semiconductor unit

By balancing impurity concentration and layer width in p-type and n-type layers, the semiconductor device addresses the challenge of maintaining high breakdown voltage in the terminal region, reducing hole current and improving overall performance.

DE112017003957B4Active Publication Date: 2024-11-28MITSUBISHI ELECTRIC CORP
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
DE112017003957
Authority / Receiving Office
DE · DE
Patent Type
Patents
Current Assignee / Owner
Priority Date
2016-08-08
Filing Date
2017-06-02
Publication Date
2024-11-28
Estimated Expiration
2037-06-02

AI Technical Summary

Technical Problem

Semiconductor devices with a super junction structure face challenges in maintaining high breakdown voltage in the terminal region due to increased hole current and reduced breakdown voltage, particularly when using silicon carbide or silicon semiconductors, which have lower hole mobility and higher resistance to hole current.

Method used

The semiconductor device incorporates alternating p-type and n-type impurity layers with varying widths in the active and terminal regions, ensuring the product of impurity concentration and layer width is balanced to reduce hole current and enhance breakdown voltage.

Benefits of technology

This configuration reduces electric potential and improves breakdown voltage in the terminal region by minimizing hole current, thereby enhancing the overall performance of the semiconductor device.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 00000000_0000_ABST
    Figure 00000000_0000_ABST
Patent Text Reader

Abstract

Semiconductor unit comprising the following: - a semiconductor substrate (3); - a semiconductor layer (4) of a first conductivity type arranged on the semiconductor substrate (3); - a first defect region (7) of a second conductivity type, which is selectively arranged on an upper layer region of the semiconductor layer (4) in an active region; - a second defect region (8) of the first conductivity type, which is selectively arranged on an upper layer region of the first defect region (7); - a first main electrode (15) which is connected to the second defect area (8); - a gate insulating layer (12) which is continuously in contact with the second defect region (8), the first defect region (7) and the semiconductor layer (4); - a gate electrode (13) arranged such that it faces the second defect region (8), the first defect region (7) and the semiconductor layer (4) across the gate insulating layer (12); and - a second main electrode (16) which is arranged on one side opposite the semiconductor layer (4) of the semiconductor substrate (3), - wherein the semiconductor layer (4) comprises first pillar layers (5a, 5b) of the first conductivity type and second pillar layers (6a, 6b) of the second conductivity type, wherein the first pillar layers (5a, 5b) and the second pillar layers (6a, 6b) are arranged such that they extend from a main surface located on a side opposite the semiconductor substrate (3) to a predetermined depth towards the semiconductor substrate (3), - wherein the first column layers (5a, 5b) and the second column layers (6a, 6b) are arranged alternately in a direction parallel to the main surface in the active region of the semiconductor layer (4) and in a peripheral region which is a region around the active region, - wherein the first defect area (7) is located in an upper layer area of ​​the second column layers (6a, 6b), - wherein the semiconductor layer (4) with which the gate insulating layer (12) comes into contact consists of the first column layers (5a, 5b), - wherein the column spacing defined by the total width of a pair of one of the first column layers (5a, 5b) and one of the second column layers (6a, 6b) in the peripheral region is specified such that it is larger than the column spacing defined by the total width of a pair of one of the first column layers (5a, 5b) and one of the second column layers (6a, 6b) in the active region, and - wherein the product of the width of one of the first column layers (5a, 5b) and the effective impurity concentration with the first conductivity type of one of the first column layers (5a, 5b) in each of the active area and the peripheral area is equal to the product of the width of one of the second column layers (6a, 6b) and the effective impurity concentration with the second conductivity type of one of the second column layers (6a, 6b).
Need to check novelty before this filing date? Find Prior Art

Description

TECHNICAL FIELD

[0001] The present invention relates to a semiconductor device, and more particularly to a semiconductor device having a super junction (SJ) structure in which p-type impurity layers and n-type impurity layers extending in a direction perpendicular to a main surface of a semiconductor substrate are alternately arranged in a direction parallel to the main surface of the semiconductor substrate. STATE OF THE ART

[0002] A vertical unit is the primary semiconductor device used in power electronics. A vertical unit has a main electrode on each of two opposing main surfaces of a semiconductor substrate, and a main current flows in a direction perpendicular to the main surfaces of the semiconductor substrate. In a typical vertical metal-oxide-semiconductor field-effect transistor (MOSFET), a depletion layer extends into a drift layer when the MOSFET is turned off, and the depletion layer acts as a voltage-blocking layer.

[0003] When the device is turned on, a current flows through the drift layer located on the semiconductor substrate. The drift layer, which has a higher resistance than the semiconductor substrate, is one of the main resistive components of the device. Therefore, a thinner drift layer thickness can reduce the drift resistance and can significantly reduce the on-resistance of the MOSFET. Furthermore, the on-resistance can also be reduced by increasing the impurity concentration in the drift layer.

[0004] The breakdown voltage of the semiconductor unit is determined by the width of the depletion layer in the drift layer. This means that a thinner drift layer reduces the breakdown voltage of the semiconductor unit. Furthermore, a higher impurity concentration in the drift layer also reduces the width of the depletion layer and thus the breakdown voltage of the semiconductor unit. In this way, a compromise relationship exists between the breakdown voltage and the on-resistance.

[0005] A superjunction structure is proposed as a way to improve the compromise relationship between breakdown voltage and on-resistance. The superjunction structure is a structure in which p-type defect layers (hereinafter referred to as p-type support layers) and n-type defect layers (hereinafter referred to as n-type support layers) are arranged alternately in the drift layer in a direction orthogonal to the direction of the main current flow.

[0006] In such an SJ structure, when the drift layer has an n-type conductivity type, a depletion layer extends from a pn junction surface between a p-type support layer and an n-type support layer, as well as from a pn junction surface or a metal junction surface present on a surface of the semiconductor device. That is, a depletion layer having the same depth as the support layer is formed in the drift layer.

[0007] Even if the impurity concentration of the n-type support layer, i.e., the impurity concentration of the drift layer, is high, it is equivalent to the impurity concentration of a p-type support layer, and the n-type support layer is thus completely depleted. This allows the breakdown voltage to be maintained. As a result, the compromise relationship between the breakdown voltage and the on-resistance in the semiconductor device is significantly improved, and the drift resistance can be reduced.

[0008] However, a semiconductor device with the SJ structure has a significant problem in ensuring a breakdown voltage in a terminal region, which is a peripheral region of an element region (an active region). A low breakdown voltage in a terminal region reduces the breakdown voltage of the entire semiconductor device. Furthermore, a concentration of avalanche breakdown current in the terminal region can cause malfunction of the semiconductor device.

[0009] Generally, the impurity concentration of the p-type support layer and the n-type support layer in the terminal region is set to be lower than the impurity concentration of the p-type support layer and the n-type support layer in the element region. This is because if the impurity concentration in the terminal region, especially the donor concentration in the n-type support layer, is reduced to a low level, the breakdown voltage in the terminal region is improved.

[0010] Patent Document 1 discloses a multi-epitaxial process technology for a silicon semiconductor device, in which epitaxial growth and ion implantation are repeated to form a large number of epitaxially grown layers in a drift layer to form an SJ structure. Here, the impurity concentration in a terminal region is set to be lower than the impurity concentration in an active region. However, such a configuration cannot be achieved with an SJ structure formed by a high-performance trench-filling process and is not suitable for a semiconductor such as silicon carbide, which has a low impurity diffusion coefficient. PRIOR ART DOCUMENTPatent document

[0011] Patent document 1: Japanese patent application disclosure JP 2001-298 190 A SHORT DESCRIPTION Problem to be solved with the invention

[0012] Furthermore, it is also effective to increase the width of the connection area to the extent that it improves the breakdown voltage of the connection area. A larger connection area reduces the electric field applied to it more effectively, thus improving the breakdown voltage of the semiconductor device.

