1. Introduction
The sizes of passive devices such as chip resistors and multi-layer ceramic capacitors (MLCCs) are shrinking rapidly as a result of the downsizing and densification of semiconductor packages. As the size of these components shrinks, so does the size of the solder joints, and the reliability of the solder joints influences the lifetime of the electronics1,2). Organic solderability preservative (OSP) and electroless nickel/immersion gold (ENIG) plating are widely used for surface finish of printed circuit boards (PCBs), and these surface finish are affecting the reliability of solder joints. OSP surface finish is causing reliability issues due to the rapid diffusion of Cu atoms, which increases the thickness of intermetallic compounds (IMCs)3). ENIG surface finish prevents IMCs growth at the joint interface by slowing Cu and Sn diffusion through the electroless Ni-P layer4-6). The IMCs formed at solder joints are brittle, and an appropriate thickness of IMCs increases the strength of the joint, but as the thickness increases, it affects the reliability of the joint. Thus, ENIG surface finish substrates are used in automotive electronics7,8).
With the high density and high integration of electronics, the gap between PCB circuits and solder joints is narrowing, and electrochemical metallic ion migration (ECM) is increasing the defects of electronics9). These ECM problems are increasing especially in the eco-friendly vehicles such as electric vehicles and hydrogen vehicles, where the operating voltage is increasing. Hence, ECM development is required. ECM is a phenomenon in which moisture is adsorbed between PCB electrodes, metal is ionized by electrolysis reaction from the anode and moved to the cathode. Precipitation occurs when electrons received from the cathode are reduced back to metal, and these precipitates grow into metal dendrites and cause an electrical short circuit10). Various causes of ECM, such as voltage, temperature, humidity, pattern spacing, metal content, and contamination have been suggested11,12).
Therefore, chip resistors of sizes 3216, 2012, 1608, and 1005 mm were joined by nitrogen (N2) reflow soldering process using Type 6 Sn-3.0Ag-0.5Cu (SAC305) solder paste, OSP and ENIG surface finish substrates in this study. Thermal shock test (TST) was performed to verify the degradation characteristics of the chip resistor solder joint and the joint reliability after bonding, and X-ray void measurement, bonding strength measurement, and microstructure analysis were performed. Furthermore, to verify the ECM properties of the solder joints, the SAC305 solder paste was printed on substrates with and without photo solder resist (PSR) and then the nitrogen reflow process was carried out. The ECM test was conducted in a high temperature and humidity environment, and optical microscope analysis, scanning electron microscopy (SEM), and energy dispersive spectroscopy (EDS) were performed to identify dendrites.
2. Experimental Method
2.1 Creating samples to evaluate solder paste properties
The solder paste used in the experiments was solder powder with Sn-3.0Ag-0.5Cu (SAC305, Mk Eletron Co. Ltd., Korea) with a powder particle size of Type 6. It was prepared using three types of fluxes developed by the manufacturer. A rosin flux was used for solder pastes A and C, whereas a water-soluble flux was used for the solder paste B. The flux types and contents are listed in Table 1. Fig. 1 shows the substrate used in the experiment, which is a 130×130 mm OSP with ENIG plating surface finish. Chip resistors of sizes 3216, 2012, 1608, and 1005 mm were mounted on the board for analysis of junction properties. The screen printing process was performed using a metal mask with a thickness of 100 ㎛ and an aperture of 100%. Fig. 2 shows the equipment and temperature profile for the surface mount technology (SMT) process. After printing the solder paste onto the substrate using a screen printer (MK-878Mx, Minami Co. Ltd., Japan), the chip resistors were mounted using a chip mounter (CP-45FV NEO, Samsung Techwin Co. Ltd., Korea) and the components were joined by the nitrogen reflow (1809UL, Heller Co. Ltd., USA) soldering process. To analyze the solder joint properties after the process, we conducted optical microscope analysis, X-ray void measurement, bonding strength measurement, and cross-sectional analysis of the 2012 mm chip resistor solder joint, as shown in Fig. 3. To evaluate the ECM properties of the solder paste, a comb pattern spacing/width of 0.318 mm was determined by referring to the ISO 9455-17 standard13), and two types of boards with and without PSR made of FR-4 were fabricated. The nitrogen reflow process was performed after printing the solder paste at 0.318 mm pattern spacing/width using a screen printer. Then, as shown Fig. 4, a test coupon sample was fabricated.
Table 1
Solder pastes specification with flux type and content
| Solder pastes | A | B | C |
|---|---|---|---|
| Flux chemical type | Rosin | Aqueous (Water soluble) | Rosin |
| Liquid flux content (wt%) | 12 | 12 | 13 |
| Solder powder content (wt%) | 88 | 88 | 87 |
Fig. 2
Photographs of (a) screen printer, (b) chip mounter, (c) hot air reflow soldering machine, and (d) temperature profile
2.2 Thermal shock test (TST)
To examine the degradation rate of chip resistor solder joints, TST was performed using TST equipment (NT- 1531W, ETAC Co. Ltd., Japan). TST was performed at –40~125°C for 1,000 cycles with a dwell time of 10 min each and a temperature change within 5 min. After TST, the shear strength of the solder joints of the chip resistor components was measured to compare the degradation rate of the joint strength compared to the initial one. After that, a cross-sectional analysis was performed.
