US3069297A

US3069297A - Semi-conductor devices

Info

Publication number: US3069297A
Application number: US787195A
Authority: US
Inventors: Beale Julian Robert Anthony
Original assignee: US Philips Corp
Current assignee: US Philips Corp; North American Philips Co Inc
Priority date: 1958-01-16
Filing date: 1959-01-16
Publication date: 1962-12-18
Anticipated expiration: 1979-12-18
Also published as: FR1225692A; NL121250C; BE574814A; DE1090770B; NL235051A; CH370165A; GB911292A

Description

Dec. 18, 1962 J. R. A. BEALE SEMI-CONDUCTOR DEVICES 2 Sheets-Sheet l Filed Jan. 16, 1959 INVENTOR JULIAN R. A. BEALE AGEN BY M Dec. 18, 1962 J, BEALE 3,069,297

SEMI-CONDUCTOR DEVICES 7 Filed Jan. 16, 1959 2 Sheets-Sheet 2 INVENTOR JULIAN R. A. BEALE BY i 8 M AGENT United States PatentOifice 3,059,297 Patented Dec. 18, 1962 SEMI-CONDUCTOR DEVICES Julian Robert Anthony Beale, Whitehall Wraysbury, near Staines, England, assignor to North American Philips Company Inc., New York, N.Y., a corporation of Delaware Filed Jan. 16, 1959, Ser. No. 787,195 Claims priority, application Great Britain Jan. 16, 1958 20 Claims. (Cl. 148-15) This invention relates to methods of manufacturing semi-conductive electrode systems or devices, more particularly transistors, the semi-conductive bodies of which contain at least two electrodes provided by fusion in close proximity to each other. It also relates to semiconductive electrode systems, more particularly transistors, manufactured by the use of such methods.

In the manufacture of many kinds of semi-conductive electrode systems, more particularly if intended for use at high frequencies, the problem is frequently involved how to provide by the fusion method two or more electrodes at a very short physical distance from each other. In fact, a decrease of the physical distance between the electrodes results in a decrease of the detrimental seriesresistance of the current path in the semi-conductor and this is beneficial to the behaviour of the semi-conductive electrode system at high frequencies. A decrease of the physical distance between the electrodes may be achieved either by decreasing the geometric distance between the electrodes, or by decreasing the specific resistance in the current path between the electrodes, or preferably by means of a combination of the two steps.

Said problems may occur in semi-conductive electrode systems in which the adjacent electrodes are of the same type, as is the case for example, in the manufacture of field-effect transistors, in which an ohmic supply or source electrode and an ohmic discharge or drain electrode are juxtaposed on one side of the semi-conductive body, the electrodes being separated by a groove in the semi-conductive body which narrows the current path between said electrodes above a blocking layer. In such a fieldefiect transistor it is important for the parts of the current path located outside the narrow portion to have as low a resistance as possible with respect to the resistance of the current path in the narrow portion which is effective for the control.

The problem is even more difficult in semi-conductive electrode systems in which the adjacent electrodes are of different types, 'for example one of the n-type and the other of the p-type, as is the case, for example, in a diffusion transistor, in which the emitter and the base which are of different types must be provided side by side on a diffused layer. In this case also, a decrease of the physical distance, for example by decreasing the geometric distance, and/ or decreasing the series-resistance of the current path in the semi-conductor, is of paramount importance since it results in a decreased resistance of the base and hence an improvement of the frequency behaviour.

For providing by fushion two or more adjacent electrodes, use is frequently made of a jig which consists, for example, of a thin plate of inert material which is disposed on the semi-conductive body and in which two or more holes of the shape desired for the electrode are provided with the desired spacing. The electrode bodies to be provided by fusion are brought through the said holes onto the semi-conductive body, the spacing between them thus being fixed during the fusion process. However, it will be evident that the shortest distance obtainable between the electrodes with such a template is limited to the minimum thickness of the wall between the holes which is permissible in view of the mechanical strength and the separate filling of the holes. In addition, the manufacture of such templates is difiicult and the use thereof expensive, inter alia, because they can be employed only a few times as a result of wear.

An object of the invention is inter alia to provide another particularly suitable method of providing by fusion two adjacent electrodes, which method is simple and may be arranged in many ways into the process of manufacturing such semi-conductive electrode systems, the said method being serviceable up to extremely small geometric distances between the electrodes. The method according to the invention as such is also very suitable for the manufacture of semi-conductive electrode systems in which the tWo adjacent electrodes are different and more particularly of'different types. The invention also provides inter alia a method which permits of obtaining in a simple manner. extremely short physical distances since it permits of reducing not only the geometric distance but also considerably decreasing the residual series-resistances between the electrodes.

