Joining of Metals by Super-Spread Wetting on Surface Fine Crevice Structure Created by Reduction-Sintering Copper Oxide Powder

Abstract

A surface fine crevice structure, created by laser irradiation, results in a region-selective super-spread wetting and the joining of metals is possible by taking advantage of this wetting. Through reduction-sintering of copper oxide (CuO) powder, we are able to obtain a sintered-Cu layer that possesses a complex sponge-like structure in which a porous network is formed. The wettability of molten tin (Sn) on the obtained sintered-Cu layer is investigated to confirm whether the super-spread wetting occurs on the surface fine crevice structure created by the method of reduction-sintering CuO powder. In addition, a joining experiment is accomplished using two Cu substrates. Therefore, a successful super-spread wetting of Sn on the surface fine crevice structure and the joining of Cu substrates are demonstrated.

Fig. 3 (a) Surface and (b) cross section of the surface fine crevice structure on a Cu substrate created by reduction-sintering of CuO powder.

1. Introduction

Tanaka and co-workers^1–3) discovered the super-spread wetting phenomenon, in which a molten metal penetrates and spreads by capillary action into the porous metal surface structure formed by the atmospheric oxidation-reduction method. They succeeded in metal-metal joining with a minimal formation of a fillet, which is the build-up of solder material, by using super-spread wetting.³⁾ The authors^4–7) found that the surface fine crevice structure created by laser irradiation on the Cu metal surface leads to region-selective super-spread wetting of molten Bi, Sn and solder materials, which could not be achieved with a porous metal surface structure created by atmospheric oxidation-reduction. Furthermore, they joined Cu substrates having the laser irradiated surface fine crevice structure through super-spread wetting. This process has the potential disadvantages of a longer processing time, in the case of a wide area, and of a lower energy absorption efficiency of the laser irradiation on a complex-shaped metal surface. Therefore, the technological innovation of laser processing and the development of a method other than laser treatment for the production of a surface fine structure, which will lead to a region-selective super-spread wetting, are required to expand the scope of applications for super-spread wetting. In terms of the development of a new method, an important point that should be considered is that the creation of a fine structure causes a capillary action, as the authors revealed.⁷⁾ An excellent candidate for this is the use of powder metallurgy for fabricating porous structures.

In this study, we propose the method of sintering CuO powder on a Cu substrate in a reducing atmosphere as a new method to create a surface fine crevice structure that will lead to a region-selective super-spread wetting. It is expected that the powder will easily spread in a specific area, and an adequate porous structure for super-spread wetting will be created by the gaps between the interconnected powders formed by the sintering. We investigated the wettability of a liquid material on the sintered Cu layer of the Cu substrate to confirm the super-spread wetting on the surface structure created by the method of reduction-sintering CuO powder. Furthermore, the joining of two Cu substrates by super-spread wetting of the joining material on the sintered Cu layer was attempted.

2. Surface Structure Produced by Reduction-Sintering of CuO Powder on a Cu Substrate

2.1 Experimental

A 20 × 15 × 1.5 mm Cu substrate (99.96% purity) and CuO powder (<10 µm, 98% purity) were used as a base material and sintered material, respectively. The average grain size of the CuO powder was confirmed to be approximately 3 µm by scanning electron microscopy (SEM). The CuO slurry was prepared by mixing the CuO powder with ethanol. The volume ratio of CuO powder to ethanol was 10 : 1. To apply the CuO slurry evenly on a Cu substrate, a metal sheet with a thickness of 0.1 mm with a 10 × 7.5 mm hole in the middle and a thin metal sheet like a razor blade were used as a frame and a doctor blade, respectively. Figure 1 illustrates the method for the application of a CuO slurry on a Cu substrate. After the frame was deposited on the Cu substrate polished with 600- and 1000-grit sandpaper, the hole of the frame was filled with the CuO slurry. Wiping the excessive CuO slurry by sliding the doctor blade on the surface of the frame enabled the top of the applied CuO slurry to be flattened. The sample was sintered and was simultaneously reduced at 1073 K for 1 h in an H₂ gas (99.995% purity) atmosphere.

Fig. 1

Illustration of the application of CuO slurry on a Cu substrate. The CuO-slurry-filled hole of the frame was flattened by sliding the doctor blade on the surface of the frame.

