Mesosurfaces can be described as surfaces which exhibit at least two surface areas which are delineated from each other by a sharp boundary and wherein the first of the at least two areas exhibits a certain defined physical property which is distinguished from the one of the at least second area. Thus, a mesosurface exhibits a surface texture of delineated, distinguished physical properties on the meso-scale, which means that the delineated surface areas are in the “meso” range of between 10 nm ² to 1000 μm ² Atypical “physical property’’ for mesosurfaces is its magnetic property i.e. the emergent magnetism from a collection of oriented electron spins.

The development of magnetic textures on the meso-scale has applications ranging from spin ice to magnetic logic and neuromorphic computing (see: e.g. https://en.wikipedia.org/wiki/Spin-ice, https://en.wikipedia.org/wiki/Magnetic logic, https://en.wikipedia.org/wiki/Neuromor-phic engineering, or references 1-7). Magnetic conduits for the controlled propagation of spin waves or magnetic racetracks that underlie a novel, high performance non-volatile magnetic memory, are other examples of magnetic meso-structures (see e.g. references 8-11). Recently, some novel phenomena such as ultra-broadband coherent perfect absorption were demonstrated, where a 0.3 nm thick film could absorb all electromagnetic waves across the RF, microwave, and terahertz frequencies. This is another application, which may be accessed by mesosurfaces.

OBJECT OF THE INVENTION

For the fabrication of optical mesosurfaces, two main approaches were used so far (see references 12, 13): i) top-down methods which rely on photolithography or electron-beam lithography and following etching, deposition, and lift-off processes; ii) bottom-up approaches that include laser/ion-beam printing, or self-assembly methods. Thus, most reported mesosurfaces need an additive or subtractive fabrication process and have static functionalities that depend on their fixed geometrical parameters and cannot be changed after fabrication. As a consequence, these techniques lead to irreversible modifications of the mesosurface material thereby excluding dynamic adaptations/modifications of the mesosurface. Moreover, most of these methods are limited to be applied to planar surfaces only, thereby excluding its application to 3D objects.

It was, therefore, an object of the present invention to avoid the above-mentioned disadvantages and to provide for an easier (less process steps) and more flexible method of manufacture of mesosurfaces, allowing for smaller and more complex structures and for a dynamic modification of the mesosurface, which inter alia means that the surface modifications are reversible.

BRIEF DESCRIPTION OF THE INVENTION

This object is achieved with a method for making mesosurfaces comprising the steps of:

I. providing a substrate which exhibits at least two distinct measurable states of an electron spin property and which can stably but reversibly be transitioned from a. a first state of the electron spin property into at least b. one second state of the electron spin property which is measurably distinct from the first state,

II. creating on a surface of the substrate at least one area a. outside of which the substrate is in the first sate of the electron spin property, and b. inside of which the substrate is in one of the at least one second state of the electron spin property, or c. wherein the states of the electron spin property of the inside area and the outside area are inverted, and

III. wherein in step II a desired to be inside area is defined and delineated on the surface of the substrate which is in the first state of the electron spin property and wherein subsequently the surface in the defined and delineated inside area is contacted with a liquid which reacts with the substrate material to yield in or to bring about the transition into the at least one second state of the electron spin property within the inside area only, or wherein in step Ell the states of the electron spin property of the surface and the inside area are inverted, wherein

IV. the individual area of the at least one area has a size of <3 μm ² and/or wherein multiple areas form one array or more arrays of areas of regular or irregular shapes of any size or sizes.

The process according to the invention has the advantage that no subtractive or additive steps are needed in making the mesosurface i.e. the composition of the mesosurface is contained entirely within the layer that is transitioned by the liquid. The process, therefore, is cheaper, and the surfaces are smoother than, for example, those formed with photolithography. Moreover, the mesosurface is dynamic, i.e. the electron spin property can be reversed or dynamically tuned by the inventive process. Another unique advantage of the new process is that it allows mesosurfaces to be formed readily on curved or flexible surfaces. Through various resist masks with tunable orifices with critical dimensions of <1μm in diameter (or a size of < 3μm ² ) up into the low hundreds of nm range, a wide range of magnetic spatial textures can be created, including an array or arrays of regular or irregularly shaped areas, e.g. checkerboard arrangements of ferromagnetic and antiferromagnetic phases as well as checkerboard structures composed only of two distinct antiferromagnetic phases. The boundaries between these regions can be just a few nanometers wide. Exemplary functional devices can be formed that are composed of spin ice structures that exhibit magnetic frustration and an operational magnetic racetrack device in which domain walls in wires can be very efficiently moved with current. The formation of complex magnetic meso- structures from homogeneous layers without the need for additive or subtractive processing using local ionic liquid gating is highly promising for a wide variety of functional magnetic spatial textures and spintronic devices. It has applications ranging from spin ice to magnetic logic and neuromorphic computing. Magnetic conduits for the controlled propagation of spin waves or magnetic racetracks that underly a novel, high performance non-volatile magnetic memory, are other examples of magnetic meso-structures.

DEFINITION OF TERMS

Within the meaning of the present invention the following terms in parenthesis, whether used in singular or plural form shall have the following meaning:

“Mesosurfaces” exhibit a surface texture of delineated, distinguished physical properties on the meso-scale, which means that the delineated surface areas are in the “meso” range of between 10 nm ² to 1000 μm ² . Mesosurfaces are to be distinguished from “metasurfaces” which are formed from periodic arrangements of matter, extended over a length scale that is typically several periods, whereas a mesosurface can be and typically are formed from an aperiodic arrangement of matter. The mesosurface can have a thickness of from a few nm to hundreds of nm.

