Chirality is a fascinating phenomenon that can manifest itself in subtle ways, for example in biochemistry (in the observed single-handedness of biomolecules) and in particle physics (in the charge-parity violation of electroweak interactions). In condensed matter, magnetic materials can also display single-handed, or homochiral, spin structures. This may be caused by the Dzyaloshinskii-Moriya interaction, which arises from spin-orbit scattering of electrons in an inversion-asymmetric crystal field. This effect is typically irrelevant in bulk metals as their crystals are inversion symmetric. However, low-dimensional systems lack structural inversion symmetry, so that homochiral spin structures may occur. Here we report the observation of magnetic order of a specific chirality in a single atomic layer of manganese on a tungsten (110) substrate. Spin-polarized scanning tunnelling microscopy reveals that adjacent spins are not perfectly antiferromagnetic but slightly canted, resulting in a spin spiral structure with a period of about 12 nm. We show by quantitative theory that this chiral order is caused by the Dzyaloshinskii-Moriya interaction and leads to a left-rotating spin cycloid. Our findings confirm the significance of this interaction for magnets in reduced dimensions. Chirality in nanoscale magnets may play a crucial role in spintronic devices, where the spin rather than the charge of an electron is used for data transmission and manipulation. For instance, a spin-polarized current flowing through chiral magnetic structures will exert a spin-torque on the magnetic structure, causing a variety of excitations or manipulations of the magnetization and giving rise to microwave emission, magnetization switching, or magnetic motors.

Single magnetic atoms on surfaces are the smallest conceivable units for two-dimensional magnetic data storage. Previous experiments on such systems have investigated magnetization curves, the many-body Kondo effect and magnetic excitations in quantum spin systems, but a stable magnetization has not yet been detected for an atom on a non-magnetic surface in the absence of a magnetic field. The spin direction of a single magnetic atom can be fixed by coupling it to an underlying magnetic substrate via the exchange interaction, but it is then difficult to differentiate between the magnetism of the atom and the surface. Here, we take advantage of the orbital symmetry of the spin-polarized density of states of single cobalt atoms to unambiguously determine their spin direction in real space using a combination of spin-resolved scanning tunnelling microscopy experiments and ab initio calculations. By laterally moving atoms on our non-collinear magnetic template, the spin direction can also be controlled while maintaining magnetic sensitivity, thereby providing an approach for constructing and characterizing artificial atomic-scale magnetic structures.

Using spin-polarized scanning tunneling microscopy we show that the magnetic order of 1 monolayer Mn on W(001) is a spin spiral propagating along h110i crystallographic directions. The spiral arises on the atomic scale with a period of about 2.2 nm, equivalent to only 10 atomic rows. Ab initio calculations identify the spin spiral as a left-handed cycloid stabilized by the Dzyaloshinskii-Moriya interaction, imposed by spin-orbit coupling, in the presence of softened ferromagnetic exchange coupling. Monte Carlo simulations explain the formation of a nanoscale labyrinth pattern, originating from the coexistence of the two possible rotational domains, that is intrinsic to the system. DOI: 10.1103/PhysRevLett.101.027201 PACS numbers: 75.70.Ak, 68.37.Ef, 71.15.Mb Magnetism-based data storage technology relies on the fact that information, i.e., the magnetic state of the area representing the bits, is stable over time. From the microscopic point of view, the direction of the magnetic moments is stabilized mainly by the spin-orbit interaction, which couples the spin to the crystal lattice and is responsible for the occurrence of easy and hard magnetization axes. Based on this picture, the need for nanoscale magnetic devices called scientists to the quest for high magnetic anisotropy materials (see, e.g., Ref.[1]).However, with decreasing magnetic bit sizes the structural inversion asymmetry of interfaces and surfaces comes into play. A surprising consequence of this fact was recently demonstrated [2], namely, that this symmetry breaking in combination with spin-orbit coupling (SOC) leads to the Dzyaloshinskii-Moriya interaction (DMI) favoring noncollinear magnetic order [3,4]. Although the DMI, because of its relativistic nature, is usually expected to be negligible compared to the nonrelativistic exchange interaction, it is sufficiently strong to impose a nanoscale leftrotating cycloidal spin spiral (SS) on the otherwise antiferromagnetic Mn monolayer on W(110): adjacent moments slightly deviate from the collinear configuration by about 7 , resulting in a long-period SS [2].Many open issues of this new phenomenon remain. In noncentrosymmetric bulk materials the DMI is the origin of the weak ferromagnetism of parent antiferromagnets [4]. What are the consequences of the DMI in thin films possessing ferromagnetic exchange coupling, e.g., can it destabilize a ferromagnetic state? And what happens to the long-range magnetic order if SS's of different propagation directions are energetically degenerate due to symmetry? Can the DMI modify the magnetic order even on the atomic scale?As shown in this Letter, 1 ML Mn=W 001 is an ideal system to address these points: As opposed to the antiferromagnetic Mn=W 110 , this system exhibits strong ferromagnetic exchange interaction [5] and a fourfoldsymmetric square surface lattice. Spin-polarized scanning tunneling microscopy (SP-STM) measurements reveal a SS, i.e., magnetic moments rotating continuously from one atom to the next, propagating along h110i directions with a period ...

