JP2002185017A - High breakdown voltage high speed diode - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a diode which is connected reverse parallel with a power semiconductor switching element and can turn on quickly, in order to protect breakdown of the switching element due to a sharp inverse current flowing successively after a forward current flows, when the switching element is turned off and pulse type large current is made to flow in the forward direction within a very short time. SOLUTION: A conductivity type cathode region 62 having impurity concentration higher than that of a semiconductor substrate is formed on one main surface of a conductivity type or an intrinsic semiconductor substrate 61, in such a manner that a sinusoidal uneven structure is formed on the surface opposite to a surface in contact with a cathode electrode 64. An opposite conductivity type anode region 63 is formed on the other main surface of the substrate, in such a manner that a sinusoidal uneven structure is formed on the surface opposite to a surface in contact with an anode electrode 65. The pitch of the uneven structure is made shorter than the diffusion length of carrier in the substrate. A protruding part of one uneven structure faces almost a recessed part of the other uneven structure.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a high-voltage, high-speed diode, and more particularly, to an extremely short time of 1 .mu.s or less.
In a laser device or the like that needs to supply a large current of several thousand A or more in a pulsed form, a power semiconductor switching element capable of applying a high voltage to a discharge electrode steeply and allowing a very small current to flow due to discharge. The present invention relates to a diode which is preferably used as a reverse conducting diode connected in antiparallel with a power semiconductor switching element for backflow protection in a used pulse power circuit.

[0002]

2. Description of the Related Art For example, a power supply device for outputting a pulse of a high voltage and a large current is used as a power supply for a pulse laser or a power supply for a pulse discharge device. FIG. 1 shows an example of a pulse generation circuit known as a power supply for a pulse laser. In this pulse generation circuit, the DC power supply 11
Charger 1 having switch 12 and current limiting resistor 13
4, an electrostatic induction thyristor 15 as a power semiconductor switching element is connected between the output terminals 14a and 14b, and the resonance coil 1 is connected in parallel with the electrostatic induction thyristor.
6 and a capacitor 17 connected in series. Usually, the coil 16 is extremely small, such as the wiring inductance. In parallel with this capacitor 17,
A capacitor 18 and a coil 19 having a large inductance value are connected in series, and a discharge gap 20 serving as a load is connected in parallel with the inductance.

First, the switch 12 is closed with the electrostatic induction thyristor 15 in a non-conductive state, and the capacitor 17 is charged via the resistor 13 and the coil 16. During this charging process, the capacitor 18 is also charged via this coil 19 since the impedance of the coil 19 is low at low frequencies. Now, assuming that the voltage of the DC power supply 11 is E, after the capacitors 17 and 18 are both charged to E, the electrostatic induction thyristor 15 is turned on by the gate drive circuit 21 connected to the gate of the electrostatic induction thyristor 15. . At this time, the electric charge of the capacitor 17 is
6 and the capacitor 17 discharges through the electrostatic induction thyristor 15 in accordance with the resonance characteristics determined by the capacitor 17 and the capacitor 1
7 is charged to approximately -E to the opposite polarity to the polarity during charging. On the other hand, the electric charge of the capacitor 18 is also discharged through the electrostatic induction thyristor 15 and the coil 19, but the discharge of the coil is very slow because the impedance of the coil is very high for high frequencies. Therefore, a voltage of about -2E is applied between the discharge gaps 20. If discharge occurs here, capacitors 17 and 1
The charge of 8 is discharged and disappears. The switch 12 is closed again to start charging.

In the above-described pulse generating circuit, if discharge is performed properly when a voltage of -2E is applied between the discharge gaps 20, the coil 16 and the capacitor 17
Since the charge accumulated in the resonance circuit composed of the static induction thyristor 15 disappears as shown by the solid line in FIG.
No current flows in the reverse direction, but if discharge does not occur properly for some reason, a ringing current flows in the resonance circuit, and flows in the reverse direction through the electrostatic induction thyristor 15 as shown by a broken line in FIG. A large current will flow. FIG. 3 shows a change in the voltage between the anode and the cathode of the electrostatic induction thyristor 15. In the event of a discharge failure, a reverse voltage is applied. In particular, a reverse current flows from the cathode of the electrostatic induction thyristor 15 to the gate. This is the same as the reverse recovery phenomenon of the diode, and an excessive reverse voltage is applied between the gate and the cathode.