2.2 Results and discussion

Figures 2(a) and (b) show the appearance of the sample after reduction-sintering and the result analyzed by X-ray diffraction (XRD), respectively. As shown in the XRD pattern of Fig. 2(b), no peak of copper oxides is observed. The CuO powder was reduced into metal Cu in the applied reducing atmosphere and sintered metal Cu was well bonded to the Cu substrate without separation. The driving force of sintering between powder and substrate as well as that between powders is a reduction in surface energy, which leads to fill up the neck space formed between powder and substrate or powders.^8–10) Thus, it is considered that the sintering between Cu powder and Cu substrate proceeds at sufficient speed comparable to that between Cu powders. Figure 3 shows the SEM image of the surface and cross section of the obtained sample. It was observed that many open connected pores took up the sintered Cu layer of approximately 70 µm thickness. There were 2 types of pore sizes; micron scale pores and submicron scale pores, as shown in Fig. 3(a). It was supposed that the micron scale pores were formed with the formation of reticular joint structures interconnected at the ends of each Cu particle. Furthermore, the submicron size pores were generated by reducing the CuO and removing the oxygen from the particles.⁸⁾ From Fig. 3(b), it was observed that a high-density sintered layer was formed at the interface of the Cu sintered layer/Cu substrate, while a lower density sintered layer was in the upper part away from the interface. The dense sintered layer at the interface can contribute to the joint between the sintered Cu layer and Cu substrate. It was considered that the positional dependence of the microstructure of the sintered Cu layer was affected by the properties of the slurry with CuO powder.^11,12) It has been reported that the cake forming behavior and the distribution of the packing fraction in a cake formed by particles in a slurry varies owing to the concentration of solid in the slurry, the size of particles, the amount of dispersant and so forth. A low solid fraction, large particle size and large amount of dispersant result in a dense cake with homogeneity arising from the sedimentation of particles being well-dispersed in the slurry. The contrasting situation, that is, a high solid fraction, small particle size and small amount of dispersant, tends to form a network structure because the particles with a low dispersibility easily agglomerated in the slurry and agglomerates were deposited. In the present work, the slurry consisted of CuO powder of approximately 3 µm, ethanol and no dispersant, with the volume ratio between the CuO powder and ethanol as 10 : 1. Therefore, the state of the slurry used in this study corresponded to the latter case, where a network structure was created. In addition to the properties of the slurry, it was also noted that the weight of the particle affects the packing fraction by the compressive pressure. In general, the lower layer of the formed cake is compressed more than the upper layer, which creates a dense layer in the bottom part. Based on these facts, it was considered that the structure formed in this study, that is, a network structure in the upper part and dense layer in the lower part, attributed to the low dispersibility of the CuO powder in the slurry and the perpendicular-position dependence of the compressing pressure.

Fig. 2

(a) Appearance and (b) X-ray diffraction pattern of the sample after the reduction-sintering of CuO powder.

Fig. 3

(a) Surface and (b) cross section of the surface fine crevice structure on a Cu substrate created by reduction-sintering of CuO powder.

3. Wetting Experiment

3.1 Experimental

Two types of wetting experiments were attempted. First, to confirm whether super-spread wetting occurred on a surface fine crevice structure created by the reduction-sintering of CuO powder, we investigated the wettability of the sintered Cu layer of a Cu plate by liquid Sn. Sn, which is the main component of soldering materials with a melting temperature of 505 K, is completely melted at 673 K.¹³⁾ A piece of Sn (99.999% purity) was prepared by ultrasonic cleaning in ethanol. Sn was deposited on the 10 mm × 7.5 mm region of the sintered Cu layer on the Cu substrate, as shown in Fig. 4. Second, we investigated the wettability of the sintered Cu layer on a vertical wall of a Cu object to confirm whether super-spread wetting occurred on a surface fine crevice structure against gravity. As shown in Fig. 5(a), the sintered Cu layer was created on the surface of an L-shaped Cu object. A piece of Sn was deposited on the bottom part of the sintered Cu layer, as shown in Fig. 5(b). The two types of samples mentioned above were respectively set up in the furnace, which was designed to observe the inside from the top, as illustrated in Fig. 6. The furnace was first vacuumed to remove the air and then H₂ gas (99.995% purity) was introduced into the furnace to prevent oxidation of the sample surface during the experiment. The experiment was conducted under an H₂ gas flow of 20 mL/min. The sample was heated to 773 K for 50 min. Once the temperature reached 773 K, the sample was cooled to room temperature. The dynamic wetting process of molten Sn was observed during heating from the top of the furnace.

