CH648693A5 - SEMICONDUCTOR ARRANGEMENT WITH AT LEAST ONE FIELD EFFECT TRANSISTOR. - Google Patents

Description

The invention relates to a semiconductor device according to the preamble of claim 1. A semiconductor device of this type is e.g. known from U.S. Patent 3,586,931.

Under the influence of an exhaustion zone for controlling the current is to be understood here, either that by changing the thickness of an exhaustion zone one of

2nd

5

10th

15

20th

25th

30th

35

40

45

50

55

j0

> 5

3rd

648 693

current channel limited in this depletion zone is narrowed or expanded, or that a current of charge carriers moving through this depletion zone is changed by changing the potential distribution in an exhaustion zone.

The field effect transistor mentioned can have different structures, depending on the shape of the source, drain and control electrodes. These electrodes can have the form of metal layers that form ohmic source and drain contacts and one or more rectifying control electrodes with Schottky contacts on the semiconductor surface. The source, drain and control electrodes can also be formed by metal layers which adjoin semiconducting electrode zones which, with the adjacent part of the semiconductor body, have pn junctions (in the case of control electrodes) or non-rectifying junctions (for the source and drain Electrodes). The control electrode, e.g. in the case of a so-called “deep depletion” field effect transistor, which have the form of a conductive layer which is separated from the semiconductor surface by an insulating layer and with which the exhaustion zone mentioned is formed in the channel region. Where reference is made in the present application to source, drain and control electrodes, the electrode zones or insulating layers which possibly belong to these electrodes must also be included.

In the known field-effect transistors of the type described, generally no high voltages can be applied across the first and the second pn junction. This includes attributable to the fact that long before the breakdown voltage of the first pn junction that can theoretically be expected on the basis of the doping profile is reached, breakdown already occurs at the second pn junction as a result of the unfavorable field distribution prevailing there. This breakdown mostly occurs on or in the immediate vicinity of the surface.

The unfavorable field distribution mentioned can be caused by a high doping of the third region and / or a high doping gradient in the vicinity of the second pn junction, but e.g. can also be brought about by the fact that the second pn junction locally has a strong curvature.

In order to increase the permissible voltage, the doping concentration of the first region can be reduced and, in addition, in order to obtain space for the depletion zone which thereby extends further in the first region, the thickness thereof can be increased. However, since the channel line is proportional to the thickness, but the pinch-off voltage is proportional to the square of the thickness of the channel area, this measure will result in the channel line being reduced if the length and width of the channel remain the same and the blocking voltage remains the same. For the pinch-off voltage Vp, namely: Vp = a2qN / 2e and for the channel line G = Wq p.Na/L, where a is the thickness of the channel region blocked by the control electrode, N the doping concentration of the channel region, W the width and L the length of the channel area, u represent the mobility of the charge carriers, q the electron charge and s the dielectric constant of the semiconductor material. If N is now reduced to a value N '= N / ß (ß> 1), Vp is found for constant pinch-off voltage:

a '= i / 2eßVp / qN = a ^ / ß and

WqUNa G ° - L, / (5 "<a

Such a reduction in the channel line is usually very disadvantageous for the good effect of the field effect transistor.

The invention has i.a. the task of creating a semiconductor arrangement with a flat surface, which contains a field effect transistor and which can be used at much higher voltages than known field effect transistors of the type described, without reducing the channel line.

The invention lies inter alia. based on the knowledge that, contrary to expectations, this can be achieved by not reducing the thickness of the first area, but reducing it.

According to the invention, a semiconductor arrangement of the type described has the features stated in the characterizing part of patent claim 1.

If the above-mentioned condition is met, the product of the doping concentration and the thickness of the first region is such that when the aforementioned blocking voltage is applied, the depletion zone is at least two-fold between the contact region and the second pn junction from the first pn junction entire thickness of the island-shaped region at a lower voltage than the breakdown voltage of the second pn junction.

Said contact area can be an electrode or electrode zone which is directly connected to the source of the reverse voltage mentioned, but can also e.g. be a semiconductor zone, which itself is not provided with a connecting conductor, but in another way, e.g. is brought to the desired potential via an adjoining semiconductor zone.

The fact that the island-like region of the first conductivity type between said contact region and the second pn junction is already completely depleted at a voltage lower than the breakdown voltage of the second pn junction means that the field strength on the surface is reduced in such a way that the breakdown voltage is no longer is determined practically entirely by this second pn junction, but to a large extent by the first pn junction running parallel to the surface.

In this way, a very high breakdown voltage can be obtained between the first and the second region, which can possibly come close to the high breakdown voltage that can theoretically be expected on the basis of the doping of the first and the second region.