“Substrate” or “phase change material” is a material, which undergoes a change in an electron spin property, such as a change in magnetic property. This change may be a change in the type of magnetic property selected from a first state of one of ferromagnetic, ferrimagnetic, anti-ferromagnetic, paramagnetic and diamagnetic property to a second state which is different from the first state and which is also selected from one of ferromagnetic, ferrimagnetic, anti-ferromagnetic, paramagnetic and diamagnetic property. The change may also be a change in degree of the magnetic property or a combination of both type and degree. The states may be chiral or achiral. The first and second state may be distinguished by their distinct chirality (e.g. clockwise or anti-clockwise) or these states may be distinguished by their domain structure, for example, the orientation of the magnetic domain along a certain direction. Thus, the substrate exhibits at least two distinct states of an electron spin property, which are all stable at ambient conditions (i.e. between 10 and 30 °C; between 950 and 1100 hPa). The transition from one state of electron spin property to another is brought about by a change in the chemical composition of the substrate material, e.g. by its stoichiometric oxygen content. This transition may be instigated by applying an electric potential to a liquid in contact with the substrate.

One prime substrate is SrCoO ₃ . SrCoO ₃ crystallizes as metallic ferromagnetic (FM) with a perovskite structure and exhibits two insulating antiferromagnetic (AF) counterparts, which exhibit a brownmillerite structure (formula: SrCoO _2.5 ). Thin layers of these compounds can be completely and reversibly transformed between each other by the addition or subtraction of oxygen via ionic liquid (IL) gating (see references14-

16). This can take place through micron sized holes in a resist layer (see reference

17). The structure of the brownmillerite phase that is formed in this way consists of periodic layers of oxygen vacancies that are ordered either vertically at low gate voltage (V) or parallel to the substrate (P) above a threshold gate voltage that is applied to the IL during the gating process. These two oxygen vacancy ordered phases, V- and P-SrCoO _2.5 , have distinct antiferromagnetic structures and properties. Taking advantage of these dual states stabilized by distinct gate voltages a series of magnetic meso-structures composed of regions formed from either of these two AF structures or the FM parent perovskite can be created. The regions can be of any arbitrary shape.

Other substrate materials for the present invention, which undergo a change in an electron spin property with change in oxygen composition includes SrFeO ₃ , REFeO ₃ (RE= La, Nd, Gd...).

“Electron spin property” or “State of electron spin property" means a magnetic property selected from one of ferromagnetic, ferrimagnetic, anti-ferromagnetic, paramagnetic and diamagnetic property. The property may also include a quantification, i.e. a degree of the magnetic property. The electron spin property can easily be determined and quantified using typical commercial measuring instruments.

Two otherwise identical substrate materials are in a different state of an electron spin property when they are measurably distinct from each other in this electron spin property. Measurably distinct means that the electron spin property varies by a statistically significant amount or factor, which is typical for the state transition in question.

“Transition” means that the material in question can be converted from one first state of electron spin property into at least one other second state of electron spin property, which is measurably distinct from the first state. The means for such a transition may be a change in pressure and/or temperature, the application of an electrical field, a chemical reaction etc.

"Stable” means that the state of electron spin property remains unchanged under ambient conditions for a minimum amount of time. Ambient conditions means between 10 and 30 °C and between 950 and 1100 hPa. A minimum amount of time means at least 100 days, preferably at least 1000 days, more preferably at least 10000 days.

"Reversibly" means that the transition from one first state of electron spin property into another second state of electron spin property can be reversed from the second state back into the first state.

An “area” created on a surface of the substrate is a part of the surface which is delineated from another part of the surface by a boundary of infinitesimal width. The area may form a closed area having a perimeter of infinitesimal width, which delineates the inside of the area from the outside of the area. Infinitesimal width means 1 -100 nm, preferably 2-50 nm, more preferred 5-10 nm. The total of the delineated surface areas are in the "meso” range of between 10 nm ² to 1000 μm ² . The critical dimension of an individual area is from <1 μm in diameter into the low hundreds nm range. Also, the area can be construed of a wide range of magnetic spatial textures of deliberate size, including an array or arrays of regular or irregularly shaped areas, e.g. checkerboard arrangements of ferromagnetic and antiferromagnetic phases as well as checkerboard structures composed only of two distinct antiferromagnetic phases, or rectangularly shaped regions that can extend to wires.

"Array” means an aggregation of two or more identical or at least nearly identical shapes. “yield in/bring about” the transition means that the applied liquid reacts with the substrate in a way that changes the chemical composition of the substrate material (=chemical reaction) so that the resulting chemical product in the area of contact is transitioned from a first state of electron spin property into another second state of electron spin property, which is measurably distinct from the first state. In an alternative embodiment or simultaneous to the chemical reaction a liquid can be used as an electrical conducting fluid allowing and/or aiding in the transport of electrons and/or ions to or from the substrate in the area of contact and/or to apply an electrical potential.

“Liquid” means a fluid medium, e.g. water, an alcohol like ethanol, methanol etc., a hydrocarbon like pentane, hexane etc., a halogen containing hydrocarbon like chloroform etc., which reacts with the substrate material to yield in or to bring about the transition into the at least one second state of electron spin property within the exposed inside area(s) only. In an alternative embodiment, the liquid can be an electrical conducting fluid allowing and/or aiding in the transport of electrons and/or ions to or from the substrate in the exposed area(s) of contact and/or to apply an electrical potential. One preferred liquid is an ionic liquid, which is an organic or inorganic liquid, which contains or consists of ions. Examples are aqueous solutions of organic or inorganic water soluble salts, e.g. NaCI, KOI, etc. or organic ionic liquids like 1-Ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide (EMIM-TFSI), 1- Propyl-3-methyl-imidazolium bis(trifluoromethylsulfonyl)imide, 1 -Butyl-3- methylimidazolium bis(tri-fluoromethylsulfonyl)imide or 1-Hexyl-3-methylimidazolium bis(trifluoromethyl-sulfonyl)imide. These imidazolinium compounds are room temperature ionic liquids (RTIL). They are non-aqueous solvents, which are advantageous over traditional aprotic polar organic solvents in electrochemical reactions. They have a low vapor pressure, high thermal stability and good electrical conductivity. The liquid may have a viscosity from a freely flowing liquid to a highly viscous gel but a common property is the ability to accept or provide oxygen to the substrate. DETAILED DESCRIPTION OF THE INVENTION

The process according to the invention transitions a surface of a substrate into a mesosurface. The process is able to create a wide range of magnetic spatial textures, including checkerboard arrangements of ferromagnetic and antiferromagnetic phases as well as checkerboard structures composed only of two distinct antiferromagnetic phases or stripes of ferromagnetic and antiferromagnetic phases for magnetic racetrack memory devices. Liquid gating, preferably ionic liquid gating (ILG), through a resist mask with holes, is used on a layer of a substrate (a phase change material) in order to locally create regions of an altered state of an electron spin property. The created regions may have various dimensions within the remaining areas of un-altered state of the electron spin property. The liquid, preferably the ionic liquid brings about a phase change, i.e. a change in an electron spin property in the local regions under the holes (see Fig. 7a and b). This approach is highly flexible, as it does not require physical (subtractive) patterning of the substrate material. This approach opens a way towards electrically reconfigurable mesomaterials.

