More than one reissue application has been filed for the reissue of U.S. Pat. No.5,777,391. These reissue applications are the present application Ser. No.09/613,541, filed Jul.7,2000; Ser. No.10/105,236, filed on Mar.26,2002; Ser. No.11/182,039 filed on Jul.15,2005; Ser. No.11/182,040, also filed on Jul.15,2005; Ser. No.11/256,620, filed on Oct.24,2005; Ser. No.11/256,621, also filed on Oct.24,2005; Ser. No.11/285,730, filed on Nov.23,2005 and Ser. No.11/285,729, filed on Nov.23,2005.
The present reissue application also claims the benefit under35 USC §120 of the filing date of Dec.11,1995 of Ser. No.08/570,646, now U.S. Pat. No.5,777,391, and benefit under35 USC §119 of Japanese Application No.6-316,444, filed on Dec.20,1994 and Japanese Application No.7-126405, filed May25,1995, the subject matter of which is incorporated herein by reference.
BACKGROUND OF THE INVENTIONThe present invention relates to a semiconductor device and a method of manufacture thereof and more particularly to a technology effectively applied to a semiconductor device and a method of manufacture thereof, the device having a structure in which a semiconductor pellet is mounted on a pellet mounting area on the main surface of a base substrate and in which a first electrode pad on the back of the base substrate is electrically connected to an external terminal on the main surface of the semiconductor pellet.
A semiconductor device with a ball grid array (BGA) structure has been introduced as a semiconductor device having a high level of integration in the Nikkei Electronics. Feb. 28, 1994, pp. 111-117, published by Nikkei McGraw-Hill. The BGA structure of such as semiconductor device, as shown inFIG. 16 (cross section of an essential part), has asemiconductor pellet2 mounted on a pellet mounting area of the main surface of thebase substrate1 and a plurality ofbump electrodes4 arranged in grid on the back of thebase substrate1 opposite the main surface.
Thebase substrate1 may be made from a printed wiring board of two-layer wiring structure.Second electrode pads1A are arranged in a peripheral area of the main surface of the base substrate1 (around the pellet mounting area), whilefirst electrode pads1B are arranged on the back of thebase substrate1 opposite the main surface. Thesecond electrode pads1A are electrically connected to through-hole conductors1C viaconductors1A1arranged on the main surface of thebase substrate1. Thefirst electrode pads1B are electrically connected to the through-hole conductors1C viaconductors1B1arranged on the back of thebase substrate1.
Thesemiconductor pellet2 may comprise mainly a semiconductor substrate2B of single-crystal silicon. On the main surface of the semiconductor substrate2B (device forming surface) is formed a logic circuit system, a memory circuit system or a combination of these. A plurality ofbonding pads2A are arranged on the main surface of the semiconductor substrate2B. Thebonding pads2A are formed in the top of the interconnect layers formed on the main surface of the semiconductor substrate2B.
Thebonding pads2A on thesemiconductor pellet2 are electrically connected to thesecond electrode pads1A on the main surface of thebase substrate1 throughbonding wires6. In other words, thebonding pads2A on thesemiconductor pellet2 are electrically connected to thefirst electrode pads1B through thebonding wires6,second electrode pads1A,conductors1A1, through-hole conductors1C andconductors1B1.
Thesemiconductor pellet2 and thebonding wires6 are sealed with aresin sealing body7 formed on the main surface of thebase substrate1. Theresin sealing body7 is formed by transfer molding.
Thebump electrodes4 are electrically and mechanically connected to the surfaces of thefirst electrode pads1B on thebase substrate1. Thebump electrodes4 may be formed from an alloy material such as Pb-Sn.
The semiconductor device of such a BGA structure is mounted on a mounting board, with thebump electrodes4 electrically and mechanically connected to electrode pads arranged on the mounting surface of the mounting board.
Another example of semiconductor device having a high circuit density is disclosed in U.S. Pat. Ser. No. 5148265, which shows a semiconductor device in which the base substrate is made from a filmlike flexible substrate. In this semiconductor device, the semiconductor pellet is mounted, with its main surface downward, on the pellet mounting area of the main surface of the base substrate made of a flexible substrate, and the bonding pads arranged on the main surface of the semiconductor pellet are electrically connected to the second electrode pads arranged on the back of the base substrate through the bonding wires. The second electrode pads on the base substrate are electrically connected to the first electrode pads on the back of the base substrate through conductors that are also arranged on the back. Bump electrodes are electrically and mechanically connected to the surfaces of the first electrode pads.
The semiconductor device of the above construction is mounted on the mounting surface of a mounting board, with its bump electrodes electrically and mechanically connected to the electrode pads arranged on the mounting surface of the mounting board.
SUMMARY OF THE INVENTIONIn the semiconductor device with the BGA structure, as shown inFIG. 16, thesecond electrode pads1A arranged on the main surface of thebase substrate1 are electrically connected through the through-hole conductors1C to thefirst electrode pads1B arranged on the back of thebase substrate1. The through-hole conductors1C comprises a hole area formed within a through-hole in thebase substrate1 and a land area (fringe portion) formed on the main surface and back surface of thebase substrate1. The inner diameter of the through-hole may be around 0.3 mm and the outer diameter of the land area of the through-hole conductor1C may be about 0.6 mm. The inner diameter of the through-hole and the outer diameter of the land area of the through-hole conductor1C are set large compared to the widths of theconductors1A1electrically connecting thesecond electrode pads1A and the through-hole conductors1C and also compared to the widths of theconductors1B1electrically connecting thefirst electrode pads1B and the through-hole conductors1C.
