Comparison of melting efficiency between blue, green, and IR lasers in pure copper welding

In recent years, the transition from motorized vehicles to electric and hybrid vehicles is currently taking place to realize a carbon-neutral society. In order to electrify vehicles, batteries to store electricity, motors to drive the vehicle, and power devices to control them are installed in the vehicle, and copper devices with high electrical and thermal conductivity are widely used for these components. Compared to motorized vehicles, electric vehicles use about three times as much copper. For the higher performance of electric vehicles, it is important to increase the efficiency of conversion from electrical energy to kinetic energy, and an advanced copper wiring joining technology is required. Tungsten inert gas (TIG) welding and swaging are the most common copper joining methods. However, TIG welding and swaging have problems like slow processing speed and large joints.¹ To solve these problems of conventional joining methods, laser welding is expected to be applied. Laser welding is also used for joining automobile bodies and can be applied to join many metals such as stainless steel and nickel alloys by using a laser in the near-infrared region.^2–4 In laser processing, optical absorption for materials is important. Although near-infrared lasers have high optical absorption for Fe and Ni, the optical absorption for copper is low at 5%–10%.⁵ Therefore, in recent years, research and development of higher output power green and blue lasers have been underway.^6–8 Blue and green lasers have a high optical absorption rate of about 60% for copper and are expected to enable efficient pure copper welding.

Although there have been an increasing number of reports in recent years on laser welding of pure copper using visible light lasers,^6–11 there have been no reports comparing blue diode lasers, IR disk lasers, and green disk lasers under uniform laser irradiation conditions. This is because the beam quality of blue diode lasers and disk lasers is different, making it difficult to compare them with the same output power and spot size. In our research, we have developed a high-brightness blue diode laser with high beam quality that enables comparison of welding phenomena at the same power and spot size as disk lasers.¹² Energy density, power divided by the spot diameter, is an important parameter in laser processing. Low energy density results in thermal conduction mode welding, while high energy density results in keyhole mode welding. In the thermal conduction mode, the laser beam is absorbed only at the surface of the material, resulting in a small weld with a small aspect ratio, while in the keyhole mode, multiple reflections produce a weld with deep penetration and a high aspect ratio. The optical absorption rate is measured at room temperature, and the wavelength dependence of the optical absorption rate during welding is not known. In our previous study, we have already reported various phenomena of keyhole mode welding of pure copper using a blue diode laser with an output power of 1.5 kW,¹² and a comparison of melting efficiency with a disk laser is required.

In this study, bead-on-plate welding is performed on a pure copper plate using a blue diode laser, an IR disk laser, and a green disk laser, and the melting efficiencies in the keyhole mode and the thermal conduction mode are compared. The pure copper plate is oxygen-free copper with a thickness of 2 mm, and the laser spot diameter is 300, 600 μm for the blue diode laser, 300, 600 μm for the IR disk laser, and 300 μm for the green disk laser at the focal point. The melting behavior is observed using a high-speed video camera, and the melting efficiency is determined from cross-sectional observation of the pure copper plate after welding.

Figure 1 shows the laser systems and experimental setup of (a-1, a-2) blue diode laser, (b-1, b-2) IR disk laser, and (c-1, c-2) green disk laser. The blue diode laser had a maximum output of 1.5 kW and was constructed by spatially coupling three 500 W blue diode lasers (B500-DM20-DW, SHIMADZU CORPORATION) using a beam combiner system. The 500 W blue diode laser had a 200 μm fiber core diameter. The fiber core diameter and numerical aperture (NA) of the 1.5 kW blue diode laser were 400 μm and 0.2, respectively. An IR disk laser with a maximum output of 16 kW and a wavelength of 1030 nm (Trudisk 16002, Trumpf. Ltd.) was employed. The laser was guided to the processing head with an optical fiber having a core diameter of 200 μm and an NA of 0.1. The IR disk laser beam was set to an angle of incidence of 10° to prevent it from being damaged by the reflected laser light. A green disk laser had a maximum output power of 3 kW and a wavelength of 515 nm (Trudisk 3022, Trumpf. Ltd). The fiber core diameter and the NA of the green disk laser were 150 μm and 0.1, respectively.

