JP6513168B2 - Semiconductor device - Google Patents

Description

  The present invention relates to a semiconductor device, and more particularly to a semiconductor device provided with an IGBT (Insulated Gate Bipolar Transistor) which is a high breakdown voltage semiconductor device.

  In the field of high breakdown voltage semiconductor devices (power devices) which control voltages exceeding several hundreds of volts, the current to be handled is also large, and therefore, device characteristics with suppressed heat generation, that is, loss are required. Moreover, as a drive system of the gate which controls those voltage * electric currents, a drive circuit is small and the voltage drive element with a small loss there is desirable.

  In recent years, for the reasons as described above, an insulated gate bipolar transistor, that is, an IGBT has become mainstream as an element capable of voltage drive and having a small loss in this field. The structure of this IGBT is a structure that can be regarded as a diode whose drain side is a diode in order to keep the breakdown voltage by lowering the impurity concentration of the drain of a MOS (Metal Oxide Semiconductor) transistor and to lower the drain resistance. .

  In such an IGBT, since the diode performs bipolar operation, in the present application, the source side of the MOS transistor of the IGBT is called the emitter side, and the drain side is called the collector side.

  In the IGBT which is a voltage driving element, generally, a voltage of several hundreds of volts is applied between the collector and the emitter, and the voltage is controlled by a gate voltage of several to several tens of volts. Also, IGBTs are often used as inverters, and although the voltage between the collector and the emitter is low when the gate is in the ON state, a large current flows, but no current flows when the gate is in the OFF state. The voltage between the collector and the emitter is high.

  Usually, since the IGBT operates in the mode as described above, the loss is divided into a steady loss which is a current-voltage product in the on state and a switching loss in the transition where the on state and the off state are switched. Be The leakage current-voltage product in the off state is so small that it can be ignored.

  On the other hand, it is also important to prevent destruction of the element even in an abnormal state, for example, when a load is shorted. In this case, while the power supply voltage of several hundred volts is applied between the collector and the emitter, the gate is turned on and a large current flows.

  In an IGBT having a structure in which a MOS transistor and a diode are connected in series, the maximum current is limited by the saturation current of the MOS transistor. For this reason, the current limitation works even at the time of the short circuit as described above, and it is possible to prevent the destruction of the element due to the heat generation for a fixed time.

The structure of the conventional IGBT is disclosed, for example, in Japanese Patent Laid-Open No. 2004-247593 (Patent Document 1). The IGBT of Patent Document 1 mainly includes a gate electrode, a source (emitter) electrode, a drain (collector) electrode, and an n-type substrate. A trench is formed on the upper surface of the n-type substrate, and the gate electrode is embedded inside the trench. A p-type base layer is formed in the upper part in the n-type substrate, and an n ⁺ -type source layer and a p ⁺ -type drain layer are formed inside the p-type base layer. The n ⁺ -type source layer and the p ⁺ -type drain layer are adjacent to each other on the surface of the n-type substrate. The gate electrode and the n ⁺ -type source layer and the p-type base layer are opposed to each other across the gate insulating film in the n-type substrate. The emitter electrode is in electrical contact with the n ⁺ -type source layer and the p ⁺ -type drain layer. A p ⁺ -type drain layer is formed on the lower surface of the n-type substrate, and the collector electrode is in contact with the p ⁺ -type drain layer on the lower surface side of the n-type substrate. An n ^-- type epitaxial layer and an n-type buffer layer are embedded between the p-type base layer and the p ⁺ -type drain layer inside the n-type substrate. The n ⁻ epitaxial layer is in contact with the p base layer and the n buffer layer, and the n buffer layer is in contact with the p ⁺ drain layer.

  In addition, IGBTs having a structure similar to that of Patent Document 1 are, for example, disclosed in Japanese Patent Application Laid-Open Nos. 2006-49933 (Patent Document 2), 2002-359373 (Patent Document 3), and Japanese Patent Application Laid-Open Nos. 9-260662 (Patents) Reference 4), U.S. Pat. No. 6,815,767 (U.S. Pat. No. 5,953,968), U.S. Pat. No. 6,953,968 (U.S. Pat. (Patent Document 7).

JP 2004-247593 A Unexamined-Japanese-Patent No. 2006-49933 JP 2002-359373 A Unexamined-Japanese-Patent No. 9-260662 U.S. Patent No. 6,815,767 U.S. Patent No. 6,953,968 U.S. Patent No. 6,781,199

In a power device, chips of a plurality of IGBTs and diodes are included in one package module, and the plurality of IGBTs are connected in parallel to one another. An important characteristic of the IGBT used for the power device is the temperature dependency of the on voltage V _CE (sat). Here, the on voltage V _CE (sat) is a voltage between the collector and the emitter required to obtain an arbitrary rated current (density) J _C. The fact that the temperature dependency of the on voltage V _CE (sat) is positive, that is, the on voltage V _CE (sat) increases with the temperature rise of the IGBT operates a plurality of IGBTs connected in parallel with one another (that is, the IGBTs Suitable for parallel operation). If the temperature dependency of the on voltage V _CE (sat) is negative, current is concentrated on the IGBT having a low on voltage V _CE (sat) when the IGBTs are operated in parallel. As a result, the package module is prone to malfunction and problems such as destruction are likely to occur.

  Therefore, an object of the present invention is to obtain a semiconductor device suitable for parallel operation.

A semiconductor device according to one aspect of the present invention includes a semiconductor substrate and an element. The semiconductor substrate has a first main surface and a second main surface facing each other. The element has a gate electrode formed on the first main surface side, a first electrode formed on the first main surface side, and a second electrode formed in contact with the second main surface. . The device generates an electric field in the channel by the voltage applied to the gate electrode, and controls the current between the first electrode and the second electrode by the electric field of the channel. The density of spikes at the interface between the semiconductor substrate and the second electrode is 0 or more and 3 × 10 ⁸ pieces / cm ² or less.

  A semiconductor device according to another aspect of the present invention includes a semiconductor substrate and an element. The semiconductor substrate has a first main surface and a second main surface facing each other. The element has a gate electrode formed on the first main surface side, a first electrode formed on the first main surface side, and a second electrode formed in contact with the second main surface. . The device generates an electric field in the channel by the voltage applied to the gate electrode, and controls the current between the first electrode and the second electrode by the electric field of the channel. The semiconductor device further includes a collector region formed on the second main surface. The collector region includes a collector diffusion layer of the first conductivity type in contact with the second electrode, a buffer diffusion layer of the second conductivity type formed closer to the first main surface than the collector diffusion layer, and a drift of the second conductivity type. And a diffusion layer. The drift diffusion layer has an impurity concentration lower than that of the buffer diffusion layer, and is formed adjacent to the buffer diffusion layer and closer to the first main surface than the buffer diffusion layer. The ratio of the number of atoms per unit area of the impurities constituting the buffer diffusion layer to the number of atoms per unit area of the impurities constituting the drift diffusion layer is 0.05 or more and 100 or less.

  According to the present invention, semiconductor devices suitable for parallel operation can be obtained.