2.3 Electrochemical metallic ion migration (ECM) test
Fig. 5 depicts the ECM test equipment and evaluation result. The ECM test was conducted at a temperature of 85±3°C and a humidity of 85±3% RH for 700 h with a DC 48 V power supply. Fig. 5 (a,c) shows the photograph of the ECM test sample in a constant temperature and humidity unit. The comb pattern of the ECM test board was connected to the unit as shown in the photo to measure the real-time insulation resistance. Fig. 5 (b) shows a real-time insulation resistance measuring instrument (NY-IM 1000, NAENG YEOL Co., Japan). As shown in Fig. 5 (d), the insulation resistance between the electrodes of the test sample was measured in real time. After completion of the ECM test, optical microscope analysis, SEM, and EDS analysis were performed to verify the dendrite formation between the electrodes.
2.4 Measuring solder joint void content and bond strength
To determine the solder joint void content, the void area relative to the area of the chip resistor solder joint was measured. The void analysis was performed by X-ray non-destructive analysis equipment (Resolution, X-Tek Co. Ltd., UK), and the void area was measured and calculated using the iSolution program (iSolution DT, IMT i-Solution Inc., USA).
The shear strength of the chip resistor was measured using a bonding test machine (Dage 4000, Nordson Co. Ltd., USA) to determine the strength of the chip resistor solder joints. The shear strength test speed was 167 ㎛/sec, and the measurement height of the test jig was 300 ㎛ for 3216, 2012, and 1608 chip resistors and 100 ㎛ for 1005 chip resistors depending on the sample size.
3. Experimental Results
3.1 Measuring the solder joint void
After measuring void content of solder joint, the void content by the substrate surface finish and solder paste type were displayed in a graph in Fig. 6. The OSP-surface finish substrate has a higher content than the ENIG-surface finish substrate. In the case of OSP substrates, if the process conditions are not suitable during the soldering process, the organic film may not migrate to the surface of the liquid solder during evaporation and solidify inside the cooled solder. This phenomenon results in a higher void content compared to the ENIG substrate14,15). However, the joint reliability of OSP surface finish boards is known to be better than that of ENIG substrate. The reason for this is attributed to the different IMCs and solder crystal structure differences at the joint interface16,17).
3.2 Solder joint shear test assessment
The joint shear test results before and after TST are shown in Fig. 7, and the degradation rate graph after TST is shown in Fig. 8. The shear test results after nitrogen reflow showed no significant difference in bonding strength between the OSP and ENIG surface finish substrates. The degradation rate graph after TST shows that less degradation was measured on the samples with ENIG surface finish than OSP. The 3216 and 2012 chip resistors with relatively large sizes showed a higher degradation rate than the 1608 and 1005 chip resistors. High and low temperatures during TST caused the warpage of substrate, which acts as a cause of such phenomena is the cause of reduced bonding strength18,19). It was found that the larger the size of the chip resistor, the greater the degradation rate because it is more affected by the warpage phenomenon20).
3.3 Analyzing the microstructure of solder joint
Cross-sectional microstructural analysis was performed to analyze solder joint cracks and IMCs. SEM photographs and EDS results of substrates with OSP and ENIG surface finish and solder joints with nitrogen reflow bonding of 2012 Chip resistors using SAC305 solder paste are shown in Fig. 9 and Fig. 10. After 500 and 1,000 cycles of TST, cracks occurred due to thermo-mechanical fatigue21). Moreover, it can be seen that the solder joints of the OSP substrate have advanced more into the solder base material than the ENIG substrate. This indicates that these cracks are responsible for the higher degradation rate on the OSP substrate.
Fig. 9
SEM micrographs of (a,b) OSP and (c,d) ENIG finish substrate after (a,c) 500 and (b,d) 1000 thermal shock cycles
Fig. 10
SEM micrographs of (a,b) OSP and (c,d) ENIG finish substrate after (a,c) as-reflowed and (b,d) 1000 thermal shock cycles and (e-f) EDS analysis results of solder joints
In the case of the substrate with OSP surface finish, the electrodes on the chip part were Ni/Sn plated, forming (Cu,Ni)6Sn5 IMC, while the Ag3Sn IMC was dispersed on the base metal part of the joint. At the board-side solder joint, (Cu,Ni)6Sn5 IMC was formed. The chip-side joints were flat and clamshell-shaped IMCs were formed at the board-side joint. For the substrate with ENIG surface finish, a (Ni,Cu)3Sn4 IMCs were formed at the solder joint of the chip and board side, and both joints had needle-shaped IMCs. Ag3Sn IMCs were formed in a dispersed manner in the parent material part of the joint. After 1,000 cycles of TST, both the OSP and ENIG surface-finish substrate exhibited coarsening and flattening of the IMCs. When the IMCs of OSP and ENIG substrate after TST were compared, it can be seen that the IMCs are thicker on the OSP substrate than the ENIG substrate. This is considered to be due to the presence of a non-reactive Ni layer between Cu and Sn that prevents diffusion. The solder joint strength is affected by the behavior of IMCs near the joint interface, and the growth of these IMCs is known to reduce shear strength22). The relatively brittle IMCs on the OSP substrate had a larger reduction in shear strength than the ENIG substrate due to its larger thickness.
