JPH0652787B2 - Shutter-barrier barrier semiconductor device - Google Patents

Description

The present invention relates to a high breakdown voltage shutter barrier semiconductor device.

[Problems to be Solved by Prior Art and Invention]

The shutter barrier diode is widely used in a high-frequency rectifier circuit and the like by taking advantage of good high-speed response (high-speed switching characteristic) and low loss. But,
The breakdown voltage of the shutter barrier barrier diode (the breakdown voltage around the shutter barrier) is significantly lower than the bulk breakdown voltage (the breakdown voltage at the center of the shutter barrier).
There is a problem that it is difficult to increase the breakdown voltage.

Providing a field plate or a guard ring in order to solve this problem is known, for example, in the second edition of "Fixix of Semiconductor Devices" by S.M.J.E., USA. Also,
The use of both field plates and guard rings has already been done.

A shutter plate barrier diode having a field plate structure is composed of an n ^{+ type} semiconductor region, an n type semiconductor region formed on the n ^{+ type} semiconductor region, and a metal electrode capable of forming a shutter barrier formed on the n type semiconductor region (hereinafter referred to as a barrier electrode). Call)
An insulating layer formed on the n-type semiconductor region so as to surround the barrier electrode, a field plate provided on the insulating layer and connected to the barrier electrode, and an ohmic contact connected to the n ^{+ type} semiconductor region. It consists of electrodes. When a reverse voltage is applied between the barrier electrode and the ohmic electrode, a depletion layer is generated between the barrier electrode and the n-type semiconductor region, and
A depletion layer is also generated in the n-type semiconductor region below the field plate due to the field effect of the field plate, and the concentration of the electric field in the semiconductor region near the periphery of the barrier electrode is relieved, and the breakdown voltage around the Schottky barrier is improved. . However, it was practically difficult to satisfactorily reduce the concentration of the electric field and significantly improve the breakdown voltage.

On the other hand, a Schottky barrier diode having a guard ring structure has a guard ring which is connected to the periphery of the barrier electrode in plan view and is formed of ap ^{+ -type} semiconductor region arranged so as to surround the barrier electrode. The p ^{+ -type} semiconductor region of the guard ring forms a pn junction with the n-type semiconductor region, and when a reverse voltage is applied to this pn junction, the depletion layer spreads more effectively than around the shutter barrier. As a result, the breakdown voltage around the barrier electrode can be improved. However, since the structure is such that the Schottky barrier diode and the pn junction diode are arranged in parallel, when a forward voltage is applied and a forward current flows, a small number of carriers are injected at the pn junction portion, and High-speed response, which is one of the features, is reduced. Although the guard ring structure is widely used in combination with the field plate structure, it is still difficult to significantly increase the breakdown voltage.

As one method for solving the above-mentioned problem, as disclosed in Japanese Patent Publication No. 49-41463, there is a method in which a resistance layer having a taper and a Schottky barrier action is provided around the Schottky barrier electrode. is there. However, it is difficult to form the tapered layer.

Therefore, an object of the present invention is to provide a Schottky barrier semiconductor device which is easy to manufacture and can surely achieve improvement in breakdown voltage.

[Means for Solving the Problems] In order to achieve the above object, the present invention is directed to a semiconductor region formed on the semiconductor region so that a Schottky barrier can be generated between the semiconductor region and the semiconductor region. A first barrier electrode and a thickness that is arranged on the semiconductor region so as to surround the first barrier electrode, is electrically connected to the first barrier electrode, and is thinner than the first barrier electrode. And a sheet resistance higher than that of the first barrier electrode and a sheet resistance of 100 KΩ / □ or less, and a Schottky barrier can be generated between the first barrier electrode and the semiconductor region. Second formed on
Of the barrier electrode and the second barrier electrode are arranged on the semiconductor region so as to surround the second barrier electrode, and are electrically connected to the second barrier electrode, and more than the first and second barrier electrodes. Large sheet resistance and 10KΩ /
□ A thin layer having a sheet resistance of the above or more and being formed so as to generate a Schottky barrier with the semiconductor region, and having a thickness thinner than that of the first barrier electrode. The first barrier electrode has a first metal layer in contact with the semiconductor region and a second metal layer formed on the first metal layer, the second barrier A Schottky barrier semiconductor device, wherein the electrode is a layer containing the same metal as the first metal layer, and the thin layer is an oxide layer of the same metal as the first metal layer. Is.

The first barrier electrode is preferably composed of a titanium thin layer and a metal layer such as aluminum, the second barrier electrode is preferably composed of a titanium thin layer, and the thin layer surrounding the second barrier electrode is preferably composed of a titanium thin layer. It preferably comprises a thin layer of titanium oxide.

[Action]

In the above invention, when a reverse voltage is applied between the first barrier electrode and the semiconductor region, a depletion layer based on a shutter barrier between the first barrier electrode and the semiconductor region,
A depletion layer due to the shutter barrier between the second barrier electrode and the semiconductor region and a depletion layer due to the shutter barrier between the thin layer and the semiconductor region are generated. First and second
When there is a semiconductor region of single conductivity type under the barrier electrode and the thin layer, the above three depletion layers are directly continuous. On the other hand, when the semiconductor region includes guard rings of different conductivity types, the above three depletion layers are continuous through the depletion layer based on the pn junction. Since the second barrier electrode has a larger sheet resistance than the first barrier electrode, and the thin layer has a larger sheet resistance than the second barrier electrode, the thin film has a sheet resistance higher than that of the first barrier electrode. A potential gradient occurs in which the potential gradually changes. As a result, the depletion layer is spread so that the electric field is less likely to be concentrated in the semiconductor region near the periphery of the barrier electrode, and the breakdown voltage is significantly improved.