Fig. 4

Appearance of setting for the wetting experiment.

Fig. 5

(a) Appearance of the sample after making the sintered Cu layer and (b) illustration of the setting of the wetting experiment.

Fig. 6

Schematic diagram of the furnace used for the wetting and joining experiments.

3.2 Results and discussion

Figure 7 shows the appearance of the sample part-way through the spreading of molten Sn at 623 K and the sample after the wetting experiment on the sintered Cu layer of the Cu substrate. It was confirmed that the Sn spread out in every direction only within the sintered Cu layer, and consequently, the whole area of the sintered Cu layer became wetted with Sn. Dark gray colored part corresponds to the initial position of Sn. It is assumed that the oxidized film formed on the surface of Sn by a slight amount of oxygen in the atmosphere or Sn itself was left on the initial Sn position. The contact angles of solid Cu and liquid Sn are 37°¹⁴⁾ at 510 K, and 23°¹⁵⁾ and 25°¹⁶⁾ at 673 K, which meant that a certain contact angle of more than 0° was present at the present experimental temperature of 773 K. Therefore, it was assumed that the molten Sn did not spread out completely on the whole surface of the normal flat Cu substrate. Thus, it was proposed that the wetting of Sn on the sintered Cu layer occurred by the same phenomenon that led to the wettability of the surface fine crevice structure created by the laser-irradiation method, that is, super-spread wetting. Therefore, it was expected that Cu–Cu joining by using the sintered Cu layer was also possible. From the cross-section of the sample shown in Fig. 8, it was observed that the intermetallic compounds of Cu₃Sn and Cu₆Sn₅ were formed at the interface between Sn and the Cu substrate. From phase diagram of the Cu–Sn¹⁷⁾ binary system, it is predicted that intermetallic compound Cu₄₁Sn₁₁ can be formed by the reaction between Cu and Cu₃Sn at our experimental temperature of 773 K. However, the Cu₄₁Sn₁₁ phase is not observed in Fig. 8. It was reported that the formation and growth of the Cu₄₁Sn₁₁ phase is considerably slow.¹⁸⁾ Therefore, we assume that the Cu₄₁Sn₁₁ was not formed, or was present but unobservable due to the short holding time at high temperature range. Considering that the thickness of part of the intermetallic compound was almost the same as the thickness of the sintered Cu layer of approximately 70 µm, it was supposed that part of the sintered Cu layer changed into intermetallic compounds after wetting of Sn. Figure 9(a) shows the surface of the edge part of the sintered Cu layer wetted with Sn. A region where the Sn was not completely filled was partially observed. As a result of the magnified observation of the surface of the region that was not completely filled with Sn, as shown in Fig. 9(b), the agglomeration of parts wetted with Sn, and an empty gap around the agglomerates were observed. It was proposed that if enough Sn was supplied, the Sn first should wet the pores of the surface fine crevice structure observed in Fig. 9(c), then wet the empty gap around the agglomerates observed in Fig. 9(b) and finally spread to cover the surface fine crevice structure.

Fig. 7

Appearance of the sample (a) part-way through wetting, and (b) after the wetting experiment.

Fig. 8

Cross section of the sintered Cu layer after the Sn wetting experiment.

Fig. 9

(a) Surface of the edge part of the sintered Cu layer wetted with Sn and (b) magnified image of the region that was not completely filled with Sn, and (c) original surface fine crevice structure of the sintered Cu layer.

The sample after the wetting experiment of Sn on the sintered Cu layer on an L-shaped Cu object is shown in Fig. 10. The metallic gray area corresponded to a Sn-wetted area. Liquid Sn completely spread out within the sintered Cu layer and reached the top of the vertical wall of the L-shaped sintered Cu layer. It was found that the super-spread wetting with Sn on the sintered Cu layer occurred not only in the horizontal direction but also in the opposite direction to gravity. This result indicated that the super-spread wetting with Sn was achievable on a sintered Cu layer created on any complex-shaped Cu object.

Fig. 10

Appearance of the sample after the wetting experiment on the sintered Cu layer of an L-shaped Cu block.

