Magnetic ordering in out-of-plane artificial spin systems based on the Archimedean lattices
A. Pac, G. M. Macauley, J. R. Massey, A. Kurenkov, F. Mila, P. M. Derlet, L. J. Heyderman · 2025
Read Magnetic ordering in out-of-plane artificial spin systems based on the Archimedean lattices by A. Pac, G. M. Macauley, J. R. Massey, A. Kurenkov, F. Mila, P. M. Derlet, L. J. Heyderman (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-08462.txt.
Introduction
Aleksandra Pac and six co-authors map in 2025 how arrays of tiny upright magnets order themselves on Archimedean lattice patterns. Artificial spin systems let physicists stage frustration and phase changes on a benchtop, one nanomagnet at a time. These magnets point out of the plane, a newer and less charted geometry.
After a demagnetizing routine, magnetic force microscopy read the frozen patterns, and simulations assigned each lattice an effective temperature. Every lattice sat above its ordering threshold, showing how hard system-wide order is here. Neighbor effects beyond the nearest were charted for four example lattices. The family splits into lattices that can settle into perfect alternating order and frustrated ones that cannot.
Catalog record
| CREATOR | A. Pac, G. M. Macauley, J. R. Massey, A. Kurenkov, F. Mila, P. M. Derlet, L. J. Heyderman |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 77 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-08462
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Magnetic ordering in out-of-plane artificial spin systems based on the Archimedean lattices A. Pac, G. M. Macauley, J. R. Massey, A. Kurenkov, F. Mila, P. M. Derlet, L. J. Heyderman Abstract Artificial spin systems, sometimes referred to as artificial spin ices, are arrays of coupled nanoscale magnets that order according to the lattice geometry, nanomagnet shape and magnetic anisotropy. Here we characterize a family of artificial spin systems that are formed by placing arrays of out-of-plane nanomagnets on the vertices of the Archimedean lattices. On demagnetizing these nanomagnet arrays using a magnetic field protocol and subsequently imaging the magnetic configuration using magnetic force microscopy, we observe different types of magnetic order. We compare our experimental results with those predicted by Monte Carlo simulations to assign an effective temperature to each lattice. We find that, for all of the lattices, the assigned effective temperature is above the transition temperature. This reflects the difficulty of obtaining system-spanning order in lattices with out-of-plane nanomagnets. We consider to what extent further-neighbor interactions affect the phase diagram and spin-spin correlations in each lattice, illustrating our results with four example lattices. We can divide the lattices into three main categories: bipartite lattices that admit a perfect antiferromagnetic ground state, frustrated lattices where ordering proceeds via a single step, and frustrated lattices with two-step-ordering. Our work highlights the diversity of magnetic ordering that can be hosted in two-dimensional artificial spin systems with out-of-plane nanomagnets, and demonstrates the importance of including long-range interactions to explain the magnetic ordering. Such insights will be important for incorporating artificial spin systems into novel computing applications. Preprint: APS/123-QED I Introduction Artificial spin systems are arrays of dipolar-coupled nanomagnets [ 1 , 2 ] that exhibit interesting phenomena including spin liquid [ 3 ] and spin glass physics [ 4 , 5 ] , vertex frustration [ 6 ] , and emergent magnetic monopoles [ 7 , 8 ] . They were initially envisaged as a two-dimensional (2D) analogue to frustrated rare earth spin-ice compounds [ 9 ] , themselves a magnetic counterpart to water ice, in which incommensurate bonding distances between hydrogen ions and oxygen centres give rise to a residual entropy [ 10 ] . In this framework, artificial spin systems are often referred to as artificial spin ices because they demonstrate similar local constraints, short-range correlations and macroscopic ground state degeneracy. However, not all artificial spin systems display ice-like physics, but rather manifest a wide variety of emergent behaviours that depend on the lattice geometry, nanomagnet shape and magnetic anisotropy. Typically, artificial spin systems are 2 D lattices of single-domain nanomagnets, whose magnetic configurations can be directly accessed in real space, with magnetic force microscopy (MFM) [ 1 ] , magneto-optical Kerr effect (MOKE) microscopy [ 11 ] , and synchrotron x-ray or electron microscopy [ 12 , 8 , 13 , 14 ] , and also in reciprocal space, with x-ray scattering [ 15 ] , for example. Analysis of these magnetic configurations is often simplified by assuming that each single-domain nanomagnet can be represented as a macrospin, with a direction set by the direction of the net moment of the nanomagnet. Beyond their fascinating physics, artificial spin systems have also attracted interest as a way to perform logic operations [ 16 , 17 ] , and for integration into unconventional computing schemes such as neuromorphic computing [ 18 , 19 , 20 , 21 ] . For such applications, arrays of nanomagnets with perpendicular magnetic anisotropy [ 12 , 22 , 23 , 24 , 25 ] are of particular interest because the magnetic state of the individual nanomagnets can be accessed via electrical means using the anomalous Hall effect [ 26 , 27 , 28 , 29 ] . This allows the direct reading or writing of the state of each nanomagnet using current, thus making it possible to interface artificial spin systems with existing CMOS architectures. However, before such out-of-plane artificial spin systems can be exploited in applications, it is important to have a greater understanding of the effect of the lattice geometry and the long-range dipolar coupling on the resulting physics. The fine balance between these two factors sets the hierarchy of intermagnet interactions, or weights [ 30 ] , which is interesting for designing neuromorphic networks [ 31 , 32 , 33 , 34 ] . Here we present a complete experimental and computational classification of the behaviour of out-of-plane artificial spin systems based on the Archimedean lattices [ 35 ] , which are shown in Fig. 1 . As can be seen in this figure and also in the Supplemental Material, Section 1 [ 36 ] , the Archimedean lattices are 11 uniform 2 D lattices made from regular polygons, such that every vertex is topologically equivalent. In all of the lattices that we fabricated, the base nanomagnet has the same diameter and is made of the same magnetic material. A given set of 11 Archimedean lattices with the same nanomagnet separation thus only differ in the number and positions of neighboring macrospins. This allows us to compare the emergent magnetic order resulting from the dipolar interactions across the whole Archimedean family by only changing the lattice geometry. For each lattice, we determine the magnetic state following a magnetic field demagnetization, and compare the observed spin-spin correlations with those predicted by simulated anneals using Monte Carlo simulations. This allows us to assign an effective temperature to each lattice, quantifying the extent to which the low-temperature spin arrangements can be accessed experimentally. In particular, by comparing the effective temperature of the experimental systems to the critical temperature of each lattice determined with Monte Carlo simulations, we can ascertain how close each artificial spin system gets to its phase transition. For the most part, we find that these lattices reach an effective temperature slightly above their critical temperature. Of the 11 lattices, four are bipartite: the Square, Honeycomb, Square Hexagonal Dodecagonal (SHD) and CaVO lattices. This means they admit a perfectly-ordered antiferromagnetic ground state. The remaining seven exhibit varying degrees of frustration, which we quantify through their entropy as determined with Monte Carlo simulations based on single spin flips. Two of these seven lattices, the Triangular and Kagome lattices, are highly frustrated and their ordering proceeds in two steps. Figure 1: SEM images of artificial spin systems with out-of-plane nanomagnets located on the vertices of the 11 Archimedean lattices. The name we give to each lattice, as well as the notation of Grünbaum and Shephard [ 35 ] , which describes the shapes meeting at each lattice point, are given in the top left corner of each panel. The building block for all lattices is a Co/Pt nanomagnet with diameter d=200 nm. Because of the strong perpendicular magnetic anisotropy, each nanomagnet behaves as a single macrospin with a magnetic moment either pointing out of or into the plane, which we refer to as up or down, respectively. While we have created arrays of nanomagnets with lattice parameters going from 220 nm to 300 nm, we show here arrays with lattice constant a=300 nm, where the separation between adjacent nanomagnets s=100 nm, because the largest lattice constant allows for better discrimination of each pattern. For each