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...