Factors limiting ferroelectric field-effect doping in complex oxide heterostructures

L. Bégon-Lours, V. Rouco, Qiao Qiao, A. Sander, M. A. Roldán, R. Bernard, J. Trastoy, A. Crassous, E. Jacquet, K. Bouzehouane, M. Bibes, J. Santamaría, A. Barthélémy, M. Varela, and Javier E. Villegas

Phys. Rev. Materials 2, 084405 – Published 14 August 2018

Abstract

Ferroelectric field-effect doping has emerged as a powerful approach to manipulate the ground state of correlated oxides, opening the door to a different class of field-effect devices. However, this potential is not fully exploited so far, since the size of the field-effect doping is generally much smaller than expected. Here we study the limiting factors through magnetotransport and scanning transmission electron and piezoresponse force microscopy in ferroelectric/superconductor $(YB a_{2} C u_{3} O_{7 − δ} / BiFe O_{3})$ heterostructures, a model system showing very strong field effects. Still, we find that they are limited in the first place by an incomplete ferroelectric switching. This can be explained by the existence of a preferential polarization direction set by the atomic terminations at the interface. More importantly, we also find that the field-effect carrier doping is accompanied by a strong modulation of the carrier mobility. Besides making quantification of field effects via Hall measurements not straightforward, this finding suggests that ferroelectric poling produces structural changes (e.g., charged defects or structural distortions) in the correlated oxide channel. Those findings have important consequences for the understanding of ferroelectric field effects and for the strategies to further enhance them.

Received 23 March 2018
Revised 21 June 2018

DOI:https://doi.org/10.1103/PhysRevMaterials.2.084405

Physics Subject Headings (PhySH)

Ferroelectricity Superconductivity

Heterostructures High-temperature superconductors Strongly correlated systems Thin films

Condensed Matter, Materials & Applied Physics

Authors & Affiliations

L. Bégon-Lours^1,*, V. Rouco^1,†, Qiao Qiao², A. Sander¹, M. A. Roldán³, R. Bernard^1,‡, J. Trastoy¹, A. Crassous¹, E. Jacquet¹, K. Bouzehouane¹, M. Bibes¹, J. Santamaría^1,3, A. Barthélémy¹, M. Varela³, and Javier E. Villegas^1,§

¹Unité Mixte de Physique, CNRS, Thales, Université Paris-Sud, Université Paris Saclay, 91767 Palaiseau, France
²Oak Ridge National Laboratories, Oak Ridge Tennessee 37840, USA
³Grupo de Física de Materiales Complejos, Dpto. Física de Materiales, Universidad Complutense de Madrid, 28040 Madrid, Spain

^*Present address: MESA+ Institute, Inorganic Materials Science, Enschede, The Netherlands.
^†Corresponding author: victor.rouco@cnrs-thales.fr
^‡Present address: Fonctions Optiques pour les Technologies de l’Information, UMR 6082, CNRS, INSA de Rennes, Rennes, France.
^§Corresponding author: javier.villegas@cnrs-thales.fr

Article Text (Subscription Required)

Click to Expand

Supplemental Material (Subscription Required)

Click to Expand

References (Subscription Required)

Click to Expand

Issue

Vol. 2, Iss. 8 — August 2018

Reuse & Permissions

Access Options

Article Available via CHORUS

Download Accepted Manuscript

Author publication services for translation and copyediting assistance advertisement

