SE437094B - SEMICONDUCTOR DEVICE WITH OTHER INCLUDING A FIELD POWER TRANSISTOR - Google Patents

Description

The gate electrodes are formed of metal layers adjacent to semiconductor electrode zones which, together with adjacent part of the semiconductor body, form pn junctions with respect to gate electrodes and non-rectifying junctions with respect to the emitter and collector electrodes. Furthermore, the control electrode may be formed by a conductive layer, which is separated from the semiconductor surface by an insulating layer and together with which the failure zone is formed within the channel area in e.g. a field effect transistor with sgk. deep emptying. Accordingly, when in the present case emitter, collector and control electrodes are indicated, this includes the electrode zones and the insulating layers which may be coordinated with these electrodes.

In view of known field effect transistors of the type indicated, in general no high voltages can be applied across the first and second pn junctions. This is due i.a. on the fact that breakthroughs, long before the breakthrough voltage is reached for the first pn junction, which can be expected theoretically on the basis of the doping profile, already occur at the second pn junction as a result of this prevailing unfavorable field distribution. This breakthrough usually occurs at or in the immediate vicinity of the surface. The unfavorable field distribution can be caused by a high doping concentration in the third region and / or by a high doping gradient near the second transition but also e.g. of any local sharp curvature of the second transition. In order to increase the allowable stress, the doping concentration of the first region can be reduced and its thickness can also be increased to make room for the failure zone, so that it can extend further into the first region. However, since the conductivity of the duct is proportional to the thickness but the voltage at completely emptied duct is proportional to the square of the thickness of the duct area, this measure will reduce the ductility of the duct when its length and width remain the same and at the same voltage at fully emptied channel. With regard to the voltage V at a completely emptied channel, it has in fact been found that Vp = a2qN / 25, while the conductivity G of the channel has been found to be equal to Wq / uNa / L, a is the thickness of the channel area displaced by the control electrode, N is the doping concentration in the channel region, W is the width, and L is the length of the channel region, / u is the mobility of the charge carriers, q is the electron charge and g is the dielectric constant of the semiconductor material. If N is reduced to a value N = N / ß (B) l), then, when the voltage Vp remains 7906289-9 \; J the same, it has been found that a '= V / š: š% íïyB = a Vríï Wq uNa / G - - <G.

L1 / ß_ -fß- But in general such a reduction in the conductivity of the channel is very unfavorable for the good operation of the transistor. and the invention is i.a. based on the task of developing a semiconductor device with a flat surface, comprising a field effect transistor with a new construction, this device can be used at considerably higher voltages than with known field effect transistors of the type described without reducing the conductivity of the channel. The invention is also i.a. based on the discovery of the circumstance, that contrary to what might be expected this can be achieved by reducing the thickness of the first area instead of increasing it.

According to the invention, therefore, a semiconductor device of the type described is characterized in that the doping concentration N in atoms / cm 3 of and the thickness di cm of the part of the island-shaped region satisfies the condition v 2,6,1o3 E EVïJ-É 5 Na S 5,1.1o5f E, where E is the relative dielectric constant and E is the critical field strength in volts / cm, at which avalanche multiplication occurs in the semiconductor material of the first region, L is the distance in cm from the contact region up to the second pn junction and VB the one-dimensional calculated the value in volts of the breakdown voltage of the first pn junction.

If this condition is met, the product of the doping concentration and the thickness of the first region is such that the burst zone, when the reverse voltage is applied, extends at least between the contact region and the second transition from the first transition over the entire thickness of the part of the island area at a voltage, lower than the breakdown voltage of the second transition. The contact area may be an electrode or an electrode zone, which is directly connected to the source of the reverse voltage, but may alternatively be e.g. a semiconductor zone, which itself is not provided with any connection conductor but is imparted to the desired potential in another way, e.g. via an adjacent semiconductor zone. e Since the part of the island-shaped region of the first conduit type between the contact region and the second junction is already completely emptied at a voltage lower than the breakdown voltage of the second junction, the field strength at the surface is reduced to such an extent that the breakdown voltage is no longer is determined by this second transition but to a considerable extent by the first transition, which extends parallel to the surface. In this way, a very high breakthrough voltage can be achieved between the first and the second area, which in certain circumstances can reach the high value, which can be expected theoretically on the basis of the dopings of the first and the second area.

Thus, due to the condition according to the invention, when increasing the voltage between the first and the second region, too high a field strength occurs prematurely at the surface between the contact area and the second transition due to the penetration of the failure zone from the second transition up to this contact area. Optimal field distribution is obtained due to the fact that with the N.d product according to the invention the maximum values in the field strength, which appear at the second transition and at the edge of the contact area, are also almost equal. In addition, if the conditions are selected in such a way that Nd is substantially equal to 3.0.l05 E and 1.2 l, ü.l0- V, ensures that the maximum field strength at the first B junction will be constantly greater than at the above-mentioned 5 maximum values that occur at the surface, so that the breakthrough constantly occurs at the first transition and not at the surface.

