A Complementary Metal Oxide Semiconductor Process-Compatible

Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)

Article

A CMOS compatible, ferroelectric tunnel junction Fabian Ambriz Vargas, Gitanjali Kolhatkar, Maxime Broyer, Azza Hadj Youssef, Rafik Nouar, Andranik Sarkissian, Reji Thomas, Carlos Gomez-Yanez, Marc A. Gauthier, and Andreas Ruediger ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.6b16173 • Publication Date (Web): 03 Apr 2017 Downloaded from http://pubs.acs.org on April 8, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Applied Materials & Interfaces is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

A CMOS compatible, ferroelectric tunnel junction Fabian Ambriz-Vargas, 1,* Gitanjali Kolhatkar, 1 Maxime Broyer,1 Azza HadjYoussef,1 Rafik Nouar,2 Andranik Sarkissian,2 Reji Thomas,1 Carlos GomezYáñez,3 Marc A. Gauthier1 and Andreas Ruediger1,*

1

Centre Énergie, Matériaux et Télécommunications, INRS, Varennes, Québec, J3X1S2, Canada.

2

Plasmionique Inc, 9092 Rimouski, J4X 2S3, Brossard, Québec, Canada

3

Departamento de Ingeniería en Metalurgia y Materiales-Instituto Politécnico Nacional,

Zacatenco, 07738, México. *

Corresponding authors e-mail: [email protected]; [email protected]

KEYWORDS: Ferroelectric Tunnel Junctions; CMOS process; Nanoscale characterization; Tunneling electroresistance effect; Electronic band alignment

1 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 20

ABSTRACT In recent years, the experimental demonstration of Ferroelectric Tunnel Junctions (FTJ) based on perovskite tunnel barriers has been reported. However, integrating these perovskite materials into conventional silicon memory technology remains challenging due to their lack of compatibility with the complementary metal oxide semiconductor process (CMOS). The present communication reports the fabrication of an FTJ based on a CMOS compatible tunnel barrier Hf0.5Zr0.5O2 (6 unit cells thick) on an equally CMOS compatible TiN electrode. Analysis of the FTJ by grazing angle incidence X-ray diffraction confirmed the formation of the noncentrosymmetric orthorhombic phase (2 , ferroelectric phase). The FTJ characterization is followed by the reconstruction of the electrostatic potential profile in the as-grown TiN/Hf0.5Zr0.5O2/Pt heterostructure. A direct tunneling current model across a trapezoidal barrier was used to correlate the electronic and electrical properties of our FTJ devices. The good agreement between the experimental and the theoretical model attests to the tunneling electroresistance effect (TER) in our FTJ device. A TER ratio of ~15 was calculated for the present FTJ device at low read voltage (+0.2 V). This study makes Hf0.5Zr0.5O2 a promising candidate for integration into conventional Si memory technology.


Page 3 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1. INTRODUCTION Nowadays, ferroelectric random access memories (Fe-RAM) are commercially available products. This type of memory combines the fast read/write access (~10 ns) of dynamic RAM (DRAM) with the non-volatility (the ability of retain information when power supply is OFF) of a FLASH memory.1 A Fe-RAM is composed of a ferroelectric capacitor where a ferroelectric thick film (~100 nm thick) is sandwiched between two electrodes. This film’s remnant polarization is then switched by an electric field between the top and bottom electrodes, resulting in two stables polarization states that are interpreted as ‘1’ and ‘0’ in the binary code.2 However, Fe-RAM faces some challenges such as the destructive read-out, which destroys the stored information during the reading process and makes it necessary to restore the data after the reading operation. A second drawback is the scaling limit. Indeed, Fe-RAMs are based on charge sensing and the lateral size of their capacitor cannot be reduced to the nanometer scale.3 Among the different noticeable types of semiconductor memories, ferroelectric tunnel junction (FTJ) devices are excellent candidates to overcome the current Fe-RAM limitations, as they combine good scalability with low operating energy, high operation speed (write time, 10 ns), high endurance (106 cycles), non-volatility and a simple structure.2 The theoretical concept of a FTJ was proposed by Esaki in 1971.4 A conventional FTJ is composed of an ultrathin ferroelectric layer (~2.8 nm thick) sandwiched between two metallic electrodes. By applying an electric field, the step barrier height at both interfaces (electrode-ferroelectric) can be modulated due to the polarization reversal of the tunnel barrier. This in turn gives rise to two different electrical resistance states (tunneling electroresistance effect, (TER)), that can be codified as a binary information (“ON” and “OFF” states).5 In spite of the advantages of the FTJs, this concept lost attention for more than 30 years due to the difficulty of producing ferroelectric films



