JPS5848964A - Forming method for junction of reverse conducting thyristor - Google Patents

Abstract

PURPOSE:To harmonize turn-off time and reverse recovered charge by making a diode section having P<+>PNiN<+> structure more than a thyrister section in the degree of the disturbance of crystal structure generated in a reverse conducting thyristor substrate after each junction of the thyristor section having NEPBNi NBPE structure and the diode section is formed when each junction is shaped into the substrate. CONSTITUTION:A layer 15 combining the PB layer of the thyristor section and the P layer of the diode section is grown to the surface of an Ni layer 12 in epitaxial shape, and the two NE layers 10 of the thyristor section are formed to the layer 15 through diffusion at an interval. A layer 13' combining the NB layer of the thyristor section and the N<+> layer of the diode section is grown to the back of the substrate 12 in epitaxial shape, the N<+> layers 16 of the diode section are formed at both end sections through diffusion, only the layer 13' between the two layers 16 is exposed, and the whole other surfaces are coated with a SiO2 film 17. The thyristor NB layer 21 having high impurity concentration is shaped into the exposed section 20 of the layer 13' through diffusion, the thyristor PE layer 14 is formed to the surface layer section while the peripheral film 17 of the layers 10 is removed, the diode P<+> layer 22 is shaped to the film 17 removed through diffusion, and crystalline disturbance is generated in the layer 12.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a reverse conduction Nylith I, and more particularly to an improvement in a manufacturing method for the purpose of harmonizing the thyristor's turn-on time and the diode's reverse recovery charge.

Hereinafter, the present invention will be explained in detail using the drawings.

Figure 1 is a vertical cross-sectional view showing the structure of a conventionally commonly used reverse conduction thyristor, Figure 2WA (a) is a current and voltage waveform diagram of the reverse conduction thyristor during commutation, and Figure 2 (b) is a diagram of the waveforms of the reverse conduction thyristor during commutation. (
The structure of the reverse conduction thyristor can be roughly divided into the structure as shown in FIG.
Separation located between Nyristor part l and diode part 2?
Consists of IF 3 and surface shaping section 4, with thyristor section 1 and diode section separated by 2%? 13 and the surface shaping portion 4 are bonded to each other within a single silicone cone 5 so as to exhibit their respective a functions.

A reinforcing support electrode 6 made of tungsten and serving as an anode electrode is integrated on one side of the silicon disk 50 via a brazing material 7 mainly made of aluminum. Also, between the other silicon disks 5.

A cathode electrode 9 is formed by the Arun vapor deposition method in a shape that short-circuits the gate electrode 8 of the thyristor 41, the thyristor portion 1, the separation band 3, and the diode portion 2.

The thyristor part l in the sirilan disk s is formed of five layers: N1 layer (10% ivy part, P1 base layer 11%, NL layer 12, N1 paste layer 13, P1 entrant layer 14), and the gate electrode 8 The part is well-known and 7 ← 1 base layer 11.8
It is formed of four layers: the first layer 12, the N1 base layer 13, and the first P1 layer 14.

On the other hand, the diode section 2 is an r layer! ! , NA layer 1! , N
It is formed from three layers: layer 16.

This is the same as the hard part 2, and the surface is increased according to the element breakdown voltage.

In this way, you can control the mushrooms! A reverse conduction thyristor used for control and other applications is characterized by a structure in which a thyristor section and a diode section are connected in antiparallel to each other with a separation band as a boundary and integrated within a single substrate. In order to cut off the thyristor current in the conductive state, the reverse conduction thyristor enters the blocking state by passing a current in the opposite direction to the thyristor current through the diode portion, and this operation is called commutation.

The second picture (1) is the current of reverse conduction cyrisk during commutation.

In the voltage waveform diagram, the current is shown as a solid line and the voltage is shown as a broken line. Qrr)
Therefore, current due to the accumulated carriers continues to flow in the layer direction when viewed from the thyristor side. This current is the reverse recovery current (I
rp), and the time product with the reverse recovery time (〒rr) is called the reverse recovery charge.