The circuit systems formed on thetypical semiconductor pellets2 have tended to grow in their level of integration and the number of functions they perform. With enhanced integration and more diversified functions of the circuit system, the number ofbonding pads2A of thesemiconductor pellet2 and the number ofsecond electrode pads1A of thebase substrate1 increase. That is, the number of through-hole conductors1C electrically connecting thesecond electrode pads1A and thefirst electrode pads1B increases as the integration and function of the circuit system are enhanced. Hence, there has been a problem that the external size of thebase substrate1 increase with the increasing number of the through-hole conductors1C, which in turn increases the size of the semiconductor device as a whole.
There is also another problem which the inventors have considered. The intervals between the through-hole conductors formed by copper foil thick film printing, etching or electroplating techniques are greater than the intervals of the bonding pads of the semiconductor pellet formed by photolithography. For this reason, in a semiconductor device with the BGA structure, as the number of the through-hole conductors1C increases, they are positioned outwardly away from thesemiconductor pellet2. This inevitably extends the length of theconductors1A1electrically connecting thesecond electrode pads1A and the through-hole conductors1C and the length of theconductors1B1, electrically connecting thefirst electrode pads1B and the through-hole conductors1C. This, in turn, increases inductance and reduces the operating speed of the semiconductor device.
In the semiconductor device using a flexible substrate for the base substrate, the flexible substrate may for example be formed of a polyester film or polyimide film. This flexible substrate has a small Young's modulus and is soft (low hardness) compared with a rigid substrate impregnated with epoxy resin or polyimide resin, as represented by the FR4 substrate according to the NEMA Standard. Therefore, when the bonding pads arranged on the main surface of the semiconductor pellet are connected with the second electrode pads arranged on the back of the base substrate through bonding wires, the bonding force applied to the second electrode pads is absorbed by the base substrate, preventing the bonding force and ultrasonic vibrations from being transmitted to the second electrode pads effectively. This gives rise to an apprehension that the connection strength between the bonding wires and the second electrode pads may decrease, leading to connection failures of bonding wires and reduced electric reliability of the semiconductor device.
In semiconductor devices that use a flexible substrate for the base substrate, the flexible substrate has a large thermal expansion coefficient in the planar direction and a small Young's modulus (small rigidity), which means that it is easy to bend, compared with the rigid substrate. Therefore, when the semiconductor device is mounted on the mounting surface of the mounting board, the reflow heat used during the process of mounting causes deformations to the base substrate, such as warping and twisting, which in turn reduces the flatness of the back of the base substrate with respect to the mounting surface of the mounting board, thereby lowering the mounting precision of the semiconductor device.
An object of this invention is to provide a technology that allows a reduction in the size of the semiconductor device.
Another object of this invention is to provide a technology that allows an increase in the operating speed of the semiconductor device.
Still another object of this invention is to provide a technology that can enhance electric reliability of the semiconductor device.
A further object of this invention is to provide a technology that can enhance mounting precision of the semiconductor device.
A further object of this invention is to provide a manufacturing technology for the semiconductor device that can accomplish the above objectives.
These and other objects and novel features of this invention will become apparent from the following description of this specification and the accompanying drawings.
Representative aspects of this invention may be briefly summarized as follows.
A semiconductor device in accordance with invention comprises a semiconductor pellet mounted on the pellet mounting area of the main surface of the base substrate, in which first electrode pads arranged on the back of the base substrate are electrically connected to the bonding pads on the main surface of the semiconductor pellet. The base substrate is formed of a rigid substrate, and its first electrode pads are electrically connected to second electrode pads also arranged on the back side of the base substrate. The semiconductor pellet is mounted, with its main surface downward, on the pellet mounting area of the main surface of the base substrate, and its bonding pads are electrically connected to the second electrode pads on the base substrate through bonding wires extending through slits formed in the base substrate.
A method of manufacturing a semiconductor device is also provided, in which a semiconductor pellet is mounted on the pellet mounting area of the main surface of the base substrate and in which first electrode pads arranged on the back of the base substrate are electrically connected to the bonding pads on the main surface of the semiconductor pellet. In particular, the method includes a step of mounting the semiconductor pellet, with its main surface downward, on the pellet mounting area of the main surface of the base substrate made of a rigid substrate, and a step of connecting the bonding pads on the semiconductor pellet to the second electrode pads electrically connected to the first electrode pads of the base substrate and arranged on the back of the base substrate through bonding wires extending through slits formed in the base substrate.
According to the above construction of this invention, the bonding pads of the semiconductor pellet and the first electrode pads of the base substrate can be electrically connected through the bonding wires and the second electrode pads, so it is possible to eliminate the through holes used to electrically connect the second electrode pads and the first electrode pads in prior structures. This allows the external size of the base substrate to be reduced by an amount corresponding to an area occupied by the through holes (land area), thus reducing the size of the semiconductor device as a whole.
Further, because the first electrode pads can be put closer to the second electrode pads by an amount corresponding to an area occupied by the through holes, the conductors of the base substrate that electrically connect the second electrode pads and the first electrode pads can be reduced in length. As a result, inductance can be reduced and the operating speed of the semiconductor device increased.
Further, the rigid substrate has a higher Young's modulus than a flexible substrate; therefore, when the bonding pads arranged on the main surface of the semiconductor pellet and the second electrode pads arranged on the back of the base substrate are electrically connected by bonding wires, the bonding force applied to the second electrode pads can be prevented from being absorbed by the base substrate. This assures effective transmission of the bonding force and the ultrasonic vibrations to the second electrode pads. Thus, the connection strength between the bonding wires and the second electrode pads is increased, making it possible to prevent connection failure of the bonding wires and to enhance electric reliability of the semiconductor device.