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 20

with a thickness compatible with coherent quantum mechanic-tunneling.2 In fact, for a long time, it was believed that the critical thickness for ferroelectricity was in the 10-100 nm range. However, recent strain engineering studies showed that the ferroelectricity of a given film can be markedly enhanced by the strain caused by the lattice-mismatch between the film and the substrate.6 In 2009, Garcia et. al. reported the fabrication of a FTJ device based on a highly strained BaTiO3 ultrathin film (3 nm thick), where a giant tunneling electro resistance effect was demonstrated.7 Since then, a large TER effect has been reported in FTJ devices based on perovskite materials such as BaTiO3,8 PbTiO39 and BiFeO3.10 Yet, integrating FTJs with ferroelectric ultrathin films having a perovskite structure into the conventional Si-based memory technology remains challenging for many reasons, such as a poor interfacing with silicon (Si), an elevated crystallization temperature, and an electrical degradation caused by a forming gas treatment.11 Hafnium zirconium oxide (Hf0.5Zr0.5O2) is a very promising material for FTJ devices, and could present a solution to overcome the actual limitations of the semiconductor memories. Just a few years ago, in 2011, its ferroelectric behavior in ultrathin film form (layer thickness below 10 nm) was first discovered.11 This was a surprise since Hafnia and Zirconia had been studied for more than a century without evidence of polar ordering. In contrast to traditional perovskite materials, Hf0.5Zr0.5O2 presents advantages such as a high compatibility with the CMOS process, a low crystallization temperature (~400 oC) and an excellent compatibility with TiN electrodes, a favored material for mass production.12-14 Therefore, owing to the importance of Hf0.5Zr0.5O2 to the semiconductor industry, this work reports the fabrication of a FTJ comprising a TiN bottom electrode, a 6-unit-cell-thick Hf0.5Zr0.5O2 tunnel barrier (~2.8 nm thick) and a Pt top electrode. We also present an


Page 5 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


investigation of the TER effect by correlating the reconstructed electrostatic potential profile with the electrical transport properties using a direct tunneling current model proposed by Brinkman.15 2. EXPERIMENTAL SECTION Sample preparation. Hf0.5Zr0.5O2/TiN bilayers were grown by radio frequency magnetron sputtering (SPT310, Plasmionique Inc.) on (100) p-type Si substrates ( = 1 − 10 ∙ ). The TiN bottom electrode was deposited at 400 oC in a N2 and Ar atmosphere ( = ⁄ + = 70%) using a sputtering pressure of 10 mTorr. The subsequent growth of Hf0.5Zr0.5O2 was performed at 425 oC. During this step, the sputtering medium consisted in a mixture of Ar and O2 ( = ⁄ + = 50%) and the RF power on the one inch in diameter polycrystalline Hf0.5Zr0.5O2 target was fixed to 20 W. In both cases, the sputtering chamber was pumped prior to deposition, to a base pressure of ~10–5 Torr, using a dry pumping station. To eliminate the contamination and maintain the target composition homogeneous, the target surface was cleaned for 15 minutes by pre-sputtering prior to all the depositions (see Supporting Information, section 1). Microstructural and structural characterization. The surface morphology of each film was analyzed by Atomic Force Microscopy (AFM, Smart SPM1000-AIST-NT Inc.), while the layer thickness of the ferroelectric films was determined by X-ray reflectivity (XRR, Philips X’Pert Materials Research Diffractometer) and the software Gen X. For films thicker than 5 nm, the structural properties were studied by high-resolution X-ray diffraction (XRD) in grazing incidence mode using a Cu Kα radiation. Electronic characterization. The bandgap of a 20 nm-thick Hf0.5Zr0.5O2 layer grown on Si quartz (0001) was calculated with the Tauc plot method using the transmission spectra obtained