[0013] However, a larger width of the terminal region increases a hole current generated by depletion of the p-type support layer in the terminal region when the semiconductor device is turned off. When the semiconductor device is fabricated using a silicon semiconductor or a silicon carbide semiconductor, the hole mobility in semiconductor crystals is lower than the electron mobility, and the resistance to a hole current is increased, resulting in an increasing electric potential at the periphery of a source contact in the terminal region.

[0014] Then, a high electrical potential is applied to an insulating layer on a surface within the element region, causing a malfunction in the semiconductor unit. Furthermore, if the impurity concentration of the n-type support layer is increased with the intention of reducing the drift resistance, the impurity concentration of the p-type support layer must be increased accordingly to equal it. Additionally, in this case, the hole current, which flows due to the turn-off, will also increase.

[0015] The present invention has been conceived to solve the problems described above, and has an object to provide a semiconductor device whose breakdown voltage in a terminal region is improved. Means to solve the problem

[0016] According to the present invention, a semiconductor unit comprises: a semiconductor substrate; a semiconductor layer of a first conductivity type arranged on the semiconductor substrate; a first impurity region of a second conductivity type selectively arranged on an upper layer region of the semiconductor layer in an active region; a second impurity region of the first conductivity type selectively arranged on an upper layer region of the first impurity region; a first main electrode connected to the second impurity region; a gate insulating layer arranged to be in continuous contact with the second impurity region, the first impurity region, and the semiconductor layer; a gate electrode arranged to face the second impurity region, the first impurity region, and the semiconductor layer across the gate insulating layer;as well as a second main electrode, which is located on one side opposite the semiconductor layer of the semiconductor substrate.;

[0017] The semiconductor layer comprises first support layers of the first conductivity type and second support layers of the second conductivity type. The first and second support layers are arranged such that they extend from a primary surface, located on a side opposite the semiconductor substrate, to a predetermined depth towards the semiconductor substrate. The first and second support layers are arranged alternately in a direction parallel to the primary surface within the active region of the semiconductor layer and in a terminal region, which is a region surrounding the active region.

[0018] The first defect region is located in an upper layer of the second support layers. The semiconductor layer with which the gate insulating layer comes into contact is the first support layer. The support distance, defined by the combined width of a pair of first support layers and a pair of second support layers in the terminal area, is set to be larger than the support distance defined by the combined width of a pair of first support layers and a pair of second support layers in the active region.

[0019] The product of the width of one of the first support layers and the effective impurity concentration of the first conductivity type of one of the first support layers is in each of the active area and the connection area equal to the product of the width of one of the second support layers and the effective impurity concentration of the second conductivity type of one of the second support layers. Effects of the invention

[0020] With the semiconductor unit according to the present invention, a semiconductor unit can be achieved whose electrical potential generated in a connection area is reduced and whose breakdown voltage in the connection area is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The figures show: Fig. 1 a graphical representation showing the extent of a depletion layer as a function of an applied voltage at a pn junction; Fig. 2 a graphical representation showing a change in hole flux as a function of the width of a support layer; Fig. 3 a perspective view showing a configuration of a semiconductor unit according to a first embodiment of the present invention; Fig. 4 a perspective view showing a manufacturing process for the semiconductor unit according to the first embodiment of the present invention; Fig. 5 a perspective view showing a manufacturing process for the semiconductor unit according to the first embodiment of the present invention; Fig. 6 a perspective view showing a manufacturing process for the semiconductor unit according to the first embodiment of the present invention; Fig. 7 a perspective view showing a manufacturing process for the semiconductor unit according to the first embodiment of the present invention; Fig. 8 a perspective view showing a manufacturing process for the semiconductor unit according to the first embodiment of the present invention; Fig. 9 a perspective view showing a manufacturing process for the semiconductor unit according to the first embodiment of the present invention; Fig. 10 a perspective view showing a manufacturing process for the semiconductor unit according to the first embodiment of the present invention; Fig. 11 a perspective view showing a manufacturing process for the semiconductor unit according to the first embodiment of the present invention; Fig. 12 a perspective view showing a manufacturing process for the semiconductor unit according to the first embodiment of the present invention; Fig. 13 a view showing a manufacturing process for the semiconductor unit according to the first embodiment of the present invention; Fig. 14 a perspective view showing the configuration of a semiconductor unit according to a second embodiment of the present invention; Fig. 15 a perspective view showing the configuration of a semiconductor unit according to a third embodiment of the present invention; Fig. 16 a perspective view showing the configuration of a semiconductor unit according to a fourth embodiment of the present invention; Fig. 17 a view showing a manufacturing process for the semiconductor unit according to the fourth embodiment of the present invention; Fig. 18 a view showing a manufacturing process for the semiconductor unit according to the fourth embodiment of the present invention; Fig. 19 a view showing a manufacturing process for the semiconductor unit according to the fourth embodiment of the present invention; Fig. 20 a view showing a manufacturing process for the semiconductor unit according to the fourth embodiment of the present invention; Fig. 21 is a perspective view illustrating a configuration of support layers formed by a multi-epitaxial process; Fig. 22 is a perspective view illustrating the configuration of a semiconductor unit according to a fifth embodiment of the present invention. DESCRIPTION OF EMBODIMENTSIntroduction

[0022] Before describing embodiments, a super junction structure (SJ structure) is first described.

[0023] In a typical MOSFET without a superjunction structure, a protection ring, a junction termination extension (JTE) structure, a field limiting ring (FLR) structure, or similar features are formed in a terminal area. Similar to an element area where a depletion layer extends from a transition surface between a base area and a drift layer towards the drift layer, a depletion layer in a terminal structure extends from a transition surface between a drift layer and a terminal structure towards the drift layer.

[0024] In this case, the number of holes emitted by the terminal structure is equal to the number of electrons present in the depleted region extending into the drift layer. Furthermore, due to an increase in drain voltage, the depletion layer has a thickness equal to the thickness of the drift layer.

[0025] In contrast, the width of the depletion layer in the superjunction structure is determined by the depth of the support layers. In each support layer, a depletion direction is a direction parallel to a main surface of a semiconductor substrate. If the width of the support layer is sufficiently small, the support layer is completely depleted with only a small drain voltage.

[0026] If the width of the support layer is increased, then the width of each support layer to be depleted is also increased, thus increasing the stress difference between a p-type and an n-type support layer, which is necessary to effect complete depletion. That is, with the same stress increase (dV / dt) per unit time, an increased support layer width can reduce the depletion path per unit time and can therefore decrease the hole current.

[0027] In short, at the time of a switchover, holes move away from the p-type support layer as majority charge carriers, and the hole current can be reduced by reducing the rate of movement of the holes per unit time.

[0028] In the element region, however, all of the p-type support layers can be connected to a source electrode, and the hole current concentrated at a contact is lower than that in the terminal region. Therefore, the width of each individual support layer in the element region can be reduced to increase the channel density and thus reduce the on-resistance in a MOSFET.