Furthermore, the rigid substrate has a small inplane thermal expansion coefficient and a high Young's modulus compared with a flexible substrate, which means the rigid substrate is harder to bend. This prevents the base substrate from being deformed (warped or twisted) due to reflow heat produced during the process of mounting the semiconductor device on the mounting surface of the mounting board. This ensures a sufficient flatness of the back of the base substrate with respect to the mounting surface of the mounting board, thus enhancing the mounting precision of the semiconductor device.
According to the above-mentioned manufacturing method of this invention, the bonding pads of the semiconductor pellet and the first electrode pads of the base substrate are electrically connected through bonding wires and second electrode pads, so the through holes electrically connecting the second electrode pads and the first electrode pads can be eliminated, making it possible to use a base substrate reduced in external size by an amount corresponding to the occupied area of the through holes. This in turn allows the manufacture of reduced-size semiconductor devices.
Further, because the bonding pads of the semiconductor pellet and the first electrode pads of the base substrate are electrically connected through bonding wires and second electrode pads, the through holes electrically connecting the second electrode pads and the first electrode pads can be eliminated, making it possible to use a base substrate whose conductors electrically connecting the second electrode pads and the first electrode pads are reduced by a length corresponding to the occupied area of the through holes. This in turn allows the manufacture of semiconductor devices with faster operating speeds.
The base substrate uses a rigid substrate having a high Young's modulus compared with a flexible substrate; therefore, when the bonding pads arranged on the main surface of the semiconductor pellet and the second electrode pads arranged on the back of the base substrate are electrically connected by bonding wires, the bonding force applied to the second electrode pads can be prevented from being absorbed by the base substrate, ensuring effective transfer of the bonding force and ultrasonic vibrations to the second electrode pads. This enhances the connection strength between the bonding wires and the second electrode pads, allowing the manufacture of the semiconductor device with high electric reliability.
Furthermore, the base substrate used is formed of a rigid substrate having a small planar thermal expansion coefficient and a high Young's modulus compared with those of a flexible substrate, which means the rigid substrate is harder to bend. As a result, the rigid base substrate is free from deformations (warping or twisting) due to reflow heat during the process of mounting the semiconductor device on the mounting surface of the mounting board. As a result, a sufficient degree of flatness of the back of the base substrate with respect to the mounting surface of the mounting board can be secured, which in turn allows the manufacture of semiconductor devices with high mounting precision.
BRIEF DESCRIPTION OF THE DRAWINGSFIG. 1 is a plan view of the main surface side of the semiconductor device, as a first embodiment of this invention, that employs a BGA structure;
FIG. 2 is a cross section taken along the line A—A ofFIG. 1;
FIG. 3 is an enlarged cross section of an essential part ofFIG. 2;
FIG. 4 is an enlarged plan view showing the state of the back side of an essential part of the semiconductor device with the resin sealing body removed;
FIG. 5 is a cross section snowing an essential part of a molding die for the resin sealing body of the semiconductor device;
FIG. 6 is a cross section showing the method of manufacturing the semiconductor device;
FIG. 7 is a cross section of an essential part of the semiconductor device showing the method of manufacture thereof;
FIG. 8 is a cross section of an essential part of the semiconductor device showing the method of manufacture thereof;
FIG. 9 is a cross section of an essential part of the semiconductor device showing the method of manufacture thereof;FIG. 10 is a cross section showing an essential part of the semiconductor device mounted on a mounting board;
FIG. 11 is a cross section showing a variation of the semiconductor device;
FIG. 12 is a cross section of the semiconductor device, as a second embodiment of this invention, that employs the BGA structure;
FIG. 13 is an enlarged plan view showing the state of the back side of an essential part of the semiconductor device with the resin sealing body removed;
FIG. 14 is a plan view showing the state of the back side of an essential part of the semiconductor device, as a third embodiment of this invention, that employs the BGA structure with the resin sealing body removed;
FIG. 15 is a plan view showing the state of the back side of an essential part of the semiconductor device, as a fourth embodiment of this invention, that employs the BGA structure and is removed of the resin sealing body removed; and
FIG. 16 is a cross section showing an essential part of the semiconductor device that employs the conventional BGA structure.
DESCRIPTION OF THE PREFERRED EMBODIMENTSThe construction of this invention is described in the following in conjunction with embodiments that apply this invention to a semiconductor device using the BGA structure.
In the drawings used for explaining the embodiments, components with identical functions are given like reference numerals and their explanations are not repeated.
Embodiment 1The outline construction of a semiconductor device, as a first embodiment of this invention, that uses the BGA structure is shown inFIG. 1 (plan view of the main surface side),FIG. 2 (cross section taken along the line A—A of FIG.1),FIG. 3 (enlarged cross section of an essential part ofFIG. 2) andFIG. 4 (enlarged plan view showing the back side of an essential part of the semiconductor device with the resin sealing body removed).
As shown inFIGS. 1,2,3 and4, the semiconductor device has asemiconductor pellet2 mounted on a pellet mounting area of the main surface of abase substrate1, with a plurality ofbump electrodes4 arranged in grid on the back of thebase substrate1 opposite the main surface.
Thebase substrate1 may be formed of a printed circuit board. The printed circuit board may, for example, have a structure in which wiring is formed over the surface of a rigid substrate of glass fiber impregnated with epoxy resin, polyimide resin or maleimide resin. In other words, thebase substrate1 is formed of a rigid substrate. The rigid substrate has a high Young's modulus and is hard compared with a flexible substrate made of polyester film or polyimide film. The rigid substrate has a small thermal expansion coefficient in a planar direction, a high Young's modulus and is difficult to bend compared with the flexible substrate. For example, the rigid substrate made of a glass fiber impregnated with epoxy resin or polyimide resin has a Young's modulus of around 16-22 GPa and a thermal expansion coefficient of about 10-20×10−61/° C. Flexible substrates made of polyester film or polyimide film have a Young's modulus of about 2-5 GPa and a thermal expansion coefficient of about 20-25×10−61/° C.