In reverse conduction thyristors, in order to improve switching characteristics, the lifetime of minority carriers within the element is shortened by gold diffusion or radiation irradiation, thereby reducing the reverse recovery charge. After flowing, when a reapply voltage (VDM) is applied in the layer direction with reverse recovery charge remaining, this reverse recovery charge, that is, accumulated carriers, causes the thyristor to fire and cause commutation failure. In order to prevent this phenomenon, it is common knowledge to design the diode to have a smaller reverse recovery charge by making the lifetime of the diode shorter than the live time of the thyristor.

Generally, when applying reverse conduction thyristors to equipment,
A saturable reactor is used by inserting it in series with the element, but if the reverse recovery charge of the diode becomes large, the following problems occur.

(1) The capacity of the saturable reactor for suppressing the reverse recovery charge increases, leading to an increase in the size and cost of the device.

(Dynamic) This causes an increase in the reapplied voltage, increases the required withstand voltage rating of the device, and reduces the yield of the device.

For these reasons as well, it is desirable to design a reverse conduction thyristor so that the reverse recovery charge of the diode is small. However, in a reverse conduction thyristor, the thyristor portion and the diode portion are formed in one silicon substrate, so from a manufacturing standpoint, it is not easy to establish the above relationship for the following reasons.

(1) Depending on the selection and combination of impurity atoms used to form and form the junction of the reverse conduction thyristor, t causes a difference in the atomic radius of the impurity atoms relative to the silicon substrate between the thyristor part and the diode part, and the crystal The way it is disturbed is different. For this reason, when a lifetime killer is introduced into a substrate using the thermal diffusion method, it is difficult to establish the above relationship because more lifetime killers are introduced to the side where the crystal structure is more disordered. (2) Thermal diffusion When introducing a lifetime killer in the method to create a lifetime difference between the thyristor part and the thyristor part, when trying to make the 2-if time of the diode part smaller than that of the thyristor part, it is necessary to It is necessary to set it higher than the diffusion temperature. Even if we try to create a difference in the lifetime between the two parts by creating a difference in thermal diffusion temperature in this way, the concentration distribution of gold, platinum, etc. in the silicon substrate, which is critical for two ifs, will not be affected by the thermal history. As a result, it is difficult to provide a difference in lifetime between the two parts within a narrow range because the heat diffusion conditions are easily redistributed.

For the above reasons, for example, if the reverse recovery charge of the diode section is made too small, the turn-off time of the thyristor section will become too short, and the characteristics of the thyristor section will deteriorate. exist. For this reason, it has been difficult to manufacture a device with a good yield in which the turn-off time of the thyristor portion and the reverse recovery charge of the diode portion are balanced.

The present invention is a method for forming a junction of a reverse conduction thyristor, which has been developed to solve the above-mentioned problems, and is particularly concerned with the turn-off time of the thyristor section and the reverse conduction of the diode section.

Hereinafter, a conventional method and a bond forming method according to the present invention will be compared and explained using the drawings.

Figures 3 (a), (b), (0), (a), and (.) are engineering drawings for explaining an example of the formation of a conventionally used reverse conduction cross-conductor junction. Note that the same reference numerals as in FIG. 1 indicate the same or similar parts.

