WO2009090776A1

WO2009090776A1 - Semiconductor device and process for producing the same

Info

Publication number: WO2009090776A1
Application number: PCT/JP2008/066928
Authority: WO
Inventors: Hisao Ishikawa; Masanori Yokoyama; Masao Kojima
Original assignee: Horizon Technology Laboratory Co., Ltd.; Oosaki Seikou Ltd.
Priority date: 2008-01-17
Filing date: 2008-09-12
Publication date: 2009-07-23
Also published as: JPWO2009090776A1; JP4389112B2

Abstract

A semiconductor device comprising a printed wiring board and a member, e.g., a semiconductor package or electronic element, mounted on the wiring board by soldering with a lead-free solder. The soldering part is inhibited from developing microvoids when the semiconductor device is exposed to a high temperature for long (high-temperature aging). The soldering part hence has improved resistance to impact fracture. Also provided is a process for producing the device. In the step of assembling an electronic part, e.g., a semiconductor device, a lead-free solder is used to bond a member to be mounted, e.g., a semiconductor package or electronic element, to a printed wiring board in the following manner. A lead-free solder at least containing 0.01-0.2 wt.% nickel or the lead-free solder further containing 0.001-0.01 wt.% germanium is used in combination with a solder flux comprising any one or more members selected among organic fatty acid nickel salts and organic cobalt salts to solder a semiconductor package, electronic element, etc. to a copper land of a printed circuit board to cause an alloy layer containing either or both of nickel and cobalt to be evenly interposed in a given thickness at the soldering interface. Thus, the soldering part is inhibited from developing microvoids when the semiconductor device is aged at a high temperature. The soldering part in the semiconductor device hence has improved resistance to impact fracture.

Description

Specification

Semiconductor device and manufacturing method thereof

Technical field

[0001] The present invention relates to a soldering apparatus and a semiconductor device for a semiconductor device in which a mounting member such as a semiconductor package or an electronic element is mounted on a printed circuit board, and solder bonding using lead-free solder for the purpose of circuit formation In particular, the present invention relates to a semiconductor device and a soldering device in which a copper land of a printed circuit board and a semiconductor package, an electronic element, etc. are soldered via a soldering material, and more specifically, the semiconductor device is in use. By suppressing the generation of voids in the microphone joints of solder joints that are formed over time when exposed to high temperatures for a long time due to heat generation of the solder (hereinafter referred to as high temperature aging), impact rupture resistance of the solder joints of the semiconductor device The present invention relates to the provision of semiconductor devices and technologies for forming solder joint films with high quality and reliability.

Background art

[0002] In recent years, electrons are increasingly required to have higher reliability and smaller size and lighter weight. Both semiconductor devices and electronic components used for these devices have become smaller and thinner, and solder joints used to form the circuit are used. The parts are also miniaturized, and strict high reliability is required.

As an example, the case of high-density mounting type BGA (ball 'grid' array) and CSP (chip 'size' package) widely used in the field of packages mounted on semiconductor devices will be described. In order to mount the CSP on the printed circuit board, it is necessary to form solder bumps using BGA and CSP leads using small solder balls. (See Figure 1)

This solder ball generally has a spherical shape, and the solder balls with a diameter of 0.76 πιιη φ have been mainly used in the past. However, these solder balls have recently become increasingly smaller and have a diameter of 0.1 0 to 0.4. 5 In addition to the fact that solder balls of πιπι φ are becoming mainstream, these solder balls are used for the tin-lead solder balls that have been widely used so far. Due to this problem, the use of bells has been prohibited or regulated, and recently, especially in the field of electronic components, so-called “lead-free solder balls” that do not contain ships are being widely used to form solder bumps for BGA and CSP. [0003] On the other hand, when assembling a semiconductor device, these BGAs and CSPs are bonded to the printed circuit board to form a circuit. First, the BGA and CSP solder bumps are aligned with the mounting position of the printed circuit board. As a technical issue in terms of bonding reliability and electrical reliability, soldering is performed using lead-free solder. As a result, a large amount of hydrogen, moisture, and other decomposed components of the activator component are generated due to the reaction between the activator component present in the solder flux and the solder metal. It is known that “large voids” (hereinafter referred to as “macrovoids”) occur, leading to poor conduction and crack breakage in the solder layer. (Patent document 1, page 4), (Non-patent document 1)

[0004] Various methods have been proposed for suppressing this macrovoid, such as a technique using a special solder flux (Patent Document 1) and a technique using a copper core ball (Patent Document 2). However, it is possible to eliminate almost no macro voids by optimizing the conditions such as flux component material 'application amount selection, soldering j & «· speed' time, riff mouth, defoaming treatment, etc. Even if there are several, it can be dispersed and confined in the solder layer, and it is almost impossible to cause cracks or breakage of the solder layer. In the case of a semiconductor device soldered under such optimum conditions, if it is stored indoors at a temperature of 40 or less and a relative humidity of 70% or less (hereinafter referred to as normal condition), the copper land of the printed circuit board is used. There are no macrovoids at the solder interface. (Patent Document 2)

