Power Semiconductor Module Busbar Welding Using a Single-Mode Fiber Laser (Part 1): Heat-Affected Zone Formation and Solidification Cracking Behavior during High-Speed Multi-Pass Welding

Article information

J Weld Join. 2026;44(3):205-215

Publication date (electronic) : 2026 June 30

doi : https://doi.org/10.5781/JWJ.2026.44.3.1

Ye-Ji Yoo^*, Keun-Jae Lee^**, Se-Min Park^**, Pil-Kyong Oh^**, Eun-Joon Chun^*,

* Department of Materials System Engineering, Pukyong National University, Busan, 48513, Korea

** Power Semiconductor Module Development Team, Hyundai Mobis Co., Ltd., Uiwang, 16082, Korea

†Corresponding author: ejchun@pknu.ac.kr

Received 2026 May 15; Revised 2026 June 5; Accepted 2026 June 8.

Abstract

1. Introduction

The electric vehicle market has entered a temporary demand slowdown phase, often referred to as the chasm, intensifying technological competition and driving the need for the development of high-performance, high-power electrification systems as a key strategy for overcoming this challenge. This trend is not limited to electric vehicles (EVs) but is expanding across the entire spectrum of electrified vehicles, including hybrid electric vehicles (HEVs), in line with global policy initiatives aimed at achieving carbon neutrality. In particular, as the automotive industry transitions toward a Software Defined Vehicle (SDV) architecture, in-vehicle power consumption is expected to increase significantly compared with conventional vehicles. Consequently, ensuring the integrity of power transmission paths and their associated joints for stable power delivery has emerged as a critical technical challenge1).

Power modules, which serve as the core power conversion units in electric vehicles, are essential systems that efficiently convert and control the electrical power required for vehicle propulsion through the use of power semiconductor devices. Power semiconductors are semiconductor devices responsible for power switching and conversion, enabling precise control of power flow under high-voltage and high-current operating conditions. A power module is therefore an integrated assembly consisting of power semiconductor devices, switching elements, substrates, busbars, and packaging structures, functioning as a key component in electric vehicles and power conversion systems2). Recently, power modules have been rapidly transitioning from conventional silicon (Si)-based power semiconductors to next-generation silicon carbide (SiC)-based devices. Although this transition enables reduced switching losses and high-voltage operation, it also leads to increased power density, resulting in higher heat generation and operating temperatures within the module. As the power output capability continues to increase, power modules are required to exhibit enhanced thermal stability and structural reliability3).

Representative welding and joining technologies used in power module packaging processes are summarized in Table 1. Soldering4), sintering5), and epoxy bonding6) are widely employed for joining semiconductor chips, ceramic insulators, and substrates within power modules. For busbar assembly connecting power modules to external power sources, ultrasonic welding and bolt fastening have traditionally been the predominant approaches. Ultrasonic welding enables joining with minimal thermal influence through friction and plastic deformation induced by high-frequency vibration7), whereas bolt fastening offers ease of assembly and disassembly through a mechanical joining mechanism but suffers from limitations associated with increased contact resistance and long-term reliability concerns8). To overcome the limitations of these conventional processes, laser-based joining technologies capable of simultaneously achieving high precision and high productivity have recently attracted increasing attention9). Among these technologies, single-mode fiber lasers provide high energy density and excellent beam quality, enabling high-speed welding and enhanced productivity10). They also offer metallurgical advantages, including the formation of weld metal under rapid heating and cooling conditions11) and the ability to suppress the formation of the heat-affected zone (HAZ) more effectively than other laser sources12). Since semiconductor chips and insulating structural components within power modules are generally composed of heat-sensitive materials, the suppression of HAZ formation and the minimization of thermal damage and distortion to surrounding components through the application of single-mode lasers are considered significant advantages13).

Table 1

Summary of power module packaging and assembly processes

As the power density of power semiconductors continues to increase, busbar welds and joints are required not only to provide electrical conduction but also to maintain reliable performance under high-current operating conditions, mechanical vibration, and thermal cycling environments. In particular, insufficient joint area can result in localized increases in electrical resistance, leading to concentrated heat generation. Over time, this may cause joint degradation and reduced reliability of the power module14). Therefore, the application of single-mode fiber lasers to power module busbar welding is expected to require large-area multi-pass welding techniques to overcome the inherently narrow bead dimensions characteristic of this laser source.

