DE1564537B2 - METHOD OF MANUFACTURING A SEMICONDUCTOR CIRCUIT MODULE - Google Patents

Description

The invention relates to a method for producing a semiconductor circuit module, based on FIG a planar disk, which is composed of a large number of similarly designed and arranged circuit components made of semiconductor material is, which by means of a heat softenable electrically insulating material from each other are separated and have contact surfaces on opposite major surfaces of the disc.

In the manufacture of certain integrated circuits using semiconductor materials and Numerous operations are partially repeated using photolithographic manufacturing processes to form the various circuit components and circuit connections. Such are the operations in some cases, for example

1. Application of a photoresist layer to a semiconductor wafer,

2. Exposing the photoresist layer and

3. Etching of the photoresist layer during the manufacture of a single circuit often too repeat.

If all circuit components to be formed on or in a semiconductor wafer are the same and are similarly oriented, is the number of photolithographic operations for their training usually less than when some or all of them are different.

In the 'manufacture of semiconductor components, So semiconductor arrangements in the form of a transistor or the like, it is known and common practice, starting from a large number of these arrangements containing disk, which then into the individual Arrangements is divided. According to a method known from the German Auslegeschrift 1188 731 is intended to contact the individual components with a metallic conductor network on the Disc glued on and divided with it. This makes the subsequent installation of the components in a circuit facilitated. The interconnection of individual components to form a circuit module, d. H. to a basic circuit required in various larger circuit arrangements is here but not provided.

A method for producing semiconductor arrangements is also known (German Auslegeschrift 1193 169), in which several semiconductor components are isolated from one another in a desired Configuration next to one another or one on top of the other joined together to form a semiconductor wafer are connected to one another, and in which selected contact areas of the circuit components can be connected to one another to form the desired semiconductor arrangements. The components are connected by oxidation. In one embodiment it is to different types of components.

Even if an arrangement of the type explained above with identical and similarly oriented circuit components is less complex than an arrangement with disparate components Usually the components of the same type are so inappropriately oriented that many types of circuit are not realized can be because the necessary circuit connections are practically impossible to produce. the The object of the invention is to avoid this disadvantage.

The invention consists in that in a method of the type mentioned, the composite Disk along the insulating material is cut into several parts, which along their parting surfaces a Have a layer of the insulating material that at least one of these parts is turned over, that at least some of the parts including the turned part reassembled and in this Arrangement by softening and connecting the insulating material by heating the parts under

a new planar disk can be attached to one another, and that from this new disk removing at least a portion of the at least two circuit components of the non-inverted Part and at least two circuit components of the turned part contains, the below Formation of the circuit module are electrically connected to one another.

The method is particularly advantageous for producing integrated solid-state circuits, for example made of diodes. As an exemplary embodiment of the invention, the method is described below in connection with the production of rectifier series circuits or rectifier bridges. In the drawing shows

F i g. 1 shows, in partial representation, a front view of a wafer of semiconductor material which is used in the process is used for the production of semiconductor devices according to the invention,

F i g. 2 shows the disk according to FIG. 2 in part. 1 after a first process step,

3 shows an induction furnace in partial cross-section and in a reduced representation, the parts of the disc and a substrate in a disassembled arrangement a further procedural stage,

F i g. 4 is a partial view of the disk and the substrate, which are joined together to form a block are connected; the dashed lines represent the planes along which the in F i g. 4 a grooves shown in the block is to be cut,

F i g. 4 a in a partial cross section along the line 4 a in Fi g. 4 and seen in the direction of the arrows grooved block,

Fig. 4b in a partial plan view of the grooved Block according to FIG. 4 a,

F i g. 5 and 6 in partial cross-sections similar to that of FIG. 4 a the semiconductor arrangement in further Processing stages, ■

F i g. 7 a composite in a partial cross-section Disk made of passivated semiconductor circuit components, which are located between the in F i g. 6 indicated horizontal planes after an insulating material has been pressed into the grooves of the block is,

F i g. 8 is a perspective view of the assembled disk of FIG. 7, being the interrupted Lines indicate the planes along which the composite disk is to be divided into Parts cuts need to be made

F i g. Figure 9 is a top perspective view of a desired configuration derived from the reassembled Dividing the disk according to FIG. 8 is made,

10 shows an embodiment of a semiconductor arrangement, which includes the stripe extending between the broken lines in FIG. 9 indicated parallel planes,

10a shows a section along the line 10a-IOa in FIG. 10, seen in the direction of the arrows,

Fi g. 11 is a schematic circuit diagram of FIG Arrangement according to Fig. 10,

F i g. 12 and 12a a perspective view of part of the assembled disk according to FIG. 9, respectively Seen from above and below, showing the electrical connections between the components are designed according to a bridge circuit, and. ') Fig. 13 is a schematic diagram of the bridge circuit the in the F i g. 12 and 12 a shown embodiment.

