Spinhenge@home

Spinhenge@home was a volunteer computing project on the BOINC platform, operated by the Bielefeld University of Applied Sciences (Fachhochschule Bielefeld), Germany, in cooperation with the University of Osnabrück and the Ames Laboratory in Iowa, United States. The project employed the Metropolis Monte Carlo algorithm to simulate spin dynamics and thermodynamic properties of nanoscale single-molecule magnets, seeking to advance the understanding of molecular magnetism and its potential applications in medicine and nanotechnology. Its public beta phase opened on 1 September 2006, and it became one of the largest BOINC-based projects in the world by participant count, ultimately enrolling close to 60,000 registered users and over 150,000 participating computers before entering an indefinite hiatus in September 2011.

Spinhenge@home
Project
Status	Completed
Category	Physics
Compute	CPU
Development
Developer	Prof. Dr. Christian Schröder; Thomas Hilbig
Author	Thomas Hilbig (Diplomarbeit)
Sponsor	Bielefeld University of Applied Sciences (Fachhochschule Bielefeld); Ames Laboratory (U.S. Department of Energy)
Maintainer	Christian Schröder
Initial release	September 1, 2006  (20 years ago)
Completed	2011 (hiatus); permanently closed c. 2013
Discontinued	c. 2013
Software
Written in	C++
Operating system	Windows (primary); Linux (experimental, slower)
Metadata
Website	http://spin.fh-bielefeld.de

Background and scientific motivation

Molecular magnetism is the study of magnetic phenomena arising from individual molecules rather than bulk materials. At the nanoscale, specially engineered coordination compounds known as single-molecule magnets (SMMs) can retain a net magnetic moment below a characteristic blocking temperature, exhibiting magnetic hysteresis of purely molecular origin. This behavior is governed by quantum mechanical spin interactions described by the Heisenberg spin Hamiltonian:

$\hat{H}=-\sum _{i<j}{J}_{ij}{\widehat{\mathbf{𝐒}}}_i\cdot {\widehat{\mathbf{𝐒}}}_j$

where $J_{ij}$ is the exchange coupling constant between magnetic ions i and j, and ${\widehat{\mathbf{𝐒}}}_i$ is the spin operator at site i. Positive $J$ denotes ferromagnetic (parallel) coupling; negative $J$ denotes antiferromagnetic (antiparallel) coupling.

At low temperatures, an SMM's magnetic relaxation time $\tau$ follows a Néel–Arrhenius-type thermal activation law,

$\tau ={\tau }_0\exp \left(\frac{U_{\text{eff}}}{k_{\text{B}}T}\right)$

where $U_{\text{eff}}$ is the effective energy barrier to spin reversal, $k_{\text{B}}$ is the Boltzmann constant, $T$ is temperature, and ${\tau }_0$ is a characteristic attempt time on the order of 10⁻⁹ to 10⁻¹⁰ s. Because $U_{\text{eff}}$ depends on both the total ground-state spin $S$ and the magnetic anisotropy, numerical simulations of many possible exchange-coupling configurations are necessary to understand and ultimately design new SMMs.

Potential applications of SMMs and nanoscale molecular magnets include:

• High-density magnetic data storage (each molecule as a single bit);
• Spintronics and molecular-scale switches;
• Quantum computing using molecular spin states;
• Biomedical uses, such as targeted cancer therapy (local tumor hyperthermia) through magnetically heated nanoparticles.

Computing the thermodynamic and dynamic properties—such as susceptibility and spin-spin correlation functions—of molecules with tens of spin centers requires a vast number of Monte Carlo steps. This computational burden made volunteer computing a natural fit.

Project history

Origins

Spinhenge@home grew out of research by Professor Christian Schröder of the Department of Electrical Engineering and Computer Science at the Bielefeld University of Applied Sciences, whose focus was Computational Materials Science and Engineering. The project was initially realized as a diploma thesis (Diplomarbeit) by Thomas Hilbig. Building on the BOINC middleware, which the University of California, Berkeley had released publicly in 2002, the project connected the physical simulation work with the then-emerging world of public resource computing.

The project website was hosted at spin.fh-bielefeld.de. Schröder described it as one of the largest BOINC-based public resource computing projects in the world.

Beta launch and growth

The project entered public beta testing on 1 September 2006. The German distributed computing community responded enthusiastically: for example, the team Planet 3DNow! joined on 12 September 2006, and a competitive "race" for credit milestones between teams was arranged in January 2008, involving participants from SETI.Germany, SETI.USA, and ESL, among others.

On 1 May 2010, the project received a five-out-of-five quality seal (Gütesiegel) from the German volunteer computing community organization Rechenkraft, marking it as absolutely recommended.

Hiatus and closure

On 28 September 2011, the project team announced a hiatus while they reviewed accumulated results and planned hardware upgrades. As of the project manager's statement in October 2013, Spinhenge@home was still described as active, but its website became unreachable for extended periods and work units ceased. By July 2022, Wikipedia's article on the project noted the hiatus was ongoing and the project was likely permanently closed. The project is now listed as a completed BOINC project in the BOINC Synergy project directory.

Computational method

The core simulation application—named metropolis in the BOINC project files—implemented the Metropolis Monte Carlo method for classical Heisenberg spin systems. At each Monte Carlo step, a trial spin flip at a lattice site is proposed; if the resulting change in energy $\mathrm{\Delta }E<0$, the trial is accepted outright; if $\mathrm{\Delta }E\ge 0$, it is accepted with probability $\exp (-\mathrm{\Delta }E/k_{\text{B}}T)$, preserving detailed balance and the Boltzmann equilibrium distribution.

