JP2004158894A - Semiconductor device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve the problem associated with the performance of a read/write amplifier IC fastened to an FCA mounted on a hard disk drive that has a large impact on the property of the hard disk drive itself, since the bad performance is stemmed from the read/write speed deterioration of the read/write amplifier IC due to the fact that the bad heat dissipation of the IC has risen its temperature. <P>SOLUTION: An electrode 15 for heat dissipation is exposed from the backside of insulating resin 13, and a metal plate 23 is fastened to the electrode 15 for heat dissipation. The backside of the metal plate 23 and the backside of a flexible sheet are arranged substancially on the same plane, and can be easily fastened to a second support member 24. Therefore, heat generated from a semiconductor device can be dissipated well through the electrode 15 for heat dissipation, the metal plate 23 and the second support member 24. <P>COPYRIGHT: (C)2004,JPO

Description

  The present invention relates to a semiconductor device, and more particularly to a semiconductor device capable of satisfactorily releasing heat from a semiconductor element and buffering a stress acting on a solder electrode.

  2. Description of the Related Art In recent years, semiconductor devices have been increasingly used in portable devices and small-sized and high-density mounting devices, and there is a demand for light, thin, small, and heat-radiating properties. Moreover, the semiconductor device is mounted on various substrates, and is mounted on various devices as a semiconductor module including the substrate. The substrate may be a ceramic substrate, a printed substrate, a flexible sheet, a metal substrate, a glass substrate, or the like. Here, an example will be described below as a semiconductor module mounted on the flexible sheet.

  FIG. 23 shows a semiconductor module using a flexible sheet mounted on a hard disk 100 (see Non-Patent Document 1 below).

  The hard disk 100 is mounted on a box 101 made of metal. A plurality of recording disks 102 are integrally mounted on a spindle motor 103. A magnetic head 104 is provided on the surface of each recording disk 102 with a small gap. Are arranged through. The magnetic head 104 is attached to the tip of a suspension 106 fixed at the end of an arm 105. Then, the magnetic head 104, the suspension 106, and the arm 105 are integrated, and the integrated object is attached to the actuator 107.

  The recording disk 102 needs to be electrically connected to a read / write amplifying IC 108 for writing and reading via the magnetic head 104. Therefore, a semiconductor module 110 in which the read / write amplification IC 108 is mounted on the flexible sheet 109 is used, and the wiring provided on the flexible sheet 110 is finally electrically connected to the magnetic head 104. The semiconductor module 110 is called a flexible circuit assembly, and is generally abbreviated as FCA.

  On the back surface of the box 101, a connector 111 attached to the semiconductor module 110 appears, and this connector (male or female) 111 and a connector (female or male) attached to the main board 112 are provided. Connected. In addition, wiring is provided on the main board 112, and a driving IC of the spindle motor 103, a buffer memory, and other driving ICs such as an ASIC are mounted.

  For example, the recording disk 102 rotates at 4500 rpm via the spindle motor 103, and the position of the magnetic head 104 is determined by the actuator 107. Since the rotating mechanism is sealed by a lid provided on the box 101, heat is inevitably retained, and the temperature of the read / write amplification IC 108 rises. Therefore, the read / write amplification IC 108 is attached to a portion having excellent heat conduction, such as the actuator 107 and the box 101. Further, the rotation of the spindle motor 103 tends to be at a high speed of 5400, 7200, or 10,000 rpm, and this heat radiation becomes more and more important.

  FIG. 24 shows the structure of the above FCA in order to further explain it. FIG. 24A is a plan view, and FIG. 24B is a cross-sectional view, in which a portion of the read / write amplification IC 108 provided at the tip is cut by AA line. Since the FCA 110 is bent and attached to a part of the inside of the box 101, the first flexible sheet 109 having a planar shape that is easily bent is employed.

  A connector 111 is attached to the left end of the FCA 110 and serves as a first connection unit. A first wiring 121 electrically connected to the connector 111 is attached to the first flexible sheet 109 and extends to the right end. The first wiring 121 is electrically connected to the read / write amplification IC 108. A lead 122 of the amplification IC 108 connected to the magnetic head 104 is connected to a second wiring 123, and the second wiring 123 is provided on the arm 105 and the suspension 106 by a second flexible sheet 124. It is electrically connected to the third wiring 126 above. That is, the right end of the first flexible sheet 109 becomes the second connection portion 127, and is connected to the second flexible sheet 124 here. Note that the first flexible sheet 109 and the second flexible sheet 124 may be provided integrally. In this case, the second wiring 123 and the third wiring 126 are provided integrally.

  A support member 128 is provided on the back surface of the first flexible sheet 109 on which the read / write amplification IC 108 is provided. As the support member 128, a ceramic substrate or an Al substrate is used. Through the support member 128, the heat is thermally coupled to the metal exposed inside the box 101, and the heat of the read / write amplification IC 108 is released to the outside.

  Subsequently, the connection structure between the read / write amplification IC 108 and the first flexible sheet 109 will be described with reference to FIG. 24B.

  The flexible sheet 109 includes a first polyimide sheet 130 (hereinafter, referred to as a first PI sheet), a first adhesive layer 131, a conductive pattern 132, a second adhesive layer 133, and a second polyimide sheet 134 from below. (Hereinafter referred to as a second PI sheet) are stacked, and the conductive pattern 132 is sandwiched between the first and second PI sheets 130 and 134.

Further, in order to connect the read / write amplification IC 108, the second PI sheet 134 and the second adhesive layer 133 at desired positions are removed, an opening 135 is formed, and the conductive pattern 132 is exposed there. . Then, as shown in the figure, the read / write amplifying IC 108 is electrically connected via the lead 122.
Nikkei Electronics June 16, 1997 (No. 691) P92-

  In FIG. 24B, the semiconductor device packaged with the insulating resin 136 is released to the outside through the heat radiation path indicated by the arrow, and in particular, the insulating resin 136 becomes a thermal resistance. The structure was such that the generated heat could not be efficiently released to the outside.

  Further, a hard disk will be described. The read / write transfer rate of this hard disk is required to be 500 MHz to 1 GHz or higher, and the read / write speed of the read / write amplifying IC 108 must be increased. For that purpose, the route of the wiring on the flexible sheet connected to the read / write amplification IC 108 must be shortened to prevent the temperature of the read / write amplification IC 108 from rising.

  In particular, since the recording disk 102 rotates at a high speed and is a space sealed by the box 101 and the lid, the temperature inside rises to about 70 to 80 degrees. On the other hand, the permissible operating temperature of a general IC is about 125 degrees, and the read / write amplification IC 125 is allowed to rise from the internal temperature of 80 degrees to about 45 degrees. However, as shown in the figure, when the thermal resistance of the semiconductor device itself and the thermal resistance of FCA are large, the read / write amplification IC 108 immediately exceeds the allowable operating temperature, and cannot achieve its original capability. Therefore, a semiconductor device having excellent heat dissipation, FCA, is required.

    In addition, since the operating frequency is further increased in the future, there is a problem that the temperature of the read / write amplification IC 108 itself rises due to heat generated by the arithmetic processing. Although the desired operating frequency can be achieved at room temperature, the operating frequency has to be reduced inside the hard disk due to the temperature rise.

  As described above, as the operating frequency increases in the future, semiconductor devices and semiconductor modules (FCAs) have been required to have more heat dissipation.

  On the other hand, the actuator 107 itself, and the arm 105, the suspension 106, and the magnetic head 104 attached thereto must be made as light as possible to reduce the moment of inertia. In particular, as shown in FIG. 17, when the read / write amplifying IC 108 is mounted on the surface of the actuator 107, the weight of the IC 108 and the weight of the FCA 110 have been required to be reduced.

The present invention has been made in view of the above-described problems, and firstly, a semiconductor element is sealed face-down integrally with an insulating resin, and a back surface thereof is electrically connected to a bonding electrode of the semiconductor element. A semiconductor device in which pads and heat radiation electrodes located on the surface of the semiconductor element are exposed,
The problem is solved by providing a metal plate on the exposed portion of the heat radiation electrode so as to protrude from the back surface of the pad.