[0005] However, even when semiconductor devices mounted with BGA and CSP soldered under such optimum conditions are exposed to a high temperature of 120 or higher for a long time (high-temperature aging), Cu (Cu) on the printed circuit board side and tin (Sn) on the solder bump side at the junction interface diffuse to form Cu ₃ Sn, which is an intermetallic compound (IMC), and the interface between Cu and Cu ₃ Sn It is widely known that “micro voids” (hereinafter referred to as “microphone mouth voids”), which are so-called Kendall voids having a diameter of 0.001 to several μm in diameter, are generated in the Cu ₃ Sn layer. (Non-Patent Documents 1 to 4)

[0006] The longer the high-temperature aging time of this microphone mouth void, the more the number of generations increases with time, and the size (volume) increases by bonding to each other. In other words, the number of microvoids increases on one side of the solder joint interface and its vicinity (in the Cu and Cu ₃ Sn interface and in the Cu ₃ Sn layer), so the void ratio of the joint interface increases and the solder joint strength itself increases. When the impact force is applied to this part, the joint breaks. This phenomenon is also widely known for conventional tin-lead solders. In recent lead-free solders (eg, tin-silver-copper solders, tin-bismuth solders, tin-copper solders), the degree of occurrence of microvoids due to high-temperature aging It is said that this is even more remarkable than tin force solder, which is a major problem in reliability of solder joints.

Patent Document 1: JP 2005-288490 A

Patent Document 2: Japanese Patent Application Laid-Open No. 2007-75856

Patent Document 3: Japanese Patent Application Laid-Open No. 2004-76030

Patent Document 4: JP-A-11-77366

Patent Document 5: Japanese Unexamined Patent Application Publication No. 2004-306092

Patent Document 6: JP-A-2005-169495

Non-Patent Document 1: R. Aspandiar, "Void in Solder Joints" SMTA Northwest Chapter

Meeting (September 21, 2005)

Non-Patent Document 2: C. Hillman: "Long-term reliability of Pb-free electronics"

Electronic Products p.69 (September 2005)

Non-Patent Document 3: Mitsuyuki Ban, Yu Shimauchi: "Reliability evaluation and failure analysis technology for electronic components",

JFE Technical Report No.13 p.97-102, August 2006

Non-Patent Document 4: Shinji Ishikawa et al .: “Generation of Kirkendall Board at the Joint of High-Temperature Solder and Cu Plate”, Journal of Japan Institute of Electronics Packaging, No. 9-4, p.269-277,

2006

Disclosure of the invention

Problems to be solved by the invention

The present invention frequently occurs in solder joints after high-temperature aging, which is a drawback of the prior art, in a semiconductor device in which electronic members such as semiconductor packages and electronic elements are solder-bonded to a printed circuit board using lead-free solder. With the aim of making technology that dramatically reduces the impact rupture resistance of the solder joints of semiconductor devices, with virtually no generation of microvoids Yes.

Means for solving the problem

An object of the present invention is to improve impact resistance after high-temperature aging of a semiconductor device in which an electronic member such as a semiconductor package or an electronic element is solder-bonded to a printed circuit board using lead-free solder. As a lead-free solder containing at least 0.1 to 0.2% by weight of nickel or nickel when soldering a semiconductor package, an electronic element, or the like in a semiconductor device to a printed circuit board or the like 0.2 wt ^_0/0 further Genoremaniumu zero. 0 0 1 to 0.0 1 of a lead-free solder and an organic fatty acid nickel salt or Yuranmune fatty acid cobalt salt containing by weight%, one, or two or By soldering semiconductor packages, electronic devices, etc. using solder flux containing more than one type, it often occurs at solder joints after high temperature aging, which is a drawback of conventional technology. The present invention provides a technique for substantially improving the impact fracture resistance of the solder joint portion of a semiconductor device with almost no occurrence of an icroboid.

[0009] Lead-free solder containing at least 0.1 to 0.2% by weight of nickel used in the present invention, or containing nickel of 0.1 to 0.2% by weight and further containing germanium in an amount of 0.01 to 0.2% by weight. Free solder containing ~ 0.11% by weight has the effect of improving the heat resistance and thermal fatigue strength by adding nickel, and the addition of germanium suppresses the oxidation of tin in the solder and improves the joint strength Although it is already known in Patent Document 4 that it has an effect, a semiconductor device in which an electronic member such as an electronic element such as a BGA is appropriately soldered to a copper land portion of a printed board using these solders in common with a normal flux. As shown in Comparative Example 2 described later, there is no large void of 30 micron or more in the normal state, and there is no microvoid of 0.0 0 1 to several microns, but after this is subjected to high temperature aging for 1 5 0 ^ 2 40 hours Obviously hundreds of microphones It has been found that the formation of voids reduces the impact resistance.