The substrate itself is a two dimensional material of negligible thickness (a thin film). The thickness is typically in the range of 1-1000 nm, preferably 5-500 nm, more preferably 5-100 nm. The planar size of the substrate is typically in the range of 1 -1000 mm ² , preferably 10-500 mm ² , more preferably 50-299 mm ² . Techniques such as chemical solution deposition (CSD), spin coating, chemical vapor deposition (CVD), plasma enhanced CVD, atomic layer deposition (ALD), molecular layer deposition (MLD), electron beam evaporation, molecular beam epitaxy (MBE), sputtering, pulsed laser deposition, cathodic arc deposition (arc-PVD) or electrohydrodynamic deposition can be used to manufacture the ultrathin films of substrate material. It is surprising that the films can even be up to about 1 μm thick and still the liquid gating, specifically ILG, at the film surface results in a dramatic change of the electron spin property in the entire thickness of the layer. This is a result of the diffusion of ions from within the interior of the film controlled by ILG. When the ionic diffusion occurs preferentially along certain crystallographic directions the cross-sectional area of the modified material is defined by the "hole” in the resist layer (the “mask”). Accordingly, it may be desirable for the transition from one state of electron spin property into another of the substrate to create the substrate surface in a preferred, predetermined direction of a crystallographic face expressed in terms of h,k, l-indices, e.g. the [100] direction, or the [001] direction, or the [111] direction etc.. This may be advantageous if e.g. the mobility of certain atoms in a certain direction of the unit cell ([h ,k, I]) of the substrate material is higher/lower than in another direction.

Such a growth of the thin film in a predetermined direction can be facilitated by epitaxy, where the substrate of interest is grown on an “inducing” surface of a second material which exhibits a physical surface which matches the desired direction [h,k, I] of the to be grown substrate and which exhibits a crystallographic structure and unit cell dimensions which are at least similar, if not close to identical (so called “lattice matching”; i.e. unit cell axis and/or angle deviation of max. 10%, preferably max. 5%, more preferred max. 3%) to the one of the to be grown substrate material. Such epitaxy methods are known to the skilled person. It is also possible to form films that are predominantly of one crystal texture but which are not crystallographically oriented with respect to the substrate. This is often referred to as polycrystalline epitaxy. Polycrystalline films with one preferred crystal orientation perpendicular to the plane of the film can be formed by use of underlayers that themselves grow with a preferred crystalline orientation and that thereby induce the epitaxial growth of the thin film of interest.

In the next step, the defined and delineated inside areas are formed on the substrate surface. The shapes of the delineated areas can take any desired form, which is suitable for the kind of mesosurface to be manufactured. E.g., they can be arranged as checkerboards or stripes. The shapes are designed and arranged on the surface of the substrate material. The predesigned delineated area shapes on the surface in which the state of electron spin property shall be transitioned from its initial state to a measurably distinct at least second state can be created e.g. by applying lithographic techniques, i.e. by creating a structured mask which covers the to be modified surface of the substrate. Typical lithographic resists, suitable for exposure using light, electron beams or heat, can be used to cover the surface of the substrate in a thin layer (thickness: 100-1000 nm, preferably 150-750 nm, more preferred 200-400 nm). Subsequently, the desired forms of the to be transitioned areas are cut out of the resist cover layer, thereby exposing the substrate surface in the cut out areas only. Cutting can be performed e.g. with precision electron beam lithography methods by use of modification of the chemical properties of the resist layer and their subsequent removal by dissolution by use of suitable chemical solvents. The resulting structures can have widths or radii that can be as small as < 500 nm, preferably < 100 nm, more preferred < 50 nm.

In the next step the state of electron spin property of the exposed area(s) of the substrate surface is (are) transitioned from its initial state of electron spin property to a measurably distinct at least second state of electron spin property. During this step the surface in the defined and delineated exposed inside area(s) is (are) contacted with a liquid which reacts with the substrate material to yield in or to bring about the transition into the at least one second state within the exposed inside area(s) only. This process is called “liquid gating” since the liquid forms ion “gates” in the exposed surface area(s). The applied liquid reacts with the substrate in a way that changes the chemical composition of the substrate material (=chemical reaction) so that the resulting chemical product in the area(s) of contact (the exposed area(s)) is (are) transitioned from a first state of electron spin property into another second state of electron spin property. In an alternative embodiment or simultaneous to the chemical reaction the liquid can be used as an electrical conducting fluid allowing and/or aiding in the transport of electrons and/or ions to or from the substrate in the exposed area(s) of contact and/or to apply an electrical potential. One preferred liquid is an ionic liquid, which is an organic or inorganic liquid, which contains or consists of ions. Examples are aqueous solutions of organic or inorganic water soluble salts, e.g. NaCI, KCI, etc. or organic ionic liquids like 1 -Ethyl- 3-methylimidazolium bis(trifluoromethylsulfonyl)imide (EMIM-TFSI), 1-Propyl-3-methyl-imidazolium bis(trifluoromethylsulfonyl)imide, 1-Butyl-3-methylimidazolium bis(tri-fluoromethylsul- fonyl)imide or 1-Hexyl-3-methylimidazolium bis(trifluoromethyl-sulfonyl)imide. These imidazolinium compounds are room temperature ionic liquids (RTIL). They are nonaqueous solvents, which are advantageous over traditional aprotic polar organic solvents in electrochemical reactions. They have a low vapor pressure, high thermal stability and good electrical conductivity. In one embodiment, where the exposed area(s) is(are) contacted with an ionic liquid the transition of state of electron spin property is preferably performed by simultaneously applying an electric voltage of 0.1-10 Volts, preferably 1-5 Volts, more preferred 2-3 Volts. In the case of applying an electric voltage, the voltage is applied against a counter electrode, which is also in contact with the ionic liquid. A preferred material for the counter electrode is a metal, more preferred a metal like Fe, Co, Ni, Pd, Pt, Cu, Ag, Au or Zn, most preferred Ag or Au or Pt.

