The surface activated bonding (SAB) method for low temperature bonding their development for heterogeneous and 3D integration is reviewed. The standard SAB method is based on surface bombardment by Ar beam in ultra-high vacuum to clean the surfaces so that they can be bonded very strongly at room temperature without heat treatment. Modifications of the surface activation have been investigated to extend the standard SAB method for various materials and applications.

The conventional bonding methods used for heterogeneous and 3D integration, such as anodic bonding, glass frit bonding, soldering, metal thermal-compression (diffusion) bonding, ultrasonic bonding, adhesive bonding, in general require high-temperature heating or annealing at 250-400 °C or even 700-1000 °C to achieve high bonding strength, which may induce great thermal stress, degrade the device reliability, and decrease manufacturing yield, especially in heterogeneous integration of dissimilar materials. Therefore, low-temperature bonding and interconnection technologies are highly desirable especially for heterogeneous integration for 2.5D or 3D applications. Although much effort has been made in recent years to attain a low temperature annealing process, most of the bonding methods including plasma activated bonding do not provide sufficient bond strength at below 200 °C or room temperature for critical applications.

The question was if such high-temperature reaction and diffusion are inevitable for bond formation. The most solid materials have high surface energy and cohesive energy. It concludes that a stable bonding can be formed just by contact at room temperature. However, the real surface is covered by oxide layer or certain contamination. Such layers, therefore, should be removed prior to the bonding. The term of the Surface Activated Bonding (SAB) was used for the first time in the article of F. S. Ohuchi and T. Suga (1994) [1], to describe the bonding process which "utilizes the fundamental nature of atomically clean surfaces created by energetic particle bombardment leading to bond formation when two surfaces are brought into contact."

Effectiveness of the surface activation for room temperature bonding had been demonstrated since the middle of '80 by several research groups in Japan. The bonding process in UHV for Al-Al and Al-Si3N4 was described in 1992 [2], and Cu-Cu for micro-bonding in 1993. The SAB method was applied successfully to the direct wafer boding of Si-Si in 1996 [3] and then expanded to heterogeneous bonding between semiconductors and metals at room temperature. The SAB method has attracted increasing interest due to its simple process flow, no need for additional intermediate materials for bonding, and compatibility with CMOS technology. However, it has failed to bond some dielectric materials, such as glass, sapphire, and silicon oxide. A modified SAB method using Si nano-intermediate layer has been developed to realize bonding of dielectrics at room temperature. novel surface activation processes have also been proposed for Cu/dielectric hybrid bonding, which is a promising approach to high-density 3D interconnects. This paper reviews the SAB methods used for heterogeneous and 3D integration with new approaches by expanding the versatility of the method.

The standard SAB method uses Ar beam bombardment to remove surface adsorption and oxidation layer to realize bonding between semiconductors when two surfaces are brought into contact. It has been studied for bonding of Si-Si, Ge-Ge, and compound semiconductors such as GaAs-Si. It has been successfully applied also to room-temperature wafer-level and chip-level Cu-Cu direct bonding [4].

The standard SAB, however, failed to bond some dielectric materials, such as glass and silicon oxide. Therefore, the modified SAB [5] was developed to solve this problem, by using a Si intermediate layer deposited on the activated surfaces. The modified SAB is now applied to bond not only SiO₂ glasses but also polymer films such as PEN and Polyimide, as well as WBG semiconductor wafers with a wide perspective of the applicability on flexible electronics and power electronics.