As described above, when a large reverse current flows between the anode and the cathode of the electrostatic induction thyristor 15 as an example of the switching element, the electrostatic induction thyristor is anti-parallel to the electrostatic induction thyristor in order to protect it from destruction. It has been proposed to connect a diode to this and to pass a reverse current through this diode. An electrostatic induction thyristor having such a diode is generally called a reverse conducting electrostatic induction thyristor. In this reverse conduction type electrostatic induction thyristor, in order to reduce the wiring inductance as much as possible, an anti-parallel diode may be integrated into a common semiconductor substrate with the electrostatic induction thyristor.
Shimizu et al. Proposes a "4000 V class reverse conducting SI thyristor (1)" on pages 4-76 to 4-77 of the proceedings of the Annual Meeting of the Institute of Electrical Engineers of Japan.

FIG. 4 is an equivalent circuit diagram of the above-mentioned reverse conduction type electrostatic induction thyristor. The cathode is connected to the anode of the thyristor in parallel with the electrostatic induction thyristor 31, and the anode is connected to the cathode of the electrostatic induction thyristor. The diode 32 is connected so as to be connected. This diode 3
2 is connected to a static induction thyristor 3 via a resistor 33.
The gate drive circuit (GC) 34 for controlling turn-on / turn-off of the electrostatic induction thyristor is connected to this gate. When the main power supply 35 indicated by the solid line between the anode and cathode of the static induction thyristor 31 is connected, a current I _T flows through the static induction thyristor, reverse polarity of the power supply 36 as shown by a broken line static induction thyristor current I _R flows through the diode 32 is protected from being destroyed when connected.

FIG. 5 is a sectional view showing the structure of the above-described reverse conducting electrostatic induction thyristor. n ^- type silicon substrate 4
A p ⁺ -type gate region 42 is formed on one surface of the substrate 1, and a buried gate region 43 is formed in the channel region. A gate electrode 45 is formed so as to be in contact with the gate region 42. Buried gate region 4
3 is formed in a comb shape so as to be surrounded by the gate region 42. An n ^{+ -type} cathode region 46 is formed above the channel region. The cathode region is connected to a cathode electrode 47 via a conductive layer, and a thyristor portion 44 is formed. A diode 49 is formed outside the thyristor section 44 via a separation band 48. The diode section 49 is formed of the p ⁺ type anode region 5.
0, and a cathode region constituted by a part 41a of the silicon substrate 41, the anode region 50 is connected to the cathode electrode 47 of the electrostatic induction thyristor via a conductive layer, and the cathode region 41a is n ⁺ It is connected to the anode electrode 52 of the electrostatic induction thyristor via the contact region 51 of the mold and the conductive layer.

[0008]

In such a reverse conduction type electrostatic induction thyristor, when a reverse voltage is applied between the anode and the cathode, the diode section 49 conducts and the thyristor section 44 is broken. The aim is to prevent the effect. However, when the reverse conducting electrostatic induction thyristor is applied to the pulse generation circuit shown in FIG. 1 described above, when a discharge error occurs, the electrostatic induction thyristor section is often destroyed by the ringing current in the resonance circuit as described above. There is a problem. In some circuits, it is necessary to supply a reverse current to the switch during normal operation. In order to clarify the mechanism by which such a problem occurs due to the reverse current, the details of how the thyristor is affected when a reverse current flows between the anode and the cathode of the reverse conduction type electrostatic induction thyristor will be described in more detail. Was considered.