[0010] In addition, the solder flux containing the columnar fatty acid nickel salt or the organic fatty acid cobalt salt of the flux used in the present invention, and nickel panolemitate, conoretate panolemitate, nickel stearate, cobalt stearate Patent Documents 5 and 6 describe that the solder flux containing solder improves the wettability of the solder during soldering. Furthermore, the nickel in the solder flux is tinned at the solder joint interface. Copper.Nicke Patent Document 6 also describes that it is effective as a measure to prevent solder erosion phenomenon by preventing the decrease in solder joint strength due to the growth of intermetallic compounds in a high-temperature environment by forming a copper intermetallic compound. .

[0011] However, the fluxes of Patent Documents 5 and 6 and ordinary lead-free solder, for example, the most widely used tin-silver-copper-based solder, are used for electronic devices such as BGA. As described in detail in Comparative Examples 3 and 4 to be described later, a semiconductor device in which an electronic member is appropriately solder-bonded to a copper land portion of a printed circuit board has a large void of 30 microns or more in a normal state. Although there was no microvoid, it was confirmed that if this was subjected to high temperature aging for 1 5 240 hours, a very large number of microvoids were generated and the impact resistance was lowered.

The invention's effect

[0012] For this reason, the inventors have repeated experiments by trial and error while combining lead-free solders of various compositions and various fluxes. As a result, at least nickel of 0.1 to 0.2% by weight is obtained. Lead-free solder containing nickel or nickel-containing organic fatty acid nickel salt or organic fatty acid cobalt salt containing 0.1-0.2% by weight of germanium and lead-free solder containing 0.01 to 0.1% by weight By soldering with a combination of solder flutters containing one or more of them, the generation of microvoids in solder joints after high-temperature aging, which is a drawback of the prior art, is almost eliminated. It was invented that the impact rupture resistance of the solder joint could be dramatically improved.

[0013] It is essential that the lead-free solder used in the invention contains nickel, and its content is preferably 0.1 to 0.2% by weight as an experimental result. The number of micro voids generated later is large, and the impact fracture resistance is also poor. This is because if the nickel layer thickness is less than 0.1% by weight, the nickel layer agglomerated at the solder joint interface is thin, so the tin-copper-nickel intermetallic compound layer (hereinafter abbreviated as IMC) formed at the interface by high-temperature aging. ) is considered to be because the C u copper land side is C u ₃ S n layer is I MC of S n ZC u Ritsuchi diffuses into the solder layer becomes thicker in order was thin. On the other hand, in the range exceeding 0.2% by weight, it is considered that Niggel's minute granular segregation is observed at the solder joint interface, and the solder joint strength is lowered. Therefore, it is desirable for the nickel content. Ku is 0. 0 2 to 0.1 weight ^_0/0. I MC of the tin-copper nickel containing N i is considered a (C u N i) 6 S n 5, which prevents the diffusion of the solder side by heat aging of the copper of the printed circuit board side, C u ₃ S n It is thought to play a role in suppressing layer growth.

[0014] Although Kichiichi Kendall Poid is generally said to be a void due to lattice defects, the microphone mouth void after high-temperature aging is as shown in Fig. 12-; I 3 The case where oxygen combines with tin 1 3 to become stannous oxide 1 2 and copper 14, ie

C u O + S n-→ C u + S n O

As a result of the volume shrinkage of about 1.5%, a micro void (void) 16 is generated, or copper oxide 11 and stannous oxide 12 as shown schematically in FIG. Case of copper 14 and stannic oxide 15

C u O + S n O → C u + S n O _z

The inventors think that microvoids (voids) 16 are generated by causing a deposition shrinkage of approximately 15% due to (a hypothesis).

In addition, when germanium is added to nickel-containing lead-free solder, the effect of tin oxidation suppression is large, so the effect of improving microvoids and impact fracture resistance after heat aging is considered to be greater.

[0015] On the other hand, the flux used in the present invention contains at least one of organic nickel salt and organic cobalt salt in the flux, and is used at the same time. ^ Lead-free solder, or nickel or cobalt, or both metals synergistically by using nickel-containing lead-free solder with nickel and / or phosphorus added to the solder High temperature aging, which is the above-mentioned difficulty, by suppressing the diffusion of copper on the printed circuit board side and tin on the solder bump side by forming an alloy layer containing copper uniformly at a predetermined thickness at the solder joint interface. The generation of microvoids in the solder joints at a later stage can be greatly suppressed to almost none, and the impact fracture resistance of the solder joints can be dramatically improved.