In order to be able to store the greater part of the charge within the burst zone in the second region, so that the minimum thickness of the first region is reduced, it is often convenient if the second region at least in the vicinity of the first region has a lower doping concentration than the first region. Although the burst zone at the first junction can in many cases easily extend over the entire thickness of the second region, it is preferably ensured in other cases that the second zone has such a thickness that the burst zone at the breakdown voltage for the first junction extends into the second s 7906289-9 area over a distance which is less than the thickness of this area. In this case it is ensured that the breakdown voltage cannot be adversely affected by the thickness of the second region. Although the described semiconductor device may be constructed in another way, a construction is particularly suitable e.g. for technological reasons, in which the first region is formed by an epitaxial layer of the first conduit type, which is arranged on the second region. The third area adjacent to the first area need not extend over the entire thickness of the first area. This is sufficient if the coordinated failure zone, at least in the operating state, extends over the entire thickness of the first area and over at least a part of its periphery and delimits an island-shaped part thereof. However, the island-shaped part of the first area is preferably limited laterally entirely by the second transition, although other embodiments are particularly suitable in some cases, as this part of the first area is limited laterally e.g. partly of the second transition and in another way with regard to the remaining part e.g. of a recessed insulating material or of a groove, such as is filled with passivating glass.

The invention is of particular importance in the case of lateral type field effect transistors, in which the current between the emitter electrode and the collector electrode flows substantially parallel to the surface. Therefore, a particularly suitable embodiment is characterized in that the emitter and collector electrodes on each side of the control electrode form non-rectifying contacts together with the first region, the contact region being formed by the collector electrode of the transistor. In this case, the control electrode is usually connected to the second area, which then serves as a second control electrode, although this is not necessary.

In some cases, an embodiment is particularly suitable in which the collector electrode is completely surrounded by the control electrode, which in turn is completely surrounded by the emitter electrode. A particularly suitable embodiment is characterized in that a semiconductor layer of the second lead type is located in the first region, that the emitter and collector electrodes consist of zones of the first lead type and the gate electrode of a zone of the second lead type and that all electrode zones extend over the entire thickness of this semiconductor layer down to the first region. 7906289-9 6 This latter particularly suitable embodiment enables placement in the middle of each other in the same semiconductor plate of complementary field effect transistors with pn-type control electrodes, ie. field effect transistors with n-channel and p-channel, as will be described below.

In addition to field effect transistors of the lateral type, the invention can also be applied to advantage to field effect transistors of so-called vertical type. In this connection, a particularly suitable embodiment is characterized in that the field effect transistor is of the vertical type, that the collector electrode forms a non-rectifying contact together with the second region, that the emitter electrode forms a rectifying contact together with the first region and that the control electrode consists of a zone of the first conduit type, which surrounds at least a part of the first area, which is coordinated with the channel area and forms the contact area.

The invention is described in more detail below with reference to the accompanying drawing, in which Fig. 1 schematically shows a known semiconductor device partly as a cross-sectional view and partly as a perspective view, Fig. 2 schematically shows a semiconductor device according to the invention partly as a cross-sectional view and partly as a perspective view, Fig. 3 shows another embodiment of the semiconductor device according to the invention, Fig. H and 5 further embodiments of the semiconductor device according to the invention, Fig. 6 a semiconductor device with a vertical field effect transistor according to the invention, Fig. 7 a field effect transistor with deep discharge according to the invention, Figs. 8A-E show the field distribution for different dimensions and dopings and Fig. 9 with regard to a particularly suitable embodiment the relationship between the doping and the dimensions of the first area. The figures are schematic and for the sake of clarity are not drawn to scale. Corresponding parts are generally denoted by the same reference numerals. Semiconductor areas of the same wire type are, as a rule, sheared in the same direction; In all embodiments, silicon has been chosen as the semiconductor material, but the invention is not limited thereto but can also be applied to the use of any other semiconductor material, e.g. germanium or a so-called III-V compound, e.g. GaAs.