  Since the protruding metal plate is positioned at the same position as the back surface of the flexible sheet serving as the first support member, the metal plate can be bonded or abutted to the inside of the housing, particularly to a portion having a flat surface, a heat sink, or the like. Become. Therefore, heat of the semiconductor element can be transmitted to the heat sink.

  Second, the problem is solved by arranging the back surface of the pad and the back surface of the heat radiation electrode on substantially the same plane.

  Third, the problem is solved by fixing the semiconductor element and the electrode for heat dissipation with an insulating material.

  Fourth, the problem is solved by fixing the heat dissipation electrode and the metal plate with an insulating material or a conductive material.

  Fifth, the problem is solved by forming the heat dissipation electrode and the metal plate integrally from the same material.

  Sixth, the problem is solved by the back surface of the insulating resin protruding from the back surface of the pad.

  Seventh, the problem is solved by drawing the same curved surface between the side surface of the pad and the back surface of the insulating resin extending from the side surface of the pad.

Eighth, a first support member provided with a conductive pattern,
A semiconductor element electrically connected to the conductive pattern is integrally sealed face down with an insulating resin, and a pad electrically connected to a bonding electrode of the semiconductor element and a front surface of the semiconductor element are formed on a back surface thereof. And a conductive pattern provided on the first support member and the pad are electrically connected to each other, and correspond to the electrode for heat dissipation. The first support member is provided with an opening, and the opening is provided with a metal plate fixed to the electrode for heat dissipation.

  Ninth, the problem is solved by attaching a second support member to which the metal plate is fixed to the back surface of the first support member.

  Tenth, the problem is solved by forming the heat dissipation electrode and the metal plate integrally with the same material.

  Eleventh, the problem is solved by providing a fixing plate made of a conductive material on the second support member corresponding to the metal plate, and thermally fixing the fixing plate and the metal plate. .

  Twelfth, the metal plate is mainly made of Cu, the second support member is mainly made of Al, and the fixing plate is mainly made of Cu formed on the second support member. The problem is solved by using a plating film.

  A thirteenth problem is solved by projecting the back surface of the insulating resin from the back surface of the pad.

  Fourteenth, the problem is solved by drawing the same curved surface between the side surface of the pad and the back surface of the insulating resin extending from the side surface of the pad.

  Fifteenth, the problem is solved by the semiconductor element being a read / write amplifying IC for a hard disk.

Sixteenth, the semiconductor element is integrally sealed face down with the insulating resin, and the back surface of the semiconductor element is extended through a pad electrically connected to the bonding electrode of the semiconductor element and a wiring integrated with the pad. External connection electrode, and a semiconductor device in which an electrode for heat radiation disposed on the surface of the semiconductor element is exposed,
The problem is solved by providing a metal plate on the exposed portion of the heat radiation electrode so as to protrude from the back surface of the external connection electrode.

  Seventeenth, the problem is solved by arranging the back surface of the external connection electrode and the back surface of the heat radiation electrode on substantially the same plane.

  Eighteenth, the problem is solved by fixing the semiconductor element and the heat radiation electrode with an insulating material.

  Nineteenth, the problem is solved by fixing the heat radiation electrode and the metal plate with an insulating material or a conductive material.

  Twentiethly, the problem is solved by forming the heat dissipation electrode and the metal plate integrally from the same material.

  Twenty-first, the problem is solved by protruding the back surface of the insulating resin from the back surface of the external connection electrode.

  Twenty-second, the problem is solved by drawing the same curved surface between the side surface of the external connection electrode and the back surface of the insulating material extending from the side surface of the external connection electrode.

23rdly, a first support member provided with a conductive pattern,
A semiconductor element electrically connected to the conductive pattern is sealed face down integrally with the insulating resin, and a pad electrically connected to the bonding electrode of the semiconductor element on the back surface thereof, and a pad integrated with the pad. A semiconductor module having an external connection electrode provided through the wiring of the above, and a semiconductor device in which a heat radiation electrode located on the surface of the semiconductor element is exposed,
The conductive pattern provided on the first support member and the external connection electrode are electrically connected, and the first support member corresponding to the electrode for heat radiation has an opening, The problem is solved by providing a metal plate fixed to the electrode for heat dissipation in the opening.

  Twenty-fourth, the problem is solved by attaching a second support member to which the metal plate is fixed on the back surface of the first support member.

  Twenty-fifth, the problem can be solved by forming the heat dissipation electrode and the metal plate integrally from the same material.

  Twenty-sixth, the second support member corresponding to the metal plate is provided with a fixing plate made of a conductive material, and the problem is solved by thermally bonding the fixing plate and the metal plate. .

  Twenty-seventh, the metal plate is mainly made of Cu, the second support member is mainly made of Al, and the fixing plate is mainly made of Cu formed on the second support member. The problem is solved by using a plating film.

  Twenty-eighth, the problem is solved by the fact that the back surface of the insulating bonding means projects beyond the back surface of the external connection electrode.

  Twenty-ninth, the problem is solved by drawing the same curved surface between the side surface of the external connection electrode and the back surface of the insulating bonding means bonded to the external connection electrode.

  Thirtieth, the problem is solved by the semiconductor element being a read / write amplification IC for a hard disk.

Thirty-first, a semiconductor device of the present invention includes a bonding pad provided corresponding to a bonding electrode of a semiconductor element, an external connection electrode provided on the back surface of the bonding pad, and a semiconductor device provided in an arrangement region of the semiconductor element. A pad, an external connection electrode for stress buffering provided on the back surface of the pad, a bonding means provided on the pad, and the bonding means fixed to the bonding means and electrically connected to the bonding pad. A semiconductor element and an insulating resin for sealing the semiconductor element so as to expose and integrate the back surface of the pad, the back surface of the external connection electrode and the back surface of the bonding means,
The external connection electrode for stress buffering is formed to be sufficiently larger than the external connection electrode, so that the stress due to the thermal expansion of the insulating resin is transmitted to the external connection electrode by buffering.

  32ndly, the pad and the external connection electrode for stress buffering are solved by being divided into four.

Thirty-third, a semiconductor device according to the present invention includes a bonding pad provided corresponding to a bonding electrode of a semiconductor element, an external connection electrode provided on a back surface of the bonding pad, and a semiconductor device provided in an arrangement region of the semiconductor element. A pad, a bonding unit provided on the pad, the semiconductor element fixed to the bonding unit and electrically connected to the bonding pad, a back surface of the pad, a back surface of the external connection electrode, and Having an insulating resin for sealing the semiconductor element so as to expose and integrate the back surface of the bonding means,
The problem is solved by forming the external connection electrode to be elongated.

34thly, the semiconductor device of the present invention is provided with a bonding pad provided corresponding to a bonding electrode of a semiconductor element, an external connection electrode provided on a back surface of the bonding pad, and an arrangement region of the semiconductor element. A pad, a bonding unit provided on the pad, the semiconductor element fixed to the bonding unit and electrically connected to the bonding pad, a back surface of the pad, a back surface of the external connection electrode, and Having an insulating resin for sealing the semiconductor element so as to expose and integrate the back surface of the bonding means,
The problem is solved when the side surfaces of the external connection electrodes have the same curved surface.

  The thirty-fifth problem is solved by that the external connection electrode is a solder electrode.

  As apparent from the above description, the present invention provides a semiconductor device in which a metal plate is fixed to a heat radiation electrode exposed on the back surface of a package and the metal plate protrudes from the back surface of an external connection electrode or pad. Accordingly, there is an advantage that mounting on the FCA becomes easy.

  In particular, an opening is provided in the FCA, and the heat dissipating electrode of the semiconductor device and the back surface of the FCA are located at the same surface position, so that the contact with the second support member is facilitated.