4. Cu–Cu Joining Using Super-Spread Wetting of Sn on a Surface Fine Crevice Structure

4.1 Experimental

Two types of joining experiments were attempted using super-spread wetting caused on the surface fine crevice structure created by the reduction-sintering of CuO powder. Two types of Cu substrates with a sintered Cu layer were prepared. As illustrated in Fig. 11(a), the sintered Cu layers were formed on a 15 × 10 mm region of the 25 × 15 mm Cu substrate and on one entire side of the 10 × 5 mm Cu substrate by the same method as mentioned in Section 2.1. The two Cu substrates were set in a furnace by contacting them with each sintered Cu layer, and a piece of Sn (99.999% purity) was placed on the sintered Cu layer part of the lower Cu substrate.

Fig. 11

Illustrations of the joining experiment. (a) The sintered Cu layers of the upper and lower Cu substrates are in contact with each other, and Sn is placed on the sintered Cu layer part of the lower Cu substrate. (b) The sintered Cu layer of the upper Cu substrate and Sn are respectively placed on the upper step and lower step of the sintered Cu layer of stair-shaped Cu block.

In another experiment, a Cu block and a stair-shaped Cu object were prepared. The sintered Cu layers were formed on one entire side of the 10 × 5 × 1.5 mm Cu block and on a 10 mm width band area, which ranged from the lower step to the upper step, on the surface of the stair-shaped Cu object with approximately a 25 mm step length, 15 mm step width and 25 mm step height, as illustrated in Fig. 11(b). The Cu block and a piece of Sn were respectively placed on the upper step and lower step of the sintered Cu layer of the stair-shaped Cu object.

These joining experiments were also carried out under the same conditions as the wetting experiment mentioned above. After vacuuming the furnace to remove any residual air, H₂ gas (99.995% purity) was introduced into the furnace. By maintaining an H₂ gas flow of 20 mL/min, the samples were heated to 773 K for 50 min. After heating, the sample was cooled to room temperature.

4.2 Results and discussion

Figure 12 shows the sample after the joining experiment with two Cu substrates. As shown in Fig. 12(a), it was confirmed that the greater part of the Sn moved from its original location to spread on the sintered Cu layer. Furthermore, the entire region of the sintered Cu layer between the upper and lower Cu substrates was wetted with Sn. Even when the sample was inverted, the upper Cu substrate did not fall off, and furthermore, a fillet was not formed at the junction, as shown in Fig. 12(b). The cross-section of the junction is shown in Fig. 13. It was observed that the space between the two Cu substrates was filled with Sn, which indicated that the molten Sn penetrated into the sintered Cu layer of the upper substrate against gravity by super-spread wetting. Thus, it was proposed that the two Cu substrates were successfully joined by the penetration of molten Sn into both the lower and upper substrates and the filling of Sn into all the pores of the sintered Cu layer. The several different colored phases observed at the joint were Cu₃Sn and Cu₆Sn₅ intermetallic compounds. It was considered that the most sintered Cu layer, that is, the surface fine crevice structure part, was transformed into intermetallic compounds. In addition, not only the intermetallic compounds were formed on the interface between Sn and Cu substrate, but also a lot of granular intermetallic compounds were interspersed with the middle layer of Sn between the 2 Cu substrates. Comparing the Fig. 8 and Fig. 13, it was found that more granular intermetallic compounds were formed in the joining experiment sample, which have the thick layer of spread Sn. It was supposed that the granular intermetallic compounds were crystallized out from liquid Sn–Cu alloy during the cooling. From these results, it was confirmed that the joining of Cu substrates using the surface fine crevice structure created by reduction-sintering of CuO powder can lead to the super-spread wetting of Sn. Additionally, this joining by using the surface fine crevice structure created by the reduction-sintering CuO powder occurred with the same joining mechanism as that observed for the laser irradiation method.

Fig. 12

(a) Appearance of the sample after the 2 Cu substrates joining experiment, and (b) inverted sample.

Fig. 13

Cross section of the junction of the sample.

Figure 14 shows the sample after the joining experiment of a Cu substrate with a stair-shaped Cu object. The metallic gray area was the Sn-wetted area. As well as in the wetting experiment of Sn on the sintered Cu layer on an L-shaped Cu object, Sn wetted and rose against gravity throughout the sintered Cu layer on the vertical wall of the Cu object to reach the top step. From Fig. 14(b), it was confirmed that the upper Cu did not fall off even when the sample was inverted. We observed that the Cu block was successfully joined to the stair-shaped Cu object by super-spread wetting of Sn on the sintered Cu layer.