Note that, at the time of ultrahigh-speed reverse commutation, a relatively large current that transiently flows causes a gradually changing potential gradient in the second barrier electrode, which causes electric field concentration in the semiconductor region near the periphery of the first barrier electrode. It's hard to happen. In the steady state after the transient current has settled down, the potential of the second barrier electrode becomes almost constant, and instead, a gradually changing potential gradient occurs in the thin layer, and the potential near the periphery of the first and second barrier electrodes is generated. Electric field concentration is less likely to occur in the semiconductor region. In this way, the presence of the second barrier electrode reduces the breakdown voltage reduction during ultra-high speed operation, and a high breakdown voltage can be obtained from DC operation to ultra-high speed operation.

Further, since the second barrier electrode is interposed between the first barrier electrode and the thin layer, the electric field concentration point does not overlap with the stress concentration point below the peripheral portion of the first barrier electrode, and the high breakdown voltage is achieved. It is possible to further ensure the effect.

[First Embodiment] FIGS. 1 (A) to (F) and FIG. 2 to describe a shutter barrier diode and a method of manufacturing the same according to a first embodiment of the present invention.
It will be described with reference to FIG.

First, as shown in FIG. 1 (A), GaAs (gallium arsenide)
A semiconductor substrate 21 made of is prepared. Semiconductor substrate 21
Has a thickness of about 300 μm and an impurity concentration of 1 to 3 × 10 ¹⁸ cm
An n-type region 23 having a thickness of 10 to 20 μm and an impurity concentration of 1 to 2 × 10 ¹⁵ cm ⁻³ is epitaxially grown on the ⁻³ n _{+ type} region 22.

Next, as shown in FIG. 1 (B), a thin layer 24 of Ti (titanium), that is, a Ti thin film, is formed by vacuum evaporation on the entire upper surface of the n-type region 23 made of n-type GaAs, and further on the entire upper surface. The Al (aluminum) layer 25 is continuously vacuum-deposited. The Ti thin layer 24 has an extremely thin thickness of 50 to 200 Å (0.005 to 0.02 μm). The thickness of the Al layer 25 is about 2 μm, and the Ti thin layer 24
Is more than 100 times. Further, Au is formed on the lower surface of the n _{+ type} region 22.
The ohmic contact electrode 26 made of an alloy of (gold) -Ge (germanium) is formed by vacuum evaporation, and then 380
Heat treatment is performed at 10 ° C. for 10 seconds.

Next, as shown in FIG. 1 (C), a part of the Al layer 25 is removed by photo-etching, and Al is made to correspond to the region where the Schottky barrier that will be the main forward current path is to be formed.
The layer 25a remains. Further, the Ti thin layer 24 is removed from the peripheral region of the device by photo-etching, and the Ti thin layer 24a under the Al layer 25a and the Ti thin layer 24b surrounding the Ti thin layer 24a are left. Since the Ti thin layer 24b is an extremely thin film even if Ti itself is a conductor, it has a sheet resistance of 20 to 400.
It is a resistance layer of Ω / □, and has an order of magnitude higher resistance than the Al layer 25a.

Next, heat treatment is performed in air at 300 ° C. for 5 to 30 minutes. As a result, as shown in FIG. 1 (D), the Al layer 25a
The Ti thin layer 24b not covered with is oxidized to form a titanium oxide thin layer 28, but the Ti thin layer 24a below the Al layer 25a is not oxidized because it is masked by the Al layer 25a.

Next, the peripheral portion of the Al layer 25a is removed by photoetching to expose the unoxidized Ti thin layer 24a 'under the Al layer 25a as shown in FIG. 1 (E). As shown in FIG. 2, the Ti thin layer 24a 'is arranged so as to surround the outer periphery of the Al layer 25b, and the titanium oxide thin layer 28 is further arranged so as to surround the Ti thin layer 24a'. There is. Al and Ti
Since both of them are metals forming a shutter barrier with GaAs, the electrode composed of the Al layer 24a ″ is covered with the first barrier electrode 27a and the Ti thin layer 24a ′ surrounding the first barrier electrode 27a adjacently. Since the Ti thin layer 24a ″ is an extremely thin film, the Al layer 25b and the Ti thin layer 2 are formed in the first barrier electrode 27a.
It is not always clear how 4a "is involved in the formation of the shutter barrier. The Ti thin layer 24a" has a role other than the formation of the shutter barrier.
Contributes to improving the adhesion of the Al layer 25b to the n-type region 23. In addition, the Ti thin layer 2 forming the second barrier electrode 27b
4a 'is the first barrier electrode 27a and the titanium oxide thin layer 2
It contributes to the electrical connection with 8 and further to the high breakdown voltage as described later. Here, it is desirable that the sheet resistance of the first barrier electrode 27a, which is the main current path of the forward current, be 1 Ω / □ or less, and in this embodiment, it is about 0.05 Ω / □. As shown in FIG. 1 (E) and FIG. 2, the titanium oxide thin layer 28 according to the present invention, which is provided in a ring shape so as to surround the second barrier electrode 27b adjacently, is the Ti thin layer 24.
It is a semi-insulating high resistance layer having a sheet resistance of 50 M to 500 MΩ / □, which is approximately 75 Å to 300 Å, which is larger than the thickness of a ′. That is, the titanium oxide thin layer 28 is not TiO ₂ (titanium dioxide) that can be regarded as a perfect insulator,
So-called oxygen-poor titanium oxide containing less oxygen than TiO ₂ .
TiO _x (where x is a numerical value smaller than 2).