[0029] Fig. Figure 1 is a graphical representation showing the extent of a depletion layer from a pn junction surface, as represented by the width of a depletion layer, as a function of a voltage applied to a pn junction where an acceptor concentration and a donor concentration are constant and equal in a region. Fig. 1, the horizontal axis represents an applied voltage (V), and the vertical axis represents the width of a depletion layer (µm), and T1 represents a characteristic when a donor concentration (ND) and an acceptor concentration (NA) are equal to 1 × 10 16 cm -3 T2 represents a characteristic where the donor concentration and the acceptor concentration are equal to 2 × 10 16 cm -3 are, and T3 represents a characteristic when the donor concentration and the acceptor concentration are equal to 3 × 10 16 cm -3 are.

[0030] Fig. Figure 1 shows that the width of the depletion layer is larger when the applied voltage is larger, and that the width of the depletion layer is smaller when the impurity concentration is higher for the same applied voltage.

[0031] Fig. Figure 1 also shows that an applied voltage of approximately 200 V is required to cause the width of the depletion layer to increase to 2.5 µm when the acceptor concentration and the donor concentration are equal to 1 x 10 16 cm -3 This means that if the total width of an n-type support layer and a p-type support layer is 5 µm, then the support layer is completely depleted when a voltage of 200 V is applied.

[0032] In contrast, an applied voltage of 600 V or more is required to achieve a depletion layer width of 4.5 µm. If the combined width of an n-type support layer and a p-type support layer is 9 µm, then at an applied voltage of 200 V, the support layer is partially undepleted and will not be fully depleted until the applied voltage is increased to approximately 600 V.

[0033] In the SJ structure, a large total width of an n-type support layer and a p-type support layer, while achieving depletion layers with the same area in an n-type support layer and a p-type support layer, reduces the depletion rate in response to a voltage increase, thereby reducing the emission rate of electrons from the n-type support layer and the emission rate of holes from the p-type support layer. That is, a slow progression of depletion reduces the emission rate of electrons and holes.

[0034] Fig. Figure 2 is a graphical representation showing calculated results of a hole current as a function of the width of a p-type support layer, where the horizontal axis represents a support width (µm) and the vertical axis represents a hole current (arbitrary units au). The calculation hypothetically holds a voltage increase (dV / dt) per unit time even when the support layer width is changed, and the donor concentration of the n-type support layer and the acceptor concentration of the p-type support layer are each equal to 1 × 10⁻⁶. 16 cm -3 .

[0035] Furthermore, a hole flux flowing from the p-type support layer is defined based on the number of holes present in a region that becomes depleted within a unit of time. Fig. Figure 2 shows that the hole current can be reduced by approximately half when the support width is increased from approximately 5 µm to 9 µm. An electric potential to be applied to a terminal region due to the hole current is proportional to the hole current, and therefore the generated electric potential is reduced by half in a similar manner. First embodiment

[0036] Now, a semiconductor device according to a first embodiment of the present invention will be described with reference to Fig. 3 to Fig. 13 described. Unit configuration

[0037] Fig. Figure 3 is a perspective view showing a configuration of a vertical MOSFET 100, which is a silicon carbide semiconductor unit that uses silicon carbide in a semiconductor substrate. It should be noted that the present invention is not limited to use with a silicon carbide semiconductor unit, but can also be used with a semiconductor unit that uses a semiconductor such as silicon (Si), gallium nitride (GaN), and diamond (C) and has the superjunction structure.

[0038] In comparison to a Si semiconductor unit, a silicon carbide semiconductor unit, which uses silicon carbide in a semiconductor substrate, can provide a semiconductor unit that has a very good breakdown voltage, a high allowable current density, very good heat resistance and suitability for operation at high temperature.

[0039] It should be noted that in the following description regarding conductivity types of impurities, a general definition is used in which an n-type is referred to as a "first conductivity type" and a p-type is referred to as a "second conductivity type", but the definition may be reversed.

[0040] As in Fig. 3, in the vertical MOSFET 100, an n-type epitaxially grown crystal layer 4 (a semiconductor layer) is disposed on one main surface of an n-type semiconductor substrate 3 made of silicon carbide, a plurality of p-type well regions 7 are selectively disposed in an upper layer portion of the epitaxially grown crystal layer 4, and a p-type contact region 10a is disposed in each well region 7 so as to penetrate the well region 7. Note that the epitaxially grown crystal layer 4 is also referred to as a drift layer.

[0041] Furthermore, in an upper layer region of each trough region 7, a source region 8 of n-type is arranged in contact with both lateral surfaces of the contact region 10a. It should be noted that the source region 8 is arranged such that its thickness is less than the thickness of the trough region 7, and the contact region 10a is arranged such that its thickness is approximately equal to or slightly greater than the thickness of the trough region 7.

[0042] A gate insulating layer 12 is selectively formed on the epitaxially grown crystal layer 4, and a gate electrode 13 is formed on the gate insulating layer 12. Specifically, the gate insulating layer 12 is arranged between adjacent source regions 8 and extends from an upper portion of one source region 8, across a trough region 7 and the epitaxially grown crystal layer 4, to an upper portion of another source region 8 of an adjacent trough region 7. The gate electrode 13 is arranged such that it covers the gate insulating layer 12.

[0043] Further, an interlayer insulating layer 14 is formed to cover the gate insulating layer 12 and the gate electrode 13, and a source electrode 15 is formed to cover the interlayer insulating layer 14. In a region of the interlayer insulating layer 14 except the region covering the gate electrode 13, a contact hole SC is formed, penetrating the interlayer insulating layer 14 in the thickness direction and reaching a portion of the source region 8 and the entire surface of the contact region 10a. Furthermore, the source electrode 15 is filled in the contact hole SC, and the source electrode 15 is thereby connected to the source region 8 and the contact region 10a.

[0044] A plurality of MOSFET elements, each formed of the source regions 8 and other regions as described above, are arranged in a direction parallel to the main surface of the semiconductor substrate 3 and connected in parallel to form an element group. A region provided with the element group is referred to as an element region (an active region) ER, and a terminal region TR for achieving a breakdown voltage of the vertical MOSFET 100 is arranged at an outer peripheral region of the element region ER.

[0045] In an upper layer region of the epitaxially grown crystal layer 4 in the terminal region TR, a p-type contact region 10b is arranged, defining an outer boundary of the element region ER. The contact region 10b is arranged such that it has a thickness equal to the thickness of the contact region 10a, but its width (a length in a direction parallel to the main surface of the semiconductor substrate 3) is greater than the width of the contact region 10a.

[0046] Furthermore, a p-type RESURF region 9 is arranged on an outer side of the contact region 10b so as to have a thickness approximately equal to the thickness of the contact region 10b.

[0047] Furthermore, a plurality of support layers 5a of n-type and of support layers 6a of p-type are arranged in the epitaxially grown crystal layer 4 such that they are arranged alternately in equal numbers in the element region ER, and a plurality of support layers 5b of n-type and of support layers 6b of p-type are arranged such that they are arranged alternately in a manufacturing region for the contact region 10b and the RESURF region 9 in the connection region TR.

[0048] It should be noted that a further plurality of n-type support layers 5a and p-type support layers 6a are alternately arranged in the terminal region TR located on an outer side with respect to an arrangement region for the n-type support layers 5b and the p-type support layers 6b.

[0049] Each support layer is arranged such that it extends in the thickness direction of the epitaxially grown crystal layer 4 from a top surface of the epitaxially grown crystal layer 4 towards the side of the semiconductor substrate 3, and the deepest region of the same is specified such that it is shallower than the thickness of the epitaxially grown crystal layer 4.

[0050] It should be noted that the width of support layer 5a (n-type) and support layer 6a (p-type) are equal, and the sum of both widths is represented by a support distance W1. Furthermore, the widths of support layer 5b (n-type) and support layer 6b (p-type) are also equal, and each width is specified such that it is greater than the combined width of both support layer 5a (n-type) and support layer 6a (p-type). A support distance W2, which is the sum of the widths of both support layer 5b (n-type) and support layer 6b (p-type), is greater than the combined width W1.