The simulations targeted molecules with dozens of magnetic ions, computing quantities such as:

• The magnetic susceptibility $\chi (T)$ as a function of temperature;
• Dynamic spin-spin correlation functions;
• Magnetization curves under applied fields.

These outputs could then be matched against experimental measurements (e.g., from SQUID magnetometry or neutron scattering) to determine or refine the exchange coupling constants $J_{ij}$ for specific molecules.

The primary computing platform was Windows, with a Linux experimental build available (though notably slower due to required 32-bit compatibility libraries for graphics rendering).

Scientific molecules studied

The icosidodecahedron geometry—identical to the arrangement of the 30 magnetic ions in Mo₇₂Fe₃₀ and Mo₇₂Cr₃₀, molecules studied using Spinhenge@home simulations.

A prominent scientific result enabled by Spinhenge@home involved the frustrated polyoxometalate molecular magnets Mo₇₂Fe₃₀ and Mo₇₂Cr₃₀—spherical cage molecules each housing 30 Fe(III) or Cr(III) magnetic ions at the vertices of an icosidodecahedron. The complete chemical formula of Mo₇₂Fe₃₀ is:

$[{\text{Mo}}_{72}{\text{Fe}}_{30}{\text{O}}_{252}({\text{CH}}_3\text{COO}{)}_{12}\{{\text{Mo}}_2{\text{O}}_7({\text{H}}_2\text{O}){\}}_2\{{\text{H}}_2{\text{Mo}}_2{\text{O}}_8({\text{H}}_2\text{O})\}({\text{H}}_2\text{O}{)}_{91}]\cdot 150{\text{H}}_2\text{O}$

Experimental measurements of the differential susceptibility $dM/dH$ of both molecules showed pronounced deviations from the predictions of standard Heisenberg models using a single exchange constant. Schröder and collaborators used the massive Monte Carlo simulations made possible by Spinhenge@home volunteers to formulate a nearest-neighbor model where the 60 nearest-neighbor exchange interactions in each molecule are described by a two-parameter probability distribution of exchange constants, achieving excellent agreement with experiment.

Earlier related work by Schröder and collaborators (predating the BOINC project) included a study of the metamagnetic phase transition of the antiferromagnetic Heisenberg icosahedron, published in Physical Review Letters in 2005, which laid the groundwork for the volunteer-computing simulations.

Institutional collaborators

• Bielefeld University of Applied Sciences (Fachhochschule Bielefeld, now Hochschule Bielefeld / HSBI) — Department of Electrical Engineering and Computer Science. Prof. Schröder led the Computational Materials Science and Engineering group there.
• University of Osnabrück — collaborated on the molecular magnetism simulation work.
• Ames Laboratory (U.S. Department of Energy, Ames, Iowa) — co-authored key papers, with work supported by the Department of Energy Basic Energy Sciences under Contract No. DE-AC02-07CH11358.

Publications

The large-scale Monte Carlo simulations produced by Spinhenge@home volunteers directly supported peer-reviewed research. Key publications include:

1. (2008-06-04).Multiple nearest-neighbor exchange model for the frustrated magnetic molecules Mo₇₂Fe₃₀ and Mo₇₂Cr₃₀. Physical Review B. pp. 224409. DOI: 10.1103/PhysRevB.77.224409. — This paper explicitly thanks the thousands of Spinhenge@home volunteers and states that the large-scale Monte Carlo simulations were made possible by volunteer computer resources.
2. (2005-01-07).Competing Spin Phases in Geometrically Frustrated Magnetic Molecules: Spin Dynamics in the Icosahedral Keplerate Mo₇₂Fe₃₀. Physical Review Letters. pp. 017205. DOI: 10.1103/PhysRevLett.94.017205. — Pre-Spinhenge foundational research on Mo₇₂Fe₃₀.

Schröder's broader body of work from the Bielefeld group, continuing after Spinhenge@home, extended into hybrid molecular and spin dynamics simulations of nanoparticle ensembles, BOINC framework modeling, and magnetoresistive systems research. A 2021 paper acknowledging the Spinhenge research lineage reported a giant spin molecule with ninety-six parallel unpaired electrons.

Additionally, Schröder presented results at the American Physical Society meetings:

• APS 75th Annual Meeting of the Southeastern Section (2008): "QMC Goes BOINC: Using Public Resource Computing to Perform Quantum Monte Carlo Calculations."
• APS 76th Annual Meeting of the Southeastern Section (2009): "How to use 100,000 PCs for studying magnetism."

Legacy and successor work

Following Spinhenge@home, Schröder's group developed the Visu@lGrid project—a framework applying UML profiling and abstraction to BOINC-based projects—and ComsolGrid, a framework for large-scale parameter studies using COMSOL Multiphysics and BOINC, presented at the COMSOL Conference 2010. These successors demonstrated how the BOINC infrastructure and lessons learned from Spinhenge@home could be applied to a broader range of computational science problems.

Schröder's group also published methodological papers on BOINC project design, including model-based generation of workunits and a UML profile for BOINC, contributing to the general infrastructure of volunteer computing research.

See also

• BOINC
• BOINC projects
• Einstein@Home
• Volunteer computing
• Single-molecule magnet
• Molecular magnetism
• Monte Carlo method
• Ames Laboratory

External links

• Spinhenge@home project website (archived, July 2006) via Wayback Machine
• More information about Spinhenge@home (archived, July 2012) via Wayback Machine
• Spinhenge@home screensaver video on YouTube
• arXiv:0801.2065 — Key research paper enabled by Spinhenge@home volunteers