[0009] Figures 6, 7 and 8, current I _ak flowing between the anode and the cathode when is pulsed involves reverse current reverse conduction type static induction thyristor, a change in the gate current I _g and the gate voltage V _g In these figures, A shows the case where the pulse width _tw is increased, and B shows the case where the pulse width is shortened. When the current _Iak is 3000A or more,
When the pulse width t _w is as long as several tens of μs or more is the reverse conducting static induction thyristor does not break, when the pulse width t _w shortened several hundred ns~ number μs is reverse conducting static induction thyristor breakdown I confirmed that there was a phenomenon that would do. Further, the destruction point is located in the electrostatic induction thyristor portion, and the feature is that no abnormality occurs in the diode portion. From this, it is inferred that whether or not the reverse conducting electrostatic induction thyristor is broken depends on the gradient of the falling portion of the current _Iak . The gradient for the long pulse width in FIG. 6A is, for example, 0.5 KA / μs, and the gradient for the short pulse width in FIG. 6B is, for example, 3 KA / μs. FIG.
As shown in B, when breakdown occurs, the gate voltage V _g
A remarkable fluctuation is observed near the peak of the reverse voltage.

Next, when a steep current was to flow through the diode, the difficulty of flowing the current was examined. 9 and 10
A forward current I _F upon applying a pulsed steep current between the anode and the cathode of the diode 32 shown in FIG. 4
And while indicating the forward voltage drop V _F, A case when the rise of the slope of the forward current is small, B shows a case where a large rise in the gradient of the forward current. Thus a forward current close correlation as shown in FIG. 11 between the rising slope and the transient ON voltage (forward recovery voltage) V _FP of I _F. That is, the withstand voltage is 40
For 00V diodes, when the rise of the gradient _dI F / dt in the forward current _{I F} is 500A / .mu.s and small, forward recovery voltage _{V FP} is substantially 70V and the lower, 1000
At A / μs, the voltage is almost 100 V, and when it is as large as 2000 A / μs, it is as high as about 170 V.

[0011] FIG. 12 is for the rise of the gradient _dI F / dt in the forward current _{I F} indicates the relationship between the breakdown voltage of the forward recovery voltage _{V FP} and diode 2000A / .mu.s at the diode, the diode the breakdown voltage has become the forward recovery voltage _{V FP} enough to become higher and higher, forward recovery voltage _{V FP} is a breakdown voltage of 4000V diodes are reached as high as 170V. In the reverse conduction type electrostatic induction thyristor, the withstand voltage of the diode portion arranged in parallel with the thyristor needs to be at least equal to the withstand voltage of the thyristor portion.
A diode having a withstand voltage of several thousand volts is also used as the diode part. Thus, the forward recovery voltage _VFP of the diode part having a high withstand voltage is high. That is, as the withstand voltage of the diode section is higher, a steep forward pulse current is more difficult to flow through the diode section.

As described above, in a conventional high withstand voltage reverse conduction type electrostatic induction thyristor having a withstand voltage of, for example, 4 KV, when a large reverse current suddenly flows after a forward current flows through the thyristor portion, a protective diode portion is provided. Cannot be conducted, and FIG.
Carriers accumulated in the channel region 44 flow rapidly from the cathode region 46 toward the gate region 43 in the opposite direction. In particular, the carrier becomes particularly large in the vicinity of the gate region 42 to which the gate current is supplied most steeply than the center gate region 45, and filamentation occurs between the channels due to the diode reverse recovery phenomenon between the gate and the cathode. Therefore, there is a problem that the thyristor portion 44 is broken.

Such a problem occurs not only in the reverse conduction type electrostatic induction thyristor but also in a switching element such as a normal thyristor, a gate turn-off (GTO) thyristor, and an insulated gate bipolar transistor (IGBT). is there.

In a pulse circuit in which a parallel circuit of a coil L and a capacitor C is cascaded as shown in FIG. 13, a diode D connected in series with a resistor R is used to prevent a reverse current. However, this diode is also required to conduct at high speed in the forward direction and have a high withstand voltage. However, conventional diodes do not fully satisfy such requirements.

