Invar Electroplating for Controlled Expansion Interconnects

Abstract

The current trends in convergent systems are to pack as many functions – digital, RF, optical, MEMS, LEDs, and other IoT devices – into a single package or module and current methods of metallizing the ever-shrinking components are rapidly becoming insufficient. A new set of metallization material technologies is necessary to meet these increasing expectations, methods that not only work with the substrates and materials of yesterday but are also ready for the future. The copper that is currently the workhorse of interconnect materials has a coefficient of thermal expansion (CTE) of 17 ppm/K while chip substrates’ CTE ranges from 2 to 7 ppm/K. The thermo-mechanical stresses induced in the package from the CTE mismatch between the die, the substrate and the buildup materials result in solder joint failure, die cracking, copper extrusion, delamination of the solder bumps and cracking of the build-up layers, leading to failure of the assemblies. Matching the CTEs of the chips and the substrates is required for future fine-pitch flip chip and wafer level package assembly in order to improve the solder joint reliability and fabricate stress-free 3D multichip systems. NANO3D Systems has developed an electroplating technique to produce uniform, low-internal-stress and controlled-expansion nickel iron alloy (INVAR) films deposited on different substrates, fabricating redistribution layers (RDL), through silicon vias (TSV), pillars, and free standing foils. INVAR electroplating was performed using a proprietary plating solution containing stabilizers that prevent the oxidation of iron, stress-reducers to reduce internal stress, and wetting agents to improve the surface morphology and eliminate pinholes. The thermal (thermal conductivity, linear thermal expansion, specific heat capacity) and mechanical (Young’s modulus, yield strength, tensile strength and internal stress) properties of the INVAR have been studied and demonstrated superior performance compared to currently available similar products. The composition of the INVAR (36 % of Nickel and 64% of Iron) electroplated films has been confirmed by utilizing SEM/EDX and SIMS. Linear thermal expansion of the films was tested using a technique based on bending of bonded layers due to thermal stress. Internal stress was studied using bent strip techniques. The linear thermal expansion down to 0.4±0.1 (x10-6, K-1) has been obtained for the INVAR plated films plated within a process window that corresponded to the lowest internal stress (30±2 (MPa)) and uniform morphology (Ra ~ 2 nm). The mechanical properties of the INVAR foils have been investigated by use of the Instron technique. The tensile strength of 500 ± 60 (MPa), yield strength of 280 ± 50 (MPa) and the ductility of 40 ± 10% have been measured for the pinhole-free INVAR films. The specific heat capacity of 0.45 ± 0.05 (J/gK)was measured by using differential scanning calorimetry (DSC) at 573 K, while thermal conductivity of as-plated INVAR was 43 ± 5 (W/mK). Smooth, uniform and flat RDL & pillars were fabricated by using through resist INVAR plating on silicon wafers. Optical profilometry studies of the morphology and thickness of plated INVAR on the silicon wafers show Ra ~ 2 nm and high thickness uniformity (of about 5 % & 3s). TSV fill of high aspect ratio features (>10:1) was also demonstrated. After being annealed at 300 0C for 2hr, SIMS studies showed no inter-diffusion in copper/Invar bi-layer foils. Trace impurities were <100 ppm in electroplated INVAR films. Analytical methods have been developed for plating solution analysis. UV-VIS and titration are the two main techniques employed for analysis of metals, organic acids, and buffers in solution. Standard curves were created by analyzing solutions with known concentrations of each analyte. Solutions of varying concentrations were prepared for analysis; low control limit (low), target, and upper control limit (high) concentrations. Blank solutions contained all components except for the one of interest. There was also a blank prepared with no metals at all to be used as the reference in UV-VIS analysis. Multiple analyses were collected for the standard solutions in order to perform a statistical analysis and confirm p/t <0.3. Using the standard curves concentrations of unknowns (active bath solutions) were calculated. The plating additives in INVAR electroplating bath such as stress-reducer, stabilizer and wetting agents were monitored by using cyclic voltammetric stripping (CVS) and FTIR techniques.