FIG. 1 is a schematic cross-sectional view showing a configuration of a semiconductor device in a first embodiment of the present invention. FIG. 7 is a schematic cross sectional view showing the first step of the method of manufacturing a semiconductor device in the first embodiment of the present invention. FIG. 14 is a schematic cross-sectional view showing a second step of the method of manufacturing a semiconductor device in the first embodiment of the present invention. FIG. 14 is a schematic cross-sectional view showing a third step of the method of manufacturing a semiconductor device in the first embodiment of the present invention. FIG. 14 is a schematic cross-sectional view showing a fourth step of the method of manufacturing a semiconductor device in the first embodiment of the present invention. FIG. 14 is a schematic cross-sectional view showing a fifth step of the method of manufacturing a semiconductor device in the first embodiment of the present invention. FIG. 16 is a schematic cross-sectional view showing a sixth step of the method of manufacturing a semiconductor device in the first embodiment of the present invention. FIG. 14 is a schematic cross-sectional view showing a seventh step of the method of manufacturing a semiconductor device in the first embodiment of the present invention. FIG. 16 is a schematic cross-sectional view showing an eighth step of the method of manufacturing a semiconductor device in the first embodiment of the present invention. FIG. 16 is a schematic cross-sectional view showing a ninth step of the method of manufacturing a semiconductor device in the first embodiment of the present invention. FIG. 16 is a schematic cross-sectional view showing a tenth step of the method of manufacturing a semiconductor device in the first embodiment of the present invention. It is sectional drawing which shows typically the state of the interface of the p-type collector area | region in which the spike was formed, and a collector electrode. It is a top view which shows typically the state of the interface of the p-type collector area | region in which the spike was formed, and a collector electrode. Is a graph showing the temperature dependence of the relationship between the collector-emitter voltage V _CE (sat) and the current density J _C of the first embodiment of the present invention. It is a figure which shows the relationship between the spike density in Embodiment 1 of this invention, and the variation | change_quantity of on voltage. It is a figure which shows the spike density dependence of the relationship between the operation temperature of the device in Embodiment 1 of this invention, and _VCE (sat). It is a figure which shows the relationship between the film thickness of a collector electrode and spike density in Embodiment 1 of this invention. It is concentration distribution along the XVIII-XVIII line of FIG. It is concentration distribution along the XIX-XIX line of FIG. It is a figure which shows the relationship between CP _{, P} / _{CP, N} , _VCE (sat), and the energy loss E _Off in Embodiment 2 of this invention. It is a figure which shows the relationship between CP _{, P} / _{CP, N} , _VCE (sat), and leak current density J _{CES in} IGBT which has a withstand voltage of 1200 V class in Embodiment 2 of this invention. C _{P, P} / C _P relationship V _CE in a second embodiment of the present invention and (sat) and J _C, illustrates the _N dependence. And S _N / S _N-in a second embodiment of the present invention and showing a relationship between V _CE (sat) and the breakdown voltage BV _CES. C _{S, P} and C _P in the second embodiment of the present _invention, a _P, a diagram showing the temperature dependence of the relationship between V _CE (sat). It is a figure which shows the CS _{, P} and _CP dependence dependence of the relationship of the operation temperature of a device and _VCE (sat) in Embodiment 2 of this invention. In a second embodiment of the present invention, showing _{^{5 × 10 15 ≦ C S,}} P, 1 × 10 16 ≦ C P, the temperature dependence of J _C -V _CE characteristics when _P. In a second embodiment of the present invention, showing a _{^{5 × 10 15> C S,}} P, 1 × 10 16> C P, temperature dependence of J _C -V _CE characteristics when _P. D _P according to the second embodiment of the present _invention, the _N or D _N-, is a diagram showing a relationship between V _CE (sat) and BV _CES. It is another example of concentration distribution along the XVIII-XVIII line of FIG. It is a figure which shows the relationship between _{SN *} / _SN and _VCE (sat) in Embodiment 2 of this invention. It is a figure which shows the relationship between depth x from a 2nd main surface in Embodiment 2 of this invention, and _VCE (sat). It is a figure which shows the relationship between (tau) _x / (tau) _N- and _VCE (sat) in Embodiment 2 of this invention. It is a figure which shows an example of the relationship of the depth x from a 2nd main surface in Embodiment 2 of this invention, and carrier lifetime. It is a figure which shows the relationship between the output of the laser annealing and the temperature of a diffusion furnace in Embodiment 2 of this invention, and carrier lifetime. It is a figure which shows the relationship between the ion implantation amount in Embodiment 2 of this invention, a carrier activation rate, _VCE (sat), and BV _CES . FIG. 21 is an enlarged cross sectional view schematically showing a second main surface of a semiconductor substrate in a third embodiment of the present invention. It is a figure which shows the relationship between centerline average roughness _Ra and largest height _Rmax , and breaking strength and carrier lifetime in Embodiment 3 of this invention. And R _a and R _max of the third embodiment of the present invention, is a diagram showing the relationship between J _CES and V _CE (sat). FIG. 21 is a cross sectional view showing a configuration of a MOS transistor portion of a semiconductor device in a fourth embodiment of the present invention. FIG. 35 is a cross sectional view showing a configuration of a first modified example of the semiconductor device in the fourth embodiment of the present invention. FIG. 35 is a cross sectional view showing a configuration of a second modified example of the semiconductor device in the fourth embodiment of the present invention. FIG. 35 is a cross sectional view showing a configuration of a third modification of the semiconductor device in the fourth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. FIG. 25 is a schematic cross sectional view showing the derivative structure of the MOS transistor structure according to the fifth embodiment of the present invention. It is a schematic sectional drawing which shows the various structures of planar gate type | mold IGBT in Embodiment 6 of this invention. It is a schematic sectional drawing which shows the various structures of planar gate type | mold IGBT in Embodiment 6 of this invention. It is a schematic sectional drawing which shows the various structures of planar gate type | mold IGBT in Embodiment 6 of this invention. It is a schematic sectional drawing which shows the various structures of planar gate type | mold IGBT in Embodiment 6 of this invention. It is a schematic sectional drawing which shows the various structures of planar gate type | mold IGBT in Embodiment 6 of this invention. FIG. 80 is a view schematically showing a concentration distribution of carriers (n-type impurities) immediately below the gate electrode 5a in the configuration shown in FIGS. 79 to 83. in the case of not forming the case of forming the n-type impurity diffusion region is a diagram showing a relationship between V _CE and J _C. And S _N14a / S _N- according to the sixth embodiment of the present invention, V _CE (sat), illustrates J _{C, Break} and V _G, and the relationship between the _Break. It is a top view which shows the layout of the semiconductor device in Embodiment 7 of this invention. FIG. 90 is a cross sectional view taken along line LXXXVIII-LXVIII of FIG. 87. FIG. 90 is a cross sectional view taken along line LXXXIX-LXXXIX of FIG. 87. It is impurity concentration distribution along the XC-XC line of FIG. It is a figure which shows the relationship between Y / X and BV _CES in Embodiment 7 of this invention. Is a diagram illustrating relationship, and D _T and E the relationship _{P / CS} or E _{P / N-and} the D _T and BV _CES according to the seventh embodiment of the present invention. D _T according to the seventh embodiment of the present invention _and showing a relationship between the _Pwell and BV _CES and .DELTA.BV _CES. It is a schematic sectional drawing which shows the various structures of planar gate type IGBT in Embodiment 7 of this invention. It is a schematic sectional drawing which shows the various structures of planar gate type IGBT in Embodiment 7 of this invention. It is a figure which shows the relationship between W _CS and X _CS, and V _CE and E _SC . FIG. 35 is a plan view showing a layout of n-type emitter region 3 and p ⁺ impurity diffusion region 6 in the semiconductor device in the seventh embodiment of the present invention. FIG. 35 is a plan view showing a modification of the layout of n-type emitter region 3 and p ⁺ impurity diffusion region 6 in the semiconductor device in the seventh embodiment of the present invention. Is a diagram showing the relationship between α and V _CE (sat) and E _SC according to a seventh embodiment of the present invention. It is a top view which shows typically the layout of the gate pad in Embodiment 8 of this invention. It is a figure for demonstrating the oscillation phenomenon of gate voltage. It is a figure for demonstrating the oscillation phenomenon of gate voltage. FIG. 21 schematically shows an electric field intensity distribution along the line XIX-XIX of FIG. 1 when a reverse bias slightly lower than the breakdown voltage is applied to the main junction of the IGBT in the ninth embodiment of the present invention. It is a figure which shows the relationship of the electric field strength of the junction plane in the Embodiment 9 of this invention, and a breakdown voltage.

Hereinafter, embodiments of the present invention will be described based on the drawings.
Embodiment 1
FIG. 1 is a schematic cross-sectional view showing the configuration of the semiconductor device in the first embodiment of the present invention. 1, a semiconductor device of this embodiment, for example, assuming a semiconductor device having a breakdown voltage of 600～6500V, are trench IGBT formed on a semiconductor substrate having a thickness t ₁ of 50~800μm . The semiconductor substrate has a first main surface (upper surface) and a second main surface (lower surface) opposed to each other. The n ⁻ drift layer (drift diffusion layer) 1 has a concentration of 1 × 10 ¹² to 1 × 10 ¹⁵ cm ⁻³ , for example, assuming a semiconductor device having a withstand voltage of 600 to 6500 V. For example, a p-type semiconductor having a concentration of about 1 × 10 ¹⁵ to 1 × 10 ¹⁸ cm ⁻³ and a diffusion depth of about 1.0 to 4.0 μm from the first main surface on the first main surface side of the semiconductor substrate A p-type body region 2 is formed. For example, the first main surface in p type body region 2 (body diffusion layer) has a concentration of 1 × 10 ¹⁸ to 1 × 10 ²⁰ cm ⁻³ and a diffusion depth of about 0.3 from the first main surface. An n-type emitter region 3 made of an n-type semiconductor of ̃2.0 μm is formed. In the first main surface adjacent to n-type emitter region 3 (second emitter diffusion layer), p ⁺ impurity diffusion region 6 (first emitter diffusion) for providing a low resistance contact to p-type body region 2 Layer) is formed at a concentration of, for example, about 1 × 10 ¹⁸ to 1 × 10 ²⁰ cm ⁻³ and the diffusion depth from the first main surface is equal to or less than the depth of n-type emitter region 3.

In the first main surface, a gate groove 1a which penetrates n-type emitter region 3 and p-type body region 2 to reach n ^- drift layer 1 is formed. The gate groove 1a has a depth of, for example, 3 to 10 μm from the first main surface, and the pitch of the gate groove 1a is, for example, 2.0 μm to 6.0 μm. A gate insulating film 4a is formed on the inner surface of the gate groove 1a. The gate insulating film 4a is formed of, for example, a silicon oxide film formed by a CVD method and a silicon oxide film or nitrogen formed by a thermal oxidation method in order to improve the characteristics, reliability and device yield of the gate insulating film. It has a laminated structure with a silicon oxynitride film segregated at the SiO ₂ interface.

A gate electrode 5a made of, for example, polycrystalline silicon in which phosphorus is introduced at a high concentration, or a metal material such as W / TiSi ₂ is formed so as to fill the inside of the gate groove 1a. A silicide layer (for example, TiSi ₂ , CoSi, etc.) may be formed on the surface of the gate electrode 5 a in order to reduce the resistance of the gate electrode 5 a. An insulating film 22A made of, for example, a silicon oxide film is formed on the upper surface of gate electrode 5a. The gate electrode 5a is electrically connected to a control electrode for applying a gate potential G. The gate electrode 5a may be formed on the first main surface side.

Thus, a gate trench is formed of the gate groove 1a, the gate insulating film 4a and the gate electrode 5a. In addition, n ^- drift layer 1, n-type emitter region 3 and gate electrode 5a, n ^- drift layer 1 is a drain, n-type emitter region 3 is a source, and opposed to gate electrode 5a with gate insulating film 4a interposed therebetween. An insulated gate field effect transistor portion (here, a MOS transistor) having a portion of p type body region 2 as a channel is formed. That is, this MOS transistor generates an electric field in the channel by the voltage applied to gate electrode 5a, and controls the current between emitter electrode 11 and collector electrode 12 by the electric field of the channel. A plurality of these MOS transistors are arranged on the first main surface.

Over the first main surface, an insulating film 9 made of, for example, silicate glass and an insulating film 22B made of a silicon oxide film formed by a CVD method are formed. A contact hole 9a reaching the main surface is provided. A barrier metal layer 10 is formed along the inner surface of contact hole 9a and the upper surfaces of insulating films 9 and 22B. A silicide layer 21a is formed in a portion where the barrier metal layer 10 and the semiconductor substrate are in contact with each other. An emitter electrode 11 (first electrode) for giving an emitter potential E is electrically connected to the n-type emitter region 3 and the p ⁺ impurity diffusion region 6 through the barrier metal layer 10 and the silicide layer 21 a. The emitter electrode 11 may be formed on the first main surface side.

Further, a p-type collector region 8 (collector diffusion layer) and an n-type buffer region 7 (buffer diffusion layer) are formed on the second main surface side of the semiconductor substrate. A collector electrode 12 (second electrode) for giving a collector potential C is electrically connected to the p-type collector region 8. The collector electrode 12 is formed on the second main surface side of the semiconductor substrate, and applies a collector potential C. The material of the collector electrode 12 is, for example, an aluminum compound. The n-type buffer region 7 is formed closer to the first major surface than the p-type collector region 8. The n ⁻ drift layer 1 has an impurity concentration lower than that of the n-type buffer region 7 and is adjacent to the n-type buffer region 7 and located closer to the first major surface than the n-type buffer region 7. The p-type collector region 8, the n-type buffer region 7 and the n ⁻ drift layer 1 constitute a collector region.

In particular, the provision of the n-type buffer region 7 reduces the main junction leak characteristics and increases the withstand voltage as compared with the case where the n-type buffer region 7 is not provided. Also, the tail current is reduced in the I _C waveform at turn-off, and as a result, the switching loss (E _OFF ) is reduced.

  Further, the diffusion depth of the n-type buffer region 7 becomes shallow because the n-type buffer region 7 is formed after the impurity diffusion region on the MOS transistor side is formed. That is, in order to suppress the adverse effect of the high temperature heat treatment on the impurity diffusion region on the MOS transistor side, when forming the n-type buffer region 7, an annealing technique in which the temperature is locally raised as in the low temperature annealing technique or laser annealing. In order to use

  In the semiconductor device of the present embodiment, for example, at the time of inverter connection, it is a pulse-like control signal which is set to -15 V in the off state and to +15 V in the on state. The collector potential C of the collector electrode 12 is set to a voltage between the power supply voltage and the saturation voltage according to the gate potential G.

Next, the manufacturing method of the present embodiment will be described.
2 to 11 are schematic cross sectional views showing the method of manufacturing the semiconductor device in the first embodiment of the present invention in the order of steps. First, referring to FIG. 2, for example, the peak concentration is 1 × 10 ¹⁵ to 1 × 10 ¹⁸ cm ⁻³ and the diffusion depth from the first main surface on the first main surface of the semiconductor substrate including n ⁻ drift layer 1. A p-type body region 2 of 1.0 to 4.0 μm is formed. Next, mask layer 31 is formed on the first major surface.