[0051] Furthermore, a field insulating layer 11 is arranged in the connection area TR on the epitaxially grown crystal layer 4, and the insulating intermediate layer 14 is arranged on the field insulating layer 11.

[0052] The source electrode 15 is arranged such that it extends from the element region ER to a position above a stacked layer consisting of the field insulating layer 11 and the insulating intermediate layer 14 in the connection region TR. Furthermore, a contact hole TC is arranged in a region of the stacked layer consisting of the field insulating layer 11 and the insulating intermediate layer 14, corresponding to a region above the contact region 10b. This contact hole TC penetrates the stacked layer in the thickness direction and extends to the contact region 10b. The source electrode 15 is also filled into the contact hole TC, thereby connecting it to the contact region 10b.

[0053] It should be noted that the stacking layer consisting of the field insulating layer 11 and the insulating intermediate layer 14 is arranged such that it covers an upper portion of the MOSFET element at the outermost periphery of the element area ER. A passivation layer 17 is arranged such that it covers an upper portion of the source electrode 15 and an upper portion of the stacking layer consisting of the field insulating layer 11 and the insulating intermediate layer 14.

[0054] Furthermore, a drain electrode 16 is arranged on the other main surface (i.e., the back side) of the semiconductor substrate 3, which is opposite to a side on which the source electrode 15 is arranged.

[0055] In the vertical MOSFET 100, which has the configuration described above, a channel is formed in an upper layer region of the well area 7, located between the epitaxially grown crystal layer 4 and the source area 8 immediately below the gate electrode 13, when a voltage is applied to the gate electrode 13. Through this channel, the source area 8 and the source electrode 15 are electrically connected via the well area 7, the n-type support layer 5a, the epitaxially grown crystal layer 4, and the semiconductor substrate 3. Manufacturing process

[0056] Next, with reference to Fig. 4 to Fig. 13 describes a manufacturing process for the vertical MOSFET 100. First, as in Fig. Figure 4 shows the epitaxially growing n-type crystal layer 4 formed by means of a method for epitaxial crystal growth on one main surface of the n-type semiconductor substrate 3 made of silicon carbide.

[0057] The orientation of the crystal plane of the semiconductor substrate may be inclined by 8° or less with respect to the direction of the c-axis, or it may be uninclined. Alternatively, the semiconductor substrate 3 may have any plane orientation. The defect concentration of the epitaxially grown crystal layer 4, for example, is in the range of 1 × 10 13 cm -3 up to 1 × 10 18 cm -3 , and the thickness of the epitaxially grown crystal layer 4 is, for example, 5 µm to 200 µm.

[0058] Thereafter, a mask MS1 made of an oxide layer is formed on the epitaxially grown crystal layer 4 by, for example, a process such as photolithography, and a surface of the epitaxially grown crystal layer 4 is etched using the mask MS1 made of an oxide layer as an etching mask. Fig. As shown in Figure 5, trenches 61a and 61b are formed to form the support layers 6a and 6b, respectively. Either a wet etching process or a dry etching process can be used to etch a surface of the epitaxially grown crystal layer 4.

[0059] However, since the aspect ratio between the width and depth of a p-type support layer used in the super-junction structure corresponds to that of the trench shape, it is desirable to use a dry etching process, such as reactive ion etching or sputter etching, which makes controlling the trench shape relatively easy.

[0060] It is preferred that the depth of the trenches 61a and 61b does not exceed the thickness of the epitaxially grown crystal layer 4. The epitaxially grown crystal layer 4 is defined as a buffer layer located below the bottom surface of the trenches 61a and 61b.

[0061] Next, at a Fig. In the process shown in Figure 6, a p-type semiconductor is filled into the trenches 61a and 61b to form the p-type support layers 6a and 6b. In this process, the p-type semiconductor is formed by epitaxial growth of crystals on the epitaxially grown crystal layer 4, thereby forming a p-type semiconductor layer 60 inside the trenches 61a and 61b and on the n-type support layers 5a and 5b. In this case, a height difference is created on the surface of the p-type semiconductor layer 60 in areas obtained by filling the trenches 61a and 61b. It should be noted that while the trenches 61a and 61b are preferably completely filled with the p-type semiconductor layer 60, complete filling is not strictly necessary.

[0062] Here, the epitaxial growth of p-type semiconductor layer 60 crystals is carried out under conditions such that the p-type support layers 6a and 6b achieve a defined acceptor concentration. Specifically, it is assumed that the acceptor concentration NA in the p-type support layer 6a maintains the charge equilibrium of the support layer in the element region ER, and this is defined such that the relation Wn1 × ND = Wp1 × NA holds, where Wn1 represents the width of the n-type support layer 5a, ND represents the impurity concentration of the n-type support layer 5a, Wp1 represents the width of the p-type support layer 6a, and NA represents the impurity concentration of the p-type support layer 6a.

[0063] It should be noted that the impurity concentration ND of the n-type support layer 5a is in a range of 1 × 10 13 cm -3 up to 1 × 10 18 cm -3lies, where this is the same as the defect concentration of the epitaxially grown crystal layer 4.

[0064] In the embodiments, the acceptor concentration and the donor concentration each represent the effective impurity concentration. The effective impurity concentration is an impurity concentration of impurities that is limited to the electrically active impurities introduced into the semiconductor substrate.

[0065] If the thickness of the buffer layer is large and thus cannot be neglected when considering a charge balance between supports, the acceptor concentration NA is specified such that the relation Wn1 × Dp1 × ND + (Wn1 + Wp1) × TB × NDB = Wp1 × Dp1 × NA applies, where a thickness DP1 of the support layer, a thickness TB of the buffer layer and an impurity concentration NDB of the buffer layer are included.

[0066] In Fig. 3. The width of support layer 5a of n-type and that of support layer 6a of p-type in element area ER are each the same, but they can differ. Furthermore, the width of support layer 5b of n-type and that of support layer 6b of p-type in connection area TR are each the same, but they can differ.

[0067] Furthermore, the donor concentrations (ND) of the n-type support layers 5a and 5b are equal in the element region ER and the terminal region TR, and the acceptor concentrations (NA) of the p-type support layers 6a and 6b are also equal. Thus, the breakdown voltage expected from the superjunction structure can be reliably obtained in both regions.

[0068] It should be noted that the donor concentrations (ND) of the n-type support layers 5a and 5b and the acceptor concentrations (NA) of the p-type support layers 6a and 6b can be either the same or different.

[0069] Furthermore, the width Wn2 of the n-type support layer 5b and the width Wp2 of the p-type support layer 6b in the connection region TR are specified such that the relation Wn2 × ND = Wp2 × NA holds. If, moreover, the donors of the buffer layer in the connection region TR cannot be neglected, the configuration is specified such that the relation Wn2 × Dp1 × ND + (Wn2 + Wp2) × TB × NDB = Wp2 × Dp1 × NA holds.

[0070] Furthermore, the support spacing W1 (Wn1 + Wp1) of the n-type support layer 5a and the p-type support layer 6a in the element region ER and the support spacing W2 (Wn2 + Wp2) in the terminal region TR are specified such that the relationship W1 < W2 holds. This configuration reduces the current generated by majority charge carriers emitted by the p-type support layers 6b in the terminal region when the MOSFET is switched off, or by majority charge carriers (in this case, holes) injected into the p-type support layers 6b in the terminal region when the MOSFET is switched on. Reducing this current decreases the electrical potential generated in the terminal region and improves the dV / dt resistance of the semiconductor unit. It should be noted that in this embodiment, W1 x 2 = W2.

