GB1600638A

GB1600638A - Semiconductor device

Info

Publication number: GB1600638A
Application number: GB21172/78A
Authority: GB
Original assignee: Philips Gloeilampenfabrieken NV
Current assignee: Koninklijke Philips NV
Priority date: 1977-05-25
Filing date: 1978-05-22
Publication date: 1981-10-21
Also published as: CA1097430A; DE2822166C2; FR2392500A1; SE7805782L; AU3635178A; DE2822166A1; NL181612B; AU518290B2; FR2392500B1; JPS53145581A; IT1094695B; NL181612C; IT7823662A0; SE433547B; NL7705729A; ES470066A1

Description

(54) SEMICONDUCTOR DEVICES (71) We, N.V. PEIILIPS GLOEILAMPEN FABRIEKEN, a limited liability Company, organised and established under the laws of the Kingdom of the Netherlands, of Emmasingel 29, Eindhoven, the Netherlands do hereby declare the invention, for which we pray that a patent may be granted to us, and the method by which it is to be performed, to be particularly described in and by the following statement: The invention relates to semiconductor devices comprising a transistor.

In a transistor having a surface-adjoining emitter of a first conductivity type, a surface-adjoining base of the opposite conductivity type, and a base-adjoining collector of the first conductivity type, it is known to form at least part of the emitter zone by a row of finger-shaped regions, hereinafter termed emitted fingers, of the first conductivity type which extend substantially parallel to each other and in a direction substantially transverse to the longitudinal direction of the row in the base zone.

Transistors having a divided emitter configuration are usually used for high powers.

As is known, in bipolar transistors, with larger currents, the greater part of the emitter current is injected in the base via those parts of the emitter-base junction which are situated nearest the base contact. Parts of the emitter-base junction situated farther from the base contact are not effective or are considerably less effective for current injection as a result of voltage drops in the base. By dividing the emitter, the base contact may be given such a configuration that an emitter-base junction is obtained having an emissive surface which is comparatively large even for larger currents. The base contact may, for example, be constructed in the form of a plurality of base contact fingers which are interdigitated with the emitter fingers.

Such power transistors often have resistors in the emitter or base track to prevent so-called "second breakdown" which may occur as a result of a local temperature rise of the emitter-base junction: at the area where such a temperature rise, even a small one, occurs the emitter current across the emitter-base junction will increase; this results in a local increase of the dissipation and hence in a further temperature rise.

In this manner an avalanche effect may occur resulting in a breakdown which may cause the destruction of the transistor. By for example, incorporating resistors (connected to the emitter fingers) in the emitter track of the transistor, the forward voltage across the emitter-base junction of the relevant emitter finger (and hence the emitter current across said emitter-base junction) can be reduced when a local temperature rise and initial increase in current associated therewith occurs.

In order to improve the protection against second breakdown for the whole operational range within which it should be possible to operate the transistor, comparatively large resistors are required. However, the operating conditions often are such that much lower resistance values could suffice, as is the case when the current is large and hence the voltage drop across the resistor is low. In general it may therefore be assumed that, because the resistance values of the said series resistors are chosen with a view to very particular operating conditions of the transistor, this method cannot ensure the optimum functioning of the transistor in other operating conditions.

It has been found that an important cause of second breakdown resides in the nonuniform temperature distribution which occurs during operation in the body. In particular, measurements have demonstrated that the temperature at the edge of the transistor is lower than in the centre. As a result of this, second breakdown will generally occur sooner in the centre than at the edge of the transistor. It has already been suggested in United States Patent Specification 3,704,398 in connection herewith to increase the distance between adjacent emitter fingers at the edge of the transistor. The underlying idea is that an emitter finger in the centre of the trans istor is heated more by the power dissipated by the adjacent emitter fingers than an emitter finger at the edge of the transistor. By suitably choosing the distances between the emitter fingers and hence the thermal resistances between the various emitter fingers, a better temperature distribution in the semiconductor body can be obtained.

However, this method requires space, that is to say, compared with a conventional transistor capable of providing a certain current, its area should be enlarged. In general, particularly when the transistor forms part of an integrated circuit, it is preferred to keep the dimensions of the transistor (as well as those of the other circuit elements) as small as possible.

