) or more. In any of the above semiconductor devices, the length of the reduction portion may be 15 μm or more and less than or equal to the thickness of the semiconductor substrate.

[0004] A second embodiment of the present invention provides a semiconductor device comprising a semiconductor substrate having an upper surface and a lower surface, and having a drift region of a first conductivity type. In the above semiconductor device, the semiconductor substrate may have a buffer region of a first conductivity type provided between the drift region and the lower surface, wherein the doping concentration is higher than that of the drift region. In any of the above semiconductor devices, the buffer region may have a decreasing portion in which the doping concentration decreases monotonically from the lower surface toward the upper surface. In any of the above semiconductor devices, the decreasing portion may contain hydrogen, and the chemical concentration of hydrogen in the decreasing portion may also decrease monotonically. In any of the above semiconductor devices, the buffer region may have an upper hydrogen peak of the chemical concentration of hydrogen on the upper surface side of the decreasing portion. In any of the above semiconductor devices, if the depth position of the upper hydrogen peak from the lower surface is x (μm), the integral concentration obtained by integrating the chemical concentration distribution of hydrogen at the upper hydrogen peak is y ( / cm²). ２ ) is y = 2.70 × 10 １３ ×exp(-2.20×10 －２ x) Above, y = 2.70 × 10 １４ ×exp(-2.20×10 －２ x) The following is acceptable:

[0005] At each position of the reduction portion of any of the above semiconductor devices, the difference between the doping concentration and the approximate straight line that approximates the logarithmic distribution of the doping concentration may be 30% or less of the value of the approximate straight line.

[0006] In any of the above semiconductor devices, the logarithmic gradient may be -50 ( / cm) or less.

[0007] In any of the above semiconductor devices, the logarithmic gradient may be -100 ( / cm) or less.

[0008] In any of the above semiconductor devices, the logarithmic gradient may be -200 ( / cm) or greater.

[0009] In any of the above semiconductor devices, the logarithmic gradient of the reduction portion may be smaller than the logarithmic gradient of the doping concentration distribution from the lower surface to the upper surface in the drift region.

[0010] In any of the above semiconductor devices, the oxygen concentration contained in the semiconductor substrate may be greater than 7.0×10 １７ ( / cm ３ ).

[0011] In any of the above semiconductor devices, the oxygen concentration contained in the semiconductor substrate may be 8.0×10 １７ ( / cm ３ ) or more.

[0012] In any of the above semiconductor devices, the oxygen concentration contained in the semiconductor substrate may be 1.0×10 １８ ( / cm ３ ) or more. <00管理できる。 In any of the above semiconductor devices, the buffer region may have a lower hydrogen peak of the chemical concentration of hydrogen on the lower surface side than the decreasing portion.

[0014] In the buffer region of any of the above semiconductor devices, there may be no peak of the chemical concentration of hydrogen on the lower surface side than the lower hydrogen peak.

[0015] In any of the above semiconductor devices, the buffer region may have an upper hydrogen peak of the chemical concentration of hydrogen on the upper surface side than the decreasing portion.

[0016] In any of the above semiconductor devices, the buffer region may have a lower peak of the doping concentration on the lower surface side than the decreasing portion.

[0017] In any of the above semiconductor devices, the buffer region may have an upper peak of the doping concentration on the upper surface side than the decreasing portion. [[ID=B]]

[0018] In any of the above semiconductor devices, the buffer region may have an upper hydrogen peak of the chemical concentration of hydrogen on the upper surface side than the decreasing portion. In any of the above semiconductor devices, when the depth position from the lower surface of the upper hydrogen peak is x (μm), the integrated concentration y ( / cm ２ ) of the chemical concentration distribution of hydrogen in the upper hydrogen peak is y = 2.70×10 １３ ×exp(−2.20×10－２ x) Above, y = 2.70 × 10 １４ ×exp(-2.20×10 －２ x) The following is acceptable:

[0019] A third embodiment of the present invention provides a method for manufacturing a semiconductor device comprising a semiconductor substrate having an upper surface and a lower surface. In the above manufacturing method, hydrogen ions may be implanted from the lower surface to a first depth position within the semiconductor substrate. In any of the above manufacturing methods, hydrogen ions may be implanted from the lower surface to a second depth position on the lower surface side of the first depth position. In any of the above manufacturing methods, the semiconductor substrate may be annealed. In any of the above manufacturing methods, the dose amount of hydrogen ions to the first depth position x (μm) from the lower surface is y ( / cm). ２ ) is y = 2.70 × 10 １３ ×exp(-2.20×10 －２ x) Above, y = 2.70 × 10 １４ ×exp(-2.20×10 －２ x) The following is acceptable:

[0020] The above summary of the invention does not enumerate all the necessary features of the present invention. Furthermore, subcombinations of these features may also constitute an invention.

[0021] This is a top view showing an example of a semiconductor device 100 according to one embodiment of the present invention. This is an enlarged view of region D in Figure 1. This is a diagram showing an example of the e-e cross section in Figure 2. This is a diagram showing an example of the doping concentration distribution and hydrogen chemical concentration distribution along the f-f line in Figure 3. This is a diagram showing another example of the doping concentration distribution and hydrogen chemical concentration distribution along the f-f line in Figure 3. This is a diagram showing another example of the doping concentration distribution and hydrogen chemical concentration distribution along the f-f line in Figure 3. This is a diagram showing an example of the manufacturing process of the semiconductor device 100 in the embodiment. This is a diagram showing the first injection process S1010. This is a diagram showing the second injection process S1020. This is a diagram showing the annealing process S1030. This is a diagram showing an example of the carrier concentration distribution of the buffer region 20 at different dose amounts. This is a diagram showing an example of the carrier concentration distribution of the buffer region 20 at different dose amounts. This is a diagram showing an example of the carrier concentration distribution of the buffer region 20 at different dose amounts. This is a diagram showing an example of the carrier concentration distribution of the buffer region 20 at different dose amounts. This is a diagram showing an example of the carrier concentration distribution of the buffer region 20 at different dose amounts. This is a diagram showing an example of the carrier concentration distribution of the buffer region 20 at different dose amounts. This figure shows the relationship between the proton injection depth and the dose. This figure shows the carrier concentration distribution in the buffer region 20 in a comparative example. This figure shows the carrier concentration distribution in the buffer region 20 in another comparative example. This figure shows the carrier concentration distribution in the buffer region 20 in semiconductor substrates 10 with different oxygen concentrations. This figure shows the relationship between oxygen concentration and the logarithmic gradient α. This figure shows the relationship between oxygen concentration and the logarithmic gradient α. This figure shows the oxygen concentration distribution in the depth direction of the semiconductor substrate 10. This figure illustrates the approximate straight line of the reduction section 210.

[0022] The present invention will be described below through embodiments, but these embodiments are not intended to limit the scope of the claims. Furthermore, not all combinations of features described in the embodiments are necessarily essential to the solution of the invention.

[0023] In this specification, one side of a semiconductor substrate parallel to its depth direction is referred to as "upper," and the other side as "lower." Of the two main surfaces of a substrate, layer, or other component, one surface is referred to as the upper surface, and the other surface as the lower surface. The directions of "upper" and "lower" are not limited to the direction of gravity or the direction in which the semiconductor device is mounted.

[0024] In this specification, technical matters may be described using the Cartesian coordinate axes X, Y, and Z. The Cartesian coordinate axes merely specify the relative positions of components and do not limit any particular direction. For example, the Z axis does not limit the direction to height relative to the ground. Note that the +Z axis direction and the -Z axis direction are opposite directions. When the sign is not specified and only the Z axis direction is written, it means the direction parallel to the +Z axis and the -Z axis.

[0025] In this specification, the orthogonal axes parallel to the top and bottom surfaces of the semiconductor substrate are defined as the X and Y axes. The axis perpendicular to the top and bottom surfaces of the semiconductor substrate is defined as the Z axis. In this specification, the direction of the Z axis may be referred to as the depth direction. In this specification, the direction parallel to the top and bottom surfaces of the semiconductor substrate, including the X and Y axes, may be referred to as the horizontal direction.

[0026] In this specification, when we refer to "identical" or "equal," we may include cases where there are errors due to manufacturing variations, etc. Such errors are, for example, within 10%.

[0027] In this specification, the conductivity type of a doped region containing impurities is described as either P-type or N-type. In this specification, impurities may specifically refer to either N-type donors or P-type acceptors, and may be referred to as dopants. In this specification, doping means introducing donors or acceptors into a semiconductor substrate to make it a semiconductor exhibiting either an N-type conductivity or a P-type conductivity.

[0028] In this specification, doping concentration means the concentration of the donor or acceptor at thermal equilibrium. In this specification, net doping concentration means the net concentration obtained by adding the charge polarity of the donor concentration (where the donor concentration is the concentration of positive ions) and the acceptor concentration (where the acceptor concentration is the concentration of negative ions). For example, if the donor concentration is ND and the acceptor concentration is NA, the net doping concentration at any given position will be ND-NA. In this specification, net doping concentration may sometimes be simply referred to as doping concentration.

[0029] Donors have the function of supplying electrons to a semiconductor. Acceptors have the function of receiving electrons from a semiconductor. Donors and acceptors are not limited to impurities themselves. For example, VOH defects, which are formed by the bonding of vacancies (V), oxygen (O), and hydrogen (H) in a semiconductor, function as electron-supplying donors. In other examples, interstitial Si-H, formed by the bonding of interstitial silicon (Si-i) and hydrogen in a silicon semiconductor, and CiOi-H, formed by the bonding of interstitial carbon (Ci), interstitial oxygen (Oi), and hydrogen, function as electron-supplying donors. In this specification, examples in which hydrogen contributes as an electron-supplying donor as described above may be collectively referred to as hydrogen donors.

[0030] In this specification, the semiconductor substrate has N-type bulk donors distributed throughout. Bulk donors are donors from dopants that are substantially uniformly contained within the ingot during the manufacturing of the semiconductor substrate ingot. In this example, the bulk donor is an element other than hydrogen. The bulk donor dopants are, for example, phosphorus, antimony, arsenic, selenium, or sulfur, but are not limited to these. In this example, the bulk donor is phosphorus. The bulk donor is also contained in the P-type region. The semiconductor substrate may be a wafer cut from a semiconductor ingot, or it may be a chip made by cutting a wafer into individual pieces. The semiconductor ingot may be manufactured by one of the following methods: the Czochralski method (CZ method), the magnetic field applied Czochralski method (MCZ method), or the float zone method (FZ method). In this example, the ingot is manufactured by the MCZ method.