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 20

from a Perkin-Elmer UV/VIS spectrometer. X-ray photoelectron spectroscopy (XPS, VG Escalab 220i XL) was employed to determine the band alignment of the Hf0.5Zr0.5O2 film with the TiN and Pt substrates, it records the shallow core level, the valence band spectra and the Fermi energies. All high resolution spectra were collected in the constant pass energy (20 eV) mode. The binding energy scale was calibrated using a pure gold standard sample, and setting Au 4f7/2 at a binding energy of 84 eV (see Supporting Information, section 2). Nanoscale characterization. The ferroelectric switching of the films was studied by piezoresponse force microscopy (PFM), using a conductive Pt-Ir coated Si cantilever tips (radius of ~30 nm). Nanoscale electrical measurements were performed in conductive AFM (C-AFM) mode. For the electrical characterization, a 30 nm thick polycrystalline Pt top electrode was deposited using DC sputtering through a shadow mask consisting of an array of 300 µm diameter holes (see Supporting Information, section 3). Local current–voltage (I-V) characteristics were measured by placing the conductive tip in contact with the platinum top electrode while the TiN bottom electrode was grounded, and performing a “DC” voltage sweep. 3. RESULTS AND DISCUSSION The surface morphology of the as grown Hf0.5Zr0.5O2 sample is shown in Figure 1a. It depicts a grainy layer uniformly covering the TiN substrate surface. The layer is thin enough that some atomic steps/terraces are still visible. The small grains characteristic of Hf0.5Zr0.5O2 are also visible, and a root-mean-square (rms) roughness of ~0.25 nm is obtained over a 11 area, attesting the atomic flatness of the Hf0.5Zr0.5O2/TiN heterostructure’s surface. As the TER effect is highly dependent on the barrier width,5 it is crucial to determine the layer thickness of our films. Using XRR measurements (Figure 1b), the layer thickness was estimated


Page 7 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


to ~2.8 nm (6 unit cells of Hf0.5Zr0.5O2) for the films used in the fabrication of the FTJ and ~5.2 nm for those used in the structural analysis. To gain insight into the structural properties of the ultrathin Hf0.5Zr0.5O2 film on the TiN substrate, a phase identification was performed using Grazing incidence Angle XRD (Figure 1c). The diffraction peaks from the (111), (200) and (220) planes belong to the non-centrosymmetric orthorhombic phase, which has ferroelectric properties. The ferroelectric phase is obtained during the initial stages of film growth. It starts with the nucleation of small grains with a high surface and volume ratio, which results in the formation of a tetragonal phase (4 /!). However, the large tensile strain along the c-axis induced by the coalescence of the nucleating grains allows the formation of the ferroelectric orthorhombic phase (2 ).16 The XRD presented in Figure 1c confirms the film composition, as a change in the Hf/Zr ratio would result in the appearance of other peaks in 2ϴ.17



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 20

Figure 1. (a) 1 × 1 µm2 AFM topography image of an Hf0.5Zr0.5O2 film on a TiN substrate, (b) XRR spectra measured on Hf0.5Zr0.5O2 films and (c) gracing incidence angle XRD spectrum of an Hf0.5Zr0.5O2 film deposited on a TiN bottom electrode.

