JP2012049207A - Method of forming bump - Google Patents

Abstract

PROBLEM TO BE SOLVED: To efficiently form a bump in each electrode pad without temporarily adhering a conductive ball using flux, when forming the bumps in the plurality of electrode pads formed on one side of a substrate in a prescribed pattern respectively using the conductive balls.SOLUTION: This method uses a fixture 1 in which a plurality of hollow parts 11 for housing the conductive balls Bs respectively are formed correspondingly to the pattern, and includes a process for mounting the conductive ball to each of the hollow parts of the fixture, a process for disposing the substrate on the fixture by vertically aligning each hollow part with each electrode pad, and a process for at least making each conductive ball closely adhere to each electrode with vertical position relation between the fixture and the substrate held, and heating and fusing these conductive balls to form the bumps. The fixture is made up from a material which is not wetted when the conductive ball is fused while having a linear thermal expansion coefficient equal to that of the substrate.

Description

  The present invention relates to a bump forming method in which bumps are respectively formed on a plurality of electrode pads formed in a predetermined pattern on one side of a substrate using a conductive ball.

  As a technique for realizing high-density integration and high-density mounting of a semiconductor device, for example, there is a flip chip mounting technique. In flip chip mounting, bump bump electrodes are formed on electrode pads which are electrode terminals formed on the surface (circuit surface) of a semiconductor chip, and are directly electrically connected to a printed circuit board or the like with the circuit surface facing downward. It has been conventionally known that a conductive ball made of metal such as gold or solder is used for forming the bump.

  In the above method, first, a conductive ball holding flux is applied onto a plurality of electrode pads formed in a predetermined pattern on one side of a substrate, and a conductive ball is placed thereon and temporarily attached. Next, bumps are formed by electrically joining the conductive balls to the electrodes by reflow (heating and melting of the conductive balls) (see, for example, the description of Patent Document 1). For the application of the flux, a printing method using a metal mask or a screen mask is generally used. On the other hand, for mounting conductive balls on electrodes, one using an arrangement jig (mask) in which a plurality of through holes (openings) are formed corresponding to each electrode is known (Patent Document 2). See description). That is, the alignment jig is positioned on the substrate so that each through hole faces each electrode pad, and the alignment jig is adsorbed from a region including the outer peripheral edge of the substrate, and in this state, each through hole of the alignment jig is disposed. A conductive ball is inserted into (installed). The array jig is detached from the substrate prior to reflow after the conductive balls are loaded.

  By the way, when using a flux for temporary attachment of the conductive ball to the electrode pad, a cleaning process is required after the reflow to remove the flux residue using, for example, an organic solvent. For this reason, there are many manufacturing processes, and it is difficult to achieve cost reduction. Also, when the pitch at which bumps are to be formed is narrow, cleaning is likely to be insufficient, and if the flux residue remains, problems such as corrosion of the electrode pad due to chlorine contained in the flux occur. There is also a fear.

JP 2006-66413 A JP 2006-303103 A

  In view of the above, the present invention has an object to provide a low-cost bump forming method capable of efficiently forming bumps on each electrode pad without temporarily attaching conductive balls using a flux. To do.

  In order to solve the above-described problems, the present invention provides a bump forming method in which bumps are respectively formed on a plurality of electrode pads formed in a predetermined pattern on one side of a substrate using a conductive ball, Using a jig in which a plurality of depressions that respectively accommodate the balls are formed corresponding to the pattern, a step of mounting conductive balls in each of the depressions of the jig, and the substrate, and each depression A process of aligning each electrode pad in the vertical direction and placing the electrode pad on the jig, and holding the positional relationship in the vertical direction between the jig and the substrate, at least each conductive ball is brought into close contact with each electrode. And a step of forming a bump by heating and melting these conductive balls, and the jig is made of a material having a linear thermal expansion coefficient equivalent to that of the substrate and not wet when the conductive balls are dissolved. This The features.