[0016] None of the nickel salt and cobalt salt added to the solder flux and organic salt are effective, but compounds that are corrosive, deteriorate over time, and are harmful to the environment and the human body are preferred. Particularly preferred are organic acid nickel or organic acid cobalt. As the organic acid constituting them, aliphatic ruponic acids such as palmitic acid, stearic acid and oleic acid, and aromatic rubonic acids such as phthalic acid and pyromellitic acid can be used. Among these, those obtained by substituting hydrogen atoms of carbonyl groups of organic fatty acids having 8 to 20 carbon atoms with metal atoms of either nickel or cobalt are excellent. The reaction mechanism is either a solder ball consisting of copper and nickel-containing lead-free solder on the substrate side by heating, or nickel-containing lead-free solder with further addition of nickel to the solder or-or BGA Flux containing any one or more of organic nickel salt or organic cobalt salt prepared between solder bumps, etc. is dissolved and activated, and then the solder balls or solder bumps are activated. When the solder ball or the solder bump is heated to a temperature higher than the melting point, the metal ions of the nickel salt or cobalt salt in the flux are also released and synergistically with the nickel in the solder ball or the solder bump. An alloy layer of tin and nickel or cono-noret is formed at the contact interface, and the alloy layer is bonded to the copper on the substrate side. It is believed that this will result in the presence of tin-copper-nickel or cobalt in the joint.

[0017] Specific examples of the above flux include, for example, nickel palmitate C (CH2) ₁₄ COONi having 16 carbon atoms, cobalt palmitate CH ₃ (CH2) wCOOCo, nickel stearate having 18 carbon atoms CH ₃ (CH2) ₁₆ COONi and stearic acid copalte CH ₃ (CH2) i ₆ COOCo are particularly effective and may be used alone or in combination of two or more of them.

“0018” Further, the total amount of organic fatty acid nickel and organic fatty acid cobalt to be blended in the solder flux is optimally 1 to 5% by weight, and if it is less than 1% by weight, the nickel in the alloy layer formed in the solder joint or Since the amount of cobalt and / or both metals is too small, the effect of suppressing the formation of microphone mouth voids after high-temperature aging is insufficient, and when the content is 5% by weight or more, the effect of suppressing the formation of microvoids is equivalent and practical. The force that inhibits the viscosity as a flux \ Ni is segregated near the joint, and the cost is also increased.

[0019] The nickel-containing lead-free solder and the organic fatty acid nickel or cobalt-containing solder flux are used in combination to form synergistically at the solder joint interface (C u N i) ₆ S "Les Li ϋ;. {Ί / ϋ ί ) · 6323 η 5 and the thickness of the I MC layer of tin-copper nickel estimated 2 5 generation of micro Boi de after heat aging in the following Ai m is still very Although the impact resistance is poor, the generation of microvoids is drastically reduced to 1 Z 10 or less at 3 m or more, and the impact resistance is remarkably improved. At 4 μm or more, there are almost no microvoids. The thickness of the IMC layer is saturated at about 6 / m within the above-mentioned range of solder and flux and within the nickel concentration range. Since it is difficult, the thickness of the IMC layer is preferably 3 to 6 μm.

Furthermore, the nickel coordination number in the IMC is determined by a synergistic effect when the nickel-containing tin-silver-copper-based lead-free solder of the present invention is used together with a shelf-containing nickel oxalate-containing flux such as nickel palmitate. More than when used alone, so the nickel in the IMC

As a result, it is estimated that the generation of microphone mouth voids after prolonged high-temperature aging is suppressed.

[0020] The constituent component of the solder flux used in the present invention is mainly composed of mouth gin, and the above organic fat

In addition to one or more acid nickel salt or organic fatty acid cobalt salt, diethanolamine

, _ Diphenyldanidine hydrobromic acid, isopropyl hydrobromic acid and other active agents, stearic acid amine and other thixotropic agents, waxes and viscosity modifiers such as cellulose, and solvent etc. It consists of a uniform mixture. There are no special conditions or restrictions on how to use these solder fluxes, and they can be used under normal soldering conditions as with conventional fluxes.

BEST MODE FOR CARRYING OUT THE INVENTION

First, as Comparative Example 1, there are 3% by weight of silver not containing nickel, phosphorus, and germanium, 0.5% by weight of copper, and the remaining solder is a lead-free solder composed of tin and a general solder flux. Flux consisting of composition, ie

WW mouth gin resin 60% by weight

Isopropyl hydrobromide (activator) 0.3 wt%

Sebacic acid (active agent) 1.0 wt ^_0/0

Amin stearate 5.0 wt ^_0/0

Ethylene glycol one Rumonobuchirue one ether 3 3.7 wt ^_0/0 Was used as the base solder flux, and soldering was performed under the conditions shown in Table 1 below, which combined various bell-free solders described later and various fluxes.