Referring to FIG. 3, mask layer 31 is patterned. Using the patterned mask layer 31 as a mask, ion implantation, for example, is performed to provide a surface concentration of 1.0 × 10 ^{18 to} 1.0 × 10 ^{20 on the first} major surface in p-type body region 2. An n-type emitter region 3 is formed with a cm ^-3 and a diffusion depth of 0.3 to 2.0 μm from the first major surface. Thereafter, the mask layer 31 is removed.

Referring to FIG. 4, a silicon oxide film 32 formed by thermal oxidation, for example, and a silicon oxide film 33 formed by CVD are sequentially formed on the first main surface. The silicon oxide films 32 and 33 are patterned by ordinary photolithography and etching techniques. The semiconductor substrate is anisotropically etched using the patterned silicon oxide films 32 and 33 as a mask. As a result, a gate groove 1a which penetrates the n-type emitter region 3 and the p-type body region 2 and reaches the n ^- drift layer 1 is formed.

  Referring to FIG. 5, by performing processing such as isotropic plasma etching and sacrificial oxidation, the opening and the bottom of gate groove 1a are rounded, and the unevenness of the side wall of gate groove 1a is flattened. Ru. Further, the sacrificial oxide film 32a is formed on the inner surface of the gate groove 1a so as to be integrated with the thermal oxide film 32 by the above-described sacrificial oxidation. By performing isotropic plasma etching and sacrificial oxidation in this manner, it becomes possible to improve the characteristics of the gate insulating film formed on the inner surface of the gate groove 1a. Thereafter, oxide films 32, 32a, 33 are removed.

  Referring to FIG. 6, the removal of the oxide film exposes the first main surface of the semiconductor substrate and the inner surface of gate groove 1a.

Referring to FIG. 7, a gate insulating film 4a made of, for example, a silicon oxide film is formed along the inner surface and the first main surface of gate groove 1a. For example, a material in which phosphorus is introduced by ion implantation into polycrystalline silicon in which phosphorus is introduced at a high concentration or polycrystalline silicon in which no impurity is introduced, or W (tungsten) / TiSi so as to fill the inside of gate groove 1a. A conductive layer 5 made of a metal material such as ₂ (titanium silicide) is formed on the entire surface.

  The gate insulating film 4a may be a silicon oxide film formed by a CVD method and a silicon oxide film formed by thermal oxidation or nitrogen, silicon and silicon oxide, for the purpose of improving the characteristics, reliability and device yield as a gate insulating film. It is preferable to use a laminated structure composed of a nitrided oxide film segregated at the interface with the above.

  Thereafter, conductive layer 5 is patterned by the usual photolithography and etching techniques.

Referring to FIG. 8, by this patterning, the conductive layer is left in gate groove 1a to form gate electrode 5a. Here, a silicide layer (for example, TiSi ₂ , COSi or the like) may be formed on the surface of the gate electrode 5 a in order to reduce the resistance of the gate electrode 5 a. Thereafter, the upper surface of gate electrode 5a is oxidized to form insulating film 22A made of, for example, a silicon oxide film. After this, for example, the surface concentration of the first main surface is 1.0 × 10 ^{18 to} 1.0 × 10 ²⁰ cm ⁻³ , and the diffusion depth from the first main surface is p ⁺ impurity shallower than n-type emitter region 3. Diffusion region 6 is formed.

  Referring to FIG. 9, an insulating film 9 made of, for example, silicate glass and an insulating film 22B made of a silicon oxide film formed by the CVD method are sequentially formed on the first main surface. In the insulating films 9 and 22B, contact holes 9a are formed by ordinary photolithography and etching techniques.

  Referring to FIG. 10, barrier metal layer 10 made of, for example, a metal layer is formed by sputtering. Thereafter, lamp annealing is performed to form a silicide layer 21a at the contact portion between the barrier metal layer 10 and the semiconductor substrate. Thereafter, emitter electrode 11 is formed.

Referring to FIG. 11, n ^- drift layer 1 on the second main surface side of the semiconductor substrate is polished. The polishing, the thickness t ₁ of the semiconductor substrate is adjusted according to the required withstand voltage of the MOS transistor. For example, to manufacture an IGBT having a withstand voltage of 600 V to 6500 V, the thickness t ₃ (FIG. 1) of the n ⁻ drift layer 1 is 50 to 800 μm. After polishing, etching or the like of the second main surface of the semiconductor substrate is performed to recover the crystallinity of the polished surface.

  Thereafter, n-type impurities and p-type impurities are implanted into the second main surface of the semiconductor substrate, for example, by ion implantation, and then the impurities are diffused. Alternatively, immediately after the n-type impurity and the p-type impurity are implanted, heat treatment is performed according to the implantation depth of each impurity. As a result, n-type buffer region 7 and p-type collector region 8 are formed. Further, a collector electrode 12 is formed to complete the semiconductor device shown in FIG. Collector electrode 12 is made of, for example, aluminum or other metal material capable of obtaining ohmic contact with p-type collector region 8.

In the present embodiment, as shown in FIG. 11, after the emitter electrode 11 is formed, the second main surface of the n ⁻ drift layer 1 is polished to form the n type buffer region 7 and the p type collector region 8. May be In addition, as shown in FIG. 2, the second main surface may be polished before the p-type body region 2 is formed. Alternatively, as shown in FIG. 9, the second main surface may be polished after or before the opening of contact hole 9a to form n-type buffer region 7 and p-type collector region 8.

In the present embodiment, the spike density at the interface between the semiconductor substrate and the collector electrode 12 (formed by the reaction between the semiconductor material forming the p-type collector region 8 and the metal material on the p-type collector region 8 side in the collector electrode 12). The density of the spikes made of an alloy is 0 or more and 3 × 10 ⁸ pieces / cm ² or less.

  12 and 13 schematically show the state of the interface between the p-type collector region where the spikes are formed and the collector electrode. FIG. 12 is a cross-sectional view, and FIG. 13 is a plan view. Referring to FIGS. 12 and 13, at the interface between p type collector region 8 and collector electrode 12, a plurality of spikes are generally formed. The spike is, for example, a protrusion (or a recess) made of an alloy of a material forming the collector electrode 12 and a material forming the p-type collector region 8 and having, for example, a quadrangular pyramid shape or an octagonal pyramid shape. Here, when the collector electrode 12 is formed of a multilayer film, the spike is formed of an alloy of a material forming the layer 12 a in direct contact with the p-type collector region 8 and a material forming the p-type collector region 8. It is formed.

  The spike density is measured, for example, by the following method. First, the collector electrode 12 is dissolved using a chemical solution and removed from the semiconductor substrate. Then, the second main surface of the exposed semiconductor substrate is observed with a microscope, and the number of concave portions, such as square pyramids and octagonal pyramids, present on the second main surface is counted. As a result, the value obtained by dividing the obtained number by the observed area is defined as the spike density.

As the spike density increases, the ionization rate of impurities in the p-type collector region 8 at low temperature (less than 298 K) decreases, and the effective injection efficiency of carriers (holes) from the p-type collector region 8 to the n-type buffer region 7 Decreases. Therefore, the J _c -V _CE characteristics of the IGBT depend on the spike density.

The following effects can be obtained by setting the spike density to 0 or more and 3 × 10 ⁸ pieces / cm ² or less. FIG. 14 is a diagram showing the temperature dependency of the relationship between collector-emitter voltage and current density in the first embodiment of the present invention. Referring to FIG. 14, V _CE (sat) is an emitter-collector voltage corresponding to any rated current density. At temperatures of 298 K and 398 K, substantially the same curve is obtained regardless of whether the spike density is 3 × 10 ⁸ pieces / cm ² or more or 3 × 10 ⁸ pieces / cm ² or less. On the other hand, at a temperature of 233 K, the emitter-collector voltage is significantly increased when the spike density is 3 × 10 ⁸ pieces / cm ² or less.

FIG. 15 is a diagram showing the relationship between the spike density and the amount of change in the on voltage in the first embodiment of the present invention. FIG. 15 shows the results when the conditions (concentration, depth) of the p-type collector region 8 and the n-type buffer region 7 are constant. Further, the variation ΔV _on of the on voltage in FIG. 15 is a value obtained by subtracting the collector-emitter voltage V _CE (sat) (233 K) at 233 K from the collector-emitter voltage V _CE (sat) at 298 K (298 K). is there. Referring to FIG. 15, when the spike density D _spike is 3 × 10 ⁸ / cm ² or less, the collector-emitter voltage V _CE at 298K (sat) is the collector-emitter of 233K voltage V _CE ( sat) or higher. On the other hand, if the spike density D _spike exceeds 3 × 10 ⁸ pieces / cm ^2, the collector-emitter voltage V _CE at 298K (sat) is the collector-emitter voltage V _CE (sat) than the value at 233K It becomes.

FIG. 16 is a graph showing the spike density dependency of the relationship between the operating temperature of the device and the voltage between the collector and the emitter in the first embodiment of the present invention. Referring to FIG. 16, when the spike density D _spike is 3 × 10 ⁸ pieces / cm ² or less, the temperature dependency of the voltage V _CE (sat) is positive, whereas the spike density D _spike is In the case of 3 × 10 ⁸ cells / cm ² or more, the temperature dependence of the voltage V _CE (sat) is negative in the region of less than 298K.

As described above, by setting the spike density at the interface between the semiconductor substrate and the collector electrode 12 to 0 or more and 3 × 10 ⁸ pieces / cm ² or less as in the present embodiment, the temperature dependency of the collector-emitter voltage V _CE Can be positive. As a result, when the IGBTs are operated in parallel, the concentration of the current to the IGBT having a low voltage V _CE is eliminated, and a semiconductor device suitable for the parallel operation can be obtained.

  The spike density can be controlled, for example, by the material of the collector electrode, the heat treatment condition, or the film thickness of the collector electrode. As a material of the collector electrode, a silicide containing Al, AlSi, Ti, and a metal is suitable. Examples of the metal-containing silicide include Ti-containing silicide, Ni-containing silicide, and Co-containing silicide. Further, as a material of the collector electrode, a material showing ohmic resistance to the contacting semiconductor layer (p-type collector region 8 in FIG. 1) such as Al or AlSi is preferable. As a material of the semiconductor substrate, Si, SiC, GaN or Ge is suitable. In particular, when silicide is used as the collector electrode, no spike exists at the interface between the semiconductor substrate and the collector electrode. The collector electrode made of silicide is formed by forming a metal made of Ti, Co, Ni or the like on the second main surface of the semiconductor substrate made of Si, SiC, GaN, Ge or the like, and performing heat treatment.

The film thickness of the collector electrode is preferably 200 nm or more. FIG. 17 is a diagram showing the relationship between the film thickness of the collector electrode and the spike density in the first embodiment of the present invention. Referring to FIG. 17, when the film thickness of the collector electrode is 200 nm or more, the spike density is 3 × 10 ⁸ / cm ² or less. On the other hand, the thickness of the collector electrode is preferably 10000 nm or less from the viewpoint of the production limit.

The spike density can be set to 0 or more and 3 × 10 ⁸ pieces / cm ² or less by appropriately combining the material of the collector electrode, the heat treatment condition, or the film thickness of the collector electrode as described above.