  According to the present invention, the jig is made of a material that does not get wet, and the conductive ball is heated and melted and transferred to the electrode pad in a state where the conductive ball is mounted in the recess formed in the jig. Wear flux is not required. As a result, a step of applying a flux at the time of bump formation and a step of cleaning a flux residue remaining after the conductive ball is heated and melted are not required, and the number of manufacturing steps can be reduced to further reduce the cost of bump formation. In addition, problems such as corrosion of electrode pads do not occur.

  Here, the shape of the recess and the depth from the jig surface are not particularly limited as long as the recess can accommodate the conductive balls. For example, when a conductive ball is mounted in the recess, the outer peripheral surface of the conductive ball may not protrude from the jig surface. In such a case, the recesses where the conductive balls are mounted and the electrode pads of the substrate are aligned in the vertical direction, and the positional relationship in the vertical direction between the jig and the substrate is maintained. If the jig and the substrate are turned upside down as they are (that is, the jig and the substrate are fixed), the conductive ball falls on the electrode pad, and the two can be brought into close contact with each other.

  By the way, when the conductive ball is heated and melted and transferred to the electrode pad, the jig and the substrate can also be heated, but in the present invention, it has a linear thermal expansion coefficient equivalent to that of the substrate. On the other hand, it is possible to prevent each conductive ball on the jig from being displaced. In the present invention, the linear thermal expansion coefficient is the same when the linear thermal expansion coefficient of the substrate and the jig are the same (both materials are the same), and when the conductive ball is heated to a predetermined temperature. When the thermal expansion of each substrate or substrate is heated, there is a difference in the coefficient of linear thermal expansion between the conductive pad mounted in the recess and the corresponding electrode pad within the range where the contact state is maintained. Including. In the present invention, since the conductive ball is heated and melted and transferred to the electrode pad, the volume of the bump does not change. For this reason, it is not necessary to make the depth and shape of the hollow part formed in the jig highly accurate.

  In this invention, it is preferable that the said hollow part is hemispherical and the curvature of the internal peripheral surface is set smaller than the curvature of an electroconductive ball. According to this, each conductive ball is positioned at a predetermined position in the recess by simply dropping the conductive ball into the recess. When the conductive ball is heated to a predetermined temperature, the jig and the substrate are heated and thermally expanded. At this time, even if there is a difference in the amount of thermal expansion between the jig and the substrate, the conductive ball remains in the recess. By following and rolling, the contact state with the electrode pad can be reliably maintained and transferred to the electrode pad when the conductive ball is heated and melted.

  Further, in order to solve the above problems, the present invention is a bump forming method in which bumps are respectively formed on a plurality of electrode pads formed in a predetermined pattern on one side of a substrate using a conductive ball, A jig in which a plurality of through holes are formed corresponding to the pattern is positioned on the substrate so that each through hole faces each electrode pad, and a conductive ball is placed on each of the through holes. A step of dropping and contacting each electrode, and a step of forming a bump by heating and melting a conductive ball while maintaining a vertical positional relationship between the jig and the substrate. The substrate may have a linear thermal expansion coefficient equivalent to that of the substrate.

  4. The bump forming method according to claim 1, wherein a circuit element is formed on a silicon wafer as a substrate, and a lead-free solder bump is provided on each electrode pad of the circuit element. When forming the jig, the jig may be made of glass or silicon.

(A) And (b) is the top view and partial sectional view which show typically the jig used for implementation of the bump formation method of this invention. The figure which shows typically the reflow furnace which can implement the bump formation method of this invention using the jig shown in FIG. (A)-(d) is a figure which illustrates a bump formation procedure roughly. (A)-(d) is a figure which illustrates schematically the bump formation procedure using the jig concerning a modification. (A)-(d) is a figure which illustrates schematically the bump formation procedure using the jig which concerns on another modification.