[0022] In Comparative Example 2, the same solder flux as in Comparative Example 1 above was used, with sharp free solder consisting of 0.05% by weight of nickel, 3% by weight of silver, 0.5% by weight of copper, and the balance being tin. Then, soldering was performed under the conditions described below.

[0023] Further, in Comparative Example 3, the same solder as in Comparative Example 1 and the base solder flux in Comparative Example 1 of 98% by weight and nickel palmitate 2% by weight were mixed and mixed uniformly to be described later. Soldering was performed under the following conditions.

[0024] It 1wt ^_0/0 and cobalt 5 parts by weight of stearic acid palmitate nickel against further base solder flux 9 8 by weight ^_0/0 of Comparative Example 1 and the same solder as Comparative Example 1 In Comparative Example 4 Soldering was performed under the conditions described below using a blended and uniformly mixed product.

On the other hand, as Example 1, the same nickel 0.5% by weight as Comparative Example 2, silver 3% by weight, copper 0.5%.

5 wt%, the same solder flux as in Comparative Example 3 with the lead-free solder and the balance being tin, i.e., by incorporating a double amount% Panoremichin, nickel relative to the base solder flux 9 8 weight ^_0/0 of Comparative Example 1 Using a uniformly mixed flux [Soldering was performed under the conditions described below.

[0026] In addition, nickel 0 as a second embodiment. 0 5 wt ^_0/0, germanium 0.0 0 5 wt%, silver 3% by weight, of copper 0.5 wt%, compared with the lead-free solder and the balance being tin Solder under the conditions described below using the same solder flux as in Example 3, that is, the base solder flux of Comparative Example 1 98% by weight and 2% by weight of nickel palmitate mixed uniformly. I did.

[0027] Further, in Example 3, 0.1% by weight of nickel, 0.05% by weight of germanium, 3% by weight of silver, and 0.5% by weight of copper are compared with a lead-free solder composed of tin of the balance, and Comparative Example 4 solder, i.e., base - later using scan solder flux 9 8 weight ^_0/0 for Panoremichin acid nickel 1 wt ^_0/0 and stearic phosphate Kobanoreto fluxes uniformly mixed by blending 1 wt% Soldering was performed under the conditions of

[0028] [Table 1] Lead-free solder composition and solder flux composition used in Examples and Comparative Examples

The preparation method of the evaluation sample and the trial production conditions are as follows.

That is, Examples 1 to 3 and Comparative Examples 1 to 4 are all BGAs of the same production lot and have external dimensions. 15mmX 15mmX 1.2mm (including solder bump thickness), number of leads (same as the number of solder bumps) 192, lead pitch force SO. 8mm, each lead part has the composition shown in each example and comparative example in advance Using a BGA with solder bumps formed by joining lead-free solder balls with a diameter of 0.45 mm φ, and burn-in via the solder flux [Table 1] with the composition shown in each example and comparative example An evaluation test sample was prepared by mounting the BGA on the corresponding copper lead portion on the test printed board. The printed circuit board for burn-in test will be described in more detail. The outer dimensions are 77mm x 132mm, the thickness is lmm, and the center of the board has the same pitch as the above BGA and 0.3mm φ copper lead. BGA continuity test circuit as 1 unit, 1 unit each up and down at 5 mm intervals, and this as 1 row in the center, 2 rows each left and right at 5 mm intervals, that is, 5 rows x 3 columns in a matrix, total Fifteen BGA-mountable circuits are formed, and the printed circuit board is covered with a solder resist film on the surface except for the copper leads. In each example and comparative example, the flux for the sample is randomly applied to each BGA unit at each n = 5 (number of repetitions), through which the above BG A is mounted, and the molten solder at a temperature of 25 0 ^ The BGA mounting side was left on the bath for 120 seconds, and each BGA solder bump was bonded to the copper land of the printed circuit board and then allowed to cool naturally.

Brief Description of Drawings

[Fig. 1] Partial cross-sectional view schematically showing the layout of BGA soldered to a printed circuit board in a semiconductor device

[FIG. 2] A cross-sectional SEM photograph of a junction in a normal state of a comparative example in a semiconductor device,

(a.) is Comparative Example 1, (b) is Comparative Example 2, and (c) is Comparative Example 3.

[FIG. 3] A cross-sectional SEM photograph of a junction part in a normal state of a semiconductor device,

(a) is Example 1 and (b) is Example 2.

[Fig. 4-1 1〗 is a cross-sectional photograph of the junction after high-temperature aging in Comparative Example 1 in a semiconductor device.