  Further, by using Al as the second support member, a fixing plate made of Cu is formed here, and a heat-radiating electrode or a metal plate is fixed to the fixing plate, so that heat generated from the semiconductor element is transferred to the second plate. It can be released to the outside via the support member.

  Therefore, a temperature rise of the semiconductor element can be prevented, and a performance close to the original capability can be obtained. In particular, the FCA mounted in the hard disk can efficiently release the heat to the outside, so that the read / write speed of the hard disk can be increased.

  Further, by providing a stress buffering electrode having the same size as the semiconductor element on the back surface of the pad in the center of the semiconductor device, the stress acting on the external connection electrode can be reduced. This can prevent the occurrence of solder cracks.

  Further, by making the external connection electrode of the semiconductor device slender and having no constriction, the stress acting on the solder electrode can be buffered and the strength of the solder electrode itself can be improved. Therefore, solder cracks can be prevented.

  The present invention provides a semiconductor device having high heat dissipation, lightness, and shortness, and at the same time, a semiconductor module mounted with the semiconductor device, for example, a semiconductor module mounted on a flexible sheet (hereinafter referred to as FCA). This implements the improvement of the characteristics of a mounted device, for example, a hard disk.

  Further, the present invention is to prevent a crack from being generated in the external connection electrode by providing an electrode for stress buffering in the center of the semiconductor device and further forming the external connection electrode in a shape without constriction.

First, as an example of a device on which the FCA is mounted, the hard disk 100 is referred to in FIG. 17, and the FCA is shown in FIG. 2 to 16 show a semiconductor device mounted on the FCA or a method of manufacturing the same.

First Embodiment Explaining Device on Which FCA is Mounted As this device, the hard disk 100 of FIG. 17 described in the section of the related art will be described again.

  Since the hard disk 100 is mounted on a computer or the like, it is mounted on the main board 112 as necessary. The main board 112 has a female (or male) connector mounted thereon. The male (or female) connector 111 mounted on the FCA and exposed from the back surface of the box 101 is connected to the connector on the main board 112. In the box 101, a plurality of recording disks 102 as recording media are stacked according to the capacity. Since the magnetic head 104 flies above the recording disk 102 at about 20 to 30 nm and is scanned, the interval between the recording disks 102 is set to an interval that does not cause a problem in this scanning. And it is attached to the spindle motor 103 at this interval. The spindle motor 103 is attached to a mounting board, and a connector arranged on the back of the mounting board has a face protruding from the back of the box 101. This connector is also connected to the connector of the main board 112. Therefore, the main board 112 includes an IC for driving the read / write amplifying IC 108 of the magnetic head 104, an IC for driving the spindle motor 103, an IC for driving the actuator, a buffer memory for temporarily storing data, and an ASIC for realizing a manufacturer's own drive. Etc. are implemented. Naturally, other passive elements and active elements may be mounted.

  The wiring for connecting the magnetic head 104 and the read / write amplification IC 108 is considered as short as possible, and the read / write amplification IC 108 is disposed on the actuator 107. However, since the semiconductor device of the present invention described below is very thin and lightweight, it may be mounted on the arm 105 or the suspension 106 in addition to the actuator. In this case, as shown in FIG. 1B, since the back surface of the semiconductor device 10 is exposed from the opening 12 of the first support member 11, the back surface of the semiconductor device 10 is thermally coupled to the arm 105 or the suspension 106, Heat of the semiconductor device 10 is released to the outside via the arm 105 and the box 101.

As shown in FIG. 17, when the read / write amplifying IC 108 is mounted on the actuator 107, a read / write circuit for each channel is formed by one chip so that a plurality of magnetic sensors can read / write. However, it is only necessary that a read / write circuit dedicated to the magnetic head 104 attached to each suspension 106 be mounted on each suspension. In this way, the wiring distance between the magnetic head 104 and the read / write amplifying IC 108 can be made much shorter than that of the structure shown in FIG. 18, so that the impedance can be reduced and the read / write speed can be improved.

Second Embodiment for Describing Semiconductor Device First, a semiconductor device of the present invention will be described with reference to FIG. 2A is a plan view of the semiconductor device, and FIG. 2B is a cross-sectional view taken along line AA.

  In FIG. 2, the following components are embedded in the insulating resin 13. That is, the pads 14, the heat radiation electrode 15 provided in the region surrounded by the pad 14, and the semiconductor element 16 provided on the heat radiation electrode 15 are embedded. The semiconductor element 16 is mounted face-down, fixed to the heat-dissipating electrode 15 via insulating adhesive means 17, and divided into four parts in consideration of adhesiveness. The separation groove formed by the four divisions is indicated by reference numeral 18A. When the gap between the semiconductor element 16 and the heat radiation electrode 15 is small and the insulating bonding means 17 is hard to penetrate, a groove shallower than the separation groove 18A is formed on the surface of the heat radiation electrode 15A as shown at 18B. May be formed on the surface.

  The bonding electrode 19 of the semiconductor element 16 and the pad 14 are electrically connected via a brazing material such as solder. Note that Au stud bumps may be used instead of solder.

  Note that there are other connection methods. For example, a bump may be attached to the bonding electrode 19 of the semiconductor element, and the bump may be connected by ultrasonic waves or pressure welding. Alternatively, solder, conductive paste, or anisotropic conductive particles may be provided around the pressed bumps. These structures will be described in detail at the end of the description of the embodiment with reference to FIG.

  In addition, the back surface of the pad 14 is exposed from the insulating resin 13 and becomes the external connection electrode 21 as it is. The side surfaces of the pads 14 are non-anisotropically etched, and are formed by wet etching. And the curved structure produces an anchor effect.

  This structure is composed of five materials: a semiconductor element 16, a plurality of conductive patterns 14, electrodes 15 for heat dissipation, insulating bonding means 17, and insulating resin 13 for embedding these. In the region where the semiconductor element 16 is arranged, the insulating bonding means 17 is formed on the heat-dissipating electrodes 15, on the pads 14, and between the pads 14, and particularly on the separation grooves 18 formed by etching. The insulating bonding means 17 is provided, and the back surface of the insulating bonding means is exposed from the back surface of the semiconductor device 10A. In addition, everything including these is sealed with the insulating resin 13. The pads 14,..., The radiation electrode 15, and the semiconductor element 16 are supported by the insulating resin 13 and the insulating bonding means 17.

  As the insulating bonding means 17, an adhesive made of an insulating material or an underfill material is preferable. In the case of an adhesive, it may be applied to the surface of the semiconductor element 16 in advance and fixed when connecting the pad 14 using an Au bump instead of the solder 20. Further, the underfill material 17 may be permeated into the gap after connecting the solder 20 (or bump) and the pad 14.

  As the insulating resin, a thermosetting resin such as an epoxy resin, or a thermoplastic resin such as a polyimide resin or polyphenylene sulfide can be used.

  As the insulating resin 13, any resin can be adopted as long as it is a resin that is hardened using a mold, a resin that can be dipped, applied and covered. Further, as the conductive pattern 14, a conductive foil mainly composed of Cu, a conductive foil mainly composed of Al, a laminate of Fe-Ni alloy, Al-Cu, a laminate of Al-Cu-Al, or the like is used. Can be. Of course, other conductive materials are possible, and in particular, a conductive material that can be etched and a conductive material that evaporates by laser are preferable. In consideration of half-etching property, plating formability, and thermal stress, a conductive material mainly formed of Cu formed by rolling is preferable.

  The present invention has a feature that the separation of the conductive pattern can be prevented because the insulating resin 13 and the insulating bonding means 17 are also filled in the separation groove 18. Further, by performing non-anisotropic etching by employing dry etching or wet etching as the etching, the side surfaces of the pads 14 can have a curved structure to generate an anchor effect. As a result, a structure in which the pad 14 and the heat radiation electrode 15 do not come off from the insulating resin 13 can be realized.