However, when implementing large-area Cu-Cu laser welds between busbars, weld integrity and HAZ behavior are expected to vary significantly depending on laser processing conditions and scanning patterns. Despite this, systematic experimental investigations remain limited. Furthermore, frequent occurrences of welding cracks have been reported in Cu-Cu laser welds15). Therefore, a detailed evaluation from the perspective of weld cracking is necessary to ensure the applicability of large-area laser welding to power module busbars and to achieve uniform electrical conductivity.

Accordingly, this study aims to establish weld integrity and optimal process guidelines for Cu-Cu busbar welding in power modules using a single-mode fiber laser. In this first part of the study, a systematic investigation was conducted focusing on HAZ formation behavior, joint strength, and weld cracking characteristics during large-area welding. In addition, a thermodynamics-based thermal analysis approach was employed to examine weldability in relation to the rapid heating and cooling behavior inherent to the high-speed welding process of single-mode lasers. By correlating experimental and computational results, fundamental processing conditions for single-mode laser welding are proposed for the fabrication of highly reliable power module busbar joints.

2. Materials and Experimental Procedures

2.1 Materials

Table 2 presents the chemical composition of the materials used in this study. High-purity C1100P Cu sheets with dimensions of 100 × 30 × 0.8 mm were employed as the base material.

Table 2

Chemical composition of materials used (mass%)

2.2 Experimental Procedures

2.2.1 Single-Mode Fiber Laser Welding

Fig. 1 illustrates the schematic configuration of the laser welding system and specimen arrangement, while the detailed welding conditions are summarized in Table 3. A single-mode fiber laser was used for the welding experiments. During welding, the laser power, laser beam diameter on the specimen surface, and scanning speed were controlled. A total of five welding conditions were selected, corresponding to heat inputs of 3.5, 4.0, 4.5, 6.7, and 20.0 J/mm, with energy densities of 121.8, 134.0, 148.9, 223.4, and 670.1 J/mm², respectively. Two welding patterns were employed: a single-pass linear weld with a length of 15 mm and a multi-pass weld covering an area of 8 × 6.3 mm² using a rectangular spiral (ㄹ-shaped) scanning pattern. For multi-pass welding, the overlap ratio along the scanning path was selected as a primary process variable in addition to the welding conditions described above. Four overlap ratio levels were investigated. To protect proprietary process information, these conditions are designated as Condition A (Low), Condition B (Medium), Condition C (Medium High), and Condition D (High). For all welding experiments, a clamping jig was used to ensure intimate contact between the upper and lower materials without any gap.

Fig. 1

Schematic illustrations of (a) the single-mode fiber laser welding system, (b) the weld specimen configuration, and (c) the laser beam scanning patterns used for single-pass and multi-pass welding

Table 3

Specific conditions for single-mode fiber laser welding

2.2.2 Microstructural Analysis and Mechanical Property Evaluation

After sectioning, the welded specimens were sequentially ground using SiC abrasive papers (#400-#2000), followed by final polishing with 3 μm and 1 μm diamond paste. For microstructural observation, the polished specimens were etched at room temperature for 30 s using a solution consisting of 50 mL distilled water and 50 mL nitric acid (HNO₃, 68-70%). Microstructural characterization of the welds was performed using an optical microscope (OM; DR IRM, Leica). Regions containing weld cracks were further examined using an Electron Probe Micro Analyzer (EPMA; JXA-8530F, JEOL) to identify and characterize the cracking features.

The mechanical properties of the laser welds were evaluated through shear tensile testing. The test specimens were prepared according to the configuration shown in Fig. 1(b), and the welding conditions were identical to those listed in Table 3. A crosshead speed of 3 mm/min was applied during testing. The maximum tensile load obtained from the load-displacement curve was used as the indicator of weld mechanical performance. For each welding condition, the maximum tensile load was measured five times, and the results were compared accordingly.