  In the present embodiment, the case of an IGBT having the configuration shown in FIG. 1 is shown. However, the semiconductor device of the present invention is not limited to the configuration of FIG. 1, and may be any device including a semiconductor substrate having a first main surface and a second main surface facing each other, and an element. This element has a gate electrode formed on the first main surface side, a first electrode formed on the first main surface side, and a second electrode formed in contact with the second main surface. ing. The element generates an electric field in the channel by the voltage applied to the gate electrode, and controls the current between the first electrode and the second electrode by the electric field of the channel. Furthermore, a device structure such as a diode may be used.

Second Embodiment
FIG. 18 is a concentration distribution along line XVIII-XVIII in FIG. FIG. 19 is a concentration distribution along line XIX-XIX in FIG. FIG. 18 also shows the conventional concentration distribution of p-type impurities or n-type impurities.

Referring to FIGS. 18 and 19, concentration C _{S, P} is the impurity concentration of p-type collector region 8 at the interface between collector electrode 12 and p-type collector region 8 (the second main surface of the semiconductor substrate), The concentration C _{P, P} is the maximum value of the impurity concentration of the p-type collector region 8. The concentration C _{P, N} is the maximum value of the impurity concentration of the n-type buffer region 7. The concentration C _sub is the impurity concentration of the n ⁻ drift layer 1. The depth D _p is the depth from the second major surface to the junction surface between the p-type collector region 8 and the n-type buffer region 7. The depth D _{P, N} is the depth from the second main surface up to the position of the concentration C _{P, N} in the n-type buffer region 7. The depth D _N− is the depth from the second main surface to the junction surface between the n-type buffer region 7 and the n ⁻ drift layer 1. When n-type intermediate layer 7a is formed as shown in FIG. 29 described later, depth _DN is equal to the second main surface of the junction surface between n-type buffer region 7 and n-type intermediate layer 7a. It is the depth from τ _P is the carrier lifetime of the p-type collector region 8, τ _N is the carrier lifetime of the n-type buffer region 7, and τ _N− is the carrier lifetime of the n ⁻ drift layer 1. τ _X is a carrier lifetime of a position at a depth of x from the second major surface. _SN is the number of atoms per unit area of the impurities constituting the n-type buffer region 7 (atom / cm ² ), and S _N− is the number of atoms per unit area of the impurities constituting the n ⁻ drift layer 1 (atom / Cm ² ). The number of atoms per unit area of impurities in the desired region is determined by integrating the impurity concentration profile in that region over the entire depth direction.

The inventor of the present application has found that the abnormal operation of the IGBT can be suppressed by setting the relationship between the p-type collector region 8, the n-type buffer region 7 and the n ⁻ drift layer 1 under the following conditions. Here, the suppression of the abnormal operation of the IGBT means the following.

a. No snap back characteristics occur in the J _c -V _CE characteristics at temperatures below 298 K.

b. IGBT should turn on even at low temperatures below 298 K.
c. Have a desired breakdown voltage, or that the IGBT does not run out of heat at 398 K or more.

FIG. 20 is a diagram showing a relationship between C _{P, P} / C _{P, N} and V _CE (sat) and energy loss E _Off at turn-off in the second embodiment of the present invention. E _Off is an energy loss when the switching device is turned off. V _snap-back is the collector-emitter voltage at point A shown in FIG. 22 when the snap back characteristic occurs. FIG. 21 is a diagram showing the relationship between C _{P, P} / C _{P, N} and V _CE (sat) and leakage current density J _CES in the IGBT according to the second embodiment of the present invention. The leak current density J _CES is the leak current density between the collector and the emitter in the state where the gate and the emitter are shorted. Referring to FIGS. 20 and 21, ratio of the maximum value of the impurity concentration of p-type collector region 8 to the maximum value of the impurity concentration of n-type buffer region 7 is C _{P, P} / C _{P, N} is C _{P, P} / In the case of C _{P, N} <1, a snapback characteristic occurs, and a snapback voltage V _snap-back occurs accordingly. As a result, as shown in FIG. 22, when C _{P, P} / C _{P, N} <1, V _CE (sat) for any current density increases. In addition, in the case of C _{P, P} / C _{P, N} > 1 × 10 ³ , J _CES increases and thermal runaway of the IGBT occurs. From the above, in order to suppress the abnormal operation of the IGBT, it is preferable that 1 ≦ _{CP, P} / _{CP, N} ≦ 1 × 10 ³ .

FIG. 23 is a diagram showing the relationship between S _N / S _N- , V _CE (sat) and breakdown voltage BV _{CES according} to the second embodiment of the present invention. The breakdown voltage BV _CES is the breakdown voltage between the collector and the emitter in the state where the collector and the emitter are shorted. Referring to FIG. 23, the number of atoms per unit area of impurity constituting n-type buffer region 7 relative to the number of atoms per unit area of impurity constituting n ^- drift layer 1 (atom / cm ² ) ² ) When the ratio S _N / S _N- is 0.05 ≦ S _N / S _N- , a high breakdown voltage BV _CES is obtained. Further, S _N / S _N-is the case of the S _N / S _N- ≦ 100 is snapback characteristic is suppressed, and is suppressed emitter-collector voltage V _CE (sat) is also low. From the above, it is preferable that 0.05 ≦ S _N / S _{N −} ≦ 100 in order to suppress the abnormal operation of the IGBT and to enable the parallel operation.

Figure 24 is a diagram showing C _{S, P} and C _P in the second embodiment of the present _invention, and _P, and the temperature dependence of the relationship between V _CE (sat). Referring to FIG. 24, the emitter-collector voltage V can be obtained by setting 5 × 10 ¹⁵ ≦ C _{S, P} and 1 × 10 ¹⁶ ≦ C _{P, P} for all temperatures of 233 K, 298 K, and 398 K. _CE (sat) is greatly reduced. Further, in consideration of the production limit, it is preferable that Cs _{, P} 1.0 1.0 x 10 ²² cm ^-3 and C _{p, P} 1.0 1.0 x 10 ²² cm ^-3 .

FIG. 25 is a diagram showing CS _{, P} and _{CP, P} dependencies of the relationship between the operating temperature of the device and V _CE (sat) in the second embodiment of the present invention. FIGS. 26 and 27 are diagrams showing the temperature dependency of the J _C -V _CE characteristic in the second embodiment of the present invention. Referring to FIGS. 24 to 27, it can be seen that the temperature dependence of V _CE (sat) is positive in the case of 5 × 10 ¹⁵ ≦ C _{S, P} and 1 × 10 ¹⁶ ≦ C _{P, P.}

From the above, in order to suppress the abnormal operation of the IGBT, it is preferable that 5 × 10 ¹⁵ ≦ C _{S, P} and 1 × 10 ¹⁶ ≦ C _{P, P.}

Figure 28 is, D _P according to the second embodiment of the present _invention, the _N or D _N-, is a diagram showing a relationship between V _CE (sat) and BV _CES. Referring to FIG. 28, depth D _{P, N} from the second main surface to the position to be concentration C _{P, N} in n-type buffer region 7 is 0.4 μm ≦ D _{P, N} , or n-type High breakdown voltage BV _CES and low emitter-collector voltage when the depth D _N− from the second main surface of the junction surface between buffer region 7 and n ⁻ drift layer 1 is 0.4 μm ≦ D _{N −} V _CE (sat) is obtained. On the other hand, when D _{P, N} > 50 μm or D _N− > 50 μm, snapback characteristics occur.

From the above, in order to suppress the abnormal operation of the IGBT, it is preferable that 0.4 μm ≦ DP _{, N} ≦ 50 μm, and 0.4 μm ≦ _DN− ≦ 50 μm.

FIG. 29 is another example of the concentration distribution along line XVIII-XVIII in FIG. Referring to FIG. 29, the collector region may further include an n-type intermediate layer 7a. The maximum value _{CP, N *} of the impurity concentration of n-type intermediate layer 7a is lower than the maximum value _{CP, N} of the impurity concentration of n-type buffer region 7, and higher than the impurity concentration _Csub of n ^- drift layer 1 . The n-type intermediate layer 7 a is in contact with both the n-type buffer region 7 and the n ⁻ drift layer 1. The depth _DN is the depth from the second main surface of the junction surface between the n-type buffer region 7 and the n-type intermediate layer 7a. The depth D _{N *} is the depth from the second main surface of the junction surface between the n-type intermediate layer 7 a and the n ⁻ drift layer 1. _{SN *} is the number of atoms per unit area of the impurity constituting the n-type intermediate layer 7a (atom / cm ² ). The n-type intermediate layer 7 a may be formed by implanting impurity ions into a part of the n-type buffer region 7. Alternatively, it may be formed by implanting ions that generate crystal defects that become lifetime killers into a part of the n-type buffer region 7 by a method such as proton irradiation.

FIG. 30 is a diagram showing the relationship between _{SN *} / _SN and V _CE (sat) in the second embodiment of the present invention. Referring to FIG. 30, the number of atoms per unit area of impurity constituting n-type intermediate layer 7a relative to the number of atoms per unit area of impurity constituting n-type buffer region 7 (atom / cm ² ) ² ) The snapback characteristic occurs when the ratio S _{N *} / S _N is 0.5 <S _{N *} / S _N.

From the above, in order to suppress the abnormal operation of the IGBT, it is preferable that 0 <S _{N *} / S _N ≦ 0.5.

FIG. 31 is a diagram showing the relationship between the depth x from the second main surface and V _CE (sat) in the second embodiment of the present invention. FIG. 32 is a diagram showing the relationship between τ _x / τ _{N −} and V _CE (sat) in the second embodiment of the present invention. FIG. 33 is a diagram showing an example of the relationship between the depth x from the second main surface and the carrier lifetime according to the second embodiment of the present invention. Referring particularly to FIG. 33, a defect is introduced into the semiconductor substrate in the vicinity of the second main surface in ion implantation for forming p type collector region 8 and n type buffer region 7. The n-type buffer region 7 needs to be annealed at a temperature higher than that of the p-type collector region 8 because the n-type buffer region 7 needs to be implanted deeper than when the p-type collector region 8 is formed. There is. As a result, thermal stress due to annealing occurs in the n-type buffer region 7, and the carrier lifetime τ _N of the n-type buffer region 7 is lower than the carrier lifetime τ _P of the p-type collector region 8. The carrier lifetimes of the n-type buffer region 7 and the p-type collector region 8 are lower than the carrier lifetime τ _N ⁻ of the n ⁻ drift layer 1.

Therefore, particularly in a region where the depth x from the second main surface is 0.50 μm ≦ x ≦ 60.0 μm, the position of the depth x from the second main surface to the carrier lifetime τ _N− of the n ⁻ drift layer 1 By setting the ratio τ _x / τ _N− of the carrier lifetime τ _x of 1 × 10 ⁻⁶ ≦ τ _x / τ _{N −} ≦ 1, in particular, as shown in FIG. 31 and FIG. The voltage V _CE (sat) is significantly reduced.

  Here, the cause of the decrease in carrier lifetime is that defects are introduced into p-type collector region 8 and n-type buffer region 7 during ion implantation for forming p-type collector region 8 and n-type buffer region 7. It is to In order to improve the carrier lifetime, it is effective to anneal the portion where the defect is introduced. Next, the relationship between the annealing technique and the carrier lifetime is shown.

