JP6269682B2 - Substrate bonding method - Google Patents

Description

  The present invention relates to a method for bonding substrates, and for example, relates to a method for bonding substrates in which metal layers formed on the substrates are bonded together.

  A joining method is known in which substrates are electrically and mechanically connected to each other using a metal layer. For example, there are known a method for stacking semiconductor elements to achieve high integration and a method for bonding semiconductor elements to a circuit board. A method is known in which a copper terminal is formed on a circuit board, and the surface of the terminal is replaced with tin (for example, Patent Document 1).

JP 2011-44563 A

  When the diameter of the metal layer is reduced, the bonding strength between the metal layers decreases, and the bonding strength between the substrates decreases. The method for bonding the substrates is aimed at improving the bonding strength between the metal layers.

And a second metal layer formed on the first metal layer and a second substrate formed on the first substrate, a step of contacting a porous body comprising a first metal, after said contacting step, said first Using a liquid containing a second metal having a higher standard electrode potential than the metal, replacing the first metal on the surface of the porous body with the second metal, and heating the second metal, And a step of bonding the first metal layer and the second metal layer .

  According to the bonding method of the present substrate, the bonding strength between the metal layers can be improved.

1A is a cross-sectional view of a semiconductor device according to Comparative Example 1, and FIGS. 1B and 1C are enlarged perspective views of a metal layer and a molten metal. 2A and 2B are cross-sectional views of the semiconductor device according to Comparative Example 2. FIG. 3A and 3B are cross-sectional views of the semiconductor device according to Comparative Example 1. FIG. 4A to 4C are enlarged views of the metal layer. FIG. 5A to FIG. 5D are cross-sectional views illustrating the substrate bonding method according to the first embodiment. FIG. 6A and FIG. 6B are cross-sectional views (part 1) illustrating the bonding method of the substrates according to the second embodiment. FIGS. 7A to 7C are cross-sectional views (part 2) illustrating the bonding method of the substrates according to the second embodiment. FIG. 8A to FIG. 8D are cross-sectional views illustrating a method for bonding substrates according to the third embodiment. FIG. 9A and FIG. 9B are cross-sectional views of the powder of Example 3. FIG. 10A to FIG. 10C are cross-sectional views illustrating the substrate bonding method according to the fourth embodiment. FIG. 11A and FIG. 11B are cross-sectional views of the powder of Example 4. FIG. 12A to FIG. 12C are cross-sectional views illustrating the substrate bonding method according to the fifth embodiment. FIG. 13A and FIG. 13B are cross-sectional views of the powder of Example 5. FIG. 14A to FIG. 14C are cross-sectional views illustrating the substrate bonding method according to the sixth embodiment. FIG. 15A to FIG. 15C are cross-sectional views (part 1) showing the bonding method of the substrates according to the seventh embodiment. FIG. 16A to FIG. 16C are cross-sectional views (part 2) illustrating the bonding method of the substrates according to the seventh embodiment. FIG. 17A and FIG. 17B are cross-sectional views (part 3) illustrating the bonding method of the substrates according to the seventh embodiment. FIG. 18A to FIG. 18C are cross-sectional views (part 4) illustrating the bonding method of the substrates according to the seventh embodiment.

  FIG. 1A is a cross-sectional view of a semiconductor device according to Comparative Example 1. FIG. 1B and FIG. 1C are enlarged perspective views of the metal layer and the molten metal. Referring to FIG. 1A, a metal layer 12 is formed on a substrate 10. An electrode 24 is formed on the substrate 20 (lower in the drawing), and a metal layer 22 is formed on the electrode 24. The metal layers 12 and 22 are joined using a molten metal 40 such as solder. FIG. 1B shows an example of the metal layer 22. Referring to FIG. 1B, the metal layer 22 is a metal terminal mainly including Cu (copper), for example. A molten metal 40 is formed on the metal layer 22. The width W1 of the metal layer 22 (diameter when the metal layer 22 is a cylinder) is, for example, 35 μm, the height H1 of the metal layer 22 is, for example, 30 μm, and the height H2 of the molten metal 40 is, for example, 13 μm. With reference to FIG. 1C, a barrier layer 23 is formed between the metal layer 22 and the molten metal 40. The barrier layer 23 mainly contains, for example, Ni (nickel). The height H3 of the barrier layer 23 is 5 μm, for example. The barrier layer 23 suppresses the reaction between the metal layer 22 and the molten metal 40. With the miniaturization of semiconductor devices and the like, the ratio of the height H1 to the width W1 of the metal layer 22 increases, for example, 1 or more. The reason for increasing this ratio is to cope with the warpage of the substrates 10 and 20 caused by the difference in the thermal expansion coefficients of the substrates 10 and 20, for example.