The reaction time of the liquid with the exposed area(s) is preferably from 1 second to 10 hours, preferably from 1 minute to 1 hour, more preferred from 10 minutes to 40 minutes.

After completion of the transition the structured resist cover layer may be removed by known techniques, e.g. washing with a solvent and the obtained mesosurface can be dried.

Brief description of figures

Figure 1

Ionic liquid gate controlled phase transitions in SrCoO _x a Sketch of the IL induced reversible phase transition between SrCoO ₃ (SCO) and V-SrCoO _2.5 or P-SrCoO _2.5

Cross-section STEM-HAADF images of: b SCO c V-SrCoO _2.5 d P-SrCoO _2.5

All images in Figures 1 b, c, and d projected along [100]. In the HAADF imaging mode, the atomic column intensity is approximately proportional to Z ² (Z is the atomic number), so that the brighter and darker columns are the Sr and Co cations, respectively e XAS of SCO, V-SrCoO _2.5 and P-SrCoO _2.5 f Phase diagram of the SCO-SrCoO _2.5 system versus \/ _G and R _G , as determined by XRD measurements

Figure 2

Oxygen planar defects in SrCoO _x

Cross-section STEM-ABF images of: a SCO, b V ^, -SrCoO _2.5 c P-SrCoO _2.5

All images in Figures 2 a, b and c are projected along [110]. The ABF images are more sensitive to low-Z elements such as oxygen so that the contrast is opposite to that in the HAADF images.

CoOx intensity profiles for: d SCO e V-SrCoO _2.5 and f P-SrCoO _2.5

The intensity profiles are taken from the areas marked by the dashed boxes in the corresponding ABF image. The Sr, Co, O, and oxygen vacancy sites (\/o) are indicated by yellow, red, green spheres, and green circles, respectively.

Figure 3

Ionic liquid gating effect on magnetic structure a M-T curve of SCO, \/-SrCoO _2.5 , and P-SrCoO _2.5 at 10 K b M-H curves of SCO, V-SrCoO _2.5 , and P-SrCoO _2.5 at 10 K c H _EB measured at 10 K for VSrCoO _2.5 /LSMO and P-SrCoO _2.5 /LSMO as a function of the in-plane angle Φ between H and [010] d M-H curves at 10 K for V-SrCoO _2.5 MO (Φ = 0°, [010] direction) and P- SrCoO _2.5 / LSMO (Φ = 45°, [110] direction) e XMLD (in total electron yield mode) of V-SrCoO _2.5 and P-SrCoO _2.5 for various <f>. A sketch of the XMLD measurement setup is shown in the inset.

Total energy for: f \/-SrCoO _2.5 and g P-SrCoO _2.5 of a 2x2x4 cell for three AF configurations (A-, C-, and G-type) as a function of seven different Co moment directions. Sketches of the AFspin configurations for V-SrCoO _2.5 and P-SrCoO _2.5 crystals are shown in the insets where the Co moment directions are indicated by arrows. The Sr, Co, O, and oxygen vacancy (Vo) sites in the sketches are indicated by yellow, red, green spheres, and green circles, respectively.

Figure 4

Creation of SrCoO _x mesostructures a Sketch of the creation of a SrCoOx mesostructured b Typical cAFM images, c Overview and d Enlarged TEM images for a SCO/V~SrCoO _2.5 (V) checkerboard mesostructure. This mesostructure is made by IL gating at V _G = +1.5 V of a SCO film through a resist mask that has several sets of orifices in the form of checkerboards. e Typical cAFM images f Overview and g Enlarged TEM images for a P-SrCoO _2.5 (P)/V-SrCoO _2.5 (V) checkerboard mesostructure. The enlarged TEM images of d and g are taken from the areas marked by the frames in c and f, respectively. This sample is made by IL gating of a SCO/V- SrCoO _2.5 (V) mesostructured without any resist mask at V _G = +2.5 V. Figure 5

SrCoO ₃ /V-SrCoO _2.5 magnetic textures a Kerr image of SCO/V-SrCoO _2.5 magnetic texture at 100 K without external magnetic field b Kerr loops of areas marked by black and red frames in a

MFM images of SCO/V-SrCoO _2.5 magnetic texture: c 200x200 nm ² checkboard and d 600x200 nm ² nanomagnets formed in a hexagonal array

The FM and AF parts are indicated by black and red frames in c. The desired patterns of the nanomagnets are marked by red frames in d. All the samples are made by IL gating on SCO sample through resist mask at V _G = +1 .5 V.

Figure 6

Magnetic racetrack formed by IL gating through a resist mask a Kerr image of racetrack device that has a width of 4 μm that is formed by resistmask IL gating (V _G = +1.5 V for 2 hours) of a SCO film. The images are taken at 76.5 K in the presence of a magnetic field of 210 Oe oriented in-plane along the racetrack channel to improve the MOKE contrast. Inset is the cAFM image at 300 K for an area including the racetrack (measured on a similar sample that was grown on a Nb-doped STO). b a sequence of Kerr microscope images of a single domain wall moved along the FM SCO wire by current pulse sequences formed from ten 20 ns long pulses; the current density is 1.07x10 ⁷ A/cm ² . No field is applied. Blue and red arrows indicate the current direction. c the domain wall velocity as a function of current pulse density J at 76.5 K and 124.4 K.