  Moreover, the back surface of the heat-dissipating electrode 15 is exposed on the back surface of the package. Therefore, the back surface of the heat-dissipating electrode 15 has a structure that can be in contact with or fixed to the metal plate 23, the second support member 24, or the fixing plate 25 formed on the second support member 24, which will be described later. Therefore, with this structure, heat generated from the semiconductor element 16 can be radiated, the temperature of the semiconductor element 16 can be prevented from rising, and the drive current and the drive frequency of the semiconductor element 16 can be increased accordingly.

  Since the semiconductor device 10A supports the pads 14 and the electrodes 15 for heat radiation with the insulating resin 13 as a sealing resin, a support substrate is not required. This configuration is a feature of the present invention. A conductive path of a conventional semiconductor device is supported by a support substrate (a flexible sheet, a printed circuit board, or a ceramic substrate) or supported by a lead frame, and thus a configuration that may not be necessary is added. However, the circuit device has a feature that it is composed of the minimum necessary components and does not require a supporting substrate, so that it is thin and lightweight, and is inexpensive because the material cost can be reduced. Therefore, as described in the first embodiment, it can be mounted on the arm or suspension of the hard disk.

  The pad 14 and the heat radiation electrode 15 are exposed on the back surface of the package. When this region is covered with a brazing material such as solder, the electrode 15 for heat radiation has a larger area, and the film thickness of the brazing material differs and wets. Therefore, the insulating film 26 is formed on the back surface of the semiconductor device 10A to make the thickness of the brazing material uniform. A dotted line 27 shown in FIG. 2A indicates an exposed portion exposed from the insulating coating 26. Here, since the back surface of the pad 14 is exposed in a rectangular shape, the same size as the exposed portion is exposed from the insulating coating 26.

  Therefore, since the wetted portion of the brazing material has substantially the same size, the thickness of the brazing material formed here becomes substantially the same. This is the same after solder printing and after reflow. The same can be said for conductive pastes such as Ag, Au and Ag-Pd. With this structure, it is possible to accurately calculate how much the rear surface of the metal plate 23 projects from the rear surface of the pad 14.

  If the metal plate 23 and the conductive pattern 32 are set on the same surface, both the pad 14 and the heat radiation electrode 15 can be soldered at one time. The exposed portion 27 of the electrode 15 for heat dissipation may be formed larger than the exposed size of the pad 14 in consideration of heat dissipation of the semiconductor element.

By providing the insulating film 26, the conductive pattern provided on the first support member 11 can be extended to the back surface of the semiconductor device 10A. Generally, the wiring provided on the first support member 11 side is arranged to bypass the fixing region of the semiconductor device 10A in order to prevent a short circuit. Can be placed without Moreover, since the insulating resin 13 and the insulating bonding means 17 protrude from the conductive pattern, the solder SD provided on the back surface of the semiconductor device 10A does not short-circuit.

Third Embodiment Explaining Semiconductor Device 10B FIG. 3 shows the present semiconductor device 10B. FIG. 3A is a plan view, and FIG. 3B is a cross-sectional view taken along line AA. Since the structure is similar to that of FIG. 2, only different parts will be described here.

  In FIG. 2, the back surface of the pad 14 directly functions as the external connection electrode 21. However, in the present embodiment, the pad 14 has the wiring 30 formed integrally and the external connection electrode 31 formed integrally with the wiring 30. Is formed.

  The rectangle indicated by the dotted line is the semiconductor element 16, and the external connection electrode 31 is arranged on the back surface of the semiconductor element 16, and the external connection electrode 31 is arranged in a ring shape as shown in the figure. This arrangement has the same or similar structure as a known BGA.

  If the semiconductor element 16 is directly disposed on the conductive patterns 14, 30, 31 and the heat radiation electrode 15, there is a possibility that both of them are short-circuited via the back surface of the semiconductor element 16. Therefore, the insulating bonding means 17 should use only an insulating material, and cannot use a conductive material.

  The conductive pattern 32 of the first support member 11 is connected to the external connection electrode 31, and the back surface of the pad 14 and the back surface of the wiring 30 are covered with the insulating film 26. A circle indicated by a dotted line on the external connection electrode 31 and a circle indicated by a dotted line on the electrode 15 for heat dissipation are portions exposed from the insulating film 26.

  Further, since the external connection electrode 31 extends on the back surface of the semiconductor element 16, the heat radiation electrode 15 is formed smaller than the heat radiation electrode 15 of FIG. The insulating bonding means 17 covers the heat radiation electrode 15, the external connection electrode 31, and the wiring 30. The insulating resin 13 covers the pad 14, the wiring 30, the semiconductor element 16, and the thin metal wire 20.

  This embodiment has an advantage that when the number of pads 14 is very large and the size is reduced, the pads 14 can be rearranged as external connection electrodes via wiring, and the size of the external connection electrodes 31 can be increased. Further, by providing the wiring, distortion applied to the bonding portion can be reduced. It is particularly preferable to form it in a wave shape. Further, since the semiconductor element 16 and the heat radiation electrode 15 are fixed by the insulating adhesive means 17 and are made of an insulating material, their thermal resistance becomes a problem. However, if the insulating bonding means is made of a silicone resin mixed with a filler that contributes to heat conduction such as Si oxide or aluminum oxide, the heat of the semiconductor element 16 can be transmitted well to the heat radiation electrode 15.

The distance between the heat radiation electrode 15 and the back surface of the semiconductor element 16 can be made uniform by unifying the diameter of the filler. Therefore, when a minute gap is formed in consideration of heat conduction, the gap can be formed by lightly pressing the semiconductor element 16 when the insulating bonding means is in a softened state and hardening as it is.

Fourth Embodiment Explaining a Manufacturing Method of Semiconductor Devices 10A and 10B This manufacturing method is different from a pattern formed by a pad 14 and a heat radiation electrode 15 in that a wiring 30 and an external connection electrode 31 are added thereto. The only difference is the pattern. In any case, since it is formed in a convex shape by half etching, it is substantially the same except for this pattern shape.

  Here, a manufacturing method thereof will be described using the semiconductor device 10B of FIG. 4 to 9 are cross-sectional views corresponding to line AA in FIG. 3A.

First, a conductive foil 40 is prepared as shown in FIG. The thickness is preferably about 10 μm to 300 μm, and here, a rolled copper foil of 70 μm was employed. Subsequently, a conductive film 41 or a photoresist is formed on the surface of the conductive foil 40 as an etching resistant mask. This pattern is the same pattern as the pads 14,..., The wirings 30,..., The external connection electrodes 31,. When a photoresist is used instead of the conductive film 41, a conductive film such as Au, Ag, Pd, or Ni is formed under the photoresist at least at a portion corresponding to the pad. This is provided to enable bonding. (See Figure 4 above)
Subsequently, the conductive foil 40 is half-etched via the conductive film 41 or the photoresist. The etching depth may be smaller than the thickness of the conductive foil 40. Note that the shallower the etching depth, the finer the pattern can be formed.

  Then, by half-etching, the conductive patterns 14, 30, 31 and the heat radiation electrode 15 appear on the surface of the conductive foil 40 in a convex shape. As described above, the conductive foil 40 used herein was a Cu foil formed mainly by rolling and having Cu as a main material. However, a conductive foil made of Al, a conductive foil made of an Fe-Ni alloy, a laminate of Cu-Al, or a laminate of Al-Cu-Al may be used. In particular, a laminate of Al-Cu-Al or Cu-Al-Cu can prevent warpage caused by a difference in thermal expansion coefficient.

Then, an insulating bonding means 17 is provided at a portion corresponding to the rectangular dotted line in FIG. The insulating bonding means 17 is provided on the separation groove 22 between the heat radiation electrode 15 and the external connection electrode 31, the separation groove between the heat radiation electrode 15 and the wiring 30, the separation groove between the wiring 30, and the like. . Reference numeral DM denotes a flow prevention film of the solder SD1 fixed here. If the flow prevention film DM is not provided, the semiconductor element 16 is inclined, and the insulating bonding means 17 cannot be injected between the semiconductor element and the conductive foil, or cleaning cannot be performed. (See Figure 5 above)
Subsequently, the semiconductor element 16 is fixed to the region where the insulating adhesive means 17 is provided, and the bonding electrode 19 of the semiconductor element 16 and the pad 14 are electrically connected. In the drawing, since the semiconductor element 16 is mounted face down, the solder SD1 is employed as the connection means.