According to the invention, for manufacturing a semiconductive electrode system, for example a transistor, the semi-condutive body of which contains two electrodes provided by fushion at close proximity to each other, an electrode is provided by fusion on the semi-conductive body over a continuous and large area of the surface, whereafter at least the metal part of the electrode is di' vided into at least two parts by forming'a narrow groove in the solidified material, which groove extends at least to the recrystallized semi-conductive zone of the electrode, whereafter the separate parts of the electrode are fused again at least partly, without allowing them to fuse together. The groove is preferably provided. to extend at least into the recrystallized zone. In certain cases it is very favourable for the groove to penetrate even more deeply than the zone under the electrode infiuenced by diffusion and/or segregated during the first treatment. The more deeply the groove is provided in the body, the higher may be the temperature during the second fushion treatment. However, it is to be noted that the depth of penetration should, of course, not be chosen greater than necessary in connection with the second fusion treatment and the electrode structure desired.

The second fusion treatment may be carried out in many favourable ways to the benefit of the semi-conductive structure. According to one particular aspect of the invention, an active impurity is added to at least one of the separate parts of the electrodes, before or during the second fusion step, whereby two adjacent different electrodes are obtained after the second fusion step. This aspect is very important inter alia in the manufacture of semi-conductive electrode systems in which the two adjacent electrodes provided by fusion are required to be of opposite types, as is the case for example, in a p-n-p or an n-p-n transistor, in which the adjacent base and emitter are of opposite types, for example, the one of the p-type and the other of the n-type. In the manufacture of such semi-conductive electrode systems, such an active impurity is added to at least one of the separate parts of the electrode before or-during the last mentioned fusion treatment, so that adjacent electrodes of opposite conductivity type are obtained.

Although it is possible to obtain the difference be tween the electrodes by adding the active impurities to one or more of the said parts only during the second fusion treatment, this addition is preferably carried out in a separate step after forming the groove and before the second fusion treatment, the second fusion treatment then being used to cause the electrode or electrodes to absorb the active impurity added by segregation or diffusion. Thus, for example, it is possible in a simple and suitable manner to obtain a semi-conductive electrode system more particularly a transistor, having two adjacent electrodes of opposite types by providing by fusion an electrode material containing donors during the first fusion treatment intended for obtaining the electrode over a continuous surface, whereby an n-type electrode is formed, and after forming the groove adding a material containing an acceptor to one of the solidified separate parts, whereafter during the subsequent treatment a p-type electrode is formed at one side of the groove due to the over compensating action of the acceptor and an n-type electrode is formed at the other side of the groove. It will be evident that the amount of acceptor added must be such that during the segregation process it can dominate the donors present in the electrode-melt to be formed. Consequently, for the active impurity to be added, use is preferably made of an impurity having a segregation constant higher than that of the impurity already available. Acceptors suitable for this purpose in germanium are, for example, the elements gallium, aluminum and boron, more particularly aluminum.

The same structure with electrodes of opposite type provided side by side by fusion may alternatively be obtained in a different manner. Thus, it is also possible to provide by fusion an electrode material containing acceptors during the first fusion treatment intended for obtaining the electrode over a continuous area of the surface, where by a p-type electrode is formed and after forming the groove to add a material containing donors to one of the separate solidified parts, whereafter during the subsequent fusion treatment an n-type electrode is formed at one side due to the overcompensating action of the donor and a p-type electrode is formed at the other side of the groove. In this case, the added amount of donors must be such that during the segregation process it can dominate the acceptors Present in the electrode melt to be formed. Consequently, in this case, use is preferably made of a donor impurity having a segregation constant higher than that of the acceptor already available.

In a further particularly suitable method according to the invention, such an electrode structure may also be obtained by providing by fusion an electrode material which is suitable as a carrier material for active impurities such, for example, as lead, bismuth, tin or similar material, during the first fusion treatment intended for obtaining the electrode over a continuous area of the surface and, after forming the groove, by adding a material containing an acceptor to the solidified material at one side of the groove and a material conatining a donor to the material at the other side of the groove, Whereafter during the subsequent fusion treatment a ptype electrode is formed at one side of the groove and an n-type electrode at the other side thereof.

It will also be evident that the invention also affords many further possibilities of acting upon the two halves of the electrode. Thus, it is possible, for example, in addition to reversing the conductivity type of one electrode, to influence at the same time the conductivity of the other electrode by adding an additional proportion of the impurity already present to the other electrode prior to the second fusion treatment.

