Discover room-temperature magnetic semiconductor candidates with agents
We’re all used to two types of magnet. Task: The author and a team of AI agents searched for Luttinger-compensated (LC) magnetic semiconductors for spintronics. They used custom agents to run Density Functional Theory (DFT) simulations (PBE+U and HSE06) on candidate materials. The workflow involved designing a new compound (YBaMnFeO₅) and re-evaluating a known 1999 material (KV[Cr(CN)₆]) to predict band gaps and spin windows. The human author guided the search, interpreted the simulation outputs, and validated the physical properties against theoretical requirements.; tools: custom-built or unnamed agent. Human role: The original post does not say how a person stayed involved.; limitations: Simulations assume perfect, dry crystals; real-world samples like the 1999 KV[Cr(CN)₆] powder contain water, which may reduce spin sorting efficiency (disagreement between PBE+U and HSE06 methods). YBaMnFeO₅ requires precise atomic ordering (checkerboard pattern) that is difficult to achieve via standard synthesis (900–1300 °C). No experimental validation of the predicted spin windows has been completed yet..
Tools
custom-built or unnamed agent
Published on Agent Work Atlas:
Overview
We’re all used to two types of magnet.
Workflow
The author and a team of AI agents searched for Luttinger-compensated (LC) magnetic semiconductors for spintronics. They used custom agents to run Density Functional Theory (DFT) simulations (PBE+U and HSE06) on candidate materials. The workflow involved designing a new compound (YBaMnFeO₅) and re-evaluating a known 1999 material (KV[Cr(CN)₆]) to predict band gaps and spin windows. The human author guided the search, interpreted the simulation outputs, and validated the physical properties against theoretical requirements.
Reported result
The agents predicted YBaMnFeO₅ has a 2.35 eV band gap and spin windows of 1.0 eV (holes) and 1.4 eV (electrons), but noted synthesis difficulties due to atomic disorder at high temperatures. For KV[Cr(CN)₆], the agents identified it as an LC semiconductor with a ~2.1 eV band gap and spin windows of 2.6 eV (holes) and 1.6 eV (electrons). The 1999 experimental data showed magnetic order up to 376 K, supporting the prediction, though direct measurement of spin sorting remains pending.
Human role
The original post does not say how a person stayed involved.
Limitations
Simulations assume perfect, dry crystals; real-world samples like the 1999 KV[Cr(CN)₆] powder contain water, which may reduce spin sorting efficiency (disagreement between PBE+U and HSE06 methods). YBaMnFeO₅ requires precise atomic ordering (checkerboard pattern) that is difficult to achieve via standard synthesis (900–1300 °C). No experimental validation of the predicted spin windows has been completed yet.
Try it
Act as a computational materials scientist. I need to identify Luttinger-compensated magnetic semiconductors for spintronics applications. 1. Define the target properties: zero net spin moment, semiconductor band gap, and spin-sorted band edges (spin window > 26 meV). 2. Use a custom agent or DFT simulation tool to evaluate candidate materials. Input: [replaceable list of candidate chemical formulas or crystal structures]. 3. Run simulations using both PBE+U (fast) and HSE06 (accurate) functionals. 4. Analyze the output for: band gap size, spin window magnitude for holes and electrons, and magnetic ordering temperature. 5. Check for synthesis feasibility: simulate atomic disorder at typical synthesis temperatures (e.g., 900-1300°C) to see if the required atomic arrangement (e.g., checkerboard) persists. 6. Compare predicted properties against known experimental data if available (e.g., magnetic ordering temperature). 7. Output a report for each candidate: predicted band gap, spin windows, stability assessment, and synthesis challenges. Before running full simulations, perform a dry run on one known material (e.g., KV[Cr(CN)₆]) to verify the agent's ability to identify spin sorting. Confirm the results with me before proceeding to new candidates.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.