  In this bonding, the pads 14 are integral with the conductive foil 40, and since the back surface of the conductive foil 40 is flat, the pads 14 abut against the table of the bonding machine. Therefore, if the conductive foil 40 is completely fixed to the bonding table, all the pads 14 and all the solder balls formed on the semiconductor element 16 are brought into contact with each other, and can be fixed without a solder defect. The bonding table can be fixed by, for example, providing a plurality of vacuum suction holes on the entire surface of the table. Note that there are other connection methods, and this structure will be described last with reference to FIG.

  In addition, the height of the semiconductor element 16 is set lower by the fact that the supporting substrate is not used and the solder ball is used instead of the thin metal wire. Therefore, the thickness of the package outer shape described later can be reduced.

When an underfill is used as the insulating bonding means 17, the semiconductor element 16 and the pad 14 are fixed and then the underfill is permeated. (See FIG. 6 above)
Then, the insulating resin 13 is formed in the entire region including the semiconductor element 16. The insulating resin may be either thermoplastic or thermosetting.

  Further, it can be realized by transfer molding, injection molding, dipping or coating. As the resin material, a thermosetting resin such as an epoxy resin can be realized by transfer molding, and a thermoplastic resin such as a liquid crystal polymer and polyphenylene sulfide can be realized by injection molding.

  In the present embodiment, the thickness of the insulating resin is adjusted so as to cover about 100 μm from the back surface of the semiconductor element. This thickness can be increased or reduced in consideration of the strength of the semiconductor device. Further, as shown in FIG. 14B, the back surface of the semiconductor element 16 may be exposed. In this case, heat radiation fins can be attached or heat of the semiconductor element can be directly radiated to the outside.

  In addition, in the resin injection, the pad 14, the wiring 30, the external connection electrode 31, and the heat-dissipating electrode 15 are integrated with the sheet-shaped conductive foil 40. There is no pattern displacement at all. Moreover, unlike the lead frame, no resin burrs are generated between these conductive patterns.

As described above, the pad 14, the wiring 30, the external connection electrode 31, the electrode 15 for heat dissipation, and the semiconductor element 16 are embedded in the insulating resin 13 and the conductive foil 40 below the convex is formed on the back surface. It is exposed from. (See Figure 7 above)
Subsequently, the conductive foil 40 exposed on the back surface of the insulating resin 13 is removed, and the pad 14, the wiring 30, the external connection electrode 31, and the heat radiation electrode 15 are individually separated.

  Various methods are conceivable for the separation step here, and the separation may be performed by removing the back surface by etching, or may be performed by polishing or grinding to separate. Further, both may be adopted. For example, if the insulating resin 13 is cut until the insulating resin 13 is exposed, shavings of the conductive foil 40 and thin burr-like metal that is thinly spread out into the insulating resin 13 and the insulating bonding means 17. There's a problem. Therefore, if they are separated by etching, they can be formed without the metal of the conductive foil 40 penetrating into the surface of the insulating resin 13 and the insulating bonding means 17 located between the Cu patterns. As a result, short-circuiting between patterns at minute intervals can be prevented.

  When a plurality of units constituting the semiconductor device 10B are integrally formed, a dicing step is added after the separation step.

  Here, a dicing apparatus is used to separate the individual pieces, but it is also possible to use a chocolate break, press or cut.

  Here, after separating the Cu pattern, an insulating film 26 is formed on the separated patterns 14, 30, 31, and 15 exposed on the back surface, and the insulating film 26 is so formed as to expose a portion shown by a dotted circle in FIG. 3A. The coating 26 is patterned. Thereafter, dicing is performed at the portion indicated by the arrow, and the semiconductor device 10B is cut out.

  The solder 42 may be formed before or after dicing.

  By the above-described manufacturing method, the pads, wiring, external connection electrodes, electrodes for heat dissipation, and semiconductor elements are embedded in the insulating resin, so that a package that is light, thin, short, and excellent in heat dissipation can be realized.

Incidentally, as shown in FIG. 9, the sealing may be performed by using the insulating bonding means 17 without using the insulating resin 13. A pattern PTN shown in FIG. 10 indicates a pad, a wiring, and an external connection electrode, and a hatched portion above this indicates a pattern for forming a solder flow prevention film. At the same time as preventing the flow of solder, it is also applied to other areas to improve the adhesion of the insulating bonding means. Although A to E are shown as types, other patterns may be employed. Also, a flow prevention film may be formed on the entire conductive foil except for the solder connection portion.

Next, effects produced by the above-described manufacturing method will be described.

  First, since the conductive pattern is half-etched and supported integrally with the conductive foil, the substrate conventionally used for support can be eliminated.

  Second, the conductive foil is formed with a conductive pattern that is half-etched to be a convex portion, so that the conductive pattern can be miniaturized. Therefore, the width and the interval can be reduced, and a package having a smaller planar size can be formed.

  Third, since it is composed of a conductive pattern, a semiconductor element, a connecting means, and a sealing material, it can be configured with the minimum necessary, and can eliminate unnecessary materials as much as possible. Can be realized.

  Fourth, pads, wirings, external connection electrodes, and heat radiation electrodes are formed as projections by half-etching, and individual separation is performed after sealing, so that tie bars and suspension leads are not required. Therefore, the formation of the tie bar (suspension lead) and the cutting of the tie bar (suspension lead) are completely unnecessary in the present invention.

  Fifth, after the conductive pattern serving as the protrusion is embedded in the insulating resin, the conductive foil is removed from the back surface of the insulating resin to separate the conductive pattern. Resin burrs generated between the leads can be eliminated.

  Sixth, the semiconductor element is fixed to the electrode for heat dissipation through the insulating adhesive means, and the electrode for heat dissipation is exposed from the back surface, so that the heat generated from the semiconductor device is removed from the front surface of the semiconductor device. It can be efficiently released to the electrode for heat radiation. Furthermore, by mixing a filler such as a Si oxide film or aluminum oxide into the insulating bonding means, the heat dissipation can be further improved. If the filler size is unified, the gap between the semiconductor element 16 and the conductive pattern can be kept constant.

Fifth Embodiment Explaining Semiconductor Devices 10A and 10B to which Metal Plate 23 is Fixed and Semiconductor Module Using the Same FIG. 1 shows a semiconductor module (FCA) 50. The mounted semiconductor device is the semiconductor device 10B shown in FIG.

  First, the first support member 11 made of a flexible sheet will be described. Here, a first PI sheet 51, a first adhesive layer 52, a conductive pattern 53, a second adhesive layer 54, and a second PI sheet 55 are sequentially stacked from the lower layer. In the case where the conductive pattern has a multilayer structure, an adhesive layer is further used, and the upper and lower conductive patterns may be electrically connected via through holes in some cases. As shown in FIG. 1C, the first support member 11 is formed with a first opening 12 that allows at least the metal plate 23 to be exposed.

  Then, a second opening 56 is formed so that the conductive pattern is exposed. The entire conductive pattern 32 corresponding to the second opening 56 may be exposed, or only the connected portion may be exposed. For example, the second PI sheet 55 and the second adhesive layer 54 may all be removed, or, as shown in the drawing, the second PI sheet 55 is entirely removed and only the second adhesive layer 54 is exposed. Parts may be removed. In this case, the solder 27 does not flow.

  In the semiconductor device of the present invention, the metal plate 23 is bonded to the back surface of the heat radiation electrode 15. Further, in the semiconductor module according to the present invention, the rear surface of the first support member and the metal plate 23 are located substantially at the same surface position.