Figure 7

Ionic liquid (IL) gating a, bSchematic diagram of ionic liquid (IL) gating for ex-situ experiments c Time evolution of gate voltage (V _G ) during the gating procedure

Figure 8

Nonvolatility of ionic liquid gating effect on transport properties a Hall bar formed by standard photolithography and argon ion milling procedures. The Hall bar is 100 μm wide and 450 μm long. A large lateral electrode is formed by lift off techniques. b Temperature dependence of the resistance of a typical SCO film in the pristine state ( V _G = 0 V) and after gating at V _G = +1 .5 V and +2.5V for 30 min. The pristine metallic SCO phase is transformed to an insulating state after these gating procedures. c V _G and d channel resistance versus time when V _G = +1 .5 V and +2.5 V are applied at a rate of 10 ⁴ V/s. Gate voltages are applied for 30 min and then removed. The resistance in both cases remains constant after V _G is removed, proving the nonvolatility of IL gating effect. Data for 90 min after V _G is set to zero are shown but the non-volatile state remains indefinitely.

Figure 9

Nonvolatility and reversibility of IL gating effect revealed by XRD a Specular XRD patterns of pristine SCO (V _G =0 V) and samples gated by V _G = +1 .5 V and +2.5 V. V _G of +1.5 V and +2.5 V drive the perovskite type SCO to brownmillerite SrCoO _2.5 phase with oxygen vacancy planes vertical (V) and parallel (P) to the surface, labelled as V-SrCoO _2.5 and P-SrCoO _2.5 , respectively. XRD patterns of gated samples were recorded twice, namely (i) 20 min and (ii) 10 days after removal of V _G as labelled in the figure. Their reproducibility indicates that the IL gating effect is nonvolatile. b Specular and c non-specular XRD patterns for samples subject to different V _G application procedures: +2.5 V→ - 3 V, +1.5 V→ - 2 V, and +1.5 V → - 2 V → +2.5 V. At each voltage V _G is applied for 2 hours. The V- and P-SrCoO _2.5 structures can be transformed back to the pristine perovskite SCO phase by applying negative V _G . A film gated at +1 .5 V (V-SrCoO _2.5 ), was transformed to P-SrCoO _2.5 by first applying a V _G of-2V to transform the film first back to SCO and then followed by V _G = +2.5 V to enable the transformation to the P- phase.

Figure 10 of SrCoOx mesostructures

Conducting atomic force microscopy (cAFM) images of: a stripe SCO/\/-SrCoO _2.5 meso-structures, and b checkerboard SCO/V-SrCoO _2.5 meso-structures c stripe V-SrCoO _2.5 /P-SrCoO _2.5 meso-structures, and d checkerboard V-SrCoO _2.5 /P-SrCoO _2.5 meso-structures.

EXAMPLES

The invention will now be described by way of examples.

The above described method of manufacture of a mesosurface is demonstrated using the metallic ferromagnetic (FM) perovskite SrCoO ₃ and its twin counterpart, the insulating antiferromagnetic (AF) brownmillerite SrCoO _2.5 . It is known that thin layers of these compounds can be completely and reversibly transformed between each other by the addition or subtraction of oxygen via ionic liquid (IL) gating (see references14- 16). It was also shown that this can take place through micron sized holes in a resist layer (see reference17). In the present invention it is shown that these transformations can take place through holes in a resist layer that are twenty times smaller, as small as 100 nm in size or even lower. The structure of the brownmillerite phase that is formed in this way, consists of periodic layers of oxygen vacancies that are ordered either vertically at low gate voltage (V) or parallel to the substrate (P) above a threshold gate voltage that is applied to the IL during the gating process. These two oxygen vacancy ordered phases, V- and P-SrCoO _2.5 , have distinct antiferromagnetic structures and properties. Taking advantage of these dual states stabilized by distinct gate voltages one can create a series of magnetic meso-structures composed of regions formed from either of these two AF structures or the FM parent perovskite. The regions can be of any arbitrary shape.

SrCoO ₃ (SCO) films, 40 nm thick, were prepared by pulsed laser deposition on (001) SrTiO ₃ (STO) substrates. IL gating of these films was carried out in devices that are shown schematically in Fig. 7a and b. Lateral gate electrodes formed from Au were used to apply a gate voltage (V _G ) to the IL and thereby to transform the SCO film. This process, depending on the sign of V _G adds or subtracts oxygen from the SCO film in a non-volatile manner.

A thin film SCO/STO sample was cut into two halves, one of which was gated at V _G = 1.5 V and the other at 2.5 V. The voltage was increased from zero at a controlled and constant rate of R _G = 10 ⁴ V/s until the target V _G was reached where it was maintained constant for 2 hours (see Fig. 7c). The resulting gated films exhibit two distinct crystal structures, as schematically shown in Fig. 1a, and as revealed by cross-sectional scanning transmission electron microscopy-high angle annular dark field (STEM- HAADF) images in Fig. 1c and d. A comparison with the structure of the ungated film shown in Fig. 1b clearly shows dramatic structural changes.

The pristine SCO film has a uniform perovskite structure (Fig. 1b). In this image the bright and dark columns are the Sr and Co cations, respectively, since the atomic column intensity is approximately proportional to Z ² (Z is the atomic number) in the HAADF imaging mode. By contrast the HAADF image for the V _G = +1.5 V sample shows dark columns along the vertical [001] direction indicating a superstructure which involves a doubling of the a-lattice parameter along [100] (see Fig. 1c). A different superstructure is found for the V _G = +2.5 V sample in which dark rows can be observed parallel to [100], the surface of the STO substrate (Fig. 1d): this corresponds to a quadrupled c-lattice parameter along [001] (see Table 1).

Such superstructures are typical of ordered oxygen vacancy planes in brownmillerite phases: the planes containing oxygen vacancies appear darker than the stoichiometric planes in the HAADF images. The superstructures observed show that positive gate voltages create oxygen planar defects in the SCO thin film and, more importantly, that the orientation of these defects can be manipulated by the magnitude of V _G . In bulk material the P phase is typically observed and this same phase has previously been detected in thin films grown on STO but the present invention for the first time provides a route to form the V phase on STO. Thus, IL gating provides a novel route to the controlled formation of both of these phases.

