High Density Hybridisation Using ENIG Bumping and Anisotropic Conductive Films

TL;DR

This paper presents a wafer-free, fine-pitch flip-chip interconnection approach by combining in-house ENIG bumping with Anisotropic Conductive Film (ACF) bonding. It demonstrates functional ENIG–ACF hybridisation at 55$~\mu$m$ pitch on Timepix3 and Ti-LGAD devices and extends the method to 25$~\mu$m$ pitch with dedicated test structures, revealing that ACF thickness relative to bump height is a key reliability limiter. A custom Python tool models isotropic ENIG-bump growth to optimise bump height for given pad geometry and ACF thickness, yielding practical guidance for high-density assemblies. The results support ENIG–ACF as a scalable, cost-effective alternative to conventional bump-bonding for future detectors, with planned work addressing reliability under thermal cycling and irradiation.

Abstract

Fine-pitch hybridisation processes are essential for next-generation pixel detectors and high-density microelectronic assemblies. Conventional bump-bonding techniques, although reliable, remain costly and difficult to implement for single-die applications. In this work, we present flip-chip hybridisation results combining Electroless Nickel Immersion Gold (ENIG) bumping with Anisotropic Conductive Film (ACF) bonding, both developed in-house. This method enables fine-pitch interconnections without requiring wafer-level processing. The feasibility of the ENIG-ACF process for functional devices was demonstrated by hybridising functional Timepix3 ASIC and Ti-LGAD (Trench Isolated Low-Gain Avalanche Detector) sensors with 55 um pitch. The approach was then extended to test structures with 25 um pitch, using aligned-particle ACFs. The analysis revealed that the main limitation originated from the mismatch between ACF thickness and bump height, underlining the importance of interface geometry optimisation. To address this, a dedicated simulation program was developed to model the isotropic growth of ENIG bumps and to determine the optimum bump height as a function of ACF thickness and pixel-matrix geometry. The results provide valuable guidelines for future high connection density assemblies and demonstrate the potential of the ENIG-ACF process as a scalable, low-cost alternative to conventional bump-bonding techniques.

Figures (6)

• Figure 1: Timepix3/Ti-LGAD hybridisation process and module pictures. a) ENIG bumps on Timepix3 chip (Scanning Electron Microscope (SEM) picture), b) Ti-LGAD sensor after ACF lamination (optical microscope), c) Timepix3/Ti-LGAD hybrid wire-bonded to PCB.
• Figure 2: Timepix3/Ti-LGAD hybrid characterisation pictures. a) Hitmap obtained with a strontium-90 source, b) Histogram representing the number of hits per pixel, c) optical-microscope picture of the lamination issue in the vacuum-groove region.
• Figure 3: (Optical-microscope pictures of test-chip hybridisation process. a) Quartz test chip, b) ENIG bumps realised on the test chips, c) Test chip after ACF lamination.
• Figure 4: Optical-microscope characterisation pictures of the test-chip assembly a) Bonding interface after separation of the chips (SEM), b) Test assembly with particles between the chains, view from the top, c) Test assembly with excess ACF outside the matrix.
• Figure 5: Simulation results for a 5 by 5 triangular matrix, with 18 µm passivation opening diameter and 50 µm pitch: a) 3D view of ENIG bumps, b) evolution of the inter-bump volume and the maximum compatible ACF thickness as a function of bump height.