[0071] Next, in a Fig. In the process shown in Figure 7, the p-type semiconductor layer is removed by means of a chemical-mechanical polishing process (CMP) until the upper areas of the n-type support layers 5a and 5b and the p-type support layers 6a and 6b are exposed in order to planarize the epitaxially grown crystal layer 4.

[0072] Then, in a Fig. In the process shown in Figure 8, defect ions are implanted using a resist mask (not shown) which is structured by photolithography, and the trough area 7, the source area 8, the RESURF area 9 and the contact areas 10a and 10b are thereby selectively formed in an upper layer area of ​​the epitaxially grown crystal layer 4.

[0073] In this process, p-type defects are introduced into the trough area 7, the RESURF area 9 and the contact areas 10a and 10b, and n-type defects are introduced into the source area 8.

[0074] The impurity concentration of the well region 7 and the RESURF region 9 is desirably, for example, in a range of 1 × 10 15 cm -3 up to 1 × 10 19 cm -3 , and the depth of the same lies, for example, in a range of 0.3 µm to 4.0 µm. The defect concentration of the source area 8 exceeds the defect concentration of the trough area 7 and is, for example, desirablely in a range of 1 × 10 18 cm -3 up to 1 × 10 21 cm -3 . Furthermore, the depth of the source region 8 is a depth that does not exceed that of the well region 7.

[0075] Regarding contact areas 10a and 10b, ions can be implanted using the same resist mask, and the defect concentration is desirablely in the range of 1 × 10 18 cm -3 up to 1 × 10 21 cm-3 It should be noted that ion implantation is preferably performed at a substrate temperature of 200 °C or higher.

[0076] It should be noted that contact areas 10a and 10b are arranged to achieve satisfactory metal-to-metal contact with the recess area 7 and the RESURF area 9, and that the semiconductor unit will function without contact areas 10a and 10b being arranged.

[0077] Furthermore, although not shown, after the introduction of the defects, a heat treatment is performed, for example, for 0.5 to 60 minutes at a temperature of 1500 °C to 2200 °C in an inert gas, such as argon or nitrogen, or in a vacuum. This electrically activates the implanted defects. Subsequently, an oxide layer is formed on the epitaxially grown crystal layer 4 by means of sacrificial oxidation. Following this, an altered surface layer in the epitaxially grown crystal layer 4 is removed by oxide layer removal using hydrofluoric acid to obtain a clean surface.

[0078] As in Fig. As shown in Figure 8, the trough area 7 in the element area ER is formed such that it covers an upper layer area of ​​the p-type support layer 6a, and the p-type support layer 6a is connected to the source electrode 15 via the contact area 10a ( Fig. 3) electrically connected. Furthermore, the contact area 10b and the RESURF area 9 in the connection area TR are formed such that they extend over an upper layer area of ​​the majority of p-type support layers 6b to cover the upper layer areas.

[0079] The majority of p-type support layers 6b covered by the contact area 10b have the same electrical potential across the contact area 10b, and the majority of p-type support layers 6b covered by the RESURF layer 9 have the same electrical potential across the RESURF area 9.

[0080] Furthermore, the contact area 10b and the RESURF area 9 are arranged such that their lateral surfaces come into contact with each other, and therefore the p-type support layers 6b covered by the contact area 10b are connected to the source electrode 15 via the contact area 10b ( Fig. 3) electrically connected.

[0081] Next, in a Fig. In the process shown in Figure 9, a silicon oxide layer is formed on the epitaxially grown crystal layer 4 by means of chemical vapor deposition (CVD). The silicon oxide layer is then structured using a resist mask (not shown) structured by photolithography to form the field insulating layer 11, which has an opening OP only in the elemental region (the active region). The field insulating layer 11 is removed by a dry etching process using carbon tetrafluoride (CF4) or the like, or by a wet etching process using hydrogen fluoride or the like.

[0082] It should be noted that the field insulating layer 11 is not limited to a silicon oxide layer and can be formed from a silicon nitride layer or another insulating layer.

[0083] Next, in a Fig. In the process illustrated in FIG. 10, an insulating layer 120 for forming the gate insulating layer 12 is formed in the element region ER exposed in the opening OP of the field insulating layer 11. The insulating layer 120 is formed on the entire substrate surface by, for example, thermal oxidation or CVD and is also formed on the field insulating layer 11, the illustration of which is omitted.

[0084] Next, a conductive layer for forming the gate electrode 13 is formed on the insulating layer 120. Generally, the gate electrode 13 is formed from polysilicon containing a high concentration of impurities. A conductive layer is formed over the entire substrate surface using CVD using polysilicon, and then an unnecessary portion is removed by dry etching.

[0085] This means, as in Fig. Figure 3 shows the gate electrode 13 arranged between adjacent source regions 8, and it extends from an upper part of a source region 8 over a trough region 7 and the epitaxially grown crystal layer 4 (an n-type support layer 5a) to an upper part of another source region 8 of an adjacent trough region 7.

[0086] Next, the insulating intermediate layer 14 is formed, for example by CVD, across the entire substrate surface to cover the gate electrode 13 and the field insulating layer 11. Then, as described in Fig. As shown in Figure 11, the contact hole SC, which penetrates the interlayer insulating layer 14 in the thickness direction and reaches a portion of the source region 8 and the entire surface of the contact region 10a, and the contact hole TC, which penetrates the stacked layer of the interlayer insulating layer 14 and the field insulating layer 11 in the thickness direction and reaches the contact region 10b, are formed. The contact holes SC and TC are formed in a slit-like shape.

[0087] It should be noted that when the contact hole SC is formed, the insulating layer 120, which remained on the epitaxially grown crystal layer 4, is also removed at the same time.

[0088] Next, in a Fig. 12, the source electrode 15 is formed, which is filled into the contact holes SC and TC and comes into contact with the source region 8 and the contact regions 10a and 10b. Furthermore, a gate wiring (not shown) to be connected to the gate electrode 13 is formed in a region outside the element region ER. The configuration of the region "A" of Fig. 12 is in Fig. 13. The drain electrode 16 is then formed on the back side of the semiconductor substrate 3.

[0089] It should be noted that the source electrode 15 and the drain electrode 16 are formed of a metal such as aluminum (Al), titanium (Ti), copper (Cu), and gold (Au), or an alloy thereof, and are formed by, for example, a sputtering method or a vapor deposition method.

[0090] It should be noted that a silicide layer, formed by a reaction of nickel (Ni) or the like and silicon carbide, can be formed on a surface of the epitaxially grown crystal layer 4 such that it comes into contact with the source electrode 15 and the drain electrode 16. A reduction in the contact resistance to each electrode can be expected by means of the silicide layer.

[0091] Finally, the passivation layer 17 is arranged to cover an upper portion of the source electrode 15 and an upper portion of the stacked layer of the field insulating layer 11 and the insulating interlayer 14, using an insulating material such as polyimide or a silicon oxide layer. Fig. The vertical MOSFET 100 shown in Figure 3 is completed. The passivation layer 17 is arranged in the terminal area TR without the source electrode and reduces surface charges, thus further improving the breakdown voltage of the terminal area TR.

[0092] When the MOSFET is turned off, each support layer emits majority carriers to the outer region of the support, so that the majority carriers are depleted. In the element region ER, holes emitted from the p-type support layer 6a move to the source electrode 15 via the contact region 10a. Meanwhile, in the terminal region TR, holes emitted from the support layer 6b move to the source electrode 15 via the contact region 10b or are emitted in phases toward the source electrode 15 via the RESURF region 9 and the contact region 10b.