The chemical composition of the samples was studied by X-ray absorption spectroscopy (XAS). The Co-L ₃ peaks for both V _G = +1 .5 V and +2.5 V shift to a lower photon energy (by ~1.1 eV), as compared to pristine SCO, indicating the Co valence is reduced by ~1 e~ (Fig. 1e). Thus, the corresponding stoichiometries of both the V and P phase are close to SrCoO _2.5 . The three phases, SCO, V-SrCoO _2.5 and P- SrCoO _2.5 , can be unambiguously identified from their x-ray diffraction patterns (XRD).

In addition to the magnitude of V _G , the rate of increase of V _G during the gating process was also varied. This allows for the phase stability diagram of SrCoO _x to be defined with respect both to \/ _G and RG. When RG is sufficiently large, and when \/ _G lies between +0.4 V and +1 .9 V, SCO is transformed to V-SrCoO _2.5 , but when V _G > +1 .9 V (but below +4.2 V, above which the sample undergoes irreversible structural changes), rather P-SrCoO _2.5 is formed. However, the situation completely changes when R _G < 2x10 ^-3 V/s, under which conditions only the V-SrCoO _2.5 phase is found, no matter how large is VG. At these low rates, the sample has sufficient time to undergo a complete transformation to the V phase before V _G ^Crit = +1.9 V is reached (see Fig. 8b). Any further increase in V _G beyond V _G ^Crit does not induce a phase transition from V to P. In order to realize such a phase transformation, it is necessary to transform the structure from V-SrCoO _2.5 back to SCO by applying a sufficiently large negative VG followed by the application of a suitably large positive V _G (see Fig. 9). For 2x10~ ³ V/s < RG < 5x10” ¹ V/s, a mixture of both the V and P phases are observed when V _G exceeds V _G ^Crit (see Fig. 1f). STEM-ABF (annular bright-field) images recorded along the [110] zone axis that is rotated by 45° as compared to the HAADF images discussed above, allows for the detailed determination of the positions of both the oxygen ions and vacancies, as shown in Fig. 2a-c. The ABF images are sensitive to low-Z elements such as oxygen. In the corresponding intensity line profiles of the CoO _x sublayers (Fig. 2d— f), the Co and filled O sites are seen as deeper and shallower troughs, respectively, while the oxygen vacancy sites are identified by broad peaks (in Fig. 2f). The ABF images of SCO and V-SrCoO _2.5 look similar (Fig. 2a and b), but the corresponding intensity profiles within a CoO _x sub-layer in V-SrCoO _2.5 (Fig. 2e) show shallower O troughs, as compared with those in SCO (Fig. 2d). Since the oxygen vacancies within the (100) plane of V-SrCoO _2.5 are imaged along the [110] zone, each imaged O column has 25%

O vacancies and thus a resultant higher ABF intensity. The CoO ₂ sublayer in P- SrCoO _2.5 shows an intensity profile similar to that of SCO (compare Fig. 2f to Fig. 2d). On the other hand, one can clearly see the pairing of the Co atoms within the oxygen deficient CoO sublayer in P-SrCoO _2.5 that results from the ordering of the oxygen vacancies (see Fig. 2c). The columns of oxygen vacancies (marked by green circles) are located between the Co-Co pairs in each CoO sublayer, which can be clearly seen as peaks in the CoO sub-layer intensity profile (Fig. 2f).

The magnetic properties of V-SrCoO _2.5 and P-SrCoO _2.5 were investigated experimentally and theoretically, which, owing to their distinct crystal structures, are also expected to differ. Temperature dependent magnetization (M-T) curves monitored in-plane along the [010] direction of the substrate are shown in Fig. 3a. These data show that SCO is a ferromagnet with a Curie temperature (Tc) of ~230 K. By contrast, both V- and P-SrCoO _2.5 do not show any evidence for FM behavior but are rather consistent with AF behavior with Neel temperatures (TN) of 323 K and 316 K, respectively.

M vs field (H) hysteresis loops of all three samples at 10 K are shown in Fig. 3b, where only pristine SCO shows a typical FM loop.

The hysteresis loops of both heterostructures of t/- and P-SrCoO _2.5 were measured along different in-plane directions, $ the angle between H and STO [010], at 10 K after cooling in an in-plane H = 2 kOe for each measurement. The loops are shifted from zero field by an exchange bias field (HEB). The angular dependence of WEB is compared for the V and P phases in Fig. 3c. Peaks in HEB are found at = 0°, 90°, and 180°, for V-SrCoO _2.5 / Lao.7Sro.3MnO ₃ (LSMO) but at = 45° and 135° for P-SrCoO _2.5 /LSMO. It is noteworthy that the maximum magnitude of HEB is about 5 times larger for V- SrCoCWLSMO than for P-SrCoO _2.5 /LSMO (Fig. 3d) and along different crystallographic directions. These substantial differences suggest the AF structures in these phases are distinct.

The AF structures were probed using x-ray magnetic linear dichroism (XMLD) in the vicinity of the Co-/_2,3 absorption edge. XMLD is the difference in absorption of an incident x-ray beam that is linearly polarized in two orthogonal directions P1 and P2, as shown in the inset to Fig. 3e. Prior to the measurements, the samples were heated to 400 K followed by in-plane field cooling (2 kOe) in different angular directions ($ to 300 K, the measurement temperature (see inset to Fig. 3e). P1 is set along the cooling field direction. The XMLD amplitude is related to the difference in the magnitude of the AF moment that is parallel to P1 and P2. Thus, the larger the amplitude of the XMLD, the more the Co moments are oriented along P1 or P2. As ^ is varied between 0° and 45°, the XMLD signal from V-SrCoO _2.5 continuously diminishes, which suggests that the easy axis of the Co moments is along [010]. By contrast, the continuously increasing XMLD signal in P-SrCoO _2.5 shows that the [110] (45°) direction corresponds to the AF easy axis in this case. Thus, the AF easy axes for V-SrCoO _2.5 and P-SrCoO _2.5 are different and are parallel to the STO directions, respectively. These findings are in good agreement with the angle dependent HEB studies discussed in Fig. 3c. As both V- and P-SrCoO _2,5 films grow in two rotational domains on the STO substrate, that are oriented at 90° to each other, the angular period of the HEB oscillation is equal to 90° but not 180°, which would otherwise be the case for a single domain.