According to the present invention there is provided a semiconductor device having a semiconductor body and comprising a transistor having a surface-adjoining emitter of a first conductivity type, a surface-adjoining base of the opposite conductivity type, and a base-adjoining collector of the first conductivity type, at least part of the emiter zone being formed by a row of finger-shaped regions, hereinafter termed emitter fingers, of the first conductivity type which extend substantially parallel to each other and in a direction substantially transverse to the longitudinal direction of the row in the base zone, which emitter fingers have different length in order to obtain during the operation of the device unequal dissipation in the longitudinal direction of the row to improve the uniformity of the temperature distribution in the semiconductor body.

The design of such a transistor structure involves the recognition of the fact that the temperature in a certain part of the transistor is first of all a function of the dissipated power in said part of the transistor and that, by simply making the current in said part larger or smaller, respectively, a temperature rise or a temperature drop, respectively, can be obtained at the area.

Therefore, by making the emitter fingers shorter at the area where in the case of an equal length of the emitter fingers the temperature would be higher than in other places (either as a result of the thermal coupling between the emitter fingers or as a result of the dissipation of adjacent circuit elements in the case of an integrated circuit, or as a result of a less good thermal dissipation) a substantially uniform temperature distribution can be obtained in a simple manner with a compact structure and substantially independently of the value of the overall current.

In a presently preferred form which proves to be particularly advantageous both because of its regular structure and its favourable electrical properties, the length of the emitter fingers decreases from the end of the row towards the centre of the row.

Each emitter finger (and even the whole emitter zone) may be formed simply by one single elongate continuous zone. If desired.

however, each emitter finger could be divided into a plurality of small partial zones which are connected by an emitter contact provided on the surface.

The base may advantageously be formed by a single continuous zone at the edge of which the collector contact is provided. The base may comprise a plurality of sub-zones which are separated from each other by intermediate, surface-adjoining parts of the collector and each of which comprises at least one emitter finger, a base contact may have a plurality of contact fingers which are provided on the said sub-zones and extended substantially parallel to the emitter fingers over the surface of the body, and a collector contact may have a plurality of collector contact fingers which are provided on the said intermediate parts of the collector and extends substantially parallel to the emitter fingers and the base contact fingers over the surface of the body. One of the advantages of this latter structure which is particularly suitable for use in an integrated circuit is a comparatively low collector-series resistance since the current need cover only a comparatively small distance in the collector. The sub-zones of the base may be conductively connected together because protective resistors between the individual sub-zones of the base and a common base contact are not necessary.Therefore, the base contact may comprise a base connection which is common to the sub-zones and which is connected to each of the sub-zones of the base via a connection provided in the collector, which connection is formed by a first sub-zone of the opposite conductivity type which adjoins an associated sub-zone of the base and by a second sub-zone of the first conductivity type which is separated from the collector by th( first sub-zone and which is connected to the common base connection and to an associated base contact finger.

Each emitter finger as viewed on the surface may be situated between a collector contact finger and a base contact finger, and an emitter contact may have a plurality of emitter con'act fingers provided towards che edge of the emitter fingers which is situated farthest from the associated base contact fingers. Because most of the charge carriers are injected in the base on the side of the emitter fingers which is situated nearest the base contact fingers. by providing the emitter contact fingers on the opposite side of the emitter fingers a resistance is introduced in the emitter current track by the emitter fingers themselves, and this contributes in producing a uniform current distribution in the longitudinal direction of the emitter fingers.

The sub-zones of the base may also have different lengths. It is particularly advan tageous from the point of view of space for the centres of the emitter fingers to be situated substantially on a straight line which extends in the longitudinal direction of the row and substantially at right angles to the emitter fingers. Advantageously, a contact pad for an external supply conductor may be formed, for example, in or near the centre of the transistor.

Embodiments of the invention will now be described, by way of example, with reference to the accompanying diagrammatic drawings, in which: Fig. 1 is a plan view of a part of a semiconductor device in accordance with the invention: Figure 2 is a cross-sectional view of the device taken on the line II-II of Fig. 1; Fig. 3 is a cross-sectional view of the device taken on the line III-III of Fig. 1, and Fig. 4 is a plan view of a part of a second semiconductor device in accordance with the invention.