(a) is an SEM photograph, (b) to (c) are EPMA pine photographs of the elements contained in the part, (b) is Sn, and (c) is Cu.

[Fig. 41-2] Cross-sectional SEM photograph of the junction after high-temperature aging in Comparative Example 1 in a semiconductor device (EPMA mapping photo of elements contained in the portion of Fig. 4-11 (a) above) (D) is Ag and (e) is Ni. (F) is a cross section of another part of Comparative Example 1.

This is the part where micro voids are bonded in SEM picture.

[Figure 5] Cross-sectional photograph of the junction after high-temperature aging in Comparative Example 2 in a semiconductor device. (A) is a SEM photograph, and (b) to (e) are EPMA mapping photographs of elements contained in that part. Yes, (b) Sn, (c) Cu, (d) Ag, (e) Ni Ni EPMA mapping photo. (F) is a cross-sectional SEM photograph of another part of Comparative Example 2 where microvoids are scattered.

[Fig. 6] A cross-sectional photograph of a junction after high temperature aging in Comparative Example 3 in a semiconductor device, (a) is an SEM photograph, and (b) to (e) are EPMA mapping photographs of elements contained in that part. (B) Sn, (c) Cu, (d) Ag, (e) Ni Ni EPMA mapping photo. In addition, (f) is a cross-sectional SEM photograph of another part of Comparative Example 3 where the microvoids are scattered.

FIG. 7 is a cross-sectional photograph of a junction after high temperature aging in Example 1 of the present invention in a semiconductor device, where (a) is an SEM photograph, and (b) to (e) are E PMA pine of elements contained in that portion. (B) is Sn, (c) is Cu, (d) is Ag, (e) is N i E

PMA mapping photo.

FIG. 8 is a cross-sectional photograph of a junction after high temperature aging in Example 2 of the present invention in a semiconductor device, where (a) is an SEM photograph, and (b) to (e) are E PMA pine of elements contained in that portion. (B) is Sn, (c) is Cu, (d) is Ag, (e) is Ni E

PMA mapping photo.

[Fig. 9] In the semiconductor device of the present invention, in the solder flux containing nickel palmitate

Ni content (horizontal axis, unit is% by weight) or in nickel-free lead-free solder

It shows the relationship between the Ni content (horizontal axis, unit is 1Z100% by weight) and the thickness of the tin copper nickel intermetallic compound (IMC) layer formed after heat aging (vertical number, unit is / zm). FIG.

[Figure 10] Cross section of the junction after high temperature aging of Comparative Examples 1 to 4 in the semiconductor device

It is sectional drawing which shows typically the state expanded 5000 times.

[Fig. 1 1] A cross-section of a junction after high temperature aging in Examples 1 to 3 of the invention in a semiconductor device It is sectional drawing which shows typically the state expanded about 500,000 times.

[Fig. 12] A diagram schematically showing copper oxide CuO and stannous oxide SnO present in the cross section of the solder joint.

[FIG. 13] A diagram schematically showing microvoids generally generated after aging around the solder joint for explaining the microvoid generation mechanism (hypothesis).

[Fig. 14] Microvoid generation mechanism (hypothesis) As an explanatory diagram of the first case, it schematically shows the volume ratio before and after the change of tin and copper oxide to stannous oxide and copper. [Fig. 15] In the semiconductor device of the present invention.

FIG. 15] Microvoid generation mechanism (hypothesis) As an explanatory diagram of the second case, it schematically shows the volume ratio before and after the change of stannous oxide and copper oxide to stannic oxide and copper.

Explanation of symbols

1 B GA

2 Printed circuit board

3 Solder bump

4 Soldering material

5 Copper lead (copper foil of printed circuit board)

6 Solder resist

7 Cu ₃ Sn layer present in the cross section of the solder joint

8 Cu ₆ Sn ₅ layer present in the cross section of the solder joint

9 Lead-free solder layer

1 0 Microphone mouth void in cross section of solder joint

1 1 Copper oxide C u O

1 2 stannous oxide S n O 2

1 3 Tin S n

1 4 Copper C u

1 5 Stannic oxide S n O

1 6 Microvoid (void)

1 7 Copper-tin-nickel alloy layer (Cu-Sn-Ni intermetallic compound layer) • 1 8 Copper-tin-nickel-cobalt alloy layer (Cu-Sn-Ni-Co intermetallic compound layer)

19 N i containing ship-free solder and normal solder flux

20 With normal lead-free solder and palmitic acid Ni-containing solder flux

21 Using normal lead-free solder and palmitic acid Ni-containing solder flux

22 Reference line indicating the thickness of the Ni-containing IMC layer at the desired solder joint interface to prevent the formation of microphone mouth voids after high temperature aging