First principles calculations of the ground state energy of both \/-SrCoO _2.5 and P- SrCoO _2.5 using a 2x2x4 supercell for three AF configurations, namely A-, C-, and G- type, were carried out. The total energy is displayed versus different orientations of the Co moments in Fig. 3f and g, respectively. In V-SrCoCh.s, the A-type AF configuration with the Co moments along [010] has the lowest energy. By contrast, the G-type AF structure with a [110] easy axis is preferred for P-SrCoO _2.5 which is in agreement with neutron powder-diffraction results on bulk material. The AF spin configurations of V- SrCoO _2.5 and P-SrCoO _2.5 are sketched in the insets to Fig. 3f and g, respectively. The calculations are consistent with the easy axis moment directions found from XMLD studies. The calculations also appear to be consistent with the experimental observation that HEB is larger in \/-SrCoO _2.5 . An A-type AF ordering in \Z-SrCoO _2.5 shows an uncompensated moment at its surface which should therefore result in a stronger HEB as compared with that of the G-type AF ordering in P-SrCoO _2.5 whose surface moment is fully compensated.

Manufacture of engineered electric and magnetic meso-structures

Several magnetic spatial meso-textures were formed from both metallic FM SCO and the two insulating AF SrCoO _2.5 phases that can be formed by IL gating. AS mentioned above micron sized SCO circular regions could already be formed in a layer of P- SrCoO _2.5 by local liquid gating (see reference 17). In this case, however, the boundary between these phases was quite broad due to substantial oxygen ion diffusion parallel to the substrate. Using resist masks with much smaller orifices, as small as ~50 nm, that are spun on top of a SCO layer and patterned by electron beam lithography, regions of one or the other of the AF phases can be formed by applying positive voltages below or above V _G to an IL that is placed on top of the resist mask, as illustrated in the sketch in Fig. 4a. An example in which a checkerboard pattern of the phase is formed in an initial uniform SCO layer is given in Fig. 4b-d. The local structure within the checkerboard has dimensions of -100x100 nm ² and the overall pattern was 20x20 μm ² in area. After IL gating and the removal of the resist layer, cAFM (conducting atomic force microscopy) is used to image the patterned structure by taking advantage of the substantially different conductivities in the SCO and V phases. The cAFM image matches perfectly the resist mask clearly demonstrating that the SCO was transformed to the V phase only within the tiny orifices. The structural transformation was confirmed by TEM analysis as shown in Fig. 4c in which dark and bright regions alternate along the [100] direction, one axis of the checkerboard pattern. A magnified TEM image in Fig. 4d shows that these regions have distinct structures of in the one case the perovskite SCO phase and in the other the V-SrCoO _2.5 (V) phase. The dimensions of these regions, of -100 nm, are consistent with the cAFM images shown in Fig. 4b. The boundary between the SCO and \/-SrCoO _2.5 region is very sharp (only -6 nm wide).

Manufacture of a magnetic meso-structure formed from the two AF SrCoOa.s phases

The two AF SrCoO _2.5 phases are formed at different IL gate voltages. First the same spatial texture as in Fig. 4b is formed. Then after the removal of the resist layer a second IL gating procedure is carried out without any resist mask at a higher VG = +2.5 V. cAFM and TEM images for the resulting meso-structure are shown in Fig. 4e-g. After this second IL gating, the previous metallic regions become highly insulating (the maximum leakage current is less than 20 nA) but, nevertheless, a checkerboard structure with distinct conductivities is visible (Fig. 4e). In the corresponding TEM images, one can now see a superstructure consisting of V-SrCoO _2.5 (V) and P- SrCoO _2.5 (P) phases, respectively (Fig. 4f and g). This means that the V-SrCoO _2.5 phase is not directly changed to P-SrCoO _2.5 phase even when a large V _G is applied (which is consistent with the phase diagram given in Fig. 1 f).

Using the same techniques a wide variety of magnetic meso-structures can be formed from any of the FM SCO and the AF V-SrCoO _2.5 and P-SrCoO _2.5 phases. In Fig. 10 several more examples are given where meso-structures were created with length scales varying from 10 μm to 200 nm in the form of checkerboards and stripes. The large area checkerboards (see Fig. 5a) allowed the local magnetic properties to be studied using magneto-optical Kerr microscopy (MOKE) that confirmed the ungated SCO regions were magnetic, as shown in Fig. 5b. Mesostructures with smaller feature sizes were studied with variable temperature magnetic force microscopy (MFM). An MFM image of a 200x200 nm ² checkerboard mesostructure is given in Fig. 5c.

To explore functional devices a periodic hexagonal array of elongated nanomagnets (600x200 nm ² ) was made using the same procedures as discussed above, as shown in Fig. 5d. In this way a spin ice structure could be formed from an initial uniform SCO layer. The magnetic frustration between the hexagonally arranged nanomagnets can clearly be seen in the MFM image shown in Fig. 5d.

Finally, a “magnetic racetrack" was formed again from an initial uniform SCO film. In this case the resist mask is used to define magnetic stripes that form the racetracks, as can be seen very clearly in the MOKE image shown in Fig. 6a. cAFM (inset to Fig. 6a) also confirms that metallic magnetic racetracks have been formed in an insulating matrix. Although the SCO is only ferromagnetic below ~230 K, nevertheless, the current induced motion of magnetic domain walls in the SCO racetracks is possible, as demonstrated in Fig. 6b at temperatures of 76.5 K and 124.4 K. The domain walls can be moved at velocities of up to -22.7 m/s at 76.5 K using a current density of only J = 1.07x10 ⁷ A/cm ² . The SCO racetracks have high resistivities of ~8.4x10 ^-6 Cm at 76.5 K and ~1.0x10 ^-5 Ωm at 124.4 K and, therefore, cannot sustain higher current densities. However, the efficiency of driving the domain walls in these SCO racetracks with current is higher even than the fastest domain wall speeds reported to date that rely on a bulk spin-transfer torque mechanism.