  2A and 2B are cross-sectional views of the semiconductor device according to Comparative Example 2. FIG. FIG. 2A is a view before joining the substrates, and FIG. 2B is an enlarged view of the metal layer after joining. Referring to FIG. 2A, the molten metal 40 is formed on the metal layer 22 (lower in the drawing) by, for example, electrolytic plating. The surface of the molten metal 40 formed by the electrolytic plating method has large irregularities. With reference to FIG.2 (b), the molten metal 40 and the metal layer 12 are made to contact, and the metal layers 12 and 22 are joined by heating. At this time, a region where the molten metal 40 is not in contact is formed at the interface with the metal layer 22. For this reason, a void 42 is generated in the molten metal 40. For example, when the flatness of the surface of the metal layer 22 is high, the void 42 is likely to occur.

  3A and 3B are cross-sectional views of the semiconductor device according to Comparative Example 1. FIG. FIG. 3A is a view before bonding of the substrates, and FIG. 3B is an enlarged view of the metal layer after bonding. Referring to FIG. 3A, in order to suppress the generation of such voids 42, the molten metal 40 is heated to a temperature equal to or higher than the melting point before the bases 10 and 20 are joined. Thereby, the molten metal 40 becomes dome shape (namely, hemispherical). Referring to FIG. 3B, after joining the metal layers 12 and 22, generation of voids in the molten metal 40 is suppressed.

  4A to 4C are enlarged views of the metal layer. 4A to 4C, for example, a semiconductor element may be mounted on a circuit board such as a silicon interposer. Alternatively, semiconductor devices may be highly integrated by three-dimensionally stacking semiconductor elements. In such a case, the wiring can be formed by using a manufacturing process used in the wiring formation process of the semiconductor element. For this reason, a wiring density of 50 times or more can be realized as compared with the resin circuit board. Therefore, the diameters of the metal layers 12 and 22 are 1/8 of the diameter of the resin circuit board.

  Thus, when the width W1 of the metal layer 22 (or the diameter when the metal layer 22 is a cylinder) is decreased, the height H2 of the molten metal 40 is decreased. This is because the diameter of the dome is substantially the same as the width of the metal layer 22, and therefore the upper limit of the volume of the spherical molten metal 40 that can be formed on the metal layer 22 is determined. On the other hand, when the molten metal 40 is heated to form a dome shape, the molten metal 40 and the metal layer 22 react to form an alloy layer 44. The height H4 of the alloy layer 44 is determined by the amount of heat during heating regardless of the width of the metal layer 22. Therefore, the height H4 of the alloy layer 44 is substantially constant. As the width W1 of the metal layer 22 decreases, the ratio of the alloy layer 44 increases. Since the melting point of the alloy layer 44 is higher than that of the molten metal 40, the metal layers 12 and 22 cannot be bonded satisfactorily.

