Semiconductor device manufacturing: process – Coating with electrically or thermally conductive material – To form ohmic contact to semiconductive material
Reexamination Certificate
1996-06-06
2001-11-20
Graybill, David E. (Department: 2814)
Semiconductor device manufacturing: process
Coating with electrically or thermally conductive material
To form ohmic contact to semiconductive material
Reexamination Certificate
active
06319810
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a method for forming solder bumps necessary for bump-connection.
2. Description of the Related Art
Recently, high-density mounting of semiconductor devices has been developed for the purpose of miniaturization and weight-reduction of electronic apparatuses such as bump grid array (BGA) packages or semiconductor devices using flip-chips. Since a space corresponding to lead sections can be eliminated by using BGA packages, which space is necessary when quad flat packages (QFP) are used, it is possible to realize a higher-density mounting. Also, according to the flip-chip mounting, it is possible to arrange a plurality of silicon chips on a circuit board to realize an ideal high-density mounting.
While the electric connection of the BGA packages or the flip chips is generally carried out by using solder bumps, it is necessary to form uniform solder bumps having identical metal compositions and/or uniform sizes and having no foreign matter or voids therein in order to realize reliable connections.
There are a plating method and a transfer method of a solder alloy for forming solder bumps. The plating method is mainly applied to the flip-chips, wherein an aluminum pad providing an electrode is thickly plated with the solder in a final step of a wafer process, which is then treated in a reflow furnace to form solder bumps.
In the transfer method, pieces of solder punched out from a thin solder sheet or cut from a solder wire are heated to form solder balls, which are then transferred to electrode pads of a semiconductor device by a transfer tool.
Another transfer method is disclosed in Japanese Unexamined Patent Publication No. 6-124953, wherein a flat plate having cavities at positions corresponding to electrode pads of a semiconductor device is used, and the cavities are filled with a solder paste by squeezing. According to this method, protruding electrodes formed of golden studs are provided on electrode pads of a semiconductor device, which electrodes are then positioned close to the corresponding cavities filled with the solder paste. By heating, a molten solder is adhered to the protruding electrode while enveloping the latter. To transfer the solder paste to the protruding electrode, the semiconductor device is positioned above the flat plate while directing the projection electrodes downward, and brought close to the flat plate to locate the projection electrodes opposite to the holes of the flat plate.
According to the method wherein the flat plate is used, having the cavities corresponding to the electrode pads of the semiconductor device, volumes of the solder paste filled in the respective cavities become equal to each other provided the capacities of the respective cavities are uniform, whereby all the solder balls (solder bumps) thus formed have the same size. Accordingly, it is possible to form a number of micro-solder bumps at once on the semiconductor device or the circuit board.
However, according to the methods wherein the solder balls are formed before the solder bumps are formed, there is a tendency that residual flux or the like sticks on the outer surfaces of the solder balls once formed, resulting that residual flux is interposed between the solder bumps and the electrode pads when the solder balls are transferred to the electrode pads and deteriorates the soldering function. Thus the adhesion of the solder balls to the electrode pads become inferior to cause the fall-off of the solder balls from the electrode pads. Also, the residual flux may be contained in the solder balls (solder bumps) transferred to the electrode pads, which generates bubbles in the solder bumps when the semiconductor device is heated again for the purpose of bonding the semiconductor device to the circuit board by a flip-chip bonding. Accordingly, it is preferable to remove the residual flux or the like stuck to the solder balls by a rinsing process after the solder balls have been formed in the cavities of the flat plate and before the transfer operation is executed. However, there is a risk in that the solder balls may be displaced from the cavities of the flat plate in the rinsing process and cannot be maintained at predetermined positions.
While, according to the method wherein the protruding electrode is provided on the electrode pad of the semiconductor device, the protruding electrode comprises a golden stud formed on the electrode pad by wire bonding. Since wire bonding is carried out one by one, a longer time period is necessary for repeating the wire bondings a number of times, and also it is difficult to carry out wire bonding for an electronic device including elements arranged at a micro-pitch. Further, according to the structure wherein the molten solder sticks to the protruding electrode while enveloping the latter, a size of the solder bump becomes larger than that of the protruding electrode, so it is difficult to form micro-solder bumps at a smaller pitch. Also, since the semiconductor device is positioned above the flat plate during the transfer process, the solder paste must move to the protruding electrode against the force of gravity.
A further transfer method is disclosed in Japanese Unexamined Patent Publication No. 4-263433. According to this method, a flat plate is used, having cavities at positions corresponding to electrodes of a semiconductor device. The cavities of the flat plate are filled with a solder paste by squeezing, and thereafter, the flat plate is heated so that the solder paste in the cavities forms solder balls.
The solder paste is a mixture of fine solder powders and a flux or a solvent. When the solder paste is heated, the solvent is evaporated and a molten solder is rounded due to surface tension to form a solder ball. Then, a semiconductor device is brought close to the flat plate having the solder balls held in the cavities, while positioning electrode pads of the semiconductor device relative to the cavities of the flat plate. Then the solder balls in the cavities are transferred to the electrode pads of the semiconductor device, using a heating and pressing device.
However, a throughput per an apparatus used in this method is low, because a considerable time is necessary to heat the flat plate and the semiconductor chip, to allow the solder to flow onto the entire surface of the electrode pad, and to cool the flat plate and the semiconductor chip, each time after the positioning step.
Also, the inventors have noted a problem that an oxidation on the surfaces of the solder bumps and the electrode pads advances when they are heated to transfer the solder bumps to the electrode pads, with the result that a low wettability of the electrode pads to the solder is deteriorated. If the electrode pad is coated with a metal layer and the melting point of the solder is lower than 200° C., the problem will not be so serious. However, if the solder bumps are directly transferred to the electrode pads made of copper or nickel which do not dissolve so much in the solder, the electrode pads are less wettable to the solder. If the solder includes Pb as a main component (for example, Pb-10Sn, Pb-5Sn) having a high melting point, the electrode pads are sometimes heated above 300° C., with the result that the copper or nickel electrode pads are easily oxidized in a short time and the flux residuals on the surface of the solder balls are carbonized. Therefore, the electrode pads are hardly wettable to the solder. Usually, after the solder balls are formed, a flux is applied to the solder balls and the solder balls are reheated above the melting point of the solder to modify the shape of the solder balls. Therefore, even when the solder balls are transferred to the electrode pads in the above described condition, it seems that the peripheries of the bottoms of the solder bumps (balls) are satisfactorily adhered or soldered to the electrode pads by the action of the flux and it appears that good solder bumps are formed on the electrode pads. However, if a force is applied
Mizukoshi Masataka
Ochiai Masayuki
Otake Koki
Watanabe Yuuji
Yamagishi Yasuo
Armstrong, Westerman, Hattori, McLeland & Naughton, LLP.
Fujitsu Limited
Graybill David E.
LandOfFree
Method for forming solder bumps does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for forming solder bumps, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for forming solder bumps will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2574912