Fig. 1 shows a known semiconductor device partly in the form of a cross-sectional view and partly in the form of a perspective view. The device comprises a semiconductor body with a field effect transistor, comprising an emitter electrode and a collector electrode with arranged electrode zones 12 and U, an intermediate channel region 1 and a control electrode with coordinated electrode zone 13, which adjoins the channel region 1. The control electrode has information to act on a failure zone by means of a control voltage supplied thereto for controlling the flow of charge carriers, in this case an electron current between the emitter electrode 12 and the collector electrode U. In this case the emitter electrode, the collector electrode and the control electrode all consist of a semiconductor zone and a metal which is arranged thereon and which is in barrier-free contact with the coordinated electrode zone and which is not shown in the figure for the sake of clarity. The channel region 1 is in this case n-conducting, the electrode zones 12 and N n-conducting with higher doping than the region 1, while the control electrode zone 13 is p-conducting and forms a rectifying pn junction 7 together with the channel region 1.

As can be seen from Fig. 1, the field effect transistor comprises a layer-shaped first region 1 of a first line type, in this case n-line type. This first region 1, which in the present case also constitutes a channel region and adjoins the control electrode, together with an underlying second region 2 of p-type forms a first pn junction 5, which extends substantially parallel to the surface 8. An island-shaped part of the area 1 is bounded laterally by a second pn junction 6 with coordinated shortage zone. This transition 6 is formed between the first region 1 and a third region 3 of p-type, which extends between the second region 2 and the surface 8 and which has a higher doping concentration than the second region 2. The transition 6 thus has a lower breakdown voltage than the transition 5. The gate leads to the island-shaped part of the area l.

According to Fig. 1, the control electrode is connected to the substrate, in this case the second region 2, although this is not necessary. When a voltage VD is applied between the connection terminals S and D on the emitter and collector electrodes, electrons flow through the region 1 from the zone 12 to the zone H.

By applying a voltage in the reverse direction between the control electrode zone 13 and the region 1 and between the second region 2 and the region 1, burst zones are formed, the boundary lines 9, 10, lü of which are indicated by dashed lines in Fig. 1. These burst zones are shown without any shading.

In the known device described above, such a doping concentration and such dimensions occur that the area 1 in the vicinity of the electrode Ä is not emptied at the breakthrough voltage 19oe2s9-9 8 for the transition 6. The voltage in the reverse direction over the transitions 6 and 7, which is highest near the collector electrode U, gives gives rise to a field strength distribution, the maximum value of which occurs in the vicinity of the places where the transitions 6 and 7 intersect the surface 8, and it is near this surface that breakthrough finally occurs at a voltage which is considerably lower than the breakdown voltage of the transition 5 inside the semiconductor body mass. .

Fig. 2 shows a semiconductor device according to the invention.

This device is very similar to the known device shown in Fig. 1. However, according to the invention, the first region 1 of the device shown in Fig. 2 has such a low doping concentration and such a small thickness that the burst zone, when a voltage is applied in the reverse direction between the second region 2 and a contact region belonging to the emitter, collector and the gate electrodes (in this case the collector electrode H) and form a non-rectifying contact together with the island-shaped area, at least between the electrode U and the transition 6 extending from the first transition 5 over the entire thickness of the area 1 at a voltage which is lower than the breakdown voltage of the junction 6. Fig. 2 illustrates the state in which the area 1 between zones 7 and U is completely emptied up to the junction 6. The voltage across the junctions 5, 6 and 7 is now completely absorbed by the coherence. the shortage zone, which extends from the zone Ä up to the boundary line 9. As a result, the field strength at the surface is considerably reduced.

Accordingly, the breakdown voltage is determined at least to a considerable extent by the properties of the junction 5, which extends within the mass of the semiconductor body. This breakdown voltage can be very high and can approach the breakthrough voltage which is theoretically expected on the basis of the doping of areas 1 and 2 '.

To achieve the described desired result according to the invention, the following dopings and dimensions were used in the device shown in Fig. 2, which comprises a semiconductor body of silicon: Zones H and 12: thickness 1 micron. Area 1: n type, doping concentration l, 5.10 thickness 5 microns Range 2: p-type, doping concentration l, 7.l0lu atoms / cm3, thickness 250 microns _ Zone 13th p-type, thickness 2.5 microns distance L from the electrode N up to the pn junction 6 = 50 microns. 3 atoms / cm, 9 9906289-9 In this case, the one-dimensionally calculated breakdown voltage VB at the first pn junction amounts to 1270 volts. The current breakthrough voltage turned out to be 700 volts. At the indicated thicknesses and doping concentrations, the burst zone within the region 2 extends over a thickness which is less than the thickness of the zone 2, while also avoiding that the burst zone at the transition 6 reaches the zone Ä at a stress value which is less than the breakdown voltage for the transition 6 taken for itself, ie. in the absence of the transition 5. By using these values for N, d, L and VB in the present case for silicon (E = 11.7 and E = 2.5.105 volts / cm), the condition VB 2 is fulfilled In the semiconductor device shown in Fig. 2, the first region 1 is formed by an epitaxial layer, which is arranged on the second region 2. In the present case, the island-shaped part of the first area is laterally completely surrounded by the second transition 6. Although other embodiments are possible, as described below, this embodiment is the most technologically simple. The island-shaped part can_e.g. over part of its periphery be limited in other ways, e.g. by means of a submerged oxide pattern or by means of a groove, such as e.g. is filled with passivating glass.