It is readily possible to manufacture a p-n-p or n-p-n transistor in the above-described manners. The electrode corresponding in type to the underlying semi-conductor may be used as the base and the electrode which is opposite thereto in type may be used as the emitter. The base zone of the transistor may be provided in different ways. Thus, it is possible, for example, to utilise a semi-conductive body which has preliminarily been provided with a zone intended as the base zone, for example a semi-conductive body of the p-type, which has a diffused zone of the n-type located at its surface. The two electrodes may be provided on this zone by the use of theinvention. Thus, it is possible first to provide by fusion a donor material on the n-type diffused zone for obtaining the electrode over the continuous area and, after forming the groove which has a depth of penetration smaller than the diffused zone, to provide one half of the electrode, which is intended as the emitter, with a proportion of an acceptor, so that a p-type electrode is formed at this side of the groove during the second fusion treatment.

According to a further aspect of the invention, which is applicable inter alia to the manufacture of a semiconductive electrode system having adjacent electrodes of opposite types, an active impurity is diffused into the semi-conductive body during one or more of the fusion treatments. Preferably, the underlying base zone is formed in the body due to the diffusion, during one or more of the fusion treatments, so that it is possible to use a semi-conductive body which is homogenously of a given type. The active impurity to be diffused into the body may be supplied during the relevant fusion treatment from the ambient atmosphere and/ or from the electrode material itself, to which it may have been added during one of the preceding steps. From there the diffusing impurity may diffuse into the body throughout its surface via the free surface of the body and via the fronts of the melts of electrode material formed. If the base zone is formed only during one of the fusion treatments, the type of the impurity to be diffused into the body is opposite to that of the initial semi-conductive body.

According to the invention, the diffusion of the active impurity is preferably effected, at least to a considerable part or substantially, during a fusion treatment after forming the groove. This affords inter alia the advantage that a low-ohmic surface is formed in the side walls of the groove so that the series-resistance and hence the physical distance between the electrodes is reduced further. This low-ohmic surface is also favorable for a low noise level and the stability of the electrode system. In addition, this method is simple and controllable and may lead to a high reproducibiiity. If the second fusion treatment is used for the diffusion and for inverting the conductivity type of one of the electrodes, the diffusing impurity is preferably chosen so that its speed of diffusion into the semi-conductor at the relevant temperature is higher than that of the impurity intended for inverting, if they are of opposite type, while for inverting the conductivity type it is necessary for the content of diffusing impurity and/or its segregation constant in the electrode material to be less than that of the segregating impurity. According to a further simple and efiicaceous embodiment of the method according to the invention, the impurity to be diffused into the body is already added to the electrode material to be provided by fusion during the first fusion treatment and diffuses from the electrode material into the body after forming the groove during the fusion treatment. Although preferably the base zone is provided in the body due to the diffusion during the second fusion treatment, the diffusion during the second fusion treatment may also advantageously be used in those cases in which the base zone has already been provided in the body beforehand, since in such cases also the diffusion permits of obtaining in the side walls of the groove a reduction of the series-resistance in the current path between the electrodes.

The method according to the invention may also advantageously be applied to the manufacture of semi-conductive electrode systems in which the adjacent electrodes provided by fusion are of the same type, as is the case, for example, in a field-effect transistor, in which the ohmic Source electrode and the ohmic drain electrode are provided side by side on a zone of a given conductivity type, a groove between said electrodes in the base Zone narrowing the current path above the p-n transition to the adjoining zone of the rectifying gate electrode. An active impurity is diffused into the body during one or both fusion treatments, but preferably to a considerable part during the second fusion treatment. As regards this diffusion, the method according to the invention affords possibilities and advantages for such semi-conductive electrode systems quite similar to those mentioned in the foregoing or hereinafter with regard to the manufacture of semi-conductive electrode systems having electrodes different in type. Thus, for such semi-conductive electrode systems, also the diffusion may be utilized in similar manners for rendering the surface of the groove low-ohmic and/or for providing the base zone of the field-effect transistor, the diffusing impurity being supplied either from the surrounding atmosphere and/or from the electrode material itself. For a field-effect transistor also the low-ohmic surface in the groove is favourable for the noise level and the stability. Only inverting one electrode can be omitted in this case.