  FIG. 34 is a diagram showing the relationship between the output of laser annealing and the temperature of the diffusion furnace and the carrier lifetime in the second embodiment of the present invention. Referring to FIG. 34, when annealing is performed in the diffusion furnace, the carrier lifetime decreases if the temperature of the diffusion furnace is too high. In addition, in the case of performing laser annealing at high output energy in the laser annealing technology, a decrease in carrier lifetime occurs. Further, since the laser has the property of being attenuated inside the semiconductor substrate, if the depth from the second main surface of the semiconductor substrate to the junction surface between p type collector region 8 and n type buffer region 7 is too deep It is necessary to increase the output of the laser annealing, and it becomes difficult to improve the carrier lifetime by the laser annealing. Taking this into consideration, the depth from the second main surface of the semiconductor substrate to the junction surface between the p-type collector region 8 and the n-type buffer region 7 is preferably more than 0 and not more than 1.0 μm.

FIG. 35 is a diagram showing the relationship between the ion implantation amount and the carrier activation rate, V _CE (sat) and BV _{CES in the} second embodiment of the present invention. Referring to FIG. 35, the activation rate of each of n-type buffer region 7 and p-type collector region 8 depends on the ion implantation amount of n-type buffer region 7 and p-type collector region 8 or the type of ions. . In FIG. 35, the activation rate in p-type collector region 8 and the activation rate in n-type buffer region 7 are different from each other, and the activation rate in p-type collector region 8 is higher than the activation rate in n-type buffer region 7 Is also lower. As a result, the IGBT can operate normally, and the breakdown voltage BV _CES can be increased. In particular, when the activation rate in p type collector region 8 is more than 0 and 90% or less, collector-emitter voltage V _CE (sat) is greatly reduced.

The activation rate is calculated by the following equation (1).
Activation rate: {(impurity concentration obtained from resistance value calculated by a method such as SR (spreading-resistance) measurement (cm ^-3 )) / (impurity concentration measured using SIMS (Secondary Ionization Mass Spectrometer) (Cm ^-3 ))} × 100 (1)
By using the above-described collector structure, it is possible to ensure normal operation of the IGBT, maintain high withstand voltage, and suppress thermal runaway of the IGBT. Further, in improving the device characteristics N ^- on the drift layer was thinned, it is possible to obtain flexibility for trade-off characteristics of the V _CE (sat) -E _OFF (control of).

Third Embodiment
In order to improve the V _CE (sat) -E _off characteristic which is an important device characteristic of the IGBT, it is effective to thin the n ^- drift layer 1. However, when the second main surface of the semiconductor substrate is polished as shown in FIG. 11, the inventor of the present invention has found that the surface roughness of the polished surface affects various characteristics of the IGBT.

FIG. 36 is an enlarged cross sectional view schematically showing a second main surface of the semiconductor substrate in the third embodiment of the present invention. Referring to FIG. 36, the center line average roughness defined in the present embodiment, JIS (Japanese Industrial Standard) to a defined by the center line average roughness R _a, of the absolute value deviation from the mean line It is an average value. Further, the maximum height is the maximum height R _max defined in JIS, and the height (R) from the lowest valley bottom in the reference length (height H _min ) to the maximum peak (height H _max ) _max = is H _max -H _min).

FIG. 37 is a diagram showing a relationship between center line average roughness and maximum height, breaking strength and carrier lifetime in the third embodiment of the present invention. Referring to FIG. 37, when 0 <R _a ≦ 200 nm and 0 <R _max ≦ 2000 nm, high breaking strength and carrier lifetime can be obtained. FIG. 38 is a diagram showing the relationship between center line average roughness and maximum height, and J _CES and V _CE (sat) in the third embodiment of the present invention. Referring to FIG. 38, in the case of 0 <R _a ≦ 200 nm and 0 <R _max ≦ 2000 nm, low collector-emitter voltage V _CE (sat) and low leak current density J _CES can be obtained.

As described above, various characteristics of the IGBT can be improved by setting 0 <R _a ≦ 200 nm or 0 <R _max ≦ 2000 nm.

Embodiment 4
In the present embodiment, the configuration of a MOS transistor which provides the same effects as the effects obtained by the configurations of the first to third embodiments will be described.

FIG. 39 is a cross sectional view showing a configuration of a MOS transistor portion of the semiconductor device in the fourth embodiment of the present invention. Referring to FIG. 39, in structure D of the MOS transistor portion of the present embodiment, relatively high concentration n-type impurity diffusion occurs in the vicinity of a region where n ^- drift layer 1 and p-type body region 2 form a pn junction. This is different from the structure C shown in FIG. 1 in that the region 14 (buried diffusion layer) is provided. The n-type impurity diffusion region 14 is formed between the p-type body region 2 and the n ⁻ drift layer 1. Although not shown, the structure A of FIG. 1 is formed under the structure D of FIG.

  The remaining structure is substantially the same as that of the structure C shown in FIG. 1, and therefore the same members are denoted by the same reference characters and description thereof is not repeated.

  The configuration in which n-type impurity diffusion region 14 is provided is not limited to the configuration of FIG. 39, and may be, for example, the configurations shown in FIGS. 40 and 41. That is, the n-type impurity diffusion region 14 may be provided in the configuration in which the emitter trench is provided.

FIG. 40 is a cross sectional view showing a configuration of a modification of the semiconductor device in the fourth embodiment of the present invention. Referring to FIG. 40, in this structure E, an emitter trench is provided in a region sandwiched between two MOS transistors. The emitter trench is composed of an emitter groove 1b, an emitter insulating film 4b, and an emitter conductive layer 5b. Emitter groove 1 b penetrates p-type body region 2 and n-type impurity diffusion region 14 to reach n ⁻ drift layer 1. Emitter insulating film 4b is formed along the inner surface of emitter groove 1b. Emitter conductive layer 5b is formed to be embedded in emitter groove 1b, and is electrically connected to emitter electrode 11 in the upper layer. Any number of emitter trenches may be formed, and a gate trench may be formed in at least one of the plurality of trenches.

  A barrier metal layer 10 is formed under the emitter electrode 11, and a silicide layer 21b is formed between the barrier metal layer 10 and the emitter conductive layer 5b.

A p ⁺ impurity diffusion region 6 for forming a low resistance contact to p type body region 2 is formed on the first main surface sandwiched between two emitter trenches, and silicide layer 21a is formed thereon. There is.

In such a configuration, a relatively high concentration n-type impurity diffusion region 14 is provided in the vicinity of the region where the n ^- drift layer 1 and the p-type body region 2 form a pn junction.

  The remaining structure is substantially the same as that of the structure D shown in FIG. 39, so the same members are denoted by the same reference characters and description thereof is not repeated.

  Structure F shown in FIG. 41 is different from structure E shown in FIG. 40 in that n-type impurity diffusion region 3 is added to the side of the emitter trench and on the first main surface.

  The remaining structure is substantially the same as that of the structure E shown in FIG. 39, so the same members are denoted by the same reference characters and description thereof is not repeated.

  In FIG. 40 and FIG. 41, the case where emitter conductive layer 5b embedded in emitter groove 1b is at the emitter potential has been described, but emitter conductive layer 5b may have a floating potential. The configuration is described below.

  Referring to FIG. 42, emitter conductive layer 5b embedded in emitter groove 1b is electrically separated from emitter electrode 11 and has a floating potential. In this case, an insulating film 22A made of, for example, a silicon oxide film, an insulating film 9 made of, for example, a silicate glass, and an insulating film 22B made of, for example, a silicon oxide film are formed on the emitter conductive layer 5b filling the emitter groove 1b. It is formed.

  The remaining structure is substantially the same as that of the structure E shown in FIG. 40, and therefore the same members are denoted by the same reference characters and description thereof is not repeated.

  The n-type impurity diffusion region 14 provided in the present embodiment is formed by ion implantation and diffusion before the p-type body region 2 is formed. Thereafter, p-type body region 2 is formed, and the same post-process as in the first embodiment is performed to manufacture various semiconductor devices (FIGS. 39 to 42) of the present embodiment.

  Also, each of the MOS transistor structures E (FIG. 40), F (FIG. 41), and G (FIG. 42) has a trench with an emitter potential or a floating potential, thereby forming the MOS transistor structures C (FIG. 1) and D (FIG. 39) The effective gate width is smaller than that of 39). As a result, the structures E, F, and G have a smaller current flow than the structures C and D, and have the effect of suppressing the saturation current.

Furthermore, in the structures E, F, and G, the ON voltage becomes larger at a lower voltage / lower current density than the structure D. In the MOS transistor structure D, the ON voltage is lowered because the n-type impurity diffusion region 14 described in US Pat. No. 6,040,599 has a carrier accumulation effect even if the n ^- drift layer 1 is thick in the collector structure A It is. In the MOS transistor structure D, even if the n ^- drift layer 1 is thicker than the conventional structure, the ON voltage can be reduced.

  In the MOS transistor structures E, F, and G, when the device is switched in a no-load state, an arbitrary current can be maintained for a longer time than the conventional structure or the MOS transistor structures C and D due to the effect of lowering the saturation current. it can. That is, the MOS transistor structures E, F, and G have the effect of suppressing the saturation current of the device and improving the breakdown tolerance.

  Furthermore, in the MOS transistor structure D having the effect of lowering the ON voltage, an oscillation phenomenon occurs at the time of switching in a no-load state. However, in the MOS transistor structures E, F, and G, even if the n-type impurity diffusion region 14 exists, the emitter conductive layer 5b which is at the emitter potential or the floating potential is effective in preventing the oscillation phenomenon.

Fifth Embodiment
43 to 78 are schematic cross sectional views showing various derived structures of the MOS transistor structure which can obtain the same effect as the fourth embodiment. With any of the structures shown in FIGS. 43 to 78, the effect of the MOS transistor structure shown in the fourth embodiment can be obtained.

Each MOS transistor structure shown in FIGS. 43 to 78 will be described below.
In the configuration shown in FIG. 43, an n-type emitter region 3 is formed only on one side of a point where one emitter trench serving as an emitter potential is provided in a region sandwiched between two MOS transistor portions and one side of gate groove 1a. 40 differs from the configuration of the structure E shown in FIG.

  In the configuration shown in FIG. 44, a plurality of emitter grooves 1b are embedded with an emitter conductive layer 5b consisting of a single integrated layer. The emitter conductive layer 5b is electrically connected to the barrier metal layer 10 and the emitter electrode 11 through the silicide layer 21b. The silicide layer 21b is formed on a bridge connecting the emitter grooves 1b. In addition, insulating films 22A, 9, 22B are formed on the emitter conductive layer 5b other than the region where the silicide layer 21b is formed.

  The configuration other than this is substantially the same as the configuration of the structure E shown in FIG. 40 described above, so the same members are denoted with the same reference numerals and description thereof is omitted.

  The configuration shown in FIG. 45 is different from the configuration shown in FIG. 44 in that n-type impurity diffusion region 3 is added to both side walls of emitter groove 1b and to the first main surface.

