Method for bonding cu pillars, and method for manufacturing cu pillar bonded body

Abstract

This method for bonding Cu pillars includes a Cu particle layer forming step for forming a Cu particle layer composed of a plurality of Cu particles on one or both of a bonding surface of a first Cu pillar and a bonding surface of a second Cu pillar, a stacking step for stacking the first Cu pillar and the second Cu pillar on one another with the Cu particle layer interposed therebetween, and a bonding step for heating the stacked first Cu pillar and second Cu pillar while applying pressure thereto in the stacking direction, to perform solid phase diffusion bonding of the first Cu pillar and the second Cu pillar, wherein, in the bonding step, a deformation amount L (μm), in the stacking direction, before and after bonding satisfies the relationship L≥1.4 μm, and the deformation amount L and a bonding surface area A (μm2) of the first Cu pillar and the second Cu pillar satisfy the relationship L / A≥0.0025.

Description

Cu pillar bonding method and Cu pillar bonded body manufacturing method

[0001] The present invention relates to a Cu pillar joining method for joining a first Cu pillar and a second Cu pillar, and a method for manufacturing a Cu pillar joined body.

[0002] In recent years, the performance of semiconductor devices has been improving, and micro-bonding technology has become increasingly important. Flip-chip mounting is widely used as a mounting technology for IC chips, and as shown in Patent Document 1, for example, a method has been proposed in which a solder layer is formed on a protruding electrode and the electrodes are bonded by solder. However, since mounting technology using solder has limitations in terms of achieving a narrower pitch, a Cu-Cu bonding technology has been proposed as a solderless bonding technology, as shown in Patent Document 2, for example, in which Cu pillars are bonded together by solid-phase diffusion.

[0003] JP 2018-046148 A JP 2020-053569 A

[0004] Incidentally, when performing Cu-Cu bonding, it is necessary to strictly control the shape of the Cu pillar to be bonded. Unlike solder bonding, Cu-Cu bonding does not generate a liquid phase during the bonding process, and therefore, particularly, improving the height uniformity of the Cu pillars and achieving high flatness of the bonding surface are necessary. That is, in solder bonding, the presence of a liquid phase allows the liquid phase to fill in differences in Cu pillar height and unevenness on the bonding surface. On the other hand, in the case of Cu-Cu bonding, since it is a solid-state bonding, differences in Cu electrode height and unevenness on the bonding surface cannot be filled in, and there is a risk of poor bonding. While CMP (Chemical Mechanical Polishing) technology exists as a method for controlling the shape of the Cu pillars, it requires extremely high-precision processing and must address issues such as warping of the substrate.

[0005] The present invention has been made in consideration of the above-mentioned circumstances, and aims to provide a Cu pillar joining method that does not require strict control of the shape of the Cu pillars to be joined and that can easily and reliably join Cu pillars together, and a method for manufacturing a Cu pillar joined body.

[0006] In order to solve the above problems, the inventors conducted extensive research and discovered that by forming a porous Cu particle layer composed of nano-sized copper particles on the bonding surface of the Cu pillar, shape conformability is improved, differences in Cu pillar height and unevenness on the bonding surface can be absorbed, and Cu pillars can be bonded to each other satisfactorily.

[0007] The present invention has been made based on the above-mentioned findings, and a Cu pillar joining method according to a first aspect of the present invention is a Cu pillar joining method for joining a first Cu pillar and a second Cu pillar, the method comprising: a Cu particle layer forming step of forming a Cu particle layer composed of a plurality of Cu particles on one or both of a joining surface of the first Cu pillar and a joining surface of the second Cu pillar; a stacking step of stacking the first Cu pillar and the second Cu pillar with the Cu particle layer interposed therebetween; and a joining step of heating the stacked first Cu pillar and the second Cu pillar while applying pressure in a stacking direction to perform solid-phase diffusion bonding between the first Cu pillar and the second Cu pillar, wherein in the joining step, a deformation amount L (μm) in the stacking direction before and after joining is 1.4 μm or more, and a joining area A (μm ２ ) and L / A≧0.0025.

[0008] According to the Cu pillar bonding method of aspect 1 of the present invention, the method includes a Cu particle layer forming step of forming a Cu particle layer on one or both of the bonding surface of the first Cu pillar and the bonding surface of the second Cu pillar, a stacking step of stacking the first Cu pillar and the second Cu pillar with the Cu particle layer interposed therebetween, and a bonding step of heating the stacked first Cu pillar and the second Cu pillar while applying pressure in the stacking direction, wherein in the bonding step, a deformation amount L (μm) in the stacking direction before and after bonding is L≧1.4 μm, and a bonding area A (μm ２ ) and L / A≧0.0025, the Cu particle layer is sufficiently deformed when pressure is applied in the bonding process, thereby absorbing differences in height of the Cu pillars and unevenness of the bonding surface, and making it possible to reliably bond the first Cu pillar and the second Cu pillar.

[0009] According to the Cu pillar bonding method of aspect 2 of the present invention, in the Cu particle layer forming step, the Cu particle layer is formed by a direct plating method or a dealloying method, and the bonding area A (μm ２ ) is 260 μm ２ The bonding area A (μm ２ ) is 260 μm ２ In addition, the first Cu pillar and the second Cu pillar can be bonded together in an air atmosphere or a formic acid atmosphere.

[0010] According to the Cu pillar bonding method of Aspect 3 of the present invention, the Cu particle layer is formed by a direct plating method in the Cu particle layer forming step, and the bonding step is performed in an air atmosphere or a formic acid atmosphere. The air atmosphere makes it easier to perform the bonding step, and the formic acid atmosphere ensures reliable bonding of the first Cu pillar and the second Cu pillar due to an oxidation-reduction effect.