On the back of thebase substrate1 are formed a plurality ofsecond electrode pads1A andfirst electrode pads1B, which are electrically interconnected throughconductors1B1on the back of thebase substrate1. Thesecond electrode pads1A,first electrode pads1B andconductors1B1are formed of a Cu film, for example.
On the surfaces of thefirst electrode pads1B are formedbump electrodes4 that are electrically and mechanically connected to them. Thebump electrodes4 may be formed of, for instance, a Pb-Sn alloy.
Thesemiconductor pellet2 is mounted, with its main surface (underside inFIG. 2 and 3) downward, on the pellet mounting area of the main surface of thebase substrate1. That is, thesemiconductor pellet2 is mounted facedown on the pellet mounting area of the main surface of thebase substrate1. Interposed between the main surface of thesemiconductor pellet2 and the pellet mounting area of the main surface of thebase substrate1 is aninsulating layer3, which may be formed of a polyimide-. epoxy- or silicon-base low-elasticity resin.
Thesemiconductor pellet2 may be rectangular and may mainly be comprised of a semiconductor substrate2B made of single-crystal silicon. On the main surface (device forming surface) of the semiconductor substrate2B are formed a logic circuit system, a memory circuit system or a combination of these. Also on the main surface of the semiconductor substrate2B, a plurality ofbonding pads2A are arranged along the sides of the rectangular surface. Thebonding pads2A are formed on the top of interconnect layers on the main surface of the semiconductor substrate2B. That is, thebonding pads2A are arranged in the periphery of the main surface of thesemiconductor pellet2 along each of the four sides.
Thebonding pads2A of thesemiconductor pellet2 and thesecond electrode pads1A of thebase substrate1 are electrically connected to each other throughbonding wires6 running inslits5 formed in thebase substrate1. Thebonding wires6 may be of gold (AU), copper (Cu) or aluminum (Al), and may be coated with insulating resin. Thebonding wires6 may be connected by a bonding method that utilizes ultrasonic vibrations in combination with thermocompression.
Theslits5 in thebase substrate1 are formed in the directions of the rows of thebonding pads2A that are arranged along each side of the main surface of thesemiconductor pellet2. That is, thebase substrate1 of this embodiment has fourslits5, each of which is located above thebonding pads2A of thesemiconductor pellet2.
Thesecond electrode pads1A of thebase substrate1 are placed in both areas of the back of thebase substrate1 divided by theslits5. Thesecond electrode pads1A located in one of the areas of the back of thebase substrate1 demarcated by the slits5 (inside the semiconductor pellet2) are supplied with a power supply such as an operation voltage (3.3 V for instance) and a reference voltage (0 V for instance). Thesecond electrode pads1A located in the other area of the back of thebase substrate1 demarcated by the slits5 (outside the semiconductor pellet2) receive a signal such as an input/output signal and a control signal.
Thesemiconductor pellet2 are provided with 100bonding pads2A on each side at a pitch of about 100 μm. The number ofbonding pads2A is increased as the level of integration and the operating speed of the circuit system mounted on thesemiconductor pellet2 increase.
The first area of the back of thebase substrate1 demarcated by theslits5 is provided with, for example, 50second electrode pads1A for each side of thesemiconductor pellet2; and the second area is provided with, for instance, 50second electrode pads1A for each side of thesemiconductor pellet2. Because thesecond electrode pads1A cannot be made as small as thebonding pads2A of thesemiconductor pellet2, the pitch of thesecond electrode pads1A is set wider than that of thebonding pads2A, for instance, at around 200 μm. That is, because thesecond electrode pads1A of thebase substrate1 are arranged in two rows for each side of thesemiconductor pellet2, the length of thesecond electrode pads1A corresponding to one side of thesemiconductor pellet2 can be made almost equal to that of thebonding pads2A arranged along one side of thesemiconductor pellet2 even if the pitch of thesecond electrode pads1A of thebase substrate1 is set to two times that of thebonding pads2A of thesemiconductor pellet2. Furthermore, thesecond electrode pads1A of thebase substrate1 can be located at positions facing thecorresponding bonding pads2A of thesemiconductor pellet2.
The peripheral area of the main surface of thebase substrate1 excluding the pellet mounting area is covered with aresin sealing body7, which seals thebonding wires6. That is, theresin sealing body7 is formed on the main surface side and the back surface side of thebase substrate1. Theresin sealing body7 is made fromepoxy resin7A containing a phenol-base hardener, silicone rubber and filler for reducing stresses.
The back of thebase substrate1 facing the main surface of thesemiconductor pellet2 is exposed from theresin sealing body7 that covers the peripheral area of thebase substrate1.
Theresin sealing body7 is formed by the transfer molding that uses amolding die10 shown inFIG. 5 (cross section of an essential part). The molding die10 has acavity11 defined by anupper die10A and alower die10B, aninflow gate13 connected to thecavity11, and, though not shown, a pot and a runner. The pot communicates with thecavity11 through the runner and theinflow gate13.
Thecavity11 comprises arecess11A formed in theupper die10A and arecess11B formed in thelower die10B. Theresin7A is supplied into therecess11A from the pot through the runner and theinflow gate13. Thebase substrate1 is placed in therecess11B.
Therecess11B is formed withrecesses12, which are located at positions facing theslits5 of thebase substrate1 and which extend in the same directions as theslits5. Placed in therecesses12 are a part ofbonding wires6 electrically connecting thebonding pads2A of thesemiconductor pellet2 and thesecond electrode pads1A of thebase substrate1, and also thesecond electrode pads1A of thebase substrate1. Theresin7A is supplied from therecess11A through theslits5 of thebase substrate1 into therecess11A.