METHODS

Sample preparation

SCO thin films, 40 nm thick, were grown on STO (001) substrates at 725 °C, in an oxygen pressure (po) of 2x10 ^-1 mbar, using pulsed laser deposition (PLD), The growth temperature and po for La _0.7 Sr _0.3 MnO ₃ (LSMO) were 600 °C and 2x10 ^-1 mbar, respectively. Undoped and 0.5 wt % Nb doped single side polished STO substrates were used. After deposition, the SCO films and the SCO/LSMO heterostructures were cooled to room temperature in an oxygen atmosphere of 500 mbar. The IL N,N-diethyl- N-(2-methoxyethyl)-N-methylammonium bis(trifluoromethylsul-fonyl)-imide (DEME- TFSI), was used for all gating experiments. The IL and the devices were separately baked at 130 °C in high vacuum (10~ ⁷ mbar) for at least 12 hours before the gating experiments were carried out. The resist masks were prepared by electron beam lithography (Raith Nanofabrication system) using a positive resist CSAR (APR6200.09 -200 nm thick). Large area thin films were gated in a probe station (pressure <10 ^-6 mbar). All the measurements were carried out ex-situ after removing the IL and any resist. Devices for transport measurements were prepared by photo-lithography and wet etching in the form of Hall-bars. Electrical contacts were formed from Au (60 nm)/Cr (10 nm) that were deposited by thermal evaporation.

Sample characterization

An FEI Titan 80-300 microscope, which is probe-corrected to achieve a point-to-point resolution of ~1 A, was used for the (S)TEM studies. The convergence semi-angle for STEM-HAADF imaging was ~22 mrad, while the collection semi-angle was 70-176 or 200 mrad for STEM-HAADF imaging, and 12-24 mrad for STEM-ABF imaging. XRD measurements were carried out using a Bruker AXS D8 ADVANCE x-ray four circle diffractometer. The magnetic properties (M-H curves at 10 K and M-T curves from 400 to 10 K) were carried out using a Superconducting quantum interference device- vibrating sample magnetometer (SQUID-VSM). The XAS and XMLD measurements in total electron yield (TEY) mode were performed at Beamline BL08U1 A at the Shanghai Synchrotron Radiation Facility (SSRF). The transport properties in Hall bar devices were measured in a Quantum Design DynaCooL All the characterization and transport experiments, unless otherwise specified, were carried out at 300 K.

First-principle calculations

First-principle calculations were carried out with the projector augmented wave implementation of the Vienna ab initio simulation package (VASP). A generalized gradient approximation plus U (GGA+D) was used as the energy functional, with a setting of the Coulomb repulsion U = 4.5 eV and the exchange interaction J = 1.0 eV for the Co d electrons. A 2x2x4 supercell was used for SrCoO _2.5 for both the vertical and parallel superstructures. All the structures are optimized with a cutoff energy of 550 eV and appropriate k-point meshes both of which were increased until convergence. References

1. Shukla, S., Deheri, P. K. & Ramanujan, R. V. Magnetic nanostructures: synthesis, properties, and applications. Springer Handbook of Nanomaterials. Springer Berlin Heidelberg: Berlin, Heidelberg, 2013, pp 473-514.

2. Martin, J. I., Nogues, J., Liu, K., Vicent, J. L. & Schuller, I. K. Ordered magnetic nanostructures; fabrication and properties. J Magn. Magn. Mater. 256, 449-501 (2003).

3. Wolf, S. A. et al. Spintronics: a spin-based electronics vision for the future. Science 294, 1488-1495 (2001).

4. Chumak, A. V., Vasyuchka, V. I., Serga, A. A. & Hillebrands, B. Magnon spintronics. Nature Phys. 11 , 453-461 (2015).

5. Grollier, J., Querlioz, D. & Stiles, M. D. Spintronic nanodevices for bioinspired computing. Proc. IEEE 104, 2024-2039 (2016).

6. Nisoli, C., Moessner, R. & Schiffer, P. Colloquium: Artificial spin ice: Designing and imaging magnetic frustration. Rev. Mod. Phys. 85, 1473-1490 (2013).

7. Torrejon, J. et al. Neuromorphic computing with nanoscale spintronic oscillators. Nature 547, 428-431 (2017).

8. Wang, Y. et al. Magnetization switching by magnon-mediated spin torque through an antiferromagnetic insulator. Science 366, 1125-1128 (2019).

9. Han, J., Zhang, P., Hou, J. T., Siddiqui, S. A. & Liu, L. Mutual control of coherent spin waves and magnetic domain walls in a magnonic device. Science 366, 1121-1125 (2019).

10. Parkin, S. S. P., Hayashi, M. & Thomas, L. Magnetic domain-wall racetrack memory. Science 320, 190-194 (2008).

11. Parkin, S. & Yang, S.-H. Memory on the racetrack. Nature Nanotechnol. 10, 195-198 (2015).

12. Ross, C. A. Patterned magnetic recording media. Ann. Rev. Mater. Res. 31 , 203-235 (2001).

13. Kim, S. et al. Nanoscale patterning of complex magnetic nanostructures by reduction with low-energy protons. Nature Nanotechnol. 7, 567-571 (2012).

14. Jeong, J., Aetukuri, N., Graf, T., Schladt, T. D., Samant, M. G. & Parkin, S. S. Suppression of metal-insulator transition in VO2 by electric field-induced oxygen vacancy formation. Science 339, 1402-1405 (2013). Lu, N. et al. Electric-field control of tri-state phase transformation with a selective dual-ion switch. Nature 546, 124-128 (2017). Altendorf, S. G., Jeong, J., Passarello, D., Aetukuri, N. B., Samant, M. G. & Parkin, S. S. P. Facet-independent electric-field-induced volume metallization of tungsten trioxide films. Adv. Mater. 28, 5284-5292 (2016). Cui, B. et al. Direct imaging of structural changes induced by ionic liquid gating leading to engineered three-dimensional meso-structures. Nature Commun. 9, 3055 (2018).