The first requirement to realize a ferroelectric tunnel junction device is for the nanometer-thick film to present ferroelectric properties.5 For this reason, the ferroelectric properties of the 6-unitcell-thick Hf0.5Zr0.5O2 film were investigated at room temperature on the bare surface of the ultrathin film as well as on the FTJ devices. Starting with the PFM response on the bare surface of the Hf0.5Zr0.5O2 film, the Figure 2a revealed a clear hysteretic behavior in both phase, as shown by the ~180$ difference, and the amplitude signals, acquired out of the resonance frequency of the cantilever, as a function of voltage. The ferroelectric switching voltages that create the downward and upward polarization were found to be +1.9 ' and −1.9 ', respectively. Similar local phase and amplitude signals were acquired on the Hf0.5Zr0.5O2 film integrated in the FTJ device (Supporting Information, Figure S4). These results confirm the ferroelectric behavior of the ultrathin Hf0.5Zr0.5O2 film. Figure 2b shows a PFM phase-contrast image depicting intentionally-written ferroelectric patterns. Three intersecting squares located from the bottom left to the top right were poled with voltages of −3', +3' and −3 ', demonstrating upward, downward and upward polarizations, respectively. The remaining area corresponds to the virgin state, which exhibits a preferential downward-polarization. This spontaneous polarization was caused by the presence of substrate-induced strains in the Hf0.5Zr0.5O2 film,18 while the intersecting squares show the capability of the material to go through a full cycle of ferroelectric polarization switching. The presented ferroelectric patterns are stable for over 60 hours, which suggests a stable and robust ferroelectric polarization of the Hf0.5Zr0.5O2 layer (Supporting Information, Figure S5). This PFM analysis is in agreement with the XRD results, confirming the formation of the non-centrosymmetric orthorhombic 8 ACS Paragon Plus Environment

Page 9 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


ferroelectric phase.

Figure 2. (a) Local PFM hysteresis loops of the phase (blue curve) and the amplitude signal (green-red curve) and (b) 3 × 3 µm2 PFM phase-contrast image.

Optical constants such as the bandgap ()* ) can be estimated from UV-visible transmission spectroscopy measurements, while the carrier transport properties at the ferroelectric /metal interface can be well understood by determining the valence band offset as well as the potential step barrier.19-20 So, accurate knowledge of these optical and electronic parameters is essential for identifying a suitable ferroelectric layer for FTJ.21 Eg is determined from transmission measurements. The corresponding Tauc plot is illustrated in Figure 3. The absorption coefficient (∝) was calculated using the following relation,19

$$

∝= , ln / %0 1,

(1)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 20

where 3 is the layer thickness and 4 is the transmission. Eg is estimated by extrapolating the linear portion of the absorption curve to the x-axis, where the absorption coefficient becomes zero, and a value of 5.06 eV was obtained for an Hf0.5Zr0.5O2 film with a thickness of 20 nm.

Figure 3. Tauc plot of a Hf0.5Zr0.5O2 film. To gain quantitative information on the electronic properties at both interfaces of the FTJ system (TiN/Hf0.5Zr0.5O2 and Hf0.5Zr0.5O2/Pt), we used the Kraut procedure.22-23 This method consists in measuring the valence band offset (VBO) by finding the core level signals provided by the film and the substrate, the Fermi energy level (EF) of the metallic bulk material as well as the valence band maximum (VBM) of the ferroelectric layer. Starting with the first interface (TiN/ Hf0.5Zr0.5O2), VBO was calculated using the following equation, '5 = 5)067 − )8 06 − 5)9:;: − '5 − 5)067 − 5)9:;: 9=>/06

(2)

The energy separation between the Ti2p centroid with respect to the leading edge of the Fermi level for the bulk TiN sample, was found to be 455.4?0.05 eV. The energy difference between Hf4f centroid and the VBM for the bulk Hf0.5Zr0.5O2 was estimated to be 14.1 eV?0.05. The binding energy difference between Ti2p and Hf4f core levels in the ultrathin film, was calculated


Page 11 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


to be 438.1?0.05 eV. Substituting these values in Eq. 2, we determined the VBO to be 3.2?0.05 eV at the TiN/ Hf0.5Zr0.5O2 interface. This analysis and the obtained values are schematically described in Figure 4. The same procedure was employed to measure the VBO at the second interface (Hf0.5Zr0.5O2/Pt, Figure 4). For the reference bulk metallic Pt sample, we obtain an energy difference equal to 70.5?0.05 eV, while the core level line separation on the Pt/ Hf0.5Zr0.5O2 ultrathin sample was found to be 53.7?0.05 eV. From these values we found the second VBO to be equal to 2.70?0.05 eV (see Supporting Information, section 5).

