JP5557158B2 - Flip chip connecting underfill agent and method of manufacturing semiconductor device using the same - Google Patents

Description

  The present invention relates to an underfill agent used for flip chip connection, a method of manufacturing a semiconductor device using the same, and a semiconductor device manufactured by the method.

  In the flip chip connection method, a surface of a silicon chip on which an electronic circuit is formed (hereinafter referred to as a “circuit surface”) faces the substrate side, and the chip is connected to the substrate through solder bumps provided on the circuit surface. It is a method to do. The underfill agent is applied between the silicon chip and the substrate in order to prevent cracks in the solder connection portion and ensure the reliability of the device.

  Conventionally, the underfill agent is filled between the substrate and the silicon chip by utilizing capillary action. However, some die sizes have a side larger than 10 mm, and more than 20 mm. In a flip chip semiconductor device using such a large die, unfilling is likely to occur. When the amount of the filler in the underfill agent is reduced, it becomes easier to fill, but since the thermal expansion coefficient of the underfill agent increases, there is a problem that peeling is likely to occur at the interface between the underfill agent and the chip or the substrate. Occur.

  Furthermore, the method of utilizing the capillary phenomenon has many man-hours, causing a decrease in productivity and an increase in manufacturing cost. Therefore, when connecting a silicon chip to a substrate, a so-called non-flow type underfill method is proposed in which an underfill agent mixed with a flux agent in advance is dropped onto the substrate and then the underfill agent is cured simultaneously with soldering. (Patent Document 1). According to this method, the productivity can be increased and the manufacturing cost can be reduced by reducing the number of steps.

  The fluxing agent is a proton donating substance such as abietic acid and reduces the oxide film on the surface of the solder bump. However, water may be generated during the reduction reaction, and the water may become water vapor in the solder reflow process, generating voids in the connection portion.

  A method of applying an underfill agent to a silicon wafer to form a B-stage and then dicing is disclosed (Patent Document 2). This is a so-called B-stage underfill or wafer level underfill.

  Furthermore, as an improved method, it has been proposed to use two layers of underfill agent, an underfill agent having a flux action around the solder bumps, and an underfill agent containing a filler on the bonding pad of the substrate. (Patent Document 3). However, a process of applying an underfill agent to each of the wafer and the substrate is necessary and cumbersome. Also, the flux component on the pad may be insufficient, resulting in poor connection. Furthermore, the layer containing a large amount of filler has poor transparency, and when applied beyond the height of the solder bumps, it becomes difficult to specify the position of the solder bumps, that is, the visibility of the solder bumps deteriorates, Positioning during dicing or connection becomes difficult. Further, the underfill agent having a flux action covers the periphery of the solder bump, and the above-described problem of void generation due to water vapor is likely to occur.

  A method is known in which the amount of underfill agent applied is set to a thickness that allows the tip of the solder ball to protrude, and the flux is applied to the tip of the solder ball after being previously B-staged (Patent Document 4). However, in this method, an underfill agent is applied to each package, followed by provisional curing, and then a flux must be applied to each individual package, resulting in a complicated process. Moreover, when the flux to be applied and the underfill agent react with each other, the molar ratio between the main agent of the underfill agent and the curing agent is shifted, and the physical properties of the cured product of the underfill agent may be reduced.

  In addition, in wafer level underfill, a method has been proposed in which a hot melt type underfill agent is ground until the top of the bump is exposed, and then dicing is performed (Patent Document 5). However, in these methods, voids are likely to be formed in the connecting portion due to cure shrinkage of the underfill agent, and a fillet covering the periphery of the chip is also difficult to form. Furthermore, an oxide film is easily formed on the exposed solder portion. In the connecting step, a commercially available flux agent is applied to a substrate or the like, but a normal commercially available flux easily contains voids because it contains a large amount of a solvent as a diluent.

  As a coating method of the underfill agent used in the above-described conventional technology, a film method, a spin coating method, a printing method, and the like are common. Although the film method and the printing method have good workability and are practical, it is difficult to form a thin film, particularly a thin film having a thickness of 20 μm or less. On the other hand, the spin coat method is effective for forming a thin film, but has a problem in terms of effective use of resources and adverse effects on the environment because the ratio of the amount of waste liquid to the amount of liquid required for forming a thin film is large.

  Conventionally, fused silica, crystalline silica, alumina, titanium oxide, silica titania, boron nitride, aluminum nitride, silicon nitride, magnesia, magnesium silicate, aluminum, etc. have been used as inorganic fillers for underfill agents (patents) Reference 6). However, the hardness of these inorganic fillers is high in the dicing process for individualizing the semiconductor wafer after the B stage and the polishing process (polishing process) for adjusting the thickness of the applied underfill agent layer. The blades of the dicing device are easily damaged and are frequently replaced, resulting in high costs. Further, since polishing must be performed at a high polishing pressure, there are problems that the semiconductor element is damaged and the wafer itself is broken.

