Overcoming the Fundamental Barrier Thickness Limits of Ferroelectric Tunnel Junctions through BaTiO₃/SrTiO₃ Composite Barriers

Abstract

ABSTRACT:Ferroelectric tunnel junctions (FTJs) have attracted increasing research interest as a promising candidate for non-volatile memories. Recently, significant enhancements of tunneling electroresistance (TER) have been realized through modifications of electrode materials.However, direct control of the FTJ performance through modifying the tunneling barrier has not been adequately explored. Here, adding a new direction to FTJ research, we fabricated FTJs with Ferroelectric tunnel junctions (FTJs), composed of a thin ferroelectric (FE) layer sandwiched in between two metallic electrodes, have been intensively investigated in recent years. This device is considered to be a promising candidate for next-generation nonvolatile memories, because it combines the advantages of both ferroelectric random-access-memory and resistiveswitching memory. 1-3 The concept of FTJ was first proposed by Esaki in 1971, 4 but the research activities have not flourished untixl this decade. [5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21] The operation of FTJs has been mainly explained in terms of interfacial screening of polarization charges. [5][6][7][8] In a metal 1(M1)/FE/metal 2 (M2) structured FTJ (Figure 1a), the two asymmetric electrodes lead to unequal screen lengths and potential changes at metal/FE interfaces. Depending on the polarizations, the electrostatic potential profile of tunneling barrier will be varied. As a result, the tunneling resistance can be switched between low (ON) and high (OFF) values by polarization reversal, leading to the socalled tunneling electroresistance (TER) effect.Extensive experimental works based on this asymmetric screening scenario have been reported during the last decade. In 2009, the TER effect was first demonstrated using conducting atomic force microscopy (CAFM). 9-11 Subsequently, FTJs with highly reproducible performance were realized in capacitor geometry, which will be useful for practical applications. 12,13 Up to recently, most studies of FTJs have focused on improving device performance by modifying the electrode materials. [14][15][16][17][18][19][20] Typical examples include the use of lightly doped semiconductors, [14][15][16] For a thick barrier, these obstacles can be avoided, but the tunneling current becomes too small for realizing a practically useful device. As an alternative, several groups have theoretically proposed a new kind of FTJ using an FE/paraelectric (FE/PE) composite barrier. 27,28 The PE layer provides a new route to control the tunneling barrier potentials, and thus the TER can be tuned and significantly enhanced. However, there have been few experimental efforts to systematically attest this theoretical prediction. 29 In this paper, we report experimental control of the TER effect by directly manipulating the tunneling barrier thickness and composition. For this purpose, we fabricated two types of FTJs: Figure S5), which further confirm good epitaxial film qualities.We estimated the lattice constants of the oxide layers from XRD reci...

Giant tunnelling electroresistance in metal/ferroelectric/semiconductor tunnel junctions by engineering the Schottky barrier

“…19), in which the ON/OFF current ratio originates from a modulation on the height of the ferroelectric barrier. Recently, efforts have been made to enhance the TER performance by incorporating extra barriers between the ferroelectric barrier and the metallic electrode, for example, the CoO x layer at Co/BTO interface20, the ultrathin dielectric SrTiO 3 barrier at BTO/SrRuO 3 interface2122 and the metal-insulator phase transition insertion (La 0.50 Ca 0.50 )MnO 3 at BTO/(La 0.70 Sr 0.30 )MnO 3 interface23. In these FTJs, the extra barriers vary in height and/or in width in response to the polarization reversal and give a more efficient modulation on the junction transport.…”

Tunnel electroresistance in BiFeO3 junctions: size does matter

“…[4][5][6] Furthermore, ferroelectric domains and their dynamics 7-10 offer additional degrees of freedom and give rise to an analog memristive response of the tunnel junction [11][12][13][14] that emulates the behavior of synapses in neuromorphic networks. 15,16 Large tunnel electroresistance values of more than 10 3 are now achievable at room temperature, [17][18][19][20][21] which makes ferroelectric tunnel junctions interesting candidates for resistive memories. 3 In addition, coupling ferroelectric materials to strongly correlated oxide electrodes can provide a local, permanent, and switchable electric field able to trigger electronic or magnetic phase transitions.…”

Tsu-Esaki modeling of tunneling currents in ferroelectric tunnel junctions

“…Efforts have also been made to increase TER by adding an extra barrier adjacent to the FE layer. 19,39,58,62 Experimental Au/Co/BTO/LSMO/NdGaO 3 (NdGaO 3 is the substrate) junction was described with a step barrier potential consisting of the BTO ferroelectric layer and a passive layer of CoO x at the Co/BTO interface. The resistive switching was attributed to the field-induced charge redistribution at the ferroelectric-electrode interface changing the CoO x barrier height.…”

“…BTO/STO composite barriers, i.e., step barrier structures, have been compared to single BTO barriers, and the composite barriers show enhanced TER and enable effective control of the barrier potentials. 39 Recent experiments 58 of a hybrid structure with a ferroelectric thin film and a 2D semiconductor layer showed an OFF/ON resistance ratio of 10 4 . The reversible accumulation-depletion of majority carriers in the 2D semiconductor occurs in response to the switching of the FE barrier, thus altering the barrier at the interface between these two layers.…”

“…37 In a study of BaTiO 3 /SrTiO 3 (BTO/ STO) junctions, the currents were modeled either using an average barrier model or by using the Landauer formalism and approximating the transmission as T total % T BTO Ã T STO . 39 In a recent paper, the NEGF method and the Landauer formula were used to model I-V curves of step barrier FTJs. 40 The results suggest that both the height and thickness of a CoO x buffer layer forming at the Co/BaTiO 3 interface may change under bias.…”