A REPORT ENTIRELY DEDICATED TO NEW MEMS PROCESSES AND MATERIALS
Athough MEMS technologies have not been driven by the same size demands as ICs, it doesn’t mean that MEMS manufacturing is just standing still. The fast growing MEMS markets, now driven by consumer applications, are:
Size-driven: for demanding consumer applications like smart phones and laptops
Performance-driven: for high end applications like aerospace
Cost-driven: for high volume applications like cell phones, automotive and games consoles
New MEMS manufacturing, packaging technologies and specific materials are necessary for solving these issues. This Yole Développement report highlights the future challenges for MEMS production and packaging. From bulk micromachining to surface micromachining to SOI, and MEMS technology has been following a well-defined evolutionary technical roadmap with 3D integration being the next possible step. In the report, you will find manufacturing trends for the different MEMS devices in terms of processes, new packaging approaches, 3D integration, CMOS MEMS integration, new materials such as structured wafers…
WHAT WILL BE THE FUTURE OF MEMS MANUFACTURING
This report analyzes the current MEMS manufacturing trends and presents some clues for understanding the next evolution in terms of die size, cost, packaging…Among other MEMS technologies to watch for the future, we have identified:
at the substrate level: SOI, glass, thin wafers, silicon...
at the MEMS die level: getters, fusion bonding, release stiction, singulation, CMOS MEMS, DRIE, trench isolation
at the packaging level: TGV, TSV, pixel-level packaging, thin film capping, active capping…
Wafer forecasts 2009-2015 by type of step (DRIE, wafer bonding, sacrificial etch, through Si vias, thin films packaging, CMOS MEMS, thin wafers) are estimated for all the analyzed MEMS technologies.
DRIE and wafer bonding are the technologies subject to major evolution: main reason is that both technologies are increasingly used for 3D TSV in the mainstream semiconductor business. Wafer bonding is the direct competitor for CMOS MEMS approach. For example, microbolometer players are more and more considering wafer bonding approach to stack the MEMS to the ROIC wafer. MEMS have been scarcely pushed by technological innovation. Most of the time, a MEMS is developed either by the use of micromachining to reduce existing sensors or the push is coming from system makers. As an example, lateral MEMS (accelerometers) have been developed by Leti because of military request from Thales. DRIE has been developed by Bosch because of automotive applications. The only exception is ADI that wanted to use its existing CMOS lines. Using CMOS is sometimes an historical choice (with the disadvantage that now the CMOS technology is evolving quicker than the MEMS technology). Indeed, CMOS MEMS is likely to be restricted to very specific applications where MEMS arrays will need very close electronic processing. For all other case, it will depend on MEMS product cycle time, flexibility, cost, integration, market demand and power consumption.
In 2011, simplification of manufacturing remains an objective: The Yole Développement MEMS law “One product, one process, one package” still rules. Will it still rule in 2020? The current work on technology and product platforms attempts to overcome the Yole Développement MEMS law. But this approach will be custom-made standard processes. By 2020, it is likely that MEMS fabs will have developed internal standard process blocks but it will be fab-specific standard tools.
KEY FEATURES OF THE STUDY
The objective of this report is to provide an understanding of current challenges of MEMS manufacturing, packaging & materials. For each MEMS manufacturing step, bottlenecks and challenges will be highlighted. It is a 350+ slide report.