32 mm in diameter and 19 in thickness as shown in Figure 3(a).
A silicon disk 5 having a diameter of 0 μm is prepared. To explain the structure of this silicon disk 5, reference numeral 12 has 81 layers, its specific resistance is about 40Ω, and its thickness is 1004 m. 15
' is a P1 paste layer in the silice part.7 In order to form the P layer in the diode part, the thickness is 7011!11 and the surface concentration is about 5X1 (l
The P layer 613' formed of atomsloa has a surface concentration of I using phosphorus as an N type impurity to form the N1 base layer of the thyristor part and the N layer of the diode part.
This is an N layer formed by an epitaxial method to have a thickness of 20 μm+ with X 10 atoms loa. 17 is an oxide film formed to a thickness of approximately 2 μm by thermal oxidation method.
Using a known photoresist technique for the oxidation [117], a window 18 is formed to form the N1 emitter layer of the thyristor section for the oxide 3117 on the 15' plane of the layer 1, and a window 18 is formed on the 13° plane of the N layer. Windows 19 are respectively opened in the oxide film 17 for forming the N layer of the diode portion. Regarding the positional relationship between the windows 18 and 19 formed in the oxide film 17, the window 19 is formed at a position farther from the center than the window 18 in order to provide a separation zone.

Figure 3 (0) shows that using phosphorus as an N-type impurity, after preliminary diffusion, oxidized J [17] is removed and additional diffusion is carried out. Layer 16 has a thickness of 415 μm 1 with a surface concentration of approximately I
A state in which 10 atoms/co were simultaneously formed is shown. At this time, a new oxide film 17 is formed on the silicon disk.

In FIG. 3(d), a window 20 is opened in the oxide film 17 on the 13° plane of the N layer in order to form a single layer of P1 emitter in the thyristor section, and then a P1 emitter is formed through this window 20 using gelon, which is a P-type impurity. One layer of industrial ivy 14 has a thickness of 10 μm 1 surface concentration I
Formed by diffusion method with X 10 atoms/co,
The reverse conduction thyristor junction formation process is completed. After the bonding is completed, the silicon disk 5 has a thyristor part 1, a diode part 2, a separation band 3, a surface shaping area 4, and a thyristor part 1 has a Nl construction ivy layer 10. Pya base layer 11. It consists of an NA layer 12, an N1 paste layer 13, a P1 emitter layer 14, and a second layer 15 for the diode section 2.
It is composed of N↓ layer 12% and N layer 16. Each of these layers undergoes heat treatment in the bonding process, and the final thickness relationship is as follows. In other words, regarding Cyrisk Department IK,
N1 construction (one layer of ivy 10 is 20 μm, PB base layer 11 is 55 layers m, N1 layer 12 is 854 mmNm base layer 13 is 20 μm, I'i + construction one layer of ivy 14 is 10 μm, diode part 2 is 2 layers 15 is 75 layers m,
The NL layer 12 has a thickness of 85 m.

The N layer 16 has 30 m layers.

Figure 3 (•) shows the state before gold is deposited on the silicon disk after the junction has been formed in order to shorten the lifetime of the thyristor and diode parts. Entsuta layer 10 and P1 base layer 11
The oxide film 17 on the exposed surface is left.

Remove all other locations.

Gold, which is a live time killer, is deposited on both sides of the silico/disk in this state using a vacuum evaporation method, and then gold diffusion is performed in an inert gas at a temperature of 380°C for 3 hours. After that, the oxide film 17 is removed, and a reverse conduction thyristor having the structure shown in FIG. 1 is manufactured in accordance with the procedure also explained in FIG.

The reverse-conducting Nylithium silicon disk formed as described above was bonded by applying an impurity as shown below.

To explain Table 1, an O mark is, for example, a diode section (P-N layer-N) with an O mark written below the P layer, indicating a P layer, and an O mark below the N layer. Those written with ``indicate'' indicate the N layer. t() is an expression of the surface concentration and the difference in radius of impurity atoms with respect to silicon (8L).

After forming the bond, the N of the thyristor part is
When examining the degree of division in the crystal structure of the L layer using a rocking curve using the X-ray two-crystal method, it was found that strong crystal disorder occurred in the direction from the P1 layer to the N1 layer in the thyristor section. The cause of this is that the atomic radius of silicon (S) in the NL layer (x, lrL), which has the smallest atomic radius of 0.88, was diffused at a high concentration into the PI layer of the thyristor part. , J[It is determined that distortion has occurred in the crystal lattice due to the difference in the child radius (75 degrees for S↓).