Though not shown inFIG. 12, therecesses12 are provided with a gas vent to prevent voids due to bubbles.
Next, the method of manufacturing the above-mentioned semiconductor device is described by referring to FIGS.6 through FIG.9.
First, abase substrate1 made of a rigid substrate is prepared. Thebase substrate1 includesslits5 as well assecond electrode pads1A,first electrode pads1B andconductors1B1on its back.
Next, as shown inFIG. 6 (cross section), thesemiconductor pellet2 is mounted on the pellet mounting area of the main surface of thebase substrate1. Thesemiconductor pellet2 is fixed to the pellet mounting area of the main surface of thebase substrate1 through an insulatinglayer3.
Next, thebase substrate1 is mounted on a bonding stage (heat block)14 with thesemiconductor pellet2 at the bottom. Thebonding stage14 has arecess14A that accommodates thesemiconductor pellet2. Thebase substrate1 and thesemiconductor pellet2 are heated to about 200° C. on thebonding stage14.
Next, as shown inFIG. 7 (cross section of an essential part), thebonding pads2A arranged on the main surface of thesemiconductor pellet2 and thesecond electrode pads1A arranged on the back of thebase substrate1 are electrically connected by thebonding wires6. Thebonding wires6 running in theslits5 are connected to thebonding pads2A of thesemiconductor pellet2 and to thesecond electrode pads1A of thebase substrate1. The connection of thebonding wires6 is accomplished by ultrasonic thermocompression bonding. In this process, thebase substrate1 is made from a rigid substrate with a high Young's modulus compared with the flexible substrate used in conventional structure, so that the bonding force applied to thesecond electrode pads1A is prevented from being absorbed by therigid base substrate1, thus allowing the bonding force and the ultrasonic vibrations to be transferred effectively to thesecond electrode pads1A. Further, because thebase substrate1 is made of a rigid substrate that has a smaller thermal expansion coefficient in the planar direction than that of a flexible substrate and a higher Young's modulus—which means it is harder to bend—it is possible to reduce positional deviations of thesecond electrode pads1A and of thebonding pads2A of thesemiconductor pellet2 due to thermal expansion of thebase substrate1.
Then, as shown inFIG. 8 (cross section of an essential part), thebase substrate1 and thesemiconductor pellet2 are put in thecavity11 defined by theupper die10A and thelower die10B of the molding die10, with thebase substrate1 fit in therecess11B of thecavity11. A part of thebonding wires6 and thesecond electrode pads1A of thebase substrate1 are placed in therecesses12 formed in therecess11B. The molding die10 is preheated to around 170°-180° C. to heighten the fluidity of theresin7A supplied into thecavity11. Because thebase substrate1 is made from a rigid substrate with a smaller thermal expansion coefficient in the planar direction than the flexible substrate and with a higher Young's modulus, which means thebase substrate1 is harder to bend, thebase substrate1 can be prevented from being deformed (warped or twisted) due to the heating of the molding die10 to about 170°-180° C. during this process.
Next, resin tablets are charged into the pot of the molding die10, nothing that they are preheated by a heater to lower the viscosity before being charged. The resin tablets in the pot are heated by the molding die10, further lowering the viscosity.
The resin is then pressurized by a plunger of the transfer molding device, forcing theresin7A from the pot through the runner and thegate13 into therecess11A and therecesses12 of thecavity11 to cover the peripheral area of the main surface of thebase substrate1, leaving the back of thesemiconductor pellet2 exposed. In this way, aresin sealing body7 that seals thebonding wires6 is formed. Theresin7A is forced into therecesses12 through theslits5 of thebase substrate1 from therecess11A. In this process, theresin7A supplied from therecess11A to therecesses12 through theslits5 flows in the axial direction of thebonding wires6, i.e., in the vertical direction, from one end side of thebonding wires6. This vertical flow of resin prevents thebonding wires6 from being deformed whereas the horizontal flow along the surface of thebase substrate1 may deform them.
Then, thebase substrate1 is taken out of the molding die10, and bumpelectrodes4 are electrically and mechanically connected to the surfaces of thefirst electrode pads1B on the back of thebase substrate1. Thus, a nearly completed semiconductor device shown inFIG. 1,2,3 and4 is obtained.
After this, the semiconductor device is shipped as a product. The semiconductor device shipped as a product is mounted on a mounting surface of a mountingboard15, with thebump electrodes4 of the semiconductor device electrically and mechanically connected toelectrode pads15A arranged on the mounting surface of the mountingboard15, as shown inFIG. 10 (cross section). The connection between thebump electrodes4 of the semiconductor device and theelectrode pads15A of the mountingboard15, although it depends on the material of thebump electrodes4, may be accomplished in an atmosphere at a reflow temperature of, for instance, around 210°-230° C. In this mounting process, because thebase substrate1 is made from a rigid substrate which has a smaller thermal expansion coefficient in the planar direction and a higher Young's modulus—which means it is more difficult to bend—than a flexible substrate, thebase substrate1 can be prevented from being deformed due to reflow heat.
This embodiment offers the following advantages.
A semiconductor device comprises asemiconductor pellet2 mounted on a pellet mounting area of the main surface of abase substrate1, in whichfirst electrode pads1B arranged on the back of thebase substrate1 are electrically connected tobonding pads2A arranged on the main surface of thesemiconductor pellet2. Thebase substrate1 is formed of a rigid substrate, and itsfirst electrode pads1B are electrically connected to thesecond electrode pads1A arranged on its reverse side. Thesemiconductor pellet2 is mounted on the pellet mounting area of the main surface of thebase substrate1, with its main surface downward, and itsbonding pads2A are electrically connected with thesecond electrode pads1A of thebase substrate1 throughbonding wires6 passing throughslits5 formed in thebase substrate1. Because with this construction thebonding pads2A of thesemiconductor pellet2 and thefirst electrode pads1B of thebase substrate1 can be electrically connected through thebonding wires6 andsecond electrode pads1A, it is possible to eliminate the through holes used to electrically connect thesecond electrode pads1A and thefirst electrode pads1B. This in turn allows thebase substrate1 to be reduced in size by an amount corresponding to the occupied area of the through holes (land area), which contributes to size reduction of the semiconductor device.