3.4 ECM test result
Fig. 11, the resistance was measured in real time while the ECM test was performed. The failure criterion was set to 1×104 Ω, and only the C solder paste/PSR-applied substrate samples showed a decrease in resistance below the failure criterion, resulting in ECM. As the solder powder particle becomes smaller, the surface oxidation increases, which increases the flux activity, causing a non-wetting phenomenon during the soldering process. The paste A used in this experiment is a Type 4-only flux, and the C paste is a Type 6-only flux. When the flux activity is high, the insulation resistance between the circuits will decrease as phenomena such as scattering and residue between patterns occur during the soldering process. This is thought to be the reason that the PSR-applied C paste showed the lowest insulation property. When substrates without and with PSR are divided, we can see that the substrate with PSR has a lower resistance value. In the case of B solder paste, the resistance value of the substrate without and with PSR did not decrease for 700 h and showed a stable resistance change.
Fig. 11
Graph of in-situ monitered insulation resistance of 0.318 mm width and space of comb pattern test coupons (a-c) without and (d-f) with PSR coating layer after 700 h electrochmeical metallic ion migration test
The optical microscope analysis result after the ECM test is shown in Fig. 12. We observed dendrites in the inter-circuit insulation zone using a stereoscopic microscope, The formation of dendrites was higher on the substrate with PSR than on the substrate without PSR, and the ECM was generated by the dendrites. For solder paste B, no dendrites could be seen on both the substrates without and with PSR, and less flux residues were seen. The flux residues in the insulation between conductors has been reported to be the cause of reduced insulation resistance and accelerated ECM in high temperature and humidity environments23). This difference in insulation is considered to be due to the composition of the flux contained in the solder paste and the amount of flux that remained after surface finish. Previous studies have shown that the fat-soluble flux shows greater reactivity than the water-soluble flux. Weak organic acids (WOA) are one of the flux residues that contribute to a decrease in surface insulation resistance (SIR). These WOA content analysis results showed that the fat-soluble flux had more than 50% higher WOA concentrations than the water-soluble flux. These results show that the fat-soluble flux promotes water adsorption, which shortens the ECM development time and significantly increases dendrite formation24,25).
Fig. 12
Optical micrographs of 0.318 mm width and space of comb pattern test coupon (a-c) without and (d-f) with PSR coating layer after 700 h electrochemical metallic ion migration test
Fig. 13 shows the results of SEM and EDS analysis to identify dendrites after ECM test. Similar to the optical microscope analysis result, we can see that more dendrites occurred on the substrate with PSR. For the B solder paste, no dendrites were observed on both substrates without and with PSR. These EDS analysis of these dendrites detected Cu and Sn components. In the case of the substrate with PSR, as the ECM test progressed, the release of moisture absorbed into the substrate was inhibited. Thus, this phenomenon is thought to be the cause of the increased dendrite formation. Moisture is present between the two electrode as humidity is applied in an ECM test. At the anode, the Sn and Cu in the metal are ionized by an electrolysis reaction. The ionized Sn and Cu ions are formed into oxides, such as SnOx and CuOx, which migrate to the cathode, receive electrons from the cathode, and are reduced and precipitated. As this precipitation reaction continues, dendrites grow from the cathode to the anode, and ECM develops. The metals most prone to ECM are Ag>Pb>Cu>Sn in descending order26). Although the ECM phenomenon was more active in Cu than Sn, the higher content of Sn in SAC305 suggests that Cu and Sn components were detected together.
4. Conclusions
In this study, 3216, 2012, 1608, and 0402 chip resistors were nitrogen reflowed onto OSP and ENIG surface finish substrates using the Type 6 SAC305 solder paste. After bonding, TST, void content, bonding strength measurement, and microstructure analysis were performed to compare solder joints with different substrate surface finish. Furthermore, substrates without and with PSR were nitrogen reflowed with solder paste for ECM property evaluation. After the ECM test, the dendrites were analyzed by optical microscope microscopy, SEM and EDS to confirm the dendrites.
1) The solder joint void content of the 2012 chip resistors showed a higher void content on the OSP than on the ENIG substrate. The OSP surface finish substrate showed a higher void content than the ENIG substrate because the organic film evaporated during reflow and formed voids inside the solder.
2) The bonding strength of chip resistors after nitrogen reflow was not significantly changed by surface finish, but the degradation rate was measured higher on the OSP than the ENIG substrate after TST. The cross-sectional analysis revealed that the OSP had more advanced crack morphology and thicker IMCs than the ENIG substrate, which explained the higher degradation rate measurement.
3) After ECM test, more dendrites developed on the substrate with PSR compared to the substrate without PSR. It was determined that the substrate with PSR was responsible for more dendrites due to the inhibition of the release of moisture absorbed into the substrate.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