Figure 4. XPS spectra taken on the (a) thick TiN film, (b) thick Hf0.5Zr0.5O2 layer, (c) TiN/ Hf0.5Zr0.5O2 interface, (d) bulk Pt film and (e) Hf0.5Zr0.5O2/Pt interface. The conduction band offset (or potential step barrier,@) is given by @ = )* − '5

(3)

Therefore, the potential step barriers at the TiN/ Hf0.5Zr0.5O2 (@ ) and at the Hf0.5Zr0.5O2/Pt (@ ) interfaces were calculated to be 1.86?0.07eV and 2.36?0.07eV, respectively. The reconstructed 11 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 20

electronic band diagram for the TiN/ Hf0.5Zr0.5O2/Pt heterostructure with spontaneous downward polarization is shown in Figure 5.

Figure 5. Schematic representation of the electronic band diagram for the TiN/ Hf0.5Zr0.5O2/Pt heterostructure. The TER effect in our system was investigated using nanoscale electrical measurements. A representative I-V curve measured on the TiN/Hf0.5Zr0.5O2/Pt devices is presented in Figure 6a. It depicts a hysteretic I-V behavior characterized by an initial high resistance state (“OFF” state, green line), which switched to a low resistance state (“ON” state, blue line). The hysteretic electrical behavior as well as the large change in electrical resistance between the two states suggest a TER effect.8, 18, 24


Page 13 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 6. (a) Typical I-V characteristic measured on a 300 µm diameter the TiN/Hf0.5Zr0.5O2/Pt heterostructure displaying low and high resistance states, (b) Electrostatic potential barrier as a function of the polarization direction. In order to confirm that the resistive switching observed in Figure 6a is due to the TER effect rather than another resistive switching mechanism, a method proposed by Zenkevich et. al.24 was used. It correlates the interfacial electronic properties with the electrical transport properties of the TiN/Hf0.5Zr0.5O2/Pt heterostructure by employing a direct tunneling current model across a trapezoidal barrier.15 This model consists in calculating the tunneling current density (A) as a function of voltage ('), using the potential barrier steps at both ferroelectric-electrode interfaces (@ and @ ), layer thickness (d), electron charge (B) and the effective electron mass (m) as described by the Brinkman model8, 15 ;CD

A = − /EFG ћI1 J

I I TU TU G SRV W G XY G G V V TU TU OG PQRG S G SRV W G XG G G

KLM NOPQRG S

_

Z [ sinh N `'Q@ −

CP V

G − /@ +

CP V CP

G a

Y,

(4)

where ∝ b = c432⁄ ⁄Q3ћ @ + B' − @ d and ћ is the reduced Planck constant (see Supporting Information, section 6). The theoretical I-V curve for the “ON” state (downward polarization) was plotted using the potential barrier values obtained by XPS (@ =1.86 eV and @ =2.36 eV), and adding only the scaling factor, which was assumed to be the same for both



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 20

states, following the method described in the literature.8, 24 Then, using the same scaling factor and Eq. 4, the I-V curve for the “OFF” state was fitted to obtain the potential barriers for that state. As shown in Figure 6a, a good agreement was obtained during the fitting of the I-V curve with the theoretical model described in Eq. 4 for both the “OFF” and “ON” states. From this model, we obtained two potential barriers (@ = 2.75 B'and @ = 2.20 B'), which produce the “OFF” state in the I-V curve (upward polarization). Hence, the polarization reversal changes at the TiN/Hf0.5Zr0.5O2 and the Hf0.5Zr0.5O2/Pt interfaces are ∆@ = 0.89 B' and ∆@ = 0.16 B', respectively. The derived changes in the potential barrier profile are shown in Figure 6b. Indeed, the electrostatic potential profile was modulated by the polarization reversal of the ultrathin layer. When the polarization vector of Hf0.5Zr0.5O2 points to the TiN bottom electrode, the conduction band will bend downward due to the build-up of negative screening charges in the TiN electrode. This will be followed by the lowering of the potential barrier height at Hf0.5Zr0.5O2/TiN interface, resulting in a low-resistance state (Figure 6b, downward state). On the other hand, when the polarization vector of Hf0.5Zr0.5O2 points to the Pt top electrode, positive screening charges will build up at the Hf0.5Zr0.5O2/TiN interface, bending the conduction band upward and resulting in a high-resistance state (Figure 6b, upward state).The TER effect in our FTJ was quantified by the TER ratio (Ag − Ah ⁄Ah ), resulting in a 4)i~ 15 at +0.2 V. The I-V curve illustrated in Figure S11 (Supporting Information) showed a change from “ON” state to the “OFF” at a read voltage of +2.2 V while the change from “OFF” state to “ON” state takes place at a read voltage of -1.7 V. The endurance of the TiN/Hf0.5Zr0.5O2/Pt heterostructures was evaluated by recording 1000 I-V cycles under quasi-static conditions on the same FTJ memory cell. Given that contacting by the AFM was required, the acquisition speed was limited by the AFM specifications and a 1000 cycle quasi-static measurement already took a day. For