Because thefirst electrode pads1B can be put closer to thesecond electrode pads1A by a distance corresponding to the occupied area of the through holes, it is possible to shorten the length of theconductors1B, of thebase substrate1 that electrically connect thesecond electrode pads1A and thefirst electrode pads1B. This reduces the inductance, increasing the operation speed of the semiconductor device.
Further, because the rigid substrate has a higher Young's modulus and is harder than the flexible substrate of the conventional structure, the bonding force applied to thesecond electrode pads1A is not absorbed by thebase substrate1 when electrically connecting thebonding pads2A on the main surface of thesemiconductor pellet2 and thesecond electrode pads1A on the back of thebase substrate1 by thebonding wires6. As a result, the bonding force and the ultrasonic vibrations are effectively transferred to thesecond electrode pads1A. This in turn increases the connection strength between thebonding wires6 and thesecond electrode pads1A, preventing possible connection failures of thebonding wires6, enhancing the electric reliability of the semiconductor device.
Moreover, because the rigid substrate has a smaller thermal expansion coefficient in the planar direction and a higher Young's modulus than a flexible substrate, which means it is more resistant to bending, thebase substrate1 is free from deformations (warping and twisting) due to reflow heat when the semiconductor device is mounted on the mounting surface of the mountingboard15. As a result, a sufficient degree of flatness of the back of thebase substrate1 with respect to the mounting surface of the mountingboard15 can be secured, enhancing the mounting precision of the semiconductor device.
Further, because the rigid substrate has a smaller thermal expansion coefficient in the planar direction and a higher Young's modulus than the flexible substrate, which means it is more resistant to bending, the warping of thebase substrate1 can be limited to less than 100 μm even when the external size of thebase substrate1 increases with the increasing number of thefirst electrode pads1B.
With the warping of thebase substrate1 limited to within 100 μm, it is possible to eliminate a reinforcement substrate intended to prevent warping of thebase substrate1. This reduces the manufacture cost of the semiconductor device compared with that of a semiconductor device having a reinforcement substrate.
Furthermore, because thebase substrate1 can be formed of a printed wiring board of a single layer structure having thesecond electrode pads1A,first electrode pads1B andconductors1B1arranged only on the back of a rigid substrate, the parts cost of thebase substrate1 can be reduced compared with that of a base substrate formed of a two-layer printed wiring board which has circuits formed on both the main and back surfaces of the rigid substrate. This means that the overall cost of semiconductor device manufacture can be lowered.
Another feature of this embodiment is that theslits5 formed in thebase substrate1 extend in the directions of rows ofbonding pads2A arranged on the main surface of thesemiconductor pellet2 and are located at positions over thebonding pads2A. With this construction, theslits5 are arranged within the area occupied by thesemiconductor pellet2, so that thebase substrate1 requires no increase in size corresponding to theslits5.
A further feature of this embodiment is that thesecond electrode pads1A are arranged in two opposite areas of the back of thebase substrate1 divided by theslits5. This construction allows an increase in the number of power supply paths for electrically connecting thebonding pads2A of thesemiconductor pellet2 and thesecond electrode pads1A of thebase substrate1. This in turn makes it possible to reduce power supply noise generated at time of simultaneous switching of signals, thereby preventing malfunctions of the semiconductor device.
Further, even when the pitch of thesecond electrode pads1A of thebase substrate1 is set larger than that of thebonding pads2A of thesemiconductor pellet2, the length of the row of thesecond electrode pads1A for each side of thesemiconductor pellet2 can be made almost equal to the length of the row of thebonding pads2A for each side of thesemiconductor pellet2. This prevents an increase in the length of thebonding wires6, which is dependent on the length of the row of thesecond electrode pads1A. As a result, it is possible to prevent thebonding wires6 from being deformed by the flow of resin when thebonding wire6 are sealed by theresin sealing body7 according to the transfer molding.
Further, because thesecond electrode pads1A can be located at positions on thebase substrate1 facing thebonding pads2A of thesemiconductor pellet2, the lengths of thebonding wires6 can be made uniform, which in turn makes uniform the inductances of the signal paths between thebonding pads2A of thesemiconductor pellet2 and thesecond electrode pads1A of thebase substrate1.
A further feature of this embodiment is the structure in which the back of thesemiconductor pellet2 opposing its main surface is exposed from theresin sealing body7 that covers the peripheral area around the main surface of thebase substrate1. This structure allows the heat generated by the operation of the circuit system mounted on thesemiconductor pellet2 to be released from the back of thesemiconductor pellet2, thus enhancing the heat dissipation efficiency of the semiconductor device.
Further, because the mechanical strength of thebase substrate1 can be reinforced by the mechanical strength of theresin sealing body7, deformations of the base substrate1 (warping and twisting) due to reflow heat during mounting can be prevented.
A further feature of this embodiment is that thebonding wires6 are sealed with theresin sealing body7. This structure prevents thebonding wires6 from being deformed due to external impacts and contacts, thus enhancing the electric reliability of the semiconductor device.
A still further feature of this embodiment is that theresin sealing body7 is formed both on the main surface side and the back surface side of thebase substrate1. This structure prevents theresin sealing body7 from becoming separated from thebase substrate1 due to the thermal stresses generated during a temperature cycle test or when thebump electrodes4 are connected. This in turn enhances the reliability of the semiconductor device.
