JP3388999B2 - Setting method of breakover current or holding current in surge protection device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a surge protection device for protecting an electric circuit system from an abnormally high voltage or an abnormally large current due to various surge factors such as lightning and switching surge, and more particularly, to design a breakover current or a holding current thereof. It relates to a method for setting the specification value as close as possible.

[0002]

2. Description of the Related Art A surge protection element as a unit element has been variously devised up to now, even if it is limited to a two-terminal type.
There is a constant voltage diode type (breakdown type) surge protection element that clamps the voltage across the element to a certain breakdown voltage when it breaks down due to the application of a surge. On the other hand, in addition to simply following the breakdown mechanism, when the element breaks down due to the application of surge and the element current begins to flow, and when it increases above the breakover current value in absolute value, it exhibits a negative characteristic. There is also a breakover type surge protection device that can absorb a large current surge by breaking over and the voltage across the device transitions to a clamp voltage lower than the breakdown voltage.

The latter element has many advantages in that it consumes less electric power (heat generation) and can absorb a large surge. However, this also includes avalanche breakdown or zener breakdown as the first breakdown initiation mechanism. There are those that use the punch-through phenomenon and those that do not. As will be described later, the present invention can be applied to any device whose initial breakdown mechanism conforms to any principle as long as it is a breakover type. A breakover type surge protection element utilizing the punch-through phenomenon is advantageous in that a breakdown voltage can be obtained and various electric characteristics such as junction capacitance and resistance can be independently designed. For this reason, the applicant of the present invention has, to date, listed in the order of application, publicly known document 1: Japanese Patent Publication No. 7-77268, publicly known document 2: Japanese Patent Publication No. 1-33951, publicly known document 3: Japanese Patent Publication No. 2-52862. Publication, publicly known document 4: Japanese Patent Publication No. 4-78186, publicly known document 5: Japanese Patent Publication 6-38507, publicly known document 6: Japanese Patent Publication 6-38508, publicly known document 7: Japanese Patent Publication 6-56885 , Public Reference 8: Japanese Patent Publication No. 7-7837, Public Reference 9: Japanese Publication No. 7-70740, Public Publication 10: Japanese Patent Laid-Open No. 4-320067, Public Publication 11: Japanese Publication No. 7-93423, Public Publication Through reference 12: Japanese Patent Publication No. 7-93424, etc., various proposals have been made that are effective when applied to a breakover type surge protection device using punch-through for initial breakdown (however, the above-mentioned known reference 7 Improvements disclosed in ~ 11 include the use of avalanche and Zena yields in the initial yielding phenomenon. It can also be advantageously applied to the breakover type surge protection device used.

First, a basic structural example and operation of the punch-through type surge protection device will be described with reference to FIGS. 6 (A) and 6 (B). As shown, there is a first semiconductor region 11, generally provided as a semiconductor wafer or substrate, the conductivity type of which is
Selected as either p or n. However, in the case of the punch-through type, the n-type is preferable as shown in view of various manufacturing conditions of other regions. A second semiconductor region 12 and a third region 13 are sequentially formed on one main surface side of the first semiconductor region 11 by generally using an appropriate impurity introduction technique such as an impurity double diffusion technique or an ion implantation technique. . The second semiconductor region 12 is a rectifying junction (pn junction) with the first semiconductor region 11.
In the case shown in the figure, the p-type is selected because it is necessary to form a p-type. However, when the device is a punch-through type, it is preferable to use the p ^- type with a slightly lower concentration, that is, the p ^- type. On the other hand, the third region 13 is a second semiconductor region with respect to the second semiconductor region 12.
It is sufficient for the region 12 to have minority carrier injection properties, that is, a region capable of forming a minority carrier injection junction together with the second semiconductor region 12, for example, when the second semiconductor region 12 is n-type as shown in the figure. Can be made of silicide that can inject holes, or a metal that can inject electrons when p-type is used. However, in general, the third region 13 is also a semiconductor region.
Such an example is also shown in the illustrated case, in which the third region 13 has a conductivity type opposite to that of the second semiconductor region 12 and forms a rectifying junction (pn junction) with the second semiconductor region 12. N-type semiconductor region. However, as can be seen from an operation example described later, since the third semiconductor region 13 also forms one end of the main current line in the element after the breakover, it is desirable that the semiconductor region has a high conductivity, that is, a high concentration. It is preferably an n-type (n ⁺ ) semiconductor region. Such a point can be applied to an element to which the present invention is applied, which will be described later.

The other main surface side of the first semiconductor region 11 (in the figure,
The lower side or the back side), facing the second semiconductor region 12,
A fourth region 14 having physical properties capable of injecting minority carriers is provided in the first semiconductor region 11. Therefore, the aforementioned third area
As in 13, the first semiconductor region 11 is
It is also possible to fabricate the fourth region 14 with a metal in the case of a n-type semiconductor and with a silicide or the like in the case of an n-type semiconductor, which is also the principle in the element to which the present invention is applied. However, this is also generally a semiconductor region, and a rectifying junction (pn junction) is usually formed with the first semiconductor region 11. Therefore, also in the case shown in the figure, this fourth region 14 where a minority carrier injection junction should be formed with the first semiconductor region 11 is a p-type semiconductor region 1
4. For the same reason as for the third semiconductor region 13 described above, it is preferably a high-concentration p-type (p ⁺ -type) semiconductor region. This also applies to the element to which the present invention is applied, which will be described later. Based on these points, in this document, the regions 13 and 14, which are not limited to the semiconductor regions, will be referred to as the regions 1 to be configured as the semiconductor regions.
1 and 12 are collectively referred to simply as “area”.

In view of the structure of FIG. 6 (A), the surface of the second region 12 and the surface of the third region 13 are common through the opening formed in the insulating film 16. Is provided with a first electrode E1 which makes ohmic contact, and a second electrode which makes ohmic contact with the fourth region 14 through an opening formed in the insulating film 17.
E2 is provided. The second electrode E2 also has a first area 11
The electrode portion in contact with is also electrically connected, and an ohmic contact region 18 is formed by this and the first semiconductor region. Particularly preferably, in order to improve the ohmic contact in this portion, the ohmic contact region 18 has the same conductivity type as the first semiconductor region 11 and has a high impurity concentration (thus, an n ⁺ type region in the case shown) 19 have. The surge protection device having such a cross-sectional structure is applicable to all areas 11, 12, 13, 14,
18 (19) has a vertical stacking relationship along the thickness direction of the first region 11, and, as is clear from the operation described below, the device current as a result of absorbing the surge is also the third , The first area 11 between the fourth areas 13 and 14 is
Because it flows in the thick direction of, it is generally called vertical type or vertical type. In contrast, the fourth area 14 is the second area
There is also a lateral type or a lateral type provided on the same main surface side of the first region 11 in a side-by-side relationship with 12. Such surge protection elements are not so different in operation principle, but the present invention does not apply such a lateral type surge protection element, and a description thereof will be omitted.

Next, the surge absorbing operation of the surge protection element shown in FIG. 6A will be described. First, for convenience, the ohmic contact region shown in FIG.
18 is absent (that is, the second electrode E2 is in the fourth region 14).
Contact only the surface of). However, a surge voltage is applied between the first and second electrodes E1 and E2, which causes a pn junction (rectifying junction) between the first region 11 and the second region 12.
At the phase of applying a reverse bias to (the phase where the side of the first electrode E1 becomes negative when each region has the conductivity type shown in the figure),
If the size is equal to or larger than a predetermined size, the upper end of the depletion layer in the pn junction between the first and second regions 11 and 12 reaches the third region 13 and punch-through occurs. . It is desirable that the second region 12 is a low concentration p ⁻ type region because the extension of the depletion layer at this time is mainly directed to the third region 13.

When such a punch-through mechanism occurs, the fourth region is transferred through the minority carrier injection junction (pn junction in this case) between the fourth and first regions 14 and 11 which is forward biased at this time. Minority carriers for the first region 11 injected into the first region 11 from 14 are the third region 13
From the second region 12 is punched through to the first region 11 and a part of the different polarity carrier that has flowed into the first region 11 and then disappears.
Many reach the second region 12, which is the space charge layer,
In addition, the second region 12
As a result of the current path being established between the first electrode E1 and the first electrode E1 that is in contact with the surface of the second region 12, the third region 13 moves laterally while licking the lower surface and contacts the surface of the second region 12. It reaches the first electrode E1 which is
That is, the current resulting from absorbing the surge is the first and second electrodes E
Start flowing at 1, E2. The starting voltage of such an operation corresponds to the point shown as "breakdown voltage V _BR " on the voltage axis in FIG. 6B showing an example of the voltage-current (V-I) characteristic of the illustrated surge protection element. Breakdown voltage V _BR is often commonly referred to as the “operating voltage”.