When using a method according to the invention in which a base zone is provided by diffusion during the second fusion treatment the depth of penetration of the melt front of the electrode material into the semi-conductive body during the fusion treatment after forming the groove, is preferably chosen greater than that of. the melt front during the first fusion treatment. This may be achieved, for example, by choosing the temperature of the second fusion treatment to be sufficiently higher than that of the first fusion treatment. This affords during diffusion inter alia the advantage that the base zone is diffused from the melt front newly formed, so that the thickness of the base zone is substantially independent of the depth of penetration of the melt front and hence extremely reproducible. In addition, more generally the advantage is obtained that the active portion of the system is displaced to penetrate the semi-conductive body more deeply so that there is less risk of the electric properties being detrimentally influenced by any residual disturbances in the crystal lattice near the groove. However, it will be evident that the depth of penetration of the groove must be greater than that of the melt front during the second fusion treatment in order to prevent the two parts from fusing together.

The groove may be formed in any suitable manner. Thus, for example, it has been found particularly favourable to use for this purpose an ultrasonic cutting method which utilises a thin ultrasonic head in combination with a fine abrasive or abrasive slurry. Another methodis one wherein a thin wire coated with a fine abrasive or in combination with an abrasive, for instance an abrasive slurry, is reciprocated at the area concerned. Said methods may be combined, for example, with an after etching treatment for the groove. Widths of 25 microns in the narrowest part of the groove may thus readily be obtained. It is thus also possible for the depth of penetration of the groove to be chosen greater than that of the melt front or of the recrystallized zone of the electrode, in order topermit the depth of penetration of the melt front during the second temperature treatment to be chosen greater than that during the first fusion treatment.

A third electrode, for example the collector electrode in the p-n-p or n-p-n transistor, or the gate electrode in the field-effect transistor, may be provided in a simple manner by alloying on the opposite side of the semi-conductive body.

A material containing a donor or an acceptor may be either a donor impurity or an acceptor impurity itself or alloys or mixtures thereof with other suitable elements. Thus, for example, in those cases in which, during the fusion treatment, a donor material is to be alloyed as well as diffused, it is possible to use one and the same suitable donor impurity for both purposes, or to use, for example, an electrode material containing two donors, one of which has a dominant function during alloying because of its higher segregation constant, and the other of which has a dominant function during diffusion because of its greater diffusion velocity. In addition, use

may be made with great advantage of an electrode material which substantially consists of a material which itself need not be suitable as an active impurity, but is particularly suitable, for example, on account of the low solubility of the semi-conductor in this material or because of its suitable mechanical properties as a carrier material for the active impurities. Examples of such carrier materials in connection with germanium are, for example, lead, indium and bismuth, and in connection with silicon, for example lead.

In order that the invention may be readily carried into effect, several aspects of the invention will now be explained in detail by way of example, with reference to the accompanying diagrammatic drawings in which:

FIGS. 1 to 5 show in section the sequential stages of a transistor during its manufacture by a method according to the invention;

FIG. 6 is a plan view of another embodiment of a transister in a given stage of the manufacture according to the invention.

In FIGS. 1 to 5, the cross-hatching is omitted for the sake of clarity.

A thin disc of electrode material is provided by melting on, and thus adherent to, a rectangular mono-crystalline semi-conductive slice 1 of p-type germanium having a specific resistance of 2 ohms/ cm. The dimensions of the semi-conductive slice are about 1 mm. by 2 mms. by 150 microns. The disc of electrode material has a diameter of about 200 microns and a thickness of about 50 microns and it consists of lead, to which 1% by weight of antimony has been added. The electrode material may be provided, for example, by heating the semi-conductive slice and the disc of electrode material placed on it approximately centrally of one of the large sides in an atmosphere of hydrogen to about 700 C. for about 3 minutes.

FIG. 1 shows the stage obtained after heating. On the unchanged p-type part 1 of the semi-conductive body an n-type zone 2 has recrystallized during cooling due to segregation of the antimony. This n-type zone 2 is has been added. The layer 3 constitutes the metal part of the electrode. Line 4 marks how deeply the molten electrode material has penetrated the otherwise solid body. Dimension a in the figure, is about microns and dimension b is about 200 microns. During heating, the antimony can diffuse along the surface of the plate 1 and thus penetrate the semi-conductive plate via its surface. In addition, the antimony can penetrate the zone 1 via the junction surface 4 between the melt of electrode material and the semi-conductive plate. This is dependent upon the temperature and the duration of the fusion treatment. However, the depth of penetration of the diffusion is small at the given temperature and duration and hence the diffused layer under the surface is not shown for the sake of clarity, but indicated only under the melt front 4 in the zone 5.

A thin groove 6 radially provided through the layer 3 penetrates the zone 1 of the plate via layer 3 and zone 2 Said groove is formed by means of an ultrasonic cutting method in which use is made of a thin cutting blade and a paste of a very fine aluminum-oxide abrasive. The groove has a width of only about 25 microns at its bottom and is slightly V-shaped due to abrasion of the sides of the groove as the cutting treatment proceeds further.