Japanese Patent Laid-Open No. 04-280443 Japanese Patent Laid-Open No. 2000-174044 JP 2003-243449 A JP 2005-268704 A JP 2006-229199 A JP 2009-135308 A

  Accordingly, an object of the present invention is to provide an underfill agent composition capable of stably forming a thin film having a large area and providing good and durable solder connectivity as an underfill agent, and a semiconductor device using the composition. It is to provide a manufacturing method.

  As a result of intensive studies, the present inventors have found that the above-described object can be achieved by using an epoxy resin composition that includes a specific spherical porous inorganic filler and is compatible with the B-stage.

That is, the present invention, first, as a means for achieving the above object,
(A) epoxy resin,
(B) a curing agent having flux properties;
(C) an imidazole catalyst,
(D) A volatile organic solvent, and (E) a porous inorganic filler having a pore volume of 0.3 to 0.7 cm ³ / g, and a spherical inorganic filler whose surface is coated with silicon oxide. A liquid wafer level underfill composition capable of being B-staged is provided.

In addition, the present invention secondly,
A method of manufacturing a semiconductor device comprising a substrate and a silicon chip flip-chip connected to the substrate,
(1) Spray application of the above-mentioned liquid wafer level underfill composition (underfill agent) compatible with the B-stage on the surface of the silicon wafer provided with solder bumps for flip chip connection. And then forming a B-stage underfill agent layer having a thickness of 1.0 to 1.3 times the height of the solder bump by converting the applied underfill agent into a B-stage,
(2) A step of cutting the silicon wafer on which the underfill agent layer in the B-stage state is formed into pieces into silicon chips,
(3) The step of positioning the piece obtained in step (2) at the position to be connected on the substrate, and (4) melting the piece of the solder bump positioned in step (3). The process of connecting to the substrate,
A method for manufacturing a semiconductor device including the following (hereinafter referred to as “manufacturing method 1”) is provided.

Furthermore, the present invention thirdly,
A method of manufacturing a semiconductor device comprising a substrate and a silicon chip flip-chip connected to the substrate,
(I) The above-mentioned liquid wafer level underfill composition that can be applied to the B-stage is used as the first underfill agent on the surface of the silicon wafer provided with solder bumps for flip-chip connection. Spraying and then forming the applied underfill agent into a B-stage to have a thickness of 0.5 to 1.0 times the height of the solder bump;
(Ii) A second underfill liquid that is different from the first underfill agent is spray-applied on the first underfill agent layer in the B-stage state, and then the applied second underfill is applied. The second underfill agent layer in the B-stage state by converting the agent into a B-stage is such that the total thickness of the first and second underfill agent layers is 1.0 to 1.3 times the height of the solder bump. Forming a process to be
(Iii) a step of cutting the silicon wafer on which the B-stage underfill agent layer is formed in steps (i) and (ii) into individual silicon chips;
(Iv) a step of positioning the individual pieces obtained in step (iii) at a position to be connected on the substrate, and (v) melting the solder bumps of the individual pieces positioned in step (iv). The process of connecting to the substrate,
A method for manufacturing a semiconductor device including the following (hereinafter referred to as “manufacturing method 2”) is provided.

According to the present invention, the following effects can be obtained.
-Since the composition of the present invention contains a specific spherical porous inorganic filler, no increase in viscosity is observed compared to conventional porous silica even when the composition is highly filled.
-Since the composition is excellent in mechanical strength after curing, it is an underfill material excellent in ability to protect a semiconductor element from impacts such as a temperature cycle.
-Since the composition of the present invention has low water absorption, it is a highly reliable underfill material in which package cracking or interface peeling hardly occurs in the solder reflow process during mounting.
The composition of the present invention is suitable for forming a thin film as an underfill agent, and can cope with the underfill of flip chip devices having a narrow gap or fine pitch.

According to the semiconductor device manufacturing method of the present invention, a thin film can be formed efficiently and stably even in a large area of a large wafer. In particular, a thin film having a thickness of 20 μm or less can be easily formed.
-According to the manufacturing method of the present invention, since the B-stage underfill layer is 1.0 to 1.3 times the height of the solder bump, there is no gap due to curing shrinkage as in the prior art.
-The composition of the present invention may contain a solvent, but it is removed in the B-stage stage and there is no risk of void formation in the B-stage stage.
-The spray coating used in the production method of the present invention is environmentally friendly because the amount of solvent used and the amount of waste can be reduced compared to the spin coating method.