  FIG. 5A to FIG. 5D are cross-sectional views illustrating the substrate bonding method according to the first embodiment. With reference to FIG. 5A, the electrode 14 is formed on the upper surface of the substrate 10. A metal layer 12 is formed on the electrode 14. An electrode 24 is formed on the lower surface of the substrate 20. A metal layer 22 is formed under the electrode 24. The bases 10 and 20 are, for example, a semiconductor substrate such as a silicon substrate or an insulating substrate such as a resin substrate. The electrodes 14 and 24 mainly contain a metal such as Cu or Al (aluminum). The metal layers 12 and 22 mainly contain Cu or the like, for example. A plurality of powders 32 are fixed to the lower surface of the metal layer 22. The powder 32 is particulate and includes a metal 35 (first metal) such as Cu, for example. The particle size of the powder 32 is, for example, 1 μm to several μm. The plurality of powders 32 may be a single layer of powder 32, but may be two or more layers of powder 32. The powder 32 is formed using, for example, an atomizing method. The powder 32 is fixed to the surface of the metal layer 22 using, for example, ultrasonic waves.

  With reference to FIG. 5B, the metal layers 12 and 22 are opposed to and in contact with each other through the plurality of powders 32 so that the powder 32 contacts the upper surface of the metal layer 22. Referring to FIG. 5C, the surfaces of the plurality of powders 33 are replaced with molten metal 36 (second metal). For example, the entire metal 35 in the powder 32 is replaced with a molten metal 36 to obtain a powder 33. For the replacement, for example, a replacement plating method using electroless plating is used. The displacement plating method is a method of replacing a metal having a low standard electrode potential with a metal having a high standard electrode potential. For example, as a metal substituted for Sn (tin), Cu, In (indium), Ni, Fe (iron), Mn (manganese), Zr (zirconium), Ti (titanium), Al, Cr (chromium) and Zn ( Zinc). Therefore, the metal 35 includes Sn, and the molten metal 36 may be a metal including at least one of Cu, In, Ni, Fe, Mn, Zr, Ti, Al, Cr, and Zn. When the powder 32, the metal layers 12 and 22, and the electrodes 14 and 24 are the same metal, the metal on the side surfaces of the metal layers 12 and 22 and the electrodes 14 and 24 is replaced with molten metal, and the replacement layers 16 and 26 are formed. It is formed.

  With reference to FIG. 5 (d), the molten metal 36 is melted to become the metal layer 34 by heating the powder 33 to the melting point of the molten metal 36 or more and to the melting point of the metal layer 34 or less. By cooling, the metal layers 12 and 22 are joined by the metal layer 34.

  According to Example 1, as shown in FIG. 5B, the metal layers 12 and 22 formed on the plurality of substrates 10 and 20 are opposed to each other through the porous body including the plurality of powders 32. Let As shown in FIG. 5C, the metal 35 on the surface of the plurality of powders 32 is replaced with molten metal 36. The metal layers 12 and 22 are joined together by heating the molten metal 36. Thereby, like the comparative example 1, it is not necessary to melt the molten metal 40 before joining the metal layers 12 and 22 together. For this reason, the alloy layer 44 is not formed as shown in FIGS. Therefore, the bonding strength between the metal layers 12 and 22 can be improved.

  Further, since the porous body including the plurality of powders 32 has a large surface area per unit volume, the metal 35 can be easily replaced with the molten metal 36.

  Furthermore, the porous body can be easily fixed to the surface of the metal layer 12 or 22 by using the plurality of powders 32 as the porous body.

  Example 2 is an example in which substrates are laminated. FIG. 6A to FIG. 7C are cross-sectional views illustrating a method for bonding substrates according to the second embodiment. Referring to FIG. 6A, the bases 10a and 20a, the metal layers 12a and 22a, the electrodes 14a and 24a, and the powder 32a are the same as the bases 10 and 20 and the metal layers 12 and 22 shown in FIG. Same as electrodes 14 and 24 and powder 32. An electrode 14b is formed on the upper surface of the base body 20a, and a metal layer 12b is formed on the electrode 14b. The electrode 14b and the electrode 24a are electrically connected through a through electrode formed in the base body 20a. The electrode 14b and the electrode 24a may be electrically connected to a circuit formed in the base body 20a.

  With reference to FIG. 6B, the base body 20b is laminated on the base body 20a. Each metal layer 12b is opposed to each metal layer 22b with a plurality of powders 32b interposed therebetween. With reference to FIG. 7A, the base body 20c is laminated on the base body 20b. Each metal layer 12c is opposed to each metal layer 22c via a plurality of powders 32c.