[0011] A Cu pillar joining method of Aspect 4 of the present invention is characterized in that, in the Cu pillar joining method of Aspect 1, the Cu particle layer has an average porosity in the range of 5% to 71%. According to the Cu pillar joining method of Aspect 4 of the present invention, the Cu particle layer has an average porosity in the range of 5% to 71%, so that the Cu particle layer is sufficiently deformed in the joining step, and the strength of the Cu particle layer is ensured, making it possible to reliably join the first Cu pillar and the second Cu pillar.

[0012] A Cu pillar joining method of Aspect 5 of the present invention is characterized in that, in the Cu pillar joining method of Aspect 1 or Aspect 2, the Cu particle layer has a Vickers hardness of 95 HV or less. According to the Cu pillar joining method of Aspect 5 of the present invention, since the Cu particle layer has a Vickers hardness of 95 HV or less, the Cu particle layer is sufficiently deformed in the joining step, thereby ensuring sufficient shape followability and enabling the first Cu pillar and the second Cu pillar to be reliably joined.

[0013] A Cu pillar joining method according to Aspect 6 of the present invention is characterized in that, in the Cu pillar joining method according to any one of Aspects 1 to 3, the pressure load in the joining step is 30 MPa or less. According to the Cu pillar joining method according to Aspect 6 of the present invention, the pressure load in the joining step is 30 MPa or less, so that the occurrence of shape defects in the first Cu pillar and the second Cu pillar after joining can be suppressed.

[0014] A Cu pillar joining method of Aspect 7 of the present invention is characterized in that in the Cu pillar joining method of any one of Aspects 1 to 4, the joining temperature in the joining step is 370° C. or less. According to the Cu pillar joining method of Aspect 7 of the present invention, the joining temperature in the joining step is 370° C. or less, so that deterioration of the substrates on which the first Cu pillar and the second Cu pillar are formed after joining can be suppressed.

[0015] A Cu pillar bonding method according to Aspect 8 of the present invention is characterized in that the load application time in the bonding step is 10 minutes or less in the Cu pillar bonding method according to any one of Aspects 1 to 5. According to the Cu pillar bonding method according to Aspect 8 of the present invention, the load application time in the bonding step is 10 minutes or less, and therefore an improvement in throughput can be expected.

[0016] A Cu pillar bonding method according to Aspect 9 of the present invention is the Cu pillar bonding method according to any one of Aspects 1 to 6, characterized in that in the Cu particle layer forming step, the Cu particle layer is formed by electroplating. According to the Cu pillar bonding method according to Aspect 9 of the present invention, since the Cu particle layer is formed by electroplating in the Cu particle layer forming step, a Cu particle layer composed of a plurality of Cu particles can be reliably formed. Furthermore, by controlling the electroplating conditions, it is possible to adjust the average porosity and Vickers hardness of the Cu particle layer.

[0017] The manufacturing method of a Cu pillar bonded body of aspect 10 of the present invention is a manufacturing method of a Cu pillar bonded body having a structure in which a first Cu pillar and a second Cu pillar are bonded, and is characterized in that the first Cu pillar and the second Cu pillar are bonded by any one of the Cu pillar bonding methods of aspects 1 to 7.

[0018] According to the manufacturing method of a Cu pillar bonded body of aspect 10 of the present invention, the first Cu pillar and the second Cu pillar are bonded by any one of the Cu pillar bonding methods of aspects 1 to 7, so that the first Cu pillar and the second Cu pillar can be reliably bonded, and a Cu pillar bonded body with excellent bonding reliability can be obtained.

[0019] According to the present invention, it is possible to provide a Cu pillar joining method that does not require strict control of the shape of the Cu pillars to be joined and that can easily and reliably join Cu pillars together, and a method for manufacturing a Cu pillar bonded body.

[0020] Fig. 1 is a cross-sectional view of a Cu pillar bonded body according to one embodiment of the present invention; Fig. 2 is a flow diagram of a Cu pillar bonding method (a method for manufacturing a Cu pillar bonded body) according to one embodiment of the present invention; Fig. 3 is a cross-sectional view of each step of a Cu pillar bonding method (a method for manufacturing a Cu pillar bonded body) according to one embodiment of the present invention; Fig. 4 is a cross-sectional view illustrating a Cu particle layer formed in the Cu pillar bonding method (a method for manufacturing a Cu pillar bonded body) according to one embodiment of the present invention.

[0021] Hereinafter, a Cu pillar bonding method (a method for manufacturing a Cu pillar bonded body) according to an embodiment of the present invention will be described with reference to the drawings.

[0022] The Cu pillar bonding method (method for manufacturing a Cu pillar bonded structure) according to this embodiment is for bonding protruding electrodes (Cu pillars) formed on semiconductor chips in a semiconductor device. As shown in Fig. 1, the pillar bonded structure 1 according to this embodiment has a structure in which a first Cu pillar 11 formed on a first semiconductor chip 10 and a second Cu pillar 21 formed on a second semiconductor chip 20 are bonded via a bonding layer 30. The thickness t1 of the bonding layer 30 is preferably in the range of 0.6 µm to 18 µm.

[0023] Next, a Cu pillar bonding method (a method for manufacturing a Cu pillar bonded body) according to this embodiment will be described with reference to Fig. 2 to Fig. 4. As shown in Fig. 2, the Cu pillar bonding method (a method for manufacturing a Cu pillar bonded body) according to this embodiment includes a Cu particle layer forming step S01, a stacking step S02, and a bonding step S03.