Fig. 14

(a) Appearance of the sample after the joining experiment on the stair-shaped Cu block, and (b) inverted sample.

5. Conclusion

In this study, we propose a reduction-sintering CuO powder method for creating a surface fine crevice structure. To confirm whether super-spread wetting occurs on the structure created by the reduction-sintering of CuO powder method, we investigated the wettability of the molten Sn on the sintered Cu layer. Furthermore, the joining of two Cu substrates was attempted. These experiments led to the following findings.

1. (1)    A surface fine crevice structure, which consists of pores formed by the formation of a joint structure interconnecting Cu particles and a loss of oxygen from the particles, is created by the method of reduction-sintering of CuO powder.
2. (2)    A super-spread wetting of the molten Sn occurs on the sintered Cu layer of a vertical wall against the gravitational force as well as on the horizontal surface.
3. (3)    The two Cu substrates with the sintered Cu layer are successfully joined by using the super-spread wetting of molten Sn. In addition, the joining of the Cu block on the upper step of the stair-shaped Cu object is achieved by the super-spread wetting of molten Sn starting from the bottom step of the Cu object.

REFERENCES

• 1)   N.  Takahira,  T.  Tanaka,  S.  Hara and  J.  Lee: Mater. Trans. 46 (2005) 3008–3014.
• 2)   N.  Takahira,  T.  Yoshikawa,  T.  Tanaka and  L.  Holappa: Mater. Trans. 48 (2007) 2708–2711.
• 3)   N.  Takahira,  T.  Yoshikawa,  T.  Tanaka and  L.  Holappa: Mater. Trans. 48 (2007) 3126–3131.
• 4)   A.  Fukuda,  H.  Matsukawa,  H.  Goto,  M.  Suzuki,  M.  Nakamoto,  R.  Matsumoto,  H.  Utsunomiya and  T.  Tanaka: Mater. Trans. 56 (2015) 1852–1856.
• 5)   M.  Nakamoto,  A.  Fukuda,  J.  Pinkham,  S.  Vilakazi,  H.  Goto,  R.  Matsumoto,  H.  Utsunomiya and  T.  Tanaka: Mater. Trans. 57 (2016) 973–977.
• 6)   A.  Fukuda,  Y.  Satake,  H.  Goto,  M.  Suzuki,  M.  Nakamoto,  R.  Matsumoto,  H.  Utsunomiya and  T.  Tanaka: J. Smart Process 5 (2016) 153–158.
• 7)   V.  Siboniso,  J.  Yeon,  C.  Grozescu,  H.  Goto,  M.  Nakamoto,  R.  Matsumoto,  H.  Utsunomiya and  T.  Tanaka: Mater. Trans. 58 (2017) 1227–1230.
• 8)   G.C.  Kuczynski: Trans. Am. Inst. Mining Met. Engrs. 185 (1949) 169–178.
• 9)   C.  Herring: J. Appl. Phys. 21 (1950) 301–303.
• 10)   G.C.  Kuczynski: J. Appl. Phys. 21 (1950) 632–635.
• 11)   H.-J.  Kim,  T.  Mori and  J.  Tsubaki: J. Ceram. Soc. Jpn. 113 (2005) 761–767.
• 12)   H.  Yamakawa,  K.  Ato and  H.  Sakamoto: Funtai Gijutsu. 3 (2011) 40–44.
• 13)   H.  Ohtani,  K.  Okuda and  K.  Ishida: J. Phase Equilibria 16 (1995) 416–429.
• 14)   Y.  Gu,  P.  Shen,  N.-N.  Yang and  K.-Z.  Cao: J. Alloys Compd. 586 (2014) 80–86.
• 15)   G.  Bailey and  H.  Watkins: J. Inst. Met. 80 (1951) 57–76.
• 16)   Q.Q.  Lai,  L.  Zhang and  N.  Eustathopoulos: Acta Mater. 61 (2013) 4127–4134.
• 17)   H.  Okamoto: J. Phase Equilibria Diffus. 35 (2014) 208–219.
• 18)   J.  Liu,  H.  Zhao,  Z.  Li,  X.  Song,  Y.  Zhao,  H.  Niu,  H.  Tian,  H.  Dong and  J.  Feng: Mater. Charact. 135 (2018) 238–244.