It is to be noted that the Figures are only diagrammatic and are not drawn to scale.

In addition, for clarity, passivating layers on the surface of the semiconductor body of Fig. 1 are not shown.

Figures 1 to 3 show a power transistor which may form part of a monolithic integrated circuit. This circuit, not further shown in the drawing, may be, for example, an amplifier device in which the present transistor belongs to the output stage of the amplifier. Of course, the transistor may alternatively be used advantageously in other circuit arrangements.

The device comprises a semiconductor body 1 of a form which is usual for conventional integrated circuits having a p-type silicon substrate 2 and an n-type epitaxial silicon layer 3 deposited thereon. The surface 4 of the body 1 is covered with an insulating passivating layer 9, usually of silicon oxide, having apertures at the areas where the body 1 or parts thereof are to be contacted.

The transistor comprises an n-type emitter 5 adjoining the surface 4 of the body 1, a surface-adioining p-type base zone 6. and an n-type collector 7 which in the present case also adjoins the surface 4. The collector 7 comprises a part of the epitaxial layer 3, a low-ohmic buried n-type collector zone 8 formed between the epitaxial layer 3 and the substrate 2, and n-type collector contact zones 17 extending from the surface 4 into the buried layer 8.

In order to obtain an emissive emitter surface which is as large as possible, the emitter is divided into a row of digital areas, hereinafter referred to as emitter fingers. These emitter fingers, referred to hereinafter for distinction by the suffixes !a, b, c, d etc., extend parallel to each other and in a direction substantially transverse to the row a, b, c, d etc in the base 6.

As shown in the plan view of Fig. 1, the length of all the emitter fingers is not approximately equal as is usual, but in accordance with the invention the emitter fingers have different lengths. As will be explained in detail hereinafter, this special emitter configuration and the associated non-uniform temperature distribution in the transistor during operation as compared with the case in which the emitter fingers would have the same lengths. In very many cases a non-uniform temperature distribution across the emitter-base junction may cause second breakdown and hence ultimately the destruction of the transistor. In transistors in accordance with the invention, the operating range within which the transistor can be operated without breakdown (SOAR= safe operating area) can be increased in a very simple manner and, as experiments have demonstrated, in a very efficacious manner. Therefore conventional and generally comparatively large series resistors in the base and/or in the emitter current track which are used in conventional power transistors to prevent second breakdown are not necessary in a transistor in accordance with the invention.

As shown in Fig. 1 the length of the emitter fingers 5 gradually decrease from the edge of the transistor on the left-hand and right-hand sides of the Figure towards the centre of the transistor centrally in the Figure. In most of the cases in which the temperature distribution in the semiconductor body is determined predominantly by the dissipation in the transistor, this configuration proves to be favourable.

In order to obtain a low collector series resistance, the base 6 or at least the active or intrinsic part of the base, is divided into a plurality of sub-zones. In order to distinguish these sub-zones from each other, the reference numerals of the individual sub-zones are provided with the suffixes a, b, c, d etc. from the left to the right in Figs. 1 and 2. Only one emitter finger 5a, Sb, Sc etc. is situated in each sub-zone 6a, 6b, 6c etc. Of course, the emitter fingers might also be distributed over the various base sub-zones 6a, b, c etc. in different manners, for example, two emitter fingers per base sub-zone.

The base comprises a base contact having a plurality of base contact fingers 10 of a suitable metal, for example, alumin ium or of a combination of metals provided on the base sub-zones 6 via usual contact windows formed in the oxide layer 9. As viewed on the surface 4, the base contact windows extend substantially parallel to the emitter fingers over the surface 4 of the body 1 and are interconnected by a common base contact part 11 of the same metal or metals as the base contact fingers 10. Because no base series resistances are necessary to prevent second breakdown, the base contact fingers 10 may be connected conductively to the common contact 11. This low ohmic connection is formed by a so-called underpass having p-type zones 12 provided in the collector 7 and adjoining the base sub-zones or base fingers 6a, b, c etc. and n-type surface zones 13 situated in the zones 12 and separated thereby from the n-type collector. During the manufacture of the device the zones 12 and 13 can be provided simultaneously with the base 6 and the emitter 5, respectively. In the embodiment shown each sub-zone 6 of the base is hence connected to the common base contact 11 via a separate underpass 12, 13.