  Referring to FIG. 7B, the surfaces of the plurality of powders 32a to 32c are replaced with molten metal, as in FIG. 5C of the first embodiment. Thereby, powders 33a to 33c containing molten metal are formed. Referring to FIG. 7C, the powders 33a to 33b are heated to the melting point or higher to melt the powders 33a to 33c and form the metal layers 34a to 34c. By cooling, the metal layers 12a and 22a are joined by the metal layer 34a. At the same time, the metal layers 12b and 22b are joined by the metal layer 34b, and the metal layers 12c and 22c are joined by the metal layer 34c.

  When three or more substrates are stacked, it is also conceivable to join a metal layer each time the substrates are stacked. However, the metal layer bonded at the initial stage of the lamination undergoes many heating processes for bonding. For example, the metal layer 34a goes through three heating steps. For this reason, the joining metal changes in quality, becomes brittle, and is liable to crack. In this manner, the quality of the bonding varies depending on the metal layers 34a to 34c. According to Example 2, as shown in FIG. 7A, the metal layers 12a and 22a, 12b and 22b, and 12c and 22c are fixed by the powders 32a to 32c. For this reason, even if the bases 20b to 20c are laminated before the metal layers 12a and 22a are heated, the bases 20a to 20c are fixed to each other. Therefore, after laminating the bases 20a to 20c, the metal layers 12a and 22a, 12b and 22b, and 12c and 22c are bonded together by heating the plurality of powders 33a to 33c as shown in FIG. 7B. Can be joined. Thereby, the dispersion | variation in the quality of joining can be suppressed.

  The base 10 can be a circuit board such as a silicon interposer or a resin substrate, and the bases 20a to 20c can be semiconductor chips, for example. The number of stacked base bodies 20a to 20c can be set to 2, 4, or 8, for example. When the bases 20a to 20c are semiconductor chips, the film thickness is, for example, 20 μm to 100 μm.

  Example 3 is an example in which a part of the surface of the powder 32 is replaced with molten metal. FIG. 8A to FIG. 8D are cross-sectional views illustrating a method for bonding substrates according to the third embodiment. FIG. 9A and FIG. 9B are cross-sectional views of the powder of Example 3. Referring to FIG. 8A, a protective film 28 is formed on the base 20 (lower surface). The protective film 28 is a protective film of the base 20 and is an insulator containing a resin such as polyimide or epoxy. Other steps are the same as those in FIG. 5B of the first embodiment, and a description thereof will be omitted.

  With reference to FIG. 9A, the powder 32 is entirely a metal 35. The metal 35 is, for example, Cu. With reference to FIG. 8B and FIG. 9B, the powder 33 is formed by replacing the metal 35 on the surface of the powder 32 with a molten metal 36. The surface of the powder 33 is a molten metal 36, and the metal 35 remains inside the powder 33. For example, the diameter of the metal 35 remaining inside is about ½ to ¼ of the diameter of the powder 33. For example, when the metal 35 is Cu and the molten metal 36 is Sn, Sn can remain in the powder 33 by performing substitutional electroless Sn plating at 60 ° C. for 15 minutes.

  With reference to FIG. 8C, the molten metal 36 is melted by heating the plurality of powders 33. For example, when the metal 35 is Cu, the molten metal 36 is Sn, and the metal layers 12 and 22 are Cu, the powder 33 is heated at 300 ° C. for 3 seconds. Thereby, the metal layer 34 including the remaining metal 35 is formed.

  Referring to FIG. 8D, the molten metal 36 and the metal 35 are alloyed by heating. For example, when the metal 35 is Cu, the molten metal 36 is Sn, and the metal layers 12 and 22 are Cu, the peak top temperature is 240 ° C. or higher using a reflow furnace, and the temperature above the melting point of the molten metal 36 is 30 seconds or longer. To heat. As a result, the molten metal 36 and the metal 35 react to form an alloy layer 45 of Cu and Sn. Since the alloy layer 45 is harder than the metal layer 34, the bonding strength can be made higher than that of the first embodiment.