Furthermore, according to the method of the present invention, the underfill agent can be continuously applied on the wafer, and there is no problem in the alignment of the solder bumps.
-Since the composition of this invention contains the hardening | curing agent which has flux property, it can form favorable connection with a solder bump and the pad on a board | substrate, without using a flux agent separately.

  Hereinafter, the present invention will be described in detail according to embodiments.

-Underfill agent composition-
The underfill agent composition of the present invention is used as an underfill agent in production method 1 and as a first underfill agent in production method 2. Hereinafter, the components will be described in order.
(A) Epoxy resin:
The epoxy resin used is not particularly limited as long as it is bifunctional or higher. Examples include novolak type epoxy resins such as bisphenol type epoxy resin, phenol novolac type epoxy resin, cresol novolak type epoxy resin, naphthalene type epoxy resin, biphenyl type epoxy resin, cyclopentadiene type epoxy resin and the like. These epoxy resins can be used individually by 1 type or in mixture of 2 or more types.

  A material that is solid at room temperature is desirable because it is convenient in the subsequent process that it is tack-free in the B-stage state, that is, the surface is not sticky. If a liquid material is used at room temperature, the adhesive remains in the B-stage state, and if heated and semi-cured to make it tack-free, the viscosity does not become low at the time of solder connection and the solder connectivity may be hindered. is there. More preferably, a grade resin with few impurities, especially ionic impurities, is used.

(B) Curing agent having flux properties:
The curing agent used as the component (B) in the composition of the present invention has flux performance. Specifically, having flux performance means having the ability to reduce the oxide film on the solder surface. Examples of the curing agent having such flux performance include a phenol resin, an acid anhydride, and an amine curing agent. Among them, those having few impurities and being solid at room temperature are desirable. These can be used alone or in combination of two or more.

  Examples of the phenolic curing agent include a phenol resin having at least two phenolic hydroxyl groups in one molecule, and more specifically, a novolak type phenol resin such as a phenol novolak resin or a cresol novolak resin; Xylylene-modified novolak resins such as novolak resin, metaxylylene-modified novolak resin, orthoxylylene-modified novolak resin; bisphenol-type phenol resins such as bisphenol A-type resin and bisphenol F-type resin; biphenyl-type phenol resin, resol-type phenol resin, phenol-aralkyl type Resin, phenolic resin such as biphenyl aralkyl type resin; triphenolalkane type resin such as triphenolmethane type resin and triphenolpropane type resin and polymers thereof Phenol resins; naphthalene ring-containing phenolic resins, dicyclopentadiene-modified phenolic resins can be used either.

  Examples of amine curing agents include aliphatic amine curing agents such as diethylenetriamine, tetraethylenetetramine and tetraethylenepentamine, alicyclic amine curing agents such as isophoronediamine, and aromatic amines such as diaminodiphenylmethane and phenylenediamine. Examples thereof include a system curing agent and dicyandiimide.

  Examples of the acid anhydride curing agent include phthalic anhydride, pyromellitic acid anhydride, trimellitic acid anhydride, hexahydrophthalic anhydride, and the like.

  These curing agents have [(B) component in terms of obtaining good flux properties and, in turn, good connectivity, and preventing generation of voids due to volatile gas during curing and physical properties of the cured product. The molar ratio of epoxy group to reactive group] / [epoxy group in component (A)] is 0.95 to 1.25, preferably 0.95 to 1.05.

(C) Imidazole catalyst:
Examples of the imidazole catalyst used as the component (C) in the composition of the present invention include 2-methylimidazole, 2-ethylimidazole, 2-undecylimidazole, 2,4-dimethylimidazole, 2-heptadecylimidazole, 1,2-dimethylimidazole, 1,2-diethylimidazole, 2-phenyl-4-methylimidazole, 2,4,5-triphenylimidazole, 2-ethyl-4-methylimidazole, 2-phenylimidazole, 1-benzyl 2-phenylimidazole, 1-cyanoethyl-2-methylimidazole, 1-cyanoethyl-2-undecylimidazole, 1-benzyl-2-methylimidazole, 2-phenyl-4,5-dihydroxymethylimidazole, 2-allyl- 4,5-diphenylimida Lumpur, 2-phenyl-4-methyl-5-hydroxymethyl imidazole, and the like.
Among these, 2-phenyl-4-methyl-5-hydroxymethylimidazole, 2-phenyl-4,5-dihydroxymethylimidazole, 2-methylimidazole, 2-ethyl-4-methylimidazole, 2-ethylimidazole, 1 , 2-dimethylimidazole, 1,2-diethylimidazole, 2,4-dimethylimidazole and 2-phenyl-4-methylimidazole are preferred.
The amount of the imidazole catalyst used as the component (C) is 0.05 to 2 parts by mass per 100 parts by mass in total of the component (A) and the component (B) from the viewpoint of the balance between curability and workability of the composition. Is preferable, and 0.1-1 mass part is still more preferable.