One of the advantages of this configuration is that stray capacitances in particular between the base and the collector of the transistor can be kept comparatively low.

It will be obvious, however, that particularly in those cases in which less stringent requirements as regards stray capacitances are imposed, the zones 12, 13 may also be constructed in the form of a continuous region. The p-n junctions between the zones 12 and 13 are short-circuited at least on the base contact side (and as shown in the present embodiment also on the other side) namely by the common base contact 11 and by the base contact fingers 10 which, at the area of said shortcircuit, are provided with laterally projecting parts 14 as is shown in Figure 1.

The collector 7, 8 has a collector contact 15 with a number of collector contact fingers 16 which are contacted to the parts 17 of the collector situated between the base fingers 6a. b, c etc. and which, viewed on the surface 4, extend interdigitally between the emitter fingers 5 and the base contact fingers 10. As usual, highly doped n-type contact zones 17 are provided at the area of the contact between the collector contact fingers 16 and the collector 7, 8.

The base and collector contacts (10, 11; 15, 16) are shown in broken lines in the plan view of Figure 1. The places where the common base contact 11 is contacted to the underpass 12, 13 are denoted by an x in Figure 1.

In the configuration shown, each emitter finger 5, viewed on the surface 4, is situated between a base contact finger 10 and a collector contact finger 16. The emitter 5 has an emitter contact 18 only part of which is shown in Figure 1 by means of dot-and-dash lines. The contact 18 comprises a plurality of emitter contact fingers 19 which are situated interdigitally between the base contact fingers 10 and the collector contact fingers 17. The fingers 19 are situated asymmetrically with respect to the emitter finger 5 namely towards the edge of the emitter fingers which is situated farthest from the associated base contact fingers 10. Since, at least with larger currents, almost only that part of each emitter finger 5 which is situated nearest to the adjacent base contact finger injects electrons into the base, the internal resistance in the emitter fingers (which results from the asymmetric location of the emitter contact fingers 19) provides a contribution to the uniformity of the current distribution in the longitudinal direction of the emitter fingers.

As is furthermore shown in Figure 1, the base fingers 6a, b, c etc. also have different lengths, as have the emitter fingers S. The centres of the emitter fingers 5a, b, c etc. coincide substantially with the lineII-II of Figure 1. Such a symmetric configuration is advantageous thermally because in a symmetrical configuration the temperature distribution in the semiconductor body usually shows a greater extent of uniformity than in an asymmetric configuration.

The part of the semiconductor body 1 situated beside the centre of the transistor where the emitter fingers 5 and the base fingers 6 are smallest, may be used advantageously, for example, for conductor tracks. In the present embodiment this part of the semiconductor body 1 is used to provide a contact pad 20 to which an external supply conductor 21 can be connected. The contact pad 20 is situated above a part 22 of the epitaxial layer 3 which is separated by a p-type isolation zone 23 from the island in which the transistor is situated. If desired, further circuit elements or parts thereof may be provided in the part 22 of the epitaxial layer 3. As shown in Fig. 3, the contact pad 20 is separated from the epitaxial layer 3 by the oxide layer 9.

The device can be manufactured by means of generally known methods which are not explained here. In this respect, reference may be made to handbooks in which the manufacture of integrated circuits is described in detail.

