BaTiO 3 thin films on platinized silicon: Growth, characterization and resistive memory behavior A. Román a,b , M. Rengifo a,b , L.M. Saleh Medina b,c , M. Reinoso a,b,d , R.M. Negri b,c , L.B. Steren a,b , D. Rubi a,b,d, ⁎ a Gereencia de Investigación y Aplicaciones, CNEA, Av. Gral Paz 1499 (1650), San Martín, Buenos Aires, Argentina b Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Argentina. c Instituto de Química Física de Materiales, Ambiente y Energía (INQUIMAE) and Departamento de Química Inorgánica, Analítica y Química Física, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Buenos Aires, Argentina d Escuela de Ciencia y Tecnología, UNSAM, Campus Miguelete (1650), San Martín, Buenos Aires, Argentina abstract article info Article history: Received 11 October 2016 Received in revised form 8 February 2017 Accepted 17 March 2017 Available online 18 March 2017 We report on the fabrication and characterization of Ti/BaTiO 3 /Pt memristive devices. BaTiO 3 films were grown on platinized silicon by pulsed laser deposition with different laser pulse energies. We prove the existence of a correlation between the fabrication conditions and the microstructure and stoichiometry of the films. It is sug- gested that the small grain size found on our BaTiO 3 films destabilizes the structural tetragonal distortion and in- hibits the appearance of long-range ferroelectric ordering. We show that even in absence of ferroelectric resistive switching (RS), two different RS mechanisms (metallic filament formation and oxidation/reduction of the Ti top electrode) compete, and can be selected by controlling the films stoichiometry and microstructure. © 2017 Elsevier B.V. All rights reserved. Keywords: Memristive oxides BaTiO 3 Thin films on silicon 1. Introduction The sustained increase of memory densities over the last decades has been, to a large extent, responsible for the remarkable evolution of micro- and nanoelectronics present in laptops, digital cameras, mobile phones, etc. Further progress, especially in portable appliances, crucially depends on the development of new solid-state memories. Today, dy- namic random access memories (DRAM) and Flash represent the dom- inant memory technologies. DRAMs are fast and show almost unlimited cycle times for write and read operation, but they are volatile, i.e. they lose the stored information when the supply voltage is cut off. In addi- tion, a further increase in storage density faces huge obstacles since their storage principle is charge-based. Although Flash memories are non-volatile and show better scaling potential than DRAMs, their write access is relatively slow and the number of write cycles is limited. The “holy grail” of the memory technology would be a RAM bringing to- gether non-volatility, speed, durability and extended scaling. Different technologies have been proposed to accomplish these requirements, i.e. phase-change memories, magnetic or ferroelectric random access memories and resistive random access memories (RRAM). The latter are based on materials which show a significant, non-volatile, change in their resistance upon application of electrical stress, a mechanism that from early 00′s is usually called resistive switching (RS) (see re- views [1,2] and references therein). RS has been found to ubiquitously exist in a huge variety of simple and complex transition metal oxides. A typical RS device, usually called “memristor”, consists in a metal/ oxide/metal stack, usually laterally confined by means of standard micro or nanofabrication techniques. Although still immature compared to other approaches, RS materials show promising properties in terms of scalability, low power consumption and fast write/read access times. In addition, memristors were shown to exhibit a similar behavior than the bit-cells of the human brain (synapses) [3], suggesting the pos- sibility of developing disruptive devices with neuromorphic behavior. Up to now, the microscopic physical and chemical mechanisms un- derneath RS have not been fully understood. Different RS mechanisms have been proposed, being the most widely accepted related to the cre- ation/disruption of conducting filaments bridging both electrodes, or to the drift-diffusion of oxygen vacancies that modulate the height of the Schottky barrier at the oxide/metal interface [1]. More recently, it has been proposed that in the case of ferroelectric oxides, the direction of the spontaneous polarization can also modulate the Schottky barrier at the oxide/metal interface, and therefore the device resistance [4–12]. These devices have the advantage of not relying on voltage- induced migration of ions -i.e. cations coming from the metallic elec- trode or oxygen vacancies- but on a pure electronic mechanism, being expected to display superior reliability and endurance. Most reports on ferroelectric memristors deal with the room tem- perature multiferroic BiFeO 3 [6, 10–13] or with the canonical room Thin Solid Films 628 (2017) 208–213 ⁎ Corresponding author at: Gerencia de Investigación y Aplicaciones, CNEA, Av. Gral Paz 1499 (1650), San Martín, Buenos Aires, Argentina. E-mail address: rubi@tandar.cnea.gov.ar (D. Rubi). http://dx.doi.org/10.1016/j.tsf.2017.03.038 0040-6090/© 2017 Elsevier B.V. All rights reserved. Contents lists available at ScienceDirect Thin Solid Films journal homepage: www.elsevier.com/locate/tsf