(D) Volatile organic solvent:
(D) The solvent of a component is mix | blended in order to make this underfill agent suitable for spray application. This improves the workability in the coating process. Examples of the solvent include ketones, esters, alcohols, ethers, γ-butyrolactone, propylene glycol methyl ethyl acetate (PGMEA) and diethylene glycol monoethyl ether acetate (EDGAC), and a mixed solvent composed of two or more kinds thereof. Can be used.

  The amount of the organic solvent is 30 to 100 parts by mass, preferably 50 to 80 parts by mass per 100 parts by mass in total of the components (A) and (B).

(E) Inorganic filler:
The inorganic filler used as component (E) is a porous inorganic filler having a pore volume of 0.3 to 0.7 cm ³ / g, and the surface thereof is coated with silicon oxide (SiO ₂ ). It is an inorganic filler. When the pore volume is less than 0.3 cm ³ / g, the mechanical strength is low, and the dicing property and the polishing property are good. However, the mechanical strength of the cured product of this composition is extremely lowered, and the solder reflow or temperature Cracks may occur in the cured product during the cycle. When it exceeds 0.7 cm ³ / g, the mechanical strength of the cured product becomes too high, and the dicing property and the polishing property are remarkably lowered. The pore volume is preferably in the range of 0.4 to 0.6 cm ³ / g.

  Here, the pore volume is measured by a gas adsorption method.

  As the inorganic filler of the component (E), in order to improve the mechanical strength after curing of the underfill agent and reduce the thermal expansion coefficient, it is particularly limited if the pore volume is in the above range. Examples thereof include fused silica, crystalline silica, alumina, titanium oxide, silica titania, boron nitride, aluminum nitride, silicon nitride, magnesia, magnesium silicate, and aluminum. These can be used singly or in combination of two or more. Preferably, silica is preferred from the viewpoint of filler processability.

  Such an inorganic filler used as the component (E) can be obtained, for example, under the trade name P15C (manufactured by JGC Catalysts and Chemicals) of silica.

The inorganic filler preferably has an average particle diameter (D ₅₀ ) of 0.05 to 10 μm in order to satisfactorily fill the space between the solder bumps. The maximum particle size (D ₉₀ ) is desirably set to ½ or less of the bump height. In addition, the underfill agent is preferably highly thixotropic so that an underfill agent layer having a uniform thickness is formed. The average particle size can be measured by a centrifugal sedimentation method and a laser diffraction method, and the maximum particle size can be measured by a sieving method.

  (E) As for the compounding quantity of the inorganic filler of a component, 50-300 mass parts is preferable per 100 mass parts of total of (A) component and (B) component, More preferably, it is 50-150 mass parts. If the amount of the inorganic filler is too small, the effects of improving the mechanical strength and reducing the thermal expansion coefficient are insufficient. If the amount is too large, the filler is likely to bite into the bumps. Moreover, in the case of the manufacturing method 1, it is desirable to adjust a compounding quantity so that the visibility of a solder bump can be obtained.

In addition, the composition of this invention can contain fillers other than (E) component. For example, it is desirable to contain a fine powder filler having an average particle diameter of 0.005 μm or more and 0.2 μm or less, particularly fine powder silica. Examples of the fine silica include fumed silica such as Aerosil 130, Aerosil 200, Aerosil 300, Aerosil 380 (all trade names manufactured by Nippon Aerosil Co., Ltd., specific surface areas are 130, 200, 300, 380 m ² / g in this order). It is preferable to use a trimethylsilylated product. That is, silica powder having an average particle size of 0.005 μm or more and 0.2 μm or less and having a trimethylsilylated surface is preferable. When these are used, there is an advantage that the shape of the fillet can be easily controlled. These differential fillers preferably have a solid refractive index of 2.35 to 2.55. When using a material having such a refractive index, it becomes easy to specify the position of the solder bump, that is, the visibility of the solder bump is improved, and positioning at the time of dicing or connection is facilitated. Is advantageous.

  When other fillers other than the component (E) are used in combination, the amount thereof is preferably 10% by mass or less, more preferably 7% by mass or less, based on the total amount of all fillers. Moreover, it is 1-20 mass parts with respect to a total of 100 mass parts of (A) component and (B) component, Preferably, it is the range of 1-10 mass parts.