[0031] The semiconductor substrate may contain oxygen. Oxygen may be present throughout the semiconductor substrate. The oxygen concentration in the semiconductor substrate will be discussed later. Higher oxygen concentrations tend to generate hydrogen donors more easily. The bulk donor concentration may be the chemical concentration of bulk donors distributed throughout the semiconductor substrate, and may be a value between 90% and 100% of that chemical concentration. In addition, the semiconductor substrate may be a non-doped substrate that does not contain dopants such as phosphorus. In that case, the bulk donor concentration (D0) of the non-doped substrate may be, for example, 1 × 10⁻⁶. １０ / cm ３ The above 5 x 10 １２ / cm ３ The following applies: The bulk donor concentration (D0) of the non-doped substrate is preferably 1 × 10⁻⁶. １１ / cm ３ That concludes the explanation. The bulk donor concentration (D0) of the non-doped substrate is preferably 5 × 10⁻⁶. １２ / cm ３ The following applies. Note that the concentrations in this invention may be values ​​at room temperature. For example, the values ​​at room temperature may be those at 300 K (Kelvin) (approximately 26.9°C).

[0032] In this specification, when P+ type or N+ type is mentioned, it means a higher doping concentration than P type or N type, and when P- type or N- type is mentioned, it means a lower doping concentration than P type or N type. Furthermore, when P++ type or N++ type is mentioned in this specification, it means a higher doping concentration than P+ type or N+ type. Unless otherwise specified, the units used in this specification are the SI units. Although units of length may be expressed in cm, calculations may be performed after converting to meters (m).

[0033] In this specification, chemical concentration refers to the atomic density of impurities measured independently of the electrical activation state. Chemical concentration can be measured, for example, by secondary ion mass spectrometry (SIMS). The net doping concentration described above can be measured by voltage-capacitance (CV) method. Alternatively, the carrier concentration measured by broadening resistance (SR) method may be used as the net doping concentration. The carrier concentration measured by CV or SR method may be the value at thermal equilibrium. Furthermore, in the N-type region, since the donor concentration is sufficiently larger than the acceptor concentration, the carrier concentration in that region may be used as the donor concentration. Similarly, in the P-type region, the carrier concentration in that region may be used as the acceptor concentration. In this specification, the doping concentration in the N-type region may be referred to as the donor concentration, and the doping concentration in the P-type region may be referred to as the acceptor concentration.

[0034] If the concentration distribution of the donor, acceptor, or net doping has a peak, the peak value may be used as the concentration of the donor, acceptor, or net doping in that region. If the concentrations of the donor, acceptor, or net doping are nearly uniform, the average value of the concentrations of the donor, acceptor, or net doping in that region may be used as the concentration of the donor, acceptor, or net doping. In this specification, the concentration per unit volume is expressed as / cm³. ３ , or atoms / cm ３ This unit is used for donor or acceptor concentrations in semiconductor substrates, or for chemical concentrations. The atom notation may be omitted.

[0035] The carrier concentration measured by the SR method may be lower than the donor or acceptor concentration. When measuring spreading resistance, the carrier mobility of the semiconductor substrate may be lower than the value for the crystalline state in the range where current flows. The decrease in carrier mobility occurs because carriers are scattered due to disorder in the crystal structure caused by lattice defects, etc.

[0036] The donor or acceptor concentrations calculated from carrier concentrations measured by the CV or SR method may be lower than the chemical concentrations of the elements exhibiting donor or acceptor properties. For example, in silicon semiconductors, the donor concentrations of phosphorus or arsenic, or the acceptor concentration of boron, are approximately 99% of their respective chemical concentrations. On the other hand, the donor concentration of hydrogen, which is also a donor in silicon semiconductors, is approximately 0.1% to 10% of the hydrogen chemical concentration.

[0037] Figure 1 is a top view showing an example of a semiconductor device 100 according to one embodiment of the present invention. In Figure 1, the positions of each component projected onto the upper surface of the semiconductor substrate 10 are shown. In Figure 1, only some components of the semiconductor device 100 are shown, and some components are omitted.

[0038] The semiconductor device 100 includes a semiconductor substrate 10. The semiconductor substrate 10 is a substrate formed of a semiconductor material. As an example, the semiconductor substrate 10 is a silicon substrate. The semiconductor substrate 10 has edges 162 when viewed from above. In this specification, when simply referred to as "viewed from above," it means viewed from the top side of the semiconductor substrate 10. In this example, the semiconductor substrate 10 has two pairs of edges 162 that face each other when viewed from above. In Figure 1, the X and Y axes are parallel to one of the edges 162. The Z axis is perpendicular to the top surface of the semiconductor substrate 10.

[0039] The semiconductor substrate 10 is provided with an active area 160. The active area 160 is a region in which the main current flows in the depth direction between the upper and lower surfaces of the semiconductor substrate 100 when the semiconductor device 100 is operating. An emitter electrode is provided above the active area 160, but it is omitted in Figure 1. The active area 160 may refer to the region that overlaps with the emitter electrode when viewed from above. Also, the region sandwiched between the active areas 160 when viewed from above may be included in the active area 160. For example, if two emitter electrodes are provided, the area between the two emitter electrodes may also be considered as the active area 160.

[0040] The active section 160 is provided with at least one of a transistor section 70 including a transistor element such as an IGBT (Insulated Gate Bipolar Transistor), and a diode section 80 including a diode element such as a freewheeling diode (FWD). In the example shown in Figure 1, the transistor section 70 and the diode section 80 are arranged alternately along a predetermined arrangement direction (in this example, the X-axis direction) on the upper surface of the semiconductor substrate 10. The semiconductor device 100 in this example is a reverse-conducting IGBT (RC-IGBT).

[0041] In Figure 1, the region where the transistor section 70 is located is denoted by the symbol "I," and the region where the diode section 80 is located is denoted by the symbol "F." In this specification, the direction perpendicular to the arrangement direction in a top view may be referred to as the extension direction (Y-axis direction in Figure 1). The transistor section 70 and the diode section 80 may each have a longitudinal length in the extension direction. That is, the length of the transistor section 70 in the Y-axis direction is greater than its width in the X-axis direction. Similarly, the length of the diode section 80 in the Y-axis direction is greater than its width in the X-axis direction. The extension direction of the transistor section 70 and the diode section 80 may be the same as the longitudinal direction of each trench section described later.

[0042] The diode portion 80 has an N+ type cathode region in the area in contact with the lower surface of the semiconductor substrate 10. In this specification, the region in which the cathode region is provided is referred to as the diode portion 80. In other words, the diode portion 80 is the region that overlaps with the cathode region when viewed from above. A P+ type collector region may be provided on the lower surface of the semiconductor substrate 10 in areas other than the cathode region. In this specification, an extension region 81, which is an extension of the diode portion 80 in the Y-axis direction to the gate wiring described later, may also be included in the diode portion 80. A collector region is provided on the lower surface of the extension region 81.

[0043] The transistor section 70 has a P+ type collector region in the area that is in contact with the lower surface of the semiconductor substrate 10. Furthermore, the transistor section 70 has a gate structure periodically arranged on the upper surface side of the semiconductor substrate 10, which includes an N type emitter region, a P type base region, a gate conductive portion, and a gate insulating film.

[0044] The semiconductor device 100 may have one or more pads on the semiconductor substrate 10. In this example, the semiconductor device 100 has a gate pad 164. The semiconductor device 100 may also have pads such as an anode pad, a cathode pad, and a current detection pad. Each pad is located near the edge 162. The vicinity of the edge 162 refers to the area between the edge 162 and the emitter electrode in a top view. When the semiconductor device 100 is mounted, each pad may be connected to an external circuit via wiring such as wires.

[0045] A gate potential is applied to the gate pad 164. The gate pad 164 is electrically connected to the conductive portion of the gate trench of the active portion 160. The semiconductor device 100 is provided with gate wiring that connects the gate pad 164 to the gate trench. In Figure 1, the gate wiring is hatched with diagonal lines.

[0046] The gate wiring in this example has an outer perimeter gate wiring 130 and an active-side gate wiring 131. The outer perimeter gate wiring 130 is positioned between the active portion 160 and the edge 162 of the semiconductor substrate 10 in a top view. In this example, the outer perimeter gate wiring 130 surrounds the active portion 160 in a top view. The area surrounded by the outer perimeter gate wiring 130 in a top view may be considered the active portion 160. Furthermore, a well region is formed below the gate wiring. The well region is a P-type region with a higher concentration than the base region, which will be described later, and is formed from the top surface of the semiconductor substrate 10 to a position deeper than the base region. The area surrounded by the well region in a top view may be considered the active portion 160.

[0047] The outer perimeter gate wiring 130 is connected to the gate pad 164. The outer perimeter gate wiring 130 is positioned above the semiconductor substrate 10. The outer perimeter gate wiring 130 may be a metal wiring containing aluminum or the like.

[0048] The active gate wiring 131 is provided in the active section 160. By providing the active gate wiring 131 in the active section 160, variations in the wiring length from the gate pad 164 can be reduced for each region of the semiconductor substrate 10.

[0049] The outer periphery gate wiring 130 and the active side gate wiring 131 are connected to the gate trench portion of the active portion 160. The outer periphery gate wiring 130 and the active side gate wiring 131 are positioned above the semiconductor substrate 10. The outer periphery gate wiring 130 and the active side gate wiring 131 may be wiring formed from a semiconductor such as polysilicon doped with impurities, or metal wiring containing aluminum, or both.

[0050] The active gate wiring 131 may be connected to the outer gate wiring 130. In this example, the active gate wiring 131 extends in the X-axis direction from one outer gate wiring 130 to the other outer gate wiring 130 that sandwiches the active portion 160, crossing the active portion 160 approximately in the center in the Y-axis direction. When the active portion 160 is divided by the active gate wiring 131, the transistor portion 70 and the diode portion 80 may be arranged alternately in the X-axis direction in each divided region.

[0051] The semiconductor device 100 may include a temperature sensing unit (not shown) which is a PN junction diode made of polysilicon or the like, and a current detection unit (not shown) which simulates the operation of a transistor unit provided in the active unit 160.

[0052] In this example, the semiconductor device 100 includes an edge termination structure 90 between the active portion 160 and the edge 162 when viewed from above. The edge termination structure 90 in this example is located between the outer peripheral gate wiring 130 and the edge 162. The edge termination structure 90 mitigates electric field concentration on the upper surface side of the semiconductor substrate 10. The edge termination structure 90 may include at least one of a guard ring, a field plate, and a resurf, which are provided in an annular shape surrounding the active portion 160.

[0053] Figure 2 is an enlarged view of region D in Figure 1. Region D is the region including the transistor section 70, the diode section 80, and the active gate wiring 131. The semiconductor device 100 in this example includes a gate trench section 40, a dummy trench section 30, a well section 17, an emitter section 12, a base section 14, and a contact section 15 provided inside the upper surface of the semiconductor substrate 10. The gate trench section 40 and the dummy trench section 30 are examples of trench sections. The semiconductor device 100 in this example also includes an emitter electrode 52 and an active gate wiring 131 provided above the upper surface of the semiconductor substrate 10. The emitter electrode 52 and the active gate wiring 131 are provided separately from each other.

