DocumentCode :
1786833
Title :
Computing with hybrid CMOS/STO circuits
Author :
Kabir, Muhammad ; Stan, Mircea
Author_Institution :
Charles L. Brown Dept. of Electr. & Comput. Eng., Univ. of Virginia, Charlottesville, VA, USA
fYear :
2014
fDate :
1-5 June 2014
Firstpage :
1
Lastpage :
6
Abstract :
Recent research in spin torque nano-oscillators (STNO) have opened the possibility of using electron spin to generate sustained microwave oscillations. Furthermore, the experimental verification of synchronization of STNOs could allow communication and computation with nanoscaled oscillators. In this paper, we propose a hybrid MOSFET/STNO array which can be used for pattern recognition applications. First, we show that an array of electrically coupled STNOs obey the dynamics of Kuramoto´s weakly coupled oscillators [1]. This behavior allows us to use the STNO array to implement the oscillatory neurocomputer proposed by Hoppensteadt et. al. [2]. We next consider practical STNO device geometries which can be used in a parallel-connected array. We propose using a dual barrier magnetic tunnel junction (DMTJ) to produce strong, harmonic oscillation signals in the absence on an external magnetic field. Finally, we perform HSPICE simulations of a hybrid MOSFET/STNO array to show how it can be used for pattern recognition.
Keywords :
MOSFET; magnetic tunnelling; nanoelectronics; oscillators; DMTJ; HSPICE simulations; STNO device geometries; dual barrier magnetic tunnel junction; electron spin; external magnetic field; harmonic oscillation signals; hybrid CMOS-STO circuits; hybrid MOSFET-STNO array; oscillatory neurocomputer; parallel-connected array; pattern recognition applications; spin torque nanooscillators; sustained microwave oscillations; weakly coupled oscillators; Arrays; Geometry; Magnetic tunneling; Mathematical model; Oscillators; Synchronization; Torque; associative memory; spin torque oscillators; synchronization;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Design Automation Conference (DAC), 2014 51st ACM/EDAC/IEEE
Conference_Location :
San Francisco, CA
Type :
conf
DOI :
10.1145/2593069.2596673
Filename :
6881400
Link To Document :
بازگشت