The whole is subsequently subjected to an etching treatment at 70 C. for about 5 minutes in a bath of 20 vol. percent of hydrogen peroxide. The etching agent removes about 2.5 microns from the surface of the germanium and hence semi-conductive material damaged during. the ultrasonic cutting treatment is also removed from the groove. Under these conditions, said etching treatment also substantially removes the superficial n-type diffused layer formed by the diffusion of antimony along the surface, from the surface of zone 1 of the semi-conductive body.

FIG. 2 shows the semi-conductive body with the electrode cut through at the stage of the etching treatment. The narrow groove 6 divides the metal layer 3, the zone 2 and the transition 4 into two halves. In FIGURE 2, the parts of the left-hand half of the electrode are indicated by 3a, 2a and 4a, and the parts of the right-hand half are indicated by 3b, 2b and 4b. The new surface of the plate is marked by line 7.

An active impurity of opposite type is added to the right-hand half of the electrode. The two halves were of the n-type. Aluminum is particularly suitable as an acceptor impurity on account of its high segregation constant. The aluminum may be added, for example, to the right-hand half by providing it by vaporisation onto the surface of the layer 3b, the surface of the semi-conductive body and that of the electrode 3a being shielded during evaporation by means of a mask.

The active impurity may alternatively be added in a simple manner, for example, by providing it in the form of a dispersion in a binder, for example by means of a brush, on the relevant electrode. A binder suitable for aluminum is, for example, a solution of methacrylate in xylene.

The whole is subsequently heated in an atmosphere of hydrogen at 950 C. for about minutes, whereby the two halves of the electrode are again fused. After the second fusion treatment, the stage shown in FIG. 3 is reached.

The second fusion treatment is carried out at a temperature sufficiently high to cause the melt front to penetrate the germanium plate more deeply than was the case during the first fusion treatment. The additional parts of the electrode halves provided during the second fusion treatment are indicated by 9a and 912. Line 10a marks the depth of penetration of the melt front during the second fusion treatment, while the depth of penetration of the first fusion treatment marked by line 4a in FIG. 2 is represented by a dotted line 4a in FIG. 3. After recrystallisation, both the zone 2a and the prolongation thereof, the zone 9a, are of n-type. However, in the right-hand half of the electrode, the zone 9b and the zone 21), after recrystallisation, have been converted into p-

type zones

9b and 2b due to the aluminum, during recrystallisation, having overcompensated the initial action of the antimony due to the high solubility and segregation constant of aluminum. In this connection it is to be noted that for overcompensation it is not necessary for the last impurity added to have a segregation constant higher than that of the first impurity. Over compensating may also be obtained with approximately the same value of the segregation constant or even with a higher segregation constant of the first impurity by choosing the content of the second impurity in the melt of electrode material to be correspondingly higher than that of the first impurity. However, it is usually preferable for the segregation constant and the solubility of the second impurity to be higher than that of the first impurity.

The coagulated layer 3b constitutes the metal part of the p-type electrode (317', 2b, 9b) and consists of lead, aluminum and antimony and possibly a small content of germanium. Line 4b of FIG. 2 is represented as a dotted line 41; in FIG. 3.

In addition to recrystallisation and alloying, diffusion also occurs during the second fusion treatment. The antimony upon being provided by fusion, diffuses both from the right-hand part and the left-hand part of the electrode via the melt front into the body, while the aluminum only diffuses from the right-hand part of the electrode. As a result of this diffusion, the p-n transition (not shown) in the right-hand parts lies a little below the line 10b, which marks the depth of penetration of the melt front in the right-hand electrode. In addition, during the second fusion treatment, due to the diffusion of the antimony which diffuses much more rapidly than does aluminum, an n-type zone 12 is formed which is internally bounded by line 11, and which extends substantially via the surface of the groove and below the p-n transition of the right-hand electrode. Due to the diffusion during the second temperature treatment, which took place at a higher temperature and for a longer period than the first temperature treatment, a properly defined diffused layer 12 and transition 11 are formed as compared to the weak diffusion during the first temperature treatment. During this second temperature treatment, the parts 3:: and 3b of the electrodes, as shown in FIG. 2, undergo a variation, that is to say, assume the shape of the

parts

3a and 3b of FIG. 3. It also appears from FIG. 3 that the electrode material upon being provided does not flow into the groove although the groove is very narrow. In this connection it is noted that the dimensions of the coagulated material after the second fusion treatment, of which the boundary line with the solid material, or in other 'words, the maximum depth of penetration of the melt front is indicated by the