  When the fine powder filler having an average particle size of 0.005 μm or more and 0.2 μm or less is used in combination as another filler, the blending amount thereof is 10% by mass or less with respect to the total amount of all fillers. Preferably, it is 3-7 mass% more preferably. When the fine powder filler is added, thin film moldability is improved, but when it is too much, the viscosity of the composition becomes too high, and the spray coatability of the composition is remarkably lowered.

・ Other ingredients:
If necessary, other components can be added to the composition of the present invention. For example, a thermoplastic polymer can be added to improve the mechanical strength during dicing of the cured underfill agent layer. In addition, a polymerization catalyst, a silane coupling agent, an ion trap material, and the like may be added as long as the objects and effects of the present invention are not impaired.

-Manufacturing method 1
Next, the manufacturing method 1 described above will be described in detail.
-Process (1):
An underfill agent is spray-applied to the circuit surface of the wafer using a spraying apparatus. The spray spray device used for spray coating uses a publicly known and existing device, but the conditions for the optimization of the thickness in thin film formation are the distance to the wafer of the spray spray part, the moving range, the pitch, the speed, and the coating amount. It is desirable that adjustment and setting can be easily performed by air pressure and application time.

  The wafer is formed with scribes that are divided into a plurality of pieces (chips), and each piece is provided with a circuit and a plurality of solder bumps for flip chip connection. The underfill agent needs to be liquid at room temperature. If it is solid, it must be heated at a fluid temperature, which may accelerate curing of the composition.

  Next, the applied underfill agent is made into a B-stage, that is, a state in which the composition has no fluidity but is not completely cured. B-staging is performed by heating the underfill agent to volatilize the solvent. In any case, the curing is performed at a temperature sufficiently lower than the curing start temperature, preferably about 40 ° C. or more from the curing start temperature (onset temperature) so as not to proceed with curing of the underfill agent.

  Preferably, after forming the B-stage, the thickness of the underfill agent layer is adjusted by etching or ashing until the top of the solder bump appears. When this step is added, solder connectivity and solder bump visibility are improved.

  The thickness of the resulting B-stage underfill agent layer is 1.0 to 1.3 times the height of the solder bump, and preferably 1.1 to 1.2 times. If the thickness is less than the lower limit, the underfill agent flows and spreads during solder connection in step (4), so that the gap between the silicon chip and the substrate is not satisfied, and a void may be generated. On the other hand, if the upper limit is exceeded, the underfill agent may wrap around the back surface (non-circuit surface) of the semiconductor element, contaminate the element, and the fillet may become too large, thereby hindering high density of the element.

-Process (2):
In the step (2), the silicon wafer is separated into pieces (dicing). Dicing may be performed using a known dicer. At this time, the B-stage underfill agent is also singulated together with the wafer.
-Process (3):
Next, in step (3), individual pieces (hereinafter referred to as “chips”) are picked up and positioned on pads on the circuit board to which the chips are to be connected. The positioning may be performed continuously with the following step (4) using a known flip chip bonder.

-Process (4):
In step (4), the solder bump is melted and the chip is solder-connected to the substrate (including the package). Solder connection is a method in which rapid heating to 200 ° C. or higher is performed using a pulse heat type flip chip bonder, and the above positioning and solder connection are performed in the same apparatus; the surface of the underfill agent layer in the B stage state becomes liquid The substrate and / or the chip can be heated to the above temperature, the surface of the underfill agent can be pressure-bonded to the substrate, and then soldered in an IR reflow apparatus. By heating, the underfill agent has a low viscosity, the solder bump and the pad on the substrate are brought into contact and soldered, and the curing of the underfill agent starts. By selecting a curing catalyst or the like, the curability can be improved, and the resin curing process can be completed simultaneously with the end of the solder connection. In the case of an underfill agent having a low reactivity, it is also possible to complete curing by heating in an oven after soldering.

-Manufacturing method 2-
Next, the manufacturing method 2 will be described in detail.
-Process (i):
Step (i) uses the above-described underfill agent composition of the present invention as the first underfill agent. In step (i), the thickness of the first underfill agent layer in the B stage state when the underfill agent is changed to the B stage is 0.5 to 1.0 times the height of the solder bump. This is carried out in the same manner as in step (1) of production method 1.

-Process (ii):
In the step (ii), a liquid second underfill agent different from the first underfill agent is spray-applied on the first underfill agent layer in the B-stage state, and then the applied second underfill agent is sprayed. The B-staged second underfill agent layer is formed into a B-stage, and the total thickness of the first and second underfill agent layers is 1.0-1. It is formed to be 3 times, preferably 1.1 to 1.2 times. If the thickness is too thin, the underfill agent flows and spreads when the solder is connected in step (v), so that the gap between the silicon chip and the substrate is not filled, and voids may be generated. On the other hand, if it is too thick, the underfill agent may wrap around the back surface (non-circuit surface) of the semiconductor element, contaminating the element, and the fillet may become large, thereby hindering high density of the element.