Page 15 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


each I-V curve, we obtained the “ON” and “OFF” resistances at a read voltage of + 0.2 V (Figure 7a). They remain stable over 1000 cycles, indicating a good endurance of the FTJs. The reproducibility of the TER effect in the FTJs under study was investigated by recording I-V cycles from 30 different FTJ devices. For each memory cell, the resistance values for both the “ON” and “OFF” states were recorded at 0.2 V. The 30 memories cells presented an ON/OFF ratio, or TER ratio, of 15±3 on average (Figure 7b). This result attests to the high reproducibility of the TER effect in the TiN/ Hf0.5Zr0.5O2/Pt FTJ memory devices. The retention properties of the present FTJ were evaluated on a single memory cell, by recording the “ON” resistance state value over time at a read voltage of +0.4 V (following the method described by M. Abuwasib et. al.26). As shown in Figure 7c the “ON” resistance state can still be read after eight hours, demonstrating that our FTJ devices present a long-time retention.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 20

Figure 7. Memory performances of the FTJ devices: (a) Fatigue measurements on a TiN/ Hf0.5Zr0.5O2/Pt (diameter of 300 µm) heterostructure displaying the ‘OFF’ (green) and ‘ON’ (blue) resistance values over 1000 write/read cycles. (b) ‘ON’ and ‘OFF’ resistances state (upper panel) and ON/OFF ratios (bottom panel) of 30 different FTJ memory cells and (c) Retention properties, where the “ON” resistance state was measured as function of time at a read voltage of 0.4 V.


Page 17 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


4. CONCLUSIONS In summary, we reported the successful fabrication of an FTJ device based on a CMOS compatible tunnel barrier (Hf0.5Zr0.5O2) and a TiN bottom electrode. The formation of the metastable ferroelectric orthorhombic phase was confirmed by correlating the results of grazing angle incidence XRD with those acquired by PFM. By combining XPS and UV-Visible spectroscopies, the potential step barriers were determined to be 1.86 eV and 2.32 eV at the bottom and top interfaces of the TiN/Hf0.5Zr0.5O2/Pt heterostructure, respectively. Using the Brinkman model, we confirmed the presence of a large electroresistance effect, with a TER ratio equal to ~15 (at +0.2 V). Our results hold the promise for novel semiconductor memory devices, since FTJs have the potential to overcome the limitation of conventional semiconductor memory devices based on charge storage such as Fe-RAM and DRAM. Supporting Information. See supporting information section for additional information related to the synthesis of the Hf0.5Zr0.5O2 and TiN films, complete XPS binding energy calibration, structural properties of the platinum top electrode, complete ferroelectric studies, complete band alignment analysis and an additional information related to the Brinkman model. ACKNOWLEDGEMENTS F.A.V. is thankful for an individual FRQNT MELS PBEEE 1M scholarship and the financial support of CONACYT-Mexico. G.K. acknowledges a postdoctoral FRQNT research scholarship. A.R. and M.G. acknowledge generous support through NSERC discovery grants. The collaboration with C. Gomez-Yañez was generously sponsored through an MRIFE grant.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 20

