1. Introduction
With the increasing use of eco-friendly electric vehicles, the use of power modules, such as inverters and converters, is also increasing for power conversion of direct current (DC) and alternative current (AC) that are used in vehicles, and related research has been actively conducted1-3). Inverters and converters convert output to a specific phase and frequency using the switching characteristics of the power semiconductor. Since eco-friendly electric vehicles exhibit high electric energy consumption compared to conventional internal combustion engines, high efficiency, high frequency, high reliability, and high power density of the power conversion system are required. For this reason, research has been conducted to replace existing silicon (Si)-based devices with devices having faster switching speed based on wide band gap semiconductors (WBG semiconductors), such as silicon carbide (SiC) and gallium nitride (GaN). With the growing demand for such compound semiconductors, probe cards for power semiconductors have been actively developed to produce power semiconductor devices that use high voltage and current4,5).
A probe card is used as a signal transmission medium at the final stage of the semiconductor manufacturing process to inspect the electrical operation of individual devices in the completed wafer. It has a structure in which probe pins are connected to the printed circuit board (PCB) at regular intervals. These pins come into contact with the electrode pad at the top of the chip to examine the quality of the chip. Probe cards are classified as high value-added products because they must be designed and manufactured individually to meet the specifications of semiconductor devices to be tested. They are divided into the horizontal type, vertical type, and micro-electro-mechanical system (MEMS). The horizontal probe card has a structure in which thousands of probe pins created by polishing tungsten (W) wires with a diameter of 150-250 ㎛ like needles are stacked horizontally to fit the position of the chip pad. It, however, involves the short-circuit problem between the probe pins because the spacing between the pins decreases as the number of pads and chips increases due to structural limitations when thousands of probe pins are stacked6). To address this problem, the vertical probe card with probe pins vertically installed was developed. The vertical probe card was designed to give bending to tungsten probe pins for the spring operation of the pins7). This vertical probe card also has the short-circuit problem with adjacent probe pins because distortion occurs at the bended part of the tungsten wires vertically installed when the spacing between the probe pins decreases due to the increase in the number of test pads and high density. To address the problems with the horizontal and vertical probe cards, the MEMS probe card was developed8-11). The cantilever wire-type probe card that uses the MEMS process has been utilized in in-circuit testing of semiconductor devices that use high voltage and current12,13).
As shown in the schematic diagram of Fig. 1, the cantilever wire probe card that uses wires has limitations in high-speed signal transmission because the probe pins are long, but it is suitable for the inspection of semiconductor devices that use high voltage and current, such as power semiconductor devices. Stress, however, is continuously applied to the wires of the probe card as they come into contact with the measurement terminals (In/Out, I/O) on the power semiconductor wafer surface tens of thousands of times. Such stress causes the fatigue fracture of the wires. Therefore, bending and fatigue strength characteristics need to be improved for an improvement in the durability reliability of the probe card. To meet the required performance, the tensile and fatigue properties of cantilever wires that used tungsten and rhenium tungsten with high melting point and tensile strength properties and palladium alloys (Pd-Ag- Cu) with lower tensile strength were analyzed in previous studies. The test results showed that W, ReW, and BeCu were superior to Pd alloy wire materials1). Pd alloys, however, can be used as probe pins due to their high electrical conductivity and oxidation resistance. In addition, according to studies by Liu, B.X and Gao, K. metals laminated in multiple layers have higher mechanical properties than metals without lamination by inhibiting strain localization and stress concentration14,15).
Therefore, in this study, dog-bone specimens in which a palladium-cobalt (PdCo) alloy and a copper (Cu) layer are laminated alternately were fabricated using the MEMS process to develop Pd alloy wire materials with improved mechanical properties, and the tensile strength and elongation according to the laminated structure were compared and analyzed. In addition, after the tensile test, the transition of the fracture mode was compared and analyzed through the scanning electron microscope (SEM) analysis of the fracture surface of the specimen.
2. Experimental Method
2.1 MEMS Probe Specimen
Fig. 2 shows the photograph of a tensile test specimen used in the experiment. Tensile test specimens were fabricated in accordance with No. 5 specimen standard of KS B 0801. The entire specimen size was 11.0 (L) × 3.0 (W) × 0.05 (D) mm, and the central part of a tensile test specimen had a length of 2.5 mm. Fig. 3 shows the cross-sectional SEM images of the tensile test specimens with 5, 7, 11, 15, and 21 layers. PdCo-Cu layers were laminated as shown in Fig. 4. In addition, the tensile strength was measured after isothermal aging for 45 minutes at 220°C to observe changes in mechanical properties according to isothermal aging for the five types of laminated specimens. Table 1 shows the inter-layer lamination thickness of MEMS probe tensile specimens according to ten PdCo-Cu lamination types. The thickness ratio of PdCo:Cu was approximately 2:1.
Fig. 3
SEM cross-sectional micrographs of PdCo (white grey)-Cu (dark grey) multi-layered specimens before isothermal aging: (a) 5, (b) 7, (c) 11, (d) 15 and (e) 21 layers
2.2 Tensile strength measurement and fracture surface analysis
For the tensile test, a universal testing machine (QUASAR 50, Galdabini Co., Ltd., Italy) was used as shown in Fig. 5. The tensile test was conducted after fixing a MEMS probe specimen at the jig as shown in Fig. 5(b) and 5(c). As shown in Table 2, the tensile test was conducted at a tensile test speed of 1.0 mm/min in accordance with the ASTM E8 standard. After the test, tensile strength and elongation were compared between the specimens according to the number of PdCo/Cu layers and the presence/absence of heat treatment. The fracture locations of the tensile specimens were identified after the tensile test for ten types of samples, and the fracture surface was observed using SEM to examine the fracture pattern of each specimen.