F i g. 1 shows a portion of a wafer 10 made from a single crystal of semiconductor material, e.g. B. silicon, germanium or gallium arsenide. The disk 10 may have a size of about 6.5 cm ² and a thickness between 0.1 and 0.3 mm and has upper and lower major surfaces 12 and 14. The description information, such as B. "upper" ίο and "lower" are only used relatively and here only to facilitate the description and not in the sense of a restriction of the invention.

Acceptor and donor impurities can through the opposing surfaces 12 and 14 of the Disk 10 can be diffused in by any known double diffusion process to each to form p-conductive and η-conductive zones 16 and 18 in the disk. This double diffusion can are controlled in a known "manner that the p-zone 16 extends from surface 12 to a depth of about 0.05 mm and has a pn-junction 20 forms with the η zone. The pn junction 20 comprises a plane which is essentially parallel to the opposing main surfaces 12 and 14 of the disc 10 is located.

. Instead of the aforementioned double diffusion process, the pn junction 20 can be through a simple Diffusion of a p-conduction generating impurity can be formed in an η-conductive disk. The pn junction 20 can also be produced using the known method of epitaxial growth will.

The disc 10 is mounted on a suitable substrate 22 (Figs. 3 and 4) to facilitate its handling and subsequent treatment. The substrate 22 can consist of a heavily doped p-conducting wafer made of silicon ( denoted by P + ) and can also be part of a finished semiconductor arrangement. The disk 10 and the substrate 22 should consist of materials with essentially the same coefficient of thermal expansion in order to obtain a stable structure in the finished product.

The disk 10 and the substrate 22 can be joined together by hot pressing. if the abutting surfaces of the wafer 10 and the substrate 22 are etched, is preferably one the contact surfaces covered with a layer 24 of metal, e.g. B. with a chrome, nickel, Niobium, zirconium or titanium layer. The metal layer 24 can be applied to the surface 12 of the disk 10 (Fig. 2) by either vapor deposition, plating, or dipping the disk 10 in a fine metal powder or by inserting a thin metal foil between the disk 10 and the substrate 22 will. The thickness of the layer 24 and also the dimensions of the other parts are for the sake of clarity shown enlarged. Preferably, the thickness of the metal layer 24 is between 0.1 and 10 µm.

The hot pressing of the wafer 10 onto the substrate 22 can be carried out under pressure in an induction furnace 26 be carried out, as shown in the illustration in a disassembled arrangement according to · F i g. 3 is indicated. the Metal layer 24 on disk 10 is placed against adjacent surface 28 of substrate 22, to form a block 30 (Fig. 4). the Disc 10 and substrate 22 are placed between carbon plates 32 and 34 so that in direction of arrows 36 and 38 pressure on the opposite

overlying sides 14 and 48 of the block 30 can be exerted while sufficient heat is supplied by the induction furnace 26 so that the metal layer 24 diffuses into the disk 10 and into the substrate 22 and thus connects them to the block 30. If the layer 24 consists of chromium or titanium, the hot pressing can be carried out at a temperature between 950 and 1400 ^° C. and at a pressure between 14 and 350 kp / cm ² . The hot pressing should be carried out in a vacuum or in a neutral or reducing atmosphere, e.g. B. in an argon or hydrogen atmosphere. Lower temperatures and pressures can be used for germanium or III-V semiconductor materials, e.g. B. gallium arsenide can be used.

After the block 30 has been produced, numerous mesas 40 are formed, with the formation of a plurality of mesas 40 Grooves 42 cut into block 30 as shown in FIGS. 4a, 4b and 5. The dashed Lines at the block in FIG. 4 show the planes along which the cuts for forming the mesas 40 are shown must be carried out. Each of the grooves 42 extend completely through the disc 10 and partially through the substrate 22, wherein a plurality of substantially similar mesas 40 in one result in largely regular patterns, as can be seen from the top view according to FIG. 4b. The grooves 42 can be made of the block 30, for example by chemical and electrolytic etching processes Sandblasting, sawing, grinding or ultrasonic machining are excluded.