Table I shows experimental conditions. The spot size was measured with a beam profiler (BPF-S400 and BPF-S1000, BPF laser innovation Corp.). The processing point was the laser focal point, and a processing head with optics to make the spot diameter 300 μm was used for each laser. When the spot diameter of the IR disk laser is 300 μm, the welding is only in the keyhole mode regardless of the output power, and pure copper welding in the thermal conduction mode cannot be compared. Therefore, an optical system with a spot diameter of 600 μm was also used. Since the optics for the green disk laser with a spot diameter of 600 μm could not be prepared, pure copper welding with a spot diameter of 600 μm was compared using the blue diode laser and the IR disk laser. The output power was varied from 1.5 kW for the blue diode laser and green disk laser to 8 kW for the IR disk laser. A pure copper plate (>99.96%, oxygen-free-copper, Nilaco), which measured 10^w× 30^l× 2^t mm³, was used. The blue diode laser was focused on a pure copper plate and irradiated vertically at the focal point. The pure copper plate was placed on a scanning stage, and Ar was used as a shielding gas. The scanning velocity and shielding Ar gas flow rate were 25 mm/s and 20 l/min, respectively. To fix irradiation conditions other than the laser wavelength, the same jig for fixing the pure copper plate and shielding gas nozzle was used in all experiments.

Figure 2 shows microscopy images of the cross-sectional copper plate at an output power of (a) 1.5 kW of the blue diode laser and (b) 7.4 kW of the IR disk laser with a spot size of 600 μm. The yellow dashed line indicates the melting area. The weld cross sections of both the blue diode laser and the IR disk laser are bowl-shaped, indicating that welding is in the thermal conduction mode.

Figure 3 shows the relationship between output power and (a) bead width, (b) penetration depth, (c) melted area, and (d) melting efficiency. Blue and red full circles indicate the blue diode laser and the IR disk laser, respectively. When a blue diode laser was used, the pure copper surface melted at an output of 0.95 W or higher, and the bead width, melting depth, and melting area increased with increasing output power. The experimental results show that as the output power increased by 0.1 kW, and the penetration depth and area increased by 20 μm and 0.009 mm², respectively. The aspect ratio, which is the melt depth divided by the bead width, is averaged 0.23. With the IR disk laser, the surface is melted at an output power of 7.4 kW or higher but transitioned to the keyhole mode at an output power of 8 kW or higher; the IR disk laser could not control the melting depth or area while maintaining the heat conduction mode. In terms of melting efficiency, the blue diode laser had a higher melting efficiency than the IR disk laser under all conditions; for the IR disk laser, the melting efficiency was around 0.17%, regardless of the output power, but for the blue diode laser, the melting efficiency increased with increasing output power. It is thought that the high thermal conductivity of pure copper causes heat to diffuse from the laser irradiation point to the pure copper substrate, resulting in a smaller melting area and lower melting efficiency. In the thermal conduction mode, laser light absorption occurs only at the topmost surface of pure copper, and the difference in optical absorption by laser wavelength is thought to be reflected in difference in melting efficiency.

Figure 4 shows microscopy cross-sectional images for blue diode laser irradiation at an output power of (a) 1.0 kW and (b) 1.5 kW, IR disk laser irradiation at the output power of (c) 3.25 and 3.4 kW, and green disk laser irradiation at the output power of (d) 1.0 and (e) 1.5 kW. The spot size of all lasers was kept at 300 μm. The yellow dashed line indicates the melting area. At 1 kW output power, the weld cross sections using the blue diode laser and the green disk laser both showed thermal conduction welding, and the bead width was 370 μm for both. The melt area was 0.026 mm² for the green disk laser and 0.038 mm² for the blue diode laser. When the output power was 1.5 kW, both the blue diode laser and the welded cross section with the green disk laser showed keyhole mode welding. The melt area was 0.77 mm² for the green disk laser and 0.63 mm² for the blue diode laser. The bead width and penetration depth were also larger for the green disk laser than for the blue diode laser. When using the IR disk laser, no melting occurred at an output power of 3.25 kW, but keyhole mode welding occurred at an increased output power of 50 W, 3.3 kW. The melted area was large but underfilled, which was attributed to jump in optical absorptivity due to the transition to the keyhole mode.