On the other hand, the results of a similar investigation on the diode section showed that there was no disorder in the crystal structure in the NA layer. It wasn't noticed. The reason for this is that gallium (107% of the atomic radius of 8A) is formed on one side of the N1 layer, and phosphorus (949G of the atomic radius of 8A) is formed on the other side, so there is a difference in the atomic radius between them. It is thought that they absorb each other.

As a result, the degree of disorder in the crystal structure of the thyristor part and diode part of the NL layer is in the relationship na<ns, where M is the degree of disorder in the diode part and n# is the degree of disorder in the thyristor part. There is.

For this reason, even if gold, which is a lifetime killer, is simultaneously diffused into the diode and thyristor parts by thermal diffusion of 1[1] into the silicon disk above, the amount of gold that is a lifetime killer will be absorbed into the silicon disk, that is, where the silicon crystal structure is disordered. It follows the principle of gold diffusion. For this reason, the relationship between the lifetime t in the silicon substrate is such that the area where a large amount of gold has diffused becomes smaller, so the relationship between the lifetime of the thyristor part and the diode part is as follows: τd is the lifetime of the diode part. If the lifetime of the thyristor section is T-, then the relationship τd〉 τ- is obtained, and 1 < T is desired as a reverse conducting thyristor.
It becomes difficult to secure the relationship a. While maintaining the above relationship, gold is added to reduce the reverse recovery charge in the diode part.
If the amount of doping is increased, the on-voltage of the thyristor section will increase and the single-ignition characteristics will deteriorate, which is not preferable. Therefore, when manufacturing a reverse conduction thyristor, it is necessary to set an appropriate am to the degree of disorder of the crystal structure in the N1 layer of the diode part and the thyristor part when forming the junction.

The present invention has been made in consideration of the above-mentioned points.

Fig. 4 (a), (b), (o), (d), and (.) are process diagrams showing one embodiment of the method for forming a joint of reverse conduction cyrisk of the present invention; The same reference numerals as in FIG. 3 indicate the same or corresponding parts.

In Figure 4, the process in Figure 4 (IL) is shown in Figure 3 (a).
4(b).
4 (1)) shows that, as shown in FIG.
After forming the N' layer 16 in the diode section, an oxide film 17 is formed to form the P1 emitter layer in the silicon risk section.
The window 2G is shown open. This 920
From the 13th plane of the N layer, antimony trioxide (Sb
An N' layer 21 having a thickness of 1 μm and a surface concentration of Sb) of I×10 atoms is formed as shown in FIG. 4(6).

The reason why antimony was used to form the N° layer 21 is explained below. Silico Y to form a single layer of P1 electrical ivy in the process
, boron has the smallest atomic radius (0,88λ)
The purpose is to have both impurity atoms cancel out the difference in atomic radius when the impurity atoms are diffused. The purpose is to increase the average atomic radius from 0.88λ of boron alone and to reduce the degree of disorder in the crystal structure of the N sublayer in one thyristor part by overlapping diffusion of both. (2) Next, we will discuss the diffusion coefficient of antimono. is about 1 from &0
Since it is an order of magnitude smaller, antimony diffusion hardly progresses during heat treatment due to boron diffusion, so it can be ignored. Therefore, if boron is diffused from the previously diffused N° layer 21 surface at a surface X concentration higher than the surface concentration of antimony, the type of the N-type antimony diffusion layer is converted to P-type, and the P layer is almost the same as the conventional method. A single layer of ivy can be formed.