After the minority carrier flow from the fourth region 14 is generated in this way, the second region 12 and the third region 13 are electrically short-circuited to each other on their surfaces by the first electrode E1. However, in FIG. 6 (B), as shown in the characteristic curve portion that suddenly rises in the positive direction of the current axis, the second region 12
A voltage value (voltage drop in the second region) that is the product of the current value of the element current that increases and the resistance value along the current path in the second region 12 after the current starts flowing through However, the junction turns on from the portion where the forward voltage of the minority carrier injection junction (pn junction in the figure) formed by the second region 12 and the third region 13 becomes equal, and the third region 13 Therefore, minority carrier injection for the second region 12 occurs in the second region 12. Then, the injection of minority carriers into the second region 12 results in a further increase in the device current flowing between the first and second electrodes E1 and E2, and this also causes a change from the fourth region 14 to the first region 11. Of the minority carrier injection junction is promoted, and the positive feedback phenomenon of widening the turn-on portion of the minority carrier injection junction between the second and third regions 12 and 13 is caused. When the turn-on of almost the entire surface is reached, the main current path inside the device is established, and the third and fourth regions 1
A large current can be absorbed between 3 and 14.

Therefore, in the characteristic diagram shown in FIG. 6 (B), an element having a certain value or more, which is shown as "breakover current I _BO ", between the first and second electrodes E1 and E2. When a current flows, a negative resistance characteristic occurs as a sign that a positive feedback phenomenon occurs inside the element, and the voltage across the element that appears between the first and second electrodes E1 and E2 is the voltage at the time when the breakover starts. It is possible to shift to a “clamp voltage (also simply referred to as ON voltage) V _P ”, which is lower than the value “breakover voltage V _BO ”, and lower than the breakdown voltage V _BR when the breakdown first starts. As a result, a large surge current can be absorbed while suppressing heat generation of the element.

With such a surge protection element,
The maximum current value that can be absorbed through the first and second electrodes E1 and E2 is generally called "surge withstand voltage I _PP ", and the minimum device current value required for the turned-on device to maintain its on-state. Is generally called “holding current I _H ”. In other words,
When the surge disappears and the current more than the holding current I _H stops flowing in the element, the element self-recovers (turns off) and returns to the initial state before this description. Therefore, the holding current I _H is also called “turn-off current I _H ”.

When the second region 12 and the third region 13 are not short-circuited on their surface at the first electrode E1, terminals are independently taken out from them and short-circuited outside the element. Although the above-mentioned operation basically occurs, then the time differential value (dV / dt) of the voltage at the rise of the applied surge depends on the resistance value or the inductance value expected in the short-circuit line or short-circuit means. It is highly possible that the breakdown voltage V _BR (and thus the breakover voltage V _BO ) fluctuates considerably depending on the magnitude of). In other words, if the second region 12 and the third region 13 are short-circuited on their surfaces by the first electrode E1 as shown in the figure, such a fear is reduced, and the breakdown voltage V _BR (breakover voltage V _BO ) can be stabilized.

However, from the above description, it is understood that the surge protection element shown in the figure specifies the polarity of the surge to be absorbed. That is, in the illustrated conductivity type relationship for each of the regions 11 to 14, unless the second electrode E2 has a positive polarity surge, it cannot be absorbed with breakover characteristics, and as described below, the second electrode Since E2 is in direct contact with the first region 11 via the ohmic contact region 18, it does not exhibit a significant reverse breakdown voltage against the application of a reverse polarity voltage. Simply first,
The result is equivalent to that the forward diode composed of the second semiconductor regions 11 and 12 is inserted between the first and second electrodes E1 and E2. In that sense, the illustrated device is unipolar or "unipolar", with a limited polarity of the surge that can be absorbed.
It is a surge protection element. On the other hand, a "bipolar" surge protection element capable of absorbing both surges of positive polarity regardless of the polarity of the surge, that is, either of the first and second electrodes E1 and E2 is also disclosed in the present application. Both unipolar elements in the group of publications available to humans have already been disclosed (except known documents 3, 6 and 10). However, in the present invention, such a bipolar surge protection element, more specifically, a surge protection element having a reverse breakdown voltage, is not applied, and therefore its description is omitted.

As to the surge protection element shown in FIGS. 6A and 6B, the explanation of the principle operation regarding the surge absorption is completed, and as shown in FIG. , The second electrode E2 is not only in electrical contact with only the fourth region 14, but the first region as shown by "Ohmic contact region 18" next to the fourth region 14. Explain the reason why you are in ohmic contact with 11.

This type of vertical surge protection element should not respond to a surge having an absolute value voltage equal to or lower than the breakover voltage V _{BO described} above when it is operating properly. However, in the device structure that does not have the second electrode E2 in ohmic contact with the first region 11 on the side of the fourth region 14, the voltage of the surge applied between the first and second electrodes E1 and E2 is the breakover voltage V. Even though it was within the absolute range smaller than _BO , sometimes it caused a malfunction that caused a breakover. This can be explained as follows.

First, since a junction capacitance Cj is usually expected in a pn junction formed by the first region 11 and the second region 12 and reverse biased when a surge is applied, the first and second electrodes Time differential of voltage rise between E1 and E2 is dV / dt
Is applied, a displacement current iT of iT = (dV / dt) Cj flows as a transient current that charges the junction capacitance Cj.

However, the junction capacitance Cj in the above equation often becomes considerably large when the area of each region is enlarged in order to obtain a sufficiently large surge withstand capability.
For example, a value of about 100 pF or more is usually considered. On the other hand, the nature and behavior of various surges have already been extensively studied and studied in various ways, and the results show that, for example, when a lightning surge is applied to a telephone communication line, etc. As for the peak value of the voltage value of the induced noise voltage to the circuit system, the sharpness of the surge (dV /
As for dt), a value up to about 100 V / μS is sufficiently conceivable. Therefore, as is clear by substituting these values into the above equation, the transient current value iT that charges the junction capacitance is 10 mA.
It can be a degree. The larger the dV / dt value, the larger it becomes, and in any case, the displacement current iT having a fairly large value can flow though it is instantaneous.

Further, in the surge protection element actually manufactured according to the sectional structure of FIG. 6 (A), high speed operation may be required, so that the distance between the fourth region 14 and the second region 12 may be increased. May be designed to be quite short, and such an element tends to have a too large breakover current I _BO , and in an absolute sense, the variation due to the manufacturing parameters is not sufficiently small. Depending on the case, the breakover current I
_{In some} cases, the value of _BO did not change, or even fell below the displacement current value iT at the time of surge application, which was obtained as described above.

As a result of these factors, the peak voltage value of the surge is the design breakover voltage V.
Even if it is a "small surge" that does not reach _BO and therefore does not need to be absorbed in particular, the element breaks if the rise is extremely steep and the voltage time derivative dV / dt is considerably high. In some cases, the phenomenon of causing overshooting occurred. Speaking on the characteristic diagram of FIG. 6 (B), the effective breakover voltage V _BO when such a malfunction occurs is much smaller than the value shown on the characteristic diagram (left side). It is equivalent to moving to.

On the other hand, as shown in FIG. 6A, the second electrode E2 electrically connected to the fourth region 14 is
At the same time, when the ohmic contact region 18 is provided by electrically contacting the main surface of the first region 11 in the vicinity of the fourth region 14, the first region 11 and the second region 12 are reverse biased. Even when a surge having the polarity to be applied is applied and therefore the minority carrier injection junction between the first region 11 and the fourth region 14 is in a forward biased relationship, a forward voltage is applied to the junction before it turns on. Since a majority carrier for the first region 11 can be poured into the first region 11 from the second electrode E2 via the ohmic contact region 18, a pn junction composed of the first region 11 and the second region 12 is formed. Junction capacity of Cj
Can be quickly charged, and thus the dV / dt resistance can be enhanced.

At this time, that is, before the minority carrier injection junction composed of the first region 11 and the fourth region 14 is still turned on, the third region 13 and the ohmic contact region 18 are connected to each other. The situation when a majority carrier flow (electron flow flowing from the third region 13 into the ohmic contact region 18 through the second region 12 in the case of the conductivity type shown in the figure) to one region 11 is flowing is shown in FIG. However, I will come back to this later.

