Session W19: Invited Session: Spin Coupling and Kondo Screening in Individual Magnetic Spins

The Impact of the Local Environment on the Kondo Screening of a High-Spin Atom

Spin 1/2 Kondo systems have been investigated extensively in theory and in a variety of experimental geometries. However the magnetic atoms that give rise to the Kondo effect in metals often have a larger spin, which makes the properties of the system more complex. Using low-temperature scanning tunneling microscopy and spectroscopy, we explore the Kondo effect of individual high-spin magnetic atoms on small islands of the thin insulator copper nitride (Cu$_{2}$N) in Cu(100) surfaces. Using a combination of elastic spectroscopy to probe the local density of states features arising from the Kondo screening and inelastic tunneling spectroscopy to study the higher energy spin excitations, we determined the spin of the atom and explore its impact on the Kondo resonance [1]. We find that the local magnetic anisotropy plays a decisive role in the physics of Kondo screening. In addition, we find that the splitting of the Kondo peak matches the splitting of the underlying unscreened spin levels, and surprisingly does not show any evidence of a renormalization of energy scales even though large renormalizations have been predicted for lower spin system. In addition, we find remarkably large variations in the strength of both the Kondo screening and magnetic anisotropy for Co atoms on both small and large Cu$_{2}$N islands. In both cases, the anisotropy and Kondo screening are inversely related: the Kondo resonance weakens as the anisotropy increases. For small islands, the Kondo screening is strongest near the center of the island, while for large island this trend is reversed. We examine the possible origins of this phenomenon, including variations in the physical and electronic structure of the Cu$_{2}$N surface. \\[4pt] [1] AF Otte et al., Nature Physics 4, 847 (2008).   

Mechanical Control of Spin States in Spin-1 Molecules and the Underscreened Kondo Effect

The ability to make electrical contact to single molecules creates opportunities to examine fundamental processes governing electron flow on the smallest possible length scales. We report experiments in which we controllably stretched individual molecules with spin S = 1/2 and S = 1 while simultaneously measuring Kondo-assisted current flow through the molecule as a probe of its spin states. For molecules with S = 1/2, the temperature dependence of the Kondo signal is in excellent agreement with the predicted universal scaling curves for the S = 1/2 Kondo effect and the molecular spin states exhibit no energy splitting with stretching, consistent with Kramers' theorem. However, for cobalt complexes we observe temperature scaling curves that are very different from the S = 1/2 Kondo predictions, and instead are in quantitative agreement with the predictions of the underscreened Kondo model for S = 1, in which conduction electrons only partially compensate the molecular spin. This allows us to identify the spin state as S = 1. As a function of stretching, the S = 1 molecules exhibit energy splittings of the spin states in the absence of magnetic field due to magnetic anisotropy arising from modification of the molecular symmetry. These findings demonstrate a mechanism of spin control in single-molecule devices and establish that they can serve as model systems for making precision tests of correlated-electron theories.   

STEM in Kondo Lattices: a new window on correlated electron materials

The tremendous developments in scanning tunneling electron spectroscopy over the past decade, applied with tremendous success to the cuprate superconductors, are now beginning to be applied to other strongly correlated electron systems. One area where they offer tremendous potential, is in the context of heavy fermion materials. In the last few years, it has become possible to start probing the physics of the Kondo lattice using STEM methods. In this talk I will review this field, discussing the physics of tunneling into the Kondo lattice, showing how tunneling involves a co-operative process of electron transfer and spin-flip, called ``cotunnelling'' [1,2]. I will provide an overview of latest results in this field, especially URu2Si2 [3,4], YbRh2Si2 [5] and CeCoIn5 [6], discussing how STEM can be used to probe various new theoretical proposals [7,8] for the exotic order and critical behavior. \\[4pt] [1] M. Maltseva, M. Dzero, and P. Coleman, Phys. Rev. Lett. 103, 206402 (2009).\\[0pt] [2] J. Figgins and D. Morr, Phys. Rev. Lett. 104, 187202 (2010).\\[0pt] [3] A. R. Schmidt et al, Nature 465, 570-576 (2010).\\[0pt] [4] P. Aynajian et al., Proc. Natl. Acad. Sci. U.S.A. 107, 10383 (2010).\\[0pt] [5] S. Ernst et al, Nature (2011).\\[0pt] [6] S. Ernst et al, Physica Status Solidi 247, 624 (2010).\\[0pt] [7] Y. Dubi and A.V. Balatsky, Phys. Rev. Lett. 106, 196407 (2011).\\[0pt] [8] P. Chandra, P. Coleman and R. Flint, to be published (2012).   

Pseudo-spin Resolved Transport Spectroscopy of the Double Dot Kondo Effect

The Kondo effect is a paradigmatic example of a highly correlated many-body state, describing how conduction electrons screen a localized spin via spin-flip scattering. Experimentally probing the spin physics of the Kondo effect is challenging, as spin-resolved measurements require either large magnetic fields that break the spin degeneracy of the localized site or ferromagnetic contacts that give the spin states different tunnel rates to the localized site. We demonstrate how the desired spin resolution can be achieved for arbitrary tunnel rates and fields in a double quantum dot system. A quantum dot consists of a confined droplet of electrons connected by tunnel barriers to conducting leads. We study two dots that are capacitively coupled with negligible inter-dot tunneling. In this system, one can have a degeneracy associated with an electron being in either dot 1 or dot 2, and this degeneracy acts as a pseudo-spin degree of freedom that gives rise to Kondo screening. We present transport spectroscopy measurements that show the zero-bias peak that is the hallmark of the Kondo effect. We use gate voltages to break the degeneracy between the pseudo-spin states, creating a pseudo-magnetic field. We show that spectroscopy measurements in this regime are analogous to measurements of the spin Kondo effect in a magnetic field. Finally, we demonstrate how measuring transport through each dot individually as a function of the bias voltage on the measured dot gives us the ability to perform pseudo-spin resolved measurements that are difficult to achieve in spin systems.   

Evolution of Kondo Resonance from a Single Impurity Molecule to the Two-Dimensional Kondo Lattice

We investigated the Kondo resonance formed by the adsorption of magnetic molecule, iron(II) phthalocyanine (denoted as FePc), on Au(111) from a single impurity regime to the two-dimensional Kondo lattice by using scanning tunneling microscopy, the density functional theory (DFT) and numerical renormalization group (NRG). In the single molecule regime, FePc takes ontop and bridge configurations. These species show characteristic Kondo signatures in their one-particle energy spectra depending on the adsorption site. The ontop species shows a broad peak accompanied by a sharp dip, while the bridge species only a broad peak. The origin of these features comes from the difference in the local symmetry around the Fe ion. The two-stage Kondo screening occurs for two localized electrons in the different 3d orbitals characterized by low and high Kondo temperatures, reflecting the different coupling strength of these orbitals with Au(111). For the ontop species, highly-symmetric SU(4) Kondo effect is realized, leading to the sharp dip. The dip observed for the ontop species is also evolved from the single impurity regime to the two-dimensional lattice. The spectral evolution and the quantum phase of the lattice are discussed by the competition of the Kondo effect with antiferromagnetic RKKY coupling between the molecular spins.