An object of the present invention is to solve the above-mentioned problems of the conventional diode and to provide a diode having high withstand voltage characteristics and capable of conducting at high speed. In particular, the present invention is to turn on the power semiconductor switching element, to forward a large pulsed current in a very short time in a very short time, the power semiconductor switching element by a steep reverse current that flows after the forward current flow It is an object of the present invention to provide a diode suitable for use as a diode connected in anti-parallel with a power semiconductor switching element to protect against breakdown.

[0016]

According to the present invention, there is provided a high-withstand-voltage high-speed diode according to the present invention, wherein a first conductive type or a first conductive type having an impurity concentration higher than that of a semiconductor substrate is provided on one main surface of an intrinsic semiconductor substrate. Is formed so as to have a concavo-convex structure on a surface opposite to a surface in contact with one main electrode, and a first main electrode region of the opposite conductivity type is formed on the other main surface of the semiconductor substrate. The pitch of the uneven structure of the first and second main electrode regions is determined by the diffusion length of the carrier in the semiconductor substrate. And one of the protrusions is substantially opposed to the other of the recesses.

In such a diode according to the present invention, the first and second main electrode regions of the one conductivity type are
The first and second main electrodes can be formed so as not to form a Schottky junction, in other words, so as to form an ohmic junction.
And the impurity concentration of the surface of the second main electrode region in contact with the first and second main electrodes is set to approximately 1 × 10 ¹⁷ atoms / c.
It can be m ³ or more.

In a preferred embodiment of the diode according to the present invention, the concavo-convex structure of the first main electrode region of one conductivity type is formed in a sine wave shape, and the peak of one main electrode region is formed in the other main electrode region. It can be formed to be substantially aligned with the valley. Such a sinusoidal uneven structure can be formed by selective diffusion. Further, the above-described uneven structure can be formed into a triangular wave shape, and such a triangular wave uneven structure can be formed by anisotropic etching.

Further, in the diode according to the present invention, the semiconductor substrate is formed of a semiconductor substrate having an impurity concentration of approximately 1 × 10 ¹³ atoms / cm ³ , and is in contact with one main electrode of the first main electrode region. The impurity concentration on the surface is approximately 1 × 10 ²⁰ atoms / cm ^3, and the impurity concentration on the surface of the second main electrode region that contacts the other main electrode is approximately 1 × 10 ²⁰ atoms / cm ^3.
It is preferable to set to × 10 ¹⁸ atoms / cm ³ .

Further, in the diode according to the present invention, the semiconductor substrate is an intrinsic semiconductor substrate, and has a first conductivity type and a first main type so as to be in contact with the surface of the first main electrode region on which the uneven structure is formed. A buffer layer having an impurity concentration lower than that of the electrode region can be formed. When such a buffer layer is provided, the thickness of the intrinsic semiconductor substrate can be reduced without lowering the breakdown voltage, and as a result, the on-voltage can be reduced.

[0021]

FIG. 14 is a sectional view showing the structure of an embodiment of a high withstand voltage and high speed diode according to the present invention. n ^-
An n ⁺ -type cathode region 62 is formed on one main surface of a silicon semiconductor substrate 61 and a p ⁺ -type anode region 63 is formed on the other main surface of the semiconductor substrate. n
^A cathode electrode 64 made of aluminum is formed on the surface of the ⁺ type cathode region 62 and ap ⁺ type anode region 6 is formed.
On the surface of No. 3, an anode electrode 65 made of aluminum is formed.

In the present invention, the surface of the n ⁺ -type cathode region 62 opposite to the surface in contact with the cathode electrode 64, that is, the surface facing the bulk of the semiconductor substrate 61 is formed in a wavy shape. The surface of the p ⁺ type anode region 63 opposite to the surface in contact with the anode electrode 65, that is, the surface facing the bulk of the semiconductor substrate is formed in a wavy shape. In the present example, the shapes of the wavy surfaces of the cathode region 62 and the anode region 63 are formed into a sine wave shape that smoothly changes, and the peak of the wavy shape of the cathode region 62 viewed from the cathode electrode 64 side is viewed from the anode electrode 65 side. The anode region 63 is configured so as to face the wavy valley of the anode region 63, and the cathode region 62 is configured to face the wavy valley of the anode region 63.