Next, as shown in FIG. 1 (F), a titanium chip thin layer 28 and a Ti thin layer 24a 'forming the second barrier electrode 27b are covered with an insulating layer 29 to form a semiconductor chip having a shutter barrier. Completed the power supply barrier diode chip. The insulating layer 29 is formed by plasma CVD.
It is made of a silicon oxide film formed by the (chemical vapor deposition) method. The insulating layer 29 may be replaced with a silicon nitride film formed by plasma CVD or photo CVD method, a polyimide resin film formed by coating method, or the like, but a silicon oxide film formed by plasma CVD method or photo CVD method is preferable. It was.

Although illustration is omitted, it is usual that, for example, a Ti layer and an Au layer are sequentially provided on the upper surface of the Al layer 25b and used as a connecting electrode for the lead member.

The dimensions of each part in FIG. 2 are exemplified as follows. First
The width a of the barrier electrode 27a is about 910 μm, and the width b of the second barrier electrode 27b is about 20 μm.
The width c of 8 is about 140 μm, and the width d between the titanium oxide thin layer 28 and the edge of the n-type region 23 is about 150 μm.

In this shutter barrier diode, a first shutter barrier and a second shutter barrier are generated between the first barrier electrode 27a and the n-type region 23 and between the second barrier electrode 27b and the n-type region 23, respectively. Not only that, a third shutter barrier occurs between the thin titanium oxide layer 28 and the n-type region 23. Titanium oxide thin layer 2
The occurrence of a shutter barrier between the No. 8 and the n-type region 23 was confirmed by the rectification characteristics, capacitance characteristics, saturation current characteristics, etc. of the shutter barrier diode. For example, when the area of the titanium oxide thin layer 28 is increased from zero, the saturation current I _S becomes equal to the titanium oxide thin layer 28 and the first and second barrier currents 27.
It increases substantially in proportion to the sum of the areas of a and 27b. It has been confirmed that this proportional relationship can be obtained at various temperatures of the shutter barrier diode. Titanium oxide thin layer 2
That the saturation current I _S changes substantially in proportion to the sum of the areas of 8 and the first and second barrier electrodes 27a and 27b is substantially the same as the first and second barrier electrodes 27a and 27b. It means that a reverse current flows through the titanium oxide thin layer 28 at the current density. This phenomenon directly shows that the thin titanium oxide layer 28 forms a Schottky barrier having a barrier height φ _B substantially the same as that of the first and second barrier electrodes 27a and 27b made of metal.

The solid line characteristic curve of FIG. 3 shows the reverse voltage-reverse current characteristic of the shutter barrier diode of FIG. 1 (F) according to the present invention, and the broken line characteristic curve of FIG. 1 (F) for comparison. From the Tsutuki barrier diode to the titanium oxide thin layer 28 and the second
The reverse voltage-reverse current characteristic of the Schottky barrier diode having the structure in which the barrier electrode 27b of FIG. As is apparent from the comparison of the two characteristic curves, the breakdown voltage of the Schottky barrier diode having the titanium oxide thin layer 28 and the second barrier electrode 27b according to the present invention is about 250V.
Therefore, the breakdown voltage of the conventional Schottky barrier diode having no titanium oxide thin layer 28 is about 60V.
And the titanium oxide thin layer 28 and the second barrier electrode 2
It can be seen that 7b is responsible for the significant improvement in breakdown voltage. The value of the breakdown voltage (about 250 V) of the Schottky barrier diode having the titanium oxide thin layer 28 and the second barrier electrode 27b is at a level substantially equal to the bulk breakdown voltage (the breakdown voltage at the center of the first barrier electrode 27a). Is believed to have reached.

Next, the reverse voltage-reverse current characteristics of the Schottky barrier diode according to the present invention will be described in more detail with reference to FIGS. 3 and 4. When the reverse voltage applied to the Schottky barrier diode is gradually increased from 0 volt, first, an extremely small saturation current I _S flows as shown in a region I of FIG.
At this time, a reverse current flows through the first and second shutter barriers based on the first and second barrier electrodes 27a and 27b, and also a reverse current flows through the third shutter barrier based on the titanium oxide thin layer 28. . A reverse voltage application circuit (not shown) is connected to the first barrier electrode 27a or anode and the ohmic electrode 26 or cathode, and is not directly connected to the titanium oxide thin layer 28. Therefore, the reverse current passing through the titanium oxide thin layer 28 is applied to the second barrier electrode 27.
b, and then into the first barrier electrode 27a. In the region I of FIG. 3, the value of the reverse current flowing through the titanium oxide thin layer 28 is small, so that the voltage difference between the point of the titanium oxide thin layer 28 near the second barrier electrode 27b and the point distant from the second barrier electrode 27b is small. Not so big. That is, the potential gradient in the lateral direction of the titanium oxide thin layer 28 is small, and the potential of each part of the titanium oxide thin layer 28 is substantially equal to the potential of the second barrier electrode 27b. Here, the Ti thin layer 24 forming the second barrier electrode 27b
Although a'has a higher resistance than the Al layer 25b, the voltage value at the second barrier electrode 27b can be ignored because the current value is small. Therefore, the potentials of the respective portions of the second barrier electrode 27b and the titanium oxide thin layer 28 are the same as those of the first barrier electrode 27.
It can be said that it is almost equal to the potential of a.