[0054] An interlayer insulating film is provided between the emitter electrode 52 and the active gate wiring 131 and the upper surface of the semiconductor substrate 10, but this is omitted in Figure 2. Contact holes 54 are provided in the interlayer insulating film in this example, penetrating the film. In Figure 2, each contact hole 54 is hatched with diagonal lines.

[0055] The emitter electrode 52 is provided above the gate trench 40, dummy trench 30, well region 17, emitter region 12, base region 14, and contact region 15. The emitter electrode 52 contacts the emitter region 12, contact region 15, and base region 14 on the upper surface of the semiconductor substrate 10 through a contact hole 54. The emitter electrode 52 is also connected to a dummy conductive portion in the dummy trench 30 through a contact hole provided in the interlayer insulating film. The emitter electrode 52 may be connected to a dummy conductive portion of the dummy trench 30 at its tip in the Y-axis direction. The dummy conductive portion of the dummy trench 30 does not need to be connected to the emitter electrode 52 and the gate conductive portion, and may be controlled to a potential different from the potential of the emitter electrode 52 and the gate conductive portion.

[0056] The active gate wiring 131 connects to the gate trench portion 40 through a contact hole provided in the interlayer insulating film. The active gate wiring 131 may be connected to the gate conductive portion of the gate trench portion 40 at the tip portion 41 of the gate trench portion 40 in the Y-axis direction. The active gate wiring 131 is not connected to the dummy conductive portion in the dummy trench portion 30.

[0057] The emitter electrode 52 is formed from a material containing metal. Figure 2 shows the area in which the emitter electrode 52 is provided. For example, at least a portion of the emitter electrode 52 is formed from aluminum or an aluminum-silicon alloy, such as AlSi, AlSiCu, or other metal alloys. The emitter electrode 52 may have a barrier metal formed from titanium or a titanium compound in the layer below the region formed from aluminum, etc. Furthermore, it may have a plug formed by embedding tungsten or the like in the contact hole so as to be in contact with the barrier metal and the aluminum, etc.

[0058] The well region 17 is provided overlapping with the active gate wiring 131. The well region 17 also extends to a predetermined width in an area that does not overlap with the active gate wiring 131. In this example, the well region 17 is provided away from the Y-axis end of the contact hole 54 towards the active gate wiring 131. The well region 17 is a second conductivity type region with a higher doping concentration than the base region 14. In this example, the base region 14 is P-type, and the well region 17 is P+-type.

[0059] Each of the transistor section 70 and the diode section 80 has multiple trench sections arranged in the direction of arrangement. In this example, the transistor section 70 has one or more gate trench sections 40 and one or more dummy trench sections 30 alternately provided along the direction of arrangement. In this example, the diode section 80 has multiple dummy trench sections 30 provided along the direction of arrangement. In this example, the diode section 80 does not have gate trench sections 40.

[0060] The gate trench portion 40 in this example may have two straight portions 39 (the trench portion which is linear along the extension direction) that extend along an extension direction perpendicular to the alignment direction, and a tip portion 41 that connects the two straight portions 39. In Figure 2, the extension direction is the Y-axis direction.

[0061] Preferably, at least a portion of the tip portion 41 is provided in a curved shape when viewed from above. By connecting the ends of the two straight portions 39 in the Y-axis direction with the tip portion 41, electric field concentration at the ends of the straight portions 39 can be mitigated.

[0062] In the transistor section 70, the dummy trench section 30 is provided between each of the straight sections 39 of the gate trench section 40. There may be one dummy trench section 30 between each of the straight sections 39, or there may be multiple dummy trench sections 30. The dummy trench section 30 may have a straight shape extending in the extension direction, and like the gate trench section 40, it may have a straight section 29 and a tip section 31. The semiconductor device 100 shown in Figure 2 includes both a dummy trench section 30 with a straight shape without a tip section 31 and a dummy trench section 30 with a tip section 31.

[0063] The diffusion depth of the well region 17 may be deeper than the depth of the gate trench portion 40 and the dummy trench portion 30. The Y-axis ends of the gate trench portion 40 and the dummy trench portion 30 are located in the well region 17 when viewed from above. In other words, at the Y-axis end of each trench portion, the bottom in the depth direction of each trench portion is covered by the well region 17. This makes it possible to mitigate electric field concentration at the bottom of each trench portion.

[0064] Mesa portions are provided between each trench portion in the arrangement direction. A mesa portion refers to a region within the semiconductor substrate 10 that is sandwiched between the trench portions. For example, the upper end of a mesa portion is the upper surface of the semiconductor substrate 10. The depth position of the lower end of a mesa portion is the same as the depth position of the lower end of a trench portion. In this example, the mesa portion is provided on the upper surface of the semiconductor substrate 10, extending along the trench in the extension direction (Y-axis direction). In this example, a mesa portion 60 is provided in the transistor portion 70, and a mesa portion 61 is provided in the diode portion 80. In this specification, when simply referred to as a mesa portion, it refers to mesa portion 60 and mesa portion 61, respectively.

[0065] Each mesa portion is provided with a base region 14. Of the base regions 14 exposed on the upper surface of the semiconductor substrate 10 in the mesa portion, the region closest to the active gate wiring 131 is defined as base region 14-e. Figure 2 shows the base region 14-e located at one end of each mesa portion in the extending direction, but a base region 14-e is also located at the other end of each mesa portion. In each mesa portion, at least one of a first conductivity type emitter region 12 and a second conductivity type contact region 15 may be provided in the region sandwiched between the base regions 14-e in a top view. In this example, the emitter region 12 is N+ type and the contact region 15 is P+ type. The emitter region 12 and the contact region 15 may be provided in the depth direction between the base region 14 and the upper surface of the semiconductor substrate 10.

[0066] The mesa portion 60 of the transistor portion 70 has an emitter region 12 exposed on the upper surface of the semiconductor substrate 10. The emitter region 12 is provided in contact with the gate trench portion 40. The mesa portion 60 in contact with the gate trench portion 40 may be provided with a contact region 15 exposed on the upper surface of the semiconductor substrate 10.

[0067] Each of the contact region 15 and emitter region 12 in the mesa portion 60 extends from one trench portion to the other in the X-axis direction. As an example, the contact region 15 and emitter region 12 of the mesa portion 60 are arranged alternately along the extension direction (Y-axis direction) of the trench portion.

[0068] In other examples, the contact region 15 and emitter region 12 of the mesa portion 60 may be arranged in a stripe pattern along the extension direction (Y-axis direction) of the trench portion. For example, the emitter region 12 may be provided in the region in contact with the trench portion, and the contact region 15 may be provided in the region sandwiched between the emitter regions 12.

[0069] The mesa portion 61 of the diode portion 80 does not have an emitter region 12. A base region 14 and a contact region 15 may be provided on the upper surface of the mesa portion 61. Contact regions 15 may be provided in contact with each base region 14-e in the region sandwiched between the base regions 14-e on the upper surface of the mesa portion 61. A base region 14 may be provided in the region sandwiched between the contact regions 15 on the upper surface of the mesa portion 61. The base region 14 may be arranged in the entire region sandwiched between the contact regions 15.

[0070] A contact hole 54 is provided above each mesa portion. The contact hole 54 is located in the region sandwiched between the base region 14-e. In this example, the contact hole 54 is provided above the contact region 15, the base region 14, and the emitter region 12. The contact hole 54 is not provided in the region corresponding to the base region 14-e and the well region 17. The contact hole 54 may be located in the center in the alignment direction (X-axis direction) of the mesa portions 60.

[0071] In the diode section 80, an N+ type cathode region 82 is provided in the region adjacent to the lower surface of the semiconductor substrate 10. In the region on the lower surface of the semiconductor substrate 10 where the cathode region 82 is not provided, a P+ type collector region 22 may be provided. The cathode region 82 and the collector region 22 are provided between the lower surface 23 of the semiconductor substrate 10 and the buffer region 20. In Figure 2, the boundary between the cathode region 82 and the collector region 22 is shown by a dotted line.

[0072] The cathode region 82 is positioned away from the well region 17 in the Y-axis direction. This ensures a distance between the well region 17, which has a relatively high doping concentration and is formed to a deep position, and the cathode region 82, thereby improving pressure resistance. In this example, the end of the cathode region 82 in the Y-axis direction is positioned further from the well region 17 than the end of the contact hole 54 in the Y-axis direction. In other examples, the end of the cathode region 82 in the Y-axis direction may be positioned between the well region 17 and the contact hole 54.

[0073] Figure 3 shows an example of the e-e cross-section in Figure 2. The e-e cross-section is the XZ plane passing through the emitter region 12 and the cathode region 82. In this example, the semiconductor device 100 has a semiconductor substrate 10, an interlayer insulating film 38, an emitter electrode 52, and a collector electrode 24 in this cross-section.

[0074] The interlayer insulating film 38 is provided on the upper surface 21 of the semiconductor substrate 10. The interlayer insulating film 38 is a film that includes at least one layer of insulating film such as silicate glass with impurities such as boron or phosphorus added, a thermal oxide film, and other insulating films. The interlayer insulating film 38 is provided with contact holes 54 as described in Figure 2.

[0075] The emitter electrode 52 is provided above the interlayer insulating film 38. The emitter electrode 52 is in contact with the upper surface 21 of the semiconductor substrate 10 through a contact hole 54 in the interlayer insulating film 38. The collector electrode 24 is provided on the lower surface 23 of the semiconductor substrate 10. The emitter electrode 52 and the collector electrode 24 are made of a metallic material such as aluminum. In this specification, the direction connecting the emitter electrode 52 and the collector electrode 24 (Z-axis direction) is referred to as the depth direction.

[0076] The semiconductor substrate 10 has an N-type drift region 18. The drift region 18 is provided in both the transistor section 70 and the diode section 80.

[0077] The mesa portion 60 of the transistor portion 70 has an N+ type emitter region 12 and a P type base region 14, which are provided in order from the upper surface 21 side of the semiconductor substrate 10. A drift region 18 is provided below the base region 14. An N type storage region 16 may also be provided in the mesa portion 60. The storage region 16 is located between the base region 14 and the drift region 18.

[0078] The emitter region 12 is exposed on the upper surface 21 of the semiconductor substrate 10 and is provided in contact with the gate trench portion 40. The emitter region 12 may be in contact with the trench portions on both sides of the mesa portion 60. The doping concentration in the emitter region 12 is higher than that in the drift region 18.

[0079] The base region 14 is located below the emitter region 12. In this example, the base region 14 is located in contact with the emitter region 12. The base region 14 may be in contact with the trenches on both sides of the mesa region 60.