However, such an ohmic contact region
The provision of 18 makes it possible to obtain a surge protection device that does not respond to "small surges", while at the same time, injecting majority carriers into the initial first region 11 for such junction capacitance charging. The phenomenon did not have a bad influence on the basic operation itself after the occurrence of the above-mentioned breakdown phenomenon. After the punching through of the first region 11 and the third region 13 as described above, the current due to the majority carriers increases, and the voltage drop in the fourth region 14 mainly in the thickness direction (depth direction) causes the fourth region 14 to fall. When the forward voltage of the minority carrier injection junction between the region 14 and the first region 11 becomes equal to this junction, the junction turns on, and from then on, minority carriers for the first region 11 begin to be injected from the fourth region 14. ,
This is because the breakdown from the device to the breakover can be performed by the mechanism already described. Also,
After the breakover, the main current path of the device current between the first and second electrodes E1 and E2 is not the path through the ohmic contact region 18 between the second electrode E2 and the first region 11, but the third region 13
And a path through the fourth region 14, which is almost equivalent to the state in the device that does not have the ohmic contact region 18.

Up to now, the punch-through type conventional element has been described. However, even if the structure shown in FIG. 6A is almost the same as the structure shown in FIG. According to the knowledge of the second area 12 and the third area 13
If the geometrical parameters and impurity concentration parameters of each region are appropriately selected in addition to increasing the thickness of the avalanche, the initial mechanism of the yield initiation is the avalanche breakdown and zener between the first and second regions 11 and 12. It has been found that a surge protection element having a positive feedback mechanism similar to that of the punch-through type described above can be manufactured by utilizing the breakdown and the breakover. Then, if such a surge protection element or, in addition, another known surge protection element is of a type that breaks over through the positive feedback phenomenon accompanying the injection of minority carriers, the above-mentioned " Problems with response to "small surges" may occur as well, so
It has also been found that the above measures against it can also be applied to them. However, as avalanche breakdown and zena breakdown are generally called "point phenomenon" (local phenomenon), the point where breakdown begins or the point where the electric field concentrates after breakdown tends to be local, so surge It is difficult to obtain a large withstand voltage I _PP , which is disadvantageous as compared with the punch-through type element described above, and also has a small degree of freedom in design and poor tolerance to manufacturing parameters. However, if we do not compare such advantages and disadvantages but only consider the dV / dt resistance, which is the problem here, the above discussion is almost applicable to such avalanche breakdown type and zener breakdown type surge protection devices. In fact, the invention described below is likewise applicable to such unipolar surge protection devices.

However, in the surge protection device utilizing the avalanche breakdown or the zener breakdown between the first and second regions 11 and 12 in the initial breakdown mechanism, the above-mentioned known reference 8 has been described in FIG. 6 and its description. As disclosed, the first and second areas 11, 12
If a plurality of impurity regions having the same conductivity type as that of the first region 11 and a higher concentration are formed in the contact area region of, the avalanche breakdown or the zener breakdown is caused despite the use of the local phenomenon. This is desirable because it can occur simultaneously in the plurality of high-concentration impurity regions, and the current distribution inside the device can be made uniform as a whole. Of course, the invention is likewise applicable to such devices.

Further, regardless of the initial breakdown mechanism, in the practical structure, the second region 12 is also used for the purpose of making the current distribution inside the element uniform and, as a result, increasing the surge withstand value I _PP , for example. Third area 13 provided inside
Alternatively, the fourth region 14 forming the minority carrier injection junction with the first region 11 may be arranged in parallel. A unipolar surge protection element that employs such a device is also a target to which the present invention is applied, as can be seen in a preferred embodiment of the present invention described later.

By the way, in the surge protection device of the kind described above, there are various kinds of characteristic specification values required for supplying to the market, but the most important ones are the breakover current I _BO and the holding current I. There is _H. However, since these parameters fluctuate because various parameters such as thickness, area size, impurity concentration, and resistivity of each region are intricately intertwined with each other, it is necessary to cut at the manufacturing site in order to obtain the values according to the design specifications. In many cases, I had to rely on And Try. Therefore, the present applicant has found that the above-mentioned known document 10: JP-A-4-32
In the publication, paying attention to the correlation between the lateral cross-sectional dimension of the ohmic contact region 18 and the thickness hc of the fourth region 14 shown in the figure in the structure shown in FIG. The dimension range of the ohmic contact region 18 that does not change the target design specification value of the breakover current I _BO or the holding current I _H has been disclosed. Simply put, if the cross-sectional dimension of the ohmic contact region 18 is made wider than a predetermined dimension specified by the relationship with the thickness hc of the fourth region 14, when the ohmic contact region 18 is actually manufactured. Even if some dimensional fluctuations occur, the breakover current in the manufactured device
This means that I _BO and holding current I _H do not have to undergo too large fluctuations.

Further, in this known document 10, as a problem independent of the cross-sectional width dimension of the ohmic contact region 18, the fourth region
Regarding the cross-sectional width dimension 14 of itself, it was also disclosed that the larger the break width I _BO or the holding current I _H, the lower the tendency. If the cross-sectional width of the fourth region 14 becomes wider, the path of the majority carrier flow that forward-biases the minority carrier injection junction constituted by the fourth region 14 and the first region 11 becomes longer, and the order of this junction becomes faster. This is because the direction ON voltage is reached.

[0028]

Certainly, the disclosure according to the known document 10 is significant in that it gives a useful design standard in view of the design circumstances before that. However, it has been found from the subsequent research that the disclosure in the known document 10 does not always apply. That is, first, in the known document 10, as described above, the thickness of the fourth region 14
It was analyzed as hc cannot be ignored. for that reason,
As shown in FIG. 6A, when the ohmic contact region 18 has the high-concentration impurity region 19, the thickness hc of the fourth region 14 is considered to be almost zero, or a negative sign is given. There was a situation in which it had to be calculated as the dimension value that it had, and there was a situation in which it could not be applied. That is, the effective thickness hc of the fourth region 14 is the dimension of the portion protruding from the interface between the region 19 and the first region 11 to the inside of the first region 11 when the high-concentration impurity region 19 is present. Therefore, if the thickness of the high-concentration impurity region 19 is about the same as the thickness hc of the fourth region 14, the thickness hc of the fourth region 14 must be almost zero. On the contrary, when the thickness of the high concentration impurity region 19 is thicker, the effective thickness of the fourth region 14 becomes a value having a negative sign. In this, the known document 10
, For example, the most basic expression regarding the cross-sectional width dimension xs of the ohmic contact region 18: xs> αhc (0.5 ≦ α
Also, ≦ 1) cannot be adopted (the symbol Ls is used for the cross-sectional width dimension of the ohmic contact region in the present invention described later).

Further, even if the thickness hc of the fourth region 14 takes a significant positive value, when the width dimension of the fourth region 14 becomes large, the thickness hc becomes substantially negligible. Often. Even in such a case, the above equation disclosed in the known document 10 cannot be adopted.

In the end, in the known document 10, the thickness hc of the fourth region 14 cannot be ignored, in other words, when the width of the fourth region 14 is not so large as compared with the thickness, the thickness hc of the fourth region 14 is not so large. The majority carrier flow in the first region that forward biases the minority carrier injection junction composed of the four regions 14 and the first region 11 is such that the interface between the thick region of the fourth region 14 and the first region 11 is licked. This is because it was thought that the voltage drop in the thick portion hc cannot be ignored.

Certainly, as already mentioned, there are actually situations where this argument holds, and within the scope of the known document
The method disclosed in 10 is significant. However, as I touched on a little bit earlier, the main part is enlarged in FIG. 5, and as shown schematically, whether avalanche breakdown or zener breakdown occurs between the first and second regions 11 and 12 due to the application of surge. Or, even if the first and third regions 11 and 13 are punched through between the first and third regions 11 and 13 to start breakdown as in the above-described desirable device operation, minority carrier injection composed of the first region 11 and the fourth region 14 is performed. Before the junction is still turned on, the third region 13 and the ohmic contact region 18
The majority carrier flow for the first region 11 flowing between the first region 11 and the second region 1
2, the electron flow flowing into the ohmic contact region 18 via the first region 11) is considered to be almost uniform in the vicinity of the interface between the second region 12 and the first region 11, but the fourth region 14 should be avoided. As a result, there is no choice but to flow toward the ohmic contact region 18, and as a result, the upper portion of the fourth region 14 has an apex angle as shown schematically.
A "blank" portion having a triangular shape with a cross section of 2θ is generated, while a concentrated portion of the majority carrier flow is generated in the ohmic contact region 18 as shown. In other words, in terms of current density J,
Than in the vicinity of the interface between the second region 12 and the first region 11,
The current density of the portion flowing into the ohmic contact region 18 becomes high.