The condition also prevents that when the voltage between the first and the second region is increased, an excessive field strength occurs prematurely on the surface between the contact region and the second pn junction due to the fact that the depletion zone of the second pn junction penetrates to this contact area. An optimal field strength distribution can be achieved by the fact that the maxima in the field strength that occur at the second pn junction and at the edge of the mentioned contact area are approximately of the same order of magnitude for the N d product.

If the conditions are also selected in a preferred manner such that N • d «3.0 • 105 eE and L> 1.4-10 ~ 5 VB, it is certain that the maximum field strength at the first pn junction is always will be larger than in the above-mentioned maxima occurring on the surface, so that the breakdown always occurs at the first pn junction and not on the surface.

Although the depletion zone of the first pn junction can in many cases extend without concern over the entire thickness of the second region, it is preferably ensured that the second region is so thick that the breakdown voltage of the first pn junction increases

5

10th

15

20th

25th

30th

35

40

45

50

55

60

65

648 693

4th

Exhaustion zone in the second area extends for a distance which is less than the thickness of this area. In this case, it is certain that the breakdown voltage cannot be adversely affected by the thickness of the second region.

Although the semiconductor structure described can also be formed in other ways, i.a. for technological reasons, the embodiment is preferred in which the first region is formed by an epitaxial layer of the first conductivity type which is produced in the second region.

The third area, which borders the first area, need not extend over the entire thickness of the first area. It is sufficient that, at least in the operating state, the associated exhaustion zone extends over the entire thickness of the first region and delimits an island-shaped part thereof over at least part of its circumference. However, the island-shaped part of the first region is preferably laterally completely delimited by the second pn junction, although other structures are sometimes preferred in which the first region is laterally e.g. partly from the second pn junction and the rest partly in another way, e.g. from a sunken insulation material or from one with e.g. passivating glass filled groove, is limited.

The invention is of particular importance in the case of lateral field effect transistors in which the current between the source electrode and the drain electrode flows practically parallel to the surface. A preferred embodiment is also characterized in that the source and drain electrodes form non-rectifying contacts with the first region on both sides of the control electrode,

said contact area being the drain electrode of the transistor. In most cases, the control electrode is connected to the second region, which then acts as a second control electrode, although this is not necessary.

In certain cases, an embodiment is preferred in which the drain electrode is virtually completely surrounded by the control electrode and the latter electrode is virtually completely surrounded by the source electrode. A particularly preferred embodiment is characterized in that a semiconductor layer of the second conductivity type is located in the first area, that the source and drain electrodes contain electrode zones of the first conductivity type and the control electrode contains an electrode zone of the second conductivity type, and that all the electrode zones mentioned are located extend through the entire thickness of said semiconductor layer to the first region. The latter preferred embodiment makes it possible to use complementary transition field effect transistors in the same semiconductor wafer, i.e. to produce n-channel and p-channel field effect transistors side by side, as will be described below.

In addition to lateral transition field effect transistors, the invention can also be used to advantage with so-called vertical type transition field effect transistors. In this context, a preferred embodiment is characterized in that the field effect transistor is of the vertical type, that the drain electrode forms a non-rectifying contact with the second region, that the source electrode forms a rectifying contact with the first region and that Control electrode contains an electrode zone of the first conductivity type, the min. least surrounds a part of the first area belonging to the channel area and forms the said contact area.

Some embodiments of the invention are shown in the drawing and are described in more detail below. Show it :

1 schematically, partly in cross section and partly in perspective, a known semiconductor arrangement,

2 schematically, partly in cross section and partly in perspective, a semiconductor arrangement according to the invention, FIG. 3 another embodiment of the semiconductor arrangement according to the invention,

4 and 5 further embodiments of the semiconductor device according to the invention,

6 shows a semiconductor arrangement with a vertical field effect transistor according to the invention,

7 shows a “deep depletion” field effect transistor according to the invention,

! 8A to E the field distribution with different dimensions and doping, and

9 shows the relationship between the doping and the dimensions of the first region for a preferred embodiment.

The figures are schematic and are not drawn to scale for the sake of clarity. Corresponding parts are generally designated by the same reference numerals. Semiconductor regions of the same conductivity type are usually hatched in the same direction.

In all exemplary embodiments, silicon is selected as the semiconductor material. However, the invention is not limited to this, but can be made using any other suitable semiconductor material, e.g. Germanium or a so-called III-V compound, such as GaAs, can be used.

1 shows a known semiconductor arrangement, partly in section and partly in perspective. The arrangement contains a semiconductor body with a field effect transistor, which is provided with a source and a drain electrode with associated electrode zones 12 and 4, an intermediate channel region 1 and a control electrode bordering on the channel region 1 with associated electrode zone 13. This control electrode serves to influence an exhaustion zone for controlling a current of charge carriers, in this example an electron current, between the source electrode 12 and the drain electrode 4 by means of a control voltage applied to the control electrode. In the present example, the source electrode, the drain electrode and the control electrode all consist of a semiconductor zone and a metal layer attached thereon, which forms an ohmic contact with the associated electrode zone and is not shown in the drawing for the sake of clarity. The channel region 1 is n-conducting in the present example; the electrode zones 12 and 4 are n-conducting with a higher doping than the region 1 and the control electrode zone 13 is p-conducting and forms a rectifying pn junction 7 with the channel region 1.