A method of manufacturing a semiconductor device, in which asemiconductor pellet2 is mounted on a pellet mounting area of the main surface of abase substrate1 and in whichfirst electrode pads1B arranged on the back of thebase substrate1 are electrically connected tobonding pads2A arranged on the main surface of thesemiconductor pellet2, comprises a step of mounting thesemiconductor pellet2, with its main surface downward, on the pellet mounting area of the main surface of thebase substrate1 formed of a rigid substrate, and a step of electrically connecting thebonding pads2A to thesecond electrode pads1A, which are electrically connected to thefirst electrode pads1B of thebase substrate1 and arranged on the back of thebase substrate1, throughbonding wires6 passing throughslits5 formed in thebase substrate1. Thebonding pads2A of thesemiconductor pellet2 and thefirst electrode pads1B of thebase substrate1 therefore are electrically connected through thebonding wires6 and thesecond electrode pads1A, so that throughholes1C used for electrically connecting thesecond electrode pads1A and thefirst electrode pads1B can be eliminated, reducing the external size of thebase substrate1 by an amount corresponding to the occupied area of the through holes. As a result, the overall external size of the semiconductor device can be reduced.
Further, because thebonding pads2A of thesemiconductor pellet2 and thefirst electrode pads1B of thebase substrate1 are electrically connected through thebonding wires6 and thesecond electrode pads1A, there is no need for throughholes1C to electrically connect thesecond electrode pads1A with thefirst electrode pads1B. This makes it possible to use abase substrate1 in which theconductors1B1electrically connecting thesecond electrode pads1A and thefirst electrode pads1B are shorter by a length corresponding to the occupied area of the through holes. As a result it is possible to fabricate a semiconductor device with fast operating speeds.
Because thebase substrate1 used is formed of a rigid substrate having a higher Young's modulus—which means it is harder—than a flexible substrate, the bonding force applied to thebonding pads2A when electrically connecting thebonding pads2A arranged on the main surface of thesemiconductor pellet2 and thesecond electrode pads1A arranged on the back of thebase substrate1 through thebonding wires6 is not absorbed by thebase substrate1, effectively transmitting the bonding force and ultrasonic vibrations to thesecond electrode pads1A. As a result, the connection strength between thebonding wires6 and thesecond electrode pads1A can be increased, which in turn allows the manufacture of a semiconductor device with high electric reliability.
Because thebase substrate1 is formed of a rigid substrate having a smaller thermal expansion coefficient in the planar direction and a higher Young's modulus—which means it is more resistant to bending—than a flexible substrate, thebase substrate1 is prevented from being deformed (warped or twisted) due to reflow heat during the process of mounting the semiconductor device on the mounting surface of the mountingboard15. This allows the back surface of thebase substrate1 to have a sufficient degree of Harness with respect to the mounting surface of the mountingboard15, thus enhancing the mounting precision of the semiconductor device.
Following the process of electrically connecting with thebonding wires6, the method of manufacture includes a process of transfer molding of aresin sealing body7 that covers the peripheral area of the main surface of thebase substrate1 and seals thebonding wires6. Because thebase substrate1 uses a rigid substrate which has a smaller thermal expansion coefficient in the planar direction and a higher Young's modulus and is more resistant to beading than a flexible substrate, this method prevents thebase substrate1 from being deformed (warped or twisted) due to heating of the molding die10.
Because theresin7A supplied from therecess11A into therecesses12 through theslits5 flows from one end side of thebonding wires6 in their axial direction, i.e., in the vertical direction, thebonding wires6 are not deformed by the flow of theresin7A, whereas they can be deformed when the resin flows along the surface of thebase substrate1, i.e., in the lateral direction.
As shown inFIG. 11 (cross section), theresin sealing body7 may be formed on the back surface of thebase substrate1 excluding the surfaces of thesecond electrode pads1A andfirst electrode pads1B. In this case, thebase substrate1 is held and clamped from both sides by theresin sealing body7 and therefore prevented from being warped.
Thebase substrate1 may, though not shown, be formed in a multilayer structure in which a plurality of rigid substrates are stacked together. This structure can reduce the manufacture cost as compared with a base substrate made up of a plurality of flexible substrates stacked together.
Embodiment 2The outline configuration of a semiconductor device as the second embodiment of this invention that employs a BGA structure is shown inFIG. 12 (cross section) andFIG. 13 (enlarged plan view of an essential part of the back side showing the state of the back side removed of the resin sealing body).
As shown inFIG. 12 and 13, the semiconductor device has thesemiconductor pellet2 mounted facedown on the pellet mounting area of the main surface of thebase substrate1 with an insulatinglayer3 in between. A plurality ofbump electrodes4 are arranged in grid on the back of thebase substrate1.
Arranged in the central area of the main surface of thesemiconductor pellet2 along the longer sides thereof is a row ofbonding pads2A, which are electrically connected to thesecond electrode pads1A arranged on the back of thebase substrate1 through thebonding wires6 passing through theslits5 formed in thebase substrate1. Thesecond electrode pads1A are electrically connected to the correspondingfirst electrode pads1B arranged on the back of thebase substrate1 throughconductors1B1.Bump electrodes4 are electrically and mechanically connected to the surfaces of thefirst electrode pads1B. That is, thebonding pads2A of thesemiconductor pellet2 are electrically connected to thefirst electrode pads1B through thebonding wires6,second electrode pads1A andconductors1B1.
Theslits5 of thebase substrate1 are formed in the central area of the main surface of thesemiconductor pellet2 along the direction of the row of thebonding pads2A arranged along the longer side of thesemiconductor pellet2. Theslits5 are tapered so that its opening on the back side of thebase substrate1 is greater than the opening on the main surface side.