Therefore, if the width dimension of the ohmic contact region 18 is set within the range shorter than the width of the majority carrier concentrated portion, the change in the width dimension due to a slight manufacturing error causes a large breakover current I _BO . It appears as an error,
Correspondingly, the holding current I _H also fluctuates greatly. On the contrary, within a range wider than the width of the majority carrier concentrated portion, even if the size of the ohmic contact region 18 actually manufactured is slightly changed with respect to an arbitrary design size, the characteristics are hardly affected. You don't have to give it. The idea in the known document 10 is just like this.

However, when the thickness of the fourth region 14 is negligible or must be considered as a value having a negative value, and when the first region 11 having a certain thickness or more is used. , This blank area becomes quite large, and the majority carrier flow in the first area 11 becomes the fourth area.
Rather than licking the interface between 14 and the first region 11, they tend to flow rather apart. The voltage drop along the thickness of the fourth region 14 may not be great. Unfortunately, the analysis in such a case has not yet been made at the time of the disclosure of the known document 10, and as a result, as described above, the breakover current I _BO or the holding current irrelevant to the thickness hc of the fourth region 14 is obtained. The design criteria for I _H could not be derived.

The present invention has been made under such circumstances. First, as a basic purpose, the thick hc of the fourth region 14
It is intended to disclose a more general setting method of these characteristic values that can set the breakover current I _BO or the holding current I _H of the surge protection element to be manufactured without using the numerical information of.

In addition, the resistivity ρ of the first region 11 depends on, for example, the operating voltage (breakdown voltage) of the device to be manufactured.
It is another object to provide a method capable of setting the breakover current I _BO or the holding current I _H in consideration of the influence of this resistivity parameter even when it is necessary to change.

The change in the value of the holding current I _H is almost proportional to the change in the breakover current I _BO . Therefore, setting the breakover current I _BO to a certain value means substantially setting the holding current I _H to a certain value, and vice versa. In other words, setting the breakover current I _BO “or” holding current I _H to a certain design specification value, as referred to in this document, ultimately means that the breakover current I _BO “and” holding current I _{H H}
Is substantially synonymous with setting both to a certain design specification value.

[0037]

In order to achieve the above object, the present invention has a surge protection element having a basic structure as described with reference to FIG. 6 (however, as described above, this is also the third region. 13, one or both of the fourth region 14 also includes a case where a plurality of one or both of the fourth region 14 are arranged in parallel while being separated from each other in one cross section along one in-plane direction with respect to the main surface of the substrate)
In the first, the plane shape of the fourth region is a rectangle having a short side and a long side, and the ohmic contact region has one of the fourth region or one of the plurality of fourth regions on one side or both sides in the short side direction thereof. When provided along the long side direction, (a) Lc is the dimension of the short side of each of the one fourth region or multiple fourth regions, and the dimension of the long side is Lc or more. (B) Regarding the dimension along the short side direction of the ohmic contact region on both sides in the short side direction of each fourth region, the dimension on one side on both sides of the fourth region is kLs (0 ≦ k ≦ 1), and the other side. The dimension on the side of is (1−k) Ls (the sum of them is Ls); (c) By changing (Lc + Ls) / Lc ² , the breakover current or holding current is set in proportion to this. Break-down in surge protection devices characterized by: We propose a method of setting bus current or the holding current.

Further, according to still another aspect of the present invention, the first semiconductor region and the fourth region are formed so that the same design reference can be obtained even when devices having different operating voltages (breakdown voltages) are produced. The resistivity in the first semiconductor region along the path of the majority carrier flow for the first semiconductor region flowing between the third region and the ohmic contact region to turn on the minority carrier injection junction
Instead of (c), by changing (c) '{(Lc + Ls) / Lc ² } (1 / ρ), the breakover current or holding current is set in proportion to this; Also suggest.

In the above structure, it is desirable that the value of k is 0.5 in order to make the current distribution in the device uniform. This means that the ohmic contact regions each having the same width (Ls / 2) in the short side direction extend in the long side direction along both side edges of the fourth region (if there are a plurality, respectively). It is effective in satisfying geometric symmetry.

In the present invention, instead of the above group of constituent elements, in the case where the fourth region has a shape other than a rectangle and is surrounded by an ohmic contact region, that is, each plane of one or a plurality of fourth regions In the case where the shape is a square or a substantially square, or a circle or a substantially circle, or an outer contour envelope shape is a circle or a substantially circle, and the ohmic contact region is provided to surround the polygon, (d) The area of each of the fourth regions or the plurality of fourth regions is Sc; (e) the area of the ohmic contact region surrounding the periphery of each fourth region is Ss; (f) ( Sc + Ss) / Sc by ^3/2 varying the, this proportion is not in the breakover current or to set the holding current; break in the surge protection device, wherein the over current or the holding current To propose how to set up.

In this case as well, minority carrier injection constituted by the first semiconductor region and the fourth region is performed so that the same design reference can be obtained even when devices having different operating voltages (breakdown voltages) are produced. The resistivity in the first semiconductor region along the path of the majority carrier flow for the first semiconductor region flowing between the third region and the ohmic contact region for turning on the junction is defined as ρ, and instead of the above configuration requirement (f). In addition, by changing (f) '{(Sc + Ss) / Sc ^3/2 } (1 / ρ), the breakover current or holding current is set in proportion to this, and a method characterized by .

In any of the configurations of the present invention, the ohmic contact region has the same conductivity type as the first semiconductor region in the portion in contact with the first semiconductor region, but has a higher impurity concentration than the first semiconductor region. It may have a high-concentration impurity region, which is desirable for obtaining good ohmic contact. At this time, when the thickness of the high-concentration impurity region is about the same as that of the fourth region, or conversely, when the thickness of the fourth region becomes thicker, the thickness of the fourth region may be substantially zero to a negative value. , Even in such a case, the above-mentioned known document 10
Unlike the conventional method according to the present invention, since the thickness parameter of the fourth region is not involved in the present invention, there is no problem and a more general design standard can be given.

[0043]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows a cross-sectional structure of a main part of an example of a two-terminal breakover type vertical surge protection device to which the present invention can be applied. The basic structure of this element is the same as that of the element already described with reference to FIG. 6 (A), and its current-voltage characteristic is also as shown in FIG. 6 (B). Therefore, the contents already described with reference to FIG. 6 can be applied to the following description almost as they are, and the same reference numerals denote the same or similar components, and some may not be described again. Further, the initial breakdown phenomenon is not limited, and for the reasons described above, it is preferable to use the punch-through phenomenon between the first and third regions 11 and 13, and it is assumed that such a device is used here as well. As explained, the first and second areas 11,
An element of the type that utilizes avalanche breakdown and zener breakdown of the pn junction formed by 12.

However, in the comparison on the drawing, FIG. 6 (A)
In the element cross-sectional structure shown in FIG. 1, in contrast to the basic structure shown in FIG. 1, the fourth region 14 is formed in order to make the current distribution (carrier flow distribution) inside the element uniform, as already described. It is composed of a plurality of elements arranged in parallel with each other, and an ohmic contact region 18 is formed between them. Further, here, the plane shape of each fourth region 14 has a short side and a long side, and a cross section along the short side direction is visible in the illustrated cross section, and the long side direction orthogonal to this is the front and back of the drawing paper. It will be the direction to exit. The ohmic contact region 18 extends along the long side direction of the fourth region located on both sides or the fourth region located on one side. Actually, the fourth region 14 and the ohmic contact region 18 are long sides with respect to the short side dimension. The dimension in the direction is quite large, and it is often patterned in stripes rather than rectangles. This is especially true when the desired dimensional relationship for the third region 13 is met in order to increase the dV / dt resistance in accordance with the teaching of the cited document 8 and in response thereto the planar shape of the fourth region 14 and ohmic contact region 18 Is often the case. However, since the present invention is not directly related to this point, further detailed description will be omitted.

Furthermore, each ohmic contact region 18 has an electrode portion continuous with the second electrode E2 in contact with the surface of the fourth region 14 (here, the second electrode E2 is “solid” with respect to the back surface of the first region 11). As a result of being formed by vapor deposition or the like, a high-concentration impurity region 19 of the same conductivity type as that of the first region 11 is formed in order to bring the first region 11 into good ohmic contact with the electrode portion that can be seen as the second electrode itself). have. Therefore, the interface of the ohmic contact region 18 is substantially the high-concentration impurity region.
It can be considered as the boundary between 19 and the first area 11.

