Issue link: https://iconnect007.uberflip.com/i/1156271
66 PCB007 MAGAZINE I AUGUST 2019 Internal Stress Under the influence of additives, the plat- ed metal deposit will have some residual in- ternal stress. Many factors—such as tempera- ture, thickness, additives, and annealing—will affect stress. Stress could be either tensile or compressive. In both cases, high stress is det- rimental to the PCB board, causing distortion in the final board called warping. We used an internal stress analyzer to measure the stress of the deposit as plated and after annealing. To measure the stress, first, a test strip was immersed in a cleaner solution at 45°C for up to 30 seconds and rinsed with water. Then, the strip was dried completely and weighed. Next, the strip was plated at the desired current den- sity for the desired time to achieve the neces- sary Cu thickness. Finally, the strip was rinsed with water and dried very carefully with low- pressure air. Then the strip was mounted on the measuring stand (deposit stress analyzer). The value for U was measured and recorded as the sum of the total number of measure- ment increments on both sides of the zero on the measuring stand. The plated test strip was weighed, and the final weight was recorded. After the deposit thickness is known and the number of increments spread between the test strip leg tips has been determined, the deposit stress can be calculated using the equation S = UKM ÷ 3T where S = pounds per square inch, U = measured number of increments spread, T = deposit thickness in inches, K is the strip calibration constant, and M equals the modu - lus of elasticity of the deposit divided by the modulus of elasticity of the substrate material. After the initial measurement was done, the strips were annealed at 130°C for one hour. Figure 9 summarizes the internal stress data for Process I with fresh and aged baths with both plated and annealed strips. Low internal stress under 1000 psi was observed for both plated and annealed conditions, and this did not change significantly as the bath ages. Deposit Grain Structure The grain structure of the deposit was stud- ied using focused ion beam (FIB) microscopy techniques. Plated copper samples from fresh and aged bath were evaluated. Figure 10 shows the grain structure of the deposit. In the fig- ures, the top portion of the fine grain struc- ture is the plated copper from Process I and the bottom portion with larger grains are from the internal stress test strip substrate. Both images are at 5000x magnification. According to the data, the grain structure remained unchanged even after aging the bath. Embedded Trench Plating Formulation Process II is tailored towards embedded trench plating applications with higher acid (200 g/L) than CuSO 4 (100 g/L) in contrast to Process I, which promotes the via fill. Results are summarized in Figure 11. However, we tested Process II and its via fill capability by changing the VMS by increasing CuSO4 to 250 g/L. According to Figure 11, an average dimple around 3–4 µm was seen in vias of 60 x 35 µm size with surface Cu of 10–15 µm. However, Process II showed excellent embedded trench plate capability with high coplanarity with the Figure 9: Internal stress of the copper deposit plated using Process I, both as plated and annealed data for fresh and aged baths. Figure 10: Grain structure with bath age, (a) fresh bath, (b) 100 Ah/L.