It is now necessary to use an electrically insulating material, e.g. B. softened glass, to be pressed into the grooves 42; However, since most types of glass contain impurities which can adversely affect the pn junction 20, it is expedient to have the plateau surface 14 of each mesa 40 and the areas of the pane 10 and the substrate 22 delimiting the grooves 42 with a layer 44 of an electrically inert material To cover material, e.g. B. with a silicon dioxide layer. The silicon dioxide layer 44 can be applied by any known method, e.g. B. by direct oxidation of the disk 10, if it consists of silicon itself, by vapor deposition of silicon dioxide (SiO, or SiO in O ₂ ), by vapor phase deposition of organosilanes or by hydrolysis or oxidation of silicon halides.

The silicon dioxide can be mixed with other oxides, e.g. B. phosphorus silicate, borosilicate or lead silicate changed when the disc 10 is made of gallium arsenide or germanium. The thickness of the Layer 44 is preferably between 0.3 and 1 μm. If a glass of high purity (inert to the pn junction) is available, the layer can 44 can be omitted.

Thereafter, a sheet of glass 46 (FIG. 6), which has been heated to the softening point, is pressed into the grooves 42 coated with oxide. The in F i g. The induction furnace 26 shown in FIG. 3 can be used to soften the glass 46. Pressure can be applied between the glass 46 and the substrate 22 by placing the glass 46 on the surface 14 of each of the coated mesas 40 and this group between the blocks 32 and 34 at a pressure of between 1.4 and 350 kgf / cm ² and at a temperature sufficient to soften the glass 46. Although the introduction of the glass 46 into the grooves 42 can easily be done by the above-described hot pressing, the glass can also be introduced in other ways, for example by allowing it to sink into these spaces or by sedimentation, melting or vapor deposition. Different types of glass have favorable thermal expansion characteristics for use with silicon or other commercially available semiconductor materials.

ίο After the glass 46 has cooled and has hardened in the grooves 42, the opposing main surfaces of the block 30 are lapped up to the planes 47 and 49, which are indicated by the broken lines in FIG. 6, a first composite disk 50 (FIG. 7) being obtained. The assembled pane 50 now has a multiplicity of rectifier components or diodes 52 (the earlier mesas 40) which are oriented in the same way and which are spatially and electrically separated from one another by the glass 46. The opposing major surfaces 54 and 56 of the composite plate 50 are flat and each include opposed exposed surfaces, that is, contact areas or terminals 14 and α 48 a of the diode 52, as shown in F i g. 7 and 8 is shown.

The similarly oriented diodes 52 are not appropriately oriented in the assembled disk 50, to make electrical connections between them and thus certain circuit arrangements to manufacture such. B. the stack 60 (Fig. 10 and 10 a) of series-connected rectifiers. Therefore, the composite disk 50 is made into a plurality of parts, e.g. B. the parts 62, 64, 66 and 68, by sections along broken lines 55, 57 and 59 (Fig. 8) passing through the glass 46 lead, disassembled.

The parts 62 through 68 are now reassembled in a desired configuration. For example the parts 64 and 68 are turned over, i. H.

rotated by 180 ° about its longitudinal axis, as shown in FIG. The parts 62 to 68 are now reassembled to form a flat composite disk 70, forming the desired configuration. The parts 62-68 can be bonded together to form the composite disk 70 by reheating them in the induction furnace 26 until the glass 46 softens and the parts are held together. If necessary, sufficient pressure can be applied to the parts 62-68 to cause them to stick together as the glass softens. The parts 62 to 68 can also be used in other ways, e.g. B. by means of suitable putty, plastics or metallic conductors, e.g. B. lead, are attached to each other. Although, as described above, the composite disk 50 has been separated into four parts and. If these parts have been joined together in the configuration according to the assembled disk 70, the assembled disk 50 can also be disassembled into two or more parts in order to be reassembled in any desired configuration. Furthermore, the composite disk 70 can also be made from parts from various disassembled disks. For example, parts 62, 64, 66, and 68 can each come from four different composite disks 50.