As can be seen from FIG. 1, the field effect transistor contains a layered first region 1 of a first (in this case the n-) conduction type. This first region 1, which in the case described here is at the same time the channel region bordering the control electrode, forms, with an underlying p-conducting second region 2, a first pn junction 5 which is practically parallel to the surface 8. An island-shaped part of the region 1 is laterally delimited by a second pn junction 6 with the associated exhaustion zone. This second pn junction 6 is formed between the first region 1 and a p-type third region 3 which extends between the second region 2 and the surface 8 and has a higher doping concentration than the second region 2. The pn junction 6 thus has a lower breakdown voltage than the first pn junction 5. The control electrode borders on the island-shaped part of area 1.

As shown in FIG. 1, the control electrode is connected to the substrate (in this case the second region 2)

10th

15

20th

25th

30th

35

40

45

50

55

60

65

5

648 693

the, although this is not necessary. When a voltage VD is applied between terminals S and D of the source and drain electrodes, electrons flow through region 1 from zone 12 to zone 4. By applying a voltage in the reverse direction between control electrode zone 13 and region 1 and between the second area 2 and area 1, exhaustion zones are obtained, the boundaries (9, 10, 14) of which are indicated by dashed lines in FIG. 1. These zones of exhaustion are shown without hatching.

In the known arrangement described above, the doping concentrations and the dimensions are such that the area 1 at the drain electrode 4 is not exhausted at the breakdown voltage of the pn junction 6. The voltage in the blocking direction across the pn junctions 6 and 7, which is highest in the vicinity of the drain electrode 4, leads to a field strength distribution in which the maximum value of the field strength occurs in the vicinity of the point at which the pn- Transitions 6 and 7 intersect the surface 8, and finally, in the vicinity of this surface, breakdown occurs at a voltage which is considerably lower than the breakdown voltage of the pn junction 5 within the volume of the semiconductor body.

2 shows a semiconductor arrangement according to the invention. This arrangement is largely the same as the known arrangement according to FIG. 1. According to the invention, however, the doping concentration and the thickness of the first region 1 are so small in the arrangement according to FIG. 2 that when a voltage is applied in the blocking direction between the second region 2 and one belonging to the source, drain and control electrodes in the island-shaped area forming a non-rectifying contact, in this case the drain electrode 4, the field effect transistor, the depletion zone is at least between the drain electrode

4 and the second pn junction 6 extends from the first pn junction 5 over the entire thickness of the island-shaped region 1 at a voltage which is lower than the breakdown voltage of the second pn junction 6. 2 shows the state in which the area 1 between the zones 7 and 4 up to the pn junction 6 is completely exhausted. The voltage across the pn junctions 5, 6 and 7 is now completely absorbed by the coherent, wide exhaustion zone, which extends from the drain zone 4 to the limit 9. As a result, the field strength on the surface is considerably reduced. The breakdown voltage is then determined to a considerable extent, if not primarily, by the properties of the pn junction running within the volume of the semiconductor body

5 determined. This breakdown voltage can be very high and very close to the breakdown voltage that can theoretically be expected due to the doping of regions 1 and 2.

In order to achieve the result described by the invention, the following doping and dimensions are used in the arrangement according to FIG. 2, which contains a semiconductor body made of silicon:

Zones 4 and 12: thickness 1 (in;

Area 1: n-conducting, doping concentration 1.5 * 1015 atoms / cm3, thickness 5 [im;

Area 2: p-type, doping concentration 1.7 * 1014 atoms / cm3, thickness 250 (im;

Zone 13: p-type, 2.5 µm thick;

Distance L of the drain electrode 4 from the pn junction 6:50 Jim.

The one-dimensionally calculated breakdown voltage VB of the first pn junction is 1270 V in the present case. The real breakdown voltage has been found to be 700 V. At the given thicknesses and doping concentrations, the depletion zone in region 2 extends over a thickness that is less than the thickness of the area 2 is, while also preventing the depletion zone of the pn junction 6 from reaching the zone 4 at a voltage value that is smaller than the breakdown voltage of the pn junction 6 per se (i.e. in the absence of the pn junction) 5), is.

With the stated values for N, d, L and VB, the condition for silicon (e = 11.7; E = 2.5 • 105 V / cm) is thus fulfilled:

2.6-102 eE / VB7L <N-d <5, l • 105 eE.