In the devices shown in Figs. 1 and 2, the control electrode forms rectifying contact, while the emitter and collector electrodes form non-rectifying contacts together with the area 1 by means of the doped surface zones 12, H and 13. But the presence of these surface zones is not absolutely necessary. Instead of zones 12 and U, barrier-free metal-semiconductor contacts can be arranged and instead of zone 15, a rectifying metal-semiconductor contact or Schottky ~ contact can be arranged in the area 1. Instead of a control electrode with a rectifying transition, a conductive layer can also be used. , which is separated from the semiconductor surface 8 by means of an insulating layer, together with which conductive layer a gap zone is formed in the epitaxial layer 1, such as e.g. is the case with a transistor with deep discharge.

Fig. 3 illustrates the manner in which the invention can be applied to design in the same monolithic integrated circuit centered field effect transistors with pn control electrodes and p- and n-channel. A field effect transistor with p-channel, which is in principle similar to the transistor described with reference to Fig. 2, but in which all corresponding semiconductor zones have the opposite lead type to those shown in Fig. 2, is denoted by I. the region 2 in this transistor is formed by an n-type epitaxial layer, which is arranged on a p-type substrate 3U.

A heavily doped n-type immersed layer 36 is located between the layer 2 and the substrate 34 to prevent penetration of the burst zone, which is coordinated with the transition 5, down to the substrate 3Ä.

A second field effect transistor II with pn control electrode is located next to the transistor I. This is also a field effect transistor according to the invention. This second transistor II also comprises an island-shaped part 32, which is formed by a part of the same epitaxial layer, of which the region 2 in the transistor I is formed. The n-type emitter zone 22, the n-type collector zone 2ü and the p-type gate electrode zone 23 extend over the entire thickness of the p-type semiconductor layer 21 located on the island 23 and from which the region li of transistor I has also been formed. down to area 32 of n-type. Zones 22 and 2H form pn junctions 26 and 26A together with region 21. while regions 21 and 32 form pn junction 39. At this second transistor, the channel region is formed by region 32. Due to mutual isolation of transistors I and II is the heavily doped zone 33 of the p-type, which surrounds both the region 2 and the region 32 completely and together with the region 32 forms the pn junction 38.

When a suitable voltage is applied between the zones 22 and 24, electrons move from the emitter zone to the collector zone via the area 32. This electron current can be influenced by the application of a control voltage in the reverse direction between the zone 23 and the area 32 and possibly also the reverse voltage between the areas 32 and 3U. As in the embodiment according to Fig. 2, the doping concentration and thickness of the layer 2, 32 have been selected according to the invention in such a way that the area 1, long before breakthrough occurs, is completely emptied at least between the collector zone H and the transition 6 and the area 32 is completely emptied at least between the collector zone 2H and the transition 27. As a result, the field strength is considerably reduced at the surface 8 and at the transistor II at the surface 39 between the regions 21 and 32 and is considerably increased by the breaking voltage.

In Fig. 3 as in Fig. 2 the insulating (oxide) layers and the contact layers present at the surface are not shown. The connections for the emitter, collector and control electrodes are schematically indicated by S, D and G.

Fig. H shows a further modified embodiment of the semiconductor device according to the invention. As with the second transistor II in Fig. 3, the n-type collector zone MH is surrounded by the p-type gate electrode zone H3, which in turn is surrounded by the n-type emitter zone H2. All electrode zones are located within a first region 1, which together with an underlying second region 2 of p-type forms a first pn junction 5 and together with a strongly doped region H7 of p-type a pn junction H8, which ends at surface 8. The emitter, collector and gate electrode zones H2, UU and H3 extend only over a part of the thickness of the first region 1. The field effect transistor can be driven in the same way as previously described transistors, the boundary lines H9 and H0 for the burst zone in the figure refers to a reverse voltage between areas 1 and 2, which is lower than the breakdown voltage. The area l is completely emptied between the zone H3 and the zone ÄH. As with the transistor II in Fig. 3, the island-shaped part of the first region is surrounded by control electrodes, which in this case serve as a third region, while the transition H6 between the gate electrode zone and the region 1 forms the second pn junction. Since the doping and thickness of the first region have been selected according to the invention, so that this region is completely emptied with rising gate electrode collector voltage, before breakthrough occurs at junction 6, this transistor can be used at very high voltage between the gate and the collector electrode.