  Hereinafter, with reference to the drawings, it is assumed that the substrate has electrode pads P formed on the surface of a silicon wafer (hereinafter referred to as “wafer W”) with a pitch (predetermined pattern) in the range of, for example, 100 to 200 μm. The bump forming method according to the embodiment of the present invention will be described by taking as an example the case where the conductive ball B is a solder ball of, for example, 100 to 150 μm and the solder bump is formed on the electrode pad P using a reflow furnace.

  Referring to FIG. 1, reference numeral 1 denotes a jig used for carrying out the bump forming method of the present invention. The jig 1 is composed of a borosilicate glass plate having a predetermined thickness. The outline of the jig 1 is not particularly limited, and can be a square having one side longer than the wafer diameter or a circle larger than the wafer diameter. On one side of the jig 1 (the upper surface in FIG. 1), recesses 11 are formed corresponding to the electrode pads P formed on the surface of the wafer W on which bumps are to be formed. Each hollow part 11 is formed in the same shape, for example, hemispherical. Further, the curvature of the inner peripheral surface of the recess 11 is smaller than the curvature of the solder ball B, and when the solder ball B is arranged, the outer peripheral surface is sized so as to protrude from the upper surface of the jig 1. The depressions 11 corresponding to the electrode pads P on the wafer W can be formed through, for example, a photolithography process conventionally used in a semiconductor device manufacturing process. In this case, a photoresist, a metal film, or a laminated film of a photoresist and a metal film is used as an etching mask, and wet etching using hydrofluoric acid or the like is used for etching. The solder ball B is lead-free and formed in a spherical shape by a known method. Note that other metal balls such as lead-containing solder and gold can also be used.

  Referring to FIG. 2, reference numeral 2 denotes an example of a reflow furnace used for carrying out the bump forming method of the present invention. The reflow furnace 2 is a sealed container that can form a vacuum atmosphere or an inert gas atmosphere. Heating means 21 such as an infrared lamp or an electric heater is provided in the reflow furnace 2 so that the inside of the reflow furnace 2 can be heated and held at a predetermined temperature. Further, an XY stage 22 having a known structure is provided at the bottom in the reflow furnace 2. A cylindrical support member 23 is installed on the XY stage 22, and the jig 1 is held by placing the lower peripheral edge of the lower surface of the jig 1 on the upper surface of the support member 23. The XY stage 22 may be configured to be able to rotate in the circumferential direction (θ direction) with its center as the rotation center.

  On the upper surface of the XY stage 22, an image pickup unit 3 such as a CCD camera is provided inside the support member 23. Here, in the present embodiment, since the jig 1 is made of a borosilicate glass plate, that is, made of a transparent material, the wafer W is imaged through the jig 1 or the conductive material mounted in each of the depressions 11. The ball can be imaged. This is advantageous when the alignment of the wafer W with respect to the jig 1 and the confirmation of whether or not the conductive ball B is securely mounted in each recess 11. In the reflow furnace, a ring-shaped pressing member 24 that presses the outer peripheral edge of the jig 1 and the wafer W downward is provided so as to face the support member 23. A driving shaft 25 of a driving source such as an air cylinder or a motor (not shown) is connected to the pressing member 24 and moves up and down in the reflow furnace 2.

  Next, a procedure for forming solder bumps will be described with reference to FIGS. First, the jig 1 is positioned and placed on the support member 23 in the reflow furnace 2 so as to be supported by the outer peripheral edge of the lower surface (state shown in FIG. 2). At this time, the pressing member 24 is in an upper position separated from the jig 1. When the jig 1 is placed, the solder balls B are mounted on the respective recessed portions 11 of the jig 1. In this case, a plurality of solder balls B are placed on the upper surface of the jig 1, the solder balls B are dispersed with a brush or the like, and the solder balls B are dropped into the respective recessed portions 11.

  In addition, since the hollow part 11 is formed in the hemispherical shape, when the solder ball B is dropped into each hollow part 11, each solder ball B is positioned at a predetermined position in the hollow part 11. Then, the jig 1 is imaged by the imaging means 3 and image processing is performed by a known method to check whether one solder ball B is mounted in all the recessed portions 11. Note that when the plurality of solder balls B are mounted in each recess 11 or no solder ball B is mounted in any of the recesses 11, the excess solder balls B are removed by the imaging unit 3. Alternatively, the solder ball B is mounted in the recess 11 that is not mounted.