Here, the dimensional relationship of the fourth region 14 and the ohmic contact region 18 is defined. First, as shown on the left side of FIG. Let Lc be the dimension of each of the regions 14 in the illustrated cross section, that is, the short side dimension, and Ls be the dimension of each ohmic contact region 18 in the illustrated cross section (also the short side dimension), and let the sum thereof be the unit length Lu for convenience. However, each ohmic contact region 18 sandwiched between the pair of adjacent fourth regions 14, 14.
As will be apparent from the explanation of the majority carrier flow in the first region described later, the right half of the ohmic contact region 18 becomes the ohmic contact region associated with the fourth region 14 located on the right side, and the left side. Half of the area is related to the fourth area 14 located on the left side.

Therefore, in order to maintain the geometrical symmetry and the uniformity of the current distribution or the carrier flow distribution inside the element, as is most commonly considered in the actual element, the short side of each fourth region 14 is considered. When the dimensions Lc are all the same, and the short-side dimension Ls of the ohmic contact region 18 sandwiched between them is also the same, when considering each fourth region 14 as the center, In the ohmic contact regions located on both sides of the four regions 14, half of each short side dimension Ls
Ls / 2 becomes the ohmic contact region 18 associated with each of these respective fourth regions 14. Therefore, it can be considered that the unit length Lu described above is constituted by a dimension portion of Lu = (Ls / 2 + Lc / 2) × 2, as shown in the central portion of FIG. In the following, first, a case where the present invention is applied according to this dimensional condition will be described.

However, with the surge having a specific polarity between the first and second electrodes E1 and E2, that is, when the respective regions have the illustrated conductivity type, the second electrode E2 has a positive polarity, the above-mentioned breakdown voltage.
When a surge of V _BR or more is applied, this also punches through between the first and third regions 11 and 13 via the second region 12 as described above with reference to FIG. The majority carrier flow (electron flow in this case) for the first region 11 injected into the first region 11 through the second region 12 should avoid the fourth region 14 as schematically shown in FIG. Ohmic contact area 18
Flow into.

Regarding the situation at this time, FIG.
(This corresponds to an enlarged view of the main part in the vicinity of the center in FIG. 1, but the high-concentration impurity region 19 is not shown in the figure: it is better if the high-concentration impurity region 19 is actually present or not. ) Again, first, the breakover current I _BO and the holding current are changed by changing the short side dimension or the short side width Lc of the fourth region 14.
It can be seen that I _H changes. That is, the width Lc of the fourth region 14
Becomes larger, the vertices of the "blank" area described above become the fourth area 14
The apex angle 2θ tends to become narrower away from, and conversely, the path Le of the majority carrier flow tends to become longer. Therefore, even if the resistivity ρ along the route Le in the first region 11 is constant, the resistance value of the route itself becomes high, so that the minority carrier composed of the fourth region 14 and the first region 11 The injection junction is more likely to be forward biased. Therefore, in other words, as the width dimension Lc of the fourth region 14 increases, the breakover current I _BO and the holding current I _H tend to decrease.
If this junction is turned on, by the mechanism already described,
This is because the positive feedback phenomenon is generated all at once. Although it is a completely schematic diagram, in FIG. 5, the majority carrier flow goes from the point a at the surface potential of 0.0 V on the surface of the ohmic contact region 18 toward the inside of the first region 11 and away from the fourth region 14. It is schematically shown that the potential difference gradually increases at various points b to e along the path Le of, and, for example, the fourth region 11
As shown in the figure, when the minority carrier injection junction is a pn junction in the p-type semiconductor region, the point at which the forward on-state voltage of this pn junction is generally 0.5 V or higher is shown as point f.

In order to analyze this in a little more detail, but for simplification, consider the upper left half of the fourth region 14 where the majority carrier flow density (substantially the same as the current density), which is the electron flow. Therefore, the reference numeral J is used), and it can be considered that it is substantially uniform in the dimensional range along the cross section in the drawing in the vicinity of the interface between the second region 12 and the first region 11. That is, Ls in FIG.
Within the size range of / 2 + Lc / 2, the density distribution of the majority carrier flow for the first region 11 exiting the second region 12 can be regarded as uniform. However, in the density J of the majority carriers injected into the portion located immediately above the fourth region 14, that is, the total majority carrier flow J injected in the size range of Ls / 2 + Lc / 2,
The portion of J · {Lc / (Lc + Ls)} flows so as to avoid the fourth region 14, and its route Le can be approximated as Le = (Lc / 2) / sin θ from the geometrical relationship in FIG.

Therefore, the majority carrier component J · {Lc / (Lc + Ls)} along the lengthened route Le first forward biases the first and fourth inter-region minority carrier injection junctions.
As a result, Vf is generated, so that the first region 11 in the route Le is eventually
Let ρ be the resistivity of, then Vf = J · {Lc / (Lc + Ls)} · {(Lc / 2) / sin θ} · ρ ....

On the other hand, since the majority carrier flow density J actually increases from the zero state at the beginning with the application of surge,
From the above equation, the larger the Lc ² / (Lc + Ls), the earlier the forward voltage Vf is reached. In other words, the breakover current I _BO or the holding current I _H proportional thereto is reduced. Therefore, conversely, the breakover current
From the viewpoint of I _BO or holding current I _H , the following equation holds. I _BO , I _H ∝ Vf ・ {(Lc ＋ Ls) / Lc ² } (1 / ρ) ・ ・ ・ ・

Therefore, as is clear from this equation, unlike the description in the above-mentioned known document 10, it is a generalized fact that it is not necessary to take the thickness hc (FIG. 6) of the fourth region 14 into consideration. , Breakover current I _BO or holding current I _H
First, it can be seen that it is not simply inversely proportional to the short side width Lc of the fourth region 14, but is proportional to (Lc + Ls) / Lc ² ...

To confirm the validity of such knowledge,
The applicant of the present invention produced many device samples and conducted experiments, and obtained good matching results. FIG. 2 shows the results of about two of them, but the sample group 1 shown by a white circle also has a sample group 2 shown by a black circle.
However, when (Lc + Ls) / Lc ² was changed linearly, the holding current I _H of the fabricated device could also be changed in proportion to this. Moreover, this result was independent of the absolute value of the dimension Lc of the fourth region 14. Therefore, if necessary, by independently designing the area size of the fourth region 14, it is possible to obtain the desired surge withstand amount I _PP that generally increases in proportion to the area, while still effectively applying the present invention. Thus, the holding current I _H or the breakover current I _BO can be set to a desired value. Of course, as described above, the holding current I _H and the breakover current I _BO are generally linearly proportional, so that FIG. 2 shows a proportional relationship of (Lc + Ls) / Lc ² with respect to the breakover current I _BO if the scale values are not used. You can see.

Further, as also shown in FIG. 2 and as understood from the above formula, the holding current I _H also changes in inverse proportion to the resistivity ρ along the path Le of the first region 11. That is, in proportion to {(Lc + Ls) / Lc ² } (1 / ρ) ..., Breakover current I _BO or holding current I _H
Can be set. Further, even if the other parameters are the same, as the thickness t of the first region 11 becomes thicker, the path Le becomes longer and the resistance value along the path Le becomes higher, so that the holding current I _H
Is also inversely proportional to the thickness t of the first region 11. This is also the result of the applicant's experiment, which is in good agreement.

By the way, the above equation is the sum of the respective dimensions Ls / 2 involved in the fourth region 14 in the ohmic contact regions 18 located on both sides in the short side direction of one fourth region 14.
It is a function of Ls. This means that ohmic contact regions 18, 18 of equal width on both sides of the fourth region 14 are provided.
But there are generally ohmic contact regions 18,18 having kLs (0 ≦ k ≦ 1) on one side and (1−k) Ls short-side dimensions on the other side. formula,
, Holds. This is because the density J of the majority carrier flow involved in the above-mentioned forward bias also changes proportionally according to their size ratio. Therefore, k is zero, that is, the fourth region 14 of interest is located, for example, at the outermost side of the plurality of juxtaposed regions, and the ohmic contact region 18 does not exist outside thereof. When a device is produced, the dimension related to the outer ohmic contact region 18 is considered to be kLs and regarded as zero, and the other (inner) short side direction dimension is regarded as (1−k) Ls = Ls. You can do it. The same is true when k is 1. Even in such a case, a result that is in good agreement with the experimental result shown in FIG. 2 is obtained. Therefore, as described above, it is considered that the ohmic contact regions 18 and 18 each having the dimension Ls / 2 exist on both sides of one fourth region 14, respectively, although it is a desirable symmetrical structure as an actual device, In a sense, it is a special case of k = 0.5, and generally k can take any value from 0 to 1.