  The structure shown in FIG. 46 is different from the structure shown in FIG. 44 in that the emitter conductive layer 5b filling the emitter groove 1b has a floating potential. In this case, the insulating films 22A, 9, 22B are formed on the entire surface of the emitter conductive layer 5b, and the emitter conductive layer 5b is electrically insulated from the emitter electrode 11.

  The structure shown in FIG. 47 is different from the structure shown in FIG. 43 in that n-type impurity diffusion region 3 is added to both side walls of emitter groove 1b and to the first main surface.

  The configuration shown in FIG. 48 is different from the configuration shown in FIG. 43 in that the upper surface of emitter conductive layer 5b protrudes above emitter groove 1b. In this case, the emitter conductive layer 5b is electrically connected to the barrier metal layer 10 and the emitter electrode 11 through the silicide layer 21b formed on a part of the surface. In addition, insulating films 22A, 9, 22B are formed on the emitter conductive layer 5b other than the region where the silicide layer 21b is formed.

  The configuration shown in FIG. 49 is different from the configuration shown in FIG. 48 in that n-type impurity diffusion region 3 is added to both sides of emitter groove 1b and to the first main surface.

  The configuration shown in FIG. 50 is different from the configuration of structure E shown in FIG. 40 in that p-type body region 2 is formed only in the vicinity of the side wall of gate groove 1a.

  The structure shown in FIG. 51 differs from the structure of structure F shown in FIG. 41 in that p-type body region 2 is formed only in the vicinity of the side wall of gate groove 1a.

  The structure shown in FIG. 52 is different from the structure shown in FIG. 50 in that the emitter conductive layer 5b filling the emitter groove 1b has a floating potential. In this case, the insulating films 22A, 9, 22B are formed on the emitter conductive layer 5b.

  The configuration shown in FIG. 53 is different from the configuration shown in FIG. 43 in that p-type body region 2 is formed only in a region sandwiched by two gate trenches.

  The configuration shown in FIG. 54 is different from the configuration shown in FIG. 44 in that p-type body region 2 is formed only in the vicinity of the sidewall of gate groove 1a.

  The configuration shown in FIG. 55 is different from the configuration shown in FIG. 45 in that p type body region 2 is formed only in the vicinity of the side wall of gate groove 1a.

  The configuration shown in FIG. 56 is different from the configuration shown in FIG. 46 in that p type body region 2 is formed only in the vicinity of the sidewall of gate groove 1a.

  The structure shown in FIG. 57 is different from the structure shown in FIG. 53 in that n-type impurity diffusion region 3 is added to both side walls of emitter groove 1b and to the first main surface.

  The configuration shown in FIG. 58 differs from the configuration shown in FIG. 48 in that p-type body region 2 is formed only in the region sandwiched by two gate trenches.

  The configuration shown in FIG. 59 is different from the configuration shown in FIG. 49 in that p-type body region 2 is formed only in a region sandwiched by two gate trenches.

  The structure shown in FIG. 60 is a gate trench so that the gate width (W) becomes the same as that of MOS transistor structures E to G described above without forming a trench in the region where the emitter trench exists in structure E shown in FIG. Is formed, that is, a configuration in which the space between the gate trenches is expanded to an arbitrary size so as to be an emitter potential.

In this case, ap ⁺ impurity diffusion region 6 for taking a low resistance contact with the p-type body region extends on the first main surface sandwiched between the two gate trenches. Silicide layer 21 a is formed in contact with p ⁺ impurity diffusion region 6 and n-type emitter region 3. The p ⁺ impurity diffusion region 6 and the n-type emitter region 3 are electrically connected to the emitter electrode 11 via the silicide layer 21 a and the barrier metal layer 10.

  The configuration other than this is substantially the same as the configuration of FIG. 40 described above, so the same members are denoted with the same reference numerals and description thereof is omitted.

  The structure shown in FIG. 61 is a structure in which the gate trench is formed so that the gate width becomes equal to that of MOS transistor structures E to G described above without forming the trench in the region where the emitter trench exists in FIG. It is the structure expanded to arbitrary dimensions so that it may become an emitter electric potential between trenches.

Also in this configuration, the p ⁺ impurity diffusion region 6 extends on the first main surface sandwiched by the gate trenches in order to have a low resistance contact to the p-type body region. Silicide layer 21 a is formed in contact with p ⁺ impurity diffusion region 6 and n-type emitter region 3. The p ⁺ impurity diffusion region 6 and the n-type emitter region 3 are electrically connected to the emitter electrode 11 via the silicide layer 21 a and the barrier metal layer 10.

  The configuration other than this is substantially the same as the configuration of FIG. 43 described above, so the same members are denoted with the same reference numerals and description thereof is omitted.

  62 differs from the configuration shown in FIG. 60 in that p-type body region 2 is formed only in the vicinity of the side wall of gate groove 1a.

  The configuration shown in FIG. 63 is different from the configuration shown in FIG. 61 in that p-type body region 2 is formed only in a region sandwiched by two gate trenches.

  In the above, although the case where the upper surface of gate electrode 5a was located in gate groove 1a was explained, it may project on gate groove 1a. The configuration in which the upper surface of the gate electrode 5a protrudes on the upper surface of the gate groove 1a is shown in FIGS.

  64 shows the configuration of the structure E shown in FIG. 40, FIG. 65 shows the configuration shown in FIG. 41, FIG. 66 shows the configuration shown in FIG. 42, FIG. 67 shows the configuration shown in FIG. 45 shows the configuration shown in FIG. 46, FIG. 71 shows the configuration shown in FIG. 47, FIG. 72 shows the configuration shown in FIG. 48, FIG. 73 shows the configuration shown in FIG. In the configuration, the upper surface of the gate electrode 5a corresponds to a configuration in which it protrudes above the gate groove 1a. In the structure shown in FIG. 66, the upper surface of emitter conductive layer 5b filling the emitter groove 1b also protrudes above the emitter groove 1b.

  Although the trench type gate structure has been described above, the configurations of the first to fourth embodiments can be applied to a planar gate type IGBT. 75 to 78 are schematic cross-sectional views showing the configuration of a planar gate type IGBT.

Referring to FIG. 75, the planar gate type IGBT is formed, for example, on a semiconductor substrate having a thickness of about 50 μm to 250 μm. For example, a p-type body region 2 made of a p-type semiconductor is selectively formed on the first principal surface side of n ^- drift layer 1 having a concentration of 1 × 10 ¹⁴ cm ^-3 . For example, p type body region 2 has a concentration of 1 × 10 ¹⁵ to 1 × 10 ¹⁸ cm ⁻³ and a diffusion depth of about 1.0 to 4.0 μm from the first major surface. The first main surface in p-type body region 2 has a concentration of, for example, 1 × 10 ¹⁸ to 1 × 10 ²⁰ cm ⁻³ or more, and a diffusion depth of about 0.3 to 2.0 μm from the first main surface. An n-type emitter region 3 made of an n-type semiconductor is formed. Next to the n-type emitter region 3, the p ⁺ impurity diffusion region 6 for providing a low resistance contact to the p-type body region 2 is, for example, about 1 × 10 ¹⁸ to 1 × 10 ²⁰ cm ⁻³ . The diffusion depth from the main surface is equal to or less than the depth of n-type emitter region 3.

A gate electrode 5a is formed on the first main surface via a gate insulating film 4 so as to face the p-type body region 2 sandwiched between the n ^- drift layer 1 and the n-type emitter region 3.

The drift layer 1 and the n-type emitter region 3 and the gate electrode 5a, n ^{- -} the n drift layer 1 and the drain, the n-type emitter region 3 and the source, facing the gate electrode 5a through the gate insulating film 4 p An insulated gate field effect transistor portion (here, a MOS transistor portion) in which a portion of the mold body region 2 is a channel is formed.

  An emitter conductive layer 5b serving as an emitter potential is formed on the first main surface sandwiched between the two MOS transistor portions. As a material of the conductive layer 5b for the emitter and the gate electrode 5a, for example, polycrystalline silicon in which phosphorus is introduced at a high concentration, high melting point metal material, high melting point metal silicide, or a composite film of them is used.

An insulating film 9 is formed on the first main surface, and a contact hole 9 a reaching a part of the surface of the first main surface is formed in the insulating film 9. A barrier metal layer 10 is formed at the bottom of the contact hole 9a. Emitter electrode 11 giving emitter potential E is electrically connected to emitter conductive layer 5b, p ⁺ impurity diffusion region 6 and n-type emitter region 3 through barrier metal layer 10.

Further, an n-type buffer region 7 and a p-type collector region 8 are formed in order on the second main surface side of the n ⁻ drift layer 1. A collector electrode 12 for giving a collector potential C is electrically connected to the p-type collector region 8. The material of the collector electrode 12 is, for example, an aluminum compound.

In the present embodiment, the spike density at the interface between the semiconductor substrate and the collector electrode 12 (that is, the interface between the p-type collector region 8 and the collector electrode 12) is 0 or more and 3 × 10 ⁸ pieces / cm ² or less.

  It is to be noted that n-type impurity diffusion region 14 may be added as shown in FIG. 76 to the configuration of FIG. 75, and n-type buffer region 7 may be omitted as shown in FIG. As shown at 78, an n-type impurity diffusion region 14 may be added and the n-type buffer region 7 may be omitted.

Sixth Embodiment
In the present embodiment, another configuration of the planar gate type IGBT shown in FIGS. 75 to 78 will be described. 79 to 83 are schematic cross sectional views showing various configurations of the planar gate IGBT in the sixth embodiment of the present invention.

Referring to FIG. 79, the planar gate type IGBT is formed on, for example, a semiconductor substrate having a thickness of about 50 μm to 800 μm. A p-type body region 2 made of a p-type semiconductor is selectively formed on the first main surface on the left side in the drawing of the n ⁻ drift layer 1. For example, p type body region 2 has a concentration of 1 × 10 ¹⁵ to 1 × 10 ¹⁸ cm ⁻³ and a diffusion depth of about 1.0 to 4.0 μm from the first major surface. The first main surface in p-type body region 2 has a concentration of, for example, 1 × 10 ¹⁸ to 1 × 10 ²⁰ cm ⁻³ or more, and a diffusion depth of about 0.3 to 2.0 μm from the first main surface. An n-type emitter region 3 made of an n-type semiconductor is formed. On the left side of the n-type emitter region 3 in the drawing, ap ⁺ impurity diffusion region 6 for forming a low resistance contact to the p-type body region 2 is formed at a distance from the n-type emitter region 3. The p ⁺ impurity diffusion region 6 is formed, for example, at about 1 × 10 ¹⁸ to 1 × 10 ²⁰ cm ⁻³ and the diffusion depth from the first main surface is equal to or less than the depth of the n-type emitter region 3.

A gate electrode 5a is formed on the first main surface via a gate insulating film 4 so as to face the p-type body region 2 sandwiched between the n ^- drift layer 1 and the n-type emitter region 3. The gate electrode 5a extends to the right end in the drawing, and is opposed to the n ⁻ drift layer 1 via the gate insulating film 4 on the right side in the drawing.