  The operation of forming the second underfill agent layer is different in the type of the underfill agent used, forming the underfill agent layer on the first underfill agent layer, and the thickness of the underfill agent layer. Except for the above, this is the same as step (1) or step (i).

As the second underfill agent,
(A) epoxy resin,
(B) a curing agent having flux properties;
A composition containing (C) an imidazole catalyst and (D) a volatile organic solvent is preferred.

  Here, since the components (A) to (D) are the same as the components described for the underfill agent composition of the present invention, a description thereof will be omitted.

The composition does not necessarily contain a filler, but a filler is blended from the viewpoints of improvement of thin film formability, moderate viscosity and spray application, visibility of bumps in step (iv), and solder connectivity. can do. As the filler, the above-described porous inorganic filler having a pore volume of 0.3 to 0.7 cm ³ / g, the spherical inorganic filler whose surface is coated with silicon oxide, and the above-mentioned average A fine powder filler having a particle size of 0.005 μm to 0.2 μm can be mentioned. These fillers are preferably contained in an amount of 1% by mass to 20% by mass, particularly 1% by mass to 10% by mass with respect to the total of the component (A) and the component (B).
In order to improve the solder connectivity between the chip and the substrate, after the formation of the B-staged first underfill agent layer in the step (i) and / or in the step (ii), the B-stage was formed. It is desirable to clean the bump surface with ultraviolet rays or plasma after the formation of the second underfill agent layer.

Step (iii), step (iv) and step (v):
These steps are substantially the same as step (2), step (3) and step (4) of production method 1, respectively.

  In the manufacturing method 2 described above, when high bump visibility and solder connectivity are particularly required, the first underfill layer is formed just below the top of the solder bump, and then the top periphery of the solder bump does not contain an inorganic filler. The formation of the second underfill layer having a relatively small content is advantageous in that high bump visibility and solder connectivity can be obtained in step (v) without impairing reliability.

  EXAMPLES Hereinafter, although this invention is demonstrated in detail using an Example and a comparative example, this invention is not restrict | limited to the following Example.

  In each example, the components shown in Table 1 were mixed with the amount shown in the table (parts means parts by mass) at room temperature with a planetary mixer, passed through a three-roll mill, and again planetary. It mixed with the mixer and underfill agent composition AG was obtained.

The materials used and the characteristics evaluation are as follows.
-Material-
-(A) Epoxy resin: Cresol novolak type epoxy resin (manufactured by Nippon Kayaku Co., Ltd., trade name: EOCN1050 (55), epoxy equivalent 200, solid at room temperature)
-(B) Curing agent: Phenol novolac resin (Maywa Kasei Co., Ltd., trade name: MEHC7800-4S, phenol equivalent 170, solid at room temperature)
(C) Curing catalyst: 2-phenyl-4,5-dihydroxymethylimidazole (2PHZ, manufactured by Shikoku Chemicals)
(D) Volatile solvent: Diethylene glycol monoethyl ether acetate (E) Spherical inorganic filler:

  Here, the pore volume is measured by a gas adsorption method, the average particle size is measured by a laser diffraction method, and the maximum particle size is a value measured by a sieving method.

Filler a: True spherical silica having an average particle diameter of 2 μm and a maximum particle diameter of 10 μm, porous and having a surface coated with SiO ₂ (pore volume: 0.01 cm ³ / g)
Filler b: True spherical silica having an average particle diameter of 2 μm and a maximum particle diameter of 10 μm, porous and having a surface coated with SiO ₂ (pore volume: 0.35 cm ³ / g)
Filler c: True spherical silica having an average particle size of 2 μm and a maximum particle size of 10 μm, porous and having a surface coated with SiO ₂ (pore volume 0.55 cm ³ / g)
Filler d: True spherical silica having an average particle diameter of 2 μm and a maximum particle diameter of 10 μm, porous and having a surface coated with SiO ₂ (pore volume 0.85 cm ³ / g)
Filler e: True spherical silica having an average particle diameter of 2 μm and a maximum particle diameter of 10 μm, porous and having a surface coated with SiO ₂ (pore volume: 1.15 cm ³ / g)
Filler f: True spherical silica having an average particle diameter of 2 μm and a maximum particle diameter of 10 μm, porous and having no surface coated (pore volume 0.55 cm ³ / g)

・ Optional ingredients:
Fine silica: {(CH ₃ ) ₃ Si} ₂ NH and silica powder treated with CH ₃ CH ₂ Si (OCH ₃ ) ₃ with an average particle diameter (d ₅₀ ) of 0.08 μm Thermoplastic resin: Weight Phenoxy resin with an average molecular weight of 10,000

-Characteristic-
(1) Water absorption rate For each material, a 3 mm thick and 50 mm diameter disk molded product was produced under curing conditions of 100 ° C./1 hour + 150 ° C./4 hours and left under conditions of 85 ° C./85% RH for 500 hours. Thereafter, the water absorption rate (mass%) was determined from the formula (weight after water absorption-initial weight) / initial weight × 100.
(2) Bending strength Each material was measured in accordance with JIS K6911.