No abstract

We prove that the magnetic ground state of a single monolayer Fe on W(001) is c2 2 antiferromagnetic, i.e., a checkerboard arrangement of antiparallel magnetic moments. Real space images of this magnetic structure have been obtained with spin-polarized scanning tunneling microscopy. An out-ofplane easy magnetization axis is concluded from measurements in an external magnetic field. The magnetic ground state and anisotropy axis are explained based on first-principles calculations. DOI: 10.1103/PhysRevLett.94.087204 PACS numbers: 75.70.Ak, 68.37.Ef, 71.15.Mb The 3d transition metals bcc Fe, hcp Co, and fcc Ni are known as the prototypical ferromagnets (FM). Much excitement was raised as experiments [1,2] indicated that Fe becomes an antiferromagnet (AFM) when stabilized in the metastable fcc phase. These experiments opened the vista that a proper control of the type of magnetic order is possible, allowing, for example, to turn a ferromagnet into an antiferromagnet or into a spin glass. However, after 40 years of research, all attempts to stabilize antiferromagnetic Fe ended up in Fe phases with fairly complex magnetic structures [3][4][5]. Thus, controlling the magnetic order in solids remains a challenge to solid state physics.Low-dimensional systems offer new possibilities to tune interactions. In this Letter we propose the interface tuning of the exchange interaction as a new route to antiferromagnetism in low-dimensional systems. For one monolayer (ML) Fe on W(001) we provide a clear proof of collinear antiferromagnetic Fe. Although this system has been studied extensively in the past, experiments could show only that the Fe ML is not ferromagnetic above 100 K [6,7], while theoretical predictions are still controversial [8,9]. Employing spin-polarized scanning tunneling microscopy (SP-STM) we obtained atomic resolution images which show that a c2 2 AFM structure with an out-ofplane magnetization direction is the magnetic ground state. This is in clear contrast to the Fe ML on W(110), which is ferromagnetic with an easy axis in the film plane [10], giving evidence that the sign of the exchange interaction can be selected by the choice of the substrate orientation alone. These different experimental results are explained based on first-principles calculations.The lack of remanence of 1 ML Fe=W001 was first discovered by spin-resolved photoemission [6] and Kerr effect measurements [7] and later interpreted on the basis of density-functional theory in the local spin-density approximation (LSDA) as an AFM ground state [8]. In these calculations the AFM state was just 10 meV=Fe atom lower than the nonmagnetic state and surprisingly the FM state did not exist. More recent calculations based on the generalized gradient approximation (GGA), however, found a FM solution, but the AFM state was not considered [9]. Meanwhile, the experimental result, i.e., the absence of remanent magnetization, has been confirmed in numerous experiments [11][12][13][14]. To our knowledge, however, up to now no experimental technique wa...