[0093] If the support spacing W2 in the connection area TR is small, i.e., if the extent of the depletion layer in each support is small, an electrical potential generated in the RESURF area 9 and the contact area 10b is increased by a hole current. In this embodiment, however, such an increase is reduced.

[0094] In contrast, the support distance in the terminal area TR of this embodiment is set to be twice the support distance in the element area ER. Therefore, the electrical potential generated in the RESURF area 9 and the contact area 10b can be reduced by approximately half compared to the case where the width of the support layer in the element area ER and the terminal area TR is the same, and the dV / dt resistance of the semiconductor unit can be improved.

[0095] Furthermore, the p-type support layers 6a and 6b are formed using a trench-filling method. Therefore, the defect concentration of the n-type support layers 5a and 5b can be set to be equal to the n-type defect concentration of the epitaxially grown crystal layer 4. Moreover, productivity can be improved compared to the case where all the support layers are formed using a multi-epitaxial method. modification

[0096] In the Fig. In the vertical MOSFET 100 shown in Figure 3, one end of the source electrode 15, viewed from the side of the source electrode 15, is arranged such that it is aligned with the position of one end of the contact area 10b. However, the end of the source electrode 15 can be arranged on the inner side with respect to the contact area 10b, on the inner side with respect to the RESURF area 9, or on the outer side with respect to one end of the RESURF area 9.

[0097] Furthermore, one or more field plates can be arranged on the field insulation layer 11 on the outer side of the source electrode 15 such that they surround or encircle the source electrode 15. Adjusting the electrical potential of the field plate can control the curvature of a corner of the depletion layer, and thus the breakdown voltage in the terminal area TR can be further improved. Second embodiment

[0098] In the vertical MOSFET 100 according to the first embodiment described above, each support pitch W2 in the terminal region TR is the same in a region below the RESURF region 9 and the contact region 10b so as to improve the dV / dt value of the resistance of the semiconductor unit, but this configuration is not limitative.

[0099] In a vertical MOSFET 200 according to a second embodiment, as in Fig. Figure 14 shows specifically a support distance W22, which is a total value of the width of each of a support layer 5c of n-type and a support layer 6c of p-type that are arranged in a region close to the element region ER, and the support distance W2 of other support layers are different, and they are specified such that the relation W22 < W2 holds.

[0100] Here, the support pitch W22 of the support layers arranged in a region close to the element region ER may be set to be equal to the support pitch W1 in the element region ER, or may be set to be equal to another width.

[0101] In this way, by simply increasing at least the support pitch at a position away from a connection point with the source electrode 15 in the terminal region TR, an effect of improving the dV / dt value of the resistance of the semiconductor device can be achieved. Third embodiment

[0102] In the vertical MOSFET 100 according to the first embodiment, the support layers surround the outermost periphery of the terminal region TR, but this configuration is not limitative.

[0103] In a vertical MOSFET 300 according to a third embodiment, as shown in Fig. 15, the RESURF region 9 specifically extends to the outermost periphery of the terminal region, and no support layers are arranged at the outermost periphery of the terminal region.

[0104] Note that the support pitch in the terminal region TR is set to be twice the support pitch of the support layers in the element region ER. Therefore, the electric potential generated in the RESURF region 9 and the contact region 10b can be reduced by approximately half compared to the case where the width of the support layer in the element region ER and that in the terminal region TR are the same, and the dV / dt value of the resistance of the semiconductor device can be improved. modification

[0105] The breakdown voltage in the terminal region TR can be further improved by disposing the RESURF region 9. However, instead of the RESURF region 9, a terminal structure such as a guard ring structure, a JTE (Junction Termination Extension) structure, a field limiting ring structure (FLR structure), and a variation of lateral doping structure (VLD structure) can also be disposed at the outermost periphery of the terminal region TR.

[0106] In the protective ring structure, a ring-shaped defect region is arranged in an upper layer of a terminal area along the terminal area. A blocking bias is applied to the defect region to form a depletion layer, thereby reducing the concentration of an electric field.

[0107] In the JTE structure, a ring-shaped defect area is arranged along a connection area in such a way that the defect concentration gradually decreases towards the outer side of the connection area.

[0108] In the VLD structure, a defect area is distributed by controlling the defect concentration in one direction parallel to a substrate surface through an opening structure of an implantation mask. Fourth embodiment

[0109] In the vertical MOSFET 100 according to the first embodiment, a planar channel MOSFET is shown, in which a channel is formed in a direction parallel to a substrate. In a fourth embodiment, the present invention is applied to a trench channel MOSFET, in which a channel is formed in a direction perpendicular to a substrate.

[0110] Fig. Figure 16 is a perspective view showing a configuration of a vertical MOSFET 400 according to the fourth embodiment. It should be noted that in Fig. 16 the same components as those of the with reference to Fig. The 3 described vertical MOSFET 100 are assigned the same reference symbols and an overlapping description has been omitted.

[0111] As in Fig. As shown in FIG. 16, in the vertical MOSFET 400, a plurality of p-type well regions 7 are selectively disposed in an upper layer portion of the epitaxially grown crystal layer 4, and in each well region 7, a p-type contact region 10a is disposed so as to penetrate the well region 7. Further, in an upper layer portion of the well region 7, an n-type source region 8 is disposed in contact with both side surfaces of the contact region 10a. Note that the source region 8 is disposed so as to cover the entire well region 7.

[0112] The source area 8 is arranged such that its thickness is less than the thickness of the trough area 7, and the contact area 10a is arranged such that its thickness is approximately equal to the thickness of the trough area 7 or slightly greater than it.

[0113] Furthermore, a trench 18 is arranged in the epitaxially grown crystal layer 4 between adjacent trough regions 7 and source regions 8, and an inner surface of the trench 18 is covered with a gate insulating layer 121. The trench 18 is also configured to extend into the interior of the epitaxially grown crystal layer 4, and the lateral surfaces of the trough regions 7 and the source regions 8 come into contact with a surface of the gate insulating layer 121.

[0114] Furthermore, a gate electrode 131 is formed to fill the trench 18, the inner surface of which is covered with the gate insulating layer 121. The gate electrode 131 is formed to partially protrude from the inside of the trench 18. The insulating interlayer 14 is formed to cover the gate electrode 131, the gate insulating layer 121, and the source region 8. The source electrode 15 is formed to cover the insulating interlayer 14.

[0115] In the vertical MOSFET 400 with the configuration described above, a channel is formed on a lateral surface of the well region 7, which extends along the lateral surface of the gate electrode 131, when a voltage is applied to the gate electrode 131. Through this channel, the source region 8 and the source electrode 15 are electrically connected via the well region 7, the n-type support layer 5a, the epitaxially grown crystal layer 4, and the semiconductor substrate 3. Manufacturing process

[0116] Next, with reference to Fig. 17 to Fig. 20 describes a manufacturing process for the vertical MOSFET 400. It should be noted that a process up to the exposure of upper regions of the n-type support layers 5a and 5b and the p-type support layers 6a and 6b on a major surface of the epitaxially grown crystal layer 4 is the same as that of the first embodiment, which is described with reference to Fig. 4 to Fig. 7, and therefore a further description thereof is omitted.

[0117] In a Fig. In the process described in Figure 17, defect ions are implanted using a resist mask (not shown) structured by photolithography, and the trough region 7, the source region 8, and the contact regions 10a and 10b are selectively formed in an upper layer region of the epitaxially grown crystal layer 4. It should be noted that the resurf region 9 and the contact region 10b are selectively formed in the terminal region, but their representation is omitted. It should also be noted that the conductivity type, defect concentration, depth, and implantation conditions for the defects in each region are the same as those in the first embodiment, and a further description of these is therefore omitted.