FIG. 4(a) shows a surface concentration of I x 10 atoms loa and a thickness of 10 by boron diffusion for the 21st surface of the N° layer.
ttm P1 building with one layer of ivy 14 formed. In front of this Plx i single layer type plate, P
The oxide film 17 on the 415th surface is removed, and when the P1 process 2 ivy layer 14 is formed, it is
5x to atoms/11a thickness approx. 10μm
A P layer 22 is formed. The purpose is to have an atomic radius of 1.2
With gallium diffusion, low surface concentration (5X10
When a P+N diffusion layer of boron (0.88 people) with a small atomic radius and a high surface concentration (5X 10 atomsloa) is formed on the P layer surface formed with a The dominating power of melon atoms becomes stronger, and it becomes easier to induce disorder in the crystal structure from the P layer to the NL layer than when using gallium alone.

Considering the atomic radius of silicon, which is 1.17 mm, the P layer in the diode section is 1.07 mm, and the P layer in the thyristor section is 1.07 mm, and the P1 emitter layer in the thyristor section is is 1.12 people, and the degree of disorder of the crystal structure affecting the 81st layer is greater in the diode part than in the thyristor part.

At the stage shown in FIG. 4 (11), the first step of forming a bond in this embodiment has been completed, and the thickness relationship between each layer is almost the same as that in the conventional case shown in FIG. Regarding the diode part, the P layer is 10 μm and the P layer is 65 μm%N.
L layer is 85μm% NI! I is 30 μm.

The silicon disk of the reverse conduction cyrisk according to this embodiment formed as described above is the 21st silicon disk shown below! Impurities such as are applied to form a bond.

In the example, the P layer of the diode part and the two layers of the silice part
I explained an example using gallium with 111.26x atoms in one layer, but with other P-type impurities, half an atomic sx,
A similar effect can be obtained by using Al1=U having ze Those marked with a mark indicate the P layer, those with an ◎ mark written below the P layer indicate the P layer, and those with an O mark written below the N layer indicate the N+ffi. There is. Also, the value within 0 indicates the surface concentration and the impurity atomic difference relative to silicon (S).

With respect to the silicon disk after the tptx surface junction is formed, the degree of disorder of the crystal structure of the thyristor part and the diode part is expressed as: rl(1, the degree of disorder of the thyristor part n-) When compared with the rocking curves obtained by the X-ray two-crystal method, a relationship of na > n+3 was clearly obtained in this example. This relationship can also be understood from the reason described in the process description of this example. Ru.

The degree of disorder of the crystal structure in the NL layer in a single silicon substrate is greater in the diode part than in the thyristor part, so if gold is thermally diffused into a silicon disk, gold doping will inevitably occur. The amount of doping will be greater in the diode part than in the thyristor part. As a result, the quick im relationship is Ta<Ta, where the lifetime of the diode section is τd and the lifetime of the thyristor section is τ6. This can be easily realized by thermal diffusion, and this is the greatest feature of the present invention.

Next, the effect of this embodiment on the resistance will be explained in comparison with the conventional method.

Figures 3 (-) and 4 (-) show the state of the silicon disk after the junction has been formed, before gold is deposited to shorten the lifetime of the thyristor and diode parts. .

Cyrisk Department's N1 construction lid 1st layer 1o and P! Leaving the oxide film 17 on the exposed surface of the first paste layer 11, all other parts are removed. Gold, which is a 2-if-time killer, is deposited on both sides of the silicon disk in this state using a vacuum evaporation method.
Thereafter, gold diffusion was performed for 3 hours at a temperature of 830° C. in an inert gas, and after the gold diffusion, the oxide film 17 was removed, and the silicon disk bonded by the conventional method and this embodiment was bonded. On the other hand, in order to compare the electrical characteristics of the elements, a reverse conduction thyristor was manufactured according to the procedure explained in Fig. 1, and the turn-off time of the thyristor part and the reverse recovery charge of the diode part were compared, as shown in Table 3. becomes. The current rating of this reverse conduction thyristor is based on the thyristor part,
There are 150 people in both diode sections, and the voltage rating is 8.
It is 00v.