2, the first region 1 is formed by an epitaxial layer deposited on the second region 2. In the present example, the island-shaped part of the first region is completely delimited laterally by the second pn junction 6. Technologically, this is the easiest, but is not necessary. The island-shaped area can e.g. over part of its circumference in other ways, e.g. of a buried oxide pattern or of one with e.g. passivating glass filled groove, limited.

In the arrangements according to FIGS. 1 and 2, the control electrode forms a rectifying contact and the source and drain electrodes form non-rectifying contacts with the region 1 by means of the doped surface zones 12, 4 and 13. The presence of these surface zones is but not absolutely necessary; Instead of the semiconductor zones 12 and 4, ohmic metal-semiconductor contacts and instead of the zone 13, a rectifying metal-semiconductor contact (Schottky contact) can be produced in the area 1. Instead of a control electrode with a rectifying transition, a conductive layer separated by an insulating layer from the semiconductor surface 8 can also be used, with which an exhaustion zone is formed in the epitaxial layer 1, e.g. is the case with a deep depletion transistor.

FIG. 3 shows how the invention can be used to arrange a p-channel and an n-channel junction field effect transistor (JFET) next to one another in the same monolithic integrated circuit.

I denotes a p-channel field effect transistor which is basically the same as the field effect transistor according to FIG. 2, but in which the conduction type of all corresponding semiconductor zones is opposite to that in FIG. 2. Furthermore, the second region 2 of this transistor is formed by an n-type epitaxial layer which is deposited on a p-type substrate 34. A highly doped n-type buried layer 36 is located between the epitaxial layer 2 and the substrate 34 in order to prevent the depletion zone associated with the pn junction 5 from reaching the substrate 34.

In addition to the field effect transistor I, a second transition field effect transistor II is arranged. This is also a field effect transistor according to the invention. This second transistor II also contains an island-shaped region 32 which is formed by part of the same epitaxial layer from which region 2 of transistor I is formed. The n-type source zone 22, the n-type drain zone 24 and the p-type control electrode zone 23 extend over the entire thickness of the p-type semiconductor layer 21 lying on the island 32, from which region 1 is also formed of transistor I is formed up to the n-type region 32. The source and drain zones 22 and 24 form with the region 21 the pn junctions 26 and 26A and the regions 21 and 32 form the pn junction 39 In this second field effect transistor, the channel region is formed by region 32. For the mutual isolation of transistors I and II

5

10th

15

20th

25th

30th

35

40

45

50

55

60

65

648 693

6

highly doped p-type region 33 is provided which completely surrounds both region 2 and region 32 and forms the pn junction 38 with region 32.

When a suitable voltage is applied between the source zone 22 and the drain zone 24, electrons move through the region 32 from the source zone to the drain zone. This electron current can be influenced by the application of a control voltage in the reverse direction between the zone 23 and the region 32 (and possibly also by the reverse voltage between the regions 32 and 34). The doping concentration and the thickness of the layer (2.32), as in the example according to FIG. 2, are chosen in accordance with the condition according to the invention, so that a long time before breakdown occurs, the region 1 at least between the drain zone 4 and the pn junction 6 and the region 32 are at least completely exhausted between the drain zone 24 and the pn junction 27. As a result, the field strength on the surface 8 and, in the case of the transistor II, the field strength on the surface 39 between the regions 21 and 32 is considerably reduced and the breakdown voltage is considerably increased.

3, like in FIG. 2, the insulating (oxide) layers and contact layers present on the surface are not shown. The source, drain and control electrode connections are shown schematically with S, D and G.

4 shows a further development of the semiconductor arrangement according to the invention. Here, as in the second field effect transistor II of FIG. 3, the n-type drain zone 44 is surrounded by the p-type control electrode zone 43, which in turn is surrounded by the n-type source zone 42. All electrode zones are arranged within an island-shaped n-type first region 1, which forms a first pn junction 5 with an underlying second p-type region 2 and with a highly doped p-type region 47 a pn junction 48 ending at the surface 8 . The source, drain and control electrode zones 42, 44 and 43 only extend over part of the thickness of the first region 1. The field effect transistor can be operated in the same way as the preceding transistor; the limits (49 and 40) of the exhaustion zone indicated in the figure are shown for a reverse voltage between regions 1 and 2 which is lower than the breakdown voltage. The area 1 between the control electrode zone 43 and the drain zone 44 is completely exhausted. The island-shaped part of the first area is here, as in the field effect transistor II of FIG. 3, surrounded by the control electrode, which in this case fulfills the function of the "third area"; the pn junction 46 between the control electrode zone and the area 1 form the “second” pn junction. By choosing the doping and the thickness of the first region 1 in accordance with the invention, so that said island-shaped region is completely exhausted with increasing gate-drain voltage before breakdown of the pn junction 46 occurs, the field effect transistor can very high voltage can be used between the control electrode and drain electrode.