Therefore, the pitches of the corrugated structures of the cathode region 62 and the anode region 63 are equal to each other. In the present invention, the pitch of the wavy structure is formed to be shorter than the carrier diffusion length of the bulk portion of the n ⁻ type silicon semiconductor substrate 61. The diffusion length of the semiconductor substrate 61 is represented by (Dτ) ^1/2 when the impurity concentration is D and the average carrier lifetime is τ. Since the carrier diffusion length of the bulk portion of the n ⁻ type silicon semiconductor substrate 61 is, for example, about 70 μm, the pitch of the wavy structure of the cathode region 62 and the anode region 63 is shorter than 70 μm, and is about 50 μm in this example. I do.

Further, in this embodiment, the anode region 6
2, the n-type impurity concentration on the surface in contact with the anode electrode 64 is approximately 1 × 10 ²⁰ atoms / cm ^3, and the p-type impurity concentration on the surface of the cathode region 63 in contact with the cathode electrode 65 is approximately 1 × 10 ¹⁸ Atoms / cm ³ . As described above, by increasing the impurity concentration on the surfaces of the anode region 62 and the cathode region 63, these regions do not form a Schottky junction with the electrodes but form an ohmic junction. In order to form an ohmic junction in this manner, the impurity concentration on the surfaces of the anode region 62 and the cathode region 63 needs to be approximately 1 × 10 ¹⁷ atoms / cm ³ or more.

Further, in this embodiment, the cathode region 6
2, the distance from the surface in contact with the cathode electrode 64 to the peak is approximately 10 μm, and the distance to the valley is approximately 1-2 μm. On the other hand, the distance from the surface of the anode region 63 in contact with the anode electrode 65 to the valley is approximately 15 μm, and the distance to the peak is approximately 1 to 5 μm.

Next, the operation of the diode of this embodiment will be described with reference to FIG. In FIG. 15, the operation of the conventional diode is also shown in comparison, but the semiconductor substrate, the cathode region, the anode region, the cathode electrode, and the anode electrode are indicated by reference numerals obtained by adding 100 to the reference numerals shown in FIG. In the conventional diode, the surfaces of the cathode region 162 and the anode region 163 that are in contact with the bulk of the semiconductor substrate 161 do not have a wavy structure, but are flat. In FIG. 15, black circles represent electrons, and white circles represent holes.

FIG. 15A shows a state where no forward voltage is applied between the anode and the cathode. In the conventional diode, in this state, thermal diffusion of carriers occurs at the junction interface between the p ⁺ -type anode region 163 and the n ⁻ -type semiconductor substrate 161, that is, the diffusion of holes from the p-type region to the n-type region. In addition, a potential is generated in such a direction as to block the diffusion of electrons from the n-type region to the p-type region. This potential is called a diffusion potential or a built-in voltage. In other words, at the junction interface between the p ⁺ type anode region 163 and the n ⁻ type semiconductor substrate 161, the p emitter region side formed of the anode region is charged to +, and the n ⁻ type semiconductor substrate 161 side is charged to −. Thus, the built-in voltage Eb from the n ⁻ type semiconductor substrate 161 at the junction interface toward the p emitter region
Will occur. A similar built-in voltage Eb is n
^- type semiconductor substrate 161 and the ^{n +} -type cathode region 162
Also appears at the junction interface with That is, the semiconductor substrate side is-
And the n emitter region side comprising the cathode region is +
And a built-in voltage Eb is generated in the direction from the cathode region toward the semiconductor substrate.