[0118] After the introduction of the defects, a heat treatment is performed, for example, for 0.5 to 60 minutes at a temperature of 1500 °C to 2200 °C in an inert gas, such as argon or nitrogen, or in a vacuum, to activate the implanted defects. Subsequently, an oxide layer is formed on the epitaxially grown crystal layer 4 by means of sacrificial oxidation. Following this, an altered surface layer in the epitaxially grown crystal layer 4 is removed by oxide layer removal using hydrofluoric acid to obtain a clean surface.

[0119] Subsequently, the trench 18 is formed, penetrating the source region 8 and the trough region 7, which are located in an upper region of a fabrication area for the n-type support layer 5a, and extending into the interior of the epitaxially grown crystal layer 4. The trench 18 is preferably formed by a dry etching process, which allows for easy control of its shape.

[0120] Furthermore, the width of the trench 18 is reduced to be smaller than the width of the n-type support layer 5a to avoid contact between a side surface of the trench 18 and the p-type support layer 6a, so that no channel is formed in the p-type support layer 6a.

[0121] Next, in a Fig. In the process shown in Figure 18, the insulating layer 120 is formed to create the gate insulating layer 121 such that it covers the inner surface of the trench 18. The insulating layer 120 is formed, for example, by thermal oxidation or CVD on the entire substrate surface.

[0122] Next, a conductive layer 130 is formed to create the gate electrode 131, filling the trench 18 whose inner surface is covered with the insulating layer 120. The conductive layer 130 is formed from polysilicon containing a high concentration of impurities and is produced by CVD over the entire substrate surface.

[0123] Next, an unnecessary area of ​​the conductive layer 130 is removed by means of a dry etching process. This, as described in Fig. Figure 19 shows the gate electrode 131, which protrudes partially from the inside of the trench 18. When the gate electrode 131 protrudes partially from the inside of the trench 18, as described above, a connection to a gate wiring (not shown) can be facilitated.

[0124] Next, the insulating intermediate layer 14 is formed, for example by CVD, on the entire substrate surface so that it covers the gate electrode 131. Then, as described in Fig. 20 shows the contact hole SC formed, which penetrates the insulating intermediate layer 14 in the thickness direction and extends to a region of the source area 8 and the entire surface of the contact area 10a.

[0125] It should be noted that the contact hole TC, which penetrates the stack layer consisting of the insulating intermediate layer 14 and the field insulating layer 11 in the thickness direction and extends to the contact region 10b, is formed in the connection region TR, although a representation of this is omitted. It should also be noted that when the contact hole SC is formed, the insulating layer 120, which remains on the epitaxially grown crystal layer 4, is simultaneously removed.

[0126] After arranging the contact holes SC and TC in the insulating intermediate layer 14, the source electrode 15 and the gate wiring (not shown) are formed, and the drain electrode 16 is formed on the back side of the semiconductor substrate 3. The material and manufacturing process are described in the first embodiment, and therefore a description of the same is omitted here.

[0127] Finally, the passivation layer 17 is arranged such that it covers an upper part of the source electrode 15 and an upper part of the stack layer consisting of the field insulating layer 11 and the insulating intermediate layer 14. Thus, the Fig. 16 shown vertical MOSFET 400 completed.

[0128] In comparison to the planar channel MOSFET, the trench channel MOSFET allows for a reduction in the width of the n-type support layer, resulting in a MOSFET with lower resistance. Fifth embodiment

[0129] In the vertical MOSFETs 100 to 400 described in the first to fourth embodiments, the support layers are formed using a so-called trench-filling process, in which the support layers are formed by filling trenches arranged in the epitaxially grown crystal layer 4 with a p-type semiconductor. Alternatively, the support layers can be formed using a multi-epitaxial process, in which the implantation of p-type impurity ions and the epitaxial growth of n-type crystals are repeated several times.

[0130] Fig. Figure 21 is a perspective view showing a configuration of support layers formed using a multi-epitaxial technique. The n-type support layers 5a and 5b and the p-type support layers 6a and 6b, which are shown in Fig. 21 are shown, are obtained by stacking a plurality of individual support layers MP in the thickness direction of the substrate, each formed by means of an implantation of p-type defect ions.

[0131] Specifically, the n-type epitaxial growth layer 4 is formed by first epitaxially growing crystals on the n-type semiconductor substrate 3, and then a single support layer MP is formed by implanting p-type impurity ions into a region to form a p-type support layer. The thickness of the epitaxially grown crystal layer 4 in this case is less than the thickness of the epitaxially grown crystal layer 4 used when the support layers are formed by the trench-filling method.

[0132] In ion implantation, a resist mask or an oxide layer mask formed by photolithography is used. In particular, the use of a thick oxide layer mask enables ion implantation with high acceleration energy, and the thickness of each individual support layer (MP) that can be formed by ion implantation can be increased. Therefore, the number of ion implantation and epitaxial crystal growth repetitions can be reduced.

[0133] Following ion implantation, an oxide layer is formed on the epitaxially grown crystal layer 4 by means of sacrificial oxidation. Subsequently, a modified surface layer in the epitaxially grown crystal layer 4 is removed by oxide layer removal using hydrofluoric acid to obtain a clean surface. Then, epitaxial crystal growth is performed again to form a second layer of the epitaxially grown crystal layer 4.

[0134] The thickness of the epitaxially grown crystal layer 4 in the second layer and subsequent layers is approximately equal to the thickness of the individual support layer MP and is predetermined based on the acceleration energy of the ion implantation, the material of the mask used, etc.

[0135] In this way, the formation of the epitaxially grown crystal layers 4 and the formation of the individual support layers MP are repeated such that a plurality of the individual support layers MP are stacked. Thus, support layers with a thickness similar to that of the support layers formed by a trench-filling method can be obtained.

[0136] It should be noted that the configuration in which the support distance W1 of the support layer 5a of the n-type and the support layer 6a of the p-type in the element area ER and the support distance W2 of the support layer 5b of the n-type and the support layer 6b of the p-type in the connection area TR are specified such that the relation W1 < W2 holds, and other configurations are also the same as in the first embodiment.

[0137] Fig. 22 is a perspective view of a vertical MOSFET 500 with support layers formed using a multi-epitaxial process. Note that the configuration is the same as that of the vertical MOSFET 100 according to the first embodiment, except that the p-type support layers 6a and 6b are formed by stacking individual support layers MP. Note that the passivation layer 17 is not shown.

[0138] When support layers are formed using the multi-epitaxial technique, the width of a support layer can be reduced compared to the trench-fill technique. This is because the width of a support layer is determined by the dimensions of a mask used during ion implantation, and thus the width of the support layer can be reduced to the limit of the mask's structuring capabilities. Furthermore, the depth of a support layer in the multi-epitaxial technique can be determined based on the number of repetitions of epitaxial crystal growth and ion implantation.

[0139] Therefore, the depth of a support layer in the multi-epitaxial method can be increased to a greater extent compared to the trench-fill method, and a MOSFET can be obtained that has a higher breakdown voltage and lower resistance compared to the trench-fill method.