lines

10a and 10b, are shown in vertical direction with exaggeration for the sake of clarity. It is not necessary during the second fusion treatment to alloy into the semi-conductive plate more deeply than during the first fusion treatment. Nevertheless this is preferably done, since in this case the additional advantage is obtained that the base thickness of the transistor is substantially independent of the depth of penetration of the electrode material, the thickness of the base zone being determined substantially by the diffusion during the second fusion treatment, which diffusion then takes place from the newly formed

melt fronts

10a and 10b. In determining the temperature difference between the first and second fusion treatments, as is necessary for obtaining the greater depth of penetration of the melt front during the second fusion treatment, it is neces sary to make allowance for the fact that loss of electrode material occurs in forming the groove 6. In the example under consideration, comparatively more lead than antimony is removed in forming the groove as a result of the difference between the contents of the two elements in the electrode material.

It will also readily be evident that the groove 6 must be deep enough to avoid that, during the second fusion treatment, the molten material does not close off the groove. The depth of the groove must therefore be chosen suitably in connection with the temperature to be used during the second fusion treatment.

The electrode system shown in FIG. 3 may be worked into a p-n-p transistor in the following manner. The surface of the body of FIG. 3 located above the dotted line 13, is covered with an etch-resistant lacquer layer consisting of a solution of polystyrene in ethyl methyl ketone, the whole subsequently being immersed into a 20% hydrogen-peroxide solution heated to 70 C. The treatment is continued until the portion of the body heneath the dotted line 13 has been removed by etching. The lacquer layer is then removed by immersing the whole into a bath of ethyl methyl ketone.

Next, a collector is provided on the body by alloying a thin disc of indium, to which 1% by weight of gallium has been added, on the etched side of the body opposite the

electrodes

3a and 3b. The alloying of the collector may be effected, for example, by heating the whole in an atmosphere of hydrogen to about 500 C. for 5 minutes. Substantially no further diffusion takes place at this comparatively low temperature. The position of the collector disc is not critical, but the collector is preferably provided approximately opposite the layers 3a. and 3b. In FIGURE 4, the reference numeral 14 indicates the recrystalized semi-conductive zone of the collector and zone 15 constitutes the metal part of the collector, which constitutes of an alloy of indium-gallium and a small content of germanium. Soldered on the layer 15, by means of an indium solder 17, is a rigid nickel member 16 which serves as a supply Wire and also as a support.

Thin nickel members

18 and 19 are also soldered on the

metal layers

3a and 3b of the base and the emitter by means of an

indium solder

20, 21, respectively. The

soldering process is carried out by means of a small soldering iron.

A transistor system is thus obtained, the

supply wires

16, 18 and 19 of which constitute conductors to the collector, the base and the emitter, respectively.

The groove 6 is subsequently filled with a lacquer layer 22 up to a level located above the

zones

2a and 2b by means of a drop of a solution of polystyrene in ethyl methyl ketone. The lacquer is diluted so that it can flow freely along the surface of the groove 6 and projects only slightly above its ends. After filling with the lacquer up to the level indicated by a dotted line in FIG. 4 the lacquer is allowed to dry. The three

supply wires

16, 18 and 19 are then connected to the positive terminal of a source of supply, the whole subsequently being placed in an etching bath containing a 5% aqueous NaOH-solution. A platinum electrode is suspended in the etching bath and connected to the negative'terminal of the source of supply. A current of ma. is adjusted and maintained for about 10 minutes, so that more than 25 microns of the surface is removed, as shown in FIGURE 5. This figure also shows that the etching agent has also etched partly below i the

metal parts

3a and 3b of the electrodes. In addition,

during etching, the superficial part of the n-type diffused layer has been removed.

The lacquer layer is subsequently removed from the groove 6 by dissolution in ethyl methyl ketone, the whole being immersed in an etching bath of 20% of hydrogenv peroxide for about seconds at 70 C. The transistor is subsequently mounted in known manner in an envelope- The transistor thus obtained has a low resistance of the base since the geometrical distance between the base contact 3a and the emitter is small and, in addition, a current path of a low specific resistance exists over this extremely, small distance along the surface of the bottom of the groove. The low specific resistance of the surface is brought about by the diffusion of antimony during the second fusion treatment, since upon diffusion into a surface there is always a concentration in the surfaces con siderably higher than at some distance below the surface. In the case under consideration, the antimony for the diffusion is supplied from the molten electrode material and the antimony diffuses from there to a high degree along the surface. The transistor also has a very low noise level and .a high stability. The above-described p-n-p transistor also has a low emitter-base capacity and a low base-collector capacity due to the limitation of the surface of the p-n junctions during etching, whereby even a portion below the metal parts of the emitter and the collector has been removed. Due to the aforementioned exceptional properties, the transistor is very suitable for use at high frequencies.