When the reverse voltage is further increased to about 60 to 100 V, breakdown occurs in a plurality of minute regions on the outer peripheral edge of the titanium oxide thin layer 28, and the reverse current increases stepwise as shown in a region II in FIG. To do. One step of this staircase corresponds to a breakdown at one outer peripheral edge of the titanium oxide thin layer 28. In the conventional Schottky barrier diode, a large reverse current flows due to a breakdown in a minute region, but a large reverse current does not flow in the Schottky barrier diode according to the present invention. That is, since the titanium oxide thin layer 28 is a semi-insulating high resistance layer, current limitation is caused by the resistance component of the titanium oxide thin layer 28, and a large increase in reverse current is suppressed. At the end of Region II, a thin layer of titanium oxide 2
8 between the inner peripheral side end P ₁ in contact with the second barrier electrode 27b and the outer peripheral side end P ₂ far from the second barrier electrode 27b increases, and as a result, the titanium oxide thin layer The potential difference between the outer peripheral end P _{2 of} 28 and the ohmic electrode 26 does not increase so much even if the reverse direction applied voltage is increased. The sheet resistance of the second barrier electrode 27b is sufficiently smaller than the sheet resistance of the titanium oxide thin layer 28,
It can be said that the voltage drop at the second barrier electrode 27b can be ignored. Here, the breakdown already generated at the outer peripheral edge P ₂ is maintained as it is, and the reverse current due to this breakdown continues to flow through the titanium oxide thin layer 28. Area II
In I, the first and second barrier electrodes 27a, 27b based on the first and second barrier electrodes 27a, 27b are formed according to the increase in the reverse voltage because no new breakdown occurs at the outer peripheral edge P ₂ of the titanium oxide thin layer 28. The reverse current through the third barrier and the third barrier based on thin titanium oxide layer 28 is gradually increased. If, titanium does not form Shiyotsutokibaria a portion of thin oxide layer 28 resistor layer for example n ^- replaced by form high-resistance GaAs layer, and an end portion of the resistor layer second
If the barrier electrode 27b is brought into ohmic contact with the barrier electrode 27b, the reverse current (leakage current) increases significantly as the reverse voltage increases, and the breakdown voltage eventually decreases. Since the thin titanium oxide layer 28 according to the present invention not only has a high resistance but also forms a Schottky barrier, the leakage current suppressing effect is larger than that of the above-mentioned resistor layer.

Not only the first and second barrier electrodes 27a and 27b,
Since the reverse voltage is also applied to the titanium oxide thin layer 28, the depletion layer 30 schematically shown in FIG. 4 is formed as the n layer under the titanium oxide thin layer 28 and the first and second barrier electrodes 27a and 27b. It occurs in the shaped area 23. Titanium oxide thin layer 28 and n ^{+ type} region 2
The potential difference between 2 and ₂ decreases from the inner peripheral edge P ₁ toward the outer peripheral edge P ₂ . In addition, the first and second barrier electrodes 27
The surface of the n-type region 23 from a, 27b to the titanium oxide thin layer 28 is continuous as a shutter barrier.
As a result, the gentle depletion layer 30 capable of relaxing the electric field concentration is obtained, and the electric field concentration in the semiconductor region near the periphery of the second barrier electrode 27b is relaxed. Therefore, as shown as a region III, a pair of electrodes 26, 27
A region where breakdown does not occur even if the reverse voltage applied between a increases increases continues widely.

When the reverse voltage becomes about 250V, the second barrier electrode 27b
Of the critical electric field strength Ecri between the periphery of the electrode and the ohmic electrode 26.
A point that exceeds t occurs and breakdown occurs.
The reverse current increases as shown in.

For comparison, when the reverse voltage-reverse current characteristics were measured before the Ti thin layer 24b shown in FIG. 1 (C) was oxidized, it was found that the Ti thin layer 24b did not become a sufficiently high resistance layer. Therefore, the reverse current cannot be suppressed as shown in the region II of FIG. 3, and the breakdown as shown by the broken line occurs as in the conventional case.

When the shutter barrier diode of the present invention was used as a rectifying diode of an output rectifying / smoothing circuit of a switching regulator having a switching frequency of 500 kHz, it was confirmed that switching loss and noise were extremely small. No decrease in the switching speed (high-speed response) due to the provision of the titanium oxide thin layer 28 was observed.

The advantages of this embodiment are as follows.

(1) Since the titanium oxide thin layer 28 is both a resistor and an object capable of producing a shutter barrier, it can favorably generate a depletion layer for relaxing the concentration of an electric field in the semiconductor region near the periphery of the second barrier electrode 27b. Therefore, the breakdown voltage can be significantly increased as compared with the conventional Schottky barrier diode having the structure in which only the field plate or the guard ring is provided and further the structure in which both the field plate and the guard ring are provided.

(2) Superior in high-speed response compared to the conventional Shoutoki barrier diode having a guard ring.

(3) The thermal instability of the characteristics found in the conventional Schottky barrier diode having a field plate with an insulating layer interposed is eliminated. In addition, the titanium oxide thin layer 28
Has a strong shielding property against ions that enter from the outside. In addition, the breakdown voltage does not decrease as the temperature rises. Therefore, the stability and reliability of breakdown voltage are extremely high.

(4) The titanium oxide thin layer 28 is obtained by oxidizing the Ti thin layer 24b, which is the extended portion of the Ti thin layer 24a provided immediately below the Al layer 25a, so that the titanium oxide thin layer 28 can be easily obtained. You can Moreover, electrical connection between the first barrier electrode 27a, the second barrier electrode 27b, and the titanium oxide thin layer 28 can be achieved easily and reliably.