3. Experiment Results and Discussion
3.1 Tensile strength measurement results for PdCo- Cu specimens
In the tensile test results, the tensile strengths of the 5, 7, 11, 15, and 21-layer specimens without heat treatment were 1.2, 1.0, 1.4, 1.5, and 1.4 GPa, respectively, as shown in Fig. 6. When the tensile strength values according to the number of PdCo-Cu layers were compared, the tensile strengths of the 11, 15, and 21-layer specimens without heat treatment were measured to be higher than that of the 5-layer specimen. In addition, the elongation values of the 11, 15, and 21-layer specimens were 8.7%, 8.0%, and 8.1%, which were lower than the elongation (9.6%) of the 5-layer specimen. It appears that tensile strength increased and elongation decreased as the number of PdCo-Cu layers increased because the barrier layers of PdCo-Cu increased during crack generation and propagation. The tensile strengths of the 5, 7, 11, 15, and 21-layer specimens with heat treatment were found to be similar with values of 1.6, 1.5, 1.4, 1.4, and 1.4 GPa, respectively. Their elongation was also similar with values of 7.8%, 7.9%, 7.0%, 7.1%, and 6.8%.
Fig. 6
Tensile strength and elongation comparison graph of PdCo-Cu multi-layered tensile test specimens with isothermal aging effect
When the tensile strength values according to the presence/absence of isothermal aging heat treatment were compared, the specimens with heat treatment exhibited higher tensile strength than those without heat treatment for the 5 and 7-layer specimens, but there was no significant difference in tensile strength depending on heat treatment for the 11, 15, and 21-layer specimens as shown in Fig. 6. It was also found that the specimens without heat treatment had higher elongation than those with heat treatment for the 5, 11, 15, and 21-layer specimens.
3.2 Fracture surface analysis
Fig. 7 shows the results of analyzing the exterior of the fractured specimens after the tensile test. Fig. 7(a), 7(c), 7(e), 7(g), and 7(i) show the images of the fractured specimens without heat treatment. The cup-and- cone shape was observed, indicating ductile fracture properties. The occurrence of rapid fracture was confirmed in the specimens with heat treatment as shown in Fig. 7(b), 7(d), 7(f), 7(h), and 7(j), indicating that the specimens had brittle fracture properties.
Fig. 7
Stereo-micrographs of multi-layer tensile specimens (a, c, e, g, i) before and (b, d, f, h, j) after isothermal aging : (a, b) 5, (c, d) 7, (e, f) 11, (g, h) 15 and (i, j) 21 layers
As shown in Fig. 8 and Fig. 9, the fracture surface of the tensile specimens was observed using SEM. A number of dimples, which are typical characteristics of ductile fracture, were observed from the fracture surface of the tensile specimens without heat treatment, and the cleavage fracture surface, which is a brittle fracture pattern, was observed in some areas as shown in Fig. 8. This confirmed that the MEMS probe with Pd-Co lamination exhibited a ductile-brittle mixed fracture mode. In addition, dimples larger than the dimples formed in the center of the PdCo layer were observed from the barrier layer between the PdCo and Cu layers in Fig. 8(h), 8(l), 8(p), and 8(t). It appears that larger dimples were formed at the barrier layer between the PdCo and Cu layers because the second phase formed during the deformation process by tensile stress in the tensile test acts as a major factor in dimple formation16). In Fig. 9, the cleavage fracture surface was mostly observed from the fracture surface of the specimens with heat treatment, indicating that the brittle fracture pattern increased compared to the specimens without heat treatment. This showed the same pattern as the result of Fig. 6 that the elongation of the 7-layer specimen was similar after heat treatment while the elongation of the 5, 11, 15, and 21-layer specimens decreased.
4. Conclusions
In this study, the optimal lamination conditions were derived by comparing the tensile strength and elongation of micro-electro-mechanical system (MEMS) probe pins according to the PdCo-Cu lamination conditions of 5/7/11/15/21 layers. In addition, the tensile strength and fracture mode of the MEMS probe according to isothermal aging were compared and analyzed. The results are as follows.
1) The MEMS probe tensile specimens without heat treatment in which PdCo-Cu was laminated in 11/15/21 layers exhibited higher tensile strength and lower elongation than the specimens with 5 layers. It appears that tensile strength increased and toughness decreased as the number of laminated PdCo-Cu layers increased because the barrier layers of PdCo and Cu increased.
2) The MEMS probe tensile specimens with 7 layers of PdCo-Cu exhibited higher tensile strength after heat treatment.
3) The fracture surface of the tensile specimens before isothermal aging heat treatment showed a fracture pattern that mixed ductile fracture and brittle fracture. The fracture surface of the specimens with 5, 11, 15, and 21 PdCo-Cu layers after heat treatment exhibited more brittle fracture surfaces than the specimens without heat treatment. The fracture surface of the 7-layer specimen with heat treatment, however, showed a fracture pattern that mixed ductile fracture and brittle fracture.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