  The thickness of the metal plate 23 is determined in consideration of the thickness of the first support member 11 and the fixing plate 25. Then, when the pad 14 and the conductive pattern 32 are fixed via the solder 27, each metal plate 23 exposed from the first opening 12 is substantially flush with the back surface of the first support member 11. The thickness has been determined. Therefore, it is possible to make contact with the second support member, and further, it is possible to make contact with and fix to the second support member having the fixing plate 25.

  Some specific examples of this connection structure will be described.

  In the first example, a structure in which a lightweight metal plate such as Al or stainless steel or a ceramic substrate is adopted as the second support member 24, and the metal plate 23 fixed to the back surface of the semiconductor device 10 is brought into contact therewith. It is. That is, this is a structure in which the second support member 24 is directly contacted without the interposition of the fixing plate 25. The electrode 15 for heat dissipation and the metal plate 23, and the metal plate 23 and the second support member 24 are fixedly selected by using a brazing material such as solder or an insulating bonding means containing filler and having excellent heat conductivity. .

  In the second example, a lightweight metal plate such as Al or stainless steel or a ceramic substrate is adopted as the second support member 24, and a fixing plate 25 is formed thereon, and the fixing plate 25 and the metal plate 23 are fixed. Structure.

  For example, when Al is used as the second support member 24, the fixing plate 25 is preferably made of Cu. This is because Cu plating is possible on Al. The film thickness may be about 10 μm. Moreover, since it is a plated film, it can be formed in close contact with the second support member 24, and the thermal resistance between the fixing plate 25 and the second support member 24 can be extremely small. Alternatively, a conductive paste may be applied as the fixing plate 25 and used instead.

  On the other hand, the Cu fixing plate 25 and the Al substrate can be fixed via an adhesive, but in this case, the thermal resistance increases.

  When a ceramic substrate is used as the second support member 24, the fixing plate 25 is formed on an electrode formed by printing and firing a conductive paste.

  The second support member 24 and the first support member 11 are fixed by a third adhesive layer 57.

For example,
First PI sheet 51: 25 μm
Second PI sheet 55: 25 μm
First to third adhesive layers 52, 54, 57: 25 μm (after firing)
Acrylic adhesive is used as the material Solder 27: 50 μm
Third adhesive layer 57: 25 μm
Acrylic adhesive is used as the material. Fixing plate 25: about 25 μm.
As described above, if the respective film thicknesses are adjusted and determined, after the semiconductor device 10A is fixed to the first support member 11, the fixing plate 25 is easily provided and the competing second support member 24 is bonded. I can do it.

  Also, a module is prepared in which the second support member 24 is bonded to the first support member 11, the semiconductor device 10 is placed in the opening 56 formed in this module, and then the solder is melted. Can be melted and fixed without poor connection.

Therefore, heat generated from the semiconductor element 16 can be released to the second support member 24 through the heat radiation electrode 15, the metal plate 23, and the fixing plate 25. Moreover, since the thermal resistance is significantly reduced as compared with the conventional structure (FIG. 17B), the drive current and drive frequency of the semiconductor element 16 can be increased. The back surface of the second support member 24 can be attached to the actuator 107, the bottom surface of the box 101, or the arm 105 shown in FIG. Therefore, finally, the heat of the semiconductor element can be released to the outside via the box 101. Therefore, even when the semiconductor module is mounted on the hard disk 100, the semiconductor element itself does not reach a relatively high temperature, and the read / write speed of the hard disk 100 can be further increased. Note that the FCA may be mounted on a device other than the hard disk. In this case, the second support member is in contact with a member having low thermal resistance. When mounted on another device, a printed circuit board or a ceramic substrate may be used instead of the flexible sheet.

Sixth Embodiment for Explaining a Semiconductor Device 10C and a Semiconductor Module 50A That Have a Heat Dissipating Electrode 15 Protruding Instead of the Metal Plate 23 In FIG. The structure in which the heat radiation electrode 15 and the fixing plate 25 are integrated is shown.

  First, this manufacturing method will be described with reference to FIGS. 4 to 8 are the same manufacturing method, and the description up to this point is omitted.

  FIG. 12 shows a state in which the insulating resin 13 is coated on the conductive foil 40, and a portion corresponding to the electrode 15 for heat dissipation is coated with a photoresist PR. When the conductive foil 40 is etched through the photoresist PR, the heat radiation electrode 15A can have a structure protruding from the back surface of the pad 14, as shown in FIG. Note that, instead of the photoresist PR, a conductive film such as Ag or Au may be selectively formed and used as a mask. This coating also functions as an antioxidant film.

  In the structure in which the metal plate 23 is bonded as shown in FIG. 1, the workability is very poor because the metal plate 23 is very thin, about 125 μm. However, if the protruding heat-radiating electrode 15A is formed in this manner, the bonding of the metal plate 23 described above becomes unnecessary.

  Then, as shown in FIG. 14, after the pad 14, the wiring 30, and the external connection electrode 31 are completely separated, the insulating film 26 is covered, and the portion where the solder is arranged is exposed. After the solder 42 is fixed, dicing is performed at a portion indicated by an arrow.

  The separated semiconductor device is mounted on the first support member 11 as shown in FIG. Then, as described above, the second support member 24 is fixed. At this time, since the heat radiation electrode 15A protrudes, it can be easily joined to the fixing plate 25 via solder or the like.

FIG. 14B shows the back surface of the semiconductor element 16 exposed from the insulating resin. For example, if the upper surface of the semiconductor element is brought into contact with the upper mold and molded, a sealing structure as shown in the figure can be realized.

Seventh Embodiment for Explaining Semiconductor Device FIG. 15A is a plan view of a semiconductor device according to the present invention, and FIG. 15B is a cross-sectional view corresponding to line AA of FIG. 15A.

  According to the present invention, the first heat radiation electrode 70A and the second heat radiation electrode 70B are arranged on substantially the same plane, and a pad 14 is provided around the first heat radiation electrode 70A and the second heat radiation electrode 70B. The back surface of the pad 14 serves as an external connection electrode as it is, and is located immediately below the bonding electrode 19 of the semiconductor chip. Also, as shown in FIG. 3, wiring for rearrangement may be adopted. At least one bridge 71 is provided between the chips. Pads 14 are integrally formed on both ends of the bridge 71, and the pads 14 are also connected to the bonding electrodes 19.

  Further, a first semiconductor chip 16A is fixed on the first heat radiation electrode 70A, and a second semiconductor chip 16B is fixed on the second heat radiation electrode 70B and connected via solder. ing.

  As is clear from the above description of the manufacturing method, since the conductive foil is half-etched and molded and supported by the insulating resin 13 before being completely separated, the bridge 71 does not need to be dropped or detached at all. .

  As in the present embodiment, the present invention also makes it possible to package a plurality of chips into one package.

  As described above, in the embodiments described above, the structure of the read / write amplifying IC has been described in consideration of the heat radiation of one IC. However, when various devices are assumed, there may be a case where heat dissipation of a plurality of semiconductor elements must be considered in order to improve the characteristics. Of course, each may be packaged, but a plurality of semiconductor elements may be packaged as shown in FIG.

Naturally, the metal plate may be connected to the heat-dissipating electrode as shown in FIG. 1 or may be configured as shown in FIG. 11 in which the heat-dissipating electrode itself protrudes. These are mounted on a flexible sheet or on a flexible sheet to which a second support member is attached.

FIG. 16 shows a method of connecting the bumps B formed on the semiconductor element 16 and the pads 14, which can be applied to all the embodiments. P indicates a plating film of Au, Ag, etc., and is formed as necessary.

  A is an ACP (anisotropic conductive paste / film) method in which conductive particles are interposed between the bump B and the pad 14 (or the plating film P), and electrical conduction is established by pressing.

  B is an SBB (stand bump bonding) method in which the bump B is connected to the pad 14 (or a plating film), and at the same time, a conductive paste CP is arranged around the bump B.

  C denotes an ESP method (epoxy-encapsulated solder connection) in which the solder SD is melted and arranged around the bump B when the bump B is pressed and fixed.