[0024] (Cu Particle Layer Forming Step S01) First, as shown in FIG. 3 , in the Cu particle layer forming step S01, a Cu particle layer 40 is formed on one or both of the bonding surface of the first Cu pillar 11 and the bonding surface of the second Cu pillar 21. In FIG. 3 , Cu particle layers 40 (40A, 40B) are formed on both the bonding surface of the first Cu pillar 11 and the bonding surface of the second Cu pillar 21. Here, the bonding surface of the first Cu pillar 11 is the surface that bonds with the Cu particle layer 40 (40A) of the first Cu pillar 11, and the bonding surface of the second Cu pillar 21 is the surface that bonds with the Cu particle layer 40 (40B) of the second Cu pillar 21. Here, as shown in FIG. 4 , the Cu particle layer 40 is configured in the form of an aggregate in which a plurality of Cu particles 41 are stacked. Holes are formed between the Cu particles 41, and the Cu particle layer 40 has a porous structure.

[0025] In this embodiment, the average porosity P ａｖｅ It is preferable that the average porosity P ａｖｅOn the other hand, when the variation in height of the first Cu pillar or the second Cu pillar is 5% or more, the average porosity P ａｖｅ may be 45% or more, preferably 50% or more, and more preferably 60% or more. When the average porosity of the Cu particle layer 40 is 45% or more, it is possible to crush the porous structure and perform bonding even if the load during bonding is reduced, which is preferable. The pillar height variation (%) is calculated as follows by measuring the heights of multiple pillars using a laser microscope or the like: (pillar height variation (%)) = (maximum height - minimum height) / (2 x average height) x 100 Average porosity P ａｖｅ is the average value of the porosity P calculated as follows. In this embodiment, the porosity P is calculated at three locations, and the average value of the porosity P is taken as the average porosity P ａｖｅ is calculated.

[0026] The porosity P is calculated by determining the total area S1 of the Cu particle layer 40 and the area S2 of the pores in the Cu particle layer 40 through image analysis of the cross section of the Cu particle layer 40 using a scanning electron microscope, and then using the following formula: Porosity P (%) = (S2 / S1) × 100

[0027] In this embodiment, the Cu particle layer 40 preferably has a Vickers hardness of 95 HV or less.

[0028] Furthermore, in this embodiment, the thickness of the Cu particle layer 40 is preferably in the range of 2 μm to 20 μm. The thickness of the Cu particle layer 40 is more preferably 3 μm or more, and even more preferably 4 μm or more. On the other hand, the thickness of the Cu particle layer 40 is more preferably 18 μm or less, and even more preferably 16 μm or less.

[0029] Here, a description will be given of a method for forming the Cu particle layer 40. Examples of the method for forming the Cu particle layer 40 include a dealloying method and a direct plating method.

[0030] In the dealloying method, copper and a metal species electrochemically less noble than copper are co-deposited by electroplating on the bonding surfaces of the first Cu pillar 11 and the second Cu pillar 21 to form a copper alloy plating film, and then the less noble metal species in the copper alloy plating film are dealloyed to form a copper particle layer 40 with a porous structure. Here, less noble metal species are metals with a higher ionization tendency than copper, including zinc, iron, magnesium, etc. In this dealloying method, by controlling the deposition ratio and deposition form of copper and the less noble metal species, respectively, it is possible to form a Cu particle layer 40 with a desired porosity and shape. In addition, in the dealloying method, before copper alloy plating, the bonding surfaces of the first Cu pillar 11 and the second Cu pillar 21 are subjected to a hydrophilization treatment and a 10% H pretreatment as pretreatment. ２ SO ４ It is preferable to carry out pickling treatment using

[0031] Next, copper alloy plating and dealloying methods will be described in detail. Copper alloy plating is formed using a copper-zinc alloy plating solution containing, for example, copper salt, zinc salt, additives that control the deposition of copper and zinc, and a solvent. This copper alloy plating can be performed by electroless plating or electrolytic plating. Metal species that are electrochemically less noble than copper (e.g., Fe, Mn, etc.) can also be selected as alloy species.

[0032] The copper ion concentration of the copper-zinc alloy plating solution is preferably in the range of 0.0025 mol / L to 0.1 mol / L, and the zinc ion concentration is preferably in the range of 0.1 mol / L to 0.8 mol / L. The zinc ion concentration is made higher than the copper ion concentration because copper is preferentially deposited over zinc due to the difference in standard oxidation-reduction potential. The pH of the plating solution is preferably 2.5 or higher to adjust the deposition balance between copper and zinc. Furthermore, the cathode current density is set to 0.3 A / dm ２ 0.8A / dm or more ２ Set within the following range.

[0033] The copper ion and zinc ion sources for copper alloy plating can be copper salts and zinc salts known as metal ion sources for plating systems. Examples include sulfates, pyrophosphates, acetates, chlorides, and sulfamates. Trisodium citrate and potassium pyrophosphate are used as conductive and supporting salts as additives for controlling the deposition of copper and zinc to form copper-zinc alloy plating films with smooth surfaces. Brighteners can include surfactants such as amino acids, compounds selected from their salts, and alkanolamines. An example of a surfactant is (ethylenedinitrilo)tetrakis(2-propanol). Amino acids can be used as long as they are water-soluble and do not cause precipitation with copper salts (copper ions) or zinc salts (zinc ions) at any concentration. Examples include glycine, serine, alanine, tyrosine, aspartic acid, glutamic acid, histidine, and the like, or their respective salts.

[0034] The dealloying of the formed copper-zinc alloy plating film can be achieved by, for example, an etching reaction using a chemical solution or an electrochemical anodic reaction. In this embodiment, acid dealloying is performed by immersing and stirring the copper alloy film in a solution containing hydrochloric acid at a concentration of 0.002 mol / L to 0.5 mol / L at a temperature ranging from 20°C to 35°C for 30 minutes or longer, depending on the thickness of the plating film, to remove zinc from the copper-zinc alloy plating film. This results in the formation of a copper particle layer 40 with a porous structure composed of copper particles on the bonding surfaces of the first Cu pillar 11 and the second Cu pillar 21. It is preferable that the dealloying be performed so that the zinc concentration in the copper particle layer 40 after dealloying is 0.6 at% or less, as measured by energy dispersive X-ray analysis (EDX).