  When the solder ball B is mounted in each recess 11, the wafer W is transferred to an upper position of the jig 1 by a transfer robot having a known structure. At this time, the transfer robot holds the wafer W so that the electrode pad P faces downward. Then, the imaging means 3 images the jig 1 and the wafer W through the jig 1, performs image processing by a known method, and each electrode pad P of the wafer W is directly above the solder ball B mounted in each depression. Is confirmed (see FIG. 3A). If there is a deviation in the vertical positional relationship between the solder ball B and the electrode pad P, the correction amount is calculated from the image processing result, and the XY stage is moved in accordance with this, and the alignment (position (See FIG. 3B).

  When the alignment of the wafer W with respect to the jig 1 is completed, the wafer W is placed on the jig 1 while the positional relationship in the vertical direction between the jig 1 and the wafer is maintained by the transfer robot. Thereby, each solder ball B is brought into close contact with each electrode pad P (see FIG. 3C). Then, the pressing member 24 is moved downward, and the wafer W and the jig 1 are pressed toward the upper surface of the support member 23 to fix them. In this state, the heating means 21 is operated to heat the inside of the reflow furnace 2 to a joining temperature that matches the solder material used (for example, in the case of Sn-3Ag-0.5Cu, 225 ° C.). As a result, the solder balls B are heated and melted and transferred to the electrode pads P, and solder bumps are formed on the electrode pads P (see FIG. 3D). At this time, since the jig 1 is composed of a borosilicate glass plate that does not get wet when the solder balls B melt, the volume when the solder balls B are reliably transferred to the electrode pads P and solder bumps are formed. Will not change. Prior to heating the wafer W, a vacuum atmosphere or an inert gas atmosphere may be formed in the reflow. Finally, the wafer W on which the solder bumps are formed is taken out from the reflow furnace 2 by the transfer robot after the pressing by the pressing means 24 is released.

  According to this embodiment, the temporary wearing flux is not necessary. This eliminates the need for a flux application process at the time of solder bump formation and a flux residue cleaning process remaining after the solder balls are heated and melted, thereby reducing the number of manufacturing processes and further reducing the cost of bump formation. In addition, there is no problem such as corrosion of the electrode pad after bump formation.

  In the above embodiment, since the jig 1 is made of a borosilicate glass plate having a linear thermal expansion coefficient approximate to that of the wafer W, unlike the metal one, the solder ball B is melted in the reflow furnace 2. The jig 1 and the wafer W are also heated by heating the inside thereof, and even if the jig 1 and the wafer W are thermally expanded, the solder balls B brought into close contact with the electrode pads P are portions where no electrode pads are present. It is possible to prevent the positional deviation from occurring. At this time, each recess 11 of the jig 1 has a hemispherical shape in which the curvature of the inner peripheral surface is set to be smaller than the curvature of the solder ball B, so that a difference in thermal expansion amount occurs between the jig 1 and the wafer W. In this case, the solder ball B rolls following the inside of the recessed portion 11 so that the contact state with the electrode pad P is reliably maintained and transferred to the electrode pad P.

  Although the bump forming method of the present embodiment has been described above, the present invention is not limited to the above. In the said embodiment, although the hollow part 11 was hemispherical and when the conductive ball B was arrange | positioned, what was set so that the outer peripheral surface might protrude the jig 1 upper surface upward, in the hollow part 11 was demonstrated. As long as the conductive ball B can be accommodated, the shape and depth of the recess 11 are not particularly limited.