  D indicates that an insulating bonding means is arranged around the bumps which have been brought into pressure contact by the NCP method (non-conductive paste).

  E is a GGI method (gold-gold interconnection) for bonding the Au bumps and the Au plating film P by ultrasonic waves.

  Finally, F is a solder bump method in which solder bonding is performed, and an insulating adhesive means and an underfill are penetrated therebetween. The present application employs this method.

As described above, there are various connection methods, and the connection method is selected from these in consideration of the connection strength. Such a structure can also be adopted between the back surface of the external connection electrode and the first support member 11.

Eighth Embodiment for Describing Semiconductor Device FIG. 17 shows a semiconductor device according to this embodiment. FIG. 17A is a plan view, and FIG. 17B is a cross-sectional view taken along line AA when the semiconductor device 10E is mounted on the mounting board 43.

  In the above description, the electrode provided on the back surface of the region of the semiconductor element 16 is provided to enhance the heat radiation effect. However, these electrodes not only have a heat radiation effect, but also have a function of buffering the stress acting on the external connection electrodes 21 provided on the periphery of the semiconductor device.

  As shown in FIG. 17, a feature of the semiconductor device 10E is that a stress buffering electrode 21B is provided on the back surface of the heat radiation electrode 15 surrounded by the external connection electrode 21 so as to cover the entire back surface. is there. That is, the solder electrode 21B having a size larger than that of the external connection electrode 21 is provided. This size may be equal to or larger than the semiconductor element. Also, it may be slightly smaller. The effect of this will be described below. The external connection electrode may be a solder, a solder bump, or a conductive adhesive.

  The semiconductor device 10E according to the present invention is supported by the insulating resin 13 as a whole. Since the thermal expansion coefficients of the mounting board 43 and the insulating resin 13 are often different, the difference is made as small as possible. However, it is very difficult to make the same, so that the thermal expansion coefficients of the two are inevitably different. Therefore, when the temperature of both the semiconductor device 10E and the mounting board 43 rises, stress acts on the external connection electrode 21 that connects them. For example, when the semiconductor device is solder-fixed only by the external connection electrodes, the magnitude of the stress increases as the package size increases. Specifically, it is proportional to the distance from the center of the semiconductor element to the periphery of the semiconductor device.

  In the present application, the stress can be relaxed by making the stress buffering electrode 21B substantially the same size as the semiconductor element 16. The region is firmly fixed to the mounting substrate by the stress buffering electrode 21B. Therefore, the stress acting on the external connection electrode 21 is proportional to the distance from the periphery of the stress buffering electrode 21B to the center of the external connection electrode 21. According to the result of the stress simulation, the maximum stress applied to the external connection electrode 21 by about 25 to 30% is reduced. Therefore, the occurrence of cracks in the external connection electrodes 21 can be prevented. The stress buffering electrode 21B may be slightly larger or slightly smaller than the semiconductor element 16.

  FIG. 22 is a graph showing the relationship between the number of cycles (horizontal axis) and the crack occurrence rate (vertical axis) in the heat cycle test. Here, the method of the heat cycle test will be described. First, a semiconductor device mounted on a mounting substrate is exposed to a gas phase via a solder electrode. Next, the temperature of the gas phase is changed, and the number of semiconductor devices having cracks in the solder electrodes due to the temperature change is measured. By performing the above operations, the life of the solder electrode with respect to the temperature change can be evaluated. In addition, the range of the temperature change of the gas phase is −40 ° C. to 125 ° C., and the time of one cycle is about 1 hour.

  The structure of Case 1 and Case 2 in FIG. 22 will be described below.

  Case 1: The size of the heat radiation electrode is the same as that of the pad back surface exposed through the insulating film. Therefore, a large number of solders are provided on the exposed portions of the heat radiation electrodes, and the solder electrodes are mounted on the mounting board as shown in FIG.

  Case 2: In Case 1, the back surface of the heat radiation electrode is exposed over substantially the entire area, and the heat radiation electrode and the electrode of the mounting board are fixed on the entire surface.

  In case 1, the crack generation rate increased from the point in time when the number of cycles exceeded 250, and reached 100% when the number of cycles reached 400. That is, when the number of cycles reaches 400, cracks have occurred in the solder electrodes of all the semiconductor devices.

  In the case 2, the crack generation rate increased from the time when the number of cycles exceeded 450 times, and reached 100% when the number of cycles reached 600 times.

  From this, it can be said that the semiconductor device of Case 2 has a structure in which solder cracks are less likely to occur than the semiconductor device of Case 1. Therefore, it is more effective to use a large-sized solder electrode having substantially the same size as the chip as an electrode for stress buffering. This is because the distance between the solder electrodes is reduced as described above.

Here, the solder electrode is made of a solder material, and the material here may be a generally-known conductive material such as brazing material, Ag, or gold, or a conductive adhesive. The material for fixing the heat radiation electrode and the mounting substrate may be an insulating adhesive as long as the heat radiation electrode 15 is not electrically connected to the chip back surface.

Ninth Embodiment for Describing Semiconductor Device FIG. 18 shows a semiconductor device according to this embodiment. FIG. 18A is a plan view, and FIG. 18B is a cross-sectional view taken along line AA when the semiconductor device 10F is mounted on the mounting board 43.

  As shown in FIGS. 18A and 18B, a feature of this semiconductor device 10F is that a stress relaxation electrode 21B is provided on the heat radiation electrode 15 divided by the groove 44. That is, the back surface of the semiconductor device 10F and the mounting substrate 43 are firmly connected via the stress buffering electrode 21B. Therefore, the stress acting on the external connection electrode 21 is in proportion to the distance from the periphery of the stress buffering electrode 21B to the center of the external connection electrode 21. Accordingly, when the semiconductor device 10F thermally expands, the stress acting on the external connection electrodes 21 can be buffered, and the occurrence of cracks in the external connection electrodes 21 can be prevented.

  Further, by providing the groove 44, the adhesive strength between the insulating adhesive means 17 and the heat radiation electrode 15 can be improved.

Although a cycle test of this structure was not performed, it is considered that the same effect as that of the structure of FIG. 17 occurs. As the number of divisions of the stress buffering electrode 21B increases, the structure shown in FIG. 2 is obtained, but the number of divisions having this effect is at most about 2 to 8 divisions. By performing this division, the amount (thickness) of the applied solder is reduced as compared with the structure where the entire surface is fixed, and the mountability is improved.

Tenth Embodiment for Describing Semiconductor Device FIG. 19 shows a semiconductor device according to this embodiment. FIG. 19A is a plan view, and FIG. 19B is a sectional view taken along line AA when the semiconductor device 10F is mounted on the mounting board 43.

As shown in FIG. 19A, the feature of this semiconductor device is that the external connection electrode 21 has an elongated shape. Again, there are two sources of stress dampening.
First, the bonding area of the external connection electrode 21 has increased.
Secondly, since the crack occurrence location CK occurs on the outer peripheral portion of the solder and on the semiconductor element side, the separation distance of the occurrence location CK is shortened.

  In order to prevent solder bridges between the electrodes, the distance between the heat radiation electrode 15 and the pad 14 needs to be about 0.3 m. However, the shortest length of the thin metal wire 20 necessary for bonding is required to be about 1 mm to 0.5 mm, depending on the thickness of the semiconductor element 16. For example, in the semiconductor element 16 having a thickness of 0.33 mm, the shortest length of the thin metal wire is 1 mm. Further, when the thickness of the semiconductor element is 0.1 mm, a minimum length of about 0.5 mm is required. Therefore, the pad 14 needs to enter the bonding portion of the pad 14 by about 0.2 to 0.4 mm inside. In other words, the crack occurrence location CK needs to be as deep as possible, but it must be as long as possible due to wire bonding restrictions. Moreover, since the number of pads 14 is small in the drawing, a square shape is possible. However, when a pad exceeding 200 pins is required, the width of the bonding pad needs to be reduced. Therefore, the shape of the pad 14 is necessarily elongated.