(Note) * indicates that the composition does not satisfy the conditions of the present invention.

  As a semiconductor chip for evaluation, JTEG Phase2E175 lead-free specification of flip chip kit manufactured by Hitachi Ultra LSI Systems Co., Ltd., on which 576 solder bumps are mounted, bump height of 70 μm (hereinafter referred to as “evaluation TEG”) was used. JKIT TYPE-III manufactured by the same company was used as the substrate. The semiconductor chip and the substrate form a daisy chain by the connection between them, and conduction is possible when all the solder bumps in the chip are connected. That is, if even one of the 576 bumps cannot be connected, the continuity test shows insulation, and connection is difficult unless a mounting method with good connectivity is used.

<Example 1>
(Example of production method 1)
The composition C was used as an underfill agent, and was applied to a TEG for evaluation which was in the state of a wafer, that is, a wafer, using a spray method. The apparatus used was a Nordson Co., Ltd. fine swirl spray gun. At that time, the thickness setting after the B stage is 80 μm, so that the X-axis speed (mm / s), Y-axis pitch speed (mm / s), spray gun distance (mm), micro opening (mm) The atomizing air pressure (MPa) and the swirl air pressure (MPa) were adjusted. The wafer thus coated with the composition C was placed in an oven at 110 ° C. for 10 minutes, and the solvent was removed to obtain a B-stage state. When the average film thickness in the B-stage state was measured with a film thickness meter, it was 80 ± 5 μm as set.

<Example 2>
(Example of production method 2)
A first underfill agent layer in a B-stage state having a thickness of 60 ± 5 μm was used in the same manner as in Example 1 except that the composition C was used as the first underfill agent, and the thickness setting after the B stage was changed to 60 μm. Formed. On the first underfill agent layer, after applying the composition G as a second underfill agent by the spray method using the same application apparatus as in Example 1, under the same conditions as the first underfill agent B-staged. The total thickness of the two-layer underfill agent layer thus obtained was 80 ± 5 μm.
<Example 3>
A B-stage thin film having a thickness of 80 ± 5 μm was formed in the same manner as in Example 1 except that the composition B was used as the underfill material.

<Comparative Example 1>
A thin film in a B-stage state having a thickness of 80 ± 5 μm was formed in the same manner as in Example 1 except that the composition A was used as the underfill material.
<Comparative example 2>
A B-stage thin film having a thickness of 80 ± 5 μm was formed in the same manner as in Example 1 except that the composition D was used as the underfill material.
<Comparative Example 3>
A B-stage thin film having a thickness of 80 ± 5 μm was formed in the same manner as in Example 1 except that the composition E was used as the underfill material.
<Comparative Example 4>
A B-stage thin film having a thickness of 80 ± 5 μm was formed in the same manner as in Example 1 except that the composition F was used as the underfill material.

-Evaluation test-
The following evaluation was performed on each wafer processed in the above-described Examples and Comparative Examples. The results are shown in Table 2.
(1) Thin film stability The standard deviation σ of the film thickness measured for the B-stage underfill agent layer was determined. And it was considered good if the film thickness was in the range of 3σ ± 2 μm.
(2) Dicing characteristics A wafer having an underfill agent layer in a B-stage state was subjected to dicing, and blade reduction was evaluated. Five 8-inch wafers were continuously diced, and the reduction was ranked as follows.

○: Reduction is 5% or less.
Δ: Decrease exceeds 5% and 10% or less.

  X: The reduction exceeded 10%.

(3) Solder connectivity After positioning with a flip chip bonder and disposing the chip, the chip manter tool temperature was set at 260 ° C. and chip attachment was performed at 1 to 10N. Next, the underfill agent was cured in an oven at 150 ° C. for 4 hours. The conductivity of 10 obtained devices was examined with a tester. In the case of the number n of good conduction, it is shown in Table 2 as “n / 10”.
(4) Endurance of connection The device after being subjected to the test (2) was subjected to a thermal cycle test (TCT) under the following conditions. After 10 devices were subjected to a thermal cycle, the conductivity was examined with a tester. In the case of the number n of good conduction, it is shown in Table 2 as “n / 10”.