The device shown in Fig. Ä is also very interesting, since it can be used as a switching diode for high voltages with insignificant change. Such a switching diode is shown in Fig. 5. This semiconductor device can be constructed in the same way as the one shown in Fig. H, with the only difference that the zone H2 in this case does not have to be contacted and thus can be covered everywhere. of an insulating layer H1 and that it ensures that the breakdown voltage between the areas H7 and U2 is low. For this purpose, the area H2 is arranged at a small distance from the area H7 and possibly also from the area 47 or penetrates into the area H7.

A voltage V1 in the reverse direction is applied across the transition 5 via barrier-free contacts on the zones HU and 2. An impet- 7906289-9 1? dance, in the present case a resistor R, is connected in series with the voltage source V1. A variable voltage V2 in the reverse direction is applied across the transition H6. Fig. 5 illustrates the state in which the voltage V1 is still small and such a high voltage V2 is applied to the control electrode that the coordinated burst zone with the boundary line H5 reaches the boundary line H0 of the burst zone at the transition 5. Under these circumstances an island-shaped part 1A is surrounded of the failure zones and electrically blocked from the remaining part of the first area 1. The voltage V1 can now be increased to very high values, since the part 1A is completely emptied from the transition 5 up to the surface 8 already at a comparatively low voltage V1 and the breakdown voltage, when the voltage V1 is further increased, it is no longer determined by the comparatively low breakthrough voltage at the transition Å6 but by the breakthrough voltage at the flat transition 5, which does not penetrate to the surface; In this case, therefore, the gate electrode zone H5 and not the area H7 also serves as the above-mentioned third area. The high voltage V1 now occurs substantially completely over the burst zone between the surface 8 and the boundary line H9 and has almost the same distance as in Fig. H. Substantially no voltage drop occurs across the impedance B, since only low leakage current flows through it and this impedance is of much lower value than that with the same series-connected blocked semiconductor device. When the control voltage V2 is reduced to such an extent that the burst zone no longer blocks the area 1 between the zone H3 and the transition 5, an operating field is formed, as a result of which the zone H2 tends to reach the potential of the collector zone UR. Long before this can occur, however, breakthroughs occur between the regions H7 and U2, so that the voltage across the semiconductor device disappears almost completely - and the voltage V1 appears substantially completely across the impedance R.

In this way, the voltage across the impedance R can be switched between a low and a high value by means of the control voltage V2.

Fig. 6 is a schematic cross-sectional view of a vertical field effect transistor according to the invention, which consists of an island-shaped area 1, which in the present case is p-conducting. In this case, the region 1 forms part of a p-type epitaxial layer having a thickness of H micron and a doping concentration of 1, 3.105 atoms / cm 3, which is grown on an n-type substrate 2 having a thickness of 250 microns. micron and a doping concentration of 3.2 .mu.lu atoms / cm5. The area 1 is laterally limited 7906289-9 by an n-type diffused zone 5. Within the island 1, a pattern 50 of silica, which is partially immersed in the semiconductor material, is produced by selective heat oxidation in the form of an oxide layer provided with openings, completely surrounded by the oxide.

Inside the semiconductor material, the oxide 50 is bounded by a thin, heavily doped p-type zone 5U, which contacts outside the insulating pattern 50 and forms a gate electrode zone. The shortest distance between the zone 5U and the transition 5 amounts to 2.5 microns. Furthermore, a heavily doped layer 52 of n-type of polycrystalline silicon is arranged on the surface and contacts between the partially immersed oxide parts 50 the semiconductor surface at surface zones 53, which are produced by indiffusion from the layer 52. A metal layer 51 is arranged on the layer 52, while the area 2 are contacted by means of a strongly doped semiconductor layer 55 and a metal layer 56. The connections to the emitter, collector and control electrodes are schematically indicated by S, D and G.

In the drift state, a collector voltage is applied via D via a voltage which is positive in relation to the voltage supplied via S to the emitter electrode. A voltage which is at least so negative in relation to the collector electrode that the burst zone extends from the transition 5 between the areas 1 and 2 up to the surface appears at the connection G of the control electrode so that the area 1 is completely emptied. The current of electrons, which is moved from the emitter electrode to the collector electrode, is mainly not affected by the emptied area 1. By changing the voltage at the control electrode, the potential distribution can be changed within the emptied area 1 and can e.g. a potential threshold is formed so that the electron current from the emitter electrode to the collector electrode can be controlled via the emptied area 1.

Since the region 1 is completely emptied at a voltage which is lower than the breakdown voltage of the junction 6, a vertical field effect transistor for very high voltage is obtained, because due to the above principle the voltage at which breakthrough occurs between the regions 1 and 2 can be very high. .