[0035] Next, the formed Cu particle layer 40 is washed with a cleaning solvent such as ethanol, water, or acetone, and then dried in the atmosphere using dry air. In order to prevent surface oxidation, it is preferable to immerse the layer 40 in a rust inhibitor containing benzotriazole and a surfactant as its main components for a predetermined period of time.

[0036] In the direct plating method, a Cu plating solution containing an azole-based additive, which is a copper ion electrodeposition inhibitor, is used to perform electroplating on the bonding surfaces of the first Cu pillar 11 and the second Cu pillar 21, thereby forming a porous copper particle layer 40. In this direct plating method, by controlling the type and content of the additive contained in the Cu plating solution and the plating conditions, it is possible to form a Cu particle layer 40 with a desired porosity and shape.

[0037] The Cu plating solution used is an acidic electrolytic copper plating solution containing a soluble copper salt, an azole compound having two to three nitrogen atoms in a five-membered ring, which is a copper ion electrodeposition inhibitor represented by the following formulas (1) to (4), an acid, and water. If necessary, a brightener, a surfactant, an antioxidant, etc. may also be added. The Cu plating solution used has a copper concentration of 0.1 mol / L or more, an azole compound concentration of 10 mmol / L or more and 50 mmol / L or less, and a chlorine chloride ion concentration of 10 ppm or less.

[0038]

[0039] In the above formulas (1) to (4), R １ ~R ４ may be the same or different from each other, and are any of an alkyl group having 10 or less carbon atoms, an alkenyl group having 10 or less carbon atoms, an alkynyl group having 10 or less carbon atoms, an aryl group having 10 or less carbon atoms, an aralkyl group having 10 or less carbon atoms, or an alkoxy group having 10 or less carbon atoms, or a group in which the hydrogen atom of any of these groups is substituted with a halogen atom, a hydroxyl group, a carboxyl group, an amino group, an alkyl-substituted amino group having 5 or less carbon atoms, a hydroxyalkyl-substituted amino group having 5 or less carbon atoms in the alkyl chain, or a mercapto group, or any of an amino group, an alkyl-substituted amino group having 5 or less carbon atoms, a hydroxyalkyl-substituted amino group having 5 or less carbon atoms in the alkyl chain, a mercapto group, a hydroxyl group, a carboxyl group, a halogen atom, or a hydrogen atom.

[0040] Specific examples of soluble copper salts include copper sulfate, copper oxide, copper carbonate, copper alkanesulfonates such as copper methanesulfonate and copper propanoate, copper alkanolsulfonates such as copper isethionate and copper propanolsulfonate, and copper organic acids such as copper acetate, copper citrate, and copper tartrate, which may be used alone or in combination of two or more.

[0041] Furthermore, the acid may be an organic acid or an inorganic acid. Examples of these include sulfuric acid, alkanesulfonic acids such as methanesulfonic acid and propanesulfonic acid, alkanolsulfonic acids such as isethionic acid and propanolsulfonic acid, and organic acids such as citric acid and tartaric acid. These may be used alone or in combination of two or more. The water may be pure water such as ion-exchanged water or distilled water.

[0042] The plating conditions are, for example, a DC power supply with a current density of 0.1 A / dm ２ ~5 A / dm ２ Approximately, preferably 0.4 A / dm ２ ~1.0 A / dm ２ It is preferable to set the current density to 0.4 A / dm and select a plating time that allows the desired copper particle layer 40 to be formed. ２ ~2.0 A / dm ２ and may be in the range of 0.4 A / dm ２ ~1.0 A / dm ２ and may be in the range of 1.0 A / dm ２ ~2.0 A / dm ２ When Cu plating is performed under the above conditions, the azole compound, which is a copper ion electrodeposition inhibitor, as well as the copper ions are adsorbed on the cathode surface, which is the joining surface of the first Cu pillar 11 and the joining surface of the second Cu pillar 21. The presence of the azole compound strongly suppresses the electrodeposition of copper ions, favoring copper nucleation, and a porous Cu particle layer 40 made of copper particles 41 is formed on the cathode surface as a copper plating film.

[0043] 3 , the first Cu pillar 11 having a Cu particle layer 40A formed on its joining surface and the second Cu pillar 21 having a Cu particle layer 40B formed on its joining surface are stacked together with the Cu particle layers 40A and 40B interposed therebetween. At this time, the thickness t0 of the Cu particle layer 40 in the stack (the total thickness of the Cu particle layers 40A and 40B) is preferably within a range of 2 μm to 20 μm.

[0044] 3 , the stacked first Cu pillar 11 and second Cu pillar 21 are heated while being pressurized in the stacking direction, thereby solid-phase diffusion bonding the first Cu pillar 11 and the second Cu pillar 21. This produces a pillar bonded body 1 in which the first Cu pillar 11 and the second Cu pillar 21 are bonded via the bonding layer 30. In this embodiment, since the Cu particle layers 40 (40A, 40B) are interposed between the first Cu pillar 11 and the second Cu pillar 21, the bonding layer 30 is formed by these Cu particle layers 40 (40A, 40B).

[0045] In the bonding step S03, the deformation amount L (μm) in the stacking direction before and after bonding is L≧1.4 μm, and the deformation amount L and the area A (μm ２ ) have a relationship of L / A≧0.0025. In this embodiment, since the Cu particle layers 40 (40A, 40B) having a porous structure are stacked via the Cu particle layers 40, when pressure is applied in the stacking direction, the Cu particle layers 40 (40A, 40B) are mainly deformed in the stacking direction. Therefore, the deformation amount L is calculated as L=t0−t1 from the thickness t1 of the bonding layer 30 and the thickness t0 of the Cu particle layers 40 (40A, 40B). The deformation amount L may be a value (a+b) obtained by subtracting the thickness Z of the stack of the first Cu pillar 11, the bonding layer 30, and the second Cu pillar 21 obtained after the bonding process from the sum (a+b) of the thickness a of the stack of the first Cu pillar 11 and the Cu particle layer 40 (40A) and the thickness b of the stack of the second Cu pillar 21 and the Cu particle layer 40 (40A).