[0032] As a method for evaluating the presence or absence of voids near the interface of the solder joint, the specimen for the evaluation test is kept in a normal state and left in a constant temperature oven for 150 hours for 240 hours. Then, the number and size of voids in the vicinity of the solder joint were observed, analyzed and compared with a scanning electron microscope (SEM) and an X-ray microanalyzer (EP MA). In addition, the samples of each of the above examples and comparative examples, which were subjected to accelerated heat aging tests under the same conditions at the same time, were tested using JEDEC (Joint Electron Device Enineer ^ Council) According to the standard Νο.22 · Β111, repeatedly drop from a height of 100 Omm at approximately 1300 G, and conduct a continuity test for each test sample each time. Examined.

[0033] As a result, from the SEM and E PMA analysis results of the cross section of the solder joint after the high-temperature aging acceleration test, it was found that the solder joint interface of Examples 1-2 and Comparative Examples 2 and 3 of the present invention Compared to the ternary alloy layer 17 of tin-copper containing nickel, Nikkenore and quaternary alloy layer 18 of tin-copper containing cobalt were detected at the joint interface of Example 3 and Comparative Example 4. As a matter of course, only the Cu ₃ Sn layer 7 was detected at the bonding interface of Example 1, and nickel and cobalt were not detected. It should be noted that a layer 8 of CueSn 5 (on the solder bump side) was detected in all of the examples and comparative examples above each of the alloy layers at the solder joint interface.

[00341 In addition, no voids were found in any of the Examples 1 to 3 and Comparative Examples 1 to 4 before the high temperature aging in the cross section of the solder joint. As an example of the SEM photograph, Fig. 2- (a) shows Comparative Example 1, Fig. 2- (b) shows Comparative Example 2, Fig. 2- (c) shows Comparative Example 3, Fig.

3— (a) shows the cross-sectional state of the solder joint of Example 1 and FIG.

[0035] On the other hand, after accelerated high-temperature aging under the above conditions, in each of Comparative Examples 1 to 4, a microphone mouth void was observed in the vicinity of the solder joint boundary section. The SEM photograph and the As an example of the E PMA mapping photograph of the Ag, Cu, Ni, and Sn elements, the case of Comparative Example 1 is shown in Fig. 4-1, Fig. 4-12. Fig. 6 shows the cross-sectional state of the solder joint of Comparative Example 3, Fig. 7 shows Example 1, and Fig. 8 shows Example 2.

According to this, the number of voids generated per 0.2 mm cross-sectional length near the joint boundary was very large, and in Comparative Examples 1 to 4, there were many microphone mouth voids. On the other hand, in Examples 1, 2 and 3, the generation of microvoids was overwhelmingly small, and the size of the microvoids was all microvoids of about 1 to 0.5 / On the other hand, in the comparative example, a larger number of 0.5 to 2 micron voids are observed, which are relatively larger than those of the example, and some of the voids are bonded horizontally in parallel to the bonding interface. Changes in macrovoids ranging from 5 to 10 m in length were also observed. (Fig. 4-2 (f) and Table 2)

[0036] The thickness of the IMC layer of tin copper nickel formed after high-temperature aging was 3 μιη or more in all the examples, whereas in Comparative Example 1, the IMC layer in which nickel was present None of Comparative Examples 2 to 4 was 2.5 μπι or less.

[0037] (i) When using normal solder and Ni flux, (B) When using Ni-containing tin silver copper solder (lead-free) and normal solder flux, and (C) Ni Containing Tin Silver Copper Solder (Lead Free) and Ni Ni Flux When Using Ni Ni Flux and IMC Layer Thickness of Tin-Nickel Nickel Formed after High Temperature Aging Case (B) Is shown in curve 19 of Fig. 9, and even if the Ni content is 0.05 wt%, the thickness of the IMC layer of tin-copper-nickel is only 1 / m at most. Similarly it was as curve 20 in FIG. 9 in the case of (A), and the thickness of I MC layer of the tin-copper nickel N i content 2 wt ^_0/0 in said flux is about 2. 4 μτα, When the Ni content is 5% by weight, the tin copper nickel I MC layer is saturated to about 3.5 μιη, but at the same time, many Ni segregation occurs at the joint interface from the Ni content of around 3% by weight. As a result, it was found that the solder joint strength and impact resistance were reduced. In contrast, in the case of (C), i.e., the nickel-containing when palmitic acid Nikkeno ^ chromatic content in the flux is used together with 2 wt ^_0/0 of flux and nickel-containing lead-free Ichisuzu silver-copper solder The relationship between the Ni content in the tin-silver-copper solder and the thickness of the IMC layer of tin-copper-nickel formed after heat aging is shown in curve 21 of Fig. 9. Ni content in the tin-silver-copper solder 0 0 2% by weight tin and copper The thickness of Kell's IMC layer is increased to nearly 4 / zm. Above 0.05%, the thickness of the IMC layer exceeds, and it is uniform without segregation of Ni, and microvoids are generated. It was found that the drastic reduction and impact resistance improved dramatically.