-1 cycle: -55 ° C x 10 minutes, 125 ° C x 10 minutes.
・ Number of cycles: 1000

  According to the method of the present invention, an underfill agent can be formed into a thin film, and solder connection and void generation can be easily reduced, and a semiconductor device having a highly reliable solder connection portion can be manufactured.

Claims (12)

(A) epoxy resin,
(B) a curing agent having flux properties;
(C) an imidazole catalyst,
(D) A volatile organic solvent, and (E) a porous inorganic filler having a pore volume of 0.3 to 0.7 cm ³ / g, and a spherical inorganic filler whose surface is coated with silicon oxide. comprise Naru, B- stage compatible liquid wafer level underfill composition.   The component (B) is included in an amount such that the molar ratio of [the epoxy group of the (B) component and the reactive group] / [the epoxy group in the component (A)] is 0.95 to 1.25. The composition according to claim 1.   The composition according to any one of claims 1 and 2, wherein the inorganic filler of the component (E) is contained in an amount of 50 to 300 parts by mass per 100 parts by mass in total of the component (A) and the component (B).   The composition which concerns on any one of Claims 1-3 whose hardening agent which has the flux property of (B) component is a phenol resin, an acid anhydride hardening | curing agent, an amine hardening | curing agent, or a combination of these 2 or more types. .   Furthermore, 1 to 20 parts by mass of silica powder having an average particle diameter of 0.005 μm or more and 0.2 μm or less and whose surface is trimethylsilylated with respect to a total of 100 parts by mass of the component (A) and the component (B) Item 5. The composition according to any one of Items 1 to 4. A method of manufacturing a semiconductor device comprising a substrate and a silicon chip flip-chip connected to the substrate,
(1) The underfill agent composition according to any one of claims 1 to 5 is spray-coated on a surface of the silicon wafer provided with solder bumps for flip chip connection, the solder bumps being provided, Next, the applied underfill agent composition is B-staged to form a B-stage underfill agent layer having a thickness of 1.0 to 1.3 times the height of the solder bump.
(2) A step of cutting the silicon wafer on which the underfill agent layer in the B-stage state is formed into pieces into silicon chips,
(3) The step of positioning the piece obtained in step (2) at the position to be connected on the substrate, and (4) melting the piece of the solder bump positioned in step (3). The process of connecting to the substrate,
A method of manufacturing a semiconductor device including:   In the step (4), the individual pieces and / or the substrate are heated to a temperature at which the surface of the B-stage underfill agent layer is liquefied, and the underfill agent layer surface is pressure-bonded to the substrate. Such a manufacturing method. A method of manufacturing a semiconductor device comprising a substrate and a silicon chip flip-chip connected to the substrate,
(I) The underfill agent composition according to any one of claims 1 to 5 is applied to the surface of the silicon wafer provided with solder bumps for flip chip connection provided with the solder bumps. Spray coating is applied as a fill agent, and then the applied underfill composition is B-staged to form a first underfill agent layer having a thickness of 0.5 to 1.0 times the height of the solder bump. The process of
(Ii) A second underfill liquid that is different from the first underfill agent is spray-applied on the first underfill agent layer in the B-stage state, and then the applied second underfill is applied. The second underfill agent layer in the B-stage state by converting the agent into a B-stage is such that the total thickness of the first and second underfill agent layers is 1.0 to 1.3 times the height of the solder bump. Forming a process to be
(Iii) a step of cutting the silicon wafer on which the B-stage underfill agent layer is formed in steps (i) and (ii) into individual silicon chips;
(Iv) a step of positioning the individual pieces obtained in step (iii) at a position to be connected on the substrate, and (v) melting the solder bumps of the individual pieces positioned in step (iv). The process of connecting to the substrate,
A method of manufacturing a semiconductor device including:   After the formation of the B-staged first underfill agent layer in step (i) and / or after the formation of the B-staged second underfill agent layer in step (ii), ultraviolet rays or The manufacturing method according to claim 8, wherein the bump surface is cleaned with plasma. The second underfill agent used in step (ii) is
(A) epoxy resin,
(B) a curing agent having flux properties;
The production method according to claim 8 or 9, which is a composition containing (C) an imidazole catalyst and (D) a volatile organic solvent. The second underfill agent further contains a spherical inorganic filler whose surface is coated with silicon oxide, in the form of a porous inorganic filler having a pore volume of 0.3 to 0.7 cm ³ / g. The manufacturing method according to claim 10. The second underfill agent further has an average particle size of 0.005 μm or more and 0.2 μm or less, and the silica powder whose surface is trimethylsilylated is 100 parts by mass in total of the component (A) and the component (B). The manufacturing method according to claim 10 or 11, wherein 1 to 20 parts by mass is contained.