Extended abstract of a paper presented at Microscopy and Microanalysis 2010 in Portland, Oregon, USA, August 1 – August 5, 2010.

Interfaces between perovskite oxides provide one of the most exciting test beds for stabilizing exotic physical phases not present in bulk materials. The emergence of metallic conductivity, and low temperature superconductivity, between the two insulators LaAlO 3 (LAO) and SrTiO 3 (STO) is of particular interest; this conducting interface can be grown on silicon, providing a pathway to integration into modern electronic devices [1]. We have recently found the interface conducts when the La/Al ratio is at or below 0.97 ± 0.03 while samples that exceed this limit are insulating [2]. While density functional theory models have explained the conductivity of a simple stoichiometric LAO film with an electronic reconstruction at the interface, fully incorporating off-stoichiometry into the model requires a precise understanding of atomic positions, chemical intermixing and electronic structure as function of La-to-Al ratio. Here, we collect spectroscopic maps on a series of LAO/STO interfaces over a stoichiometry range to correlate the electronic structure information to the transport properties. Scanning Transmission Electron Microscopy (STEM) in combination with Electron Energy Loss Spectroscopy (EELS) allows for spatially resolved, chemically sensitive investigations of oxide interfaces [3]. We used a Nion 5th-max = 33 mrad) to collect spectroscopic images of four LAO / STO interfaces with La/Al ratios ranging from 0.91 0.02 (conducting) to 1.06 0.02 (insulating). With both high spatial resolution (~1Å) and high usable collected beam current (nearly 300 pA), the two dimensional distribution of all atomic species-La, Al, O, Sr and Ti-can be extracted in minutes. By directly comparing the spectroscopic maps of both insulating and conducting samples, we are uniquely poised to distill the microscopic characteristics that correlate with transport properties from those that do not. Notably, neither the conducting nor the insulating samples showed atomically abrupt interfaces, with interdiffusion over a few unit cells at the interface; the most diffuse interfaces showed a La concentration in the substrate dropping to < 1% by 5 u.c. and < 0.001% by 8 u.c. As shown in Fig. 1, there is no correlation between the degree of La interdiffusion at the interface and the transport properties. However, by collecting all atomic species present, we can determine the total concentration of the A-site cations, Sr and La, and the B-site cations, Ti and Al (Fig. 2) across the interface. We find a dip in the total B site cation concentration across the interface in the insulating samples not present in the conducting samples; neither series showed a dip in the total A-site cation concentration. This dip in the Bsite cation concentration in insulating films can mitigate the polar catastrophe at the interface without a corresponding electronic reconstruction, and hence without free carriers. [4] 1450

Extended abstract of a paper presented at Microscopy and Microanalysis 2010 in Portland, Oregon, USA, August 1 – August 5, 2010.

Direct measurements have been made of the effect of strain on the critical temperature of tin, aluminum, and indium films. The films (&lt;1000 Å) were vacuum deposited onto glass and mylar substrates at room temperature and were then strained by mechanically bending the substrate in liquid helium. Tc depended linearly on the strain over the range of observation, ∼10−3, for both tensile and compressive bending. The slopes of the Tc vs strain curves were 8.7° and 9.1°K per unit strain, respectively for tin and aluminum films on glass substrates and 4.9° and 3.8°K per unit strain for tin and indium films on mylar substrates. Earlier measurements1,2 have indicated considerable variation in the Tc dependence on strain from film to film, but our data are quite reproducible and do not seem to depend markedly on the evaporation conditions. For an isotropic material or an anisotropic material for which the relative orientation of the strain axis and the axes of crystal symmetry are known it is possible to relate the change in Tc to a corresponding volume change. On this basis one can compare the results on thin films with strain and pressure measurements on bulk samples. Comparing d ln Tc/d ln V for thin films and bulk samples3 we find that the values for aluminum agree within five percent, those for tin films on glass are lower by thirty percent, and those for indium films are higher by a factor of three. The change in Tc due solely to a change in sample volume can be calculated as a function of the Debye temperature on the basis of the McMillan theory1 by using the Gruneisen relation and taking λ, the phonon-mediated electron interaction, to be inversely proportional to the square of the Debye temperature. Comparing experimental results with these calculations we find that the experimental results for dTc/dθD lie within the limits imposed by μ* = 0.1 to 0.2 for bulk aluminum, tin, and indium and thin-film aluminum and tin. (Materials for which the Coulomb pseudopotential, μ*, has been measured lie within these limits.)

Strain and dimensional confinement can be used to tune magnetic and ferroelectric properties or enhance device performance. In epitaxial films, the different lattice spacing of an underlying substrate can be used to impose a biaxial strain. We have studied the effect of the substrate-induced biaxial strain on the electrical conductivity and magnetic properties of manganite compounds. Epitaxial (001) La0.7Sr0.3MnO3 thin films have been grown by reactive molecular-beam epitaxy (MBE) on single crystalline substrates, varying the substrate-induced biaxial strain from −2.3% to +3.2%. The strain caused the Curie temperature to decrease by up to 100 K in very good agreement with the predictions of Millis et al. [A. J. Millis, T. Darling, and A. Migliori, J. Appl. Phys. 83, 1588 (1998)]. The effect of dimensional confinement on manganites was also investigated in thin films by synthesizing a superlattice of two formula-unit-thick layers of CaMnO3 separated by CaO double layers, i.e., the n=2 Ruddlesden-Popper phase Ca3Mn2O7. Magnetization measurements on 30 nm thick (001)-oriented Ca3Mn2O7 thin films grown on (110) YAlO3 substrates by MBE reveal a Neél transition temperature of TN=115 K, similar to bulk Ca3Mn2O7, but 10 K lower than thick CaMnO3 films grown on this same substrate.