The arrangement according to FIG. 4 is also particularly interesting because, with a small change, it can be used as a switching diode for high voltages. Such a switching diode is shown in Fig. 5. The semiconductor structure of this arrangement can be the same as that of FIG. 4, with the difference that in this case the zone 42 need not be contacted and can therefore be covered everywhere by an insulating layer 41 and that the breakdown voltage is ensured between areas 47 and 42 is low. In order to achieve the latter, the area 42 is applied at a short distance from the area 47, possibly even adjacent to the area 47 or penetrating into the area 47.

A voltage Vi in the reverse direction is applied across the pn junction 5 via ohmic contacts on the zones 44 and 2. An impedance, in this example a resistor R, is connected in series with the voltage source Vi. Furthermore, a variable voltage V2 is applied in the blocking direction across the pn junction 46.

5 shows the state in which the voltage Vi is still low and in which a voltage V2 is applied to the control electrode such that the associated fatigue zone (limit 45) has reached the fatigue zone limit 40 of the pn junction 5. Under these conditions, an island-shaped part 1A is surrounded by the exhaustion zones and is electrically isolated from the remaining part of the first region 1.

The voltage Vi can now be increased to very high values, because even at a relatively low voltage Vi, the island-shaped region 1A is completely exhausted from the pn junction 5 to the surface 8, and if the voltage Vi is further increased, the breakdown voltage no longer results from that relatively low breakdown voltage of the pn junction 46, but is determined by that of the flat pn junction 5 which does not appear on the surface. In this case, too, the control electrode zone 43 and not the region 47 fulfills the function of the aforementioned «third region».

The high voltage Vi is now practically completely above the exhaustion zone between the surface 8 and the boundary 49 and the exhaustion zone has approximately the course shown in FIG. 4. There is practically no voltage across the impedance R because it is only traversed by a small leakage current and is much smaller than that of the blocked semiconductor arrangement lying in series with it.

If the control voltage V2 is now reduced to such an extent that the depletion zone no longer closes the area 1 between the control electrode zone 43 and the pn junction 5, a drift field arises, as a result of which the source zone 42 will attempt to test the potential of the drain zone 44 to reach. Long before this can occur, breakdown occurs between the regions 47 and 42, as a result of which the voltage across the semiconductor arrangement is practically completely neutralized and the voltage Vi comes to be virtually completely above the impedance R.

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

In Fig. 6, a vertical field effect transistor according to the invention is shown schematically in cross section. This consists of an island-shaped region 1, which is p-conducting in the present example. Region 1 here is part of a p-type epitaxial layer with a thickness of 4 μm and a doping concentration of 1.3 · 1015 atoms / cm 3, which is on an n-type substrate 2 with a thickness of 250 Jim and a doping concentration of 3.2-1014 atoms / cm3 has grown. The island-like region 1 is laterally delimited by an n-type diffused zone 3. Within the island 1, a pattern of silicon oxide 50 partially buried in the semiconductor material is produced in the form of an oxide layer by selective thermal oxidation, in which openings are completely surrounded by the oxide. The oxide 50 is delimited within the semiconductor material by a thin, highly doped p-conducting zone 54, which is contacted outside the insulating pattern 50 and forms the control electrode zone. The smallest distance between zone 54 and pn junction 5 is 2.5 im.

Furthermore, a highly doped n-type layer 52 made of polycrystalline silicon is produced on the surface, which has a semiconductor surface between the buried oxide parts 50

5

10th

15

20th

25th

30th

35

40

45

50

55

60

65

7

648 693

is in contact with surface zones 53 which are obtained by diffusion from the layer 52. A metal layer 51 is produced on the layer 52, while the region 2 is contacted via a highly doped semiconducting contact layer 55 and a metal layer 56. The connections of the source, drain and control electrodes are denoted schematically by S, D and G, respectively.

In the operating state, a voltage which is positive with respect to the source electrode S is applied to the drain electrode D. At least one negative voltage is present on the control electrode G with respect to the drain electrode such that the exhaustion zone extends from the pn junction 5 between the regions 1 and 2 to the surface, so that the region 1 is completely exhausted . The electron current which moves from the source electrode to the drain electrode is then practically not hindered by the exhausted area 1. By changing the voltage on the control electrode, the potential distribution within the depleted area 1 can be changed and e.g. a potential threshold is formed, whereby the electron current from the source electrode to the drain electrode can be regulated via the depleted region 1. Because region 1 is completely depleted at a voltage below the breakdown voltage of pn junction 6, a vertical field effect transistor for very high voltage can be obtained because, due to the principle set out above, the voltage at breakdown between regions 1 and 2 occurs can be very high.

6 can be manufactured in the following way. An n-type substrate 2 with a p-type epitaxial layer with the aforementioned dopings and thicknesses is assumed. By usual diffusion methods, e.g. the island isolation zone 3 is produced by phosphorus diffusion. At the same time, the highly doped n-type contact layer 55 is diffused on the underside.