FIG. 15B shows a state immediately after a forward bias is applied in a state where a built-in voltage is generated as described above. At this time, holes injected from the anode region 163 are prevented from being injected by the built-in voltage Eb generated at the bonding interface with the semiconductor substrate 161, and electrons injected from the cathode region 162 are injected into the bonding interface between the cathode region and the semiconductor substrate. The injection is hindered by the generated built-in voltage Eb, the movement of holes from the anode region to the cathode region and the movement of electrons from the cathode region to the anode region are significantly restricted, and a forward current hardly flows through the diode. in this way,
In the conventional diode, the movement of carriers is suppressed due to the built-in voltage, and high-speed turn-on cannot be performed.

FIG. 15C shows a state when, for example, about 1 μs elapses after the application of the forward voltage, and occurs at the respective junction interfaces between the semiconductor substrate 161 and the cathode region 162 and the anode region 163. The built-in voltage is neutralized by carriers injected from the cathode region and the anode region, respectively, so that holes move from the anode region 163 to the cathode region 162 and electrons move from the cathode region 162 to the anode region 163. This shows a state in which the movement is started and a forward current is flowing.

On the other hand, in the diode according to the present invention, when in the off state, as shown in FIG.
At the junction interface between the semiconductor substrate 61 and the anode region 63 as in the conventional diode, the semiconductor substrate side is charged to +,
The anode region side is negatively charged, and at the bonding interface between the semiconductor substrate 61 and the cathode region 62, the semiconductor substrate side is negatively charged.
, And the cathode region side is charged to +, so that a built-in voltage of about 0.7 eV as shown by an arrow is generated. However, in the present invention, these junction interfaces are not flat as in the prior art, but have a wavy structure, so that the built-in voltage is not uniform, and the charging at the valleys becomes strong, so that the bonding interface is large. Built-in voltage occurs. On the other hand, the charging at the peak portion is weak, so that the built-in voltage is low.

Next, in the state immediately after the application of the forward voltage shown in FIG. 15B, in the valley portion of the wavy structure where the strong charge is generated, similarly to the conventional diode, the valley is injected from the p emitter region composed of the anode region 63. The holes neutralize the negative charge on the semiconductor substrate 61 side, and the electrons injected from the n-emitter composed of the cathode region 62 neutralize the + charge on the semiconductor substrate side at the junction interface. Therefore, the movement of holes from the anode region to the cathode region and the movement of electrons from the cathode region to the anode region are limited at the beginning of energization until the charge is neutralized, that is, until the built-in voltage decreases. You. However, there is little charge at the slope from the valley to the peak of the wavy structure, and carriers injected from the emitter on the opposite side reach the valley via the slope. Further, since the junction is selectively shallow, the emitter injection efficiency from the peak is selectively increased.
Here, the carrier injection efficiency α is proportional to the peak value, the depth and the gradient of the impurity concentration at the junction, but in the diode according to the present invention, the junction area of the cathode region 62 and the anode region 63 in contact with the semiconductor substrate 61. In the sloping portion from the valley to the peak of the wavy structure, the peak value of the impurity concentration is large, the depth is deep, and the gradient of the impurity concentration is large, so that the carrier injection efficiency α is large, and as a result, the cathode region 62 , A large amount of holes are injected from the anode region 63 and a large forward current flows.

Further, from the start of conduction shown in FIG.
At the time when 1 μs has elapsed, a large amount of electrons are injected from the inclined portion and the mountain of the cathode region 62, and a large amount of holes are injected from the inclined portion and the mountain of the anode region 63. A region is generated, a current multiplication effect appears, so-called conductivity modulation occurs, and a large amount of current flows.

As described above, in the conventional diode, after the conduction starts, the built-in voltage of about 0.5 to 0.7 eV at the junction interface between the cathode region 162 and the anode region 163 and the bulk of the semiconductor substrate 161 causes
Although the injection of electrons and holes is prevented and the forward current does not flow rapidly, in the diode according to the present invention, the cathode region 62 and the anode region 63 and the semiconductor substrate 6
Since the interface with the bulk of No. 1 has a wavy structure, the carrier injection efficiency is selectively improved, the forward current rises sharply, and a large amount of electrons and holes are injected from the cathode region and the anode region. As a result, a carrier-rich region is generated in the bulk of the semiconductor substrate 61, so that a current multiplication action by conductivity modulation is performed, and a larger amount of current flows.