As described above, this embodiment offers similar effects and advantages to those of the first embodiment. With theslits5 tapered, it is possible to prevent contact between thebase substrate1 and a bonding tool when one end of thebonding wires6 is bonded to thebonding pads2A of thesemiconductor pellet2. This in turn raises the yield of semiconductor device assembly in the bonding process.
Embodiment 3The outline configuration of a semiconductor device as the third embodiment of this invention that employs a BGA structure is shown inFIG. 14 (plan view of an essential part of the back side showing the state of the back side removed of the resin sealing body).
As shown inFIG. 14, the semiconductor device has asemiconductor pellet2 mounted facedown on a pellet mounting area of the main surface of thebase substrate1, with an insulatinglayer3 in between.Bump electrodes4 are arranged in grid on the back of thebase substrate1.
At the outer periphery of the main surface of thesemiconductor pellet2, a plurality ofbonding pads2A are arranged along the sides of the pellet. At the central portion of the main surface of thesemiconductor pellet2, a plurality ofbonding pads2A are arranged along the longer or shorter side of the pellet. Thebonding pads2A are electrically connected to thesecond electrode pads1A arranged on the back of thebase substrate1 bybonding wires6 passing throughslits5 formed in thebase substrate1. Thesecond electrode pads1A are electrically connected tofirst electrode pads1B arranged on the back of thebase substrate1 throughconductors1B1.Bump electrodes4 are electrically and mechanically connected to the surfaces of the individualfirst electrode pads1B. That is, thebonding pads2A are electrically connected to thefirst electrode pads1B through thebonding wires6,second electrode pads1A andconductors1B1.
Theslits5 are arranged at each sides of thesemiconductor pellet2 and also at the central portion of the pellet. That is, thebase substrate1 of this embodiment has fiveslits5, each of which is located above thebonding pads2A of thesemiconductor pellet2.
As explained above, this embodiment offers the similar effects and advantages to those of the first embodiment. Because theslits5 are arranged at the sides and the central portion of thesemiconductor pellet2, it is possible to increase the number ofbonding pads2A arranged on the main surface of thesemiconductor pellet2 and the number ofsecond electrode pads1A arranged on the back of thebase substrate1. This allows an increase in the number of power supply paths for electrically connecting thebonding pads2A of thesemiconductor pellet2 and thesecond electrode pads1A of thebase substrate1. This is turn allows a further reduction in power supply noise generated when output signals are switched simultaneously. Furthermore, this construction makes it possible to increase the number of signal paths electrically connecting thebonding pads2A of thesemiconductor pellet2 and thesecond electrode pads1A of thobase substrate1 and therefore reduce the external size of thesemiconductor pellet2 dictated by the number ofbonding pads2A.
Although this embodiment has been shown to have only oneslit5 formed at the central portion of thesemiconductor pellet2, two ormore slits5 may be arranged parallelly or crosswise to each other at the central part of thesemiconductor pellet2. By increasing the number ofslits5 in this way, it is possible to further increase the number of thesecond electrode pads1A of thebase substrate1 and the number of thebonding pads2A of thesemiconductor pellet2.
Embodiment 4The outline configuration of a semiconductor device as the fourth embodiment of this invention that employs a BGA structure is shown inFIG. 15 (plan view of an essential part of the back side showing the state of the back side removed of the resin sealing body).
As shown inFIG. 15, the semiconductor device has asemiconductor pellet2 mounted facedown on a pellet mounting area of the main surface of thebase substrate1, with an insulatinglayer3 in between.Bump electrodes4 are arranged in grid on the back of thebase substrate1. Thebase substrate1 is formed of a printed circuit board of, for example, 3-layer wiring structure.
At the outer periphery of the main surface of thesemiconductor pellet2, a plurality ofbonding pads2A are arranged along the sides of the pellet. Thebonding pads2A are electrically connected to thesecond electrode pads1A arranged on the back of thebase substrate1 throughbonding wires6 passing throughslits5 formed in thebase substrate1.
Of thesecond electrode pads1A,electrode pads1A2are formed integral withelectrode plates8A. Theelectrode plates8A are electrically connected toother electrode plates8A via through holes (not shown) and internal wiring (not shown) in thebase substrate1. Theelectrode plates8A is connected to be at a reference voltage (0 V for example). Of thesecond electrode pads1A,electrode pads1A3are formed integral with anelectrode plate8B. Thiselectrode plate8A is connected to be at an operating, voltage (3.3 V for instance).
With this embodiment, because the throughholes1C that electrically connect thesecond electrode pads1A on the main surface of thebase substrate1 and thefirst electrode pads1B on the back are eliminated, theelectrode plates8A and theelectrode plate8B can be arranged on the back of thebase substrate1. This allows thebump electrodes4 to be freely located and shortens the distance between thebonding pads2A of thesemiconductor pellet2 and thepump electrodes4. As a result, the inductance can be reduced, thereby increasing the operating speeds of the semiconductor device.
The invention has been described in detail in connection with representative embodiments of the invention. It is noted, however, that the invention is not limited to these embodiments but that many modifications may be made without departing from the spirit of the invention.
Representative advantages of this invention may be summarized as follows.
It is possible to reduce the size of a semiconductor device in which the semiconductor pellet is mounted on the pellet mounting area of the main surface of the base substrate and in which the first electrode pads arranged on the back of the base substrate are electrically connected to the bonding pads arranged on the main surface of the semiconductor pellet.
It is possible to increase the operating speed of the semiconductor device.
It is also possible to enhance the electric reliability of the semiconductor device.
Further, it is possible to increase the mounting precision of the semiconductor device.