[0080] The storage region 16 is located below the base region 14. The storage region 16 is an N-type region with a higher doping concentration than the drift region 18. By providing a high-concentration storage region 16 between the drift region 18 and the base region 14, the carrier injection promotion effect (IE effect) can be enhanced, and the on-voltage can be reduced. The storage region 16 may be provided so as to cover the entire lower surface of the base region 14 in each mesa portion 60.

[0081] A P-shaped base region 14 is provided in the mesa portion 61 of the diode portion 80, in contact with the upper surface 21 of the semiconductor substrate 10. A drift region 18 is provided below the base region 14. A storage region 16 may also be provided below the base region 14 in the mesa portion 61.

[0082] In both the transistor section 70 and the diode section 80, an N-type buffer section 20 may be provided between the drift section 18 and the lower surface 23. The buffer section 20 may be provided in contact with the drift section 18. The buffer section 20 may function as a field stop layer that prevents the depletion layer extending from the lower end of the base section 14 from reaching the P+-type collector section 22 and the N+-type cathode section 82. The doping concentration distribution of the buffer section 20 will be described later.

[0083] In the transistor section 70, a P+ type collector region 22 is provided below the buffer region 20. The acceptor concentration of the collector region 22 is higher than that of the base region 14. The collector region 22 may contain the same acceptors as the base region 14, or it may contain different acceptors. The acceptors of the collector region 22 are, for example, boron.

[0084] In the diode section 80, an N+ type cathode region 82 is provided below the buffer region 20. The donor concentration of the cathode region 82 is higher than that of the drift region 18. The donor of the cathode region 82 is, for example, hydrogen or phosphorus. However, the elements that serve as donors and acceptors for each region are not limited to the examples described above. The collector region 22 and the cathode region 82 are exposed to the lower surface 23 of the semiconductor substrate 10 and are connected to the collector electrode 24. The collector electrode 24 may be in contact with the entire lower surface 23 of the semiconductor substrate 10. The emitter electrode 52 and the collector electrode 24 are made of a metallic material such as aluminum.

[0085] One or more gate trenches 40 and one or more dummy trenches 30 are provided on the upper surface 21 of the semiconductor substrate 10. Each trench extends from the upper surface 21 of the semiconductor substrate 10, through the base region 14, and down to below the base region 14. In regions where at least one of the emitter region 12, contact region 15, and storage region 16 is provided, each trench also penetrates these doping regions. The statement that a trench penetrates a doping region is not limited to manufacturing in the order of forming the doping region before forming the trench. Manufacturing in which doping regions are formed between the trenches after the trenches have been formed is also included in the statement that a trench penetrates a doping region.

[0086] As described above, the transistor section 70 is provided with a gate trench section 40 and a dummy trench section 30. The diode section 80 is provided with a dummy trench section 30, but not with a gate trench section 40. In this example, the boundary between the diode section 80 and the transistor section 70 in the X-axis direction is the boundary between the cathode region 82 and the collector region 22.

[0087] The gate trench portion 40 has a gate trench, a gate insulating film 42, and a gate conductive portion 44 provided on the upper surface 21 of the semiconductor substrate 10. The gate insulating film 42 is provided covering the inner wall of the gate trench. The gate insulating film 42 may be formed by oxidizing or nitriding the semiconductor on the inner wall of the gate trench. The gate conductive portion 44 is provided inside the gate trench, on the inside of the gate insulating film 42. In other words, the gate insulating film 42 insulates the gate conductive portion 44 from the semiconductor substrate 10. The gate conductive portion 44 is formed of a conductive material such as polysilicon.

[0088] The gate conductive portion 44 may be provided to be longer than the base region 14 in the depth direction. The gate trench portion 40 in this cross-section is covered by an interlayer insulating film 38 on the upper surface 21 of the semiconductor substrate 10. The gate conductive portion 44 is electrically connected to the gate wiring. When a predetermined gate voltage is applied to the gate conductive portion 44, a channel formed by an electron inversion layer is formed on the surface layer of the interface of the base region 14 that is in contact with the gate trench portion 40.

[0089] The dummy trench portion 30 may have the same structure as the gate trench portion 40 in its cross-section. The dummy trench portion 30 has a dummy trench, a dummy insulating film 32, and a dummy conductive portion 34 provided on the upper surface 21 of the semiconductor substrate 10. The dummy conductive portion 34 is electrically connected to the emitter electrode 52. The dummy insulating film 32 is provided covering the inner wall of the dummy trench. The dummy conductive portion 34 is provided inside the dummy trench and is provided inside the dummy insulating film 32. The dummy insulating film 32 insulates the dummy conductive portion 34 from the semiconductor substrate 10. The dummy conductive portion 34 may be formed from the same material as the gate conductive portion 44. For example, the dummy conductive portion 34 is formed from a conductive material such as polysilicon. The dummy conductive portion 34 may have the same length as the gate conductive portion 44 in the depth direction.

[0090] In this example, the gate trench portion 40 and the dummy trench portion 30 are covered by an interlayer insulating film 38 on the upper surface 21 of the semiconductor substrate 10. The bottom portions of the dummy trench portion 30 and the gate trench portion 40 may be curved (curved in cross-section) with a downward convex shape.

[0091] Figure 4A shows an example of the doping concentration distribution and hydrogen chemical concentration distribution along the f-f line in Figure 3. The vertical axis in Figure 4A shows the values ​​of the doping concentration and hydrogen chemical concentration. The vertical axis in Figure 4A is on a common logarithmic (log) scale. The horizontal axis in Figure 4A shows the depth position (position in the Z-axis direction) from the reference position (0), with the bottom surface 23 of the semiconductor substrate 10 as the reference position. The horizontal axis in Figure 4A is on a linear scale. The f-f line is a line that passes through the depth direction from the collector region 22 to the emitter region 12. In Figure 4A, the doping concentration distribution is shown by a solid line, and the hydrogen chemical concentration distribution is shown by a dashed line.

[0092] The doping concentration distribution has peaks in the emitter region 12, base region 14, storage region 16, and collector region 22, respectively. The semiconductor substrate 10 as a whole contains bulk donor. The concentration of the bulk donor is denoted as BD. The doping concentration in the drift region 18 may be the same as or different from the bulk donor concentration BD. The buffer region 20 has a higher doping concentration than the drift region 18.

[0093] The buffer region 20 in this example has a reduction section 210. The reduction section 210 is a region where the doping concentration decreases monotonically from the lower surface 23 to the upper surface 21 of the semiconductor substrate 10. A monotonically decreasing doping concentration means that there are no regions where the doping concentration increases from the lower surface 23 to the upper surface 21. In Figure 4A, more than 80% of the depth of the buffer region 20 is the reduction section 210. The reduction section 210 does not need to have a peak in doping concentration.

[0094] In this example, the reduction portion 210 contains hydrogen. In other words, the buffer region 20 has hydrogen in the portion corresponding to the reduction portion 210. The entire buffer region 20 may contain hydrogen. In the buffer region 20 of this example, the chemical concentration of hydrogen also decreases monotonically in the reduction portion 210. The buffer region 20 of this example contains a hydrogen donor. The reduction portion 210 of this example contains a hydrogen donor. In the buffer region 20, the hydrogen chemical concentration may be higher than the doping concentration. In the reduction portion 210, the hydrogen chemical concentration may be higher than the doping concentration. In this example, the hydrogen chemical concentration is higher than the doping concentration by the amount of the activation rate.

[0095] Let α ( / cm) be the logarithmic gradient of the doping concentration distribution from the lower surface 23 to the upper surface 21. The logarithmic gradient α is the slope when the horizontal axis is the linear depth position from the lower surface 23 and the vertical axis is the doping concentration on a common logarithmic (log) scale.

[0096] The logarithmic gradient α may be the slope of an approximate line obtained by approximating the reduction section 210 with a straight line. The approximation method may be the least squares method. Alternatively, the logarithmic gradient α may be the slope of a straight line connecting two points in the reduction section 210. The two points may be the two ends of the reduction section 210, or any points within the reduction section 210 that are at a distance of half or one-quarter of the length of the reduction section 210. As an example, the logarithmic gradient α of the straight line connecting the first position and the second position may be calculated by setting the depth from the bottom surface 23 of the first position to a (cm) and the doping concentration to A ( / cm). ３ ) and the depth from the lower surface 23 of the second depth position is b (cm), and the doping concentration is B ( / cm). ３ If we let ), then it is given by the following equation: α = (log １０ A-log １０ B) / (a ​​- b) For example, if a = 0.0005 cm (5 μm), b = 0.004 (40 μm), A = 1 × 10 １５ / cm ３ B = 2 × 10 １４ / cm ３ Therefore, the logarithmic gradient α between the two positions is calculated using the following formula: α = (log) １０ 1 x 10 １５ -log １０ 2 x 10 １４ ) / (0.0005 - 0.004) ≈ (15 - 14.3) / (-0.0035) ≈ -200

[0097] In the decreasing section 210, the logarithmic gradient α may be less than 0 and greater than or equal to -250. Alternatively, taking the absolute value, the absolute value of the logarithmic gradient |α| in the decreasing section 210 may be greater than 0 and less than or equal to 250. In the decreasing section 210, the doping concentration decreases from the lower surface 23 to the upper surface 21, so the logarithmic gradient α is a negative value. In the decreasing section 210, the logarithmic gradient α may be less than or equal to -10, less than or equal to -50, less than or equal to -100, or less than or equal to -150. In the decreasing section 210, the logarithmic gradient α may be greater than or equal to -200, or greater than or equal to -180. Alternatively, in the decreasing section 210, the absolute value of the logarithmic gradient |α| may be greater than or equal to 10, greater than or equal to 50, greater than or equal to 100, or greater than or equal to 150. In the decreasing section 210, the absolute value of the logarithmic gradient |α| may be 200 or less, or 180 or less.

[0098] The depth of the reduction portion 210 may be 15 μm or more. The depth of the reduction portion 210 may be 20 μm or more, 30 μm or more, or 40 μm or more. The depth of the reduction portion 210 may be less than or equal to the thickness of the semiconductor substrate 10. The length of the reduction portion 210 may be less than or equal to half the thickness of the semiconductor substrate 10, less than or equal to one-quarter, less than or equal to 150 μm, or less than or equal to 100 μm. The reduction portion 210 may be formed in the depth direction over more than half of the buffer region 20, less than or equal to 70%, less than or equal to 80%, or less than or equal to 90%. The depth of the reduction portion 210 or its proportion may be for a single reduction portion 210.

[0099] The reduction portion 210 may be a region where the logarithmic gradient α or |α| in the depth direction is within the above range. The reduction portion 210 may be a region where the portion where the logarithmic gradient α or |α| is within the above range is continuous. At the position in contact with the end of the reduction portion 210, the logarithmic gradient α or |α| may be outside the above range. The reduction portion 210 may be a region without an inflection point. An inflection point is, for example, the point where the sign of the second derivative changes when the doping concentration is differentiated twice at the depth position.