Figure 5 shows a relationship between the output power and (a) bead width, (b) penetration depth, (c) melted area, and (d) melting efficiency. Blue and red full circles and the green empty circle indicate the blue diode laser, the IR disk laser, and the green disk laser, respectively. The focal spot size of each laser was 300 μm. Focusing on bead width, penetration depth, and melted area, the pure copper surface was melted at an output power of 0.9 kW and higher with the blue diode laser and green disk laser, and both were thermal conduction mode welding up to 1 kW. At output powers of 1.1 kW and above, the welding transitioned to keyhole mode welding. In the thermal conduction mode, the bead widths were comparable, but the penetration depth and melt area were about 10% greater when the blue diode laser was used. At output powers of 1.1 kW and higher, the green disk laser had about 20% greater penetration depth and melt area than the blue diode laser. When using optics with a spot diameter of 300 μm, the blue diode laser has a light input with a divergence angle of ±15°, while the green disk laser has a smaller divergence angle of ±3.8°. In keyhole mode welding, it has been reported that multiple reflections differ depending on the angle of incidence of the light, which may be caused by factors other than the difference in absorptivity due to wavelength. With the IR disk laser, the pure copper surface did not melt up to 3.25 kW output power but transitioned to keyhole mode welding above 3.3 kW output power, and no thermal conduction mode welding occurred. Bead width, penetration depth, and melted area tended to increase with increasing output power when using the IR disk laser, but had large deviations and low linearity. This is thought to be due to the lack of welding stability. For the blue diode laser and green disk laser, these values increased linearly with the output power.

The melting efficiencies were similar for melting efficiencies from 1 to 1.5 kW power output with the blue diode laser and green disk laser and from 3.3 to 3.75 kW output with the IR disk laser. Melting efficiencies in keyhole mode welding ranged from 2% to 7%, 10 times higher than those obtained with thermal conduction mode welding; with the IR disk laser, the maximum thermal efficiency was 6.7% at output powers above 3.75 kW because the molten part penetrated the copper plate. When the blue diode laser or green disk laser was used, the welding depth was 1–1.3 mm even at an output power of 1.5 kW, and the melting efficiency was equal to or better than that of an infrared disk laser with an output power of 3.75 kW, where the welding depth was 2 mm. The melting efficiency tends to increase as the welding penetration depth increases. It is expected that the welding depth will reach 2 mm with a higher output power blue diode laser or green disk laser, and the melting efficiency will be much greater than with the IR disk laser.

Figure 6 (Multimedia view) shows a high-speed video observation with (a) the blue diode laser, (b) the IR disk laser, and (c) the green disk laser. The output power of the blue diode laser and the green diode laser was 1.25 kW and that of the IR disk laser was 3.4 kW. The spot size was 300 μm. A camera was used to observe the area around the point where the cross-sectional observation was performed, which was taken from 2.5 to 5 mm from the weld start point. The playing speed of the video is 1/250. When the blue diode laser is used, a keyhole is formed at the laser-irradiated area, and it can be seen that smaller sputters are scattered from the molten pool to the surrounding area. When the green disk laser is used, a keyhole is formed at the laser irradiated area and a larger spatter is scattered from behind the molten pool; when the IR disk laser is used, a keyhole is formed at the laser irradiated area, but the molten pool periodically explodes and scatters. No keyhole is observed after the molten pool has largely scattered, and the keyhole is formed again after a while. It was found that the welding behavior of pure copper differs greatly among the blue diode laser, the green disk laser, and the IR disk laser. Further investigation of these differences in behavior is needed, and x-ray transmission observation of keyholes is considered to be a suitable method.

Bead-on-plate welding was performed on a pure copper plate using a blue diode laser, an IR disk laser, and a green disk laser, and the melting efficiencies in the keyhole mode and the thermal conduction mode were compared. Comparative experiments were conducted using optical systems with spot diameters of 300 and 600 μm for each laser, and under uniform conditions of pure copper plate size, jigs, and shielding gas nozzles. As a result, comparing the blue diode laser and the IR disk laser, the blue diode laser had a better controllability of the amount of melting in thermal conduction mode welding, and the melting efficiency was about four times higher than that of the IR disk laser. In keyhole mode welding, the melting efficiencies were equal under conditions where the melting area was less than half, indicating that the melting efficiency was higher with the blue diode laser, which has a higher output power than 1.5 kW, than with the IR disk laser. Comparing the blue diode laser and the green disk laser, the blue diode laser resulted in a larger melt area and melting efficiency for thermal conduction mode welding, but the green disk laser was larger for keyhole mode welding. This is thought to be due to the difference in the beam focusing angle, but a more detailed investigation, including observation of the keyhole shape, is required.

This work was partly supported by the New Energy and Industrial Technology Development Organization (NEDO) of Japan. We are grateful to NICHIA CORPORATION for the development of the blue diode laser. The authors would like to thank Bastian Becker, TRUMPF Corporation, for using a 3 kW green disk laser.

The authors have no conflicts to disclose.

Keisuke Takenaka: Writing – original draft (equal). Yuji Sato: Formal analysis (equal); Validation (equal); Writing – review & editing (equal). Shumpei Fujio: Formal analysis (equal); Visualization (equal). Masahiro Tsukamoto: Funding acquisition (equal); Project administration (equal); Supervision (equal).