Table 3 The measurement conditions in Table 3 are: forward current 850 people, reverse current 300A, and a1/a during commutation for turn-off time measurement.
The reverse recovery charge measurement was performed at a reverse current of 300 A and a current fall rate of 50 Vμj.

The following empirical formula approximately holds true between the turn-off time, reverse recovery charge, and the lifetime τ of the NA layer.・Lifetime of the N1 layer in the thyristor section! If the lifetime in the 111-% diode section is t1a, then the turn-off time tq:10zu11...
・・・・・・(1) Reverse recovery charge Qvr″: 1001
00y ・・・・・・・・・(2) fl), (2)
When the lifetime of the 81 layers of the thyrisk part and the diode part of the reverse conduction thyristor manufactured by both manufacturing methods is evaluated from the measurement results shown in Table 3 using the relationship of the equation, the results shown in Table 4 are obtained.

As understood from Table 4, in the bond forming method of the present invention,
With only 1 l of gold diffusion, it is possible to obtain the TIN tortoise relationship desired for a reverse conduction thyristor. As a result, it has become possible to manufacture an element with a good balance between the turn-off time of the silisc section and the reverse recovery charge of the diode section.

On the other hand, in the conventional method, even if the turn-off time of the diode part is small, the reverse recovery charge of the diode part is large. In addition, this increases the required breakdown voltage of the device, which lowers the yield of the device. Therefore, even though the device has a small turn-off time, the turn-off time performance cannot be reflected in the miniaturization of the device due to the reverse recovery charge of the diode. Such a device cannot be called a reverse conduction thyristor with a good balance between turn-off time and reverse recovery charge.

The reverse conduction risk according to the present invention can reduce the reverse recovery charge by as much as 401 compared to the conventional method, thereby reducing the capacitance of the saturable reactor to 30 and the element breakdown voltage to 25. Eliminate the harmful effects of the law,
Depending on the turn-off time of the device, the device can be made smaller and the yield of the device can be improved, leading to overall cost reduction.

In view of the above points, the present invention brings about great progress in the application technology of reverse conduction thyristors, and its practical effects are extremely large.璽$1) In order to obtain a direct relationship with fil, a P layer was formed on the P layer using boron diffusion to establish the above relationship, but it is not fixed to this, and the diode A boss is applied to the N layer surface of the section with a front page density lower than the front page density of this N layer.
Those skilled in the art can easily guess that similar effects can be obtained by diffusing 2y.

[Brief explanation of the drawing]