3. Experimental Results and Discussion

3.1 Macrostructures of Single-Pass Cu-Cu Power Module Busbar Welds

Fig. 2 presents the surface and cross-sectional macrostructures of single-pass linear weld beads. The welding results obtained under heat inputs of (a) 20.0 J/mm, (b) 6.7 J/mm, and (c) 4.0 J/mm are shown in Fig. 2. As the heat input decreased, the weld bead width was reduced to approximately 0.22 mm, as shown in Fig. 2(c), while a penetration depth of approximately 1.0 mm was maintained. Despite the high beam reflectivity of Cu, the high energy density characteristic of the single-mode fiber laser (134.0 J/mm²) enabled relatively deep penetration even under the ultra-low heat input condition of 4.0 J/mm. One of the key objectives of this study was to achieve HAZ-free welding capable of minimizing thermal damage to adjacent components and regions surrounding the busbar. Therefore, in the large-area multi-pass welding experiments described in the following sections, the heat input condition shown in Fig. 2(c) (4.0 J/mm, corresponding to approximately 25% penetration into the lower material) was selected as the reference condition, and the heat input was varied slightly to investigate HAZ behavior during multi-pass welding.

Fig. 2

Bead surface and cross-sectional macrostructures of single-pass Cu-Cu welds obtained under different heat inputs, (a) 20.0 J/mm, (b) 6.7 J/mm, and (c) 4.0 J/mm

3.2 Effects of Heat Input and Overlap Ratio on the Weldability of Large-Area Multi-Pass Cu-Cu Power Module Busbar Welds

Based on the heat input condition of 4.0 J/mm identified in Section 3.1, the heat input was varied by ±0.5 J/mm to evaluate the effects of overlap ratio and heat input on weld integrity during large-area welding using the rectangular spiral scanning pattern. Weld integrity was assessed from three perspectives: weld cracking, HAZ formation behavior observed in macrostructures, and maximum load obtained from shear tensile testing.

3.2.1 Solidification Cracking Behavior of Large-Area Multi-Pass Welds as a Function of Overlap Ratio

Fig. 3 compares the solidification cracking behavior observed under different overlap ratio conditions at a heat input of 4.0 J/mm. Among the overlap ratio conditions examined in this study, weld surface cracks were observed in all conditions except Condition A. Fig. 4 presents the results of EPMA analysis performed on the fracture surface of a representative crack formed under Condition B.

Fig. 3

Solidification cracking behavior of multi-pass Cu-Cu weld bead surface at different overlap ratios: conditions of (a) A, (b) B, (c) C and (d) D (Heat input: 4.0 J/mm)

Fig. 4

Representative SEM images of the solidification cracking fracture surface observed in multi-pass Cu-Cu welds: (a) low-magnification (50×), (b) high-magnification (500×), and (c) corresponding elemental distribution obtained by EPMA analysis

The fracture surface exhibited traces of thin liquid films associated with dendritic solidification structures, and microsegregation of trace impurity elements remaining in the C1100P Cu was also observed. Based on these observations, the surface cracks identified under the high-overlap conditions (Conditions B, C, and D) in Fig. 3 were classified as solidification cracks. Even under the ultra-low heat input conditions characteristic of single-mode laser welding, repeated thermal expansion and solidification shrinkage can generate residual stresses as successive passes are deposited during high-overlap multi-pass welding. These accumulated strains are believed to have promoted crack formation within the weld solidification temperature range. Accordingly, the present results indicate that large-area single-mode laser welding of power module busbars is susceptible to solidification cracking and that minimizing the overlap ratio along the welding path, together with real-time process control, is essential for suppressing crack formation.

3.2.2 Evaluation of Macrostructures and Maximum Shear Tensile Loads of Multi-Pass Welds under Different Heat Inputs

Fig. 5 presents the results of large-area multi-pass welding performed with the overlap ratio fixed at Condition A and heat inputs of (a) 4.5, (b) 4.0, and (c) 3.5 J/mm. As described in Table 3, the welded area was 8 × 6.3 mm². Cross-sectional observations revealed an average penetration depth of approximately 1.4 mm. Under Condition A, none of the three heat input conditions exhibited the solidification cracking observed under the higher-overlap conditions shown previously, and sound welds free from welding defects such as porosity were successfully obtained. In particular, clear boundaries between individual passes were observed, and no evidence of HAZ formation, including grain coarsening, was detected throughout the weld region. These results indicate that ultra-low heat input multi-pass welding using a single-mode laser within the range of 3.5-4.5 J/mm can achieve HAZ-free welding behavior.