[0100] In this example, the reduction section 210 has a slope-shaped doping concentration distribution over a predetermined length. Therefore, when the depletion layer extending from the lower end of the base region 14 reaches the buffer region 20, oscillation of the voltage or current waveform can be suppressed.

[0101] The logarithmic gradient α of the reduction section 210 may be smaller than the logarithmic gradient α of the doping concentration distribution from the lower surface 23 to the upper surface 21 in the drift region 18. Alternatively, the absolute value |α| of the logarithmic gradient of the reduction section 210 may be larger than the absolute value |α| of the logarithmic gradient of the doping concentration distribution from the lower surface 23 to the upper surface 21 in the drift region 18. In this example, the logarithmic gradient of the drift region 18 is 0. The absolute value of the logarithmic gradient α of the reduction section 210 may be 2 times or more, 5 times or more, or 10 times or more the absolute value of the logarithmic gradient α of the drift region 18. The absolute value of the logarithmic gradient α of the reduction section 210 may be 1000 times or less the absolute value of the logarithmic gradient α of the drift region 18.

[0102] Let β ( / cm) be the logarithmic gradient of the hydrogen concentration distribution from the lower surface 23 to the upper surface 21. The logarithmic gradient β or its absolute value |β| may be determined by replacing the doping concentration distribution in the case of a logarithmic gradient α or its absolute value |α| with the hydrogen chemical concentration distribution. In the reduction section 210, the logarithmic gradient β may be within the same range as the logarithmic gradient α. In the reduction section 210, the absolute value of the logarithmic gradient |β| may be within the same range as the absolute value of the logarithmic gradient |α|. In the reduction section 210, the difference between the logarithmic gradient α and the logarithmic gradient β, or the difference between their absolute values ​​|α| and |β|, may be within 30% of the logarithmic gradient α or its absolute value |α|. That is, in the reduction section 210, the hydrogen chemical concentration distribution may also decrease in the same way as the doping concentration distribution.

[0103] The buffer region 20 may have a lower peak 212. The lower peak 212 is a doping concentration peak located on the lower surface 23 side of the reduction region 210. The lower peak 212 may or may not be in contact with the reduction region 210. The boundary between the reduction region 210 and the lower peak 212 may be a position where the logarithmic gradient α or its absolute value |α| falls outside the range of the reduction region 210. In Figure 4A, this boundary is indicated by z2.

[0104] A doping concentration peak of the collector region 22 or cathode region 82 may be located between the lower peak 212 and the bottom surface 23. Figure 4A shows the doping concentration peak of the collector region 22. No other doping concentration peaks are required between the lower peak 212 and this doping concentration peak. In other words, the lower peak 212 may be the doping concentration peak located on the bottom surface 23 side in the buffer region 20. In Figure 4A, z1 indicates the boundary (PN junction portion) between the lower peak 212 and the doping concentration peak of the collector region 22.

[0105] The buffer region 20 may have an upper peak 214. The upper peak 214 is a doping concentration peak located on the upper surface 21 side of the reduction region 210. The upper peak 214 may or may not be in contact with the reduction region 210. The boundary between the reduction region 210 and the upper peak 214 may be a position where the logarithmic gradient α or its absolute value |α| falls outside the range of the reduction region 210. Alternatively, the boundary may be a position where the logarithmic gradient α or its absolute value |α| is 0. In Figure 4A, this boundary is indicated by z3.

[0106] There does not need to be a doping concentration peak between the upper peak 214 and the drift region 18. In other words, the upper peak 214 may be the doping concentration peak located on the uppermost surface 21 side in the buffer region 20. The boundary between the upper peak 214 and the drift region 18, that is, the boundary between the buffer region 20 and the drift region 18, may be a position where the logarithmic gradient α or its absolute value |α| is within the range of the logarithmic gradient α or its absolute value |α| of the decreasing portion 210. In another example, the boundary may be a position where the doping concentration is 1.2 times that of the doping concentration in the drift region 18. In Figure 4A, the boundary between the upper peak 214 and the drift region 18 is indicated by z4.

[0107] Let P1 be the doping concentration at the upper peak 214, and P2 be the doping concentration at the lower peak 212. The peak doping concentration may be the doping concentration at the peak's highest point. Doping concentration P1 may be less than doping concentration P2. Doping concentration P1 may be less than or equal to one-tenth of doping concentration P2, and may be greater than or equal to one-hundredth. Doping concentration P1 is 1.0 × 10⁻⁶ １５ / cm ３ The following may apply: Doping concentration P1 may be greater than or equal to bulk donor concentration BD. Doping concentration P2 may be 1.0 × 10⁻⁶ １５ / cm ３ The above is 1.0 x 10 １７ / cm ３ The following is acceptable:

[0108] Let A1 be the average value of the doping concentration in the reduction section 210. The doping concentration P2 may be greater than the average value A1. The doping concentration P1 may be less than the average value A1. This reduces the doping concentration at the peak on the upper surface 21 side of the buffer region 20, thus further suppressing oscillation. However, the doping concentration P1 may be greater than the average value A1. The average value A1 is 1.0 × 10⁻⁶. １４ / cm ３ The above is 1.0 x 10 １５ / cm ３ The following is acceptable:

[0109] The buffer region 20 does not need to have doping concentration peaks other than the upper peak 214 and the lower peak 212. The buffer region 20 may have two peaks, or fewer than two, as will be described later, or even just one. Using the symbols shown in Figure 4A, the length of the buffer region 20 in the depth direction is z4-z1, and the length of the reduction portion 210 in the depth direction is z3-z2.

[0110] The buffer region 20 may have a lower hydrogen peak 312 in its hydrogen chemical concentration distribution. The lower hydrogen peak 312 is a peak of hydrogen chemical concentration located on the lower surface 23 side of the reduction portion 210. The lower hydrogen peak 312 may be in contact with the reduction portion 210, and a portion of it may be within the range of the reduction portion 210. In other examples, the lower hydrogen peak 312 may not be in contact with the reduction portion 210.

[0111] The lower hydrogen peak 312 and the lower peak 212 may be corresponding peaks. Correspondence between the lower hydrogen peak 312 and the lower peak 212 means that the apex of the other concentration peak is located within the full width at half maximum of one concentration peak. When the lower hydrogen peak 312 and the lower peak 212 correspond, the distance between the apexes of the two concentration peaks may be 2 μm or less, 1 μm or less, or 0.5 μm or less. Furthermore, the apex position of the lower hydrogen peak 312 and the apex position of the lower peak 212 may be equal. In Figure 4A, the apex position of the lower hydrogen peak 312 is indicated by z5.

[0112] In the buffer region 20, there does not need to be any hydrogen chemical concentration peaks below the lower hydrogen peak 312 on the lower surface 23 side. In other words, the lower hydrogen peak 312 may be the hydrogen chemical concentration peak located furthest down on the lower surface 23 side in the buffer region 20, and there does not need to be any other hydrogen chemical concentration peaks between the lower hydrogen peak 312 and the collector region 22 or cathode region 82.

[0113] The buffer region 20 may have an upper hydrogen peak 314 in terms of hydrogen chemical concentration. The upper hydrogen peak 314 is a peak of hydrogen chemical concentration located on the upper surface 21 side of the reduction portion 210. The upper hydrogen peak 314 may be in contact with the reduction portion 210, and a portion of it may be within the range of the reduction portion 210. In other examples, the upper hydrogen peak 314 may not be in contact with the reduction portion 210. The upper hydrogen peak 314 and the upper peak 214 may be corresponding peaks. In Figure 4A, the apex position of the upper hydrogen peak 314 is indicated by z6.

[0114] In the buffer region 20, there does not need to be any hydrogen chemical concentration peaks on the upper surface 21 side of the upper hydrogen peak 314. In other words, the upper hydrogen peak 314 may be the hydrogen chemical concentration peak located furthest towards the upper surface 21 in the buffer region 20, and there does not need to be any other hydrogen chemical concentration peaks between the upper hydrogen peak 314 and the drift region 18.

[0115] On the upper surface 21 side of the buffer region 20, the hydrogen chemical concentration may be lower than the bulk donor concentration BD. In this example, in the drift region 18, the hydrogen chemical concentration and the bulk donor concentration BD are equal, and on the upper surface 21 side of that, the hydrogen chemical concentration is lower than the bulk donor concentration BD.

[0116] Figure 4B shows another example of the doping concentration distribution and hydrogen chemical concentration distribution along the f-f line in Figure 3. The doping concentration distribution in this example does not have the upper peak 214 compared to the doping concentration distribution shown in Figure 4A. Similarly, the hydrogen chemical concentration distribution in this example does not have the upper hydrogen peak 314 compared to the hydrogen chemical concentration distribution shown in Figure 4A. Other aspects are the same as in Figure 4A. The same effect as in Figure 4A can be obtained with this type of distribution as well.

[0117] Figure 4C shows another example of the doping concentration distribution and hydrogen chemical concentration distribution along the f-f line in Figure 3. In this example, the doping concentration distribution has an upper peak 214, but the hydrogen chemical concentration distribution does not have an upper hydrogen peak 314. Otherwise, it is the same as in Figure 4A. The same effect as in Figure 4A can be obtained with this type of distribution as well.

[0118] Figure 5 shows an example of the manufacturing process for the semiconductor device 100 in this embodiment. The manufacturing process in this example comprises a first injection step S1010, a second injection step S1020, and an annealing step S1030. Figure 5 shows a part of the manufacturing process.

[0119] Figure 6A shows the first injection process S1010. Prior to the first injection process S1010, the element structure shown in Figure 3, etc., may be formed on the upper surface 21 of the semiconductor substrate 10. In Figure 6A and the subsequent manufacturing process diagrams, the upper surface 21 is omitted. Also, prior to the first injection process S1010, backgrinding may be performed to process the semiconductor substrate 10 to a desired thickness, and a collector region 22 or a cathode region 82 or both may be formed.

[0120] In the first implantation step S1010, hydrogen ions are implanted from the lower surface 23 of the semiconductor substrate 10 to a first depth position z11. The first depth position z11 may be located inside the semiconductor substrate 10. The hydrogen ions may be protons, deutrons, or tritons. In this example, they are protons. The first depth position z11 is the implantation depth from the lower surface 23. In Figure 6A, protons implanted near the first depth position z11 are indicated by an "x". The first depth position z11 may be 30 μm or more away from the lower surface 23. The first depth position z11 may be located closer to the lower surface 23 than the center in the depth direction of the semiconductor substrate 10. The implantation position of the hydrogen ions can be controlled by the acceleration energy used to accelerate the hydrogen ions.