(5) The second barrier electrode 27b composed of the Ti thin layer 24a 'between the first barrier electrode 27a and the titanium oxide thin layer 28.
Since it is arranged, there is little decrease in breakdown voltage during ultra-high speed operation. That is, a shutter barrier diode is formed without the second barrier electrode 27b shown in the first (D), and this shutter barrier diode is operated at an extremely high speed (commutation time 10
Withstand voltage of 15V to 20V
The phenomenon that the degree of deterioration has decreased appears. This phenomenon is considered as follows. During ultrafast reverse commutation, the titanium oxide thin layer 28 experiences a capacitive transient current that is much greater than in the steady state. Therefore, the potential gradient of the titanium oxide thin layer 28 becomes too large, electric field concentration occurs in the semiconductor region near the periphery of the first barrier electrode 27a, and as a result, the breakdown voltage decreases. On the other hand, in the shutter barrier diode related to the present invention of FIG. 1 (F), the second barrier electrode 27
The sheet resistance of b is significantly smaller than that of the titanium oxide thin layer 28. Therefore, the potential gradient generated in the second barrier electrode 27b due to the transient current is equal to that of the first barrier electrode 27a.
The electric field concentration in the vicinity of the peripheral edge of is gentle. As a result, the breakdown voltage reduction during ultra-high speed operation is reduced.
In the steady state, the potential gradient generated on the second barrier electrode 27b is reduced to a negligible level, and instead, the potential gradient generated on the titanium oxide thin layer 28 has an improved breakdown voltage.

(6) The second barrier electrode 27b composed of the Ti thin layer 24a 'between the first barrier electrode 27a and the titanium oxide thin layer 28.
Since it is arranged, the breakdown voltage can be further improved. That is, stress is generated in the n-type region 23 due to the difference in the thermal expansion number between the relatively thick first barrier electrode 27 a and the n-type region 23. Since the second barrier electrode 27b and the titanium oxide thin layer 28 are extremely thin, the stress due to these is applied to the first barrier electrode 27.
It can be ignored as compared with the case of a. Therefore, a portion where the stress is concentrated is formed below the peripheral portion of the first barrier electrode 27a on the surface of the n-type region 23. At this stress concentration point, since the strain is applied to the n-type region 23, the critical electric field strength Ecrit that causes the breakdown is lower than that in other portions. Therefore, if the electric field is concentrated at this stress concentration point, breakdown is likely to occur.

On the other hand, it is well known that the electric field concentration point occurs at a portion below the boundary portion of the coating film having greatly different sheet resistance.
That is, in this embodiment, an electric field concentration point occurs on the surface of the n-type region 23 below the boundary between the second barrier electrode 27b and the titanium oxide thin layer 28. The mechanism by which this electric field concentration point occurs can be understood from the following. First and second
The sheet resistance of the barrier electrodes 27a and 27b is sufficiently smaller than the sheet resistance of the titanium oxide thin layer 28. Therefore, when the reverse voltage is applied, the first and second barrier electrodes 2
It can be said that the potentials of the respective portions 7a and 27b are equal. As a result, it can be considered that the potentials are equal on the surfaces of the n-type regions 23 below the first and second barrier electrodes 27a and 27b. On the other hand, the titanium oxide thin layer 28 has a large sheet resistance, and when the reverse current increases, as described above, the inner peripheral edge P ₁ and the outer peripheral edge are increased.
The potential difference with P ₂ increases. As a result, the potential difference between the titanium oxide thin layer 28 and the ohmic electrode 26 is reduced to the outer peripheral edge.
It decreases gradually toward the P ₂ side. Strictly speaking
The reverse current density of the titanium oxide thin layer 28 is large on the side closer to the second barrier electrode 27b and decreases exponentially toward the side farther away. Therefore, the above-mentioned potential difference is also due to the outer peripheral edge P ₂
It decreases exponentially toward. Therefore, n
The electric field in the minute section of the surface portion of the n-type region 23 in the thickness direction (first direction) of the shaped region 23 has a continuous constant value under the second barrier electrode 27b, It gradually decreases exponentially from the boundary portion between the barrier electrode 27b and the titanium oxide thin layer 28 toward the outer peripheral end P ₂ of the titanium oxide thin layer 28. In addition, in a minute section of the surface portion of the n-type region 23 in a direction (second direction) that vertically crosses the boundary portion between the second barrier electrode 27b and the titanium oxide thin layer 28 along the surface of the n-type region 23. Is almost zero in the lower part of the second barrier electrode 27b, and increases sharply when reaching a minute section including the lower part of the boundary between the second barrier electrode 27b and the titanium oxide thin layer 28, Further, in the lower part of the titanium oxide thin layer 28, it gradually decreases exponentially from the lower part of the boundary part to the lower part of the outer peripheral edge P ₂ . Further, in the third direction which is perpendicular to both the first direction and the second direction along the surface of the n-type region 23, the curvature of the second barrier electrode 27b and the titanium oxide thin layer 28 is sufficiently large. Therefore, it is considered that the electric field is barely generated.
The electric field on the surface of the n-type region 23 is the vector sum of the electric field in the first direction, the electric field in the second direction, and the electric field in the third direction, and therefore, in the second direction. The peak of the electric field works and the second barrier electrode 27b and the titanium oxide thin layer 28 are
A region where the maximum electric field is generated, that is, an electric field concentration point exists below the boundary portion of the.

As described above, the stress concentration point is located below the boundary between the first barrier electrode 27a and the second barrier electrode 27b,
The electric field concentration point is located below the boundary between the second barrier electrode 27b and the titanium oxide thin layer 28. In this embodiment,
Thin second barrier electrode 27b composed of Ti thin layer 24a '
By providing the above, the stress concentration point and the electric field concentration point can be separated from each other, the frequency of products having a specified breakdown voltage or less is reduced, and the breakdown voltage yield is improved. That is, the average value of breakdown voltage was improved.