The n ^- drift layer 1, the n-type emitter region 3 and the gate electrode 5 a make the n ^- drift layer 1 the drain, the n-type emitter region 3 the source, and p facing the gate electrode 5 a with the gate insulating film 4 interposed therebetween. An insulated gate field effect transistor portion (here, a MOS transistor) having a portion of the mold body region 2 as a channel is formed.

Insulating film 9 and emitter electrode 11 are formed on the first main surface. Insulating film 9 covers n-type emitter region 3 and p-type body region 2 on the first main surface and gate electrode 5a. Emitter electrode 11 covers p ⁺ impurity diffusion region 6 and insulating film 9, and applies emitter potential E to p ⁺ impurity diffusion region 6 and n-type emitter region 3.

Further, an n-type buffer region 7 and a p-type collector region 8 are formed in order on the second main surface side of the n ⁻ drift layer 1. A collector electrode 12 for giving a collector potential C is electrically connected to the p-type collector region 8.

In the present embodiment, the spike density at the interface between the semiconductor substrate and the collector electrode 12 (that is, the interface between the p-type collector region 8 and the collector electrode 12) is 0 or more and 3 × 10 ⁸ pieces / cm ² or less.

  The configuration shown in FIG. 80 is that p type body region 2 is formed deeper (further closer to the second main surface side) in a region where insulating film 9 is not formed in plan view. It is different from the composition of Such a p-type body region 2 is formed by adding a step of implanting a p-type impurity into the first major surface using the insulating film 9 as a mask.

Configuration shown in FIG. 81, n so as to be adjacent to the side surface of the p-type body region 2 ^- in that the n-type impurity diffusion region 14a in the drift layer 1 is formed, it is different from the configuration of FIG. 79.

  In the configuration shown in FIG. 82, p region body region 2 is formed deeper (further closer to the second main surface side) in a region where insulating film 9 is not formed in plan view. It is different from the composition of

The configuration shown in FIG. 83 is different from the configuration of FIG. 81 in that n-type impurity diffusion region 14a is further formed in n ^- drift layer 1 so as to be adjacent to the bottom surface of p-type body region 2. .

By forming n-type impurity diffusion region 14a adjacent to p-type body region 2 as in the structure shown in FIGS. 81 to 83, as shown in FIG. 84, the emitter side (when IGBT is in the on state) The carrier concentration on the first main surface side increases. As a result, the characteristics of the IGBT can be improved. FIG. 85 is a diagram showing the relationship between V _CE and J _C in the case where the n-type impurity diffusion region is formed and in the case where it is not formed. Referring to FIG. 85, in the case of forming the n-type impurity diffusion region 14a, the emitter-collector voltage V _CE is reduced relative to the current density J _C.

Figure 86 is a diagram illustrating the S _N14a / S _N- according to the sixth embodiment of the present _{invention, V CE (sat), J} C, Break and V _G, and the relationship between the _Break. Here, S _N14a / S _N- A, n ^- the number of atoms per unit area of the impurity constituting the drift layer 1 (atom / cm ²⁾ units of impurities constituting the n-type impurity diffusion region 14a with respect to S _N- It is a ratio of the number of atoms per area (atom / cm ² ) _SN14a . J _{C, Break} is a current density that allows the device to shut off in RBSOA (Reverse Bias Safety Operation Area) mode, and VG _{, Break} is a gate that can shut off the device in SCOA (Short Circuit Safe Operation Area) mode It is a voltage. Referring to FIG. 86, in the case of 0 <S _N14a / S _N− ≦ 20, high blocking performance is obtained, and low collector-emitter voltage V _CE (sat) is obtained. Therefore, in order to reduce the on-voltage while ensuring the RBSOA and SCSOA are preferably n-type impurity diffusion region 14a satisfies _{0 <S N14a / S N- ≦} 20.

Seventh Embodiment
FIG. 87 is a plan view showing the layout of the semiconductor device in the seventh embodiment of the present invention. FIG. 88 is a cross-sectional view taken along line LXXXVIII-LXVIII of FIG. 87, and FIG. 89 is a cross-sectional view taken along line LXXXIX-LXXXIX of FIG. FIG. 90 is an impurity concentration distribution along the XC-XC line in FIG. The hatched portion in FIG. 87 is a region where p-type impurity diffusion region 41 is formed. Further, in FIG. 87, only the gate groove 1a (dotted line in the figure) formed along one gate electrode wiring 11a is shown, but in practice, a plurality of gate grooves are formed along each gate electrode wiring 11a. A gate groove 1a (or an emitter groove 1b) is formed. The configuration of the IGBT according to the present embodiment will be described with reference to FIGS.

  Referring particularly to FIG. 87, emitter electrodes 11 and gate electrode interconnections 11a are alternately arranged in the lateral direction in the drawing and extend in the longitudinal direction in the drawing. At the lower end portion of the gate electrode wiring 11 a in the center of the chip in the drawing, a gate pad 28 for electrically connecting to another wiring is provided. Further, each of the plurality of gate grooves 1a is arranged in the vertical direction in the drawing along the extending direction of the gate electrode wiring 11a immediately below the gate electrode wiring 11a. Each of the plurality of gate grooves 1a is arranged along the extending direction (longitudinal direction in the drawing) of the short side of the rectangular planar shape. A p-type body region 2 and an n-type impurity diffusion region 14 are formed between gate grooves 1a adjacent in the vertical direction in the drawing. Further, p-type impurity diffusion regions 41 (well layers) are formed between the emitter electrodes 11 adjacent to each other in the lateral direction in the drawing (that is, the end of the gate groove 1a). The p-type impurity diffusion region 41 extends in the longitudinal direction in the drawing along the emitter electrode 11 directly below the gate electrode wiring 11 a.

Referring particularly to FIG. 88, n-type impurity diffusion region 14 is formed between p-type body region 2 and n ⁻ drift layer 1. As shown in FIG. 90, n-type impurity diffusion region 14 has an impurity concentration higher than the impurity concentration of n ⁻ drift layer 1. When n-type impurity diffusion region 14 is present, at least one of gate groove 1a and emitter groove 1b (for example, FIG. 40) is doped with an impurity concentration of 1 × 10 ¹⁶ cm in n-type impurity diffusion region 14. ^By projecting the second main surface side from the position of ^-3 , a high withstand voltage (BV _CES ) can be held. The configuration shown in FIG. 88 is substantially the same as the configuration of structure D shown in FIG.

  Particularly referring to FIG. 89, gate electrode 5a filling the inside of gate groove 1a extends also on the first main surface outside gate groove 1a, and electrically extends from gate electrode interconnection 11a in the extended portion. It is connected. The barrier metal layer 10 is located under the gate electrode interconnection 11a, and the silicide layer 21a is formed in a region where the barrier metal layer 10 and the gate electrode 5a are in contact with each other. A passivation film 15 is formed on the gate electrode wiring 11 a and the emitter electrode 11. The p-type impurity diffusion region 41 reaches a deeper position (to the second main surface side) than the gate groove 1a.

  Although all the grooves shown in FIG. 87 are gate grooves 1a in which the gate electrode 5a is embedded, at least one of these grooves may be a gate groove, and the other grooves are for example for an emitter. It may be a groove.

Here, referring to FIG. 88, a pitch between the gate groove 1a and another groove adjacent to the groove (the gate groove 1a on the right side in the drawing) is defined as a pitch X. Further, the depth from the first main surface of the semiconductor substrate to the bottom of the gate groove 1a constituting the gate trench is defined as depth Y. Further, the junction surface between p type body region 2 and n type impurity diffusion region 14 (junction surface between p type body region 2 and n ^- drift layer 1 when n type impurity diffusion region 14 is not formed) The protrusion amount of the gate groove 1a from the above is defined as a protrusion amount _DT . Further, referring to FIG. 89, the distance (depth) from the junction surface of p type impurity diffusion region 41 and n ^- drift layer 1 to the bottom of gate groove 1a is defined as depth _{DT, Pwell} .

  The inventors of the present application have found that, in an IGBT having a trench gate structure, the breakdown voltage (breakdown voltage) of the IGBT can be improved by designing the gate trench under the following conditions.

FIG. 91 is a diagram showing the relationship between Y / X and BV _CES in the seventh embodiment of the present invention. Referring to FIG. 91, when depth Y from the first main surface of the semiconductor substrate to the bottom of gate groove 1a forming the gate trench is larger than the pitch between gate groove 1a and another groove adjacent thereto ( In other words, in the case of 1.0 ≦ Y / X), a high breakdown voltage BV _CES is obtained.

Figure 92 is a diagram showing the relationship of relationships, and D _T and E _{P / CS} or E _{P / N-and} the D _T and BV _CES according to the seventh embodiment of the present invention. Here, E _{P / CS} means the electric field strength at the junction surface between p-type body region 2 and n-type impurity diffusion region 14, and E _{P / N-} means that n-type impurity diffusion region 14 is formed. It means the electric field strength at the junction surface between the p-type body region 2 and the n ⁻ drift layer 1 when it is not. Referring to FIG. 92, electric field strength E when protrusion amount _DT of gate groove 1a from the junction surface between p type body region 2 and n type impurity diffusion region 14 is 1.0 μm ≦ _{DT. P / CS} or E _{P / N-} is reduced, and a high breakdown voltage BV _CES is obtained.

FIG. 93 is a diagram showing the relationship between _{DT, Pwell} and BV _CES and ΔBV _CES in the seventh embodiment of the present invention. Here, ΔBV _CES means a value obtained by subtracting BV _CES when the gate potential is -20 V from BV _CES when the gate potential is 0 V (the same potential as the emitter potential). Referring to FIG. 93, the depth _{DT, Pwell} from the bottom of gate groove 1a to the bottom of p-type impurity diffusion region 41 (the junction surface between p-type impurity diffusion region 41 and n ^- drift layer 1) is D _In the case of _{T, Pwell} ≦ 1.0 μm, a high breakdown voltage BV _CES is obtained, and the variation amount ΔBV _CES of the breakdown voltage is also suppressed low.

As described above, by manufacturing the groove 1a for the gate and the groove 1b for the emitter so as to satisfy the condition of 1.0 ≦ Y / X, 1.0 μm ≦ D _T , or 0 <D _{T, Pwell} ≦ 1.0 μm, The breakdown voltage of the IGBT can be improved.

  In FIG. 88, the configuration in which n-type impurity diffusion region 14 is formed over the entire space between gate grooves 1a is described. However, n-type impurity diffusion region 14 is shown in FIGS. 94 and 95 below. As such, it may be formed only in a part between the plurality of grooves.

  94 and 95 are schematic cross sectional views showing various configurations of the trench gate type IGBT in the seventh embodiment of the present invention. In the structure shown in FIG. 94, n-type impurity diffusion region 14 is formed only around the gate trench. The n-type impurity diffusion region 14 is formed so as to be in contact with the gate groove 1a and not to be in contact with the emitter groove 1b. On the other hand, in the configuration shown in FIG. 95, n-type impurity diffusion region 14 is formed only around the emitter trench. The n-type impurity diffusion region 14 is formed to be in contact with each of the two emitter grooves 1 b and not to be in contact with the gate groove 1 a.

  The remaining configuration is substantially the same as the configuration of the structure E shown in FIG. 40, so the same members are denoted by the same reference characters and description thereof is not repeated.