[0046] The area A (μm ２) is obtained by measuring the shape of either the first Cu pillar 11 or the second Cu pillar 21 in a planar view using a white light interference microscope, calculating the surface area, and averaging the areas measured at three locations to obtain the area A (μm ２ The shapes of the bonding surfaces of the first Cu pillar 11 and the second Cu pillar 21 facing each other are substantially the same, and the area A (μm ２ ) may be the area of ​​the bonding surface of the first Cu pillar 11 or the area of ​​the bonding surface of the second Cu pillar 21.

[0047] Furthermore, in this embodiment, in the Cu particle layer forming step, the Cu particle layer is formed by a direct plating method or a dealloying method, and the bonding area A (μm ２ ) is 260 μm ２ Preferably, it is 190 μm or less. ２ More preferably, it is 180 μm or less. ２ It is more preferable that the bonding area A (μm ２ ) is 20 μm ２ The bonding area A (μm ２ ) is 1.5 μm ２ It is preferable that this is equal to or greater than this.

[0048] The thickness a can be determined by arbitrarily selecting 10 stacks of the first Cu pillar 11 and the Cu particle layer 40 (40A), observing the cross section of each stack using an optical microscope after cross-section processing, measuring the thickness (or height) of the stack, and obtaining the average value as the thickness a. The thickness b can be determined by arbitrarily selecting 10 stacks of the second Cu pillar 21 and the Cu particle layer 40 (40B), observing the cross section of each stack using an optical microscope after cross-section processing, measuring the thickness (or height) of the stack, and obtaining the average value as the thickness b. The thickness Z can be determined by arbitrarily selecting 10 stacks of the second Cu pillar 21 and the Cu particle layer 40 (40B), observing the cross section of each stack using an optical microscope after cross-section processing, measuring the thickness (or height) of the stack, and obtaining the average value as the thickness Z.

[0049] In this embodiment, the pressure load in the joining step S03 is preferably 30 MPa or less. The pressure load in the joining step S03 is more preferably 24 MPa or less, and even more preferably 12 MPa or less. The lower limit of the pressure load in the joining step S03 is preferably 1 MPa or more, and more preferably 5 MPa or more.

[0050] Furthermore, in this embodiment, the bonding temperature in the bonding step S03 is preferably 370° C. or less. The bonding temperature in the bonding step S03 is more preferably 320° C. or less, and even more preferably 280° C. or less. The bonding temperature in the bonding step S03 is preferably 100° C. or more, and more preferably 150° C. or more.

[0051] In this embodiment, the load application time in the bonding step S03 is preferably 10 minutes or less. The load application time in the bonding step S03 is more preferably 9 minutes or less, and even more preferably 7 minutes or less. The load application time in the bonding step S03 is preferably 30 seconds or more, and more preferably 1 minute or more.

[0052] The Cu pillar bonding method (method for manufacturing a Cu pillar bonded body) of this embodiment configured as described above includes a Cu particle layer forming step S01 of forming a Cu particle layer 40 (40A, 40B) on one or both of the bonding surfaces of the first Cu pillar 11 and the second Cu pillar 21, a stacking step S02 of stacking the first Cu pillar 11 and the second Cu pillar 21 with the Cu particle layer 40 (40A, 40B) interposed therebetween, and a bonding step S03 of heating the stacked first Cu pillar 11 and the second Cu pillar 21 while applying pressure in the stacking direction. In the bonding step S03, the deformation amount L (μm) in the stacking direction before and after bonding is L≧1.4 μm, and the deformation amount L and the bonding area A (μm ２) and L / A≧0.0025, the Cu particle layer 40 (40A, 40B) is sufficiently deformed when pressurized in the bonding process S03, thereby absorbing unevenness on the bonding surfaces of the first Cu pillar 11 and the second Cu pillar 21 and differences in height between the first Cu pillar 11 and the second Cu pillar 21, and making it possible to reliably bond the first Cu pillar 11 and the second Cu pillar 21 without strictly controlling the shape.

[0053] In this embodiment, the average porosity P ａｖｅ is in the range of 5% or more and 71% or less, the Cu particle layer 40 (40A, 40B) is sufficiently deformed in the bonding step S03, and the strength of the Cu particle layer 40 (40A, 40B) is ensured, making it possible to reliably bond the first Cu pillar 11 and the second Cu pillar 21. That is, when the average porosity of the Cu particle layer 40 (40A, 40B) is 5% or more, the deformability of the Cu particle layer 40 (40A, 40B) is ensured, making it possible to sufficiently deform the Cu particle layer 40 (40A, 40B) in the bonding step S03. Furthermore, when the average porosity of the Cu particle layer 40 (40A, 40B) is 71% or less, the strength is ensured. Note that in this embodiment, in order to further ensure the deformability of the Cu particle layer 40 (40A, 40B), the average porosity P ａｖｅ On the other hand, in this embodiment, in order to further ensure the strength of the Cu particle layer 40, the average porosity P ａｖｅ is more preferably less than 45%, and even more preferably 32% or less. ａｖｅ When the average porosity P ａｖｅ When the area A (μm ２This is preferable because the porous structure can be crushed and bonding can be performed even if the bonding load is reduced and the thickness of the porous structure is reduced.