Then, an antioxidant mask (which also serves as an implantation mask) containing silicon nitride, hereinafter referred to as a nitride mask, is attached in the form of a square grid consisting of masking strips with a width of 4 µm which are spaced 10 µm apart. Boron is then implanted with a dose of 1015 cm-2 and an energy of 60 keV. The photoresist that is used to etch the photomask is retained and also serves as a mask against the implantation. This creates the p-type layer 54.

Then the photoresist is removed and after annealing at 900 ° C for 30 minutes, the oxide pattern 50 with a thickness of e.g. 1 µm. The techniques for generating a buried oxide pattern by selective oxidation are described in detail in “Philips Research Reports”, Volume 25, 1970, pp. 118-132. After removal of the nitride mask, a layer 52 of polycrystalline silicon with a thickness of 0.5 μm is produced, which e.g. is n-doped by phosphorus implantation. The mixture is then heated in nitrogen at 1050 ° C. for 30 minutes, the channel regions 53 being produced by diffusion from the layer 52. The aluminum metallization (51, 56, 57) is then applied by vapor deposition and masking, and the arrangement can be fully assembled in a casing.

The distance L (see FIG. 6) is 70 | i.m in the present example. The one-dimensionally calculated breakdown voltage VB of the p + p_n_ structure (54, 1.2) is approximately 688 V. With (for silicon) e = 11.7 and E = 2.5 vi05V / cm the condition is thus met:

2.6-102 eE i / Vß7L <N-d <5, l • 105 eE.

If the region 53 is lightly doped, regulation of the current between the source and drain electrodes can also occur, in that the pn junction between the regions 54 and 53 in the region 53 forms an exhaustion zone which is changed by the control voltage changes the cross section of the current path through region 53. Under certain circumstances, both this mechanism and the aforementioned mechanism can play a role.

The invention is not limited to field effect transistors with a pn junction or Schottky junction. For example, the control electrode be separated from the semiconductor surface by an insulating layer. As an example, FIG. 7 shows schematically in section a “deep depletion” transistor whose structure and effect are completely identical to those of the transistor according to FIG. 2, with the only difference that the zone of exhaustion of the control electrode (boundary 14) does not pass through a pn junction, but is formed by a control electrode which consists of an electrode layer 60 which is separated from the semiconductor surface by an insulating layer (for example an oxide layer) 61. Furthermore, the same doping concentrations and dimensions and the same switching method as in FIG. 2 can be used in the arrangement according to FIG. 7.

The above-mentioned preferred doping concentrations and dimensions will now be explained in more detail with reference to FIGS. 8A to 8E and 9.

8A to 8E schematically show in cross section five different possibilities for the field distribution in a diode, which corresponds to the island-shaped part of the first region in the previous examples. For the sake of clarity, only half of the diode is shown; the diode is intended to be rotationally symmetrical about the axis labeled "Es". The area 1 corresponds to the island-shaped part of the “first area” in each of the preceding examples, the pn junction 5 to the “first pn junction” and the pn junction 6 to the “second pn junction”. In the figures, the region is assumed to be in-conducting and region 2 to be p-conducting; however, the line types can also be reversed. The doping concentration of region 2 is the same in all figures.

If a voltage is then applied in the reverse direction across the pn junctions 5 and 6 between the n ~ region 1 (via the n + contact region 4) and the p ~ region 2, a course of the field strength distribution Es longitudinal occurs along the surface the line S, while in the vertical direction the field strength Eb runs along the line B.

8A shows the case in which the breakdown voltage does not yet completely exhaust layer 1. A high maximum value of the field strength Es occurs on the surface at the pn junction 6, which due to the high doping of the p + region 3 is higher than the maximum value of the field strength Eb that occurs at the pn junction 5, viewed in the vertical direction . When the critical field strength E is exceeded (for silicon about 2.5 • 105 V / cm and somewhat dependent on the doping), breakdown occurs on the surface at junction 6 before the exhaustion zone (dashed in FIG. 8 A with 9 and 10 designated) extends in the vertical direction from the transition 5 to the surface.

8B to 8E show cases in which the doping concentration N and the thickness d of the layer 1 are such that before the occurrence of surface breakdown at the junction 6 the layer 1 is completely exhausted from the junction 5 to the surface. Over part of the route between areas 3 and 4, the field strength Es is constant along the surface, while both at the point

5

10th

15

20th

25th

30th

35

40

45

50

55

60

65

648 693

8th

of the pn junction 6 and of the n + n ~ junction at the edge of area 4 (due to the curvature of the n + n ~ junction) form peaks in the field strength distribution.