To explain the effect of the invention, the transistor described will be compared with a conventional power transistor in which all the emitter fingers have the same or at least substantially the same length. As a result of the dissipation, the temperature of the emitter base junction will rise during operation. This rise in temperature will generally be non-uniform may be considered qualitatively as follows, it being assumed that the temperatures in the part of the semiconductor body 1 which is occupied by the transistor itself, and that the heat dissipation is substantially the same for all the emitter fingers. A first contribution to the rise in temperature occurring during operation is provided by the current through each individual emitter finger. When the emitters are of equal lengths and the above-described conditions apply, this contribution will in principle be equally large for all the emitter fingers. A second contribution to the rise in temperature of each individual emitter finger is provided by the dissipation of the immediately adjacent emitter fingers. This contribution is not equally large for all emitter fingers but for the outermost emitter fingers each having only one immediately adjacent emitter finger it is smaller than for the remaining emitter fingers which are each bounded on either side by two adjacent emitter fingers. A subsequent contribution to the rise in temperature of the individual emitter fingers is provided by the dissipation of the emitter fingers which are situated two places farther away in the row. For the same reasons, this contribution will be smaller for the emitter fingers which are situated at the edge of the transistor than for the emitter fingers in the centre of the transistor. On the basis of these considerations it may be seen that on the conditions described a temperature distribution occurs in a conventional power transistor having emitter fingers of the same lengths in which the temperature in the centre is higher than at the edge of the transistor.

This non-uniformity in the temperature distribution may be one of the most important causes for the occurrence of second breakdown and hence may often result in the destruction of the transistor; since the part of the emitter-base p-n junction having the highest temperature can pass a disproportionately large part of the overall current through the transistor, the temperature at the area of this part may rise even further, which again results in a further increase of the local current. By making the emitter fingers at the area where (in the case of emitter fingers of equal lengths) the tepmerature would become highest during operation, smaller than at the area where (in the case of emitter fingers of equal length) the temperature would become lowest, a more uniform temperature distribution can be obtained in a simple manner which does not require extra area of the semiconductor body. In the practical case in which the supposition is assumed that the temperature distribution in the part of the semiconductor body 1 in which the transistor (5, 6, 7) is situated is mainly determined by the dissipation in the transistor itself, this means that the emitter fingers 5 become gradually shorter from the edge towards the centre, so that the dissipation in the centre is smaller than at the edge. As a result of the uniform temperature distribution obtained in this manner, resistors which are provided, to prevent second breakdown, in many conventional power transistors in the current track of the base and/or emitter fingers, are not necessary.

As will be obvious, the mutual ratios between the lengths of the emitter fingers depends in particular on the thermal resistance between the various emitter fingers and hence on the distances between the emitter fingers was approximately 140 ment, the largest mutual distance between the emittr fingers was approximately 140 microns and the smallest distance was approximately 100 microns. It has been found that good results can be obtained when the emitter fingers Sa-5e have lengths of approximately 450 microns, 400 microns, 350 microns, 300 microns, and 250 microns, respectively. In the case in which the distances between the emitter fingers become larger, the differences in lengths of the emitter fingers may bcome smaller. When on the contrary the distances between the emitter fingers become smaller (for example in the case in which the collector is contacted only along the edge of the transistor and not by means of the collector contact fingers 16), the differences between the emitter fingers should be made larger.

Figure 4 is a plan view of a second embodiment of a transistor in accordance with the invention which serves for operation at smaller power levels than the transistor of Figures 1 to 3 and therefore comprises only six emitter fingers 5a-5f. It is to be noted that only the emitter fingers are shown in the plan view of Figure 4.

The emitter contact fingers, the base fingers with the base contact and the collector contact ore not shown in the Figure for clarity. The mutual distances between the emitter fingers 5 were in this case alternatively approximately 104 and 138 microns. It has been found that good results are obtained with lengths of the emitter fingers 5a-5f of 450 microns, 380 microns, 300 microns, 300 microns, 380 microns and 450 microns, respectively.

It will be obvious that many variations are possible to those skilled in the art without departing from the scope of this invention. For example, the emitter fingers themselves may also be divided into a number of parts which are separated from each other by intermediate parts of the base fingers. Such a division of the emitter may be particularly advantageous when the length of the emitter fingers is large, for example exceeding 500 microns. Furthermore, the conductivity types of the various regions may be reversed. Instead of the materials described, for example for the metallisation, other materials may advantageously be used. Instead of base contact fingers 10 which are connected to the common base contact 11 via underpasses 12, 13, a base contact in the form of a single conductor strip meandering between the emitter contact fingers 19 and the collector contact fingers 16 may alternatively be used.