FIGURE 6 shows another embodiment of a transistor which may likewise be manufactured in a similar manner by the method according to the invention. FIG. 6 is a plan view of this transistor at a manufacturing stage corresponding to FIG. 3. Instead of a straight groove through the electrode, in this embodiment an annual groove 6 is provided, which is filled with polystyrene lacquer before proceeding to the second etching treatment. The central part 3b constitutes the metal part of the emitter, whereas the outer part 3a constitutes the metal part of the base. During the second etching treatment, due to the emitter being fully surrounded by the groove filled with the polystyrene, there will be no etching below the metal part of the emitter. The emitter-base capacity is thus higher and this embodiment is not particularly suitable for use at very high frequencies, although admirably suited for use as a medium power transistor at high 10 frequencies. Manufacturing steps not specially mentioned in this example are wholly identical with those described with regard to the transistor shown in FIG. 1 to 5.

It is to be noted that many variations are possible Within the scope of the invention. Thus, for example, it is also possible, after the first fusion treatment, to divide the largearea electrode into more than tWo parts and thus obtain more than two adjacent electrodes provided by fusion. In this case the second fusion treatment may be used for acting upon the conductivity and/ or the conductivity type of one or more of the electrodes. Thus, it is also possible in those cases in which after forming the groove, the type of one electrode must be inverted and the diflfusion of the base zone must be carried out, to perform these two treatments in two separate fusion treatments. It will also readily be evident that the use of the invention is not limited to the specified semi-conductors germanium and silicon, but that it also comprises other semi-conductors, for example, the semi-conductive compounds, such as the III-V compounds, for example GaAs and InP. Further more, the invention is of course, applicable not only to the manufacture of transistors, but also to any other semiconductive electrode or devices having at least two adjacent electrodes.

What is claimed is:

l. A method for producing a semi-conductor device comprising providing on a surface of a semi-conductive body a largearea contact, dividing the contact but not the entire body into plural separated portions, thereafter adding to one of the plural portions an active impurity capable of altering the conductivity of that contact portion when incorporated therein, and thereafter fusing the separated contact portions to incorporate the active impurity into that portion to which it was added thereby to selectively alter its conductivity.

2. A method as set forth in claim 1 wherein the largearea contact is divided into two halves.

3. A method of providing adjacent regions of different a conductivity in a semi-conductive body, comprising fusing and alloying an impurity-bearing mass to a surface of the semi conductive body to produce underneath the mass a region of given conductivity type in the body, thereafter forming a groove into and through the mass and into the said region of given conductivity type to divide the mass into at least two separate parts, thereafter adding to less than all of the parts another impurity capable of altering the conductivity type of the underlying body region when incorporated therein, and thereafter refusing the separated masses to incorporate the added impurity into the selected parts and thereby alter the conductivity type of the underlying region and make it different from the adjacent region of the said given conductivity type.

4. A method as set forth in claim 3 wherein the groove extends through the said region of given conductivity type thus dividing it into at least two separate parts.

5. A method as set forth in claim 3 wherein the groove is formed by,cutting by ultrasonic means.

6. A method as set forth in claim 3 wherein the groove is cut by reciprocating a thin wire associated with a fine abrasive in contact with the mass.

7. A method as set forth in'claim 4 wherein the temperature at which the first fusion is carried out is lower than the temperature at which the refusion is carried out.

8. A method of providing adjacent regions of different conductivity in a semi-conductive body, comprising fusing and alloying a donor-impurity-bearing mass to a surface of the semi-conductive body to produce underneath the mass a region of n-type conductivity in the body, thereafter cutting a slot into and through the mass and into the n-type region to divide the mass into two separate parts, thereafter adding to one of the parts an acceptor impurity having a segregation coefficient in the semi-conductive body greater than that of the donor impurity, and thereafter refusing the masses to incorporate the acceptor impurity into the selected'part and thereby 1 1 alter the conductivity type of the underlying region and make it p-type.

9. A method as set forth in claim 8 wherein one underlying region constitutes the base region and the other underlying region constitutes the emitter region of a transistor.