FIG. 16 shows the gradient of the forward current dI _F / dt on the horizontal axis.
Taking [A / μs] and taking the forward voltage _VFP on the vertical axis, curve A shows the characteristics of the diode according to the present invention, and curve B shows the characteristics of the conventional diode. Both diodes have a breakdown voltage of 4000 V. The height of the peak of the cathode region 62 of the diode of the present invention is equal to the thickness of the cathode region 162 of the conventional diode, and the height of the peak of the anode region 63 of the diode of the present invention is equal. Is equal to the film thickness of the anode region 163 of the conventional diode. The instantaneous voltage generated at the time when a current having a high current rise rate (di / dt) flows is referred to as a forward recovery voltage (V _FP ). di / dt
When a current of 2000 A / μs flows, a forward recovery voltage of about 110 V is generated, whereas a conventional diode having a flat anode region and a cathode region generates a high forward recovery voltage of approximately 160 V when the same current flows. I do.

Further, as shown in FIG. 4, the diode of this embodiment is connected in anti-parallel with an electrostatic induction thyristor, and has a pulse width of 1 μm.
s, a pulse current of 3000 A, and a voltage of 3000 V, the device was used in a situation where a reverse current of 2500 A flows due to an abnormal operation. As a result, the breakdown rate of the electrostatic induction thyristor was almost 15%. On the other hand, when the conventional diode was used in the same situation, the destruction rate of the electrostatic induction thyristor was 100%.

FIG. 18 is a sectional view showing the structure of another embodiment of the diode according to the present invention. In this example, the silicon semiconductor substrate 61 is formed of an intrinsic semiconductor, and the surface of the n ⁺ -type cathode region 62 on which the wavy structure is formed and the semiconductor substrate 61 are formed.
And an n-type buffer layer 66 is formed between the substrate and the bulk. The n-type impurity concentration of this buffer layer 66 is approximately 1
It can be set to × 10 ^{16 to} 5 × 10 ¹⁷ atoms / cm ³ . By providing such a buffer layer 66,
The bulk thickness of the semiconductor substrate 61 made of an intrinsic semiconductor can be reduced, and the on-resistance can be reduced.

FIG. 19 is a sectional view showing still another embodiment of the diode according to the present invention. In this example, the distance between the interface between the cathode electrode 64 and the cathode region 62 and the valley of the cathode region 62 having a wavy structure is set to approximately 0.5 μm.
m and the anode electrode 65 and the anode region 6
Also, the distance from the interface with No. 3 to the valley of the anode region 63 having the wavy structure is shortened to approximately 0.5 μm.
With this configuration, there is an effect that the reverse recovery characteristic is improved and the off characteristic is also improved.

The present invention is not limited to the above-described embodiment, and many modifications and variations are possible. For example, in the above-described embodiment, the semiconductor substrate 61 is formed of silicon, but may be formed of another semiconductor material such as silicon carbide or gallium arsenide. Also,
The impurity concentration of each region described above is shown as an example, and it is a matter of course that the present invention is not limited to such numerical values.

Further, in the above-described embodiment, the wave-like structures of the cathode region and the anode region are formed in a sine wave shape that changes smoothly, but may be formed so as to change in a triangular wave shape as shown in FIG. As shown in FIG. 20, a structure in which a peak of a triangular wave and a trapezoidal valley are combined may be employed. For example, a sinusoidal wavy structure can be formed by selective diffusion, and a triangular wavy structure can be formed by anisotropic etching.

As described above, according to the diode of the present invention, for example, by being connected in anti-parallel with the electrostatic induction thyristor, the destruction rate of the electrostatic induction thyristor can be significantly reduced. Further, it can be effectively used as a high breakdown voltage diode in other pulse circuits and laser power supply circuits.

[Brief description of the drawings]

FIG. 1 is a circuit diagram showing a configuration of an example of a pulse generation circuit using an electrostatic induction thyristor.

