Hard x-ray photoemission and density functional theory study of the internal electric field in SrTiO${}_{3}$/LaAlO${}_{3}$ oxide heterostructures

Hard x-ray photoemission and density functional theory study of the internal electric field in SrTiO $_{3}$ /LaAlO $_{3}$ oxide heterostructures

E. Slooten, Zhicheng Zhong, H. J. A. Molegraaf, P. D. Eerkes, S. de Jong, F. Massee, E. van Heumen, M. K. Kruize, S. Wenderich, J. E. Kleibeuker, M. Gorgoi, H. Hilgenkamp, A. Brinkman, M. Huijben, G. Rijnders, D. H. A. Blank, G. Koster, P. J. Kelly, and M. S. Golden

Phys. Rev. B 87, 085128 – Published 25 February 2013

Abstract

A combined experimental and theoretical investigation of the electronic structure of the archetypal oxide heterointerface system LaAlO $_{3}$ on SrTiO $_{3}$ is presented. High-resolution, hard x-ray photoemission is used to uncover the occupation of Ti $3 d$ states and the relative energetic alignment—and hence internal electric fields—within the LaAlO $_{3}$ layer. First, the Ti $2 p$ core-level spectra clearly show occupation of Ti $3 d$ states already for two unit cells of LaAlO $_{3}$ . Second, the LaAlO $_{3}$ core levels were seen to shift to lower binding energy as the LaAlO $_{3}$ overlayer thickness, $n$ , was increased, agreeing with the expectations from the canonical electron transfer model for the emergence of conductivity at the interface. However, not only is the energy offset of only $\sim 300$ meV between $n = 2$ (insulating interface) and $n = 6$ (metallic interface) an order of magnitude smaller than the simple expectation, but it is also clearly not the sum of a series of unit-cell-by-unit-cell shifts within the LaAlO $_{3}$ block. Both of these facts argue against the simple charge-transfer picture involving a cumulative shift of the LaAlO $_{3}$ valence bands above the SrTiO $_{3}$ conduction bands, resulting in charge transfer only for $n \geq 4$ . We discuss effects which could frustrate this elegant and simple charge-transfer model, concluding that although it cannot be ruled out, photodoping by the x-ray beam is unlikely to be the cause of the observed behavior. Turning to the theoretical data, our density functional simulations show that the presence of oxygen vacancies at the LaAlO $_{3}$ surface at the $25 %$ level reverses the direction of the internal field in the LaAlO $_{3}$ . Therefore, taking the experimental and theoretical results together, a consistent picture emerges for real-life samples in which nature does not wait until $n = 4$ and already for $n = 2$ mechanisms other than internal-electric-field-driven electron transfer from idealized LaAlO $_{3}$ to near-interfacial states in the SrTiO $_{3}$ substrate are active in heading off the incipient polarization catastrophe that drives the physics in these systems.

Received 20 November 2012

DOI:https://doi.org/10.1103/PhysRevB.87.085128

Authors & Affiliations

E. Slooten^1,*, Zhicheng Zhong², H. J. A. Molegraaf², P. D. Eerkes², S. de Jong^1,†, F. Massee^1,‡, E. van Heumen¹, M. K. Kruize², S. Wenderich², J. E. Kleibeuker^2,§, M. Gorgoi³, H. Hilgenkamp², A. Brinkman², M. Huijben², G. Rijnders², D. H. A. Blank², G. Koster², P. J. Kelly², and M. S. Golden¹

¹Van der Waals Zeeman Institute, University of Amsterdam, Science Park 904, 1098 XH Amsterdam, The Netherlands
²Faculty of Science and Technology and MESA+ Institute for Nanotechonology, University of Twente, P.O. Box 217, 7500 AE Enschede, The Netherlands
³Helmholtz Zentrum Berlin für Materialien und Energie GmbH, Albert-Einstein-Strasse 15 12489 Berlin, Germany

^*e.slooten@uva.nl
^†Current affiliation: SLAC National Accelerator Laboratory, 2575 Sand Hill Road, Menlo Park, California 94025-7015, USA.
^‡Current affiliation: Laboratory of Solid State Physics, Department of Physics, Cornell University, Ithaca, New York 14853, USA.
^§Current affiliation: Department of Materials Science and Metallurgy, University of Cambridge, Pembroke Street, Cambridge, CB2 3QZ, United Kingdom.