WHAT WE CLAIM IS: 1. A semiconductor device having a semiconductor body and comprising a transistor having a surface-adjoining emitter of a first conductivity type a surface-adjoining base of the opposite conductivity type, and a base-adjoining collector of the first conductivity type, at least part of the emitter zone being formed by a row of fingershaped regions, hereinafter termed emitter fingers, of the first conductivity type which extend substantially parallel to each other and in a direction substantially transverse to the longitudinal direction of the row in the base zone, which emitter fingers have different lengths in order to obtain during the operation of the device unequal dissipation in the longitudinal direction of the row to improve the uniformity of the temperature distribution in the semiconductor body.

2. A semiconductor device as claimed in Claim 1, in which the length of the emitter fingers decreases from the end of the row towards the centre of the row.

3. A semiconductor device as claimed in Claim 1 or Claim 2, in which the base comprises a plurality of sub-zones which are separated from each other by intermediate, surface-adjoining parts of the collector and each of which comprises at least one emitter finger, a base contact has a plurality of contact fingers which are provided on the said sub-zones and extend substantially parallel to the emitter fingers across the surface of the body, and a collector contact has a plurality of collector contact fingers which are provided on the said intermediate parts of the collector and extend substantially parallel to the emitter fingers and the base contact fingers across the surface of the body.

4. A semiconductor device as claimed in Claim 3, in which each emitter finger as viewed on the surface is situated between a collector contact finger and a base contact finger, and an emitter contact has a plurality of emitter contact fingers provided towards the edge of the emitter fingers which is situated farthest from the associated base contact fingers.

5. A semiconductor device as claimed in any of Claims 3 to 4, in which each sub-zone of the base comprises only one emitter finger.

6. A semiconductor device as claimed in any of Claims 3 to 5, in which the subzones of the base also have different lengths.

7. A semiconductor device as claimed in any of Claims 3 to 6, in which the base fingers are connected to a common base connection.

8. A semiconductor device as claimed in any of Claims 3 to 7, in which the base contact comprises a base connection which is common to the sub-zones and which is connected to each of the sub-zones of the base via a connection provided in the collector, which connection is formed by a first sub-zone of the opposite conductivity type adjoining an associated sub-zone of the base and by a second sub-zone of the first conductivity type which is separated from the collector by the first sub-zone and which is connected to the common base connection and to an associated base contact finger.

9. A semiconductor device as claimed in any of Claims 2 to 8, in which the centres of the emitter fingers are situated substantially on a straight line which extends in the longitudinal direction of the row and substantially at right angles to the emitter fingers.

10. A semiconduc

Claims

**WARNING** start of CLMS field may overlap end of DESC **. are possible to those skilled in the art without departing from the scope of this invention. For example, the emitter fingers themselves may also be divided into a number of parts which are separated from each other by intermediate parts of the base fingers. Such a division of the emitter may be particularly advantageous when the length of the emitter fingers is large, for example exceeding 500 microns. Furthermore, the conductivity types of the various regions may be reversed. Instead of the materials described, for example for the metallisation, other materials may advantageously be used. Instead of base contact fingers 10 which are connected to the common base contact 11 via underpasses 12, 13, a base contact in the form of a single conductor strip meandering between the emitter contact fingers 19 and the collector contact fingers 16 may alternatively be used. WHAT WE CLAIM IS:

1. A semiconductor device having a semiconductor body and comprising a transistor having a surface-adjoining emitter of a first conductivity type a surface-adjoining base of the opposite conductivity type, and a base-adjoining collector of the first conductivity type, at least part of the emitter zone being formed by a row of fingershaped regions, hereinafter termed emitter fingers, of the first conductivity type which extend substantially parallel to each other and in a direction substantially transverse to the longitudinal direction of the row in the base zone, which emitter fingers have different lengths in order to obtain during the operation of the device unequal dissipation in the longitudinal direction of the row to improve the uniformity of the temperature distribution in the semiconductor body.

10. A semiconductor device as claimed in any of Claims 2 to 9, in which at least one of the emitter, base and collector zones is connected to a contact pad for an external supply conductor, and said contact pad as viewed on the surface is situated immediately beside the shortest emitter fingers present in the centre of the transistor.

11. A semiconductor device substantially as described with reference to Figures 1 to 3 of the accompanying drawings.

12. A semiconductor device substantially as described with reference to Figure 4 of the accompanying drawings.