10. A method as set forth in claim 8 wherein the slot has an annular shape.

11. A method of providing adjacent regions of different conductivity in a semi-conductive body, comprising fusing and alloying an acceptor-impurity-bearing mass to a surface of the semi-conductive body to produce underneath the mass a region of p-type conductivity in the body, thereafter cutting a slot into and through the mass and into the body to divide the mass into at least two separate parts, thereafter adding to one of the parts a donor impurity whose segregation coelficient in the semiconductive body is greater than that of the acceptor impurity, and thereafter refusing the masses to incorporate the donor impurity into the selected part and thereby alter the conductivity type of the underlying region and make it n-type.

12. A method of providing adjacent regions of different conductivity in a semi-conductive body, comprising fusing and alloying a metal mass to a surface of the semiconductive body to produce underneath the mass an alloyed region, thereafter cutting a slot into and through the mass and into the body to divide the mass into at least two separate parts, thereafter adding to one of the parts an acceptor impurity and adding to the other part a donor impurity, and thereafter refusing the masses to incorporate the donor and acceptor impurities into their associated parts and thereby make the conductivity types of the underlying and adjacent regions of opposite conductivity.

13. A method of providing adjacent regions of different conductivity in a semi-conductive body, comprising fusing and alloying an impurity-bearing mass to a surface of the semi-conductive body to produce underneath the mass a region of given conductivity type in the body, thereafter forming a groove into and through the mass and into the said region of given conductivity type to divide the mass into at least two separate parts, adding another impurity to one of the separated parts, refusing the separate parts to incorporate the added impurity into the underlying body region and thereby alter its conductivity, and, during one of the fusion steps, diffusing an impurity into the body.

14. A method as set forth in claim 13 wherein the diffused impurity forms the same conductivity type as that formed by the impurity originally present in the mass.

15. A method as set forth in claim 13 wherein the diffusion step takes place during the second fusion step.

16. A method of making a semi-conductor device, comprising providing an alloyed electrode on a surface of a semi-conductive body. separating said alloyed electrode into two closely-adjacent alloyed electrodes on the same surface of said semi-conductive body, adding to one only of the said separated electrodes a dispersion of aluminum in a binder, and thereafter fusing the separated electrodes to incorporate the aluminum into the said one electrode and thereby alter its conductivity.

17. A method for producing a semiconductor device comprising forming on a surface of a semiconductive body a large-area fused contact, thereafter forming a narrow groove in the contact which extends completely therethrough and into the underlying semiconductive body but not completely through the latter thereby to divide the contact into plural separated portions in contact with the same semiconductive body, and thereafter refusing the separated contact portions, but maintaining them separate, in the presence of an active impurity to incorporate the latter in a portion of the body thereby modifying its conductivity.

18. A method for producing a semiconductor device comprising fusing and alloying a metal mass to a surface of a semiconductive body to form a large-area fused contact, thereafter forming a narrow groove in the contact which extends completely therethrough and into the underlying semiconductive body but not completely through the latter thereby to divide the contact into plural separated portions in contact with the same semiconductive body, and thereafter refusing the separated contact portions in the presence of an active impurity to diffuse the latter into a portion of the body adjacent the contacts thereby altering its conductivity.

19. A method as set forth in claim 18, wherein the active impurity is present in the atmosphere during the refusion step.

20. A method as set forth in claim 18 wherein a contact portion contains another active impurity of the opposite-conductivity-forming type which becomes incorporated in the adjacent body portion during the refusion step forming a recrystallized region defining a junction with the body portion containing the diffused impurity.

References Cited in the file of this patent UNITED STATES PATENTS 2,794,846 Fuller June 4, 1957 2,837,704 Emeis June 3, 1958 2,846,340 Jenny Aug. 5, 1958 2,865,082 Gates Dec. 23, 1958 I UNITED STATES PATENT OFFICE CERTIFICATE OF CORRECTION Patent No. 3,069,297 December l8 1962 Julian Robert Anthony Beale It is hereby certified that error appears in the above numbered patent requiring correction and that the said Letters Patent should read as corrected below.

Column 2 lines 24 and. 140,- for fushion-"=- read fusion I column 3 line 5O,v for "conatining" read containing column 9-, line 64, for "annual" read annular column 10, line 23, after "elgectrod" insert systems -o.

Signed and sealed this 27th day of August I963 (SEAL) Attest:

ERNEST w. SWIDER DAVID A Attesting Officer Commissioner of Patents 1

Claims

1. A METHOD FOR PRODUCING A SEMI-CONDUCTOR DEVICE COMPRISING PROVIDING ON A SURFACE OF A SEMI-CONDUCTIVE BODY A LARGE-AREA CONTACT, DIVIDING THE CONTACT BUT NOT