REFERENCES (1) Jeong, D. S.; Thomas, R.; Katiyar, R. S.; Scott, J. F.; Kohlstedt, H.; Petraru, A.; Hwang, C. S. Emerging Memories: Resistive Switching Mechanisms and Current Status. Rep. Prog. Phys. 2012, 75 (7), 076502. (2) Garcia, V.; Bibes, M. Ferroelectric Tunnel Junctions for Information Storage and Processing. Nat. Commun. 2014, 5, 4289. (3) Tsymbal, E. Y.; Gruverman, A. Ferroelectric Tunnel Junctions: Beyond the Barrier. Nat. Mater. 2013, 12, 602-604. (4) Esaki, L.; Laibowitz, R. B.; Stiles, P. J. Polar Switch. IBM Tech. Discl. Bull 1971, 13, 2161. (5) Tsymbal, E. Y.; Kohlstedt, H. Applied Physics. Tunneling Across a Ferroelectric. Sci. 2006, 313 (5784), 181-183. (6) Choi, K. J.; Biegalski, M.; Li, Y. L.; Sharan, A.; Schubert, J.; Uecker, R.; Reiche, P.; Chen, Y. B.; Pan, X. Q.; Gopalan, V.; Chen, L.-Q.; Schlom, D. G.; Eom, C. B. Enhancement of Ferroelectricity in Strained BaTiO3 Thin Films. Sci. 2004, 306, 1005-1008. (7) Garcia, V.; Fusil, S.; Bouzehouane, K.; Enouz-Vedrenne, S.; Mathur, N. D.; Barthelemy, A.; Bibes, M. Giant Tunnel Electroresistance for Non-destructive Readout of Ferroelectric States. Nat. 2009, 460 (7251), 81-84. (8) Gruverman, A.; Wu, D.; Lu, H.; Wang, Y.; Jang, H. W.; Folkman, C. M.; Zhuravlev, M. Y.; Felker, D.; Rzchowski, M.; Eom, C.-B.; Tsymbal, E. Y. Tunneling Electroresistance Effect in Ferroelectric Tunnel Junctions at the Nanoscale. Nano lett. 2009, 9 (10), 3539-3543. (9) Crassous, A.; Garcia, V.; Bouzehouane, K.; Fusil, S.; Vlooswijk, A. H. G.; Rispens, G.; Noheda, B.; Bibes, M.; Barthélémy, A. Giant Tunnel Electroresistance with PbTiO3 Ferroelectric Tunnel Barriers. Appl. Phys. Lett. 2010, 96 (4), 042901. (10) Yamada, H.; Garcia, V.; Fusil, S.; Boyn, S.; Marinova, M.; Gloter, A.; Xavier, S.; Grollier, J.; Jacquet, E.; Carretero, C.; Deranlot, C.; Bibes, M.; Barthelemy, A., Giant Electroresistance of Super-Tetragonal BiFeO3‑Based Ferroelectric Tunnel Junctions. J. Am. Chem. Soc. NANO 2013, 7, 5385–5390. (11) Müller, J.; Böscke, T. S.; Bräuhaus, D.; Schröder, U.; Böttger, U.; Sundqvist, J.; Kücher, P.; Mikolajick, T.; Frey, L. Ferroelectric Zr0.5Hf0.5O2 Thin Films for Nonvolatile Memory Applications. Appl. Phys. Lett. 2011, 99 (11), 112901. (12) Hyuk, P. M.; Joon, K. H.; Jin, K. Y.; Lee, W.; Kyeom Kim, H.; Seong Hwang, C., Effect of Forming Gas Annealing on the Ferroelectric Properties of Hf0.5Zr0.5O2 Thin Films with and without Pt Electrodes. Appl. Phys. Lett. 2013, 102 (11), 112914. (13) Hyuk, P. M.; Joon, Kim. H.; Jin, K. Y.; Lee, W.; Moon, T.; Seong, H. C. Evolution of Phases and Ferroelectric Properties of Thin Hf0.5Zr0.5O2 Films According to the Thickness and Annealing Temperature. Appl. Phys. Lett. 2013, 102 (24), 242905. (14) Kim, H. J.; Park, M. H.; Kim, Y. J.; Lee, Y. H.; Jeon, W.; Gwon, T.; Moon, T.; Kim, K. D.; Hwang, C. S. Grain Size Engineering for Ferroelectric Hf0.5Zr0.5O2 Films by an Insertion of Al2O3 Interlayer. Appl. Phys. Lett. 2014, 105 (19), 192903. (15) Brinkman, W. F. Tunneling Conductance of Asymmetrical Barriers. J. Appl. Phys. 1970, 41, 1915. (16) Hyuk, P. M.; Joon, K. H.; Jin, K. Y.; Moon, T.; Seong, H. C. The Effects of Crystallographic Orientation and Strain of Thin Hf0.5Zr0.5O2 film on its Ferroelectricity. Appl. Phys. Lett. 2014, 104 (7), 072901. 18 ACS Paragon Plus Environment