Article Text (Subscription Required)

Click to Expand

References (Subscription Required)

Click to Expand

Issue

Vol. 87, Iss. 8 — 15 February 2013

Reuse & Permissions

Access Options

Author publication services for translation and copyediting assistance advertisement

Images

Figure 1
(a) The Ti $2 p$ spectrum for $n = 2$ and $n = 6$ . As a reference the spectrum for pure STO ( $n = 0$ ) is also included. The shoulder (inset) at 457 eV shows the existence of Ti $^{3 +}$ ( $3 d^{1}$ entities) for samples both above and below the critical thickness of four layers of LAO. The $n = 2$ and the $n = 6$ spectra have been shifted to lower binding energy (160 and 220 meV, respectively) in order to align the main line easing comparison between the shoulders. (b) The Ti $2 p$ spectrum for the $n = 6$ sample as a function of emission angle $θ$ between 5 $^{\circ}$ and 50 $^{\circ}$ away from normal emission (NE). Within the noise the shoulder (inset) shows marginal or no angular dependence. (c) A simulation—based on the model of Ref. 29—of the expected Ti $^{3 +}$ shoulder intensity for several degrees of spatial confinement of the q-2DEG for an electron emission angle of 50 $^{\circ}$ from normal emission (NE).Reuse & Permissions
Figure 2
(a) La $4 d$ and Sr $3 d$ core-level photoemission spectra for selected STO/(LAO) $_{n}$ samples. The Sr $3 d$ data were measured with $h ν = 2150$ eV; the La $4 d$ data were measured with $h ν = 3000$ eV. The inset shows a clear shift of the La $4 d$ maximum close to 103 eV to lower binding energies with increasing LAO layer thickness, whereas the Sr $3 d$ spectra essentially do not shift in energy between $n = 2$ and $n = 6$ . (b) A simulated $n = 3$ spectrum for the La $4 d$ core levels which are shifted by 0.9 eV per LAO UC. The colored, dashed spectra are the layer-resolved spectra for the first (interface side, red), second (orange), and third (surface side, green) LAO UC, whose $λ$ -weighted sum then gives the total (black) spectrum. (c) The La $4 d_{5 / 2}$ peak FWHM as a function of LAO layer thickness.Reuse & Permissions
Figure 3
Evolution of the energetics in the LaAlO $_{3}$ overlayer tracked using the La $4 d$ and Al $2 s$ core-level binding energies vs LaAlO $_{3}$ thickness, $n$ . (a) Data up to $n = 6$ . To indicate the spread in energy shifts encountered, in addition to the main data points (see legend), the smaller gray symbols show energy-shift data from all experiments we conducted on STO/LAO interface samples, including using Al $K_{α}$ radiation, and using samples from other thin film growth batches. For comparison with experiment, one-fourth of the expected core-level shift based on DFT calculations for a vacancy-free interface is also shown as an average weighted by the inelastic mean-free path length in HAXPES. The results of these calculations are also shown in Table . (b) Energy shift data including $n = 20$ .Reuse & Permissions
Figure 4
(a) The Ti $2 p$ spectrum for the low-quality substrate sample compared to the $n = 5$ sample from the high-quality series. The sample with the lower quality substrate shows a Ti $^{3 +}$ shoulder which is clearly much bigger than that of the high-quality sample. The spectrum of the low-quality substrate was shifted in energy to coincide with that of the high-quality film. (b) The Ti $2 p$ spectrum for the low-quality substrate, $n = 5$ sample at different excitation energies between 1486 eV and 6 keV. These spectra have been normalized on the main Ti $2 p_{3 / 2}$ peak. The intensity of the Ti $^{3 +}$ (3 $d^{1}$ ) shoulder shows a clear kinetic energy dependence, in turn related to the inelastic mean-free path length, indicating confinement of the q-2DEG over 2.5 nm, as shown in panel (c) in which the Ti $^{3 +}$ /Ti $^{4 +}$ intensity ratio is plotted as a function of $λ$ .Reuse & Permissions
Figure 5
(a) The $n$ dependence of the potential in the LaAlO $_{3}$ expressed via the relative changes in the La $5 s$ shallow core-level binding energy extracted from the DFT calculations for vacancy-free and oxygen-deficient LAO/STO systems described in the text. Points describing a positive slope reflect a trend to decreasing La core-level binding energy. (b) The $2 \times 2$ lateral supercell used in the DFT calculations. Oxygen vacancies ( $25 %$ ) were located in the outermost AlO $_{2}$ layers. (c) Layer-by-layer data from the DFT calculations for the system shown in panel (b). The right-hand panels show the $n$ dependence of the energetics in the STO substrate and the left-hand panels the evolution of the energies across the LaAlO $_{3}$ films. The yellow shading indicates regions with LAO potentials such that the La core levels move to higher binding energy, opposite to the direction of the in-built potential scenario.Reuse & Permissions

Physical Review B

covering condensed matter and materials physics