  As in the third embodiment, the metal 35 on the surface of the plurality of powders 32 may be replaced with the molten metal 36, and the metal 35 inside the powder 32 may not be replaced with the molten metal 36. Whether or not the metal 35 inside the powder 32 is replaced with the molten metal 36 can be controlled by, for example, the time of replacement plating.

  As shown in FIG. 8D, the alloy layer 45 is formed by alloying the inside of the powder 33 that is not replaced with the molten metal 36 and the molten metal 36. Since the molten metal 36 is soft, it is easily deformed. By forming the alloy layer 45, the metal layers 12 and 22 can be firmly joined. The alloy layer 45 may not be formed. Further, the protective film 28 may or may not be provided.

  Example 4 is an example in which the inside of the powder 32 is a refractory metal and the metal on the surface of the powder is replaced with a molten metal. FIG. 10A to FIG. 10C are cross-sectional views illustrating the substrate bonding method according to the fourth embodiment. FIG. 11A and FIG. 11B are cross-sectional views of the powder of Example 4. With reference to FIG. 10A, a resin film 19 is formed on the substrate 10. A resin film 29 is formed on the base 20 (lower surface). The resin films 19 and 29 are insulators including a thermosetting resin such as epoxy. In the powder 32, a metal 35 is formed around a refractory metal 37 via a barrier layer 38. Other steps are the same as those in FIG. 5B of the first embodiment, and a description thereof will be omitted.

  Referring to FIG. 11A, in the powder 32, the refractory metal 37 is surrounded by a barrier layer 38 and the barrier layer 38 is surrounded by a metal 35. The refractory metal 37 is, for example, Pd (palladium) having a diameter of 1 μm, and is formed using an atomizing method. A barrier layer 38 is formed on the surface of the refractory metal 37 using, for example, an electroless plating method or a barrel plating method. The barrier layer 38 is, for example, a Ni film having a thickness of 0.5 μm. A metal 35 is formed on the surface of the barrier layer 38 using, for example, an electroless plating method or a barrel plating method. The metal 35 is, for example, Cu having a thickness of 1 μm.

  With reference to FIG. 10B and FIG. 11B, powder 33 is formed by replacing metal 35 with molten metal 36. Since the barrier layer 38 is not replaced, the replacement plating stops without controlling the time. For example, when the metal 35 is Cu and the molten metal 36 is Sn, substitutional electroless Sn plating is performed at 60 ° C. for 15 minutes. Referring to FIG. 10C, the molten metal 36 of the powder 33 is melted to form the metal layer 34 including the refractory metal 37 and the barrier layer 38. For example, when the molten metal 36 is Sn, the powder 33 is heated at 300 ° C. for 3 seconds. Using a reflow furnace, the peak top temperature is set to 240 ° C. or higher, and heating is performed for 30 seconds or more. As a result, the thermosetting resin films 19 and 29 are cured in contact with each other.

  According to the fourth embodiment, the plurality of powders 32 mainly include the refractory metal 37 and the metal 35 formed on the surface of the refractory metal 37. Thereby, since the refractory metal 37 reinforces the metal layer 34, the alloying as in the third embodiment may not be performed.

  In Example 5, the inside of the powder 32 is an insulator, and the metal on the surface of the powder is replaced with molten metal. FIG. 12A to FIG. 12C are cross-sectional views illustrating the substrate bonding method according to the fifth embodiment. FIG. 13A and FIG. 13B are cross-sectional views of the powder of Example 5. Referring to FIG. 12A, in the powder 32, a metal 35 is formed around an insulator 39 through a barrier layer 38. Other steps are the same as those in FIG.

  Referring to FIG. 13A, in the powder 32, the insulator 39 is surrounded by a barrier layer 38, and the barrier layer 38 is surrounded by a metal 35. The insulator 39 is, for example, a resin such as polyimide having a diameter of 1 μm, and is formed by using an atomizing method. The other processes are the same as those in FIG.