[0054] In the present embodiment, when the Vickers hardness of the Cu particle layer 40 (40A, 40B) is 95 HV or less, the Cu particle layer 40 (40A, 40B) is sufficiently deformed in the joining step S03, thereby ensuring sufficient shape conformability, and the first Cu pillar 11 and the second Cu pillar 21 can be reliably joined. In the present embodiment, to further ensure the shape conformability of the Cu particle layer 40, the Vickers hardness of the Cu particle layer 40 is more preferably 92 HV or less, and even more preferably 89 HV or less. There is no particular lower limit for the Vickers hardness of the Cu particle layer 40, but it is substantially 60 HV or more.

[0055] In this embodiment, when the pressure load in the bonding step S03 is set to 30 MPa or less, the pressure load is relatively low, which can suppress deformation of the first Cu pillar 11 and the second Cu pillar 21 during bonding and can suppress the occurrence of defective shapes of the first Cu pillar 11 and the second Cu pillar 21 after bonding. Note that in the bonding step S03, because the porous Cu particle layer 40 is interposed between the first Cu pillar 11 and the second Cu pillar 21, even if the pressure load is 30 MPa or less, the deformation amount L in the stacking direction can be ensured and the first Cu pillar 11 and the second Cu pillar 21 can be reliably bonded.

[0056] In this embodiment, when the bonding temperature in the bonding step S03 is set to 370° C. or less, the bonding temperature is relatively low, and it is possible to suppress deterioration of the substrates on which the first Cu pillar 11 and the second Cu pillar 21 are formed after bonding. Note that in the bonding step S03, the porous Cu particle layer 40 is interposed between the first Cu pillar 11 and the second Cu pillar 21, so that even if the bonding temperature is 370° C. or less, diffusion of Cu atoms can be promoted and the first Cu pillar 11 and the second Cu pillar 21 can be reliably bonded.

[0057] In this embodiment, when the load application time in the bonding step S03 is 10 minutes or less, the load application time is relatively short, which is expected to improve throughput during bonding and enable bonding of the first Cu pillar 11 and the second Cu pillar 21 after bonding in a short time. Note that in the bonding step S03, the porous Cu particle layer 40 is interposed between the first Cu pillar 11 and the second Cu pillar 21. Therefore, even if the load application time is 10 minutes or less, the deformation amount L in the stacking direction can be ensured, and diffusion of Cu atoms can be promoted, enabling reliable bonding of the first Cu pillar 11 and the second Cu pillar 21.

[0058] In this embodiment, in the Cu particle layer forming step, the Cu particle layer is formed by a direct plating method, and the bonding area A (μm ２ ) is 260 μm ２ If the following is true, the first Cu pillar 11 and the second Cu pillar 21 can be reliably bonded even if the bonding atmosphere in the bonding step S03 is air or formic acid, which is preferable; and further, if the atmosphere in the bonding step S03 is a formic acid atmosphere, the first Cu pillar 11 and the second Cu pillar 21 can be more reliably bonded, which is even more preferable.

[0059] In the present embodiment, when the Cu particle layer 40 is formed by electroplating in the Cu particle layer forming step S01, the Cu particle layer 40 composed of a plurality of Cu particles 41 can be reliably formed. In addition, by controlling the electroplating conditions, the average porosity P ａｖｅ It is also possible to adjust the Vickers hardness.

[0060] In this embodiment, pickling may be performed before the bonding step S03. If pickling is performed, even if an oxide film is formed at the bonding interface, the oxide film can be removed, which is preferable because it enables the first Cu pillar 11 and the second Cu pillar 21 to be bonded more reliably. The pickling may be performed in, for example, 10% H ２ SO ４ It is recommended to use the following.

[0061] Although the embodiment of the present invention has been described above, the present invention is not limited to this and can be modified as appropriate within the scope of the technical idea of ​​the invention.

[0062] The results of confirmation experiments conducted to confirm the effectiveness of the present invention will be described below.

[0063] First, a 100 nm thick Ti layer was formed on a silicon wafer, and then a 300 nm thick Cu layer was formed as a seed layer by sputtering. Next, bump patterns with diameters of 15 μm, 25 μm, and 5 μm were formed using photoresist. The resulting bump patterns were hydrophilized and then immersed in 10% H ２ SO ４ The steel was subjected to pickling treatment using the method described above.

[0064] Next, a copper sulfate-based bump-forming plating solution was used for plating, and Cu pillars were formed to a height of 10 μm for bump patterns with a diameter of 15 μm and 25 μm. For bump patterns with a diameter of 5 μm, Cu pillars were formed to a height of 5 μm. Thereafter, as shown in Table 1, a Cu particle layer was formed on the bonding surface of the Cu pillar for Samples 1 to 14. As shown in Table 1, the dealloying method and direct plating method were used to form the Cu particle layer. Note that plating was performed for Sample 14 using a copper sulfate-based plating solution.

[0065] In the dealloying method, a Cu-Zn alloy was plated, and after plating, a dealloying treatment was performed using 2 M HCl to form a Cu particle layer. In this embodiment, two types of plating solutions for dealloying (dealloying-1 and dealloying-2) were prepared as follows.

[0066] (Dealloying-1) Copper sulfate pentahydrate (Cu ２＋ as): 0.01 mol / L zinc sulfate heptahydrate (Zn ２＋ as citric acid): 0.25 mol / L Trisodium citrate dihydrate: 0.3 mol / L Glycine, an amino acid compound: 0.1 mol / L Ion-exchanged water: balance The pH of the plating solution was adjusted to 4.2 with citric acid.

[0067] (Dealloying-2) Copper sulfate pentahydrate (Cu２＋ as): 0.01 mol / L zinc sulfate heptahydrate (Zn ２＋ as citric acid): 0.25 mol / L Disodium citrate dihydrate: 0.3 mol / L Ion-exchanged water: balance The pH of the plating solution was adjusted to 4.5 with citric acid.