Page 19 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(17) Muller, J.; Boscke, T. S.; Schroder, U.; Mueller, S.; Brauhaus, D.; Bottger, U.; Frey, L.; Mikolajick, T. Ferroelectricity in Simple Binary ZrO2 and HfO2. Nano Lett. 2012, 12 (8), 43184323. (18) Soni, R.; Petraru, A.; Meuffels, P.; Vavra, O.; Ziegler, M.; Kim, S. K.; Jeong, D. S.; Pertsev, N. A.; Kohlstedt, H. Giant Electrode Effect on Tunnelling Electroresistance in Ferroelectric Tunnel Junctions. Nat. Commun. 2014, 5, 5414. (19) Pavunny, S. P.; Thomas, R.; Kumar, A.; Fachini, E.; Katiyar, R. S. Optical Properties of Amorphous High-k LaGdO3 Films and its Band Alignment with Si. J. Appl. Phys. 2012, 111 (4), 044106. (20) Pavunny, S. P.; Thomas, R.; Kumar, A.; Scott, J. F.; Katiyar, R. S. Optical Dielectric Function Modeling and Electronic Band Lineup Estimation of Amorphous High-k LaGdO3 Films. ECS J. Solid State Sci. Technol. 2012, 1 (4), N53-N57. (21) Chernikova, A.; Kozodaev, M.; Markeev, A.; Negrov, D.; Spiridonov, M.; Zarubin, S.; Bak, O.; Buragohain, P.; Lu, H.; Suvorova, E.; Gruverman, A.; Zenkevich, A. Ultrathin Hf0.5Zr0.5O2 Ferroelectric Films on Si. ACS Appl. Mater. Interfaces 2016, 8 (11), 7232-7237. (22) Kraut, E. A.; Grant, R. W.; Waldrop, J. R.; Kowalczyk, S. P. Precise Determination of the Valence-Band Edge in X-Ray Photoemission Spectra: Application to Measurement of Semiconductor Interface Potentials. Phys. Rev. Lett. 1980, 44, 1620. (23) Kraut, E. A.; Grant, R. W.; Waldrop, J. R.; Kowalczyk, S. P. Semiconductor Core-Level to Valence-Band Maximum Binding-Energy Differences: Precise Determination by X-ray Photoelectron Spectroscopy. Phys. Rev. B 1983, 28, 1965. (24) Zenkevich, A.; Minnekaev, M.; Matveyev, Y.; Lebedinskii, Y.; Bulakh, K.; Chouprik, A.; Baturin, A.; Maksimova, K.; Thiess, S.; Drube, W. Electronic Band Alignment and Electron Transport in Cr/BaTiO3/Pt Ferroelectric Tunnel Junctions. Appl. Phys. Lett. 2013, 102 (6), 062907. (25) Wen, Z.; Li, C.; Wu, D.; Li, A.; Ming, N. Ferroelectric-Field-Effect-Enhanced Electroresistance in Metal/Ferroelectric/Semiconductor Tunnel Junctions. Nat. Mater. 2013, 12, 617-621. (26) Abuwasib, M.; Lu, H.; Li, T.; Buragohain, P.; Lee, H.; Eom, C.-B.; Gruverman, A.; Singisetti, U. Scaling of Electroresistance Effect in Fully Integrated Ferroelectric Tunnel Junctions. Appl. Phys. Lett. 2016, 108 (15), 152904.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 20

Table of contents/Abstract Graphic


A Complementary Metal Oxide Semiconductor Process-Compatible

Recommend Documents