[0140] In the multi-epitaxial process, a suitable use of an implantation mask can provide a difference between the acceptor concentrations NA1 and NA2 of the respective p-type support layers 6a and 6b formed by ion implantation. When the acceptor concentration NA2 of the p-type support layer 6b in the terminal region TR is lower than the acceptor concentration NA1 of the p-type support layer 6a, and the width Wn2 of the n-type support layer 5b, the donor concentration ND thereof, and the width Wp2 of the p-type support layer 6b in the terminal region TR have the relationship Wn2 × ND = Wp2 × NA2, the extent of a depletion layer in the p-type support layer 6b in the terminal region can be further increased. Further application examples

[0141] The first to fifth embodiments described above represent examples in which the present invention is applied to a MOSFET having the same conductivity type in the epitaxially grown crystal layer 4 and the semiconductor substrate 3. However, the present invention can also be applied to an insulated-gate bipolar transistor (IGBT) having different conductivity types in the epitaxially grown crystal layer 4 and the semiconductor substrate.

[0142] Based on the in Fig.In the configuration shown in Figure 3, for example, the configuration of an IGBT is obtained by modifying the semiconductor substrate 3 to have a p-type structure, while the epitaxially grown crystal layer 4 has an n-type structure. In such a case, the source region 8 and the source electrode 15 of the MOSFET correspond to an emitter region and an emitter electrode, respectively, of the IGBT, and the drain electrode 15 of the MOSFET corresponds to a collector electrode.

[0143] Furthermore, the present invention can also be applied to a freestanding substrate formed solely from the epitaxially grown crystal layer 4, which is obtained by removing the semiconductor substrate 3 by a mechanical, chemical, or other process. It should be noted that the freestanding substrate formed solely from an epitaxially grown layer can also be referred to as a "semiconductor substrate".

[0144] If a source region, source electrode, etc. are formed on one main surface of a freestanding substrate and a drain electrode is formed on the other main surface, then a MOSFET is obtained. If an emitter region, emitter electrode, etc. are formed on one main surface of a freestanding substrate and a collector region and collector electrode are formed on the other main surface, then an IGBT is obtained.

[0145] Although the invention is shown and described in detail, the foregoing description is to be understood in all aspects as merely illustrative and not as limiting. Therefore, it is understood that numerous modifications not shown can be conceived without deviating from the scope of the invention.

[0146] It should be noted that the embodiments according to the present invention can each be freely combined with one another and each embodiment can be modified within the scope of the invention, as far as is expedient, or features can be omitted.

Claims

[1] Semiconductor device comprising: - a semiconductor substrate (3); - a semiconductor layer (4) of a first conductivity type arranged on the semiconductor substrate (3); - a first impurity region (7) of a second conductivity type selectively arranged on an upper layer region of the semiconductor layer (4) in an active region; - a second impurity region (8) of the first conductivity type selectively disposed on an upper layer region of the first impurity region (7); - a first main electrode (15) connected to the second impurity region (8); - a gate insulating layer (12) continuously arranged in contact with the second impurity region (8), the first impurity region (7) and the semiconductor layer (4); - a gate electrode (13) arranged to oppose the second impurity region (8), the first impurity region (7) and the semiconductor layer (4) across the gate insulating layer (12); and - a second main electrode (16) arranged on a side opposite the semiconductor layer (4) of the semiconductor substrate (3), - wherein the semiconductor layer (4) comprises first column layers (5a, 5b) of the first conductivity type and second column layers (6a, 6b) of the second conductivity type, wherein the first column layers (5a, 5b) and the second column layers (6a, 6b) are arranged such that they extend from a main surface located on a side opposite the semiconductor substrate (3) with a predetermined depth towards the semiconductor substrate (3), - wherein the first column layers (5a, 5b) and the second column layers (6a, 6b) are arranged alternately in a direction parallel to the main surface in the active region of the semiconductor layer (4) and in a peripheral region which is a region around the active region, - wherein the first impurity region (7) is arranged in an upper layer region of the second column layers (6a, 6b), - wherein the semiconductor layer (4) with which the gate insulating layer (12) comes into contact consists of the first column layers (5a, 5b), - wherein the column pitch defined by the total width of a pair of one of the first column layers (5a, 5b) and one of the second column layers (6a, 6b) in the peripheral region is set to be greater than the column pitch defined by the total width of a pair of one of the first column layers (5a, 5b) and one of the second column layers (6a, 6b) in the active region, and - wherein the product of the width of one of the first pillar layers (5a, 5b) and the effective impurity concentration having the first conductivity type of the one of the first pillar layers (5a, 5b) in each of the active region and the peripheral region is equal to the product of the width of one of the second pillar layers (6a, 6b) and the effective impurity concentration having the second conductivity type of the one of the second pillar layers (6a, 6b). [2] Semiconductor unit according to claim 1, wherein the impurity concentration of the first column layers (5a) in the active region and the impurity concentration of the first column layers (5b) in the peripheral region are set to be equal, and wherein the impurity concentration of the second pillar layers (6a) in the active region and the impurity concentration of the second pillar layers (6b) in the peripheral region are set to be equal. [3] Semiconductor unit according to claim 1 or 2, wherein the first and second column layers (5a, 5b, 6a, 6b) are arranged by forming the second column layers (6a, 6b) with gaps in the semiconductor layer (4), wherein the semiconductor layer (4) is present below the first column layers (5a, 5b) and the second column layers (6a, 6b) as a buffer layer and wherein the sum of the product of the width of one of the first pillar layers (5a, 5b), the depth of one of the first pillar layers (5a, 5b) and the effective impurity concentration with the first conductivity type of one of the first pillar layers (5a, 5b) and the product of the one pillar spacing, the thickness of the buffer layer and the impurity concentration with the first conductivity type of the buffer layer is equal to the product of the width of one of the second pillar layers (6a, 6b), the depth of one of the second pillar layers (6a, 6b) and the effective impurity concentration with the second conductivity type of one of the second pillar layers (6a, 6b). [4] Semiconductor unit according to one of claims 1 to 3, wherein the semiconductor layer (4) further comprises a RESURF region (9) of the second conductivity type arranged to extend over upper layer regions of the first and second column layers (5b, 6b) in the peripheral region, wherein the RESURF region (9) is electrically connected to the first main electrode (15) and wherein the second column layers (6b) in the peripheral region are electrically connected to the first main electrode (15) via the RESURF region (9). [5] Semiconductor unit according to claim 4, wherein the RESURF area (9) is arranged to extend to an outermost periphery of the peripheral area and wherein the first and second column layers (5b, 6b) are not present below the RESURF region (9) at the outermost periphery of the peripheral region. [6] A semiconductor device according to any one of claims 1 to 5, wherein the first and second pillar layers (5b, 6b) in the peripheral region are predetermined such that the pillar pitches of the first pillar layers (5a, 5b) and the second pillar layers (6a, 6b) have the same value in the entire peripheral region. [7] A semiconductor device according to any one of claims 1 to 5, wherein the first and second pillar layers (5b, 6b) in the peripheral region are predetermined such that the pillar pitch defined by the total width of a pair of one of the first pillar layers (5a, 5b) and one of the second pillar layers (6a, 6b) arranged in a region close to the active region and the pillar pitch of another of the first pillar layers (5a, 5b) and another of the second pillar layers (6a, 6b) have different values. [8] A semiconductor device according to any one of claims 1 to 7, wherein the semiconductor substrate (3) uses silicon carbide as a semiconductor.

Citation Information

Patent Citations

  • Semiconductor device and manufacturing method thereof

    JP2001298190A

  • sj semiconductor device and method for its manufacture

    DE10106006A1

  • Semiconductor device and manufacturing process of the same

    DE102005059534A1

  • LOAD COMPENSATION STRUCTURE AND CORRESPONDING MANUFACTURING

    DE102015110112A1

  • JP002006186108A