(7) Ti thin layer 24a 'adjacent to the first barrier electrode 27a
By providing the second barrier electrode 27b composed of
The reverse surge resistance was improved, albeit slightly.

[Second Embodiment] Next, a Schottky barrier diode according to a second embodiment of the present invention shown in FIG. 5 will be described. However, in FIG. 5 and FIGS. 6 to 8 showing the third, fourth and fifth embodiments which will be described later, the same reference numerals are given to the substantially same parts as FIG. And their description will be omitted. Fifth
Fig. 1 (F) for the shutter barrier diode shown in the figure.
Similarly to the above, the second barrier electrode 27b made of the Ti thin layer 24d is formed so as to surround the first barrier electrode 27a made of the Al layer 25b and the Ti thin layer 24c, and surrounds the second barrier electrode 27b. Thus, the titanium oxide thin layer 28 is formed. However, in the Schottky barrier diode of FIG. 5, the layer thickness of the Ti thin layers 24c and 24d is thicker than the Ti thin layers 24a ′ and 24a ″ of the Schottky barrier diode of FIG. 1 (F). The barrier electrode 27b is formed to have a thickness of about 100 to 400 Å, which is about twice as thick as the titanium oxide thin layer 28. In order to manufacture the Schottky barrier diode of FIG. The layer 24 is provided thickly, the Ti thin layer 24 is exposed as shown in FIG. 1 (C), and then the Ti thin layer 24b is etched to a predetermined thickness and oxidized as shown in FIG. 1 (D). Later, as in Fig. 1 (E)
The Al layer 25a is etched to expose the thick Ti thin layer 24d to obtain the second barrier electrode 27b. Also, the Ti thin layer 24
It is possible to oxidize b while keeping it thick to form the titanium oxide thin layer 28 and then etch it to a predetermined thickness, and various manufacturing methods can be considered. When the thickness of the second barrier electrode 27b is increased in this way, the effect of increasing the reverse surge withstand capability becomes more remarkable than in the case of the first embodiment. That is, the second via electrode 27
If the resistance of the portion b is large, a large current at the time of breakdown flows concentratedly on the peripheral portion of the first barrier electrode 27a, so that the reverse surge resistance becomes small. In this embodiment, the layer thickness of the second barrier electrode 27b is increased and the resistance is further lowered as compared with the case of the first embodiment, so that the second barrier electrode 27b serves as a passage of the reverse surge current. ) Works more effectively than As a result, the reverse surge current becomes the second
It is possible to alleviate the fact that a large current flows in a distributed manner to the barrier electrodes 27b of the above and concentrates as described above, and the reverse surge withstand capability increases.

[Third Embodiment] The Schottky barrier diode of the third embodiment shown in FIG. 6 has a ring-shaped titanium oxide thin layer 28a on the surface of the n-type region 23, and further has a ring-shaped titanium oxide thin layer 28a. Thin titanium oxide layer 28b. The first barrier electrode 27a is the Al layer 2
5b and two Ti thin layers 24e and 24f. The second barrier electrode 27b is composed of two Ti thin layers 24g and 24h. This structure can be formed by repeating the steps shown in FIGS. 1 (B) to (D) twice. The upper titanium oxide thin layer 28b is formed to have the same thickness and degree of oxidation as the titanium oxide thin layer 28 of FIG. 1 (F). Although the lower titanium oxide thin layer 28a has almost the same thickness as the upper titanium oxide thin layer 28b, since the degree of oxidation is stronger than that of the upper titanium oxide thin layer 28b, The sheet resistance is larger than that of the titanium oxide thin layer 28b. Therefore, the current flowing through the upper thin titanium oxide layer 28b is larger than the current flowing through the lower thin titanium oxide layer 28a, and the potential gradient is mainly determined by the upper layer. Further, in the outer peripheral portion of the titanium oxide thin layer 28a where the titanium oxide thin layer 28b is not overlapped, the sheet resistance is higher than that of the two-layer portion. As a result, the region II in FIG. 3 shifts to the high voltage side and the reverse current in the region II increases less,
It is possible to provide a shutter barrier diode having a small reverse current. Further, the inner peripheral edges of the upper titanium oxide thin layer 28b and the lower titanium oxide thin layer 28a do not coincide with each other, and are in a step shape. Therefore, the Ti thin layer 24g, 24
The sheet resistance at the boundary between the second barrier electrode 27b made of h and the titanium oxide thin layers 28a and 28b changes in two steps, and the electric field concentration at the lower part of the boundary surface can be relieved and the breakdown voltage becomes higher. Has the advantage of improving. It is more preferable to continuously change the change in sheet resistance at this boundary. or,
As with the second embodiment, the reverse surge withstand capability can be improved as compared with the first embodiment.

[Fourth Embodiment] In a shutter barrier diode of the fourth embodiment shown in FIG. 7, the Ti thin layer 24a ″ and the titanium oxide thin layer 28 of the shutter barrier diode of FIG. The structure has a structure in which a Ti thin layer 31 that is thicker than the Ti thin layer 24a ″ is provided below the boundary portions of these connected. The Ti thin layer 31 is an extremely thin film having a thickness of 100 to 400 Å at all times. In this case, the Al layer 25b and the lower part thereof are covered with the first barrier electrode 27.
a, a part of the Ti thin layer 31 outside the lower part of the Al layer 25b can be considered as the second barrier electrode 27b. The Schottky barrier diode of FIG. 7 can be manufactured by a manufacturing process in which the Ti thin layer 31 is formed on the upper surface of the n-type region 23 in advance and the process of FIG. 1 (E) is omitted. The Schottky barrier diode shown in FIG. 7 can improve the reverse surge withstand capability as compared with the first embodiment as in the second and third embodiments. Further, in the peripheral portion of the Ti thin layer 31 of the first barrier electrode 27a, a Schottky barrier having a barrier height φ _B larger than that of the portion of the Ti thin layer 24a ″ is formed.
Therefore, the peripheral breakdown voltage of the main shutter barrier is improved, and the breakdown voltage of the shutter barrier diode is further improved.