  With reference to FIG. 12B and FIG. 13B, powder 33 is formed by replacing metal 35 with molten metal 36. Other steps are the same as those in FIG. 10B and FIG. 11B of the fourth embodiment, and the description thereof is omitted. With reference to FIG. 12C, the metal layer 34 including the insulator 39 and the barrier layer 38 is formed by melting the molten metal 36 of the powder 33. The other steps are the same as those in FIG.

  According to the fifth embodiment, the powder 32 includes the insulator 39 and the metal 35 formed on the surface of the insulator 39. Thereby, since the insulator 39 reinforces the metal layer 34, it is not necessary to perform alloying as in the third embodiment.

  As in Examples 4 and 5, the resin films 19 and 29 may be used to seal between the substrates 10 and 20. Thereby, it is not necessary to use an underfill material.

  As in Examples 1 to 5, the powder 32 in the porous body can be formed into particles (for example, spherical) mainly containing metal. Thereby, since the surface area of the powder 32 is increased, the metal 35 can be easily replaced with the molten metal 36.

  Example 6 is an example in which the powder is fibrous. FIG. 14A to FIG. 14C are cross-sectional views illustrating the substrate bonding method according to the sixth embodiment. Referring to FIG. 14 (a), the powder 32d is fibrous. Other steps are the same as those in FIG. 5B of the first embodiment, and a description thereof will be omitted.

  Referring to FIG. 14B, the fibrous powder 33d is formed by replacing the fibrous powder 32d with molten metal. The other steps are the same as those in FIG. Referring to FIG. 14C, the metal layer 34 is formed by melting the powder 33d. The other processes are the same as those in FIG.

  As in Example 6, the porous body may include a plurality of fibrous powders. Thus, the powders of Examples 1 to 5 may be fibrous.

  Example 7 is an example in which the substrate bonding method according to Example 3 is applied to a method for manufacturing a semiconductor device. FIG. 15A to FIG. 18C are cross-sectional views illustrating the substrate bonding method according to the seventh embodiment. Referring to FIG. 15A, the electronic circuit 50 is formed on the base 20. The substrate 20 is a silicon substrate and is in a wafer state including a plurality of chips. An insulating film 52 is formed on the substrate 20. The electronic circuit 50 includes a transistor formed on the base 20 and a wiring formed in the insulating film 52. The insulating film 52 is a silicon oxide film, for example, and includes an interlayer insulating film of multilayer wiring. The electrode 24 is formed on the insulating film 52. The electrode 24 contains Cu and is electrically connected to the electronic circuit 50. The chip size is, for example, 8.5 mm × 8.5 mm. The interval between the electrodes 24 is, for example, 50 μm, and 400 electrodes 24 are formed in one chip.

  Referring to FIG. 15B, a protective film 28 containing polyimide and having a film thickness of, for example, 5 μm is formed on the insulating film 52. With reference to FIG. 15C, an opening is formed in the protective film 28 on the electrode 24. A metal layer 22 containing Cu and having a height of, for example, 15 μm is formed on the electrode 24. The metal layer 22 is a bump (projection terminal) for connection. By grinding the back surface of the substrate 20 in the wafer state, the thickness of the substrate 20 is set to 100 μm, for example. Using a dicing method, the wafer is cut into a chip state.

  With reference to FIG. 16A, a powder 56 mainly containing Cu is disposed on a substrate 54. The powder 56 is formed using an atomizing method. The average particle diameter of the powder 56 is, for example, 3 μm. Referring to FIG. 16B, the surface of the metal layer 22 and the surface of the substrate 54 are brought into contact with each other. The powder 56 is fixed to the surface of the metal layer 22 by applying ultrasonic waves. The ultrasonic output is 100 kHz and the amplitude is 2 μm. With reference to FIG. 16C, the base body 20 is peeled from the substrate 54. A powder 32 mainly containing Cu is fixed on the metal layer 22.