To produce the rectifier stack 60

I 564 537

7 8

(Fig. 10 and 10a), the four series-connected arrangement 60 through the connections 82 and 84 geDioden 52, a section 71, which forms between.

see the two by the broken lines In Figs. 12 and 12a is an exemplary embodiment

72 and 74 (F i g. 9) indicated parallel planes play a rectifier bridge 90 shown which

lies, out of the assembled disc 70- 5 from a portion of the reassembled

taken. The section 71 can here be along the disk 70 (FIG. 9), which section is the

Lines 72 and 74 leading through the glass 46 to four diodes denoted by 52 e, 52 /, 52 g and 52 h

of any suitable type, e.g., by ultrasound. The connections 14a and 48a (Fig. 9) of the

treatment, sandblasting or chemical diodes 52 e and 52 / are each electrically through

Etching process to be cut out. 10 a conductor 92 (Fig. 12), e.g. metallic paint,

In Fig. 9 and in the subsequent figures are connected to one another; the terminals 14a and 48a some of diodes 52 with an additional reference of the diodes h 52 g and 52 provided letters in a similar manner, the one connected to the relative position via a conductor 94 to each other. The individual diodes 52 in the respective arrangement of conductors 92 and 94 can be identified more easily from the main surface 96. Suitable surfaces 98 will now run on the solid-state circuit 90 up to the adjoining section 71 taken over, in order to provide convenient access for connections in order to achieve the electrical connection lines shown in FIG. The formations are to be found on the rectifier stack 60 shown opposite the main surface 99 shown and 10 a. In this case, the connection 48 a of a diode circuit 14 a of both diodes 52 / and 52 h is connected to one another via a 52 a (FIG. 10 a) in the illustrated embodiment 20 conductor 100 , and the molding is connected to the connection 14 a of the adjacent circuits 48 a of the two diodes 52 e and 52 g, diode 526 is connected to one another by a connection 76 of both in a similar manner via a conductor 102 of, for example, metallic color, and is connected. The conductors 100 and 102 can also connection 48 α of the diode 52 b with the connection from the main surface 99 to the adjoining surface 104 connection 14 a of the adjacent diode 52 c through 25 so that the solid-state circuit is conveniently connected in a painted electrical line 78 . the desired overall circuit switched Similarly, the diode 52 may be a d.

Painted electrical line 80 connected in series. Fig. 13 shows schematically the circuit of the with the diode 52c. The end connections 82 and 84 rectifier bridge 90. In FIG. 13, the conductors of the rectifier stack 60 can represent the input terminals at the terminals 14 a and 48 a of the diodes 52 a and 52 d to be rectified and which are each painted on. As can be seen, conductors 100 and 102 can represent the output terminals that are adjacent diodes which can be conveniently removed in a stack in a known manner in one at which the rectified voltage is oriented in such a section.
Rectifier elements connected in series can be seen by merely lending adjacent ones that, according to the method described, elekindes from diodes by means of comparatively short connections from comparatively cheap disks with similarly oriented disks that are offset together on opposite sides of the section . circuit components isolated from each other

Instead of painting the connections, these 40 solid-state circuits can be made,

also by metal vapor deposition, by soldering albeit in the described embodiment

Wires, by attaching lead clips for making new rectifier assemblies

or in any other known manner a composite disk was used which

be attached. has a large number of diodes of the same type, so can

Fig. 11 shows the circuit diagram for the 45 for the production of other solid-state circuits also rectifier arrangement 60, in which the diodes 52 a composite disks, the other similar to 52 d through the connections 76 to 80 in series circuit components, z. B. transistors are switched off and the connections of the rectifier hold are used.

1 sheet of drawings

Claims (4)

Patent claims: 1. A method for producing a semiconductor circuit module, starting from a planar disk which is composed of a large number of similarly designed and arranged circuit components made of semiconductor material, which are separated from one another by an electrically insulating material that can be softened by means of heat, and contact surfaces on opposite main surfaces of the disk characterized in that the assembled disk (50) is divided along the insulating material (46) into several parts (62, 64, 66, 68) which have a layer of the insulating material along their separating surfaces, that at least one (64, 68) of these parts is turned over, that at least some of the parts, including the turned-over part, are reassembled and secured together in this arrangement by softening and bonding the insulating material by heating the parts to form a new planar disc (70), and that new planar disc (70) is formed from this Disc wh At least one section is removed which contains at least two circuit components (52 a, 52 c or 52 g, 52 o) of the un-turned part and at least two circuit components (52 b, 52 d or 52 /, 52 h) of the turned-over part are electrically connected to one another to form the circuit module (60, 90). 2. The method according to claim 1, characterized in that the circuit components as Diodes (52) are formed, through the electrical connections (92, 94,100,102) to a Rectifier bridge (90) are interconnected. 3. The method according to claim 1, characterized in that a section (60) is removed from the new disc (70) which contains a number of circuit components (52 a to 52 d) which are arranged alternately reversed with respect to their contact surfaces in the section are. 4. The method according to claim 3, characterized in that the circuit components as Diodes (52) are formed, which through the electrical connections (76, 78, 80) with one another can be connected in series.