Abstract. Since the end of the last millenium, the microelectronics industry has been facing new issues as far as CMOS devices scaling is concerned. Linear scaling will be possible in the future if new materials are introduced in CMOS device structures or if new device architectures are implemented. Innovations in the electronics history have been possible because of the strong association between devices and materials research. The demand for low voltage, low power and high performance are the great challenges for the engineering of sub 50 nm gate length CMOS devices. Functional CMOS devices in the range of 5 nm channel length have been demonstrated. The alternative architectures allowing to increase devices drivability and reduce power consumption are reviewed. The issues in the field of gate stack, channel, substrate, as well as source and drain engineering are addressed. HiK gate dielectric and metal gate are among the most strategic options to consider for power consumption and low supply voltage management. By introducing new materials (Ge, diamond/graphite carbon, HiK, . . . ), Si based CMOS will be scaled beyond the ITRS as the future System-on-Chip Platform integrating also new disruptive devices. For example, the association of C-diamond with HiK, as a combination for new functionalized Buried Insulators, will bring new ways of improving short channel effects and suppress self-heating. Because of the low parasitics required to obtain high performance circuits, alternative devices will hardly compete against logic CMOS. (Fig. 1) has accelerated the scaling of CMOS devices to lower dimensions continuously despite the difficulties that appear in device optimization.
PACSHowever, uncertainties about lithography, economics and physical limitations will probably slow down the evolution. For the first time, since the introduction of poly gate in CMOS devices process, showstoppers other than lithography appear to be attracting special attention and could require some breakthrough or evolution if we want to continue scaling at the same rate. Design could also be affected by this evolution.Which are the main showstoppers for CMOS scaling? In this paper, we focus on the possible solutions and guidelines for research in the next years in order to propose solutions to enhance CMOS performance before we need to skip to alternative devices. In other words, how can we offer a second life to CMOS?To that respect, the roadmap distinguishes today three types of products: High Performance (HP) (Fig. 1), Low a e-mail: sdeleonibus@cea.fr Operating Power (LOP) and Low Standby Power (LSTP) devices. In the HP case, a historical fact will happen by the 32 nm node: the contribution of static power dissipation will become higher than the dynamic power contribution to the total power consumption! This main fact could affect the MOSFET saturation current as can be observedArticle published by EDP Sciences and available at