[Embodiment of FIG. 5] When strict high-speed response is not required, a guard ring can be provided to provide a Schottky barrier diode with high withstand voltage and large reverse surge resistance. By providing the guard ring, the depletion layer based on the first and second barrier electrodes, the depletion layer based on the pn junction by the guard ring, and the depletion layer based on the thin titanium oxide layer are continuously spread. FIG. 8 shows an example of a shutter barrier diode according to the present invention having a guard ring. The Schottky barrier diode shown in FIG. 8 has a p ^{+ -type} (high-concentration p-type) region in the n-type region 23 below the boundary between the second barrier electrode 27b and the titanium oxide thin layer 28 as shown in the figure. A guard ring consisting of 32 is provided. In this case, the p ^{+ -type} region 32, the second barrier electrode 27b and the titanium oxide thin layer 2
The contact with No. 8 is considered to be a resistive contact having a relatively large resistance value showing a slightly non-linear resistance tendency, and is not at least a shock barrier. Therefore, the shutter barrier region based on the first and second barrier electrodes 27 a and 27 b and the shutter barrier region based on the titanium oxide thin layer 28 are separated by the p ^{+ -type} region 32. The broken line in FIG. 8 schematically shows the expansion of the depletion layer when a reverse voltage is applied.

The Schottky barrier diode shown in FIG. 8 has a second barrier electrode 27b and a titanium oxide thin layer 2 which tend to cause electric field concentration.
A p ^{+ type} region 32 was formed in the lower part of the boundary portion with 8. Therefore, higher breakdown voltage can be expected. In addition, p ^{+ type} region 32
And the first barrier electrode 27a are the first barrier electrode 27a.
Is electrically connected via the second barrier electrode 27b having a higher resistance than Therefore, p ⁺
The forward current flowing in the shaped region 32 can be limited by the second barrier electrode 27b, and the effect of accumulating a small number of carriers can be reduced.
In other words, it can be said that the structure is desirable as a Schottky barrier diode having a guard ring, which can improve the breakdown voltage without significantly impairing the high-speed response. Note that when the main purpose is to improve the reverse surge resistance, it is unavoidable that the high-speed response is impaired, and the first barrier electrode 27
The p ^{+ -type} region 3 is formed below the boundary between a and the second barrier electrode 27b.
Form 2.

[Modification]

The present invention is not limited to the above-mentioned embodiments, and the following modifications are possible, for example.

(1) The sheet resistance of the titanium oxide thin layers 28, 28a, 28b varies depending on the semiconductor chip structure and size, but is 10 kΩ / □ to 500 MΩ / □, preferably 10 MΩ / □ to 1000 MΩ / You should choose □.

(2) When the width c of the titanium oxide thin layer 28 is about 10 μm or more, the effect of improving the withstand voltage appears, and when it is 30 μm or more, the effect becomes remarkable. But,
In order to increase the yield to obtain a predetermined breakdown voltage, 100
It is more desirable to design to be μm or more. Width c 500
Even if it is set to be μm or larger, it is possible to obtain a sufficient withstand voltage improving effect. Therefore, although there is no upper limit on the width c,
Even if the width c is 500 μm or more, it is not possible to expect a relatively high breakdown voltage, and there is a problem that the semiconductor chip becomes large. Therefore, the width c is 30 to 50
It is desirable to set it in the range of 0 μm.

(3) The sheet resistance and width of the second barrier electrode 27b are 10 Ω / □ to 100 kΩ / □ and 2 μm to 50 μm, respectively.
It is appropriate to choose.

(4) The thickness of the second barrier electrode 27b, the thickness of the titanium oxide thin layer 28, and the two titanium oxide thin layers 28 of FIG.
The total thickness of a and 28b should be 5000 Å (0.5 μm) or less so as not to substantially cause a stress concentration point, and more preferably 3000 Å (0.3
μm) or less.

(5) The thickness of the Ti thin layers other than the Ti thin layers 24c, 24d and 31 is 20Å in consideration of the film thickness control, the oxidation temperature, the oxidation time, etc.
It should be above. The upper limit is not limited as long as the above-mentioned predetermined sheet resistance can be obtained, but when the Ti thin film is thermally oxidized to form a titanium oxide thin layer, it should be 300 Å or less in consideration of the oxidation temperature and the oxidation time. Is. The upper limit can be further expanded if strong oxidation such as plasma oxidation is performed.

(6) The oxidation temperature when the Ti thin layer 24b is oxidized to obtain the titanium oxide thin layer 28 is preferably 500 ° C. or lower,
When using an Au-based electrode, the temperature is 380 ° C. or lower. The lower limit of the oxidation temperature is 200 ° C. or higher when the thermal oxidation method is used, but it may be a low temperature of room temperature or lower when the plasma oxidation is used. The oxidation time depends on the thickness of the Ti thin layer 24b,
Although it changes depending on the oxidation temperature and the oxidizing atmosphere, it is desirable to set it within the range of 5 seconds to 2 hours.

