Atomic-scale probe of molecular magneto-electric coupling

TL;DR

The study demonstrates an atomic-scale probe of molecular magneto-electric coupling at a two-dimensional ferroelectric interface by placing FePc sensor spins on bilayer SnTe and interrogating them with scanning tunneling microscopy and inelastic tunneling spectroscopy. FePc LUMO energies provide a nanometer-scale marker of ferroelectric domain structure, while spin-excitation energies track domain-induced variations in magneto-crystalline anisotropy, revealing a structural coupling mechanism. A combination of molecular manipulation across domain boundaries and single-molecule tests confirms the effect is rooted in lattice distortions from domain stacking with the substrate, not inter-molecular interactions alone. The results offer a proof-of-concept for designing two-dimensional heterogeneous multiferroics by engineering MCA via controlled lattice distortions and domain structure.

Abstract

Van der Waals heterostructures are a core tool in quantum material design. The recent addition of monolayer ferroelectrics expands the possibilities of designer materials. Ferroelectric domains can be manipulated using electric fields, thus opening a route for external control over material properties. In this paper we explore the possibility of engineering magneto-electric coupling in ferroelectric heterostructures by studying the interface of bilayer SnTe with iron phthalocyanine molecules as a model system. The molecules act as sensor spins, allowing us to sample the magneto-electric coupling with nanometer precision through scanning tunneling microscopy. Our measurements uncover a structural, and therefore material-independent and intrinsic, mechanism to couple electric and magnetic degrees of freedom at the nanoscale.

Figures (9)

• Figure 1: (a) Typical sample with ML and BL islands of SnTe on HOPG. Agglomerations of FePc decorate the SnTe (1 V, 12 pA). (b) Example of a ferroelectric grain boundary in BL SnTe. The 90$^{\circ}$ degree lattice rotation leads to stark differences in the observed moiré patterns in the adjacent domains (1 V, 24 pA). (c) Representative conductance spectrum acquired on BL SnTe. The conduction band onset lies at ca +1 V, the valence band onset at ca. -0.23 V. The valence band maximum (VBM) is found at much lower energies, ca. -0.8 V. (d) Waterfall plot tracking the energy of the VBM across a grain boundary. Owing to variations in lattice strain, the VBM settles to different values in either domain.
• Figure 2: (a) STM image of an FePc island on BL SnTe. The molecules self-assemble into islands with a square lattice geometry (1 V, 50 pA). (b) Representative conductance spectrum acquired above the FePc centre ion showing the LUMO resonance with a maximum at ca. 460 mV. The LUMO maximum can be used to track the SnTe grain boundaries beneath the FePc island. (c) Representative conductance spectrum acquired above the centre ion of an FePc molecule showing the symmetric step-like increases of inelastic spin-flip excitations, encoding the spin properties of the molecules.
• Figure 3: (a) Spatial map of the LUMO maxima of the molecules inside the red contour in Fig. \ref{['fig:Figure2']}a. Each circle represents one FePc molecule. The location of the grain boundary is clearly visible as an increase in the LUMO energy. (b) Spatial map of the lowest spin excitation energy as extracted from IETS spectra of the molecules inside the red contour in Fig. \ref{['fig:Figure2']}a. Each circle represents one FePc molecule. The domain contrast of SnTe is clearly represented in the spin excitation energies. Grey circles represent outliers with abnormally low spin excitation energies. (c) Composite STM image showing some steps of the molecular manipulation of an FePc molecule across the domain boundary next to the FePc island in Fig. \ref{['fig:Figure2']}a. Red arrows indicate the the manipulation sequence. Not all steps are shown. (d) LUMO maxima (blue) and lowest spin ecitation energies (orange) extracted at each point of the manipulation sequence. The origin of the $x$-axis is centered on the domain wall. The patterns observed in panels (a) and (b) for the FePc island also hold for single molecules.
• Figure S1: Atomically resolved STM image of several SnTe domains separated by a grain boundary. The lines mark the locations of the extracted line profiles on the right hand side.
• Figure S2: Atomically resolved STM image of several SnTe domains separated by a grain boundary. The lines mark the locations of the extracted line profiles on the right hand side.