The semiconductor device shown in Fig. 6 can be manufactured in the following manner. Starting material is an n-type substrate 2 with a p-type epitaxial layer with the above-mentioned dopings and thicknesses. The island-shaped insulating zone 3 is produced according to conventional diffusion methods e.g. by indiffusion of phosphorus. At the same time, the heavily doped n-type contact layer 55 is diffused on the underside. An antioxidant mask, which at the same time serves as an implantation mask and contains silicon nitride and is hereinafter referred to as nitride mask, is then produced in the form of a square frame, consisting of masking tapes with a width of H micron, which is located at a distance of 10 microns from each other. Then boron is implanted in a dose of 1015 ions / cm) with an energy of 60 keV. The photo varnish used for etching the mask remains and also serves as a mask against implantation. In this way, the p-type layer 5ü is formed.

Then the photoresist is removed and the oxide pattern is formed after heat treatment at 900 ° C for 30 minutes by heat oxidation to a thickness of e.g. l micron. Techniques for forming a recessed oxide pattern through selective oxidation are described in detail in Philips Research Reports, Volume 25, 1970, pp. Ll8-132. After the nitride mask is removed, a layer 52 of polycrystalline silicon is prepared to a thickness of 0.5 microns, which is doped to n-type e.g. by implanting phosphorus. Thereafter, heating is performed at 105,000 for 30 minutes in nitrogen, the channel areas 53 being formed by diffusion from the layer 52. The coating 51, 56, 57 of aluminum is then performed by steam precipitation and masking, if desired, after an additional p-type doping. performed to extend the layer 54 inside its contact window, after which the device can be provided with a housing. The distance L in Fig. 6 in the present case amounts to 70 microns. The one-dimensionally calculated breakdown voltage VB for the p + p_n_ combination SH, 1, 2 amounts to almost ace-volts. With f = 11.7 and E = 235,105 voit / cm for silicon, the condition B <236.102 g E L- _ N. <1 S 5,1.1o5E E is met. When the area 53 is weakly doped, control of the current between the emitters can and the collector electrodes also occur in that the pn junction between regions 5U and 63 forms a gap zone within region 52, which by changing the control voltage changes the cross section of the current path through region 53. In some circumstances, both this and the above mechanism may participate.

The invention is not limited to field effect transistors with pn junction or Schottky junction. be separated from the semiconductor surface by means of an insulating layer.

Fig. 7 shows in the form of a schematic cross-sectional view an example 7906289-9 lb of a transistor with deep discharge, which has exactly the same construction and operates in the same way as the transistor shown in Fig. 2 with the only difference that the burst zone in control the electrode with the boundary line lü is not formed by a pn junction but by a control electrode, consisting of an electrode layer 60, which is separated from the semiconductor surface by means of an insulating layer 61 of e.g. an oxide. Furthermore, in the device shown in Fig. 7, the same doping concentrations and dimensions and the same switching method can be used as in the device according to Fig. 2.

The above particularly suitable doping concentrations and dimensions will be further explained with reference to Figs. BA-E and 9.

Figs. 8A-E are schematic cross-sectional views for illustrating five different possibilities for field distribution in the diode, which correspond to the island-shaped part of the first area in the previous example. For the sake of clarity, only one half of the diode is shown, which is assumed to be rotationally symmetrical about the axis denoted by Es. The region 1 corresponds to the island-shaped part of the first region each of the preceding examples, while the transition 5 corresponds to the first pn junction and the transition 6 corresponds to the second pn junction. In the figures, the area 1 is assumed to be of the n-type, while the area 2 is assumed to be of the p-type, but the pipe types can also be reversed. The doping concentration for the area is the same in all Figures 8A-E. When a voltage is applied between the n - region 1 via the n + contact region U and the p - region 2 in the reverse direction over the transitions 5 and 6, a change in the field strength distribution ES along the surface according to line s takes place, while the field strength Eb in the vertical direction varies according to line b .

Fig. 8A illustrates the case where complete emptying of the layer 1 does not yet occur at the breakdown voltage. A high maximum value of the field strength ES appears at the surface at the transition 6, which due to the strong doping of the p + - region 3 is higher than the maximum value of the field strength Es, which seen in the vertical direction occurs at the transition 5. When the critical field strength E is exceeded , which for silicon amounts to about 2.5.105 volts / cm and is somewhat dependent on the doping, occurs breakthroughs at the surface near the transition 6, before the failure zone, as in Fig. 8A is indicated by dashed lines 9 and 10, extends in the vertical direction from the transition 5 up to the surface. Fig. SB-8E illustrates cases where the doping concentration N and the thickness d of the layer 1 are such that the layer 1, before surface breakthrough occurs at the transition 6, is completely emptied from the transition 5 to the surface. over a part of the path between the areas 5 and U, the field strength Es along the surface is constant, while peaks are formed in the field strength distribution both at the location of the transition 6 and the n + n transition at the edge of the area U due to the curvature of the edge at the later transition.