In the case shown in FIG. 8B, the peak value at transition 6 is highest and higher than the maximum value of Eb at transition 5, so that there will be breakdown on the surface, but at relatively higher values than in the case according to FIG. 8A , because the field strength distribution on the surface is more homogeneous and the maxima decrease as a result. The case of Fig. 8B can be seen from the case of Fig. 8A e.g. can be obtained by reducing the thickness d of the layer 1 while the doping remains the same.

Fig. 8C shows the reverse case with respect to Fig. 8B. In this case the field strength peak at the edge of area 4 is much higher than at the pn junction 6. This case can e.g. result if the layer 1 has a very high specific resistance and the area 1 is already exhausted before the breakdown voltage occurs. In this case, breakdown 4 can occur at the edge of the area if the maximum field strength at this edge is higher than at the pn junction 5.

The case shown in FIG. 8D is more favorable. It is ensured here that the doping concentration and the thickness of the region 1 are such that the two field strength peaks on the surface are practically identical to one another. In this case, although breakdown will still occur on the surface if, as shown in FIG. 8D, the maximum field strength Eb at the pn junction 5 is smaller than the maxima at the surface, the field strength distribution S at the surface is made symmetrical, the maximum field strength on the surface is lower than with an asymmetrical field strength distribution, so that the breakdown occurs at higher voltage.

Finally, FIG. 8E shows a case in which, by appropriately choosing the doping and thickness of layer 1 and by increasing the distance L for a given doping concentration of region 2, the maximum field strength at the surface at any reverse voltage is lower than the maximum field strength at pn - transition is 5. In this case, the breakdown will always occur within the semiconductor body at the pn junction 5 and not on the surface.

It should also be noted that if this distance L is too small, the field strength at the surface will increase (because the total voltage between areas 3 and 4 determines the surface between curve S and line Es = 0), so that at lower Voltage breakdown occurs on the surface.

Calculations have shown that the most favorable breakdown voltage values are obtained within the area enclosed by lines A and B in FIG. 9. In FIG. 9, the product of the doping concentration N in atoms / cm 3 and the thickness d in cm of region 1 is plotted as the abscissa for silicon as semiconductor and the value of 106 • L / VB with L in cm and VB in volts is plotted as ordinate. VB is the one-dimensionally calculated value of the breakdown voltage of the pn junction 5, i.e. 8A to E the breakdown voltage of the n + p-p structure, if it is assumed that the doping concentration of regions 1 and 2 are homogeneous and thus the pn junction 5 is rugged,

that the n + region 4 has a practically negligible resistance and that the n + n-p ~ structure extends infinitely far in all directions perpendicular to the axis Es. This imaginary breakdown voltage VB can be calculated very easily using the assumptions mentioned (see, for example, S.M. Sze, "Physics of Semiconductor Devices", Wiley and Sons, New York 1969, Chapter 5).

If silicon is chosen as the semiconductor material, it then turns out that for values of N • d which lie between the lines A and B, i.e. For

7.6 • 108 / VßTL <N • d «1.5 • 1012

the condition of FIG. 8D (symmetrical field distribution on the surface) is fulfilled.

If the condition according to FIG. 8E is also to be fulfilled (symmetrical field distribution on the surface with breakdown at the pn junction 5), values for L, N and d should be chosen which are on or close to the line C of FIG. 9 . For L / VB> 1.4-10 ~ 5 the following practically applies: n-d = 9-10 "cm-2.

As already mentioned, the values in FIG. 9 apply to silicon, which has a critical field strength E of approximately 2.5 * 10s V / cm and a dielectric constant e of approximately 11.7. In general, for semiconductor materials with a relative dielectric constant 8 and a critical field strength E, 2.6-102 eE / Vg7 L <Nd <5, l · 105eE between lines A and B and practically equal to line C nd 3 • 105sE and here also L / VB> 1.4 • 10 "5.

The values e and E can be readily taken from the known literature by the person skilled in the art. For the critical field strength E, for example on S.M. Sze, "Physics of Semiconductor Devices", Wiley and Sons, New York 1969, p. 117, Fig. 25.

With the aid of the data given above with reference to FIGS. 8A to E and 9, the person skilled in the art can select the dopings and dimensions which are most favorable under the given conditions for all semiconductor structures described in the preceding examples. It will not always be necessary or desirable that surface breakthrough is avoided under all circumstances (FIG. 9, curve C) as long as one stays only within or on lines A and B of FIG. 9.

The invention is not restricted to the exemplary embodiments described. For example, semiconductor materials other than silicon, insulating layers other than silicon oxide (e.g. silicon nitride, aluminum oxide) and metal layers other than aluminum can be used. In each exemplary embodiment, the line types can also be replaced by the opposite types.

It should also be noted that, although in the examples the third region 3 is doped higher than the second region 2, this third region can also have the same doping concentration as the region 2, so that it forms a continuation of this second region. In this case, the lower breakdown voltage of the second pn junction 6 is brought about by the strong curvature in the transition region between the first pn junction 5 and the second pn junction 6.