  As shown in FIG. 4A, when the conductive ball B is mounted in the recess 11a, the outer peripheral surface of the conductive ball B may be prevented from protruding from the surface of the jig 10a. In such a case, the jigs 10a and the wafer W are fixed after the depressions 11a on which the solder balls B are mounted and the electrode pads P of the wafer W are aligned in the vertical direction (FIG. 4). Then, the jig 10a and the wafer W are turned upside down together. As a result, the conductive balls B fall on the electrode pads and adhere to each other (see FIG. 4C). In this state, the conductive ball may be heated and dissolved and transferred onto the electrode pad (see FIG. 4D).

  Moreover, in the said embodiment, although what used the borosilicate glass as a material of the jig 1 was demonstrated to the example, it is not limited to this, For example, it is good also as a product made from a silicon | silicone. In this case, if the solder ball B mounted in the recess 11 and the corresponding electrode pad P are kept in close contact with each other when heated together with the wafer W and thermally expanded, there is a difference in linear thermal expansion coefficient. In addition, when the substrate is other than the wafer W, the material of the jig 1 is appropriately selected accordingly.

  Furthermore, in the said embodiment, although what provided the hollow parts 11 and 11b was demonstrated as an example as the jig 1, the through-hole 12 corresponding to each electrode pad P on the wafer W was provided as the jig 10b. (See FIG. 5A). This jig 10b is made of borosilicate glass having a linear thermal expansion coefficient equivalent to that of the wafer W, as in the above embodiment. Then, the respective through holes 12 are positioned so as to face the respective electrode pads P and arranged on the wafer W (see FIG. 5B), and the solder balls B are respectively applied to the through holes 12 by the same method as described above. It is dropped and brought into contact with each electrode pad P (see FIG. 5C). In this state, when the solder balls B are heated and melted, bumps are formed on the electrode pads P (see FIG. 5D).

  In the above embodiment, the example using the reflow furnace 2 has been described as an example. However, there is no particular limitation as long as the conductive ball can be heated and melted. May be dissolved by heating. In the above embodiment, the case where bumps are formed on one wafer W has been described as an example. However, the wafer W and the jig 1 on which the solder balls B are mounted are aligned and fixed in advance in a place other than the reflow furnace. A plurality of fixed ones can be stored in a reflow furnace and processed simultaneously.

  DESCRIPTION OF SYMBOLS 1 ... Jig, 11 ... Recessed part, B ... Solder ball (conductive ball), P ... Electrode pad, W ... Wafer (substrate).

Claims (4)

A bump forming method for forming bumps on a plurality of electrode pads formed in a predetermined pattern on one side of a substrate using a conductive ball,
Using a jig in which a plurality of depressions respectively accommodating the conductive balls are formed corresponding to the pattern, and mounting the conductive balls in each of the depressions of the jig;
Placing the substrate on the jig by aligning the depressions and the electrode pads in the vertical direction; and
Including maintaining a vertical positional relationship between the jig and the substrate, at least each conductive ball in close contact with each electrode, and heating and melting these conductive balls to form bumps,
A bump forming method, wherein the jig has a linear thermal expansion coefficient equivalent to that of a substrate and does not get wet when the conductive ball is melted.   The bump forming method according to claim 1, wherein the recess is hemispherical, and a curvature of an inner peripheral surface thereof is set smaller than a curvature of the conductive ball. A bump forming method for forming a bump on each of a plurality of electrode pads formed in a predetermined pattern on one side of a substrate using a conductive ball,
A step in which a jig in which a plurality of through holes are formed corresponding to the pattern is positioned on the substrate so that each through hole faces each electrode pad, and
Dropping a conductive ball into each of the through holes and bringing it into contact with each electrode;
A step of heating and melting a conductive ball to form a bump while maintaining a vertical positional relationship between the jig and the substrate, and
A bump forming method comprising the jig having a linear thermal expansion coefficient equivalent to that of a substrate. 4. The bump forming method according to claim 1, wherein a circuit element is formed on a silicon wafer as a substrate, and a lead-free solder bump is provided on each electrode pad of the circuit element. When forming,
A bump forming method, wherein the jig is made of glass or silicon.