[0068] In the direct plating method, a Cu particle layer was formed by plating using a Cu plating solution containing an azole-based additive that acts strongly on the surface. In this embodiment, the following two types of plating solutions for the direct plating method (Direct Plating-1 and Direct Plating-2) were prepared.

[0069] (Direct plating-1) Copper sulfate pentahydrate (Cu ２＋ Copper ion electrodeposition inhibitor (3,5-diamino-1,2,4-triazole): 10 mmol / L Chlorine: 0 ppm Ion-exchanged water: balance pH: 2.5

[0070] (Direct plating-2) Copper sulfate pentahydrate (Cu ２＋ Copper ion electrodeposition inhibitor (3,5-diamino-1,2,4-triazole): 5 mmol / L Chlorine: 0 ppm Ion-exchanged water: balance pH: 2.5

[0071] The area of ​​the bonding surface of the formed Cu pillar was determined by measuring the shape of the surface of the Cu pillar with a white light interference microscope, calculating the surface area, and then calculating the average value of the areas measured for three bumps. The current density during plating, the thickness of the Cu particle layer, the average porosity of the Cu particle layer, and the Vickers hardness are shown in Table 1. The average porosity P of the Cu particle layer ａｖｅwas calculated as the arithmetic mean (n = 3) of the porosity (P) calculated by the following formula (1) based on the total area (S1) of the Cu particle layer calculated by image analysis of the cross section of the Cu particle layer using a scanning electron microscope and the area (S2) of the void portion in the Cu particle layer. P (%) = (S2 / S1) × 100 (1) The Vickers hardness of the Cu particle layer was measured using an ultra-microhardness tester. The thickness of the Cu particle layer was measured by processing the cross section of a laminate of Cu pillars and a Cu particle layer, viewing it from the side using an optical microscope, and determining the boundary between the Cu pillars and the Cu particle layer from the difference in particle size between the Cu pillars and the Cu particle layer, and measuring the thickness of the Cu particle layer from the boundary.

[0072]

[0073] Next, the photoresist on the formed bump pattern was removed, and then a bonding test was carried out. Before the bonding test, the pretreatment shown in Table 2 was carried out as needed. When pretreatment was carried out, it is indicated as "acid pickling treatment" in Table 2. This was carried out by pickling in 10% H ２ SO ４ The acid pickling treatment performed by acid pickling is shown in Table 2. When no pretreatment was performed, it is indicated as "no acid pickling treatment" in Table 2. The bonding test was performed using a high precision flip chip bonder T-3000-PRO manufactured by Tresky, after the top and bottom were aligned. The atmosphere in which the bonding process was performed, as well as the bonding load, bonding temperature, and bonding time (load application time) set in the bonding process are shown in Tables 2 and 3. In the bonding atmosphere shown in Table 2, "N ２ " is N ２ "Air atmosphere" indicates that the bonding step was performed in an air atmosphere, and "formic acid atmosphere" indicates that the bonding step was performed in an air atmosphere with 0.5 mL / μm of formic acid at a concentration of 88%. ２ Apply to the joining surface and then ２ This indicates that the bonding was carried out in an atmosphere.

[0074] The amount of deformation L before and after bonding was evaluated by measuring the total length of the protruding electrodes, including the Cu pillars and Cu particle layer, before bonding and the distance between the chips after bonding after cross-section processing. The total length of the protruding electrodes before bonding was determined by first measuring the heights a and b of 10 random bumps on each side and calculating the average height. The distance between the chips after bonding was determined by cross-section processing of the bonded body and calculating the average value of the cross-sectional height (Z) measured at n = 5. The value of (a + b) - Z was calculated as the amount of deformation L.

[0075] To evaluate the bonded structure, the shear strength (bond strength) of the bonded structure (pillar bonded structure) was measured. Measurements were performed using a shear strength evaluation tester (Bond Tester; Dage Series 4000, manufactured by Nordson Advanced Technologies, Inc.). Specifically, shear strength was measured as follows: The bonded structure was fixed horizontally, and a shear tool was used to press the chip-attached Si wafer horizontally from the side at a position 50 μm above the surface (top surface) of the bonding layer. The strength measured when the chip broke was measured. A pattern with 8,100 bumps within the chip was used for bonding, and the bond strength per bump was calculated by dividing the obtained bond strength by the total area of ​​all 8,100 bumps. The shear tool movement speed was 0.1 mm / sec. Three strength tests were performed per condition, and the arithmetic average of the results was used as the measured bond strength. The results are shown in Table 3.

[0076] Based on the measured bonding strength of the pillar junction, the bonding state of the pillar junction was evaluated as follows. The results are shown in Table 3. A: Excellent bonding state (bonding strength of 3.0 MPa or more) B: Good bonding state (bonding strength of 2.0 MPa or more, less than 3.0 MPa) C: Poor bonding state, difficult to use as a product (bonding strength less than 2.0 MPa)

[0077] In Examples 1 to 5, 6-1, 7 to 10, 11-1, 12-1, and 13 of the present invention and Comparative Examples 1 to 4, Cu particle layers were formed on the Cu pillars on both the top and bottom, and are shown as "both porous" in Table 2. In Examples 6-2, 11-2, and 12-2 of the present invention, only Cu pillars were used on the bottom side, and Cu particle layers were formed on the Cu pillars on the top side, and are shown as "porous on only one side" in Table 2. In addition, to facilitate shear strength measurements during bonding, the Si chips were diced and mounted so that the top side was 6 mm square and the bottom side was 10 mm square.

[0078]

[0079]

[0080] In Comparative Example 1, in the bonding process, the deformation amount L (μm) in the stacking direction before and after bonding was set to 1.0 μm, and the deformation amount L and the bonding area A (μm ２ In Comparative Example 2, the ratio L / A of the thickness of the pillar bonded structure to the thickness of the pillar bonded structure was 0.0020, and the bonding strength of the pillar bonded structure was 0.9 MPa. In Comparative Example 2, the deformation amount L (μm) in the stacking direction before and after bonding in the bonding step was 1.1 μm, and the bonding strength of the pillar bonded structure was 1.5 MPa.