FIG. 2 is a signal waveform diagram for explaining the operation.

FIG. 3 is a signal waveform diagram for explaining the operation.

FIG. 4 is a diagram showing a basic structure of a conventional reverse conduction type electrostatic induction thyristor.

FIG. 5 is a sectional view showing a detailed structure of a conventional reverse conducting electrostatic induction thyristor.

6A and 6B are signal waveform diagrams for explaining the operation of a conventional electrostatic induction thyristor.

7A and 7B are signal waveform diagrams for explaining the operation of a conventional electrostatic induction thyristor.

8A and 8B are signal waveform diagrams for explaining the operation of a conventional electrostatic induction thyristor.

9A and 9B are signal waveform diagrams showing forward recovery characteristics of a protection diode of a conventional electrostatic induction thyristor.

FIGS. 10A and 10B are signal waveform diagrams showing a transient on-voltage of a protection diode of a conventional electrostatic induction thyristor.

FIG. 11 is a graph showing a relationship between a current gradient and a forward recovery voltage.

FIG. 12 is a graph showing a relationship between a withstand voltage of a diode and a forward recovery voltage.

FIG. 13 is a circuit diagram showing a configuration of a pulse circuit to which a diode according to the present invention can be applied.

FIG. 14 is a sectional view showing a configuration of an embodiment of a diode according to the present invention.

FIGS. 15A, 15B and 15C are diagrams showing the operation of a diode according to the present invention in comparison with the operation of a conventional diode.

FIG. 16 is a graph showing characteristics of a diode according to the present invention and a conventional diode in comparison.

FIG. 17 is a sectional view showing a configuration of another embodiment of the diode according to the present invention.

FIG. 18 is a sectional view showing the configuration of still another embodiment of the diode according to the present invention.

FIG. 19 is a sectional view showing another example of a wavy structure in the diode according to the present invention.

FIG. 20 is a cross-sectional view showing still another example of the wavy structure of the diode according to the present invention.

[Explanation of symbols]

61 semiconductor substrate, 62 cathode region, 63 anode region, 64 cathode electrode, 65 cathode electrode, 66 buffer layer

Claims (6)

[Claims] A first main electrode region of one conductivity type having an impurity concentration higher than that of a semiconductor substrate is brought into contact with one main electrode on one main surface of one conductivity type or intrinsic semiconductor substrate. A first main electrode region of the opposite conductivity type is formed on the other main surface of the semiconductor substrate so as to have a concavo-convex structure on the surface opposite to the surface, and the first main electrode region is opposite to the surface in contact with the other main electrode. Are formed so as to have an uneven structure on the surface thereof, and the pitch of the uneven structure of the first and second main electrode regions is
A high-withstand-voltage high-speed diode characterized in that the length is shorter than the diffusion length of carriers in a semiconductor substrate and one convex portion is substantially opposed to the other concave portion. 2. The device according to claim 1, wherein said first and second main electrode regions of one conductivity type are formed so as to form an ohmic junction with said first and second main electrodes. High withstand voltage high speed diode. 3. The first and second main electrode regions, wherein:
^3. The high withstand voltage and high speed diode according to claim 2, wherein the impurity concentration on the surface in contact with the second main electrode is approximately 1 × 10 ¹⁷ atoms / cm ³ or more. 4. The semiconductor substrate according to claim 1, wherein said semiconductor substrate is an intrinsic semiconductor substrate, and is of a first conductivity type and in contact with a surface of said first main electrode region on which an uneven structure is formed, based on an impurity concentration of said first main electrode region. 4. The high-breakdown-voltage high-speed diode according to claim 1, wherein a buffer layer having a low impurity concentration is formed. 5. The rugged structure of said first and second main electrode regions is formed in a sine wave shape.
A high-voltage high-speed diode according to any one of the above. 6. A structure according to claim 1, wherein said first and second main electrode regions have a triangular waveform.
A high-voltage high-speed diode according to any one of the above.