Figure 1 is a vertical cross-sectional view showing the structure of a conventionally commonly used reverse conduction thyristor, Figure 2 (-) is a current and voltage waveform diagram of the reverse conduction thyristor during commutation, @Figure 2 (b) #Iz diagram Enlarged view of the mu part of (U), Figure 3 (a) (b) (+1) (1) (
@) is a process diagram for explaining an example of junction formation of a conventionally used reverse conduction thyristor, and Fig. 4 (N3 (11
) (0)(, 1)(e) is a process diagram showing an embodiment of the junction forming process of the reverse conduction cyrisk of the present invention. l... Thyristor part, 2... Diode part, 3... Separation band, 4... Surface shaping part, 5...
... Silicon disk, 6 ... Supporting electrode for reinforcement,
7...4-material, 8...gate electrode, 9
...Cando electrode, 10...N1 construction ivy layer, 11...PB heath layer, 12...N
Layer, 13...NB p layer s14...pm
Emitter single layer, 15...P layer, 16...N layer, lf...oxide film, 18, 19.20...
Window opened in oxide film, 21...N'' layer, 2t...PFa. Patent applicant Toyo Denki Manufacturing Co., Ltd. Representative Atsushi Doi No./17 Figure 2 Figure 41 Procedure Written amendment (method) February 1982 Director General of the Japanese Patent Office Mr.L. Case description 1982 Patent Application No. 146351λ Name of the invention Method for forming a junction for a reverse conduction thyristor 1 Relationship between the person making the amendment and the case Patent application Postal code 104 2-7-2 Yaesu, Chuo-ku, Tokyo January 26, 1972 6 Contents of amendment (1) Description page 14, line 15 “Figure 4 is Naka” (b
)(C)(d)(e)...'' to ``Figure 4(-)(b)
XC) (d) is corrected to ``. (2) Specification cylinder page 14, line 19 to page 15, line 1 “4th
In the figure, the process in Figure 4 (1) is shown in Figures 3 (-) to (C).
), so the explanation will be omitted, and the explanation will start from FIG. 4(b). It is omitted here because it is the same as the original. (3) On page 15 of the specification, line 2, "Fig. 4 (b) wa" is corrected to "Fig. 4 (Otori) wa." (4) Correct "As shown in Figure 4 (C)..." on line 10 of page 15 of the specification to "As shown in Figure 4 (b)..." (5) Page 16 of the specification tube In the 11th line, "Fig. 4 (d)" is corrected to "Fig. 4 (C)". (6) Specification Letter 17jf Line 7, "The state shown in Figure 4 (d)..." is corrected to "The state shown in Figure 4 (C)...". ()) Page 17, line 12 of the specification is corrected to ``At the stage of Figure 4(d)...'' and ``At the stage of Figure 44(C)...''. (8) On page 20, line 5 of the specification, "Fig. 3 (-) and Fig. 4 (e)" was changed to "Fig. 3 (-) and Fig. 4 (d)".
(9) Page 24, line 17 of the specification, “No. 4
Figure (-) (b) (C) (d) (C) is ``Figure 4 (l
Xb) (C) (d) is corrected to ``. Figure 4 of the drawings will be amended as shown in the corrected drawings. HanaIL 口14 /'E Procedural amendment (voluntary) February 3, 1980 Commissioner of the Japan Patent Office Mr. 1. Indication of the case 1981 Patent Application No. 146351 2 Name of the invention Method for forming a junction of a reverse conduction thyristor 3 , Relationship with the case of the person making the amendment Patent applicant postal code 104 2-7-2 Yaesu, Chuo-ku, Tokyo (311) Toyo Denki Seizo Co., Ltd. (!) Akira #
1tlW 11th 5th day 1 380υ 1[C......] in ``830tl in inert gas''
At the temperature of...the river...'' is corrected. (2) The following corrections have been made to Table 1 on page 12 of the same. (3) Page 21, line 20 “Reverse recovery charge Qvr: 100τyvid
......(2)" to "reverse recovery charge Qrr':: 100rNid
--12) Correct to J. (4) Page 24, lines 14-15 “Figure 2 (b)
) "Figure 2 (-) enlarged view of part A" is changed to "Figure 2 (b) is the enlarged view of the
(5) Among the drawings, correct Fig. 4 to make it appear as a corrected drawing.

Claims (1)

[Claims]
A thyristor section consisting of a B-P8 structure and ? ”-P-N
4-N When forming the respective junctions of the diode portions having the current structure, the degree of disorder of the crystal structure occurring in the substrate after the formation of the junctions is greater in the diode portion than in the thyristor portion. A method for forming a joint of a reverse conducting thyristor. (2) The junction of the thyristor part, which consists of N, -P, -N↓-N, -P, is formed by forming the N1 layer with N-type impurity glue and the Ps space layer with D-type impurity gallium. Or with aluminum, N! 1. A method of bonding a reverse conduction thyristor according to claim 1, in which the P1 base layer is formed using phosphorus as an N-type impurity, and the P1 electric ivy layer is formed using antimony as an N-type impurity and boron as a D-type impurity. . (3)? The junction of the diode part, which has a structure of +-P-N↓-N+, is formed by forming the P◆ layer with Si-type impurity bones, and forming the second layer with P-type impurities such as gallium or aluminum. A method for forming a junction in a reverse conduction thyristor according to claim 1, wherein the N4 layer is formed using phosphorus as an N-type impurity.