Images

Figure 1
(a) Resistance vs temperature of a BFO-Mn (30 nm)/YBCO(4 u.c.)/PBCO(2 u.c.)/STO/ heterostructure, measured with an injected current $J = 230 A c m^{- 2}$ in the virgin state (green curve), after poling the ferroelectric towards the YBCO (depleted state, blue curve) and away from the YBCO (accumulated state, red curve). The inset shows a sketch of the lithographed multiprobe bridge used for electrical measurements. In the central area (between the four voltage probes $V_{1} - V_{4})$ , the remnant ferroelectric polarization is set by applying $| V_{dc} |$ (typically 5 V) between the scanning AFM tip and the YBCO. (b) Critical temperature $T_{c}$ vs the $1 / e R_{H}$ for a series of samples with the same BFO thickness (30 nm) and variable YBCO thickness (symbols’ legend in unit cells). For a given sample (symbol), red/blue corresponds to the accumulated/depleted state, and green to the virgin state. The dashed curve is the $T_{c}$ vs charge-carrier density $n$ for bulk YBCO. The experimental $T_{c}$ is defined as the temperature at which the resistance drops to 90% of that at the onset of transition. (c) YBCO thickness $t_{YBCO}$ times the variation $Δ (1 / R_{H})$ between depleted and doped states, plotted as a function of $t_{YBCO}$ . The shaded area is a guide to the eye. The dashed lines correspond to simulations with the model of Fig. 5.
Reuse & Permissions
Figure 2
(a) High-resolution HAADF image of a BFO-Mn (30 nm)/YBCO(3 u.c.)/PBCO(4 u.c.)/STO sample (top), along with a sketch of the proposed atomic plane stacking sequence at the BFO/YBCO interface (bottom). Y, Cu, Ba, Fe, and Bi atomic columns are shown in dark blue, blue, green, red, and light brown, respectively. (b) EELS atomic resolution maps across the stacking, obtained from MLLS fits of raw EEL spectrum images. From top to bottom: $Ti L_{2, 3}, Fe L_{2, 3}, Ba M_{4, 5}$ , and $\Pr M_{4, 5}$ maps, along with an overlay of the Ti (yellow), Fe (red), Ba (green), and Pr (blue) signals. Some spatial drift is observed. The panel at the bottom exhibits the laterally averaged normalized signals along the direction marked with a blue arrow in (a) corresponding to the four chemical signals, same color scale. Vertical dashed lines mark the position of the interfaces, while horizontal maps highlight the 25% and 75% levels.
Reuse & Permissions
Figure 3
(a) STEM-HAADF cross-sectional image of an as-grown BFO-Mn (30 nm)/YBCO(5 u.c.)/PBCO(4 u.c.)/STO/ heterostructure. Panels (b)–(d) exhibit false color maps of the off-center displacements of Fe columns relative to the center of the Bi sublattice, measured for every unit cell (represented by a pixel) along the [001], [100] directions and in the (010) plane, respectively. A positive $d_{z}$ means that the Fe columns exhibit a net displacement that points away from the interface. The maps correspond to the BFO area framed with blue dashed lines in (a). Each color map is accompanied by a profile of the Fe displacement averaged laterally over every whole pixel line. All the scales are in angströms.
Reuse & Permissions
Figure 4
(a) PFM phase and (b) amplitude image of the surface of a BFO-Mn (30 nm) /YBCO (4 u.c.)/PBCO (4 u.c.)/STO/heterostructure. The image size is $6 μ m \times 6 μ m$ . The background corresponds to the virgin state. A square with opposite PFM phase contrast (dark) has been written by applying a dc voltage $V_{dc} = 5 V$ between the YBCO layer and the scanning tip (ground). The phase contrast is reversed again in a concentric square (bright) by applying $V_{dc} = - 5 V$ . (c) Local piezoresponse phase and amplitude (d) as a function of $V_{dc}$ (arrows and labels indicate the $V_{dc}$ cycling sequence). Black and blue symbols correspond to two subsequent cycles. Similar loops were observed upon repeated cycling, and also for different positions of the tip. (e) Cartoon of the proposed remnant ferroelectric domain structure after application of a large negative/positive voltage. This allows for an understanding the piezoresponse amplitude contrast behavior observed in (b) and (d).
Reuse & Permissions
Figure 5
(a) Cartoon of the field-effect model at the ferroelectric/superconductor interface. When the ferroelectric polarization $P$ points away from the YBCO (accumulated state), the charge (hole) accumulation leads to a carrier density increase over some length scale, locally rising $T_{c}$ . The native carrier density (and $T_{c})$ are recovered further from the interface. This is modelled by virtually separating the superconductor into two regions of thickness, $t_{1}$ and $t_{2}$ , the former having an average carrier density $n_{1}$ higher than the latter $n_{2}$ . The $T_{c}$ of the superconductor, as probed by electron-transport measurements in the direction parallel to the interface, corresponds to that of the doped region $T_{c 1}$ . On the contrary, upon polarization reversal (depleted state), the measured $T_{c}$ corresponds to that of the region far from the interface $T_{c 2}$ which is the highest because the depleted $n_{1} ≪ n_{2}$ . (b) Inverse of the Hall coefficient as a function of the total thickness of the superconductor $t_{YBCO} = t_{1} + t_{2}$ , for the doped (red) and depleted (blue) states. The symbols (squares/circles) are experimental data. The solid lines correspond to model calculations for the case in which the carrier mobility μ is uniform across layers 1 and 2. The dashed curves correspond to calculations for the case in which the mobility close to the interface $μ_{1}$ varies upon ferroelectric switching. The horizontal lines mark $n_{1}$ (red/blue respectively for the doped/depleted state) and $n_{2}$ (black, independent of the polarization state). (c) Variation between the depleted and accumulated states for the same data and calculations as in (b). (d) Critical temperature measured for the same samples as in (b). The lines are a guide to the eye.
Reuse & Permissions

Physical Review Materials