Beyond the Horizon of CMOS

For decades the density of integrated circuits has grown exponentially, according to the empirical Moore´s law published in 1965. In this period, the storage density has increased by a factor of about 100 million. Yet, this rapid development is approaching fundamental physical limits. In view of this coming end of Moore´s law of conventional complementary metal-oxide-semiconductor (CMOS) transistors, alternatives have been investigated for a long time. Already in the 1960s, there were first reports about non-volatile memory (NVM) effects in metal–metal oxide–metal cells.¹

Research was resumed in the 1990s. The Tokura group discovered electrically triggered resistive switching (RS) in Pr_xCa_1-xMO₃ (PCMO) while investigating the magnetoresistive properties of this material.² In 2000, the IBM Zurich lab reported the RS of perovskite-type zirconates.³ In both cases, the physical origin of the switching remained unexplained. It was up to our group to clarify the switching mechanism which is due ionic motion in the oxide and a local redox process changing the cation valence and, hence, the electronic conductivity of the oxide. We were also able to show that the switchable region can be limited to a filamentary range of one to two nanometers in diameter. This results have enormous potential for further miniaturization of storage elements and a corresponding extension of Moore´s law.⁴

Another NVM switching effect exists for cells with electrochemical active metal electrodes such as silver or copper and a cation conductor. The first reported appeared in 1976.⁵ In 2007, we were able to introduce a new classification of the most important variants of redox-based RS effect⁶. In hindsight, one can guess why it took so long for the mechanism of RS to be elucidated. On the one hand, the complexity of the phenomenon and the confusing variety of variants presented a significant barrier to understanding the effect. On the other hand, the decisive interdisciplinary approach for understanding and consensus to emerge had been missing. This finally helped our group to arrive to the breakthrough: Behind an apparently purely solid-state physical phenomenon of a future nanoelectronic component was an electrochemical cause.

In 2008, the Hewlett-Packard labs discovered that the electrical characteristics of RS elements can be described in terms of the nonlinear dynamical theory of memristive devices,⁷ published by Kang and Chua in 1976.⁸ It should be noted that the theory of memristors⁹ introduced in 1971 does not apply to these devices, and up to today, there is no two-terminal device which can be described by the theory of memristors.

In the meantime, many further variants have been elucidated. One can distinguish between devices which rely on the switching in a conducting filament and devices which show an area-dependent interface type switching.¹⁰ Furthermore, one can distinguish two opposite switching polarities which even can occur in one cell and which we called eightwise and counter-eightwise.¹¹ The change of the resistive state can happen in either an abrupt or in a gradual manner and very recently we learned how to control this feature,¹² which will be of paramount importance for the future use of RS devices as electronic synapses for artificial neural networks in neuromorphic computers.

Text kindly provided by:

Prof. Rainer Waser

Institute of Electronic Materials (PGI-7)
Peter Grünberg Institut
Forschungszentrum Jülich
D-52425 Jülich, Germany

Institute of Materials in Electrical Engineering
and Information Technology 2 (IWE 2)
RWTH Aachen University
D-52056 Aachen, Germany

http://www.emrl.de/

Jülich-Aachen Research Alliance,
Section Fundamentals of Future Information Technology (JARA-FIT)

http://www.jara.org/en/research/fit

References:

[1]       W. R. Hiatt, T. W. Hickmott, Appl. Phys. Lett. 1965, 6, 106.

[2]       A. Asamitsu, Y. Tomioka, H. Kuwahara, Y. Tokura, Nature 1997, 388, 50.

[3]       A. Beck, J. G. Bednorz, C. Gerber, C. Rossel, D. Widmer, Appl. Phys. Lett. 2000, 77, 139.

[4]       K. Szot, W. Speier, G. Bihlmayer, R. Waser, Nat. Mater. 2006, 5, 312.

[5]       Y. Hirose, H. Hirose, J. Appl. Phys. 1976, 47, 2767.

[6]       R. Waser, M. Aono, Nat. Mater. 2007, 6, 833.

[7]       D. B. Strukov, G. S. Snider, D. R. Stewart, R. S. Williams, Nature 2008, 453, 80.

[8]       L.O. Chua, S.M. Kang, Proc. IEEE 1976, 64, 209.

[9]       L.O. Chua, IEEE Trans. Circuit Theory CT-18 1971, 507.

[10]     A. Sawa, Mater. Today 2008, 11, 28.

[11]     R. Muenstermann, T. Menke, R. Dittmann, R. Waser, Adv. Mater. 2010, 22, 4819.

[12]     A. Hardtdegen, C. La Torre, F. Cüppers, S. Menzel, R. Waser, S. Hoffmann-Eifert, IEEE Trans. Electron Devices 2018, 65.