  The inventor of the present application has found that the voltage between the collector and the emitter can be reduced and the breakdown energy can be improved by controlling the width of the n-type impurity diffusion region 14 and the distance from the emitter groove 1b.

FIG. 96 is a diagram showing the relationship between W _CS and X _CS and V _CE and E _SC . Here, W _CS is the width of the n-type impurity diffusion region 14 in the region existing around the emitter groove 1 b in plan view, and X _CS is the width of the n-type impurity diffusion region 14 from the emitter groove 1 b. The distance to the end of the Referring to FIG. 96, when width W _{CS of} n-type impurity diffusion region 14 is 6 μm ≦ W _CS ≦ 9 μm, or distance X _CS from emitter groove 1b to the end of n-type impurity diffusion region 14 is 0. In the case of 5 μm ≦ X _CS ≦ 2 μm, the collector-emitter voltage V _CE is reduced and a high short-circuit breakdown energy _ESC can be obtained.

FIG. 97 shows a planar layout of n-type emitter region 3 and p ⁺ impurity diffusion region 6 in the semiconductor device in the seventh embodiment of the present invention. Referring to FIG. 97, each of gate electrode 5a and emitter conductive layer 5b extends in the vertical direction in the drawing, and between gate electrode 5a and emitter conductive layer 5b and between emitter conductive layers 5b. The n-type emitter region 3 is formed between them. The n-type emitter region 3 extends in the vertical direction in the figure, and p ⁺ impurity diffusion regions 6 are periodically formed in a region sandwiched by the n-type emitter regions 3. Further, as shown in FIG. 98, n-type emitter regions 3 and p ⁺ impurity diffusion regions 6 are alternately formed along the extending direction (vertical direction in the figure) of gate electrode 5a or emitter conductive layer 5b. May be

Here, as shown in FIGS. 97 and 98, the width of n-type emitter region 3 along the extending direction of gate electrode 5a is defined as W 2 _SO, and p ⁺ impurity along the extending direction of gate electrode 5a. The width of the diffusion region 6 is defined as W _PC . The present inventors have discovered that by controlling the relation between the W _SO and W _PC, possible to reduce the collector-emitter voltage was found to be able to improve the fracture energy.

FIG. 99 is a diagram showing the relationship between α and V _CE (sat) and E _SC in the seventh embodiment of the present invention. α (%) is a value defined by α = (W _SO / W _SO + W _PC ) × 100. Referring to FIG. 99, when α is in the range of 8.0% ≦ α ≦ 20.0%, a low collector-emitter voltage V _CE (sat) is obtained, and a high breakdown energy E _SC is obtained. Be

Eighth Embodiment
FIG. 100 schematically shows a planar layout of a gate pad in the eighth embodiment of the present invention. Referring to FIG. 100, in the present embodiment, a part of the current path of gate electrode interconnection 11a (FIG. 87) is formed of resistor 28a locally having high resistance. In FIG. 100, a part of the gate pad 28 for electrically connecting the wiring (surface gate wiring) and the gate electrode wiring 11a is formed by the resistor 28a. Each of the resistors 28 a protrudes so as to face each other at an opening provided at a central portion of the gate pad 28. Resistor 28a may have the same structure as, for example, gate electrode 5a shown in FIG. 1 or FIG.

FIGS. 101 and 102 are diagrams for explaining the oscillation phenomenon of the gate voltage. In such IGBT and MOS transistors of the trench gate structure, the switching speed increases, the time variation of the current I _c as shown in FIG. 101, the collector-emitter voltage V _CE oscillates. The cause of this is that the LCR circuit constant causes the device to oscillate. Therefore, by providing the resistor 28a, it becomes an LCR circuit constant which makes the device difficult to oscillate. As a result, as shown in FIG. 102, the oscillation phenomenon of the gate voltage _V.sub.ge can be suppressed.

(Embodiment 9)
In order to improve the V _CE (sat) -E _OFF characteristics in the IGBT, n ^- although it is effective to reduce the thickness of the drift layer 1, n ^- when the thickness of the drift layer 1, a high breakdown voltage It will be difficult to realize. Therefore, the inventor of the present application has determined the electric field intensity EP _{/ CS} of the junction surface between the p-type body region 2 and the n-type impurity diffusion region 14 (if the n-type impurity diffusion region 14 is not formed, n ^- the electric field strength E _{P / N-)} of the junction surface between the drift layer 1, n type buffer region 7 and n ^- noting the joint surface field strength E _{n / n-and} the relationship between the drift layer 1 It has been found that the withstand voltage of the IGBT can be improved.

  FIG. 103 schematically shows an electric field strength distribution along line XIX-XIX in FIG. 1 when a reverse bias slightly lower than the breakdown voltage is applied to the main junction of the IGBT in the ninth embodiment of the present invention FIG. FIG. 104 is a diagram showing the relation between the electric field strength of the junction and the breakdown voltage in the ninth embodiment of the present invention.

Referring to FIG. 103, when a reverse bias slightly lower than the breakdown voltage is applied to the main junction of the IGBT, the electric field in the semiconductor is p-type body region 2 and n ^- drift from the first main surface of the semiconductor substrate. rapidly increases in the region up to the junction surface of the layer 1, then, n ^- slowly decreased in the drift layer within 1, n ^- has decreased sharply in the drift layer 1 and the n-type buffer region 7. In addition, the electric field is 0 in the p-type body region 2 and the n-type buffer region 7. Referring to FIG. 104, the electric field strength E _{P / N-} of the junction surface between n ^- drift layer 1 and p-type body region 2 is 0 <E _{P / N-} ≦ 3.0 × 10 ¹⁵ (V / cm) In the case of, a high breakdown voltage BV _CES is obtained. In addition, when the electric field strength E _{N / N-} of the junction surface between the n-type buffer region 7 and the n ^- drift layer 1 is 2.0 × 10 ¹⁴ ≦ E _{N / N-} (V / cm), a high breakdown voltage BV _CES is obtained. E _{N / N-} is preferably E _{P / N-} or less.

  The structures or numerical ranges described in the first to eighth embodiments can be combined as appropriate.

  The embodiments disclosed above should be considered in all respects as illustrative and not restrictive. The scope of the present invention is shown not by the above embodiments but by the scope of claims, and is intended to include all modifications and variations within the meaning and scope equivalent to the scope of claims.

  The present invention is suitable as a high voltage semiconductor device suitable for parallel operation, in particular as a semiconductor device comprising an IGBT.

1 n ^- drift layer, 1 a gate groove, 1 b emitter groove, 2 p type body region, 3 n type emitter region or n type impurity diffusion region, 4, 4 a gate insulating film, 4 b emitter insulating film, 5 conductive layer , 5a gate electrode, 5b emitter conductive layer, 6 p ⁺ impurity diffusion region, 7 n type buffer region, 7 an n type intermediate layer, 8 p type collector region, 9, 22A, 22B insulating film, 9a contact hole, 10 barrier Metal layer, 11 emitter electrode, 11a gate electrode wiring, 12, 12a collector electrode, 14, 14a n-type impurity diffusion region, 15 passivation film, 21a, 21b silicide layer, 28 gate pad, 28a resistor, 31 mask layer, 32 , 33 silicon oxide film, 32a sacrificial oxide film, 41 p-type impurity diffusion region.

Claims (3)

A semiconductor substrate having a first main surface and a second main surface facing each other;
A gate electrode (5a) formed on the first main surface side, a first electrode (11) formed on the first main surface side, and a second electrode formed in contact with the second main surface And an element having (12)
The device generates an electric field in a channel by a voltage applied to the gate electrode, and controls a current between the first electrode and the second electrode by the electric field of the channel;
A first emitter diffusion layer (6) of the first conductivity type formed on the first main surface and in contact with the first electrode (11);
And a second emitter diffusion layer (3) of a second conductivity type formed on the first main surface and in contact with the first electrode and the first emitter diffusion layer.
And a collector region formed on the second main surface,
The collector region is a collector diffusion layer (8) of a first conductivity type in contact with the second electrode (12), and a buffer of a second conductivity type formed closer to the first main surface than the collector diffusion layer A diffusion layer (7) and a drift diffusion layer (1) of the second conductivity type, the drift diffusion layer having an impurity concentration lower than that of the buffer diffusion layer, and adjacent to the buffer diffusion layer It is formed closer to the first major surface than the buffer diffusion layer,
A first diffusion type body diffusion layer (2) to be the channel;
And a buried diffusion layer (14, 14a) of the second conductivity type formed between the body diffusion layer and the drift diffusion layer (1).
A gate groove (1a) and an emitter groove (1b) are formed in the first main surface of the semiconductor substrate, and the gate electrode (5a) is embedded in the gate groove (1a). And a conductive layer (5b) serving as an emitter potential is embedded in the emitter groove (1b),
The buried diffusion layer (14) is formed in contact with the gate groove and not in contact with the emitter groove.
The width of said first emitter diffusion layer along the extending direction of the gate electrode (5a) (W _PC) and the second width of the emitter diffusion layer along the extending direction of the gate electrode (W _SO) and the The semiconductor device, wherein the ratio (W _SO / W _SO + W _PC ) of the width (W _SO ) of the second emitter diffusion layer to the sum is 0.08 or more and 0.20 or less. Before Kido atoms per unit area of the impurity constituting the lift diffusion layers (S _N-) atoms per unit area of the impurity constituting the buffer diffusion layer to (S _N) ratio of 0.05 over 100 The semiconductor device according to claim 1, which is the following. A semiconductor substrate having a first main surface and a second main surface facing each other;
A gate electrode (5a) formed on the first main surface side, a first electrode (11) formed on the first main surface side, and a second electrode formed in contact with the second main surface And an element having (12)
The device generates an electric field in a channel by a voltage applied to the gate electrode, and controls a current between the first electrode and the second electrode by the electric field of the channel;
And a collector region formed on the second main surface,
The collector region is a collector diffusion layer (8) of a first conductivity type in contact with the second electrode (12), and a buffer of a second conductivity type formed closer to the first main surface than the collector diffusion layer A diffusion layer (7) and a drift diffusion layer (1) of the second conductivity type, the drift diffusion layer having an impurity concentration lower than that of the buffer diffusion layer, and adjacent to the buffer diffusion layer It is formed closer to the first major surface than the buffer diffusion layer,
The ratio of the number of atoms per unit area (S _N ) of the impurities constituting the buffer diffusion layer to the number of atoms per unit area (S _N− ) of the impurities constituting the drift diffusion layer is 0.05 or more and 100 or less Yes,
A first diffusion type body diffusion layer (2) to be the channel ;
And a buried diffusion layer (14, 14a) of the second conductivity type formed between the body diffusion layer and the drift diffusion layer (1) .
A gate groove (1a) and an emitter groove (1b) are formed in the first main surface of the semiconductor substrate, and the gate electrode (5a) is embedded in the gate groove (1a). And a conductive layer (5b) serving as an emitter potential is embedded in the emitter groove (1b),
The buried diffusion layer (14) is formed in contact with the gate groove and not in contact with the emitter groove.
The semiconductor device according to claim 1, wherein a protrusion amount (D _T ) of the gate groove from the bottom of the body diffusion layer is 1.0 μm or more and a depth reaching the second main surface.

Images