[0081] In Comparative Example 3, the deformation amount L (μm) in the stacking direction before and after bonding in the bonding process was set to 0.8 μm, and the deformation amount L and the bonding area A (μm ２ In Comparative Example 4, the ratio L / A of the porous Cu particle layer to the Cu particle layer was 0.0016, and the bonding strength of the pillar bonded structure was 0.8 MPa. In Comparative Example 4, no porous Cu particle layer was formed, and the bonding strength of the pillar bonded structure was 0.5 MPa.

[0082] In contrast, in the present invention example, in the bonding process, the deformation amount L (μm) in the stacking direction before and after bonding is L≧1.4 μm, and the deformation amount L and the bonding area A (μm ２) was set to 0.0025 or more, and the bonding strength of the pillar joint body was 2.4 MPa or more, which was improved compared to the comparative examples. In particular, in Inventive Examples 1 to 6-2, whether the plating method was dealloying or direct plating, a pretreatment of pickling treatment was performed, and N ２ In particular, according to the method of the present invention, as shown in Inventive Examples 4 and 5, pillar bonded bodies having stable and excellent bonding strength were obtained even when the bonding surface area was small.

[0083] In Example 7 of the present invention, although the area of ​​the bonding surface was small, a porous Cu particle layer was formed because a direct plating method was used, and the bonding strength of the pillar bonded assembly was 5.8 MPa. ａｖｅ In Example 9 of the present invention, although the area of ​​the bonding surface was small, the average porosity P ａｖｅSince the bonding strength was 70%, a pillar-bonded structure with excellent bonding strength could be obtained even when the pressure load in the bonding process was reduced to 8 MPa. In Inventive Example 10, the Cu particle layer was formed using a direct plating method using 3,5-diamino-1,2,4-triazole as a copper ion electrodeposition inhibitor, which prevented oxidation of the bonding interface and allowed a pillar-bonded structure to be obtained without performing a pretreatment such as pickling. In Inventive Examples 11-1 and 11-2, the Cu particle layer was formed using a direct plating method using 3,5-diamino-1,2,4-triazole as a copper ion electrodeposition inhibitor, which prevented oxidation of the bonding interface and allowed a pillar-bonded structure to be obtained even in an air atmosphere. In Inventive Examples 12-1 and 12-2, bonding was performed in a formic acid atmosphere, which allowed maximum utilization of the sintering effect of the porous structure of the Cu particle layer, tending to improve bonding strength. In particular, even in the case of a "one-sided porous" state in which only a Cu pillar was used on the lower side and a Cu particle layer was formed on the Cu pillar on the upper side, as in Inventive Example 12-2, the bonding strength tended to improve. On the other hand, as shown in Inventive Example 13, even when a direct plating method was used, when a pillar-joined body was obtained by performing bonding in an air atmosphere without performing a pretreatment of pickling, a pillar-joined body could be obtained, but the bonding strength tended to decrease.

[0084] From the results of the above confirmation experiments, it was confirmed that the present invention can provide a Cu pillar joining method that does not require strict control of the shape of the Cu pillars to be joined and that can easily and reliably join Cu pillars together, as well as a method for manufacturing a Cu pillar bonded body.

[0085] REFERENCE SIGNS LIST 1 Cu pillar bonded body 11 First Cu pillar 21 Second Cu pillar 30 Bonding layer 40 Cu particle layer 41 Cu particles S01 Cu particle layer forming step S02 Stacking step S03 Bonding step

Claims

1. A Cu pillar joining method for joining a first Cu pillar and a second Cu pillar, comprising: a Cu particle layer forming step of forming a Cu particle layer composed of a plurality of Cu particles on one or both of a joining surface of the first Cu pillar and a joining surface of the second Cu pillar; a stacking step of stacking the first Cu pillar and the second Cu pillar via the Cu particle layer; and a joining step of heating the stacked first Cu pillar and the second Cu pillar while applying pressure in the stacking direction, thereby solid-phase diffusion bonding the first Cu pillar and the second Cu pillar together, wherein in the joining step, a deformation amount L (μm) in the stacking direction before and after joining is L≧1.4 μm, and a relationship between the deformation amount L and a joining area A (μm ２ ) and L / A≧0.0025.

2. In the Cu particle layer forming step, the Cu particle layer is formed by a direct plating method or a dealloying method, and the bonding area A (μm ２ ) is 260 μm ２ 2. The Cu pillar bonding method according to claim 1, wherein:

3. The Cu pillar bonding method according to claim 2, characterized in that in the Cu particle layer forming step, the Cu particle layer is formed by a direct plating method, and the atmosphere in the bonding step is an air atmosphere or a formic acid atmosphere.

4. The method for joining Cu pillars according to claim 1, wherein the average porosity of the Cu particle layer is within the range of 5% to 71%.

5. The Cu pillar joining method according to claim 1, wherein the Cu particle layer has a Vickers hardness of 95 HV or less.

6. The Cu pillar joining method according to claim 1, wherein the pressure load in the joining step is 30 MPa or less.

7. The Cu pillar bonding method according to claim 1, wherein the bonding temperature in the bonding step is 370°C or less.

8. The Cu pillar bonding method according to claim 1, wherein the load application time in the bonding step is 10 minutes or less.

9. The Cu pillar bonding method according to claim 1, wherein in the Cu particle layer forming step, the Cu particle layer is formed by electroplating.

10. A method for manufacturing a Cu pillar bonded body having a structure in which a first Cu pillar and a second Cu pillar are bonded, characterized in that the first Cu pillar and the second Cu pillar are bonded by the Cu pillar bonding method described in any one of claims 1 to 7.

Images