In the case shown in Fig. BB, the peak value is highest at junction 6 and higher than the maximum value of E 'at junction 5, so that breakthroughs will occur at this point at the surface but with comparatively higher values than that of Figs. 8A, because the field strength distribution at the surface is more homogeneous and the maximum values will thus decrease.

The case according to Fig. 8B can be obtained from the case of Fig. 8A e.g. by reducing the thickness d of the layer 1, while maintaining the same doping.

Fig. 80 shows the opposite case to Fig. 8B. In this case, the field strength peak at the edge of the area 4 is considerably higher than at the transition 6. This case can e.g. occur when the layer 1 has a very high specific resistance and the area 1 is emptied,> before the breakdown voltage occurs. In this case, breakthroughs can occur at the edge of the area H, when the maximum field strength at this edge is higher than the field strength at the junction 5.

The case shown in Fig. 8D is more favorable. In this case, it is ensured that the area 1 has such a doping concentration and thickness that the two field strength peaks are substantially equal at the surface. Although breakthrough at the surface will still occur, when according to Fig. 8D the maximum field strength Eb at the transition 5 is less than the maximum value at the surface, the maximum field strength at the surface will be lower in this case by causing the field strength distribution S at the surface to be symmetrical than at asymmetric field strength distribution, so that breakthrough occurs at a higher voltage.

Fig. 8E finally illustrates a case where the maximum field strength at the surface at any back voltage is lower than the maximum field strength at the transition 5 due to the appropriate choice of doping and thickness of the layer 1 and by increasing the distance L with a certain doping concentration at the area 2. As a result, the breakthrough in this case will constantly occur inside the semiconductor body at the junction and not at the surface. Furthermore, it can be pointed out that the field strength at too low a value of this distance L will increase at the surface (in fact, the total voltage between the areas 3 and H determines the area between the curve S and the line Es = 0), so that breakthrough occurs at lower voltage at the surface.

Calculations have shown that the most favorable values for the breakthrough voltage are obtained within the surface, which in Fig. 9 is enclosed by lines A and B. In Fig. 9, the product of the doping concentration is N in atoms per cm deposited along the horizontal axis for silicon as a semiconductor and the value of 106.L / VB with L in cm and VB in volts is deposited along the vertical axis. VB is the one-dimensionally calculated value of breakthrough voltages at junction 5, ie. in Figs. 8A-E the breakdown voltage of the n + n_p_ combination, assuming that the doping concentrations in regions 1 and 2 are homogeneous, so that the transition 5 is interrupted, that region U has substantially negligible series resistance and that said combination extends itself infinitely far in all directions at right angles to the axis ES. This fictitious breakthrough voltage VB can be very easily calculated with the said assumptions. For the purpose, see e.g. the article by S.M. Sze, Physics of Semiconductof Devices, Wiley & sons, New York, 1969, chap. 5.

In the case where silicon is selected as the semiconductor material, it is found that for values of N x d, which lie between the lines A and B, i.e. for V 7.6io8 V 55 S n.d S 1.5.1o12 the condition according to Fig. 8D with symmetrical field distribution at the surface is met. If the condition according to Fig. SE is also to be fulfilled with regard to symmetrical field distribution at the surface with breakthrough at the transition 5, the values for L, N and d should be chosen in such a way that they lie on or near the line C in Fig. 9. For L / VB 2 1, u.1o'5 mainly applies, that Nd = 9.l0l1 cm'2.

As already pointed out, the values given in Fig. 9 refer to silicon, which has a critical field strength E of about 2.5.105 volts per cm and a dielectric constant of about 11.7. For semiconductor materials with a relative dielectric constant E and a critical field strength E or generally between lines A and B that 2,6.1o2É 'si / go 5 Nd 5 5,1.1o5 ÉE and for line C, that Nd is essentially equal to 3; l05 EE and that in this fail also L / VB 2 1,4.1o'5. 7 The values Éh and E are found in the available literature, whereby with regard to the critical field strength E reference can be made to the above-mentioned article p. 117, Fig. 25.

The invention is not limited to the described embodiments. Thus, e.g. semiconductor materials other than silicon, insulating layers other than silicon oxide such as silicon nitride, alumina and metal layers other than aluminum are used. In each embodiment, the cable types can also be replaced by their opposite types. It can be pointed out here that, although the third region 3 in the examples described is always more heavily doped than the second region 2, this third region can also have the same doping concentration as the second region, so that it forms an elongation of the second area.