[0121] In this specification, the region through which the implanted hydrogen ions pass is referred to as the passing region 202. In the example in Figure 6A, the passing region 202 extends from the bottom surface 23 to the first depth position z11. In the semiconductor substrate 10, lattice defects mainly consisting of vacancies, such as single-atom vacancies (V) and double-atom vacancies (VV), are formed in the passing region 202 through which the hydrogen ions pass. Atoms adjacent to vacancies have dangling bonds. Lattice defects also include interstitial atoms and dislocations, and in a broad sense, donors and acceptors may also be included, but in this specification, lattice defects mainly consisting of vacancies may be referred to as vacancy-type lattice defects, vacancy-type defects, or simply lattice defects. Furthermore, the formation of many lattice defects by hydrogen ion implantation into the semiconductor substrate 10 can strongly disrupt the crystallinity of the semiconductor substrate 10.

[0122] Figure 6B shows the second implantation step S1020. In the second implantation step S1020, hydrogen ions are implanted from the bottom surface 23 to a second depth position z21. The second depth position z21 is located closer to the bottom surface 23 than the first depth position z11. In this example as well, the hydrogen ions are protons. The second depth position z21 is the implantation depth from the bottom surface 23. The second depth position z21 may be closer to the bottom surface 23 of the semiconductor substrate 10 than the first depth position z11. The second depth position z21 may be within 5 μm from the bottom surface 23.

[0123] The proton dose in the second injection step S1020 may be greater than the proton dose in the first injection step S1010, and may be 10 times or more. The dose in the second injection step S1020 may be 1000 times or less than the dose in the first injection step S1010, and may be 100 times or less. The hydrogen injected in the second injection step S1020 becomes a source of hydrogen donors in the buffer region 20.

[0124] Figure 6C shows the annealing process S1030. In the annealing process S1030, the semiconductor substrate 10 is annealed. For example, the annealing temperature is 370°C and the annealing time is 5 hours. As a result, the hydrogen injected into the second depth position z21 diffuses toward the upper surface 21. At that time, the hydrogen (H), the oxygen (O) contained in the semiconductor substrate 10, and the vacancies (V) formed in the passage region 202 interact to form the hydrogen donor described above. As a result, the doping concentration in the passage region 202 increases, and a reduction portion 210 is formed. In addition, a lower peak 212 of the doping concentration distribution may be formed near the second depth position z21, and an upper peak 214 of the doping concentration distribution may be formed near the first depth position z11.

[0125] Conventionally, a method for forming a buffer region 20 using selenium as a dopant is known, but selenium requires thermal diffusion at a high temperature of 900°C. On the other hand, protons form hydrogen donors through annealing at relatively low temperatures. Therefore, it is possible to promote the refinement of the surface structure, mainly the gate structure, before the formation of the buffer region 20. According to the manufacturing method of this example, a buffer region 20 having a slope-shaped doping concentration distribution can be formed using protons. In addition, the number of proton injections can be reduced.

[0126] The first depth position z11 and the apex position z6 of the upper hydrogen peak 314 shown in Figures 4A to 4C may be equal, and the first depth position z11 may be located within the half-width range of the upper hydrogen peak 314. The second depth position z21 and the apex position z5 of the lower hydrogen peak 312 shown in Figures 4A to 4C may be equal, and the second depth position z21 may be located within the half-width range of the lower hydrogen peak 312. Furthermore, the first injection step S1010 may be performed after the second injection step S1020. Also, the dopant injected in the first injection step S1010 is not limited to protons. For example, a pass-through region 202 may be formed by injecting helium in the first injection step S1010.

[0127] Figure 7A shows an example of the carrier concentration distribution in the buffer region 20 at different dose levels. In Figure 7A, the vertical axis represents the carrier concentration on a common logarithmic (log) scale. In Figure 7A, the horizontal axis represents the depth position on a linear scale. In the concentration distribution graph of Figure 7A, the lower end of the buffer region 20 is set as the reference position (0) in the depth direction.

[0128] Figure 7A shows that in the first injection step S1010, the first depth position z11 is 50 μm and the proton dose is 3.0 × 10⁻¹⁰ １３ / cm ２ The carrier concentration distribution in this case is shown. A decreasing region 210 with a gradient is formed from just before 10 μm to above 40 μm. The decreasing region 210 in this example has a good shape in terms of length and gradient. In addition, a lower peak 212 is formed on the lower surface 23 side of the decreasing region 210, and an upper peak 214 is formed on the upper surface 21 side of the decreasing region 210. The upper peak 214 has its peak near 50 μm. On the upper surface 21 side of the upper peak 214, the carrier concentration decreases to the bulk donor concentration and becomes a drift region 18.

[0129] Figure 7B shows an example of the carrier concentration distribution in the buffer region 20 at different dose amounts. In Figure 7B, in the first injection step S1010, the proton dose amount is 1.0 × 10 １３ / cm ２ This shows the carrier concentration distribution under these conditions. In other words, the dose is smaller than in the case of Figure 7A. The injection depth is the same as in the case of Figure 7A.

[0130] The carrier concentration distribution in this example also has a reduction section 210. However, the logarithmic gradient α of the reduction section 210 in this example is greater (closer to zero) than the logarithmic gradient α of the reduction section 210 in Figure 7A. Alternatively, the absolute value |α| of the logarithmic gradient of the reduction section 210 in this example is less than or equal to 1 / 10 of the absolute value |α| of the logarithmic gradient of the reduction section 210 in Figure 7A, and is therefore close to zero. In other words, the reduction section 210 in this example is closer to flatter than the reduction section 210 in Figure 7A. Furthermore, a clear upper peak 214 is not observed in the carrier concentration distribution in this example. Moreover, the reduction section 210 in this example is formed beyond the proton injection depth of 50 μm to the upper surface 21 side.

[0131] If the dose in the first injection step S1010 is small, the amount of defects formed in the passage region 202 will also be small, making it easier for hydrogen to diffuse from the second depth position z21 toward the upper surface 21. As a result, the logarithmic gradient α of the reduction section 210 is large (close to zero). In addition, hydrogen diffuses beyond the first depth position z11 toward the upper surface 21, and the reduction section 210 is formed beyond the first depth position z11 toward the upper surface 21. It is thought that, for example, a thermal donor is formed beyond the first depth position z11 toward the upper surface 21.

[0132] Figure 7C shows an example of the carrier concentration distribution in the buffer region 20 at different dose amounts. In Figure 7C, in the first injection step S1010, the proton dose amount is 3.0 × 10 １２ / cm ２ This shows the carrier concentration distribution in that case. In other words, the dose is less than in the case of Figure 7B. The injection depth is the same as in the cases of Figures 7A and 7B.

[0133] The logarithmic gradient α of the reduction section 210 in this example is larger (closer to zero) than in the case of Figure 7B. Alternatively, the absolute value |α| of the logarithmic gradient of the reduction section 210 in this example is less than or equal to 1 / 10 of the value in the case of Figure 7B, and is close to zero. In other words, the reduction section 210 in this example is closer to flatter than the reduction section 210 in Figure 7B. Furthermore, the reduction section 210 is formed to the upper surface 21 side beyond the first depth position z11. That is, the effect of hydrogen diffusion becoming easier due to the smaller dose amount is more apparent than in the case of Figure 7B. In other words, if the dose amount in the first injection step S1010 is small, the gradient becomes relatively larger (the absolute value of the gradient becomes smaller and closer to zero). Also, it becomes more difficult to control the length of the buffer region 20.

[0134] Figure 8A shows an example of the carrier concentration distribution in the buffer region 20 at different dose amounts. In Figure 8A, the first depth position z11 is 100 μm. The graph is displayed in the same way as in Figure 7A. In Figure 8A, in the first injection step S1010, the proton dose amount is 1.0 × 10⁻⁶. １３ / cm ２ This shows the carrier concentration distribution under those conditions.

[0135] In this example, a decreasing section 210 with a gradient is formed from just before 10 μm to just before 90 μm. The decreasing section 210 in this example has a good shape in terms of length and gradient. Furthermore, a lower peak 212 is formed on the lower surface 23 side of the decreasing section 210, and an upper peak 214 is formed on the upper surface 21 side of the decreasing section 210. The upper peak 214 has its peak near 100 μm. On the upper surface 21 side of the upper peak 214, the carrier concentration decreases to the bulk donor concentration, becoming a drift region 18.

[0136] Figure 8B shows an example of the carrier concentration distribution in the buffer region 20 at different dose amounts. In Figure 8B, in the first injection step S1010, the proton dose amount is 3.0 × 10 １２ / cm ２ This shows the carrier concentration distribution under these conditions. In other words, the dose is smaller than in the case of Figure 8A. The injection depth is the same as in the case of Figure 8A.

[0137] The carrier concentration distribution in this example also has a reduction section 210. However, the logarithmic gradient α of the reduction section 210 in this example is greater (closer to zero) than the logarithmic gradient α of the reduction section 210 in Figure 8A. In other words, the absolute value |α| of the logarithmic gradient of the reduction section 210 in this example is smaller than the absolute value |α| of the logarithmic gradient of the reduction section 210 in Figure 8A, and is closer to zero. That is, the reduction section 210 in this example is closer to flatter than the reduction section 210 in Figure 8A. Furthermore, a clear upper peak 214 is not observed in the carrier concentration distribution in this example. The reduction section 210 in this example is formed beyond the proton injection depth of 100 μm to the upper surface 21 side. This is thought to be because, as mentioned above, the dose is small, so the amount of defects formed in the passage region 202 is also small, and hydrogen diffuses more easily.

[0138] Figure 8C shows an example of the carrier concentration distribution in the buffer region 20 at different dose amounts. In Figure 8C, in the first injection step S1010, the proton dose amount is 3.0 × 10⁻⁶. １３ / cm ２ This shows the carrier concentration distribution under these conditions. In other words, the dose is greater than in the case of Figure 8A. The injection depth is the same as in the cases of Figures 8A and 8B.

[0139] The carrier concentration distribution in this example also has a decreasing section 210 with a gradient. The logarithmic gradient α of the decreasing section 210 in this example is smaller than the logarithmic gradient α of the decreasing section 210 in Figure 8A. In other words, the absolute value |α| of the logarithmic gradient of the decreasing section 210 in this example is larger than the absolute value |α| of the logarithmic gradient of the decreasing section 210 in Figure 8A. That is, the doping concentration decreases more steeply in the decreasing section 210 in this example than in the decreasing section 210 in Figure 8A. Furthermore, the buffer region 20 in this example has a lower peak 212 and an upper peak 214. However, in the buffer region 20 in this example, a valley 216 is formed between the decreasing section 210 and the upper peak 214. The carrier concentration in the valley 216 may be lower than the carrier concentration in the drift region 18.