As a matter of course, the fourth region 14 is not a plurality of juxtaposed structures but is configured as a single region, and has an ohmic contact region with a dimension kLs on one side and a dimension (1-k) Ls on the other side. Not only when there is an ohmic contact region of equal size Ls / 2 on both sides, but also when there is an ohmic contact region on only one side. .

However, the short side dimension of the ohmic contact region 18
Regarding Ls, as is taught in the above-mentioned publicly known document 10, the upper limit is generally considered, but
Considering the operating current that generally flows in the first region 11 in the thickness direction, the extent of the lateral diffusion is generally estimated, and the dimensions kLs and (1−k) Ls of the ohmic contact region 18 are The upper limit (after all, the longer upper limit) is the first area
The distance between the fourth region 14 and the second region 12 along the thickness direction of 11 or less, or at least less than or equal to the thickness of the first region 11, it is preferable that the first region 11 The mixed environment of majority carriers and minority carriers inside can be homogenized, and stable device operation can be expected in the end, and surge withstand I _PP does not have to be reduced.

Similar to the above, FIGS. 3 (A) and 3 (B) show examples of desirable sectional structures to be described in connection with the present invention. In the configuration of FIG. 1, the fourth region 14 and the ohmic contact region 18 were in direct contact with each other at their side edges. However, it does not have to be the case, and may be as shown in FIG. 3 (A). That is, an insulating film 17 ′ having a small dimension is arranged on the back surface of the first region 11 at the lateral boundary portion between the fourth region 14 and the first region 11 along the direction in which the two regions 14, 11 are arranged in parallel. A contact opening for electrode formation is opened so as to remain, and as a result, the electrode portion E2 that makes ohmic contact with the fourth region 14 and the first region 11 are formed.
The electrode portion which makes ohmic contact with is electrically connected to each other so as to extend over the insulating film 17 '.
Of course, the structure of FIG. 1 is easier to manufacture and is generally used. However, in the case of the cross-sectional structure of FIG. 3A, the distance Dd in the in-plane direction between the ohmic contact region 18 and the fourth region 14 with respect to the first region 11 is the fourth region 14 and the first region. If it is shorter than the lateral extension distance of the depletion layer formed by the junction with 11, the fourth region 14 and the ohmic contact region 18 are substantially in direct lateral contact, as in FIG. Will not change. Even if it is further apart, the principle of the present invention can often fulfill this, but if it is too far apart, there is a problem in obtaining the concentration effect of the majority carrier flow for the first region 11 flowing beside the fourth region 14. It may occur, and it is not desirable in terms of increasing the area occupied by the device itself.

Rather, when actually manufacturing the device, as shown in FIG. 3B, the high-concentration impurity region 19 is provided in the ohmic contact region 18, and the high-concentration impurity region 19 is slightly formed in the fourth region. It often has a cross-sectional shape that laterally cuts into the 14 side. When viewed in cross section, the high-concentration impurity region 19 and the fourth region 14 have portions overlapping each other in their side edge regions. In such a case, the respective effective side edges serving as a reference for measuring the above-described dimensions Lc and kLs or (1-k) Ls are, as shown in FIG. The lateral contact boundary between the high-concentration impurity region 19 and the fourth region 14 when seen in a plane projection from the side may be used, and the above-mentioned expressions (1) to (8) can be applied almost as they are. Regarding the thickness, as described above, the thickness of the high-concentration impurity region 18 may be about the same as the fourth region 14, or may be thicker than the fourth region 14, and the cross-sectional dimensions Ls, Lc Also for that of the ohmic contact area 18
Ls is often longer.

Further, when the fourth region 14 is formed from a plurality of the fourth regions 14 arranged in parallel in the short side direction as shown in FIG. 1, it is at least in the plane of the first region 11 as shown in the sectional view. It suffices if this is the case in the cross section along the direction, and for example, the ends in the long side direction orthogonal to the illustrated cross section may be connected to each other.

The present invention is not limited to the case where the fourth region 14 is a rectangle having a short side and a long side as described above, and the ohmic contact region 18 is provided only along the long side. The planar shape of the fourth region 14 is square or almost square,
Or, it is a circular shape or a substantially circular shape, or a polygonal shape in which the envelope shape of the outer contour is a circular shape or a substantially circular shape, and is one in one cross section along one in-plane direction of the first region 11, or they are parallel to each other while being separated from each other. Even if the ohmic contact region 18 is provided so as to surround the fourth region, the breakover current I _BO or the holding current of the surge protection device to be fabricated is We present a formula equivalent to the above formula that is useful for setting I _H.

Even if the plane shape of the fourth region 11 is as described above, the cross-sectional dimension of the fourth region when the cross-section along one in-plane direction of the first region 11 is taken is the square of the area Sc. Proportional to Sc ^1/2 . Therefore, similar to the discussion about the short side dimension Lc of the rectangular fourth region 14, when the cross-sectional dimension of the fourth region 14 increases,
In other words, as the number Sc ^1/2 increases, the path Le of the majority carrier flow
Becomes longer and the resistivity ρ along this path Le in the first region 11 is
Since the resistance value of the path itself becomes high even if is constant, the minority carrier injection junction formed by the fourth region 14 and the first region 11 is likely to be forward biased. From this, first, the breakover current I _BO and the holding current I _H are proportional to 1 / Sc ^1/2 .

On the other hand, the area of the fourth region 14 is set to Sc as described above.
Further, assuming that the area of the ohmic contact region 18 surrounding the fourth region 14 is Ss, the allocation of the current distribution (majority carrier flow distribution) J is advantageous in terms of area ratio as described with reference to FIG. The current component contributing to forward biasing the minority carrier injection junction by the fourth and first regions 14 and 11 is Sc.
Since the ratio is / (Sc + Ss), the breakover current I _BO
The holding current I _H is proportional to its reciprocal (Sc + Ss) / Sc.
As a result, in combination with the above, by changing (Sc + Ss) / Sc ^3/2 ..., It is possible to set the breakover current I _BO or the holding current I _H in proportion to this.

Similarly, if the resistivity along the path Le in the first region 11 is ρ, the breakover current I _BO or the holding current I _H is inversely proportional to the resistivity ρ as described above. Since it can be set, an expression such as {(Sc + Ss) / Sc ^3/2 } (1 / ρ) ... Can also be listed as one embodiment of the present invention in addition to the above expression.

In this way, even when the ohmic contact region 18 is provided so as to surround the periphery of the fourth region, the first region 11 is provided for the same reason as described above regarding the cross-sectional dimensions kLs and (1−k) Ls. The upper limit of the cross-sectional dimension for each one side of the ohmic contact region 18 located on both sides of the fourth region 14 in one cross section along one in-plane direction, the fourth region 14 and the second region 12
Or at least the first area 11
It is better to be less than the thickness.

Of course, even when the ohmic contact region 18 surrounds the periphery of the fourth region 14, the ohmic contact region 18 has the same conductivity type as the first region 11 in the portion in contact with the first region 11, but It is desirable to have a high-concentration impurity region 19 having a higher impurity concentration than the region 11, and FIG.
As described in (B), when the high-concentration impurity region 19 and the fourth region 14 which surrounds the high-concentration impurity region 19 have portions overlapping each other in their peripheral regions, each of the above areas
Each effective peripheral part that is the calculation standard of Sc, Ss is
It may be set as a lateral contact boundary between the high-concentration impurity region 19 and the fourth region 14 when viewed in plan view from the second region side.
Conversely, when a lateral gap Dd is formed between the high-concentration impurity region 19 or the ohmic contact region 18 and the fourth region 14, the separation distance Dd between them is formed by the junction between the fourth region 14 and the first region 11. It is better to make it shorter than the lateral extension distance of the depletion layer.