[0038] Further, regarding the impact rupture resistance after high temperature aging of Examples 1 to 3 and Comparative Examples 1 to 4, continuity failure (increased resistance value), which is considered to be impact rupture of the solder joint due to repeated impact of the above drop test In each comparative example, the number of drops to a maximum was 9 times or less, but in Examples 1 to 3, the number of drops was 40 times or more (the repeated tests were stopped) [Table 2].

[0039] [Table 2]

Number and size of voids per cross-section length of 0.2 mm

As described above, the technology of the present invention clearly realizes a highly reliable semiconductor device excellent in impact resistance after high-temperature aging, which is unprecedented, and is industrially valuable.

Claims

The scope of the claims

[1] In a semiconductor device, when lead-free solder is used to solder a mounting member to a printed circuit board, lead-free solder containing at least nickel of 0.1 to 0.2% by weight and organic fatty acid Solder bonding of semiconductor packages, electronic devices, etc. to copper lands on printed circuit boards using solder flux containing either one or more of nickel salt or organic fatty acid cobalt salt, either nickel or cobalt 1 By forming an alloy layer containing seeds or both metals at a thickness of 3 to 6 microns at the solder joint interface, microvoids in the solder joints generated when the semiconductor device is exposed to high temperatures for a long time (high temperature aging) Suppressed and improved impact rupture resistance of solder joints.

[2] In a semiconductor device, when soldering a mounting member to a printed circuit board using lead-free solder, it contains at least nickel of 0.1 to 0.2% by weight, and germanium ~ 0 · 0 Lead-free solder containing 1 wt%, organic fatty acid nickel salt or organic fatty acid cobalt salt The semiconductor device is exposed to high temperature for a long time by soldering to the soldering land of the printed circuit board, and forming an alloy layer containing one or both of nickel and cobalt at the solder joint interface thickness of 3-6 microns (High-temperature aging) A semiconductor device that suppresses microvoids in solder joints and improves the impact rupture resistance of solder joints.

[3] In the semiconductor device according to claims 1 and 2, the organic fatty acid dieckel salt or cobalt oxalate salt of the flux is selected from among nickel palmitate, cobalt panolemitate, nickel stearate, and cobalt stearate. Suppresses micro-voids in solder joints when semiconductor devices are exposed to high temperatures for a long time (high temperature aging) by soldering using one or more solder fluxes containing 1 to 5% by weight. And a semiconductor device for improving the impact rupture resistance of the solder joint.

[41 The method of manufacturing a semiconductor device, when the solder bonding the mounting member to a printed circuit board using a lead-free one solder, 0.1 at least nickel 0 1-0. Lead-free one solder containing 2 weight ^_0/0 And solder paste containing one or more of organic fatty acid nickel salt or organic fatty acid cobalt salt and used for semiconductor packages, electronic devices, etc. By soldering to the copper land of the printed circuit board and forming an alloy layer containing one or both of nickel and cobalt at the solder joint interface thickness of 3 to 6 microns, the semiconductor device can be used for a long time. A method of manufacturing a semiconductor device, characterized by suppressing a microphone opening void at a solder joint when exposed to high temperature (high temperature aging) and improving impact rupture resistance of the solder joint

In a method of manufacturing a semiconductor device, when soldering a mounting member to a printed circuit board using lead-free solder, at least nickel is contained in an amount of 0.1 to 0.2% by weight, and germanium is further added. ~ 0.1% by weight of lead-free solder and organic fatty acid nickel salt or organic fatty acid cobalt salt The semiconductor device is formed by bonding to the copper land of the printed circuit board and forming a thickness of 3 to 6 microns at the bonding interface including an alloy layer containing one or both of nickel and cobalt. It is characterized by suppressing the microvoids in the solder joints when exposed to high temperatures for a long time (high temperature aging) and improving the impact fracture resistance of the solder joints. The method of manufacturing a semiconductor device.

6. The method of manufacturing a semiconductor device according to claim 4 or 5, wherein the organic fatty acid bicker salt or the organic fatty acid cobalt salt of the flux is any one of panolemitic acid dieckel, palmitic acid cobanoleto, nickel stearate, and cobalt stearate. Micro-void voids in solder joints that occur when semiconductor devices are exposed to high temperatures for a long time (high temperature aging) by soldering using one or two types of solder flux containing 1 to 5% by weight. And a method for manufacturing a semiconductor device, wherein the impact rupture resistance of the solder joint portion is improved.