In such cases, the lower breakdown voltage at the second junction 6 is caused by the sharp curvature in the transition region between transitions 5 and 6. f;

Claims (15)

xq d 1906289-9 Patent claims A semiconductor device with a semiconductor body provided with a substantially flat surface, comprising at least one field effect transistor with an emitter and a collector electrode, a channel area located between these electrodes and a control electrode located adjacent to this area in order to use a control voltage supplied thereto. affect a failure zone for controlling the flow of charge carriers between the emitter and collector electrodes, the transistor comprising a layered first region of a first lead type, which together with an underlying second region of the second lead type forms a first pn junction, which extends substantially parallel to the surface, and at least during operation an island-shaped part of the first region is at least partially bounded laterally by a second pn junction with coordinated gap zone, which is formed between the first region and a third region of the second conduit type, which borders the first area, while this second transition has a lower breakthrough rate than the first junction and at least the control electrode adjoins the part of the island-shaped region, to which between the second region and a contact region of the transistor, which belong to the emitter, collector and control electrodes and together with this part form a non- rectifying contact, a voltage is applied in the reverse direction, characterized in that the doping concentration N in atoms / cm3 of and the thickness di cm of the part of the island-shaped area satisfies the condition B 2.6.1o2ÉE -ï 5 Nd 5 5.1.1o5 than, where S is the relative dielectric constant and E is the critical field strength in volts / cm, at which avalanche multiplication occurs in the semiconductor material of the first region, L is the distance in cm from the contact region up to the second pn junction and VB the one-dimensionally calculated value in volts of the breakthrough voltage of the first pn junction. Device according to claim 1, characterized in that N.d is substantially equal to 3,0.105 E E and L Z 1,4.10_5.VB. Device according to claim 1 or 2, characterized in that at least the part of the second area located near the first area 7906289-9 m has a lower doping concentration than the first area. N. Device according to claim 1, 2 or 3, characterized in that the second region has such a thickness that the failure zone at the breakdown voltage of the first transition extends into the second region over a distance which is less than the thickness of this area. Device according to any one of claims 1-H, characterized in that the first region is formed by an epitaxial layer of the first conductor type arranged on the second region. _ ' Device according to any one of claims 1-5, characterized in that the island-shaped part of the first area laterally is completely bounded by the second transition. Device according to one of Claims 1 to 6, characterized in that the control electrode consists of a semiconductor zone which forms a pn junction together with the adjacent part of the channel region. Device according to one of Claims 1 to 6, characterized in that the control electrode consists of a metal layer which, together with the adjacent part of the channel area, forms a Schottky junction. Device according to one of Claims 1 to 6, characterized in that the control electrode consists of a conductive layer which is separated from the adjacent part of the channel region by means of an insulating layer. 10. lO. Device according to one of Claims 1 to 9, characterized in that the transistor is of the lateral type, the emitter and collector electrodes on each side of the control electrode forming non-rectifying contacts together with the first region, while the contact region is formed by carbon - the lecturer electrode. 11. ll. Device according to one of Claims 10 to 10, characterized in that the control electrode is connected to the second region. 7 7 12. Device according to claim 10 or 11, characterized in that the collector electrode is substantially completely surrounded by the control electrode, which in turn is substantially completely surrounded by the emitter electrode. Device according to claim 12, characterized in that the semiconductor layer of the second lead type is located in the first region, that the emitter and collector electrons 7906239-9 consist of zones of the first lead type and the control electrode of a zone of the second wire type and that all electrode zones extend over the entire thickness of the semiconductor layer down to the first region. leeward. Device according to claim 12, characterized in that the emitter electrode consists of a zone of the first wire type, which is not connected to any external voltage source, that a heavily doped zone of the second wire type is located on the side of the control electrode the emitter zone and extends from the surface down to the second region and is located so close to the emitter zone that the breakdown voltage between these two zones is considerably lower than the breakdown voltage of the first junction, that the collector electrode and the second region are connected to a voltage source is connected in series with a load impedance and produces reverse voltage across the first junction, and that the control electrode is connected to a voltage source which causes variable reverse voltage between the control electrode and the first region, so that the island-shaped part of the first region, which is surrounded of the control electrode and the coordinated failure zone, is temporarily electrically barrierable from it the remaining part of the first area. Device according to one of Claims 1 to 9, characterized in that the transistor is of the vertical type, in that the collector electrode forms a non-rectifying contact together with the second region, in that the emitter electrode forms a rectifying contact together with the the first region and that the control electrode consists of a zone of the first lead type, which surrounds at least a part of the first region coordinated with the channel region and forms said contact area.