FIG. 4 shows another example of the structure of the surge protection element which is preferably applied to the present invention. The same reference numerals as those used in the above description correspond to the respective constituent elements described so far, and thus the description of each element will be omitted. Further, in order to simplify the drawing, the first electrode E1 is not shown in the plan view of the device. In this structure, not only the rectangular fourth area 14 but also the rectangular third area 13
Also, according to the teaching of the above-mentioned known document 11, the plurality of third and fourth regions 13 and 14 are arranged side by side in a plan projection manner. Are intersected with each other, and in the illustrated case, they are orthogonal to each other. In this case, the carrier flow inside the device is as if it were "shower" or "sprinkler", and the distribution is more uniform, which is desirable. Incidentally, FIGS. 1 and 5 can also be regarded as a cross-sectional structure along one long side direction in the third region 13 provided in plural in the element shown in FIG. However, as a general rule, the plurality of third and fourth regions 13 and 14 may be arranged in parallel with each other.

Although the present invention has been described in detail above, when the conductivity type of all regions of the element to which the present invention is applied is opposite to the conductivity type shown in the figure, the flow of the majority carrier flow for the first region 11 will be described. The direction of is opposite to the direction of the arrow shown in FIGS. In other words, according to the present invention, when the conductivity types of all the regions are the opposite conductivity types, the third and fourth regions are also provided.
As described above, as long as 13 and 14 are regions having physical properties capable of forming a minority carrier injection junction with the second region 12 and the first region 11, respectively, they can be applied even if they are not semiconductors.

[0070]

According to the present invention, a two-terminal breakover type vertical surge protection device is provided on the back surface side of the first region and the minority carrier in order to enhance the dV / dt resistance. In a surge protection element having an electrode structure having not only a region for forming an injection junction (fourth region) but also a portion in direct ohmic contact with the first region,
Regardless of the thickness of the fourth region, the breakover current I _BO or (and) the holding current I _{H is} more commonly formed.
It is possible to provide a method capable of setting.

[Brief description of drawings]

FIG. 1 is a diagram illustrating an example of a basic structure of a two-terminal breakover type vertical surge protection device to which the present invention can be applied, a definition regarding dimensions of each region, and a majority carrier flow in a first region.

FIG. 2 is a characteristic diagram showing a variation example of a holding current caused by applying the present invention.

FIG. 3 is an explanatory diagram regarding a mutual positional relationship and a shape relationship between an ohmic contact region and a fourth region.

FIG. 4 is a schematic configuration diagram of another configuration example preferable as a surge protection element to which the present invention can be applied.

FIG. 5 is a schematic explanatory diagram regarding a majority carrier flow in the first region in which the first region and the fourth region are forward biased.

FIG. 6 is an explanatory diagram regarding a cross-sectional configuration and operation in a basic example of a surge protection element to which the present invention can be applied.

[Explanation of symbols]

11 First semiconductor region, 12 Second semiconductor region, 13 Third domain, 14 Fourth Area, 18 Ohmic contact area, 19 High concentration impurity region, E1 first electrode, E2 Second electrode.

─────────────────────────────────────────────────── ─── Continuation of front page (56) References JP-A-4-320067 (JP, A) JP-A-5-55554 (JP, A) JP-A-6-77505 (JP, A) (58) Field (Int.Cl. ⁷ , DB name) H01L 29/74

Claims (11)