5

10th

15

20th

25th

30th

35

40

45

50

55

60

G

6 sheets of drawings

Claims (14)

648 693 1. Semiconductor arrangement with a semiconductor body with a practically flat surface with at least one field effect transistor with a source electrode and a drain electrode, an intermediate channel region and a control electrode adjoining the channel region, in order to control an exhaustion zone by means of a control voltage applied to the control electrode influencing a current of charge carriers between the source and drain electrodes, the field effect transistor containing a layered first region of a first conductivity type, which forms a first pn junction which runs practically parallel to the surface with an underlying second region of the second conductivity type, wherein, at least in the operating state, an island-shaped part of the first area is laterally at least partially delimited by a second pn junction with an associated exhaustion zone, which borders between the first area and one adjacent to the first area is formed in the third region of the second conductivity type, this second pn junction having a lower breakdown voltage than that of the first pn junction, at least the control electrode being adjacent to the island-shaped part and between the second region and a contact region of the field effect transistor, which belongs to the source, drain and control electrodes and forms a non-rectifying contact with the island-shaped area, a voltage can be applied in the reverse direction, characterized in that the doping concentration N in atoms / cm3 and the thickness d in cm of the island-shaped area meet the condition: 2.6-102 eE / VbTL «N-d <5.1 -105 eE, where e is the relative dielectric constant and E is the critical field strength in V / cm at which avalanche multiplication occurs in the semiconductor material of the first region, while L is the distance in cm of said contact region from the second pn junction and VB is the one-dimensionally calculated value of the Represent the breakdown voltage of the first pn junction in volts. 2. Semiconductor arrangement according to claim 1, characterized in that N • d practically the same 3.0-105 eE and L> 1.4-10 ~ 5-VB. 3. Semiconductor arrangement according to one of the preceding claims, characterized in that the second region is so thick that at the breakdown voltage of the first pn junction, the depletion zone in the second region extends over a distance which is smaller than the thickness of this region . 4. Semiconductor arrangement according to one of the preceding claims, characterized in that the first region is formed by an epitaxial layer of the first conductivity type produced on the second region. 5. Semiconductor arrangement according to one of the preceding claims, characterized in that the island-shaped part of the first region is laterally completely delimited by the second pn junction. 6. Semiconductor arrangement according to one of the preceding claims, characterized in that the control electrode contains a semiconducting control electrode zone which forms a pn junction with the adjacent part of the channel region. 7. Semiconductor arrangement according to one of claims 1 to 5, characterized in that the control electrode contains a metal layer which with the adjacent part of the channel region a rectifying metal semiconductor Transition forms. 8. Semiconductor arrangement according to one of claims 1 to 5, characterized in that the control electrode contains a conductive layer which is separated by an insulating layer from the adjacent part of the channel region. Semiconductor device according to one of the preceding claims, characterized in that the field effect transistor is of the lateral type, the source and drain electrodes forming non-rectifying contacts on both sides of the control electrode with the first region, and wherein said contact region is formed by the drain electrode. 10. Semiconductor arrangement according to one of claims 1 to 9, characterized in that the control electrode is connected to the second region. 11. The semiconductor arrangement according to claim 9 or 10, characterized in that the drain electrode is virtually completely surrounded by the control electrode, and that the control electrode is virtually completely surrounded by the source electrode. 12. The semiconductor arrangement as claimed in claim 11, characterized in that there is a semiconductor layer of the second conductivity type in the first region, that the source and drain electrodes contain electrode zones of the first conductivity type and the control electrode contains an electrode zone of the second conductivity type, and that all of these are mentioned Electrode zones extend through the entire thickness of said semiconductor layer to the first region (FIG. 3). 13. The semiconductor arrangement according to claim 11, characterized in that the source electrode contains a source zone of the first conductivity type, which is not connected to an external voltage, that on the side of the source zone remote from the control electrode, a highly doped zone of the second There is a conduction type that extends from the surface to the second region and is so close to the source zone that the breakdown voltage between these two zones is considerably lower than that of the first pn junction, that the drain electrode and the second Area can be connected to a voltage source, which is connected in series with a load impedance and supplies a reverse voltage which is above the first pn junction, and that the control electrode can be connected to a voltage source which provides a variable reverse voltage between the control electrode and the first area, whereby the of the control electrode and the associated exhaustion zone surrounding the island-shaped part of the first area can be temporarily electrically separated from the rest of the first area (Fig. 5). 14. A semiconductor device according to one of claims 1 to 8, characterized in that the field effect transistor is of the vertical type, that the drain electrode forms a non-rectifying contact with the second region, that the source electrode makes a rectifying contact with the first region forms and that the control electrode contains an electrode zone of the first conductivity type which surrounds at least a part of the first region belonging to the channel region and forms the said contact region.