Fig. 5

Surface and cross-sectional macrostructures of multi-pass Cu-Cu welds obtained at different heat inputs, (a) 4.5 J/mm, (b) 4.0 J/mm, and (c) 3.5 J/mm (overlap ratio: condition A)

Fig. 6 shows representative load-displacement curves obtained from shear tensile testing of the single-pass weld shown in Fig. 2(c) and the multi-pass weld shown in Fig. 5(b), both produced at a heat input of 4.0 J/mm. Fig. 7 presents the corresponding maximum load values together with images of the fracture surfaces. The multi-pass welds exhibited an average maximum load of approximately 3.50 kN, corresponding to an increase of approximately 920% compared with that of the single-pass welds. In the single-pass specimens, the limited weld area resulted in stress concentration within the weld region, leading to full interfacial fracture. In contrast, the large-area multi-pass welds exhibited partial interfacial fracture without complete separation of the welded joint, demonstrating substantially improved mechanical performance.

Fig. 6

Representative load-displacement curves obtained from shear tensile tests of single-pass and multi-pass Cu-Cu welds at a heat input of 4.0 J/mm

Fig. 7

Maximum load of single-pass and multi-pass Cu-Cu welds at an overlap ratio of condition A

3.3 Thermal Analysis of Rapid Heating and Cooling Behavior during Optimized Large-Area Multi-Pass Welding

To investigate the HAZ-free behavior observed in Fig. 5, the thermal histories associated with heating and cooling during large-area multi-pass welding using a single-mode laser were calculated using the Thermo-Calc Additive Manufacturing Module. This module enables theoretical prediction and validation of the rapid melting and solidification behavior occurring during high-speed laser additive manufacturing processes, which are difficult to measure experimentally. The single-mode laser welding conditions employed in this study, including heat input and energy density, were comparable to those commonly used in laser-based additive manufacturing. Therefore, the module was utilized to analyze the rapid heating and cooling behavior occurring within the large-area welds produced using the rectangular spiral scanning pattern.

All input parameters used in the calculations were identical to the actual laser welding conditions. The cooling rate was determined from a single thermal cycle, defined as the period from the peak temperature reached at a specific location after passage of the heat source to subsequent cooling to room temperature. Figs. 8-10 present the calculated thermal histories and cooling rates obtained for selected locations corresponding to the 6th, 13th, 20th, and 25th passes on the bead surface (Fig. 8), the overlapped regions within the U-shaped segments of the rectangular spiral scanning pattern (Fig. 9), and the HAZ locations along the bead depth direction (Fig. 10), respectively.

Fig. 8

Calculated thermal histories and corresponding cooling rates at selected points on the bead surface during single-mode fiber laser multi-pass welding at a heat input of 3.5 J/mm, (a) extracted surface locations (6th, 13th, 20th, and 25th passes) and (b) corresponding cooling rate profiles

Fig. 9

Calculated thermal histories and corresponding cooling rates at overlapped regions of the multi-pass scanning path during single-mode fiber laser welding at a heat input of 3.5 J/mm: (a) extracted overlap locations and (b) corresponding cooling rate profiles

Fig. 10

Calculated thermal histories and corresponding cooling rates along the bead depth direction during single-mode fiber laser multi-pass welding at a heat input of 3.5 J/mm, (a) extracted subsurface locations and (b) corresponding cooling rate profiles

The total welding time required to complete the 8 × 6.3 mm² weld area employed in this study was calculated to be less than 0.4 s. In particular, an average cooling rate of approximately 7 × 10⁴ K/s was obtained at the bead surface. Furthermore, no increase in cooling behavior associated with thermal accumulation was observed in the overlapped regions of the scanning path. These results indicate that thermal accumulation caused by adjacent-pass overlap can be effectively suppressed even under multi-pass welding conditions. Heat generated during each pass is rapidly dissipated through the extremely high cooling rate, thereby enabling the formation of HAZ-free welds.

In addition, the calculated HAZ location immediately beneath the weld boundary along the bead depth direction exhibited a maximum temperature below 400 K. Maintaining a peak temperature below 400 K during the ultra-high-speed large-area welding process, completed within approximately 0.4 s, suggests that thermal damage to surrounding components potentially located beneath the weld bead can also be effectively minimized.

4. Conclusions

In this study, the occurrence of solidification cracking, joint strength, and HAZ formation behavior in power semiconductor module busbar welds were systematically evaluated using a single-mode fiber laser to establish weld integrity and process guidelines for busbar welding applications.

Article information Continued