(57) [Claims] 1. A first semiconductor region having a conductivity type opposite to that of the first semiconductor region, which is provided on one of the front and back main faces of the first semiconductor region, and has a conductivity type opposite to that of the first semiconductor region. The second semiconductor region forming the junction and the first semiconductor region are in contact with the second semiconductor region from the side opposite to each other to form a minority carrier injection junction, and the second semiconductor region is connected to the second semiconductor region. The minority carriers for the region can be injected, and one third region in a cross section along one in-plane direction with respect to the main surface, or a plurality of third regions in a juxtaposed relation spaced apart from each other, and the first semiconductor region. Is provided on the other main surface side opposite to the one main surface of, and forms a minority carrier injection junction in contact with the first semiconductor region, and a minority of the first semiconductor region to the first semiconductor region is formed. Carriers can be injected, and the plane shape has short sides. A rectangle having sides, one cross section along one in-plane direction with respect to the other main surface, or a plurality of fourth regions in a juxtaposed relationship while being separated from each other in the short side direction; (Ii) a first electrode electrically contacting the surface of the semiconductor region and the surface of the third region in common, and a second electrode electrically contacting each surface of the one fourth region or a plurality of fourth regions An electrode and one side or both sides in the short side direction of each of the one fourth region or the plurality of fourth regions along the long side direction of each of the one fourth region or the plurality of fourth regions. A reverse bias of the pn junction between the first and second electrodes, the electrode portion electrically connected to the second electrode and an ohmic contact region for ohmic-contacting the first semiconductor region with each other. Voltage above breakdown voltage depending on polarity When the second surge is applied, the second breakdown occurs and the surge current begins to be absorbed between the first and second electrodes, and after the breakdown starts, the second region is formed between the third region and the ohmic contact region. The minority carriers composed of the first semiconductor region and the fourth region due to an increase in majority carrier flow for the first semiconductor region that flows through the semiconductor region and the first semiconductor region so as to avoid the fourth region. Injection of the minority carriers from the fourth region to the first semiconductor region after a forward voltage is applied to the injection junction to turn on the minority carrier injection junction, and the second semiconductor from the third region. Due to the synergistic effect with the injection of the minority carriers into the region, when the magnitude of the surge current that has started to be absorbed becomes equal to or greater than the breakover current, a breakover occurs via the positive feedback phenomenon, and between the first and second electrodes. Two terminals that continue to absorb surge current while shifting the pressure to a clamp voltage that is relatively low in absolute value and return to the initial state when the magnitude of the surge current falls below the holding current as the surge falls. A method for setting the breakover current or the holding current of the surge protection device in a breakover type vertical surge protection device, comprising: a fourth region or a plurality of fourth regions. The dimension of the short side is Lc, and the dimension of the long side is Lc or more; both sides of the one fourth region in the short side direction or both sides of each of the plurality of fourth regions in the short side direction. Regarding the dimension of the ohmic contact region along the short side direction, the dimension on one side of the both sides of the fourth region is kLs (0 ≦ k ≦ 1), and the dimension on the other side is (1−k) Ls. As; (Lc + L s) / Lc ² is changed to set the breakover current or the holding current in proportion to this, and the setting method of the breakover current or the holding current in the surge protection device. 2. The first semiconductor region is provided on one main surface side of both front and back main surfaces and has a conductivity type opposite to that of the first semiconductor region and a pn between the first semiconductor region and the first semiconductor region. The second semiconductor region forming the junction and the first semiconductor region are in contact with the second semiconductor region from the side opposite to each other to form a minority carrier injection junction, and the second semiconductor region is connected to the second semiconductor region. The minority carriers for the region can be injected, and one third region in a cross section along one in-plane direction with respect to the main surface, or a plurality of third regions in a juxtaposed relation spaced apart from each other, and the first semiconductor region. Is provided on the other main surface side opposite to the one main surface of, and forms a minority carrier injection junction in contact with the first semiconductor region, and the minority of the first semiconductor region to the first semiconductor region with respect to the first semiconductor region. Carriers can be injected, and the plane shape has short sides. A rectangle having sides, one cross section along one in-plane direction with respect to the other main surface, or a plurality of fourth regions in a juxtaposed relationship while being separated from each other in the short side direction; (Ii) a first electrode electrically contacting the surface of the semiconductor region and the surface of the third region in common, and a second electrode electrically contacting each surface of the one fourth region or a plurality of fourth regions An electrode and one side or both sides in the short side direction of each of the one fourth region or the plurality of fourth regions along the long side direction of each of the one fourth region or the plurality of fourth regions. A reverse bias of the pn junction between the first and second electrodes, the electrode portion electrically connected to the second electrode and an ohmic contact region for ohmic-contacting the first semiconductor region with each other. Voltage above breakdown voltage depending on polarity When the second surge is applied, the second breakdown occurs and the surge current begins to be absorbed between the first and second electrodes, and after the breakdown starts, the second region is formed between the third region and the ohmic contact region. The minority carriers composed of the first semiconductor region and the fourth region due to an increase in majority carrier flow for the first semiconductor region that flows through the semiconductor region and the first semiconductor region so as to avoid the fourth region. Injection of the minority carriers from the fourth region to the first semiconductor region after a forward voltage is applied to the injection junction to turn on the minority carrier injection junction, and the second semiconductor from the third region. Due to the synergistic effect with the injection of the minority carriers into the region, when the magnitude of the surge current that has started to be absorbed becomes equal to or greater than the breakover current, a breakover occurs via the positive feedback phenomenon, and between the first and second electrodes. Two terminals that continue to absorb surge current while shifting the pressure to a clamp voltage that is relatively low in absolute value and return to the initial state when the magnitude of the surge current falls below the holding current as the surge falls. A method for setting the breakover current or the holding current of the surge protection device in a breakover type vertical surge protection device, comprising: a fourth region or a plurality of fourth regions. The dimension of the short side is Lc and the dimension of the long side is Lc or more; both sides of the one fourth region in the short side direction or both sides of each of the plurality of fourth regions in the short side direction. Regarding the dimension of the ohmic contact region along the short side direction, the dimension on one side of the both sides of the fourth region is kLs (0 ≦ k ≦ 1), and the dimension on the other side is (1−k) Ls. And; The majority carrier flow for the first semiconductor region flowing between the third region and the ohmic contact region to turn on the minority carrier injection junction formed of the first semiconductor region and the fourth region. The resistivity in the first semiconductor region along the path of is defined as ρ; By changing {(Lc + Ls) / Lc ² } (1 / ρ), the breakover current or holding current is set in proportion to this. A method of setting a breakover current or a holding current in a surge protection device. 3. A method according to claim 1 or 2;
The dimensions kLs and (1−k) Ls of the ohmic contact region
The upper limit of is the distance between the fourth region and the second semiconductor region. 4. The method according to claim 1 or 2, wherein
The dimensions kLs and (1−k) Ls of the ohmic contact region
The upper limit of is the thickness of the first semiconductor region. 5. The method according to claim 1, 2, 3 or 4, wherein k is 0.5. 6. The method according to claim 1, 2, 3, 4 or 5, wherein the ohmic contact region is of the same conductivity type as the first semiconductor region at a portion in contact with the first semiconductor region. Has a high-concentration impurity region having a higher impurity concentration than the first semiconductor region. 7. The first semiconductor region is provided on one main surface side of both front and back main surfaces and has a conductivity type opposite to that of the first semiconductor region and a pn between the first semiconductor region and the first semiconductor region. The second semiconductor region forming the junction and the first semiconductor region are in contact with the second semiconductor region from the side opposite to each other to form a minority carrier injection junction, and the second semiconductor region is connected to the second semiconductor region. The minority carriers for the region can be injected, and one third region in a cross section along one in-plane direction with respect to the main surface, or a plurality of third regions in a juxtaposed relation spaced apart from each other, and the first semiconductor region. Is provided on the other main surface side opposite to the one main surface of, and forms a minority carrier injection junction in contact with the first semiconductor region, and the minority of the first semiconductor region to the first semiconductor region with respect to the first semiconductor region. Carrier can be injected, and the planar shape is square Is a square, or a circle or a circle, or a polygon in which the envelope shape of the outer contour is a circle or a circle, and in one cross section along one in-plane direction with respect to the other main surface, one or each other A plurality of fourth regions that are in a juxtaposed relationship while being spaced apart from each other, a first electrode that makes common electrical contact with the surface of the second semiconductor region and the surface of the third region, and the one fourth region Alternatively, a second electrode electrically contacting each surface of the plurality of fourth regions and an electrode surrounding the periphery of the one fourth region or the plurality of fourth regions and electrically connected to the second electrode. An ohmic contact region for ohmic contact between the portion and the first semiconductor region, and the first,
When a surge having a voltage equal to or higher than the breakdown voltage is applied between the second electrodes with a polarity that reversely biases the pn junction, the first and
After starting to absorb the surge current between the second electrodes, and after starting the breakdown, the fourth region is interposed between the third region and the ohmic contact region via the second semiconductor region and the first semiconductor region. The minority carrier injection is performed by applying a forward voltage to the minority carrier injection junction formed by the first semiconductor region and the fourth region due to an increase in majority carrier flow for the first semiconductor region that flows so as to avoid By the synergistic effect of the injection of the minority carriers from the fourth region to the first semiconductor region after the junction is turned on and the injection of the minority carriers from the third region to the second semiconductor region, When the magnitude of the surge current that has started to be absorbed becomes equal to or greater than the breakover current, the surge current breaks through the positive feedback phenomenon and the voltage between the first and second electrodes is relatively low in absolute value. A two-terminal breakover type vertical surge protection device that continues absorbing surge current while shifting to the pump voltage and returns to the initial state when the magnitude of the surge current falls below the holding current as the surge falls. , A method for setting the breakover current or the holding current of the surge protection device; the area of each of the one fourth region or a plurality of fourth regions
Sc; the area of the ohmic contact region provided around the periphery of each of the one fourth region or the plurality of fourth regions as Ss; by changing (Sc + Ss) / Sc ^3/2 , Setting the breakover current or the holding current in proportion to each other; A method of setting the breakover current or the holding current in the surge protection device, which is characterized in that: 8. The first semiconductor region is provided on one main surface side of the front and back main surfaces and has a conductivity type opposite to that of the first semiconductor region and a pn between the first semiconductor region and the first semiconductor region. The second semiconductor region forming the junction and the first semiconductor region are in contact with the second semiconductor region from the side opposite to each other to form a minority carrier injection junction, and the second semiconductor region is connected to the second semiconductor region. The minority carriers for the region can be injected, and one third region in a cross section along one in-plane direction with respect to the main surface, or a plurality of third regions in a juxtaposed relation spaced apart from each other, and the first semiconductor region. Is provided on the other main surface side opposite to the one main surface of, and forms a minority carrier injection junction in contact with the first semiconductor region, and the minority of the first semiconductor region to the first semiconductor region with respect to the first semiconductor region. Carrier can be injected, and the planar shape is square It is a substantially square, or a circle or a circle, or a polygon whose outer contour has an envelope shape of a circle or a circle, and in one cross section along one in-plane direction with respect to the other main surface, one or A plurality of fourth regions that are in a juxtaposed relationship while being spaced apart from each other; a first electrode that makes common electrical contact with the surface of the second semiconductor region and the surface of the third region; A second electrode electrically contacting each surface of the region or the plurality of fourth regions and surrounding the periphery of the one fourth region or the plurality of fourth regions and electrically connected to the second electrode An ohmic contact region that makes ohmic contact between the electrode portion and the first semiconductor region is provided, and a surge having a voltage equal to or higher than the breakdown voltage is applied between the first and second electrodes with a polarity that reverse-biases the pn junction. Descend Lie down, start absorbing the surge current between the first and second electrodes,
After initiation of the breakdown, majority carriers for the first semiconductor region flow between the third region and the ohmic contact region so as to avoid the fourth region through the second semiconductor region and the first semiconductor region. A forward voltage is applied to the minority carrier injection junction composed of the first semiconductor region and the fourth region due to the increase in the flow, and the minority carrier injection junction is turned on to turn on the minority carrier injection junction. Due to the synergistic effect of the injection of the minority carriers into the first semiconductor region and the injection of the minority carriers from the third region into the second semiconductor region, the magnitude of the surge current that begins to be absorbed is the breakover current. When it becomes the above, it breaks over through the positive feedback phenomenon and absorbs the surge current while shifting the voltage between the first and second electrodes to a clamp voltage of a relatively low voltage in absolute value. In a two-terminal breakover type vertical surge protection device that returns to the initial state when the magnitude of the surge current falls below the holding current due to the fall of the surge, the above breakover current of the surge protection device or the above A method for setting a holding current, wherein the area of each of the one fourth region or the plurality of fourth regions is Sc; and surrounds each of the one fourth region or each of the plurality of fourth regions. The area of the ohmic contact region provided by Ss is Ss; and the minority carrier injection junction constituted by the first semiconductor region and the fourth region is turned on and the third region and the ohmic contact are turned on. Let ρ be the resistivity of the first semiconductor region along the path of the majority carrier flow for the first semiconductor region flowing between the contact regions; {(Sc + Ss) / Sc ^3/2 } (1 / ρ) And setting the breakover current or the holding current in proportion to this by setting the breakover current or the holding current in the surge protection device. 9. A method according to claim 7 or 8;
The upper limit of the cross-sectional dimension for each side of the ohmic contact region located on both sides of the fourth region in one cross section along one in-plane direction with respect to the other main surface of the first semiconductor region, the fourth region and A distance from the second semiconductor region. 10. The method according to claim 7, wherein the ohmic contact is located on both sides of the fourth region in a cross section along one in-plane direction with respect to the other main surface of the first semiconductor region. The upper limit of the cross-sectional dimension on each side of the region is set as the thickness of the first semiconductor region. 11. The method according to claim 7, 8, 9 or 10, wherein said ohmic contact region is of the same conductivity type as said first semiconductor region in a portion in contact with said first semiconductor region. A high concentration impurity region having an impurity concentration higher than that of the first semiconductor region;