[0140] If the dose amount in the first injection step S1010 is large, the amount of defects formed in the passage region 202 will also increase, making it difficult for hydrogen to diffuse from the second depth position z21 to the upper surface 21. As a result, the logarithmic gradient α of the reduction section 210 becomes smaller, and the absolute value of the logarithmic gradient |α| becomes larger. However, since hydrogen does not diffuse to the first depth position z11, a region where defects do not recover is created. The remaining defects create a valley 216 in the carrier concentration distribution. In other words, if the dose amount is large, defects may remain. Note that the dose amount of protons to the second depth position z21 in the second injection step S1020 is the same in all figures from 7A to 7C and from 8A to 8C.

[0141] Figure 9 shows the relationship between proton injection depth and dose. The horizontal axis of Figure 9 represents the proton injection depth, and the vertical axis represents the proton dose. The proton injection depth shown in Figure 9 is the first depth position z11 from the bottom surface 23 in the first injection step S1010. The proton dose shown in Figure 9 is the proton dose in the first injection step S1010. The proton injection depth on the horizontal axis is a linear scale with the bottom surface 23 as the reference position (0). The proton dose on the vertical axis is a common logarithmic (log) scale.

[0142] The center line in the figure represents the proton injection depth of 50 μm and dose of 3.0 × 10⁻¹⁴ as shown in Figure 7A. １３ / cm ２ The plot and the proton injection depth of 100 μm and dose of 1.0 × 10⁻¹⁰ shown in Figure 8A １３ / cm ２ This is a straight line connecting the plotted points. In other words, the center line shows the relationship between the injection depth and dose amount when the logarithmic gradient α (or the absolute value of the logarithmic gradient |α|) and length of the reduction portion 210 formed in the buffer region 20 are good. In other words, by setting the conditions near the center line in the first injection step S1010, a buffer region 20 with a good shape can be formed. If the proton injection depth is x (μm) and the proton dose amount is y ( / cm2), then the center line is y = 9.00 × 10 １３ ×exp(-2.20×10 －２ It is represented by x).

[0143] As described above, when the proton dose decreases, the logarithmic gradient α of the reduction section 210 increases (or the absolute value of the logarithmic gradient |α| decreases), and it becomes difficult to control the length of the buffer region 20. For this reason, it is preferable to set a lower limit on the proton dose. In Figure 9, y = 2.70 × 10 １３ ×exp(-2.20×10 －２ The line represented by x) is the lower limit line. The slope of the lower limit line is equal to that of the center line.

[0144] As described above, if the proton dose is too large, a defect will remain between the decrease portion 210 and the upper peak 214. Therefore, it is preferable to set an upper limit on the proton dose. In Figure 9, y = 2.70 × 10 １４ ×exp(-2.20×10 －２ The line represented by x) is the upper limit line. The slope of the upper limit line is equal to that of the center line.

[0145] The proton dose amount in the first injection step S1010 may be a value between the lower limit and the upper limit. That is, the proton dose amount y in the first injection step S1010 is y = 2.70 × 10 １３ ×exp(-2.20×10 －２ x) Above, y = 2.70 × 10 １４ ×exp(-2.20×10 －２ x) may be less than or equal to this. This makes it possible to form a buffer region 20 that has a good reduction section 210 with a logarithmic gradient α (or the absolute value of the logarithmic gradient |α|) and length, and with a small amount of remaining defects.

[0146] According to the above formula, if the proton injection depth x is 50 μm, the proton dose y is 9.0 × 10⁻⁶. １２ ( / cm ２ ) Above, 9.0 x 10 １３ ( / cm ２ ) The result is as follows. According to the above formula, when the proton injection depth x is 100 μm, the proton dose y is 3.0 × 10 １２ ( / cm ２ ) Above, 3.0 x 10 １３ ( / cm ２ ) The following applies.

[0147] The intercept of the lower limit line may be 3.00×10 １３ and may be 4.00×10 １３ and may be 5.00×10 １３ The intercept of the upper limit line may be 2.50×10 １４ and may be 2.00×10 １４ and may be 1.00×10 １４ In this specification, the dose amount refers to the number of ions per unit area injected into the semiconductor substrate 10 when ion implantation is performed. The dose amount of a predetermined region of the semiconductor substrate 10 can be the integrated concentration obtained by integrating the chemical concentration of the implanted ions over the depth direction of the predetermined region. The unit of the integrated concentration is / cm

[0148] Therefore, the dose amount and the integrated concentration may be treated as the same. That is, in the relationship between the proton implantation depth and the proton dose amount described above, the proton implantation depth x is the depth position x (μm) from the lower surface 23 of the upper hydrogen peak 314 shown in FIG. 4A, and the proton dose amount y is the integrated concentration y ( / cm ２ obtained by integrating the hydrogen chemical concentration distribution at the upper hydrogen peak 314). The depth position x from the lower surface 23 of the upper hydrogen peak 314 may be the apex position z6 of the upper hydrogen peak 314. The integrated concentration y may be obtained by integrating between both ends of the upper hydrogen peak 314. When both ends are unknown, it may be obtained by integrating within the range of the full width at half maximum of the upper hydrogen peak 314. ２ ).

[0149] FIG. 10A is a diagram showing the carrier concentration distribution in the buffer region 20 in the comparative example. The vertical axis in FIG. 10A shows the value of the carrier concentration on a scale of common logarithm (log). The horizontal axis in FIG. 10A shows the depth position on a linear scale. In the graph of the concentration distribution in FIG. 10A, the lower end position of the buffer region 20 is taken as the reference position (0) in the depth direction. FIG. 10A shows the carrier concentration distributions of the buffer regions 20 formed on different types of semiconductor substrates 10.

[0150] In this comparative example, in order to form a distribution close to the distribution of the buffer region 20 when selenium is used as a dopant, protons are implanted in four steps. However, in the carrier concentration distribution of this comparative example, the four peaks are discrete, and most of the buffer region 20 does not have a slope-like distribution. Therefore, there is a possibility that oscillation may occur when the depletion layer extends and reaches the peak. In particular, since the concentration of the deepest peak in the fourth step is high, oscillation may occur when the depletion layer reaches the peak. Further, by increasing the number of implantation times, the carrier concentration distribution can be made closer to a slope shape, but the productivity decreases and the implantation cost increases.

[0151] FIG. 10B is a diagram showing the carrier concentration distribution of the buffer region 20 in another comparative example. The display of the graph is the same as in FIG. 10A. In this comparative example, in order to suppress oscillation, the peak concentrations in the third and fourth steps are reduced. However, in that case, it becomes easy to be affected by the oxygen concentration and the carbon concentration of the semiconductor substrate 10, and a large difference occurs in the carrier concentration depending on the type of the semiconductor substrate 10 used. In the semiconductor device 100 of the embodiment, a slope-like distribution in which the carrier concentration continuously changes can be formed using protons. Further, the number of proton implantation times can be suppressed. Therefore, oscillation can be suppressed while improving productivity.

[0152] FIG. 11 is a diagram showing the carrier concentration distribution of the buffer region 20 in semiconductor substrates 10 having different oxygen concentrations. The vertical axis in FIG. 11 shows the value of the carrier concentration on a scale of common logarithm (log). The horizontal axis in FIG. 11 shows the depth position on a linear scale. In the concentration distribution graph of FIG. 11, the lower end position of the buffer region 20 is used as a reference position (0) in the depth direction. In FIG. 11, the shape of the decreasing portion 210 in the buffer region 20 is shown enlarged.

[0153] In FIG. 11, the case where the semiconductor substrate 10 having an oxygen concentration of 1.0×10 １８ / cm ３ is shown by a solid line, the case where the semiconductor substrate 10 having an oxygen concentration of 7.8×10 １７ / cm ３ is shown by a dashed-dotted line, and the case where the semiconductor substrate 10 having an oxygen concentration of 3.0×10１７ / cm ３ The case using the semiconductor substrate 10 is shown by the dotted line. When the oxygen concentration is low (dotted line), the carrier concentration is almost flat and no reduction portion 210 is formed. As the oxygen concentration increases (dotted line), the reduction portion 210 begins to form, and as the oxygen concentration increases further (solid line), the logarithmic slope of the reduction portion 210 becomes smaller (or the absolute value of the logarithmic slope becomes larger), and its length increases.

[0154] The formation of hydrogen donors in the buffer region 20 can be rate-limited by the lowest concentration of vacancies (V), oxygen (O), and hydrogen (H). In this example, a sufficient amount of vacancies (V) are formed by ion implantation in the first implantation step S1010. When the oxygen concentration is low, the formation of hydrogen donors is rate-limited by oxygen (O), and the concentration distribution of hydrogen donors reflects the concentration distribution of oxygen. Since the oxygen concentration distribution is almost uniform in the semiconductor substrate 10, when the oxygen concentration is low (dotted line), the carrier concentration is almost flat. On the other hand, as the oxygen concentration increases, the formation of hydrogen donors is rate-limited by hydrogen (H), and the concentration distribution of hydrogen donors reflects the concentration distribution of hydrogen. Since hydrogen diffuses from the second depth position z21 toward the upper surface 21, the hydrogen chemical concentration distribution has a slope that decreases from the lower surface 23 toward the upper surface 21 (see Figure 4A, etc.). It is thought that the hydrogen rate-limiting effect becomes stronger as the oxygen concentration increases, and this slope is reflected in the carrier concentration. Note that in Figures 7A to 7C and 8A to 8C, the oxygen concentration is 1.0 × 10⁻⁶. １８ / cm ３ The semiconductor substrate 10 is used.

[0155] Figure 12A shows the relationship between oxygen concentration and the absolute value |α| of the logarithmic gradient α. The vertical axis of Figure 12A shows the absolute value |α| of the logarithmic gradient α ( / cm) of the decreasing section 210 on a linear scale. The horizontal axis of Figure 12A shows the oxygen concentration ( / cm) contained in the semiconductor substrate 10. ３The values ​​are shown on a common logarithmic (log) scale. Figure 12A was created based on the measurement results from Figures 7A to 7C and Figure 11. As described above, the decreasing section 210 does not have a slope when the oxygen concentration is low, but as the oxygen concentration increases, the decreasing section 210 begins to have a slope. The oxygen concentration at the boundary is approximately 7.0 × 10⁻⁶. １７ / cm ３ That is the case.

[0156] When the oxygen concentration exceeds the boundary oxygen concentration, hydrogen becomes the rate-limiting factor, and the hydrogen chemical concentration distribution begins to be reflected, so the absolute value of the logarithmic gradient α |α| is thought to increase rapidly. As the oxygen concentration increases further, the hydrogen chemical concentration distribution is fully reflected in the decreasing section 210, and the increase in the absolute value of the logarithmic gradient α |α| is thought to become more gradual.