(7) Titanium oxide thin layers 28, 28a and 28b are formed by vapor deposition or sputtering of titanium oxide,
The Ti thin layers 24a ', 24b, 24g, and 24h may be replaced with titanium nitride layers having relatively high conductivity. The titanium nitride layer can be formed by nitriding a thin Ti layer using the Al layer as a mask.

(8) A titanium oxide thin layer is suitable as a thin layer having a high sheet resistance and generating a shock barrier, but it may be a thin oxide layer of a Ta (tantalum) -based material. Further, the Ti thin layer 24 and the titanium oxide thin layer 28 may be those to which In, Sn or the like is added.

(9) A short-circuit electrode may be formed so that the outer peripheral ends of the titanium oxide thin layers 28 and 28b are in ohmic contact with the n-type region 23. In this case, the titanium oxide thin layers 28, 28a, 28b are caused by the current flowing from the short-circuit electrode to the first barrier electrode 27a by using the titanium oxide thin layers 28, 28b as a resistance path.
Potential distribution occurs. In addition, the titanium oxide thin layers 28, 2
8a, 28b are divided into inner and outer regions by concentrically forming an annular region of relatively high resistance, and the titanium oxide thin layers 28, 28 are formed.
The potential distribution of a and 28b may be stabilized.

(10) The present invention can be applied to a shutter barrier semiconductor device using a III-V group compound such as InP (Indium Neighboride) or silicon instead of GaAs.

(11) In the case of forming a shutter barrier semiconductor device in an integrated circuit, a planer is provided in which an n ^{+ type} region 22 is provided so as to surround the n type region 23 in an island shape, and an ohmic electrode 26 is provided on the surface side of the n type region 23. It may be a structure.

(12) The n-type region 23 and the n ^{+ -type} region 22 can be replaced with the p-type region.

〔The invention's effect〕

 The present invention has the following effects.

(A) The barrier electrode is composed of a first barrier electrode having a relatively thick thickness and a second barrier electrode having a thinner thickness, and the relatively thin second barrier electrode has a relatively thick first barrier electrode and a thin layer. Since it is arranged between the first barrier electrode and the semiconductor region, which is relatively thick, the stress due to the difference in thermal expansion coefficient between the semiconductor region and the boundary point between the barrier electrode and the thin layer is different from each other. Therefore, the breakdown voltage is improved or stabilized. That is, a stress concentration point occurs in the semiconductor region below the peripheral edge of the relatively thick first barrier electrode, and the critical electric field strength that causes breakdown decreases at the stress concentration point. on the other hand,
An electric field concentration point occurs in the semiconductor region below the boundary between the barrier electrode and the thin layer. If the stress concentration point and the electric field concentration point coincide with each other, breakdown easily occurs. However, in the present invention, since the electric field concentration point is separated from the stress concentration point by providing the relatively thin second barrier electrode, the breakdown voltage is increased. Improvement or stabilization is achieved.

(B) Since the first barrier electrode is composed of the first and second metal layers and the second barrier electrode is composed of the same metal as the first metal layer, good and easy electrical connection between the two can be achieved. can do.

(C) Since the thin layer capable of forming the Schottky barrier and having a high sheet resistance is the oxide layer of the same metal as the metal of the first metal layer, it can be easily manufactured, and the second layer Good and easy electrical connection to the first barrier electrode via the barrier electrode can be achieved.

[Brief description of drawings]

1 (A) to 1 (F) are sectional views showing the Schottky barrier diode according to the first embodiment of the present invention in the order of manufacturing steps, and FIG. 2 is a plan view showing the state of FIG. 1 (E). FIG. 3 is a reverse voltage-reverse current characteristic diagram of the Schottky barrier diode of FIG. 1 (F), and FIG. 4 is a partial enlargement of the Schottky barrier diode of FIG. 1 (F) schematically showing the depletion layer. Sectional drawing, FIG. 5 is a sectional view showing a shutter barrier diode of the second embodiment, FIG. 6 is a sectional view showing a shutter barrier diode of the third embodiment, and FIG. 7 is a fourth embodiment. FIG. 8 is a sectional view showing the shutter barrier diode of FIG. 8, and FIG. 8 is a sectional view showing the shutter barrier diode of the fifth embodiment. 22 ... N ^{+ type} region, 23 ... N type region, 24a, 24a ′,
24a "... Ti thin layer, 25, 25a, 25b ... Al layer, 26
... ohmic electrode, 27a ... first barrier electrode, 27b
... second barrier electrode, 28 ... titanium oxide thin layer, 29 ...
Insulation layer.

Claims (1)

[Claims] 1. A first region formed on the semiconductor region so that a Schottky barrier can be generated between the semiconductor region and the semiconductor region.
A barrier electrode, which is arranged on the semiconductor region so as to surround the first barrier electrode, is electrically connected to the first barrier electrode, and has a thickness smaller than that of the first barrier electrode. And a sheet resistance larger than that of the first barrier electrode, and a sheet resistance of 100 KΩ / □ or less,
And a second barrier electrode formed so as to generate a Schottky barrier between the semiconductor region and the semiconductor region, the second barrier electrode being disposed on the semiconductor region so as to surround the second barrier electrode, and Is electrically connected to the second barrier electrode, has a larger sheet resistance than the first and second barrier electrodes, and has a sheet resistance of 10 kΩ / □ or more, and is between the semiconductor region and the second barrier electrode. A thin layer formed so as to generate a Schottky barrier and thinner than the first barrier electrode, wherein the first barrier electrode is in contact with the semiconductor region. A first metal layer and a second metal layer formed on the first metal layer, wherein the second barrier electrode is a layer containing the same metal as the first metal layer. The thin layer is the first Schottky barrier semiconductor device which comprises a metal layer same metal as oxide layer.