As a result, the area where the crack is generated is located at the same time as the bonding area is increased, and the generation of the crack is suppressed.

Eleventh Embodiment for Describing Semiconductor Device FIGS. 20 and 21 show a semiconductor device according to this embodiment. FIG. 20A is a plan view, and FIG. 20B is a cross-sectional view taken along line AA when the semiconductor device 10G is mounted on the mounting board 43. 21A to 21C show the structure of the external connection electrode 21. FIG.

  As shown in FIGS. 21A and 21B, a feature of the semiconductor device 10J lies in the structure of the external connection electrode 21. The problem is that, as shown in FIG. 21C, the insulating coating 26 hits the solder extending surface, so that no constriction NK is formed. If the external connection electrode 21 does not have the constriction NK, a structure as shown in FIGS. 21A and 21B may be used.

  It was found that stress was easily concentrated in the constricted NK, and cracks were easily generated. Further, in a thin package having a thickness of 0.5 mm or less, the composition ratio of the insulating resin to the constituent materials of the semiconductor device is small because of the thinness. . Since the thermal expansion coefficient of silicon, which is the material of the semiconductor element 16, and that of the mounting substrate are greatly different, the stress acting on the external connection electrode 21 that connects them is large. Therefore, it is important that the solder extending surface has a smooth curved surface without constriction in all the embodiments of the invention.

  Next, the results of the heat cycle test will be described with reference to FIG. First, the structure of the semiconductor device used in the tests of Cases 3 and 4 will be described.

  The external connection electrodes of both the case 3 and the case 4 of the semiconductor device also have no constriction. The difference between the two lies in the structure of the electrode for stress buffering. A plurality of electrodes having the same size as the external connection electrodes are provided as the stress buffering electrodes of the semiconductor device of the case 3. The size of the stress buffering electrode of the semiconductor device of Case 4 is about the same as that of the semiconductor element.

  In case 3, the test is currently underway, but no solder cracks have occurred even if the number of cycles exceeds 750.

  In case 4, the crack occurrence rate is 0% even when the number of cycles reaches 1400. That is, even if the number of cycles reaches 1400, no crack occurs in the solder electrode.

  Next, differences in the structures of Case 1 to Case 4 will be described.

  The difference between the structures of Case 1 and Case 2 lies in the shape of the stress buffering electrode. The electrode for stress buffering of the case 1 has the same size as the external connection electrode 21. On the other hand, the stress buffering electrode of the case 2 has substantially the same size as the semiconductor element. That is, the case 2 electrode for stress buffering is larger than the case 1 electrode.

  The difference between the structures of Case 1 and Case 3 lies in the shape of the external connection electrode. The external connection electrode of the case 1 has a circular and constricted structure. On the other hand, the external connection electrode of the case 2 is elongated and has no constriction.

  The difference between the structures of Case 2 and Case 4 lies in the shape of the external connection electrodes, although both have stress buffering electrodes of substantially the same size as the semiconductor element. The external connection electrode of the case 2 has a circular and constricted structure. On the other hand, the external connection electrode of the case 4 is elongated and has no constriction.

  This indicates that solder cracks can be prevented by combining the two features of the shape of the stress buffering electrode and the shape of the external connection electrode. In other words, the size of the stress buffering electrode is about the same as the size of the semiconductor element, and the shape of the external connection electrode is elongated and has no constriction.

  As described above, the feature of the present invention is that the electrode which was originally effective as a heat-dissipating electrode is firmly fixed to the mounting board by soldering through solder, thereby preventing the external connection electrode from cracking. It proved to be effective.

  Further, it has been found that the strength of the external connection electrode can be improved by forming the external connection electrode without a constriction.

  As a result, remarkable improvement in the mountability of the thin package is realized, which greatly contributes to the practical use of a light, thin and small package.

  In all of the above-described embodiments, the insulating bonding means 17 was used as the bonding means between the semiconductor element 16 and the heat radiation electrode 15. However, this bonding means is not limited to an insulating one. As long as the back surface electrode of the semiconductor device is not short-circuited, a conductive bonding means may be used.

It is a figure explaining the semiconductor module of the present invention. FIG. 3 illustrates a semiconductor device of the present invention. FIG. 3 illustrates a semiconductor device of the present invention. FIG. 4 is a diagram illustrating a method for manufacturing a semiconductor device according to the present invention. FIG. 4 is a diagram illustrating a method for manufacturing a semiconductor device according to the present invention. FIG. 4 is a diagram illustrating a method for manufacturing a semiconductor device according to the present invention. FIG. 4 is a diagram illustrating a method for manufacturing a semiconductor device according to the present invention. FIG. 4 is a diagram illustrating a method for manufacturing a semiconductor device according to the present invention. FIG. 3 illustrates a semiconductor device of the present invention. FIG. 3 is a diagram illustrating a flow prevention film formed on a conductive pattern. It is a figure explaining the semiconductor module of the present invention. FIG. 4 is a diagram illustrating a method for manufacturing a semiconductor device according to the present invention. FIG. 4 is a diagram illustrating a method for manufacturing a semiconductor device according to the present invention. FIG. 4 is a diagram illustrating a method for manufacturing a semiconductor device according to the present invention. FIG. 3 illustrates a semiconductor device of the present invention. FIG. 4 is a diagram illustrating a connection structure between a semiconductor element and a pad. FIG. 3 illustrates a semiconductor device of the present invention. FIG. 3 illustrates a semiconductor device of the present invention. FIG. 3 illustrates a semiconductor device of the present invention. FIG. 3 illustrates a semiconductor device of the present invention. FIG. 3 illustrates a semiconductor device of the present invention. FIG. 9 is a diagram illustrating experimental results of a heat cycle using the semiconductor device of the present invention. FIG. 3 is a diagram illustrating a hard disk. FIG. 24 is a diagram illustrating a conventional semiconductor module employed in FIG. 23.

Explanation of reference numerals

10 Semiconductor device
11 First support member
12 First opening
14 pads
16 Semiconductor elements
23 Metal plate
24 Second support member
56 Second opening

Claims (5)

A bonding pad provided corresponding to the bonding electrode of the semiconductor element;
An external connection electrode provided on the back surface of the bonding pad,
A pad provided in an arrangement region of the semiconductor element;
An external connection electrode for stress buffering provided on the back surface of the pad,
Bonding means provided on the pad, the semiconductor element fixed to the bonding means, and electrically connected to the bonding pad,
An insulating resin that seals the semiconductor element so as to expose and integrate the back surface of the pad, the back surface of the external connection electrode, and the back surface of the bonding unit,
The semiconductor device, wherein the external connection electrode for stress buffering is formed sufficiently larger than the external connection electrode, and a stress due to thermal expansion of the insulating resin is transmitted to the external connection electrode by buffering.   2. The semiconductor device according to claim 1, wherein the pad and the external connection electrode for stress buffering are divided into four. A bonding pad provided corresponding to the bonding electrode of the semiconductor element;
An external connection electrode provided on the back surface of the bonding pad,
A pad provided in an arrangement region of the semiconductor element;
Bonding means provided on the pad,
The semiconductor element fixed to the bonding means and electrically connected to the bonding pad;
An insulating resin that seals the semiconductor element so as to expose and integrate the back surface of the pad, the back surface of the external connection electrode, and the back surface of the bonding unit,
The semiconductor device, wherein the external connection electrode is formed to be elongated. A bonding pad provided corresponding to the bonding electrode of the semiconductor element;
An external connection electrode provided on the back surface of the bonding pad,
A pad provided in an arrangement region of the semiconductor element;
Bonding means provided on the pad,
The semiconductor element fixed to the bonding means and electrically connected to the bonding pad;
An insulating resin that seals the semiconductor element so as to expose and integrate the back surface of the pad, the back surface of the external connection electrode, and the back surface of the bonding unit,
A semiconductor device, wherein a side surface of the external connection electrode has the same curved surface.   The semiconductor device according to claim 1, wherein the external connection electrode is a solder electrode.

Images