[0157] The oxygen concentration contained in the semiconductor substrate 10 is 7.0 × 10 １７ / cm ３ It can be larger than this. This makes it easier to form the reduced portion 210. The oxygen concentration contained in the semiconductor substrate 10 is 8.0 × 10 １７ / cm ３ The above is sufficient, 1.0 × 10 １８ / cm ３ The above is sufficient. This makes it possible to increase the absolute value of the logarithmic gradient α of the reduction section 210. The oxygen concentration is 3.0 × 10 １８ / cm ３ It may be less than the average value of the semiconductor substrate 10.

[0158] Figure 12A shows the absolute values ​​of the logarithmic gradient α for each dose at an injection depth of 50 μm, as shown in Figures 7A to 7C. The absolute value of the logarithmic gradient α increases as the dose increases. This is thought to be because as the dose increases, the amount of defects formed in the passage region 202 increases, making it more difficult for hydrogen to diffuse. In other words, the logarithmic gradient α of the reduction section 210 can be adjusted by adjusting the dose.

[0159] Figure 12B shows the relationship between oxygen concentration and the absolute value of the logarithmic gradient α. Figure 12B was created based on the measurement results from Figures 8A to 8C. The axes of the graph in Figure 12B are the same as in Figure 12A. The absolute value of the logarithmic gradient α shown in Figure 12B is (3.0 × 10) when the dose is small. １２ / cm ２ Except for the case shown in Figure 12A, the absolute value of the logarithmic gradient α is smaller. This is thought to be because the proton injection depth is greater and the length of the reduction section 210 is larger. When the dose is small, fewer defects are formed, so it is thought that the influence of other donors (e.g., CiOi-H) is present. By adjusting the dose according to the injection depth, the logarithmic gradient α of the reduction section 210 or the absolute value of the logarithmic gradient |α| can be adjusted.

[0160] Figure 13 shows the oxygen concentration distribution in the depth direction of the semiconductor substrate 10. The vertical axis of Figure 13 shows the oxygen concentration on a common logarithmic scale (log). The horizontal axis of Figure 13 shows the depth position on a linear scale, with the bottom surface 23 as the reference position (0). Secondary ion mass spectrometry (SIMS) was used to measure the oxygen concentration.

[0161] Figure 13 shows that the oxygen concentration in the depth direction of the semiconductor substrate 10 is almost uniform. The oxygen concentration may be uniform at least from the bottom surface 23 to 50 μm, may be uniform up to 100 μm, may be uniform up to the center of the substrate, or may be uniform throughout the entire substrate. For the oxygen concentration to be uniform over a predetermined range, the variation in oxygen concentration within the predetermined range may be within 30% or 20% of the average value of the oxygen concentration within the predetermined range. The increase in oxygen concentration at depth 0 μm is thought to be due to the instability of the measured values ​​from the start of measurement until the measurement accuracy stabilizes. The oxygen concentration on the surface does not need to be included in the oxygen concentration of the semiconductor substrate 10.

[0162] Figure 14 is a diagram illustrating the approximate straight line of the decreasing section 210. Figure 14 shows the oxygen concentration shown in Figure 11 as 1.0 × 10⁻⁶. １８ / cm ３In this case, the carrier concentration in the decreasing section 210 is approximated by a straight line. The axes in Figure 14 are the same as in Figure 11. In this specification, a distribution shown with a common logarithmic (log) scale on the vertical axis and a linear scale on the horizontal axis is sometimes referred to as a logarithmic distribution. A logarithmic distribution is a semi-logarithmic representation with a common logarithmic vertical axis.

[0163] The dotted line in the figure represents the approximate straight line. The approximate straight line is a straight line that approximates the logarithmic distribution of doping concentrations. The least squares method may be used for the linear approximation. Let c1 be the carrier concentration at a predetermined depth position in the reduction section 210, and let c2 be the value of the approximate straight line at the same depth position. At each position in the reduction section 210, the difference between c1 and c2 may be 30% or less of c2. In other words, the maximum difference between c1 and c2 in the reduction section 210 may be 30% or less of c2. That is, the reduction section 210 may decrease linearly. This can further suppress oscillation. The above ratio may be 20% or less, or 10% or less.

[0164] Although the present invention has been described above using embodiments, the technical scope of the present invention is not limited to the scope described in the above embodiments. It will be apparent to those skilled in the art that various modifications or improvements can be made to the above embodiments. It will be clear from the claims that such modified or improved forms may also be included in the technical scope of the present invention.

[0165] It should be noted that the execution order of operations, procedures, steps, and stages in the devices, systems, programs, and methods shown in the claims, specifications, and drawings is not explicitly stated as "before," "prior to," etc., and that these can be performed in any order unless the output of a previous process is used in a later process. Even if the operation flow in the claims, specifications, and drawings is described using phrases such as "first," "next," etc. for convenience, this does not mean that it is mandatory to perform the operations in that order.

[0166] 10... Semiconductor substrate, 12... Emitter region, 14... Base region, 15... Contact region, 16... Storage region, 17... Well region, 18... Drift region, 20... Buffer region, 21... Top surface, 22... Collector region, 23... Bottom surface, 24... Collector electrode, 29... Straight section, 30... Dummy trench section, 31... Tip section, 32... Dummy insulating film, 34... Dummy conductive section, 38... Interlayer insulating film, 39... Straight section, 40... Gate trench section, 41... Tip section, 42... Gate insulating film, 44... Gate conductive section, 5 2...Emitter electrode, 54...Contact hole, 60...Mesa region, 61...Mesa region, 70...Transistor region, 80...Diode region, 81...Extended region, 82...Cathode region, 90...Edge termination structure, 100...Semiconductor device, 130...Outer gate wiring, 131...Active side gate wiring, 160...Active region, 162...Edge, 164...Gate pad, 202...Passing region, 210...Decrease region, 212...Lower peak, 214...Upper peak, 216...Valley, 312...Lower hydrogen peak, 314...Upper hydrogen peak

Claims

1. A semiconductor device comprising a semiconductor substrate having an upper surface and a lower surface and a drift region of a first conductivity type, wherein the semiconductor substrate has a buffer region of a first conductivity type provided between the drift region and the lower surface, the buffer region having a decreasing portion in which the doping concentration monotonically decreases from the lower surface toward the upper surface, the decreasing portion contains hydrogen, the chemical concentration of hydrogen also monotonically decreases in the decreasing portion, the logarithmic gradient of the distribution of the doping concentration toward the upper surface toward the lower surface toward the upper surface toward the decreasing portion is less than 0 and -250 ( / cm) or more, and the length of the decreasing portion is 15 μm or more and less than or equal to the thickness of the semiconductor substrate.

2. A semiconductor substrate having an upper surface and a lower surface, and having a drift region of a first conductivity type, wherein the semiconductor substrate has a buffer region of a first conductivity type provided between the drift region and the lower surface, having a doping concentration higher than that of the drift region, the buffer region has a decreasing portion in which the doping concentration decreases monotonically from the lower surface toward the upper surface, the decreasing portion contains hydrogen, and the chemical concentration of hydrogen also decreases monotonically in the decreasing portion, the buffer region has an upper hydrogen peak of the chemical concentration of hydrogen on the upper surface side of the decreasing portion, and when the depth position of the upper hydrogen peak from the lower surface is x (μm), the integral concentration obtained by integrating the chemical concentration distribution of hydrogen at the upper hydrogen peak is y ( / cm²). ２ ) is y = 2.70 × 10 １３ ×exp(-2.20×10 －２ x) Above, y = 2.70 × 10 １４ ×exp(-2.20×10 －２ x) Semiconductor device that is less than or equal to the following.

3. The semiconductor device according to claim 1 or 2, wherein at each position of the reduction portion, the difference between the doping concentration and the approximate straight line that approximates the logarithmic distribution of the doping concentration is 30% or less of the value of the approximate straight line.

4. The semiconductor device according to claim 1, wherein the logarithmic gradient is -50 ( / cm) or less.

5. The semiconductor device according to claim 4, wherein the logarithmic gradient is -100 ( / cm) or less.

6. The semiconductor device according to claim 5, wherein the logarithmic gradient is -200 ( / cm) or greater.

7. The semiconductor device according to claim 1, wherein the logarithmic gradient of the reduction portion is smaller than the logarithmic gradient of the doping concentration distribution from the lower surface to the upper surface in the drift region.

8. The oxygen concentration contained in the semiconductor substrate is 7.0 × 10 １７ ( / cm ３ The semiconductor device according to claim 1 or 2, which is larger than ).

9. The oxygen concentration contained in the semiconductor substrate is 8.0×10 １７ ( / cm ３ ) or more, and the semiconductor device according to claim 8.

10. The oxygen concentration contained in the semiconductor substrate is 1.0 × 10⁻⁶ １８ ( / cm ３ The semiconductor device according to claim 9, wherein the above conditions apply.

11. The semiconductor device according to claim 1 or 2, wherein the buffer region has a lower hydrogen peak of the chemical concentration of hydrogen on the lower side of the decreasing portion.

12. The semiconductor device according to claim 11, wherein in the buffer region, there is no peak of chemical concentration of hydrogen on the lower side of the lower hydrogen peak.

13. The semiconductor device according to any one of claims 1 or 4 to 7, wherein the buffer region has an upper hydrogen peak of the chemical concentration of hydrogen on the upper surface side of the reduction portion.

14. The semiconductor device according to claim 1 or 2, wherein the buffer region has a lower peak of the doping concentration on the lower side of the reduction portion.

15. The semiconductor device according to claim 1 or 2, wherein the buffer region has an upper peak of the doping concentration on the upper surface side of the reduction portion.

16. The buffer region has an upper hydrogen peak of the hydrogen chemical concentration on the upper surface side of the reduction portion, and when the depth position of the upper hydrogen peak from the lower surface is x (μm), the integral concentration obtained by integrating the chemical concentration distribution of hydrogen at the upper hydrogen peak is y ( / cm²). ２ ) is y = 2.70 × 10 １３ ×exp(-2.20×10 －２ x) Above, y = 2.70 × 10 １４ ×exp(-2.20×10 －２ x) The semiconductor device according to claim 1, wherein the following conditions apply.

17. A method for manufacturing a semiconductor device comprising a semiconductor substrate having an upper surface and a lower surface, comprising: implanting hydrogen ions from the lower surface to a first depth position in the semiconductor substrate; implanting hydrogen ions from the lower surface to a second depth position on the lower side of the first depth position; annealing the semiconductor substrate; and injecting a dose of hydrogen ions y ( / cm) from the lower surface to the first depth position x (μm). ２ ) is y = 2.70 × 10 １３ ×exp(-2.20×10 －２ x) Above, y = 2.70 × 10 １４ ×exp(-2.20×10 －２ x) A method for manufacturing a semiconductor device that is as follows:

Images