  Referring to FIG. 17A, a circuit board is prepared as the base 10. The substrate 10 mainly contains BT (Bismaleimide-Triazine) resin and has a film thickness of, for example, 0.35 mm. An electrode 14 is formed on the base 10. The electrode 14 mainly includes Cu and is disposed so as to correspond to the electrode 24. On the electrode 14, a metal layer 12 mainly containing Cu and having a height of, for example, 15 μm is formed. The metal layer 12 is a bump. An electrode 58 is formed on the back surface of the substrate 10. The electrode 58 and the electrode 14 are electrically connected via a wiring formed in the base 10. Using a flip chip bonder, the base 20 is placed on the base 10 and aligned. With reference to FIG. 17B, the metal layers 12 and 22 are brought into contact with each other through the powder 32. The metal layers 12 and 22 are mechanically connected through the powder 32 using ultrasonic waves.

  Referring to FIG. 18A, the surface of the powder 32 is replaced with Sn by using a substitutional electroless tin plating solution (for example, 580MJ manufactured by Ishihara Pharmaceutical Co., Ltd.) to form the powder 33. The plating process is performed at 60 ° C. for 15 minutes. Thereby, Cu of about 1 μm on the surface of the powder 32 is replaced with Sn, and Cu remains in the center of the powder 32. The powder 33 is heated. As an example, heat is applied for 3 seconds from the substrate 20 side so that the temperature of the powder 33 becomes 300 ° C. Thereby, Sn in the powder 33 is melted. Further, using a reflow furnace, heating is performed so that the peak top temperature is 240 ° C. and the time above the melting point of Sn is 30 seconds or more. Thereby, Sn and Cu in the powder 33 are alloyed, and the alloy layer 45 is formed.

  Referring to FIG. 18B, an underfill material 60 is injected between the substrates 10 and 20. The underfill material 60 mainly includes a thermosetting epoxy resin. The underfill material 60 is cured by heating at 150 ° C. for 2 hours using a thermostatic bath. With reference to FIG. 18C, solder balls 62 are formed on the electrodes 58. The solder ball 62 mainly contains Sn—Ag (tin silver). Thereby, the semiconductor device having the chip mounted on the circuit board is completed.

  In the seventh embodiment, the example in which the plurality of powders 32 are fixed to the surface of the metal layer 22 has been described. However, the plurality of powders 32 may be fixed to at least one surface of the metal layers 12 and 22. Further, although the substrate bonding method according to the third embodiment is used, the substrate bonding method according to the first to sixth embodiments may be used.

  In the first to seventh embodiments, bumps have been described as examples of the metal layers 12 and 22, but one may be a bump and the other may be a pad.

  Although the embodiments of the present invention have been described in detail above, the present invention is not limited to such specific embodiments, and various modifications and changes can be made within the scope of the gist of the present invention described in the claims. It can be changed.

10, 20 Base 12, 22 Metal layer 14, 24 Electrode 32, 33 Powder 34 Metal layer 45 Alloy layer

Claims (6)

Contacting the first metal layer formed on the first substrate and the second metal layer formed on the second substrate with a porous body containing the first metal;
After the contacting step, using a liquid containing a second metal having a higher standard electrode potential than the first metal, replacing the first metal on the surface of the porous body with a second metal;
Joining the first metal layer and the second metal layer by heating the second metal ;
A method for bonding substrates, comprising: The porous body, method of bonding substrates according to claim 1, characterized in that the plurality of powder. The method for bonding substrates according to claim 2 , further comprising a step of fixing the plurality of powders to at least one surface of the first metal layer and the second metal layer . Step method of bonding substrates according to any one of claims 1 to 3, characterized in that not replace the inside of the porous body to the second metal to replace the first metal to the second metal . Process according to claim 4, characterized in that it comprises a step of alloying the interior of said first metal and said second metal of said porous body to bonding the second metal layer and the first metal layer The method for bonding substrates as described. The contacting step includes a step of laminating three or more substrates including the first substrate and the second substrate ,
Wherein the step of the first metal layer for bonding the second metal layer 5 a second metal laminated least three base from claim 1, characterized in that it comprises the step of heating collectively The method for bonding substrates according to any one of the above.

Images